Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

The Well-Being of the Elderly in Asia

The Well-Being of the Elderly in Asia The well-being of the elderly in Asia : a four-country comparative study / Albert I. Hermalin, editor. Ann Arbor : University of Michigan Press, c2002. http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial http://www.hathitrust.org/access_use#cc-by-nc-3.0 This work is protected by copyright law (which includes certain exceptions to the rights of the copyright holder that users may make, such as fair use where applicable under U.S. law), but made available under a Creative Commons Attribution-NonCommercial license. You must attribute this work in the manner specified by the author or licensor (but not in any way that suggests that they endorse you or your use of the work). This work may be copied, distributed, displayed, and performed - and derivative works based upon it - but for non-commercial purposes only. For details, see the full license deed at http://creativecommons.org/licenses/by-nc/3.0/. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 The Well-Being of the Elderly in Asia ^e-. A Four-Country Comparative Study Albert I. Hermalin, Editor Ann Arbor THE UNIVERSITY OF MICHIGAN PRESS Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Copyright © by the University of Michigan 2002 All rights reserved Published in the United States of America by The University of Michigan Press Manufactured in the United States of America © Printed on acid-free paper 2005 2004 2003 2002 4321 No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, or otherwise, without the written permission of the publisher. A CIP catalog record for this book is available from the British Library. US CIP data applied for. ISBN 0-472-11280-5 Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Contents List of Figures ix List of Tables xi Authors KJX Preface xxi 1 Aging in Asia: Facing the Crossroads 1 Albert I. Hermalin andLora G Myers The Source and Nature of the Challenges 3 Demographic Responses and Their Limitations 6 Non-Demographic Responses 12 Goals of the Volume 15 Tools and Perspectives 17 History of the Project, Data Collected, and Other Products 18 Impact 21 2 The Demographic, Socioeconomic, and Cultural Context of the Four Study Countries 25 John Knodel, Mary Beth Ofstedal, and Albert I. Hermalin Population Aging and Sociodemographic Change in World Context 26 Demographic, Socioeconomic, and Cultural Context of the Four Study Countries 35 Political and Cultural Setting 36 Demographic Change 44 Socioeconomic Change 49 Changing Characteristics of the Elderly 52 Demographic and Social Characteristics of the Sample Elderly 54 Conclusions 59 3 Policies and Programs in Place and under Development 65 Mary Beth Ofstedal, Angelique Chan, Napaporn Chayovan, Yi-Li Chuang, Aurora Perez, Kalyani Mehta, and Albert I. Hermalin Social Security Retirement Programs 68 Health Care Programs 75 Other Programs and Policies Targeted to the Older Population 85 Conclusions 94 Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 vi Contents 4 Theoretical Perspectives, Measurement Issues, and Related Research 101 Albert I. Hermalin The Demography of Aging 101 Related Disciplines and Their Perspectives 104 The Conceptual Framework 107 Societal Changes and Their Impact on the Elderly 110 The Effect of Population Aging on Economic Development 114 Asian Interests in Research on Aging 119 Measuring and Understanding Living Arrangements 122 Identifying and Measuring Intergenerational Support Dimensions 123 Measuring Work and Leisure 128 The Economic Status of the Elderly 129 Health Status and Health Care Utilization 130 Vulnerable Groups and Their Levels of Disadvantage 131 Cross-Cutting Issues, Transitions, and Policy Recommendations 132 5 Patterns and Determinants of Living Arrangements 143 John Knodel and Mary Beth Ofstedal Measurement Issues 144 Trends in Coresidence 147 General Patterns of Household Composition 149 Coresidence in Relation to Children's Gender and Marital Status 154 Determinants of Living Arrangements 162 Multivariate Results 170 Conclusions 179 6 Intergenerational Support and Transfers 185 Ann Biddlecom, Napaporn Chayovan, and Mary Beth Ofstedal Measurement Issues 187 Background 190 Norms and Attitudes about Family Support 194 Availability of Children and Social Contact 196 Current Flows of Intergenerational Support 197 Conclusions 219 7 Work, Retirement, and Leisure 231 Albert I. Hermalin, Angelique Chan, Ann Biddlecom, and Mary Beth Ofstedal Trend and Levels of Labor Force Participation 237 Modeling the Retirement Process 247 Correlates of Labor Force Participation 251 Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Contents vii Attitudes and Preferences Related to Work and Retirement 265 Leisure Activities of the Elderly 275 Conclusions 283 8 Economic Well-Being: Insights from Multiple Measures of Income and Assets 295 Albert I. Hermalin, Ming-Cheng Chang, and Carol Roan Measurement Issues Related to Economic Well-Being 296 Levels of Income in Absolute and Relative Terms 301 Major Sources of Income 306 Factors Affecting Levels of Income 308 Household Income and Indirect Sources of Support 319 Subjective Measures of Economic Well-Being 324 Assets and Home Ownership 327 Changes in Economic Well-Being over Time 344 Conclusions 348 9 Physical and Mental Health of the Elderly 361 Zachary Zimmer, Josefina N. Natividad, Mary Beth Ofstedal, and Hui-Sheng Lin Methodological Issues 362 Health Risk Behaviors 366 Physical Health 371 Emotional and Cognitive Health 385 Self-Assessed Health 388 Conclusions 398 10 Patterns of Health Care Utilization 413 Mary Beth Ofstedal and Josefina N. Natividad Health Insurance Coverage 417 Health-Seeking Behaviors 425 Use of Specific Medical Services 429 Unmet Need for Health Services 450 Conclusions 455 11 The Vulnerable and Their Potential Disadvantages 461 Albert I. Hermalin, Mary Beth Ofstedal, and Kalyani Mehta Methods and Measurement Issues 462 Research Findings 469 Cross-Country Profiles of Relative Risks 486 Multivariate Results 492 Summary of Disadvantages Faced by the Sociodemographic Groups 499 Policy and Programmatic Implications 503 Conclusions 508 Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 viii Contents 12 Capturing Change: Transitions at Older Ages and Cohort Succession 519 Albert I. Hermalin Cohort Succession and Its Implications for Population Aging 521 Transitions over the Later Life Course 532 Conclusions 539 13 Making the Choices: Policies and Research for the Coming Years 543 Albert I. Hermalin Major Findings and Their Implications 544 Anticipating the Future: Demographic, Social, and Economic Trends and Their Interconnections 555 Policy Strategies in an Uncertain World 563 Developing a Research Agenda 574 Looking Backward and Looking Forward 577 Appendix A: Survey Design Information and Methodological Notes 583 Appendix B: Guide to Archived Data 590 Appendix C: Project-Related Publications 592 Index 599 Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Figures 1.1. Factors Affecting the Weil-Being of the Older Population, Policy Developments, and Their Interrelationships 7 4.1. Conceptual Framework of Factors Affecting Well-Being of the Elderly and Others, and Policy Formation 108 7.1. Labor Force Participation Rates for the Philippines, Thailand, Taiwan, and Singapore: 1970-1990 238 7.2. Percent of Older Males and Females in Labor Force, by Age and Country: 1980s and 1995-1996 242 8.1. Interrelationships among Various Dimensions of Economic Well-Being 300 8.2. Ratio of Mean Income in Sociodemographic Groups to Overall Mean Income, by Country 312 8.3. Sufficiency of Income for the Cohort 60 and Older, 1989 and 1996 346 8.4. Distribution of Sufficiency of Income in 1996 by Level of Sufficiency in 1989 347 10.1 Determinants of Health Services Utilization 415 11.1A. Philippines: Relative Risks of Economic Disadvantages for Sociodemographic Groups 471 11. IB. Philippines: Relative Risks of Health Disadvantages for Sociodemographic Groups 472 11.1C. Philippines: Relative Risks of Social Disadvantages for Sociodemographic Groups 473 11.2 A. Thailand: Relative Risks of Economic Disadvantages for Sociodemographic Groups 475 11.2B. Thailand: Relative Risks of Health Disadvantages for Sociodemographic Groups 477 11.2C. Thailand: Relative Risks of Social Disadvantages for Sociodemographic Groups 478 Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 x Figures 11.3 A. Taiwan: Relative Risks of Economic Disadvantages for Sociodemographic Groups 480 11. 3B. Taiwan: Relative Risks of Health Disadvantages for Sociodemographic Groups 481 11.3C. Taiwan: Relative Risks of Social Disadvantages for Sociodemographic Groups 482 11.4A. Singapore: Relative Risks of Economic Disadvantages for Sociodemographic Groups 483 11.4B. Singapore: Relative Risks of Health Disadvantages for Sociodemographic Groups 484 11.4C. Singapore: Relative Risks of Social Disadvantages for Sociodemographic Groups 485 11.5. Average Relative Risks Associated with Each Dimension of Disadvantage, by Sociodemographic Group and Country 489 12.1. Proportion of Adult Males with at Least a Secondary Education, by Birth Cohort and Country 525 12.2. Proportion of Adult Females with at Least a Secondary Education, by Birth Cohort and Country 526 Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Tables 2.1. Fertility, Mortality, and Percentage of Population 60 and Older in Selected Asian Countries, 1965 to 2030 27 2.2. Measures of Population Aging in the United States, Europe, East Asia, and Southeast Asia, 1975, 1995, 2030 30 2.3. Dependency Ratios in the United States, Europe, and Subregions of Asia, 1965, 1995, and 2030 34 2.4. Demographic Trends and Projections by Country: 1950-2030 46 2.5. Key Social and Economic Indicators, 1970-95 51 2.6. Demographic and Socioeconomic Indicators of the Population Aged 60 and Over, by Gender, 1970-90 53 2.7. Basic Characteristics of Survey Respondents by Gender: Philippines, Thailand, Taiwan, and Singapore 55 2.8. Basic Characteristics of Survey Respondents by Age: Philippines, Thailand, Taiwan, and Singapore 57 2.9. Percent of Respondents with Various Types of Living Kin 59 3.1. Social Security Retirement Programs 70 3.2. Health Care Programs 78 4.1. Outline of Intergenerational Support System 127 5.1. Trends in Household Composition among Those Age 60 or Over 148 5.2. Percent of Elderly Living with Specific Types of Household Members 151 5.3. Percent Distribution According to Household Composition 153 5.4. Percent Distribution by Number of Generations and Generational Composition 154 5.5. Coresidence with Sons and Daughters Conditioned on Availability 155 Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 xii Tables 5.6. Percent Coresident among Elderly Whose Children Are All Married, by Sex Composition of Children 157 5.7. Percent Distribution of Elderly Parents According to Location and Contact with Children, among Those with Any Living Children 160 5.8. Mean Number of Living Children by Sex and Marital Status of Children and Respondent's Place of Residence 161 5.9. Percent Living Alone (among unmarried) or with Spouse Only (among married), by Selected Sociodemographic Characteristics 165 5.10. Percent Living with Adult Children, Conditional on Availability of Child, by Marital Status of Child and Selected Sociodemographic Characteristics of Respondents 168 5.11. Odds-Ratios for the Effects of Sociodemographic and Health Characteristics on Living Alone or with Spouse Only: Results from Pooled Logistic Regression Models 172 5.12. Odds-Ratios for the Effects of Sociodemographic and Health Characteristics on Living with Unmarried and Married Children: Results from Pooled Logistic Regression Models 176 6.1. Percent Distribution of Weekly Visits between Parents and Children among Elderly Adults with at Least One Non- Coresident Child 198 6.2. Percent Distribution of Money and Material Transfers Involving Older Adults 199 6.3. Percent of Older Persons Receiving Money and Material Goods from Another Person, by Age, Gender, Marital Status, Place of Residence, Education, and Living Arrangements 200 6.4. Percent Receiving Money and Material Goods from Another Person, by Provider Type and Number 203 6.5. Percent of Older Persons Providing Money and Material Goods to Another Person, by Age, Gender, Marital Status, Place of Residence, Education, and Living Arrangements 206 6.6. Percent Providing Money and Material Goods to Another Person, by Recipient Type and Number 208 Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Tables xiii 6.7. Percent Distribution (of those with grandchildren) Providing Child Care for Grandchildren 209 6.8. Amount of Money Transfers Older Persons Received and Provided in the Year Prior to Survey Interview 211 6.9. Odds-Ratios for the Effects of Sociodemographic, Health, and Family Structure Factors on Directional Flow of Money Transfers 214 6.10. Odds-Ratios for the Effects of Sociodemographic, Health, and Family Structure Factors on Directional Flow of Material Transfers 218 7.1. Percentage Distribution of Type of Current or Past Employ ment (if not currently working), by Gender, Age and Place of Residence 234 7.2. Percent Currently in Labor Force, by Age and Gender 240 7.3. Current Economic Activity (in percent), by Gender, Age and Place of Residence 244 7.4. Job Characteristics by Gender, Country, and Work Status for Adults Age 60 Years and Older 248 7.5. Distributions (and mean values) of Independent Variables by Gender, Retirement Status, and Country 254 7.6. Odds-Ratios from Logistic Regression Coefficients for the Effects of Demographic, Socioeconomic, Health and Family Characteristics on Retirement, among Males Who Ever Worked 258 7.7. Odds-Ratios from Logistic Regression Coefficients for the Effects of Demographic, Socioeconomic, Health and Family Characteristics on Retirement, among Females Who Ever Worked 260 7.8. Odds-Ratios from Pooled Logistic Regressions for the Effects of Country and Demographic, Socioeconomic, Health and Family Characteristics on Retirement, by Urban-Rural and Gender 263 7.9. Predicted Probabilities of Retirement by Urban-Rural Residence, Gender, Pension Status and Country 265 Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 xiv Tables 7.10. Main Reason for Stopping Work, among Persons Who Ever Worked but Are Not Currently Working 266 7.11. Percent Distribution of Reasons Given for Retirement by Those Not Working, Mid-1980s vs. Mid-1990s 269 7.12. Percent Distribution of Opinions about Stopping Work in Old Age among Adults 64 Years and Older 272 7.13. Percentage of Adults Age 60 Years and Older Engaged in Leisure Activities by Work Status, Gender, Age, and Self- Reported Health Status 277 8.1. Mean, Median, and Percentage Distribution of Yearly Income by Economic Unit: Philippines, Thailand, Taiwan, and Singapore 302 8.2A. Mean and Median Income in Local Currency and $US and National Income Measures: Philippines, Thailand, Taiwan, and Singapore 305 8.2B. Income Indexes and Ratios: Philippines, Thailand, Taiwan, and Singapore 305 8.3. Major Source of Income: Percentage Reporting Each Source by Type of Economic Unit and Country 307 8.4. Ratio of Mean/Median Income in Sociodemographic Groups to Overall Mean/Median Income, by Country 310 8.5. Percent of Respondents in Bottom Income Tier by Sociodemographic Characteristics and Country 316 8.6. Odds-Ratios for the Effects of Sociodemographic Characteristics on Being in the Bottom Tier of the Income Distribution, by Country 317 8.7A. Percent Who Have Others Pay for All or Some of Their Household Expenses (indirect support) by Type of Economic Unit and Living Arrangement 321 8.7B. Ratios of Those in Lowest Income Tier Receiving Indirect Support to All Economic Units in Same Living Arrangement 321 8.8A. Data on Income of Households and Financial Arrangements of Elderly Units, by Country 323 Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Tables xv 8.8B. Income and Structure of Household for Those Living with Others 323 8.8C. Additional Details on Pooling Arrangements for Economic Units Living with Others 325 8.9A. Distribution of Sufficiency of Income: Philippines, Taiwan, and Singapore 327 8.9B. Percent Experiencing Some or Considerable Difficulty by Living Arrangement, Receipt of Indirect Support and Type of Economic Unit: Philippines, Taiwan, and Singapore 328 8.9C. Distribution of Satisfaction with Current Economic Situation by Income Position, Type of Economic Unit and Living Arrangement: Thailand and Taiwan 329 8.10. Percent of Elderly Units Owning Various Types of Assets, by Country 331 8.11. Percentage Distribution of Home Ownership of Elderly Residence by Sociodemographic Characteristic and Country 336 8.12. Median House Value for Respondents or Couples Who Own Their Home by Sociodemographic Characteristic and Country 342 8.13. Percentage Distribution of Major Sources of Income in 1996 by Major Source in 1989 345 App. Unweighted Sample Size and Percent Missing Income, by 8A. Sociodemographic Characteristics Used in Table 8.5 355 App. Alternate Specifications of Factors Affecting Being in the 8B. Bottom Tier of the Income Distribution. Panel A: Odds- Ratios for the Effects of Demographic, Socioeconomic Characteristics and Health Status. Panel B: Odds-Ratios for Reduced Models 357 9.1. Percent Currently Engaging in Health Risk Behaviors, by Selected Sociodemographic Characteristics 368 9.2. Percent Reporting Selected Chronic Conditions, by Age and Gender 373 9.3 A. Percent with One or More Life-threatening Condition, by Selected Sociodemographic Characteristics (age standardized) 380 Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 xvi Tables 9.3B. Percent with One or More Debilitating Condition, by Selected Sociodemographic Characteristics (age standardized) 381 9.4. Percent Reporting Difficulty Performing Selected Activities of Daily Living (ADL), by Age and Gender 382 9.5. Percent Reporting Difficulty Performing One or More ADL, by Selected Sociodemographic Characteristics (age standardized) 385 9.6. Ratio of Mean Scores on Indicators of Depression and Loneliness, by Selected Sociodemographic Characteristics 386 9.7. Ratio of Mean Cognition Scores, by Selected Sociodemo graphic Characteristics (age standardized) 388 9.8. Self-Assessed Health Status, by Age and Gender 390 9.9. Percent Reporting Low Assessments of Health, by Selected Sociodemographic Characteristics (age standardized) 392 9.10. Trends over Time in Self-Assessed Health 393 9.11. Odds-Ratios for the Effects of Sociodemographic, Social Support, and Health Characteristics on Self-Assessed Health: Results from Ordinal Logistic Regression Models 396 10.1. Percent for Whom Medical Care Costs in Past Year Were Covered by Health Insurance or Provided at No Charge, by Selected Sociodemographic Characteristics: Philippines 418 10.2. Percent Who Are Aware of and Have Used Public Health Insurance Programs, by Selected Sociodemographic Characteristics: Thailand 420 10.3. Percent Enrolled in Universal Health Insurance Program, by Selected Sociodemographic Characteristics: Taiwan 422 10.4. Percent Distribution for Main Source of Coverage for Health and Medical Care Expenses, by Selected Socio demographic Characteristics: Singapore 424 10.5. Percent Distribution for Usual Method of Treatment When 111, by Selected Sociodemographic Characteristics: Thailand 426 10.6. Percent Distribution for Usual Place/Method of Treatment When Iil, by Selected Sociodemographic Characteristics: Singapore 428 Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Tables xvii 10.7. Percent Using Specified Preventive Services during Past 6-12 Months, by Age and Gender 432 10.8. Percent Using Any Preventive Service during Prior 6-12 Months, by Selected Sociodemographic Characteristics 434 10.9. Percent Seeking Outpatient Treatment from Traditional and Western Medical Providers during Prior Month or Year, by Selected Sociodemographic Characteristics 436 10.10. Percent Hospitalized during Prior Year, by Selected Sociodemographic Characteristics 440 10.11. Odds-Ratios for the Effects of Predisposing, Need and Enabling Factors on Hospitalization: Results from Country-specific Logistic Regression Models 442 10.12. Odds-Ratios for the Effects of Predisposing, Need and Enabling Factors on Hospitalization: Results from Pooled Logistic Regression Models 446 10.13. Percent with Unmet Need for Medical Care, by Selected Sociodemographic Characteristics 451 10.14. Percent Reporting Specified Reasons for Unmet Need, among Those Reporting Unmet Need 453 10.15. Odds-Ratios for the Effects of Predisposing, Need and Enabling Factors on Unmet Need: Results from Country-specific Logistic Regression Models 454 11.1A. Vulnerable Sociodemographic Groups and Their Percentage of All Elderly, by Country 465 11. IB. Indicators of Economic, Health and Social Disadvantage and Their Percentage of All Elderly, by Country 466 11.2. Profile of Economic Disadvantages 468 11.3. Average Relative Risks for Each Sociodemographic Group, by Broad Dimension of Disadvantage and Country 487 11.4. Percent of Each Sociodemographic Group That Experiences Disadvantages in Two or More Dimensions, by Country 493 11.5. OLS Regression Coefficients for the Effect of Socio demographic Characteristics on Number of Economic Disadvantages, by Country 495 Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 xviii Tables 11.6. OLS Regression Coefficients for the Effect of Socio- demographic Characteristics on Number of Health Disadvantages, by Country 497 11.7. OLS Regression Coefficients for the Effect of Socio- demographic Characteristics on Number of Social Disadvantages, by Country 498 11.8. OLS Regression Coefficients for the Effect of Socio- demographic Characteristics on Number of Different Dimensions Disadvantaged, by Country 500 11.9. Summary of Each Sociodeinographic Group's Significant Disadvantages, by Broad Dimension and Country 502 11.10. Knowledge and Use of Select Specialized Programs for the Elderly, by Country 506 App Definitions for Indicators of Economic, Health 11A. and Social Disadvantages 512 12.1. Projected Number of Children Ever Born among Women 60+: Taiwan, 1985-2020 523 12.2. Projected Outcomes of the Likelihood of Completing Senior High School in Randomly Selected Pairs of Older and Younger Taiwanese Males, 1980-2000 527 12.3. Female Labor Force Participation Rates at Ages 45-49, circa 1970, 1980, and 1990: Select Asian and OECD Countries 529 12.4. Transitions in Living Arrangements over Three- to Four-Year Periods for Those 55 Years and Older at Baseline: Indonesia, Singapore, and Taiwan 534 12.5. Transitions in the Severity of Limitations between 1989 and 1996 536 12.6. Gross and Net Changes in Income Levels over Four-Year Periods in Singapore and Taiwan 537 Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Authors and Affiliations Ann Biddlecom abiddle@umich.edu Angelique Chan socchana@nus.edu.sg Ming-Cheng Chang jchang@bhp.doh.gov.tw Yi-Li Chuang yilic@bhp.doh.gov.tw Napaporn Chayovan cnapapor@chula.ac.th Albert I. Hermalin al berth @ umich.edu John Knodel jknodel@umich.edu Hui-Sheng Lin harveyl @bhp.doh.gov.tw Kalyani K. Mehta swkkkm@nus.edu. sg Lora Myers loram@umich.edu Josefina Natividad joyn@philonline.com Mary Beth Ofstedal mbo@umich.edu Aurora Perez aeptfg@compass.com.ph Carol Roan roansche we @ charter.net Zachary Zimmer zzimmer@popcouncil.org Research Investigator Population Studies Center University of Michigan Assistant Professor of Sociology National University of Singapore Deputy Director-General Bureau of Health Promotion, Department of Health Taiwan, ROC Section Chief Center for Population and Health Survey Research Bureau of Health Promotion, Department of Health Taiwan, ROC Associate Professor College of Population Studies Chulalongkorn University, Bangkok Emeritus Professor of Sociology and Research Scientist Population Studies Center University of Michigan Professor of Sociology and Research Scientist Population Studies Center University of Michigan Director Center for Population and Health Survey Research Bureau of Health Promotion, Department of Health Taiwan, ROC Associate Professor of Social Work Department of Social Work and Psychology National University of Singapore Administrative Associate Population Studies Center University of Michigan Professor of Sociology University of the Philippines Research Investigator Population Studies Center and Survey Research Center University of Michigan Professor (on leave), Population Institute, University of the Philippines Research Associate Population Studies Center University of Michigan (former affiliation) Research Associate Policy Research Division Population Council Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Preface The well known Chinese saying that the longest journey starts with a single step is an elegant, motivational adage, but it fails to convey that the full panoply of rea sons and preparations for that journey are often obscured in any retrospective reckoning. So it is with this work. The many steps by the many actors covering more than a decade of collaborative research and activity are numerous and com plex but can be enumerated in a reasonably linear fashion. It is more daunting to convey the original motivations for the journey and the groundwork that allowed for the human and financial resources to be mobilized relatively rapidly. A major impetus can be traced to my long involvement with the study of fertility and family planning in Asia, especially Taiwan. By the late 1980s, I had witnessed, over a period of 20 years, the rapid decline of fertility in much of East Asia and several countries of Southeast Asia and the growing success of their family planning programs. The speed of the decline exceeded the expectations that I and many other observers held in the early 1970s and led to reflections about whether the emerging low rates could be sustained into the future in societ ies that relied so heavily on children for old age support. Reflections on that rather specific relationship broadened quickly into wider concerns about the growing older populations in these countries and their potential sources of support in soci eties that were undergoing such rapid demographic, cultural, and socioeconomic change. It soon became apparent that there was a unique opportunity to study both these ongoing major transformations and their repercussions on several key soci etal institutions. This insight was aided by the ethos of social and economic demo graphic research at the Population Studies Center, set in motion by Ronald Freed- man, that emphasized the study of demographic phenomena in the context of broader family and social dynamics. My experience with several large fertility and family surveys in Asia also demonstrated the feasibility of effectively using this important mode of data collection and identified several organizations with outstanding capabilities within the region. The potential for mobilizing a strong collaborative team for a detailed ex amination of population aging also grew out of the research culture of the Popula tion Studies Center. Starting with Dr. Freedman's pioneering study in Taiwan in the early 1960s, it has been Center practice to work closely with strong organiza tions in developing countries over a long period, exchanging personnel and offer ing opportunities for advanced training to outstanding young staff members. As a Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 xxii Preface result it was relatively easy to identify a number of close collaborators in the U.S. and Asia, who had related interests in population aging and could devote time and some organizational resources to the emerging venture. Though there have been some inevitable changes in personnel over the 12-year period between the official start of the project and the completion of this volume, the team, as described below, has remained remarkably intact over the life of this study, attesting to the strong collegiality and mutuality of interest that has guided this project from the outset. Major credit for outfitting the expedition as it started its journey in 1989 must go to the National Institute on Aging for granting me a Merit Award (R37 AG07637) which provided stable, flexible, and long-term funding over the period 1989-2001.1 am very grateful to Richard Suzman, currently Associate Director of Behavioral and Social Research at the Institute, for encouraging me to undertake this research and for his long-term support and interest in our activities. I also appreciate the fine staff support at NIA over the years from Rose Li, Georgeanne Patmios, and David Reiter. Though the NIA grant covered the direct expenditures of the research, the training and core support grants to the Population Studies Center from the National Institute on Aging and the National Institute of Child Health and Human Development were also important resources. The training grants provided the funds for our pre- and postdoctoral trainees, many of whom became closely involved with the project, as detailed below. The core support grants con tributed significantly to maintaining the outstanding administrative, computing, library, and editing infrastructure at the Center, which is vital for the smooth con duct of any large-scale project. From the beginning, the project has aimed to understand the status of the elderly within each country and then conduct comparative analyses that would further elucidate the cultural and economic forces at work. A full discussion of our approach is given in Chapter 1. To achieve our goals required a diverse set of countries and collaborators who shared a long-term commitment to the research. The countries we chose were the Philippines, Taiwan, Thailand and Singapore. The three Southeast Asian countries had carried out important preliminary re search on the status of the elderly as part of an ASEAN project, and Taiwan was prepared to launch a large-scale survey on the same topic. The three U.S. investi gators associated with the project at the outset, John Casterline (at Brown Univer sity and later a visitor to the University of Michigan), John Knodel, and I, collec tively had research experience and close ties with this group of countries. The key organizations and personnel who initiated the project include the following: • University of Philippines Population Institute - Dr. Lita Domingo • The Taiwan Provincial Institute of Family Planning (now constituted as the Center for Population and Health Survey Research) - Dr. Ming-cheng Chang, Dr. Hui-sheng Lin, Yi-Li Chuang Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Preface xxiii • Department of Population Studies (now the College of Population Studies), Chulalongkorn University, Bangkok - Dr. Napaporn Chayovan, Dr. Werasit Sit tit rai and Chanpen Saengtienchai • Department of Statistics, Government of Singapore and National University of Singapore - Dr. Paul P. L. Cheung and Dr. Kalyani Mehta, respectively Though the organizations named above have remained the core collaborators over the life of the project, there have been inevitable shifts in personnel over time as well as some augmentation. A most sad event for all those engaged in the project was the death of Dr. Lita Domingo in 1996 from complications due to kidney disease. Lita brought great intelligence, energy, and warmth to all our delibera tions and her death was a profound professional and personal loss. Colleagues at the University of Philippines who worked closely with Lita on this project—Dr. Josefina Natividad, Dr. Aurora Perez, and Prof. Grace Cruz—have been most generous in their time and efforts in carrying on the work she so ably started. In 1994, Dr. John Casterline left the project due to the press of other com mitments but his efforts in helping shape the project in its earliest years have been very influential. Similarly, Dr. Paul Cheung found that his duties as Chief Statisti cian of Singapore were increasingly demanding and found it necessary to cut back after 1998. Fortunately Dr. Kalyani Mehta and Dr. Angelique Chan, both at the National University of Singapore, were able to take up many of the analyses and related tasks for Singapore. As noted, our affiliations were with the organizations named as well as with the individuals identified. In each case, the organization played a strong role in carrying out or facilitating a key survey or other data collection effort, and often were the hosts for the annual workshops which rotated across venues. These workshops were our major mechanism for planning each stage of the project, for reporting on research completed or underway, and for discussing some of the broader issues associated with population aging in Asia. Over time these workshops were attended by representatives of other countries in Asia who were interested in launching similar research (including China, Indonesia, South Korea, and Vietnam) and by U.S. researchers who were interested in the development of the project or who were invited to present recent developments of relevance to the project— such as the launching of the Health and Retirement Survey in the United States. I wish to thank the Directors of the organizations who assisted the project in one or more of these ways and also by facilitating the archiving of data collected in the course of this project (as described in Appendix B). In addition to those who played a dual role as director and investigator, like Dr. M. C. Chang, Dr. P. L. Cheung, and Dr. A. Perez, these include: Dr. Corazon Raymundo, Dr. Josefina Cabigon, Dr. Vipan Ruffolo and Dr. Kua Wongboonsin. I also appreciate the joint invitations from the Department of Sociology and the Center for Advanced Studies Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 xxiv Preface (CAS) of the National University of Singapore to spend time there in 1999 and 2000 to pursue research and writing relevant to this manuscript. Much of the material in Chapter 7 first appeared as a CAS working paper and we appreciate the permission to incorporate it here. Within each organization we also offer our thanks to the many staff members who played a valuable role in conducting the surveys within each country, and they are recognized to the extent known to us in Appendix A, which provides the details of the surveys within each country. At the Population Studies Center there are a large number of people who played a special role in the project and the production of the volume. Special thanks go to my co-investigator, John Knodel, who has been close to the project from its planning stage onward. John took the leadership in organizing and help ing carry out the innovative set of comparative focus group discussions in each country, in helping design and write up the results of the large representative sample survey conducted in Thailand, and he has contributed substantially to analyzing the results of our efforts through numerous papers and his role in several chapters. As noted at the outset, the project has been greatly enriched and aided by the large number of pre- and postdoctoral trainees at the Center who expressed interest in our research and chose to become affiliated with our work to some degree. Among the predoctoral trainees these include Megan Beckett, Jennifer Cornman, Xian Liu (as both a pre- and postdoc), Mary Beth Ofstedal, James Raymo, Michael Schoenbaum, Shiau-ping Shih and Zachary Zimmer. It is a source of great satisfaction to me as an advisor or mentor to these students, that five used data from the project for their dissertation and all have been involved in one or more papers stemming from the project. It is also rewarding to note that most of them have continued research careers closely involved with population aging. The long and special involvement of Mary Beth Ofstedal with this project and volume merits special mention and appreciation. Mary Beth was the Center's first predoctoral trainee in population aging per se. She arrived in 1987 just as plans were getting underway to mount the first intensive survey of aging in Tai wan. Her considerable skills in survey research were most valuable in helping fashion the complex questionnaire, and in assisting with the first round of analy sis. From 1991 to 1997 Mary Beth was a staff member at the National Center for Health Statistics, specializing in the Longitudinal Survey of Aging while complet ing her dissertation. In 1997 she returned to Michigan as an assistant research scientist at the Population Studies Center and the Survey Research Center. Since returning she has been closely involved with this project and in the preparation of this volume, writing or co-authoring several chapters, reviewing others, and play ing a very active role in all phases of planning and production. It is a source of great satisfaction that Mary Beth and Zachary Zimmer as co-principal investiga tors were awarded an R01 grant from the NIA in 2001 to continue aspects of this project in Asia, with a special focus on health transitions, and that they will be working with many of the same colleagues. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Preface xxv There have been eight post-doctoral fellows associated with the project: Emily Agree, Ann Biddlecom, Angelique Chan, Bruce Christensen, Ellen Kramarow, Rose Li, Xian Liu, Lindy Williams. Each has contributed one or more papers related to the project and has been actively involved with project develop ments through workshops or related efforts. Their substantial contributions are gratefully acknowledged. As with the predocs, many continue to work in this area and several have become more closely involved with the project: Rose Li has been a program officer at NIA and NICHD; Angelique Chan took up a post at the National University of Singapore and has become a key collaborator of the project there; and Ann Biddlecom, after several years with the Population Council, has returned to the Population Studies Center as Research Investigator, devoting part of her time to analyses from this project and contributing to this volume as author and in a number of other helpful ways. As noted at the outset, in addition to the direct financial support for project activities, the success of the project and the preparation of this volume are also due to a large extent to the excellent infrastructure at the Center. Lora Myers has been the key administrator from the outset of the project, handling the numerous reports and responsibilities associated with the grants, arranging the subcontracts and facilitating the workshops with utmost skill and effectiveness. Her rapport with all project collaborators contributed greatly to the esprit of the group and beyond her many administrative duties she was able to contribute to our scientific deliberations through her good judgement and ideas about aging issues. Ingrid Naaman as secretary to the project from its outset has also been invaluable to its smooth operation. She has effectively prepared the many proposals, reports, and papers emanating from the project and cheerfully assisted in hosting workshops and visitors. Her skill and patience were particularly manifest in the preparation of this volume, as she processed a high proportion of the material and learned new software to facilitate the conversion to camera ready copy. Several others at the Center also played important roles in preparing this volume. Fran Heitz was of great value in assisting with the analysis of several chapters and in producing standard formats for the many analytic tables. N. E. Barr, the Population Studies Center editor, played a key role at several stages. First she served as copy editor for the volume, contributing substantially to the clarity and readability of each chapter through her effective changes and sharp eye for ambiguity. Secondly, she took the lead in producing camera ready copy by establishing the protocols and actively processing each chapter to meet space and appearance criteria. Another major contributor to the volume was our librarian Yan Fu. Van's skill in hunting down often recondite Asian and other sources that reported on important policy and program developments in the region, as well as locating many other relevant references has greatly enriched the breadth of the volume and expedited its completion. Pearl Johnson, who assisted Yan in the li brary until 2001, was also most helpful on the numerous library matters arising. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 xxvi Preface I very much appreciate the encouragement of Ellen McCarthy, our editor at the University of Michigan Press, to undertake this volume and her advice at each of the critical stages. Kathy Osborn assisted ably in applying the software needed to convert the text toward a final format. Thanks are also due to Ronald Freedman, Linda Martin, Robert Schoeni, Peerasit Kamnuansilpa, and S. C. Wu for their helpful comments on specific points or earlier drafts of several chapters, and to an anonymous reviewer for several useful suggestions. The specific goals of the project and of this volume were described above and are treated at greater length in Chapter 1, which also reviews the structure and content of the succeeding chapters. It is worth noting here, however, the broader philosophy that guided our work. Although this is technically an edited volume, we have striven to achieve an integrated monograph. This was facilitated through our joint work at several workshops, in which chapter outlines and early drafts were reviewed and critiqued, and by my editorial efforts to achieve coherence across chapters. Accordingly one will find many cross-references within the chap ters as well as special chapters devoted to the underlying framework as well as to a synthesis of the results. Although this volume is informed by more than 10 years of research, the analyses presented were newly undertaken in order to build sys tematically on previous work and to achieve appropriate uniformity in structure and format. A conscious effort was made to make the volume as up to date and broadly relevant as possible. To this end, policy issues and research references through the end of 2001 and into early 2002 are reflected, and broader develop ments on aging in Asia and elsewhere are introduced. The project has benefited a great deal from having colleagues in each country with considerable research experience and with close knowledge of the culture and of ongoing developments with special relevance to population aging. Although many are in influential governmental or university posts and/or have served on important committees dealing with the topic, it must be stressed that the opinions expressed in the volume are those of the individual authors and do not represent official country positions. It is our hope, however, that the research and policy implications raised throughout the volume will contribute to ongoing discussions and to the development of effective policy, program, and research strategies. The content and style of presentation undertaken have been chosen to serve a number of potential audiences: specialists in Asia who wish to learn more about population aging in the region, demographers with growing interest in population aging as a specialized area of inquiry: and gerontologists who wish to pursue the developments in East and Southeast Asia. We hope our efforts will prompt contin ued attention to the rapid demographic and socioeconomic changes ongoing and their implications for the welfare of the older population in the region. These transformations are unique in the history of Asia and merit close scrutiny by the social science community in the years ahead. Albert I. Hermalin Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Chapter 1 AGING IN ASIA: FACING THE CROSSROADS Albert I. Hermalin and Lor a G. Myers In the broadest sense, this monograph is a study in social change and in anticipating future social change. It examines the effects on the current and future elderly of the rapid demographic transition that has occurred over much of Asia, accompanied in many countries by equally dramatic social and economic transformations. Policymakers in Asia have been aware of these trends and their potentially deleterious consequences for a decade or more, and have sought to fashion appropriate policies and programs that anticipate and mitigate their effects. Accordingly, any study of the sociodemographic trends and their impact must be examined in the light of policies put in place and under development. This convergence of demographic and socioeconomic changes in Asia has led to concern for the vitality of several key institutions, the strength of intergenerational relations, and the well-being of a large segment of the popu lation. For example, one strand of literature and popular opinion holds that these changes have weakened respect and support for elders, a viewpoint buttressed by reports in the mass media of isolated older individuals aban doned by their children and living in dire straits. A very different school of thought, which draws on limited data, holds that longstanding family and cultural arrangements in Asia will protect the elderly from the deleterious effects of these changes. Meanwhile, policymakers must simultaneously strive to maintain strong institutional arrangements and make realistic assessments of sociodemographic trends in order to effectively meet evolving needs. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 2 The Well-Being of the Elderly in Asia A second major theme that concerns both researchers and policy-mak ers is the potential impact of population aging on economic development. Will emerging older age structures adversely affect numbers of workers and their productivity and, if so, what policy levers are available to mitigate these effects? Will the costs of new pension and health programs associated with population aging detract significantly from improvements in consumption and savings levels? Governments are also concerned that the economic growth they need to develop more services and support for the elderly, as well as to maintain and expand other social programs, may not be forthcoming as a result of population aging. Our study is an empirical analysis of the effects of population aging to date and an attempt to discern some of the future patterns and issues. How well off are the current elderly and what factors affect their situation? To the extent that Asian societies are accommodating the rapid cultural, demographic, and socioeconomic changes, what forms do these adjustments take and how robust might they be under future pressures? How have policymakers been responding and what major challenges lie ahead? Current techniques of data collection and analysis allow a more complete answer to the first question and providing it will occupy much of this volume. At the same time by re viewing current policies and programs in relation to our analytic findings, we hope to shed some light on the other key issues as well. The analyses in pursuit of these goals relies to a great extent on detailed surveys and related studies carried out over a several-year period in four Asian countries—the Philippines, Taiwan, Thailand, and Singapore. Though no small group of countries can represent the diversity of Asia, these four focal countries exhibit considerable variety in terms of cultural heritage, ethnicities, and history, as well as in their levels of demographic and socioeconomic change. Each chapter contains analyses that compare and contrast these countries, with an eye to identifying commonalities and differences in their responses to the underlying forces. Though the focus on four countries sacrifices some of the generality to be gained from including a larger number, it facilitates these comparative analyses that would not otherwise be possible. As such this work should be seen as complementing other literature on the topic that includes more countries, but usually in the form of a country-by-country review without a common framework. (Phillips, 2000; Bengston et al., 2000; Phillips, 1992). To contextualize our detailed analyses of the four countries, at various points we relate our findings to wider trends in Asia and elsewhere. In this introductory chapter we review the nature of the challenges facing many Asian countries as a result of population aging and socioeco nomic trends and the potential levers available, as well as their limitations. We also provide a brief introduction to the surveys and material that form the Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Aging in Asia: Facing the Crossroads 3 core of our analyses, and a guide to the subsequent chapters that take up specific themes stemming from our broad goals. The Source and Nature of the Challenges Population aging in its most direct manifestation refers to alterations in the age structure of a population, which results in increasing proportions of the population at older ages (conventionally demarcated at age 60 or 65), and consequently, decreasing proportions at younger ages. It occurs first through declines in fertility levels: as adults have fewer children, the number at younger ages relative to those at older ages decreases. Once low levels of fertility are reached, declines in death rates contribute further to an older age structure. In the post-World War II period, East and Southeast Asia were the leaders among developing regions in the speed and magnitude of their fertility de cline. As described in greater detail in Chapter 2, fertility rates in a number of countries in these regions declined from an average of 5 to 6 children per woman in the 1960s to less than 2 children per woman by 1990. These changes have led to a rapid increase in the proportion of the population 60 or older and have set the stage for even faster growth of the older population in the future. The current older population is very much a "transitional" generation in terms of their demographic history and on many other dimensions. (Phillips [2000, p. 13] refers to the current older population as an "interim" generation in a somewhat different sense.) They raised their families during periods of much higher fertility and as infant and childhood mortality rates were falling. As a result they have larger numbers of living children on average than previous cohorts of elderly, and, given current fertility rates, have significantly more living children than will future cohorts of elderly. In societies in which older people rely to a large extent on children for economic, physical, and emotional support, a sharp reduction in number of children raises concern about the continued welfare of the elderly. Chapter 2 also traces the noteworthy improvements in life expectancy that occurred throughout East and Southeast Asia over the past 30 years. These improvements in mortality serve to enhance both the numbers at older ages and their average age, leading to a higher proportion at the oldest ages (those above ages 75 and 80). The differential in life expectancy between men and women means that women predominate at the older ages and particularly at the oldest ages, while the changing levels of mortality will have an impact on the proportions widowed at each age. These gender differences in the older population intersect with differences in the labor force histories of men and Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 4 The Well-Being of the Elderly in Asia women, as well as with differences in patterns of authority and inheritance arrangements, in affecting well-being, service demand, and policy needs. Equally or more challenging to the welfare of the elderly, in the eyes of many observers, are the rapid socioeconomic transformations that have oc curred over much of Asia during the last 40 years. Many countries in the region experienced major, sustained economic growth that led to substantial increases in per capita income, shifts from a reliance on agricultural produc tion to the manufacturing and service sectors, dramatic increases in the levels of school attendance and educational attainment, and greater participation of women in higher education and the labor force. (Chapter 2 also documents the magnitude and nature of many of these changes.) While these develop ments have brought considerable attention to the economic prowess of Asia, leading to such sobriquets as the "four Tigers" and moving a number of coun tries from the "developing" to the "newly industrialized" country category, they have also put pressure on a number of traditional social and cultural institutions. One outcome of these socioeconomic changes was another demographic transformation—a rapid urbanization in many countries of Asia, fueled by high rates of rural to urban migration. To the extent that these migrants were predominantly younger and better educated than their rural neighbors, these movements led to aging population structures in rural areas, a phenomenon sometimes referred to as "aging in place" (Morrison, 1990). These changes conjure up images of isolated rural elderly residing in economically stagnant agricultural areas. Countering this image however is the argument that older people prefer, and are better off, remaining in their familiar rural surroundings than following their children to the strange environs of the city. Consideration must also be given to the likelihood that decisions of younger family members to migrate to cities are not independent of the needs of older parents, and that often one child may stay behind to assist, as illustrated by a community study in Thailand (Knodel and Saengtienchai, 1996). The issue is far from resolved, and accordingly our analyses pay close attention to urban-rural differences on a number of dimensions of well-being. An extensive literature has theorized, on the basis of experience in the West, that industrialization is inevitably accompanied by a movement away from extended family organization to a nuclear living arrangement and greater independence among individuals in their social and economic lives (Goode, 1963). Although this thesis has been modified in several important respects (see discussion in Chapter 4), the underlying association between moderniza tion and changing forms of family organization remains influential. In the context of East and Southeast Asia, the demographic forces making for smaller numbers of children and increasing numbers of elderly, combined with the Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Aging in Asia: Facing the Crossroads 5 socioeconomic forces allowing children to gain more education and pursue new jobs in new locations, raise the specter that many older people will lose the physical, economic, and social support that was customary in earlier times. At the same time, insofar as children remain attentive to the needs of their parents, the improved economic conditions of the children can work to the parents' benefit. Close analyses of the extent and nature of parental support and their effects on economic well-being are needed to help resolve the issue. A fundamental question for Asia is thus anticipating the kinds of changes that will occur in terms of longstanding societal and familial norms that his torically have provided much of the support required by older persons. In this regard, it is important to note that much of the prior demographic and socio- economic change occurred while these societies were quite different in their family structures than the pre-industrial West. These changes have had sub stantial impact on several dimensions of family life and other societal norms while leaving other aspects relatively intact, as discussed further in Chapter 4. It remains to be seen how continued changes will affect the intergenerational support system and other dynamics of family life, and how variable these responses will be over time and across countries. Policymakers have been aware of these trends and cross-currents for some time, and have undertaken several measures to address them while pon dering a broader and more extensive set of programs. The 1982 United Na tions World Assembly on Aging (United Nations, 1982) was significant in raising awareness of emerging trends in population aging. It is possible to identify since then a sharp upswing, particularly in Asia, in the number of conferences, books and reports, and new investigations by demographers and other social and health scientists regarding various facets of this topic. In addressing these issues, policymakers in Asia are facing several di lemmas. Leaders are seeking approaches that will address the emerging needs of their growing older populations while preserving their societal structures. They also recognize the high costs and related problems associated with the social welfare programs for the older populations in the West. In a broad sense, policymakers in Asia are exploring whether economic "modernization" in evitably means "Westernization," in terms of social programs, or whether they can fashion programs at reasonable costs without undermining the traditional family arrangements that already provide substantial support. Several other features of the Asian landscape and the dynamics of population aging there heighten the implicit challenges. The speed of demographic transformations over much of East and Southeast Asia and the recency of economic successes mean that many of the countries in the region have experienced much faster growth in their older population than occurred among the currently industrialized countries, and some are doing so while at Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 6 The Well-Being of the Elderly in Asia an earlier stage of economic development. For example, it has taken (or will take) between 21 and 27 years for Thailand, Singapore, and China to increase the proportion of their population 65 years and older from 7 to 14 percent, while the same change took around 45 years in Spain and the United Kingdom and 68 years in the United States (Kinsella and Gist, 1995). Another concern, as noted, is that the very process of population aging can impede economic development by reducing the size of the population in the economically productive ages relative to those at the younger and older ends of the age continuum. (See Chapters 2 and 4 for further discussion of these trends and the relevant theory.) In addition this region, which over much of the past 30 years has been noteworthy for its economic successes and confidence in its ability to address social and economic problems, experienced a serious economic recession beginning in mid-1997. Although the recession has had different effects in different countries, it did undermine the ability of several countries to move ahead as scheduled with expanding existing social welfare programs or initiating new programs, and it has engendered caution throughout the region about the speed and magnitude with which new policies and programs designed to meet various social goals, including the needs of the older population, can be undertaken. (It should be noted that almost all our data collection and analyses precede the economic crisis, though some post- 1997 data are used in the concluding chapters when examining transitions.) Figure 1 illustrates the broad dynamics at work and the inherent chal lenges. (A more detailed version of this figure guides the theoretical discus sion in Chapter 4). Within each country long-standing societal and cultural arrangements guide the familial support and exchange institutions, which in turn largely determine the well-being of the elderly in terms of health, eco nomics, and work. These traditional arrangements are under pressure from the rapidly changing demographic and socioeconomic conditions, which also affect the resources and the range of policies and programs that policymakers can undertake in response. The bi-directional arrow between well-being of the elderly (and the underlying family and related structures) and the policies and programs suggests the dynamic interplay occurring over time. As policies and programs affect levels of well-being, the adequacy of the response, along with the political influence of the elderly and other forces, guide future pro gram development. Demographic Responses and Their Limitations To the extent that population aging, in its narrowest sense, is a demographic phenomenon, it is inevitable that consideration would be given to the possi bility of demographic responses that would alleviate the situation. The most Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 0. cg J -o M£u 2z g3 -aB ™•s f.tfl O.2 :§£i5.55 C£Rt3^^-Cau1 1cSobj-u u3 ^w u ^3O 2 ^*,leii2 "*a a £ £ oo _SI.I g S 8 o =5 £ K K £a3e0OlBuUvi So ^3>• AC trt H£ § |1 1 *U, U H Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 8 The Well-Being of the Elderly in Asia common types of responses include efforts to raise birth rates, encourage marriage, and influence immigration patterns. Some countries have experimented with policies aimed at raising birth rates to some degree as a way of slowing the aging of their populations. Singapore, prompted in large part by concerns about population aging and its consequences, formulated a new population policy in 1987 that promoted fertility. Using the slogan "Have Three, and More If You Can Afford It," Singapore offered several fertility incentives, including tax rebates upon the birth of a third child, subsidies for childcare center fees, priorities in housing allocation, and a special leave program for female civil servants with young children (Cheung, 1989). Associated with the emphasis on demographic responses is the ten dency to view population aging as a problem or crisis, whether or not the proposed solutions point to higher fertility or other responses. This is illus trated by titles such as Averting the Old Age Crisis (World Bank, 1994) or Gray Dawn: How the Coming Age Wave Will Transform America and the World (Peterson, 1997). As Hermalin (1995) and others have noted, this line of thought has several drawbacks. An older age structure should be viewed as a recent human triumph, reflecting on one hand couples' success in achieving desired small numbers of children, and on the other as gains toward the uni versal aspiration for longer and healthier lives. Moreover, the attempt to fash ion a demographic solution to the perceived problem assumes that a society's demographic behavior can be fine-tuned up or down—contrary to widespread evidence on the outcome of efforts to adjust fertility rates through exhortation or financial incentives (Uhlenberg, 1992; Hohn, 1987). It also overlooks the potential social and economic dislocations caused by frequently shifting fer tility rates, as societies, for example, cope first with the costs of expanding schooling and training capacities after an upturn in fertility, and then cope with the costs of excess capacity after a downturn. Nonetheless, a number of countries in Europe and elsewhere have pro grams of child allowances, and other financial and non-financial inducements, to maintain or enhance fertility levels. Hohn (1987) presents an overview of the plans in several European countries, along with estimates of their efficacy and suggestions for future initiatives that are responsive to a wider array of social and economic factors associated with parenthood. (For a historical com parative analysis of family policies in industrialized countries see Gauthier, 1996.) New approaches can be expected from time to time. Singapore, for example, introduced a new bonus plan for births starting in April 2001. Under this plan the government will contribute a specified amount for second and third births into a special Child Development Account for five years and will also match parents' contributions into that account up to a stipulated limit. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Aging in Asia: Facing the Crossroads 9 The family can spend the funds in the account for education and developmen tal expenses like childcare and nursery fees (Ministry of Community Devel opment and Sports, 2001). In addition, programs that expand the availability of childcare have been adopted in Japan and other countries, both to encour age childbearing as well as female labor force participation. It is not expected that these plans will radically alter the levels of fertility and hence future age structures, but rather that they will slow down the rapid rate of population aging by modestly increasing the average number of births from the very low levels now prevailing. In a number of countries in Asia and Europe, total fertility rates per woman are in the neighborhood of 1.5 births, so that even a 20 percent increase would leave fertility levels below the population replace ment level of 2.1 children. A prime contributor to the low fertility rates in many countries has been a decline in the proportion of young adults who marry, and the later ages at marriage for those who do. This change is particularly notable in several Asian countries where universal marriage was the norm until recently. For example, in Taiwan 19 percent of women aged 35-39 were unmarried as of 1999, compared to 13 percent in 1990 and 7 percent in 1980. (Chapter 2 shows how the average age at marriage has advanced in the four countries under study.) In Japan, the trends are similar but at a somewhat different level, with 10 percent of women single at ages 35-39 in 1995, double the proportion 20 years earlier (Retherford et al., 2001). These trends reveal that greater proportions of the population are delaying and possibly foregoing marriage, and they have implications for the family arrangements and needs of these cohorts as they age (Hermalin, 1993). Lower marriage rates are a response to a number of ongoing demo graphic, socioeconomic, and cultural factors. For example, increasing educa tional and career opportunities for women open up alternatives to marriage and also increase the difficulty of finding suitable partners. And the decline of arranged marriages and the greater acceptance of premarital sex have helped reduce some of the pressure to marry. (For Japan, see the discussion by Retherford et al., 2001; Raymo, 1998.) Though policies directed toward enhancing marriage rates appear less common than those directed toward births within marriage, several instances can be cited. Singapore, starting in 1984, has promoted interactions between young men and women through various organized activities as well as initiat ing a computerized match-making service for university graduates and oth ers. And the initiatives in other countries that reduce the opportunity costs of motherhood, such as enhanced maternity leave programs and subsidized childcare, also serve to make the prospect of marriage and parenthood more attractive. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 10 The Well-Being of the Elderly in Asia Demographers and others studying marriage trends often employ the analogy of the market, where the number of eligible brides and grooms at each age, and at different levels of education and other characteristics, help determine the numbers of each gender who marry at each age. One result, for example, of Singapore's decreased fertility rates and rapid rise in the educa tional level of women over the past several decades, is an insufficient number of eligible women in the appropriate age range for men in their late 20s with modest education. This has led many of these men to secure brides from other countries, to the extent that the proportion of marriages between Singaporean citizens and foreigners has increased to one quarter in recent years. This influx of foreigners as a result of marriage raises the more general question of immigration as a demographic response by countries undergoing rapid population aging. Technically speaking, policies focusing on migration are not demographic policies if they do not have as their goal a change in population size or population age structure, although policies designed to re cruit labor or reunite families may have direct demographic responses (H6hn, 1987). From this narrower demographic perspective, the United Nations (2001) has calculated the rate of immigration needed under various assumptions for countries with older age structures to achieve demographic goals in terms of numbers or proportions of people in various age categories. Lesthaeghe (2000) presents a detailed analysis of replacement migration for Europe in relation to other demographic and social factors. Most migration policy in recent years has been more narrowly focused on recruiting workers to supplement the existing labor force and/or to fill certain occupational niches. As such migration is a policy for addressing the potentially deleterious economic consequences of population aging. An analy sis by McDonald and Kippen (2001) reviews the labor supply prospects of 16 more developed countries in terms of the potential of immigration as well as other steps. But migration policy is often a complex and contentious issue within countries, intersecting with long-standing traditions, cultural values, and the competing interests of various ethnic and socioeconomic classes. International migration on the world scene has taken a number of different forms over the last 500 years, and has varied considerably in intensity and underlying dynamics over this period (Massey et al., 1998). The migration patterns that emerged in the 1960s were quite different from those of the past in the variety of receiving and sending countries, and in the shift of the supply of migrants from Europe to the developing world. Much of the driving force for migration over the last 40 years has been labor scarcity in countries with capital-intensive industries and rapid economic development, the availability of labor in densely settled countries in earlier stages of industrialization, and the wage differentials across these countries. In the 1960s, the countries of Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Aging in Asia: Facing the Crossroads 11 Western Europe started to receive so-called "guest workers" from Southern Europe and elsewhere; by the 1970s several Southern European countries were importing migrant workers from the Middle East and North Africa, and the oil-rich nations in the Gulf region also sought labor migrants from Asia as well as the Middle East (Massey et al., 1998, Chapters 1 and 5). By the late 1980s, Japan as well as several of the rapidly industrialized countries in the region were receiving international migrants, mostly from other countries in the region. Japan, as of 1990, had slightly more than a million registered foreigners, mostly from Korea, but during the 1980-90 de cade most of the increase among East Asians was due to migrants from China (Massey et al., 1998, Table 6.1 and page 163). Okunishi (1995) estimates that as of 1993 there were 950,000 foreign workers in Japan (including 300,000 illegal unskilled workers) but representing only about 1.5 percent of the total labor force, and McDonald and Kippen (2001) do not expect that Japan can absorb a large number of immigrants in the future. Atoh (2000), while noting that the net inflow of foreigners increased somewhat in the latter part of the 1990s and that demand for foreign labor in some industries is likely to persist, concurs that there is little likelihood of a major change in the level of immi gration in Japan in the coming years. The four focal countries of our study represent a spectrum of migra tion policies and behaviors: Singapore and Taiwan receive a considerable number of labor migrants; Thailand both receives migrants and sends its own labor to other countries; while the Philippines is a country with a relatively large outflow of their own workers. It is of interest that a substantial number of migrants from Thailand and the Philippines work in Singapore and Tai wan, enhancing the value of treating the four countries from a close compara tive perspective. Singapore has developed a pro-immigration policy, welcoming profes sionals (those with stipulated levels of education, earnings, and experience) to become permanent residents and at the same time admitting a large number under a guest worker policy. As of 2000, these two groups constituted 26 percent of the country's population. Many of the workers in Singapore are from the Philippines, predominantly women employed as domestic workers, and Thailand also supplies a substantial number of laborers. As of 2000, Taiwan had over 300,000 foreign workers, about double the number in 1994, representing about 3 percent of the Taiwan labor force. The Philippines and Thailand together supply approximately 80 percent of the total. Although most foreign workers are employed in manufacturing and other traditional industries, increasing attention has been focused on labor shortages within Taiwan's large high-tech sector and the possible need to re cruit more highly trained workers. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 12 The Well-Being of the Elderly in Asia Thailand is both a sending and receiving nation with regard to immi gration. As noted, substantial numbers of Thais are working in Singapore and Taiwan, and in total the number of workers legally registered to work abroad was 184,000 as of 1997, with the number of illegal emigres probably exceed ing this number (Sobieszczyk, 2000). At the same time there are estimated to be over one million foreign workers in Thailand—200,000 professional ex patriates and over 900,000 unskilled farm and construction workers from the poorer neighboring countries (Martin et al., 1995; Stern, 1998). The Philippines is a major labor exporter—in fact, according to Martin et al. (1995) it is second only to Mexico in this respect. Battistella (1995) reports government estimates of 4.2 million workers abroad in the early 1990s, including 2.4 million documented workers and 1.8 million undocumented, but more conservative estimates place the number of documented workers at 1.6 million. The remittances from the overseas workers are significant inputs into the families and communities as well as the economy as a whole, and labor emigration policy has become an explicit factor in economic develop ment planning. Immigration policy impinges directly on the size of the labor force and economic development, so it is likely to receive continuing attention from governments. McDonald and Kippen (2001) project that sustaining current immigration levels in Singapore would serve to maintain the size of their labor force through 2050. Immigration also has a direct effect on age struc ture, but in most cases this effect is likely to be secondary and only marginally alter the overall trend toward an older population. Accordingly, nothing sug gests that the rapid levels of population aging projected by the United Nations and others for many countries of the world over the next 30 to 50 years will be substantially changed. Singapore, for example, despite policies promoting births and migration is still projected to have 30 percent of its population over age 60 by 2030, if the estimated levels of future fertility, mortality, and migra tion are borne out. (See Chapter 2 for more details on these and related data.) Non-Demographic Responses Because of the limited potential impact of demographic responses in slowing population aging, policy attention is being devoted to programs that will protect and enhance the well-being of the elderly and will mitigate the potentially deleterious effects of population aging on economic development. Policies addressed to the first issue involve programs that provide direct assistance to the elderly in areas of perceived need as well as initiatives that will sustain family arrangements for the support of older parents. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Aging in Asia: Facing the Crossroads 13 Chapter 3 provides details on the large-scale social security and health insurance programs in place and under development in the four study coun tries. All four countries are currently reviewing and amending these programs in the face of their changing demographic profiles and the observed shortfalls in existing programs. Many of these programs initially served special groups such as civil servants or employees of large firms; others, with more general intent, have proved to be inadequate due to problems of enforcement or the presence of provisions that have diminished their effectiveness for current needs. As described in Chapter 3, Taiwan launched a new national health in surance program in 1995, referred to as Universal Health Insurance, replac ing a number of partial systems tied to classes of employees, and is set to launch a new national pension program in the near future. Thailand is phasing in a new national pension program, delayed as to starting date and scope by the economic crisis which affected that country strongly in 1997. Singapore has altered provisions of its Central Provident Fund to enhance its effective ness as a source of retirement income, and has added health insurance fea tures for older ages. In the Philippines a new National Health Insurance pro gram was introduced in 1995. In addition to these broad pension and health care programs, many of the countries in Asia, including the four study countries, are expanding the range and availability of governmental services for the elderly. As described in Chapter 3, these include provisions for long-term care, community ser vices such as daycare, senior centers, and services within the home, programs to enhance cultural and recreational activities, and attention to those with special economic problems. In fashioning new programs and amending existing ones, policymakers are constrained not only by the costs involved, especially in periods of uncer tain economic growth, but by the desire to insure that these programs do not undermine existing social arrangements that provide a substantial share of the economic, physical, and emotional support to the elderly. The aforemen tioned community service enhancements, for example, are intended to help older people remain with their families while they serve their perceived needs. In addition explicit steps to sustain family support are underway in several countries, ranging from laws requiring children to care for parents (as de scribed for Singapore in Chapter 3) to incentive programs that promote shared living arrangements or related support. How well such efforts to sustain exist ing family arrangements succeed in the face of ongoing demographic and socioeconomic trends and the advent of broader pension and health programs is a critical question for the coming years and one meriting major research attention. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 14 The Well-Being of the Elderly in Asia The second major concern for policymakers concerns the effect of popu lation aging on economic growth. The policy levers here center on steps that will enhance labor force participation and productivity. These include poli cies that will increase the labor force participation of women by upgrading their skill levels through education and life-time learning programs. Also im portant are policies that address retirement age issues and the structure of pension program—insuring that they maximize savings and do not contain incentives for early retirement (World Bank, 1994; Gruber and Wise, 1999). Several features of recent demographic and socioeconomic history pose a number of more specific challenges. Past trends in birth rates and life ex pectancy mean the numbers of older people, especially those at advanced ages, will rise sharply in the coming years. This will create additional de mands for long-term care, geriatric facilities, personnel, and expertise in the health infrastructure—all while maintaining existing capability in the areas of infectious diseases and the needs of mothers and children. Past and con tinuing improvements in life expectancy at older ages also mean that growing numbers of elderly will encounter a series of major transitions over their later years in terms of work, health, wealth, and living arrangements. Government policies and programs can anticipate some of these and seek to enhance well- being at each stage, but often difficult choices will be involved. In the realm of work and retirement, for example, current statutory retirement ages among civil servants and other covered employees are relatively young in much of Asia, and governments may be tempted to move these back to reduce the social welfare costs and maintain productivity. But in the short run, this may run counter to the need to create more jobs for the rapid rise in the young adult population, reflecting high past fertility and the changing skill require ments of many jobs. Dramatic increases in social and economic development over the last quarter century in much of East and Southeast Asia also pose several dilem mas. As a result of these changes, the younger generation is much more edu cated, urban, and professional than their elders, attributes that can benefit society even while they present potential problems. The sharp differences across age groups is not only a potential source of intergenerational tension and conflict, but it means that the future elderly will have sharply different characteristics than the current elderly. They are likely to demand different, more sophisticated, and costly services, as well as bring different resources to providing for their own needs. And they are likely to have expectations that are different from those of their parents. As a result, policymakers need to fashion programs to assist the current elderly as appropriate, but avoid getting locked into policies and approaches that might be rendered ill-suited in the Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Aging in Asia: Facing the Crossroads 15 coming years given the changing socioeconomic characteristics of the nature elderly. In some countries, policymakers will also be challenged to ensure adequate implementation and enforcement of new or amended programs. Countries with large sectors of the economy devoted to agricultural, small businesses, and informal labor may find it difficult to set and enforce contributions to social programs from owners, employers, and employees, particularly where past standards have been lax. In such cases projected revenues may not match actual benefits, jeopardizing the system and causing financial strain for the country. Both the changing conditions under which programs will operate and the newness of the programs themselves suggest that monitoring program effectiveness and making appropriate adjustments will be regularly required in the coming years. As addressed further in Chapter 13, developing an effec tive dialogue between policies and research will be important to anticipate problems and develop guidelines for future actions. Goals of the Volume The strength and recency of the demographic and socioeconomic forces at work make it difficult to distinguish short-term adjustments from possible long-term transformations, and we make no claim to having a crystal ball. At the same time, policymakers are looking to research to help guide future actions. Sufficient data are now available to allow us to take stock of the current trends and to draw out the key issues for future policy and research consideration. Accordingly, our first goal is to set forth a clear picture of the status of the elderly in the four Asian countries we have studied intensively: the Philippines, Thailand, Taiwan, and Singapore. More specifically, we will examine the well-being of the elderly in terms of their physical, mental, and emotional health; their economic sufficiency as reflected in income, assets, and living conditions; and their patterns of work, retirement, and leisure. We will describe both the current distributions and the recent changes on a number of measures by comparing our results to earlier survey data. A careful delineation of the actual conditions facing the elderly is important to counter less informed opinion that tends to focus on two extremes: that ongoing changes have had severe deleterious effects on the older population, or that pervasive family support arrangements have successfully buffered the elderly from adverse consequences. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 16 The Well-Being of the Elderly in Asia A second goal is to understand the factors that influence the well-being of the elderly in these four countries. As suggested by Figure 1, these factors include not only the personal characteristics of the elderly but their living arrangements and support systems, and the policies and programs in place in the governmental and private sectors. In carrying out these goals, each analytic chapter will focus on com parative analyses across the four countries to assess the importance of cul tural variation as well as other conditions on the differences in outcomes. In so doing we will be able to determine the extent to which a single Asian pattern is evident versus multiple context-specific patterns. The analyses will include both detailed descriptive patterns and multivariate techniques designed to identify the major forces at work. To better identify country-specific ten dencies, the multivariate analyses will be carried out separately for each country and across countries by pooling the data. An important unknown in anticipating the future well-being of the elderly and the programs needed is the degree to which current societal systems may shift in the coming years. Accordingly, a third goal will be to assess the strength of key societal elements that have supported the elderly in the past, to discern the level of recent trends, and to anticipate future changes. The detailed analyses in pursuit of these goals are presented in Chap ters 5-11. Chapters 5 and 6 take up the third goal first, describing current patterns of living arrangements and support levels, and contrasting these with those observed some years earlier. As stated above, and depicted in Figure 1, these basic institutional and familial arrangements are key determinants of the well-being of older persons on several dimensions. In these chapters we also analyze the characteristics of the elderly and their children that affect these arrangements. Though cross-sectional surveys are limited in their abil ity to trace change, knowledge of current interrelationships combined with the ability to anticipate the social and demographic characteristics of future elderly can provide useful insights on the patterns likely to emerge. Chapter 7 initiates the analysis of the dimensions of well-being, investigating both the level of labor force activity of older persons and some of their leisure activities. For those still working we describe their main occupation, whether they are subject to mandatory retirement practices, and whether they have pension coverage. For retirees, we investigate the nature of their previous work, their retirement process, and the reasons for retirement. Chapter 8 focuses on the economic well-being of older individuals and couples, employing multiple measures of income and relative income posi tion, taking into account individual versus household measures and indirect sources of support. Data on asset holdings are also employed to present a fuller picture of overall economic levels. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Aging in Asia: Facing the Crossroads 17 Chapters 9 and 10 turn to health-related measures of the older popula tion. Chapter 9 examines the health status of the elderly, through their reports of select chronic diseases, levels of difficulty in eating, dressing, and other activities of daily living, measures of depression and cognition, and levels of self-reported health. Several health behaviors are also reported. Chapter 10 focuses on the levels of health care utilization and perceived needs by exam ining the proportions using various services, the costs involved, and the ex tent to which those with stated problems were limited in obtaining care. Chapter 11 reverses the analytic strategy by identifying the sociodemographic groups often thought to be particularly vulnerable to the emerging trends of population aging in these countries and examining the degree to which they are relatively disadvantaged compared to other older persons on a variety of well-being measures. Chapter 12 looks to the future using cohort analysis to illustrate how the elderly in the coming years will differ from the current elderly and the policy considerations that arise as a result. It also looks at changes within cohorts, using transition rates to exam ine shifts in economic status and living arrangements as older people age. The final broad goal of the volume is to interpret the policy and pro gram challenges that emerge from the trends and interrelationships observed in our analyses. Chapter 3 describes in some detail the current policies within each country. In a number of the chapters, the degree of involvement of the current older population with various programs is described and the effect of that involvement on their well-being is analyzed. Each analytic chapter touches on the policy implications of the findings and the concluding chapter brings together information on this topic—tracing the meaning of our results for policy and program development. The goal is to identify those areas needing more urgent attention and some of the policy levers that appear most promis ing, as well as those areas in which governments may have more time to develop policy. Chapter 13 also provides a summary of key findings and dis cusses promising lines of future research and ways of strengthening ties be tween policy and research. Tools and Perspectives Gerontology is a large and diffuse field with theoretical, substantive, and methodological contributions from many disciplines. The analyses presented here are from the perspective of the demography of aging, itself a wide-rang ing and eclectic field, informed by the theory and methods of demography, by the set of problems addressed, and by the methods and data it shares with other fields (Martin and Preston, 1994). In the analytic chapters that follow, Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 18 The Well-Being of the Elderly in Asia this perspective is manifest in attention to the underlying quality of the data, a focus on the constituent elements of complex constructs, and an attempt to detect the influence of possible causal factors by controlling for other forces through standardization or statistical multivariate techniques. The breadth of gerontological interest precludes a single encompass ing theoretical basis for these analyses. (Even individual disciplines with gerontological concerns maintain a number of distinct theories and hypoth eses.) Our strategy, therefore, as developed at some length in Chapter 4, is to provide a broad heuristic framework (expanded from Figure 1) within which a number of specific theories and hypotheses can be located, and which can serve effectively as a guide to the specific analyses of well-being forming the core of this work. In carrying out these analyses the goal is to provide the best estimates possible of current well-being in each dimension, to identify the factors associated with these levels, and to draw careful comparisons across countries as to levels and dynamics wherever possible. As to conventions, we use the word "elderly" as synonymous with "older person" for stylistic variety. In some quarters, "elderly" has a connotation of someone frail or dependent. As our research makes clear, the "elderly" in our sense of the word are often independent and providing goods and services to others. In presenting data for the four focal countries, we have decided to sequence them as the Philippines, Thailand, Taiwan, and Singapore rather than alphabetically. This order has the virtue of pairing the Philippines and Thailand, which are the most rural of the four countries and at the lower end of the per capita income scale, and Taiwan and Singapore, which are pre dominantly Chinese and at the upper end of the income scale. History of the Project, Data Collected, and Other Products Although the analyses presented in the following chapters were newly under taken for this volume, from another perspective this work attempts to distill ten years of observation by a diverse team of researchers in the United States and several Asian nations. The very broad research framework described above has led us to employ a variety of data sources, including censuses, surveys, and ethnographic and other qualitative data sources. The following is a brief review of these sources, how they were obtained, and where future investiga tors can find and use them. Early Data At the outset of the study in 1989, we took stock of existing data relevant to the social and economic status of the elderly in the Philippines, Thailand, Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Aging in Asia: Facing the Crossroads 19 Taiwan, and Singapore. These included censuses, other official statistics, and a set of surveys sponsored by the Association of Southeast Asian Nations (ASEAN) in three of the countries (see Chen and Jones, 1989). In 1989 we also initiated a new panel study in Taiwan, the Survey of Health and Living Status of the Elderly in Taiwan (Taiwan Provincial Institute of Family Plan ning, et al., 1989), which demonstrated the practical need for data on mea sures of health, economic status, and social support. The early sources proved valuable for a variety of analyses. Census data were employed to analyze labor force trends and to compare current and future cohorts on educational attainment and other traits (Raymo and Cornman, 1997; Hermalin and Christenson, 1992). Casterline, et al. (1991) used ASEAN data and the 1989 survey in Taiwan to compare living arrangements of the elderly in the four countries. They also set the stage for a new round of surveys that would expand the scope of coverage and enhance comparability across countries. Data Collected 1990-1999 In 1990 the project began data collection efforts that continued through the decade. In 1991-1992 a series of focus group sessions were held in each of the four study countries with both elderly participants and adults with elderly parents. The goal was to learn about current attitudes toward the elderly, their role in the family as givers and receivers of support, and how these roles might be changing over time. Specific topics included living arrangements, support exchanges, social contacts, work and economic status, health status, utilization of health programs, and role in the community. These efforts repre sented a pioneering use of focus groups in comparative social science re search and resulted in detailed analyses of cross-generational attitudes and expectations concerning living and support arrangements for the elderly (see the special issue of Journal of Cross Cultural Gerontology, Vol. 10:1-2,1995.) During 1993-1996, new surveys of the elderly were conducted in the Philippines, Thailand, Taiwan, and Singapore. Each of the surveys is nationally representative, with more extensive subject matter and broader age coverage than the ASEAN surveys and other early efforts. They incorporate lessons from the successful 1989 Taiwan survey, from the focus groups, and from the U.S. Health and Retirement Survey (HRS), which was launched in 1992. A core questionnaire was adopted with modification by each country. Because of multiple auspices and country-specific concerns, the exact timing and content of these surveys vary across the countries. Issues of comparability are addressed at many points in the chapters that follow. The surveys are listed below; for additional detail, see Appendix A. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 20 The Well-Being of the Elderly in Asia The surveys conducted in collaboration with the project and which form the major data base for the analyses which follow are listed below. Some of the surveys covered those aged 50 or 55 and older but only the interviews with those 60 and older are analyzed here. Additional details about their de sign and execution are presented in Appendix A: 1989 Survey of Health and Living Status of the Elderly in Taiwan (AM.049) 1993 Follow-up Survey of the Health and Living Status of the Elderly in Taiwan (#=3,155) 1995 National Survey of Senior Citizens in Singapore (#=4,750) 1995 Survey of the Welfare of the Elderly in Thailand (#=7,708) 1996 Survey of the Health and Living Status of the Middle Aged (#=2,462) and Elderly (#=2,669) in Taiwan 1996 Philippine Survey of the Elderly (#=2,285) Detailed analyses of these surveys form the core of this monograph. How ever, we also make select use of more recent panel data collected in 1999 when Taiwan conducted another round of interviews with the elderly and near-elderly, and Singapore reinterviewed those first studied in 1995. Archived Data and Publications As much as possible, the data and publications that form the background for this monograph have been made available for general use. A focus group data archive is based at the Population Studies Center, University of Michigan. The full set of transcripts, both in original language and English translation is available by request. In addition many of the data files from the surveys also are available for further analyses. See Appendix B for additional information about the content of and access to these data archives. A comprehensive bibliography for the project is given in Appendix C. "Research Reports on the Elderly in Asia" was established at the University of Michigan Population Studies Center (PSC) in 1990 as a working paper series for the project. As of mid-2001, the series includes more than 50 items representing a range of topics and analytic approaches. Many of these Re search Reports, as well as other papers originating from the project, have appeared as articles in peer-reviewed journals. In addition, overseas colleagues have also produced a large number of reports and publications for in-country circulation. The appendix presents a list of the Research Reports and describes a website for identifying the papers, dissertations, and other publications ema nating from this project. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Aging in Asia: Facing the Crossroads 21 Impact Going beyond the products we can claim as our own, we believe that the project has had an important impact on the field. Demographers have been recruited to study the elderly in the four countries and the United States Our colleagues abroad have presented their work to a wide range of public, scien tific, and government assemblies, and several are now involved with influen tial government committees on the topic of aging. In addition, a number of investigators initiating research on population aging in other countries have drawn on our questionnaires as starting points. These include Bangladesh, the People's Republic of China, Sri Lanka, South Korea, Malaysia, South Africa, Egypt, and the multi-country Pan-American Health Organization spon sored SABE project. The PSC Research Reports numbered 1,49 and 50 pro vide complete questionnaires for the Taiwan surveys in 1989 and 1996. These items are among the most frequently requested in the series. REFERENCES Atoh, Makoto. 2000. "The Coming of a Hyper-Aged and Depopulating Society and Population Policies: The Case of Japan." Paper presented at Expert Group Meeting on Policy Responses to Population Ageing and Population Decline. New York, United Nations Population Division. Battistella, Graziano. 1995. "Philippine Overseas Labour: From Export to Man agement." ASEAN Economic Bulletin 12(l):257-273. Bengtson, Vern L., Kyong-Dong Kim, George C. Myers, and Ki-Soo Eun, eds. 2000. Aging in East and West: Families, States, and the Elderly. New York: Springer Publishing Company. Casterline, John B., Linda Williams, Albert I. Hermalin, Ming-Cheng Chang. Napaporn Chayovan, Paul Cheung, Lita Domingo, John Knodel, and Mary Beth Ofstedal. 1991. "Differences in the Living Arrangements of the Elderly in Four Asian Countries: The Interplay of Constraints and Preferences." Com parative Study of the Elderly in Asia Research Report No. 91 -10, Population Studies Center, University of Michigan. Chen, Ai Ju, and Gavin Jones. 1989. Ageing in ASEAN: Its Socio-economic Con sequences. Singapore: Institute of Southeast Asian Studies. Cheung, Paul P. L. 1989. "Beyond Demographic Transition: Industrialization and Population Change in Singapore," Asia-Pacific Population Journal 4( 1 ):35- Gauthier, Anne H. 1996. The State and the Family: A Comparative Analysis of Family Policies in Industrialized Countries. New York: Clarendon Press. Goode, William J. 1963. World Revolution and the Family. New York: Free Press. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 22 The Well-Being of the Elderly in Asia Gruber, Jonathan, and David A. Wise, eds. 1999. Social Security and Retirement Around the World. Chicago: University of Chicago Press. Hermalin, Albert I. 1993. "Fertility and Family Planning among the Elderly in Taiwan, or Integrating the Demography of Aging into Population Studies," Demography 30(4):507-17. . 1995. "Aging in Asia: Setting the Research Foundation." Asia Pacific Population Research Reports 4 (April). Hermalin, Albert I., and Bruce A. Christenson. 1992. "Census Based Approaches for Studying Aggregate Changes in Characteristics of the Elderly." Asian and Pacific Population Forum 6 (2):35-42, 58-68. H6hn, Charlotte. 1987. "Population Policies in Advanced Societies: Pronatalist and Migration Strategies." European Journal of Population 3:459-481. Kinsella, Kevin and Yvonne J. Gist. 1995. Older Workers, Retirement, and Pen sions: A Comparative International Chartbook. IPC/95-2, Bureau of the Cen sus, Economics and Statistics Administration. Knodel, John, and Chanpen Saengtienchai. 1996. "Family Care for Rural Elderly in the Midst of Rapid Social Change: The Case of Thailand." Social Change 26(2):98-115. Lesthaeghe, Ron. 2000. "Europe's Demographic Issues: Fertility, Household For mation and Replacement Migration." Paper prepared for Expert Group Meet ing on Policy Responses to Population Ageing and Population Decline, Popu lation Division, Department of Economic and Social Affairs, United Nations Secretariat, New York, 16-18 October. Martin, Linda G, and Samuel H. Preston (eds.). 1994. Demography of Aging. Washington, DC: National Academy Press. Martin, Philip L., Andrew Mason, and Ching-lung Tsai. 1995. "Overview." ASEAN Economic Bulletin 12( 1): 117-124. Massey, Douglas S., Joaquin Arango, Graeme Hugo, Ali Kouaouci, Adela Pellegrino, and J. Edward Taylor, eds. 1998. World in Motion: Understand ing International Migration at the End of the Millennium. Oxford: Clarendon Press. McDonald, Peter and Rebecca Kippen. 2001. "Labor Supply Prospects in 16 Developed Countries, 2000-2050." Population and Development Review 27(l):l-32. Ministry of Community Development and Sports. 2001. "Baby Bonus." Singapore: Singapore Government. Retrieved March 22,2001 from the World Wide Web: http://www.mcds.gov.sig/HTML/home/faq.html. Morrison, Peter A. 1990. "Demographic Factors Reshaping Ties to Family and Place." Research on Aging 12(4):399-408. Okunishi, Yoshio. 1995. "Japan." ASEAN Economic Bulletin 12(2): 139-162. Peterson, Peter C. 1999. Gray Dawn: How the Coming Age Wave Will Transform America and the World. New York: Random House. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Aging in Asia: Facing the Crossroads 23 Phillips, David R., ed. 1992. Ageing in East and South-East Asia. London: Edward Arnold. . 2000. Ageing in the Asia-Pacific Region: Issues, Policies and Future Trends. London: Routledge. Raymo, James M. 1998. "Later Marriages or Fewer? Changes in the Marital Be havior of Japanese Women." Journal of Marriage and the Family 60:1023- Raymo, James M., and Jennifer C. Comman. 1997. "Labor Force Status Transi tions across the Life Course in Taiwan, Singapore, Thailand, and the Philip pines: 1970-1990." Comparative Study of the Elderly in Asia Research Re- /jorrNo. 97-44, Population Studies Center, University of Michigan. Retherford, Robert D., Naohiro Ogawa and Rikiya Matsukura. 2001. "Late Mar riage and Less Marriage in Japan," Population and Development Review 27(1):65-102. Sobieszczyk, Teresa. 2000. "Pathways Abroad: Gender and International Migra tion Institutions in Northern Thailand." Unpublished Dissertation, Cornell University. Stern, Aaron. 1998. "Thailand's Migration Situation and Its Relation with APEC Members and Other Countries in Southeast Asia." Human Resources Devel opment Working Group Paper No. 011 (January). Bangkok: Institute of Asian Studies, Chulalongkorn University. Taiwan Provincial Institute of Family Planning, Population Studies Center, Uni versity of Michigan, Institute of Gerontology, University of Michigan. 1989. "1989 Survey of Health and Living Status of the Elderly in Taiwan: Question naire and Survey Design." Comparative Study of the Elderly in Four Asian Countries Research Report No. 1, Population Studies Center, University of Michigan, Ann Arbor. Uhlenberg, Peter. 1992. "Population Aging and Social Policy." Annual Review of Sociology 18:449-474. United Nations. 1982. World Assembly on Aging, (Vienna, Austria). New York: United Nations. United Nations. 2001. Replacement Migration: Is It a Solution to Declining and Ageing Populations? ST/ESA/SER.A/206. New York: United Nations. World Bank. 1994. Averting the Old Age Crisis. New York: Oxford University Press. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Chapter 2 THE DEMOGRAPHIC, SOCIOECONOMIC, AND CULTURAL CONTEXT OF THE FOUR STUDY COUNTRIES JohnKnodel, Mary Beth Ofstedal, and Albert I Hermalin In Chapter 1 we singled out three factors that have increased the focus on population aging throughout East and Southeast Asia: the rapid demographic changes, particularly sharply lower fertility, that have led to a rapid rise in the proportions at older ages; the concurrently rapid socioeco- nomic changes that challenge many traditional family arrangements; and the interest of policymakers in these regions in fashioning appropriate policies and programs to address these changes. In this chapter we provide an overview of the magnitude and nature of the demographic and socioeconomic changes, reserving a discussion of the policies and programs in place and under discussion for the next chapter. Though our analytic focus throughout the volume is on a comparison of the four study countries, it is useful to start by placing the demographic dynamics in the larger context of changes ongoing in Asia and to contrast the Asian profile with that of Europe and the United States. Subsequent sections of the chapter will then provide a more detailed demographic profile of the four study countries as well as their socioeconomic changes. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 26 The Well-Being of the Elderly in Asia Population Aging and Sociodemographic Change in World Context Table 2.1 provides trends in three key demographic parameters for the three broad subregions of Asia and for select countries within each, including the four study countries. As is well known, the age structure responds primarily to fertility levels and secondarily, once low fertility levels are reached, to mortality. Accordingly the table presents 30-year trends in fertility and mor tality, and the past and projected proportions for those aged 60 or older. The most dramatic declines in fertility have taken place in East Asia, where over a span of 25 years the total fertility rate fell from an average of 5.5 children per women to just 1.9, with little variation among countries. All countries in this subregion currently have fertility rates below the population replacement level. Southeast Asia also shows a sharp fertility decline on average between 1965-70 and 1990-95, but the region exhibits considerable variability, with some countries below the replacement level and others remaining at high fertility. South-Central Asia shows the least reduction overall, but significant fertility declines in the last 10 to 15 years are apparent in each of the countries. Improvements in life expectancy across Asia are no less dramatic. Be tween 1965-70 and 1990-95, life expectancy at birth rose in each subregion by about 10 years or more. East Asia shows the lowest level of mortality in 1995, with life expectancy at about 70 years. Southeast Asia has a life expect ancy of 64 years overall, but considerable variation exists among countries. South-Central Asia's life expectancy is about 60 years. The effects of these rapid changes in fertility and mortality are reflected in the proportions 60 or older, with the countries or subregions showing the largest and earliest declines generally experiencing the highest proportions at the older ages. In 1995, more than one-tenth of the East-Asian population was 60 or older, with Japan exceeding one-fifth. By 2030, it is projected that al most one-fourth of the entire subregion will be in the older age category, with Japan exceeding one-third and many other countries not far behind. Taiwan is somewhat lower in fertility and higher in expectation of life than the regional average in the most recent period, and its proportion 60 or older is somewhat higher as well, and projected to remain so. In both Southeast and South-Central Asia the proportion of the popula tion 60 or older was approaching 7 percent in 1995 and is projected to ap proximately double that by 2030. Reflecting their diverse fertility and mortal ity trajectories, the countries within these regions will differ considerably in age structure. For example, by 2030 Singapore and Thailand are projected to exceed 30 and 20 percent above age 60, respectively, in keeping with their low fertility levels. The other countries in Southeast Asia have lower projec- Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 rt •o f^ Cs in vO .I •o CJ •5 in -o •o NO .o a. a. JS •o Cs BQ ON ON ao S! •o OCN-HCN-^p -^ ON CN in ** fi oq p f) CN oq in ^ CO CO Tj* in NO in ^ in CN •—' *—' ^ CO O ^ ON ON CNCNCOCOCNCN -H-^-^ — CNCO-H -H-^—i -H Ovoor-OvO•* ootsCNONOin inOCNQTtino ONONO-HO _:__;„_;_: t^txiCO^-^-^^ firiSt^inCN — ro ro oq ON oo ro m CN —i CN CN OONO-HONCNOOlO CNNOOOOOO NO NO CN o —i NO —i r- Tt ^ — NO in —i —i ON o •* O O ^5 r** ••T oo ^i ^^ CD ^^ ^r ^^ ^^ ^^ vo O Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 28 The Well-Being of the Elderly in Asia tions corresponding to their higher fertility rates. In the Philippines, which has the highest fertility in the subregion, only about 13 percent of the popula tion is expected to be over 60 in 2030—a proportion similar to that found in the other higher fertility countries in the region. Sri Lanka, which had a fertil ity rate close to replacement level in 1990-95, is the only country in South- Central Asia for which the estimate of those age 60 or older in 2030 approaches 20 percent. The United Nations classifies the countries of the world as more devel oped, less developed, and least developed, with the latter category a sub-class of less developed (United Nations, 1999). A strong association exists between level of development and population aging. In 1995,18.4 percent of the popu lation in the more developed countries was aged 60 or over, as compared to 7.3 percent in the less developed countries and 4.8 percent in the least devel oped. The medium-variant projection for 2030 estimates percentages of 29.2, 14.6, and 7.0, respectively, displaying sharp increases for all but the least developed countries. All the countries of Asia, with the exception of Japan, are in the less developed category, as are all countries of Latin America and the Caribbean, and Africa. Among these three geographic regions, Asia has the highest proportion of those 60 or older and is projected to maintain this lead into 2030, although the differences between it and Latin America are quite small as shown in the table below. Percent of Population Age 60 or Over *Medium Variant Projection Source: United Nations, 1999, Vol. II, Table A.I AfricaAsiaLatin America and the Caribbean19954.98.27.42030*6.917.215.7 A more detailed demographic comparison of Asia and Latin America, including a discussion of differences in the level of attention given to popula tion aging, is given in Hermalin (1995). But given the generally greater famil iarity with trends and issues in population aging in the United States and Eu rope, the tables that follow focus on comparisons with the more developed regions. Table 2.2 compares the United States, Europe, East Asia, and South eastern Asia on a number of measures of population aging. The measures chosen are designed to look at the age distribution of the older population, the Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Demographic, Socioeconomic, and Cultural Context 29 growth rates at different advanced ages, and elements of population size. In the first panel we see that the proportion of the population 60 or older is higher in the older industrial countries of the United States and Europe than in Asia. In Europe, which is somewhat higher on this measure than the United States and is projected to remain so, the proportion over 60 in 1995, at 19 percent, is almost twice as high as East Asia (at 10 percent) and almost three times as high as Southeast Asia. The second panel of Table 2.2 presents the age distribution within the older population (age 65 or older). As shown here, there is also a much higher concentration of the "oldest-old"—those above age 80—in Europe and the United States, with one in five of the older popu lation above 80 in the former and almost one in four in the latter, in contrast to about one in seven in the two Asian subregions. But these differentials are projected to diminish considerably over the next 30 years. By 2030, the pro portion 60 and over will approach 30 percent in Europe and 26 percent in the United States, compared to 24 percent in East Asia and 15 percent in South east Asia. The proportions aged 75 or older which are currently over 5 percent in the United States and Europe will advance to 9 and 10 percent, respec tively, by 2030; in East Asia, the proportion will triple from 2.2 percent to 6.6 percent, and in Southeast Asia it will advance two and a half times from 1.3 to 3.2 percent. The implied growth rates and resulting population sizes shown in the lower part of the table are quite revealing. In each of the four areas^ the growth rates among the older population will be fastest at the oldest ages, leading to further increases in the mean age of the 60-plus population. The rates of growth will be considerably higher in East and Southeast Asia than in Europe and the United States. Between 1995 and 2030, the average annual growth rate for the 60-plus population is projected to be around 3 percent for the two subregions of Asia, compared to 2 percent for the United States and 1 percent for Europe— rates similar to those experienced between 1975 and 1995. The population 80 and older will be growing at an average annual rate of about 4 percent over the next 30 years for the Asian areas and only about 2 percent in United States and Europe. Though these rapid rates of growth for the older population in parts of Asia and elsewhere are often talked about from the standpoint of the speed with which these areas are aging (for example, the number of years it takes a country to double its population over age 65 from 7 to 14 percent is quite frequently shown), from a policy standpoint attention must also be given to the numbers involved. Although future fertility levels will largely determine the proportion of elderly within each country, the number of elderly at any given point is essentially a function of past births and mortality levels. Because of the past large birth cohorts, Asian countries must anticipate a sharp increase Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 o uf Ov ON r^ Ov .a •o •i J2 tN IN 0l rt of)om£inON•S to 2IIgue8 S.5 2 3 oo•5 a 2J 3 < 3 st2 ^|He8 u•a 33 Co— ~ -IMS 2«S<8eeS 8.5 23 "5— — CN ont — O 00 —2S*« S3!:2ONONOCN OtseOCNONCNOW-i ^^VlO*^* CN ts — CN CN CN tN-r o r l — . — momvOOONVi ON'OO*TJ* ts CN CN CN ts CNt^mr^JNO ^OOONONvO ^ •— . o o4r-NOts fiootso oomvoo^3 ^3 CN *™* ON OO t** Tf*— • m rs s-i*•iOONCNO f^OONOOOi CO IO CO *n C^ *O ^^ — ' — ' CO CN — CN^irlviO ooNO^^r^r^i CO ^ r^i O ^^ rv1 r**lOflOO W^ONNO^^Tt CN --^ r*- ro NO O N fi r^ ^3r^TJ-^-o tsONmtsTr CN ^ s^TtTtoooo or- — mvO-m— r-oooovi — * -s m ts — -^ooins-i-H mmO*^TtO•^r4 ininOvo•e3 P — v ,*•* JS O IT) O3 « e * 2.~ - .- o— .^ d'2bMON^*ONC- .i ^cavvor^r^rtaj— *i'S'flTjlOSOmino £;S.3inOinO C&Ovor^oo tftOi®vor^i - oo u O X 9 IK C ft O S. eo808o§08o8o80§o8o808o8o8a Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 a -n rt %errtnt in<EL,r- TrsiS2mvO•*tsef00^NOsvor^m!N:1:iHiite> 9t -In. n1s2*mCN•*tO00„ONON8NO^00^-rrN0m111iS3 •ort0 and orMm5Nminl2t^Tf.Ot000_;**'*t•11iTJeoNOrsir^Svo00y.Soiro«-t*-ONOii:iIeo•aa r-00 00r-ingr-0°^—^™ONr-m002r-in00m8'iii|80 and o-^ ^; w-i r-; 00^^ m m m mm ^^ ts r- oO o1 rsi m •*— — i m oo r-9NO O O -^ Cs)O ts CN CN CNp ^ ^ m O r-oi m m ^* ^*m p CN oo r-;— m m m mm-^^^-O — — — mo sn ^^ Tt r-— — — . CN mi : i : :: : i i i: i i : :: i i i :IIIIltllll..o •i£ o"3 O "3 "3 15 w a.— — _. _•a - •« c c c 2 g & 3 n n n n2 e o Q m in o g g H vO vo r- oo< iNO CNS200 TT2^m in=S ?!CNI- 00r-: vboo —CN r*oi rjmm vo^ 2OO tNm tNr- ONro r-m O-^ •-r-ino vb00e Ov^ ^•—vo r*-rvi TJrm1 ||III!= S o o&^ vO CXIe GEL. £00o r-tsSpONNONOCNmmino00mCNONr^r-ei$oSme1^.NOvoCNOCNC Orti20 Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 32 The Well-Being of the Elderly in Asia in the number of elderly. Some of the magnitudes involved are shown in the bottom panel of Table 2.2. In 1995 the number age 60 and over in Europe and the United States combined, 182 million, was about equal to the number in East and Southeast Asia combined, at 179 million. By 2030, East and Southeast Asia are projected to have 518 million over age 60, 76 percent more than the combined total of the United States and Europe. The changes at 80 and over are in some ways even more dramatic. The numbers in Europe and the United States combined are projected to nearly double, from 30 million to 56 million, but the numbers in East Asia and South east Asia will be four times as large, advancing from 16 million to 65 million. As a consequence, the number of oldest-old in the two subregions combined will be 15 percent greater than the combined total of the United States and Europe by 2030. It is also worth noting that in 2000, about 46 percent of the world's population aged 75 or older was estimated to reside in Asia; and in 1998, the number aged 80 and over in China, India, and Japan combined was estimated to be more than 22 million. In addition to the rapid growth of numbers at older ages, the gender imbalance that occurs at the older ages due to the differential in life expect ancy between men and women merits attention. As of 1995, slightly more than half of all those 60 or older were women but this proportion rises to two out of three for those 80 or older in East Asia and six out of ten in Southeast Asia. Older women in many parts of Asia currently have lower education and labor force experience than men, less income and access to assets, and dimin ished authority within the family. Though these conditions are changing to some extent, these disadvantages make many women dependent on family members and on public programs, especially at advanced ages and under con ditions of illness or disability. These needs often place pressures on family members for caregiving and may tax insufficiently developed public programs. The current gender composition at older ages is not expected to vary greatly over the next 50 years according to United Nations projections, given current estimates of future life expectancy for each sex. They reveal that by 2050 both East and Southeast Asia will have about 85 males per 100 females aged 60 or older, but only 59 males per 100 females among those 80 and older. The improvements in life expectancy that have occurred in the past and are projected for the future do alter the rates of widowhood, so that the pro portion of the older population who are widows will continue to decline. The changes that have occurred over the last 20 years for the four study countries are shown below (in Table 2.6). These changing proportions and numbers related to older populations have strong policy implications, particularly in the provision of health and social services. In many countries in Asia (and elsewhere in the developing Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Demographic, Socioeconomic, and Cultural Context 33 world), the health infrastructure has been largely attuned to addressing infant and child mortality and infectious diseases, and providing family planning and other reproductive services. The number of new medical, public health, and social service programs, personnel, and facilities required to address the needs of the older population are likely to put substantial strain on many econo mies. (See Chapters 3 and 10 for additional discussion.) By contrast, insofar as the relevant European and American infrastructures have expanded to meet the past growth in their older populations, they may be able to adjust more easily to the more modest rates of growth at the oldest ages projected for the near future. The question of costs and services is part of the larger question of the effect of population aging on economic growth and productivity raised ear lier. The demographic metric most often invoked as an indicator of popula tion age structure effects on the economy is the dependency ratio. This ratio relates the age segments of the population with little economic productivity (defined in Table 2.3 as under age 15 or over age 64) to the age segment most economically active (15-65). Although like many single measures it has its controversial and ambiguous aspects (Schulz et al., 1991, Chap 3; World Bank, 1994, p. 35; Johnston, 1999), the differences between Asia, Europe, and the United States in this regard are worthy of notice. Table 2.3 contrasts the United States, Europe, and the two subregions of Asia on the overall dependency ratio and ratios for both the younger and older age components in the 1960s, the 1990s, and as projected to 2025-2030. In the United States, as of 1990-95, the total dependency ratio was 52 people either under age 15 or over age 64 to every 100 people between the ages of 15-64, compared to 50 in Europe, 49 in East Asia, and 68 in Southeast Asia. As expected because of declining fertility rates, the portion of the depen dency ratio due to the older population has been increasing over the past 30 years and the portion due to the younger population has been decreasing in all regions. The overall ratio is affected by the magnitude of each component and the rate of change. For the past 30 years, the decline in the younger age popu lation more than offset the rise in the older population in the United States and Europe, causing the overall ratio to fall. But over the next 25-30 years it is expected that the decline in the younger age population will be limited while the older age group will continue to grow, producing ratios similar to 1960 levels. The story in East Asia is similar, with the steep decline in the overall dependency ratio expected to level off. In Southeast Asia, however, the later onset of the fertility decline means that the overall dependency ratio is ex pected to continue to decline through 2025-2030, as the decrease at younger ages more than compensates for the increase at the older ages. It is of interest that in the next 25 to 30 years, all four regions are expected to be fairly similar Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 o IN •o in vO 'o .5b •o (U .ti •o 0l .2«13O00.2Iii§utAa55 1•33I0SCNti11sONvASrl0Qi,1ooON8ONO CNOON£ONOS'05sONO CO CN1i02isoCNmON ONINO2sotN.ON 2I0NO—8•a2uBV*S|ONOoosq_iooJoON—IO,-In0ONiO00 00CN CNNO1(2«.™NOONs--^-*~i00ON ONNOsONCN CNd5ON od CNS2*'B Ori^0q^3q rsON•f:oop.t~i_ONCOON ON tNONod**)•8ODOvNOq-eodONr-oo tsoo2ooSOi0NOroiooNOvi00 ONCN06.I-Index o Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Demographic, Socioeconomic, and Cultural Context 35 in their total dependency ratios (ranging from 46 in Southeast Asia to 59 in the U.S.) but sharply different in their indexes of aging, which represent the ratio of the aged dependency ratio to the young dependency ratio (see row 4). The changes in the overall dependency ratio and its components in each region are a reflection of the magnitude and speed of the demographic transition experienced in terms of fertility and mortality rates. An important line of economic research has combined these resultant changes in age struc ture with life cycle variations in savings and consumption to trace aggregate effects on economic growth and investments (Coale and Hoover, 1958; Ma son, 1987; Bloom and Williamson, 1998). A number of these analyses view the sharp declines in the dependency ratios in East Asia experienced between the 1960s and the 1990s as a "demographic gift" which contributed signifi cantly to the high savings and economic growth rates this region generally enjoyed over this period (Mason et al., 2001; Bloom and Williamson, 1998). From this standpoint, those countries in Southeast Asia (and in other develop ing regions) that are projected to have overall declines of dependency ratios in the coming years can expect a stimulus to their economic growth if appro priate policies are in place. For the United States, Europe, and portions of East Asia where the ratios are projected to level off or increase, their gener ally higher per capita incomes should make it easier to adjust to the changes in age composition. A more detailed discussion of these and other economic models, which relate the effects of population aging to aspects of economic growth, is presented in Chapter 4. Demographic, Socioeconomic, and Cultural Context of the Four Study Countries We turn now to the social, cultural, economic, and demographic settings of the four study countries as they bear on the circumstances of the elderly within their populations. In all four countries the current elderly populations have lived through rapid demographic, social, and economic changes. In important respects these changes have operated in similar directions in each of the coun tries. Yet, because they occurred at different paces, were conditioned by dif ferent cultural settings and political conditions, and evolved from different levels of Socioeconomic development, these changes have resulted in vari able circumstances for the elderly populations in the four countries. The cur rent demographic, socioeconomic, and cultural contexts are characterized by a combination of common and contrasting features that shape the behavior of the elderly and condition their relations with their children, relatives, and com munities. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 36 The Well-Being of the Elderly in Asia In addition to reviewing the broad settings, we describe some basic socioeconomic and demographic characteristics of the elderly who constitute our samples. We draw both on the four major surveys that provide the basis of the analyses and on other more standard sources. Political and Cultural Setting The political and cultural context has played an important role in defining the nature and magnitude of demographic, social, and economic changes in each of these countries, and in shaping the family, social, and productive lives of their populations. The four countries represented in this study have distinc tive political histories and cultural underpinnings, yet they share a number of significant cultural mores—such as the centrality of the family in social life and in support for the elderly. In this section we provide a brief review of the political history and cultural context for each country. Philippines The Philippines is an archipelago of 7,100 islands stretching almost a thou sand miles. It is comprised of three major island groups, Luzon in the north, Mindanao in the south, and Visayas in between, and it is separated from the mainland Asian continent by the South China Sea to the west and the Philip pine Sea to the east. The Philippines was originally inhabited by Malay and Negrito groups. Centuries of pre-colonial cultural and commercial contact with other coun tries in Asia, including China and India, and almost four hundred years of western colonization have produced a culture that is a blend of eastern and western influences. In the 16th century the Philippines became a colony of Spain, some 40 years after the historic circumnavigation of the globe by Ferdinand Magellan and his documentation of the archipelago, which he named after the Spanish King Philip II in 1521. Spain occupied the country for 333 years (1565-1898). After a brief period of independence following the Philip pine revolution of 1898 the country was next colonized by the United States from 1899 up to the start of World War II, when the Japanese invaded the country and ruled from 1942 to 1945. The Philippines became an independent nation in 1946 (Agoncillo, 1990). The most tangible legacy of Spanish rule is Catholicism, which is pro fessed by a large majority of the Filipino population. The Catholic religion Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Demographic, Socioeconomic, and Cultural Context 37 has had a pervasive influence on Philippine culture. For example, the Church actively advocates against artificial contraception and the Philippines is one of the few countries where divorce has not been legalized. Although Islam was introduced to the country in the 14th century by Arab traders via the Malay Peninsula to the south, the subsequent Spanish colonial settlements in the northern and middle parts of the archipelago arrested its northward spread. To this day Mindanao, which was never colonized by Spain, is the only Muslim area in the Philippines. The Philippine^ island geography and mountainous topography have contributed to the development of dozens of distinct languages, all belonging to the Austronesian family of languages. Of the many languages, five may be considered major, and one, Tagalog, became the basis of the national lan guage Filipino. English is also used widely throughout the archipelago, a legacy of American rule. A major characteristic that Filipinos share with their Asian neighbors is the primacy of the family in social life. Kinship is of bilateral descent with both maternal and paternal kin given equal importance in defining links be tween generations. This system ensures a large pool of potential sources of economic, emotional, and social support among kin (Castillo, 1977; Lopez, 1991; Medina, 1991). Other cultural values support the kinship system as a social alliance system by which members enter into reciprocal ties. These values include the maintenance of smooth interpersonal relations manifest in the avoidance of open disagreement, the observance of debts of gratitude (utang na loob) to those owed favors, and the avoidance of behaviors that bring shame (hiyd) to one's self and to family (Lopez, 1991). Utang na loob are debts generally repaid "with interest" rather than discharged in kind (Hollnsteiner, 1973). This type of debt is most evident in the parent-child relationship, whereby children are expected to be grateful to parents for having given them life itself. This lays the groundwork for the cultural expectation that children should care for elderly parents. Failure to repay utang na loob is sanctioned by hiya, the feeling of shame and social discomfort that comes from knowing that one is in a socially unacceptable position (Lynch, 1973). Respect for all older persons is also a fundamental value. The culture does not proscribe specific differential treatments of chil dren according to their gender, and makes no gender-based distinctions in matters of property ownership and inheritance. All children inherit equally from parents. Similarly, all are expected to repay debts of gratitude to their parents by caring for them in their old age (Medina, 1991). Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 38 The Well-Being of the Elderly in Asia Thailand Thailand is the only Southeast Asian country never to have been colonized by a Western power. The government has been a constitutional monarchy since 1932 when the absolute monarchy was ended though a bloodless revolution. Despite the democratic aims of the 1932 revolution, Thailand has since been ruled under various degrees of military dictatorships with frequent coup d'etats replacing one military clique with another (10 successful coups out of 19 attempts have occurred since 1932). Thailand experienced a brief period of relative democracy from the early to mid-1970s, which ended in a violent overthrow of the elected government and the installation of a short-lived ultra right wing government by the military. During the 1980s, military influence weakened, although again an elected government was overthrown by a coup in 1991. Since 1992, when massive popular demonstrations against an unelected military prime minister resulted in his resignation, parliamentary democracy has become more firmly established. The Thai population is relatively homogeneous in major cultural as pects. The vast majority are ethnic Thais and speak some form of the Thai language; about 95 percent profess Buddhism, typically of the Theravada branch. Still, numerous minorities can be defined in terms of ethnicity, lan guage, or religion. Muslims constitute approximately 4 percent of the popula tion and make up the largest and most notable religious minority. Christians and Hindus constitute less than one percent each. According to the 1990 cen sus, most Muslims (81 percent) live in the southern region, and about half speak Malay rather than Thai; the remaining Muslims reside mainly in Bangkok and the central region (Knodel et al. 1999). The Malay-speaking Muslims are ethnically and culturally quite distinct from the ethnic Thai majority and con siderably less assimilated than Thai-speaking Muslims. According to the 1995 Survey of Elderly in Thailand, 3 percent of the respondents age 60 and over indicated they were Muslims and about half of these indicated they spoke Malay. A particularly influential ethnic group is the Chinese and Thai-Chinese, though it is difficult to estimate their numbers with any precision. Indeed, because considerable assimilation including extensive intermarriage has taken place, there is no agreed-upon definition of Chinese ethnicity in Thailand. Various estimates suggest they constitute from 7 to 12 percent of the total population (Kaplan, 1981). They tend to reside in the urban areas and to be involved in commercial activities. According to the 1995 Survey of Elderly in Thailand, about 4 percent of the population aged 60 and over identified themselves as Chinese and another 9 percent as mixed Thai and Chinese. Although some older Chinese still speak their native dialect, these cohorts are Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Demographic, Socioeconomic, and Cultural Context 39 dying out. Younger cohorts are all fluent in Thai and indeed many no longer can speak Chinese. Other numerically smaller minorities include Indians and Pakistanis, who are engaged largely in trade, Cambodians concentrated largely along the border areas near Kampuchea, Vietnamese living largely in the northeast, and a variety of hill tribes located mainly in the mountainous regions, especially in the northwest. Four regional dialects are spoken in Thailand: central, northeastern, northern, and southern versions of the Thai language. The official language is the central dialect, which is spoken throughout the central region, the lower north and parts of the upper south. It is generally understood throughout the country, although older people are more likely than younger persons to feel uncomfortable conversing in it. The other regional dialects are only partially mutually intelligible. The northeastern dialect is very similar to Laotian. Nu merous small minorities speak their own distinctive (not mutually intelligible) languages including versions of Cambodian and numerous tribal languages such as Hmong and Karen. Because all residents must attend school taught in the Thai language, younger generation members of non-Thai ethnic groups understand Thai, but their elders often have only minimal mastery. As in other southeast Asian countries, in Thailand the family tradition ally assumes primary responsibility for the elderly. A moral obligation for adult children to support and care for their elderly parents pervades Thai cul tural values and provides a strong normative basis for the prevailing pattern of familial support. As noted in Chapter 3, this responsibility is reflected in both the laws of the land and in the formulation of social policy. Focus group research indicates that both elderly parents and their adult children share similar views regarding this sense of responsibility and that it is found in all regions, transcending economic status or rural-urban residence (Knodel, Saengtienchai and Sittitrai 1995). Age relationships are an important part of Thai culture firmly embed ded in the language. Forms of address typically incorporate acknowledgement of who is more senior, and many terms identifying kinship distinguish who is older and younger than the reference person. In general, respect for seniority in age is an integral part of Thai culture and traditionally older persons are to be treated with respect. These traditional values have likely changed in recent decades, especially in practice, but systematic evidence to determine the ex tent and nature of such change is still lacking. Ethnic Thais tend to be matrilocal, preferring to live with a married daughter than a married son (especially those in the northeast and upper north, essentially the areas coterminous with the regional dialects) (Knodel, Saengtienchai, and Sittitrai, 1995). At the same time Thais tend to be to rela tively flexible in this matter, living with a son if no daughter is available Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 40 The Well-Being of the Elderly in Asia (Knodel, Chayovan and Siriboon, 1992). Inheritance is bilateral with chil dren typically inheriting equally, except for children remaining with their parents, who commonly gain the house and perhaps an extra share of the land. Ethnic Chinese and those Thai-Chinese who still identify strongly with their Chinese heritage, however, tend to be patrilocal. Repaying parents is generally viewed by Thais as a continual obliga tion that starts when the children are old enough to provide meaningful help— long before parents reach old age. Accordingly, the care and support provided by children when their parents are too old to take care of themselves is viewed as the culmination of this process. Underlying the obligation to repay parents are the concepts of katanyu katawethi and bunkun, both of which are firmly ingrained in Thai Buddhist culture and have no simple English equivalents (Rabibhadana, 1984; Podhisita, 1985). Both terms relate to reciprocation of actions around a sense of gratitude and debt. Katanyu refers to a constant sense of awareness on the part of someone for benefits that another person has bestowed upon him. Katawethi refers to doing something in return for them. Bunkun characterizes the person who bestows favors that incur a sense of gratitude and debt on the beneficiary. The concept of katanyu katawethi usually refers specifically to parent- child relationships while bunkun extends to many realms of life. Both con cepts, however, characterize the essence of the relationship between parents and children. Giving life to and raising a child provides parents with the epitome of bunkun and instills a sense of gratitude and debt in the child that is virtually impossible to repay completely. This bunkun parents have in relation to chil dren directly leads to the sense of obligation adult children have to provide support and care to them when the parents are in their elderly years. Although Thai Muslims do not describe filial piety in these same terms, they also have a strong sense of obligation to parents that is deeply rooted in their religion. Indeed throughout much of east and southeast Asia, there are similar cultural prescriptions related to the obligation to repay parents especially when they reach older ages and can no longer support and take full care of themselves (Asisetal., 1995). Taiwan Taiwan is an island approximately 100 miles from the Chinese mainland that comprises about 4.25 million square miles of land area. The island is mostly mountainous with dense population concentrations in cities. The total popula tion of Taiwan was 22 million in 1999, 5 million of whom lived in one of the three major cities in Taiwan: Taipei, Kaoshiung, and Taichung. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Demographic, Socioeconomic, and Cultural Context 41 For many thousands of years prior to 1500, Taiwan was inhabited ex clusively by the Taiwan Aborigines, a population group that shares cultural features with certain groups in Oceania and the Philippines (Fricke, Chang, and Yang, 1994). Taiwan was annexed by China only in 1682. Major waves of Chinese settlement from the 17th through the 19th centuries brought two main Chinese dialect groups to the island: the dominant Hokkien speakers who migrated in from neighboring Fujian province and tended to settle in the coastal and valley regions, and the Hakka speakers who predominantly came from Guangdong province and who, as a result of their later arrival, tended to clus ter in upland and hilly areas. Customarily these two groups are referred to as Taiwanese. By the 19"1 century the aborigines had been decimated, with sur viving communities driven into remote mountainous areas. For the most part, the Chinese settlers tried to recreate on the island the social and cultural forms of their home communities on the mainland, and they considered themselves part of China. By the end of the 19th century the Chinese population of Taiwan numbered over 2.5 million, and in 1885 the island was elevated to provincial status within China, with its capital in Taipei. With China's loss in the Sino-Japanese War of 1894-95, the second distinctive phase began. Taiwan was ceded to Japan, which ruled it as a colony from 1895 to 1945. Japan introduced a number of kinds of modernization to the island during its period of occupation, including expanded transportation and communications, increased education, and broadened public health ef forts. Yet, the occupation of Taiwan by Japan did little to change basic family organization or to alter the basic structure of rural society (Barclay, 1954; Cohen, 1976; Hermalin, 1976). Following World War II, the period of Chi nese territorial unity lasted only five years, until 1949, at which point the victory of Mao Zedong and his followers in the Chinese Civil War drove the remnant Nationalist forces into exile in Taiwan, where they formed the Re public of China, a rival regime that continues to exist. The Taiwanese family system is based on the patrilineal Chinese kinship system that was brought to Taiwan by the migrants from Southeastern China. Under this system, inheritance is passed through the male line and the availability of male heirs is vital to the continuance of the lineage. (Although daughters often received a substantial dowry, they were not entitled to a share of the inheritance until after post-World War II, and even then many women voluntarily signed away their share (Lee, 1999). Whereas daughters typically leave their natal families to join their husbands' families upon marriage, sons were expected to remain in the family household, eventually assume responsibility for the family farm or business, and provide care for their parents in old age. The household ideal called for several married sons and their families Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 42 The Well-Being of the Elderly in Asia living together in the same household with the older parents (Cohen, 1976; Wolf, 1968). For various reasons including economic considerations this ideal was rarely achieved (Cohen, 1976; Wolf, 1968; Freedman, 1958), but it was and still is quite common for older parents to live with one married son and his family. The intergenerational contract in Chinese society is governed by strong norms of filial piety, which stress the primacy of children's relationships with their parents (Cornman, 1999). Children are expected to be obedient and sub servient to their parents, particularly their father, throughout their lives. Par ents provide care for children with the expectation that their children will support them in old age, making it a moral obligation. Given the gender dif ferences in family membership under the patrilineal family system, parents primarily expect old-age support from their sons, although women as daugh ters-in-law are expected to show respect to and support their husbands' par ents. Despite this strong patrilineal system, Chapter 6 reveals that daughters also provide considerable assistance to older parents. Over 90 percent of the Taiwan population are Buddhist, Confucian, or Taoist, with about 5 percent Christian. Singapore Singapore is a city-state located at the southern tip of the Malaysian Penin sula. The country is comprised of one main island, which is separated from Peninsula Malaysia by the Straits of Johor, and roughly 60 smaller islands for a total land area of about 250 square miles. During the early stages of its history (around the end of the 14th cen tury), Singapore was caught in the struggles between Thailand (then Siam) and the Java-based Majapahit empire for control over the Malay Peninsula, and eventually the island became part of the Malacca Sultanate. In 1819 the British established a trading post in Singapore and in 1824 Singapore's status as a British possession was formalized, an arrangement which continued through 1959, except for a 3-year period during World War II when Singapore was occupied by Japanese forces. An unsuccessful attempt by the Communist part of Malaya to take over Singapore by force in 1948 led to a 12-year state of emergency, but Singapore eventually attained self-government in 1959. In 1963 Singapore merged with Malaya, Sarawak, and North Borneo to form Malaysia, but the merger was short-lived and Singapore was separated from the rest of Malaysia in 1965. Since 1965 Singapore has thrived as a sovereign, democratic, and independent nation. Singapore's population is multicultural, descended primarily from im migrants from the Malay Peninsula, China, the Indian sub-continent, and Sri Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Demographic, Socioeconomic, and Cultural Context 43 Lanka. The total population of approximately 3 million consists of 79% Chi nese, 14% Malays, 6% Indians, and 1% Eurasians (General Household Sur vey, 1995). Social policies and services for the population tend to be concep tualized from a multicultural perspective, and sensitivity to the worldviews of these various cultures is considered important. Cultural aspects that are of particular relevance to this study are highlighted for each of the major ethnic groups below. The Chinese have a pragmatic, achievement-oriented world view em phasizing the importance of contributing to family and society. Confucian values and principles such as a strong work ethic and maintenance of hierar chical relations between the individual, family, and state form the basis of society. The present cohort of elderly Chinese in Singapore consists largely of believers in Taoism, Buddhism, and Confucianism, with a small minority of Christians. Many of them believe in ancestor worship, a system of beliefs and rituals that ascribe high status to ancestors. Older family members are re spected because they are perceived as "future ancestors." The elderly are ex pected to behave with propriety and their descendents, especially the sons, are expected to be filial. The Chinese subscribe to a patriarchal and patrilineal family system, under which males in the household are the main providers and decision makers. However, as a result of social changes the balance of power between the genders is being re-negotiated. The centrality of Islamic religion and family is a distinguishing cul tural feature of the Malays. The bilateral, egalitarian family system differenti ates this community from the patrilineal Chinese and Indian ethnic communi ties. Malay elderly are respected for their religious knowledge and actions; their status is increased, for example, if they have performed the haj (pilgrim age) to Mecca. The Muslim doctrine advocates each devotee to remember that after death he will have to face judgment before God, and therefore he must conduct his life according to the principles of the religion. Older Malays tend to avoid being burdensome to their children by striving to be as indepen dent as possible, although a large proportion of elderly Malays live with mar ried daughters and sons-in-law. Although the Indian ethnic group is characterized by a high level of heterogeneity across linguistic, religious, and regional boundaries, certain cultural similarities such as dressing, food, and music bind its members. The majority of Indians follow the Hindu religion, which emphasizes reincarna tion and the law of karma (or principle of cause and effect). Within Hinduism, the spiritual search for the upliftment of one's soul is emphasized. In daily living, the family assumes great importance in the life of each person. Apart from Hinduism, a considerable proportion of Indians are Muslims, followed by Christians. Among Indians, some differences exist between those who Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 44 The Well-Being of the Elderly in Asia have their roots in North India and those whose ancestors came from South India. North Indians tend to be more conservative and the dominance of the male in the household is emphasized. The traditional norms seem to be chang ing faster among South Indians, and women tend to be more liberal. Never theless, the patriarchal system is still prevalent among most Indian house holds, except for the Malayalee sub-community. The Indian community is similar to the Malay community in the emphasis of religion in daily life. How ever, it is more similar to the Chinese community in the linguistic divisions that exist, unlike the Malay community which is united by a common lan guage (although some variations exist such as Javanese and Boyanese Malay). Demographic Change Table 2.4 shows the past and future trends of key demographic indicators related to population aging in the four countries. For all but Taiwan they are based on the United Nations medium variant projections as presented in the 1998 revision (United Nations, 1999). Although projections of population dynamics for 30-50 years in the future of necessity rely on strong assump tions, they can still illustrate some of the major patterns likely to emerge. As indicated by trends in the total fertility rate, all four countries have experienced significant fertility decline. Indeed this is the primary demographic force underlying the inevitable population aging that is underway. Thailand, Taiwan, and Singapore were all experiencing total fertility rates slightly below 2 births per woman in 1995, and thus have already fallen below replacement level fertility (which, at present mortality levels, corresponds to a total fertility rate between 2.1 and 2.2). The Philippines, however, has not experienced a similar degree of decline and in 1995 had a total fertility rate approximately twice the level of the other three countries. Current fertility levels in Singapore, Taiwan, and Thailand are very similar, and are projected to remain so for some decades. Nevertheless, the timing of the declines in these three countries have differed significantly over the past few decades, with Singapore reaching low levels first and Thailand last. The different paths toward low fertility have important ramifications for the course of population aging, as is evident from the differences among the four countries in the proportions of their populations aged 60 and older. Most striking is the much slower and more moderate levels of aging that char acterize the Philippines' population compared to the other countries. Even by 2020, the projections suggest the elderly will still constitute less than one in ten Filipinos, whereas both Taiwan and Singapore will have exceeded this Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Demographic, Socioeconomic, and Cultural Context 45 level two decades earlier. The concentration of Thailand's fertility decline in a more recent period means that population aging there will lag behind Singapore and Taiwan by about a decade. Taiwan's age pyramid is influenced by massive migration of military personnel in 1949 when the Nationalist sol diers left the communist mainland for Taiwan. The migrants were predomi nantly males in their 20s and 30s who married at low rates. This event has contributed significantly to population aging in Taiwan and accounts for why Taiwan currently has a higher percentage aged 60-plus than Singapore, even though Singapore experienced a somewhat earlier decline in fertility. Never theless, with the possible exception of the Philippines, the pace of population aging to be experienced by the countries under study will be extremely rapid compared to the historic experience of the West (Jones, 1993). If the projec tions are correct, Singapore, Taiwan, and Thailand will see the share of their 60-plus population double, from 10 to 20 percent, within about two decades, with this change taking place somewhat earlier in Singapore and Taiwan (2000- 2020) than in Thailand (2010-2030). Major decreases in mortality levels have occurred in the study coun tries over the last half of the 20th century. Life expectancy at birth has risen 20 years in all four, and both Singapore and Taiwan now have life expectancies comparable to those in North America. Although mortality levels are not as favorable in the Philippines and Thailand, life expectancy at birth in both is somewhat above the regional average for Southeast Asia (see Table 2.1). Con tinued mortality improvement is expected in the next few decades, although at a more moderate pace, with much of it concentrated at the older ages since levels at younger ages are already quite low. The past and future improvements in mortality translate into increased survival rates to older ages and thus an increasing number of elderly persons. Combined with the high fertility rates of the past, the result is unprecedented growth in the numbers of elders in each country. For example, during the quarter century between 1970 and 1995, the size of the population age 60- plus more than tripled in Taiwan (reflecting in part the aging of the special cohorts of military migrants referred to above) and much more than doubled in the other three countries. These rates of growth are far higher than those for the overall population, accounting for the elderly's increasing share of the population, and are anticipated to continue unabated for some time (see the last three columns of Table 2.4). Thus, the number of persons 60 and older is projected to more than double once again during the quarter century between 1995 and 2020 in all four countries, and perhaps more than triple in Singapore, despite falling rates of overall population growth. In addition, the elderly popu lation itself is beginning to age in each country and will continue to do so Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 O N^ C^) ON •o cDOUI tssts—o"'o^5tss~1o oo ONOr-ONs—oONndicatm— oCN NOr-— ONCN ror-ifv 1*-ts ~* -fel OvNOoo rovotN ONOON ON^Soq pin r-ON oovo tsinro inr^ r^,tJOf;eJlippinsTotal fertility rate1 Lifesexpestancy atJS£°iOr^,ofmCOts_NOCNtsroe*O OONOeoTOo tsvo"ts5r-—ovO^^•r.|# persons a-e rts+tsmON^^vor-¥^in^^inm0-;vi1n o133 CL D.'oOmvo^1-Jrf ,*•^r-00mr-?cjm00CJvdr^p-i+% of elderly a-e _2NeoSOOovoo0r-mp 'n^.inqeo0H O•ag>. ugT31•8uooa•a O3(NONr-ONinp.O inr-^-00r-00^CNONr,roOv00201Total dependemeyON ON—• in^^ oq™ -ter**OO Ov— f>Hr~r- —— ONNO00 00voro mts 06vo -; voNO C^in 06in••* mNO CNinvo ovo r^"f;clailandTotal fertility rate' Lifesexpestancy aljaH00tsCOin,—^NOr*'O ^mots.r-00CO*'NOm^NOr-r- tswSrsm-00 00— *m11# permens a-e rte+SitN min t-I00 —O In•— i ^r^* rooo —•*vq inCOr- voNO ONmTt ro in 0600 ON•* inin Tt00 TtTt NOro»i% of pop.lation aj % of elderly a-e _00 O ON I-1ts qro tso tso —— •*O — ioo r-TTts pinvo Ovo — NOO m vo •*00oo ooin vo ONJJJNO vqin rooo£JO uC «£• 2S >vI =11 •S H.oj U 00 T3110 H Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Pvl t= O m "^ "p "in ts -;CN 06 •* vo -^ o r- r- vo CN in m invo"u **^^ CN t^^ m vo m r*-— r- CN — CN d r-r^ Tt CN Tt CN Ttinoo CN vo in m — r-— vo ^ Tt 06 Tt tsr- ^i ^ ^ Tt -- Ttoo — • CN ™ r*; m r^- £ $ £J 3 £J TJCNoo r- in ov r- o ts-^ Tt -H o o « vdCNOO O Tt NO O CN — '— • Tt ^ ^^ vo c> om o O r-; oo vo r^ts CN ts vo m vo r-r*^ s CO inO vo O CN O ooTt 00 2 Tt — Tt —oo Tt — • r- m oin Tt °° Tt C*S Tt OOp vq ^ CN O^ Tt O;m r^t n r-*uv *^8± -82 |8§ 2^0.0 r", o i >-i td — ^* So 0 t-j r o a + ua r- c >,Ulllf |Ii!iillH^^ V^ PN C*i ir i "^ OM ^ r* ^ rs vb on^ oo ^l m m m vo™40N NO - Tt vo NO— O SS Tt C> C> TfOO J5v CN Tt CN T>^^ to — 0N m —^H Ov « i/S ^" CN Or- vo m vo m vo- pi ?2S3» ?r- Tt mr- CN m— NO & ON — OO OOr- ° •* mOO CO Tt CO ^3 NO^^ in '/^ oo CN r- ONr- tVj Tt mOO CO ^ CN — f; NO— • — r*-' r- i— o rir- — - m inm 06 ?J vi d m. ^^vo ^ m r-in in 1 ^ m oo oin Tt CN m oi m d vo vo fi ooTt Tt r-; Tt CN 0vd o 3 m ve i Tt m vo ^ m r-oin-aporeTotal fertility rateb Lifesexpectansy at birth # permens a-e rt0+ ( m e s_0; % of pop.lation a-e rts+ % of elderly a-e _+ Old-a-e dependemey ratio Total dependemey ratio .sCA I.i III Q OS 5 JT -e t S.».f •adS •-.as. S.1 5 ( S5 !««' fajs 111? 3l.2f Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 48 The Well-Being of the Elderly in Asia through the early decades of the 21st century. Singapore, Taiwan, and Thai land have already experienced some aging of their elderly populations be tween 1970 and 1995, with the increase in the percent of elderly age 70-plus being most marked in Singapore and least in Thailand. Although the elderly population in the Philippines has aged little during the 20th century, it is pro jected to begin aging quite rapidly beginning in 2020. Two of the dependency ratios presented earlier for several regions are also presented in Table 2.4 for the four countries under study. These are the old-age dependency ratio, which relates the population aged 65-plus to the working age population (15-64), and the overall dependency ratio, which re lates the population under age 15 and aged 65-plus to the working age popu lation. The changes in age structure that are taking place in the study coun tries translate into changes in both dependency ratio measures. The faster growth of the older population relative to population in the working ages leads to increases in the old-age dependency ratio. Although the ratio was fairly low in all four countries up to 1995, by the end of the projection period in 2030 it has increased substantially in each, although to varying extents. On one extreme is Singapore, where by 2030 the older population is anticipated to be about 36 percent the size of the working-age population. On the other extreme is the Philippines where the older population is anticipated to be only 11 percent the size of the working-age population. Changes in the total dependency ratio do not necessarily follow those for the old-age dependency ratio, because recent fertility trends have a large impact on the number of persons under age 15. In all four countries the total dependency ratio has fallen during the course of the recent fertility declines. So far, lower fertility has more than compensated for the rapid growth of the elderly. Thus, the overall dependency ratios in 1995 are substantially below those of several decades earlier. However, in all but the Philippines, the fertil ity declines appear to have plateaued and are projected to remain low and relatively stable, which means that the anticipated continued growth of the older population will cause the overall dependency ratios to rise in Singapore, Taiwan, and Thailand. Within the timeframe shown, this will occur first in Singapore and last in Thailand; only the Philippines, where fertility decline lags, will fail to experience a rise in the ratio. Even given these increases, it is worth noting that the overall dependency ratios projected for all four coun tries will be lower than the ratios these countries experienced in 1950, prior to the onset of their fertility decline. As discussed in Chapter 4, these past trends and future projections have a number of important implications for under standing past economic growth trends and possible future trajectories. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Demographic, Socioeconomic, and Cultural Context 49 Socioeconomic Change The demographic trends of the past decades that have precipitated unprec edented increases in the elderly population have been accompanied by rapid social and economic changes with potentially profound implications for the circumstances under which the future elderly will live. Many of these changes are anticipated to alter the patterns of familial support for the elderly that have traditionally prevailed in much of the world, including the four countries un der study. The forces most commonly cited as potentially undermining famil ial support include the increased economic activity outside the home by women (the predominant caregivers); the physical separation of parents and adult children associated with urbanization and increased migration; and ideational changes, especially Western-style individualism spread through mass media and public education (Mason, 1992; Martin, 1989 and 1990; Caldwell, 1982). At the same time, some believe that rising incomes will permit adaptations to past arrangements that might leave both generations of elderly parents and adult children better oile For example, higher incomes might allow a shift in living arrangements from literal coresidence to separate but nearby dwell ings, thus continuing to meet the same needs of both generations but enhanc ing their privacy as well. The economic crisis that began in mid-1997 and swept much of the region notwithstanding, all of the study countries except the Philippines have experienced rapid economic growth over the past few decades. As shown in Table 2.5, the gross national product (GNP) grew by more than 5 percent per year on average in Singapore and Thailand and more than 8 percent per year in Taiwan during the quarter century between 1970 and 1995. Only in the Philippines has growth been rather stagnant. By 1995, however, substantial differences are evident among the four countries in the level of economic progress achieved. The per capita GNP in that year (in U.S. dollars) varied from just over $1,000 in the Philippines to almost $27,000 in Singapore. Al though per capita GNP in Taiwan is only about half that of Singapore, both countries have become newly industrialized and enjoy levels of income far above those of either the Philippines or Thailand. To some extent the differ ences in average income as measured by per capita GNP are moderated by differences in purchasing power, as indicated by the purchasing power parity indices. In both the Philippines and Thailand, relatively low costs of living increase their purchasing power, while in Singapore the opposite holds. Nev ertheless, substantial differences remain among the four countries even after purchasing power is taken into account. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 50 The Well-Being of the Elderly in Asia The countries also vary substantially in terms of the other social and economic indicators provided in Table 2.5. In all, the percentage of the labor force in agriculture decreased between 1970 and 1990, although in the city state of Singapore agriculture was of negligible importance even at the earlier date. As with indicators of per capita economic levels, the Philippines and Thailand differ considerably from Singapore or Taiwan in agricultural labor, with the former countries maintaining a significant portion of their labor force in agriculture compared to a decline to low or no proportion in the latter countries. The Philippines, Taiwan, and Thailand have experienced substantial increases in the level of urbanization over the last few decades, reflecting net rural-to-urban migration as well as rural-to-urban reclassification of some areas. Nevertheless, the three countries vary substantially with respect to their present levels of urbanization, which correspond roughly to the percentage engaged in agriculture. Although definitional differences also affect such figures, it is clear that at present Thailand is the most rural of the three with only a minority of the population in cities and towns, the Philippines is more intermediate, and Taiwan is already substantially urbanized. Definitional differences create problems for comparing the four coun tries with respect to women's participation in the labor force. Nevertheless, in all but Thailand, where the participation rates are by far the highest, Table 2.5 shows a trend towards increased labor force participation among women. In 1970, about one-quarter of women in Singapore and one-third of those in the Philippines and Taiwan were in the labor force; by 1995 the portion had risen to about half in all three countries. Although all four countries have seen rapid increases in education in recent decades, they vary in terms of the secondary school enrollment ratio, a measure that relates the number of secondary school students to the entire population within the relevant age range. In 1995, the proportion of eligible children enrolled in secondary education was 96 percent in Taiwan, more than three-quarters in the Philippines and Singapore, and more than half in Thailand. It should be noted that, in all four countries, primary school atten dance is virtually universal and has been so for several decades. With regard to the average age at which men and women enter their first marriage, the trend in all four countries has been towards a rising age at marriage, particularly for women. Increases in women's age at marriage have been particularly pronounced in Singapore and Taiwan. In none of the coun tries is early marriage prevalent; most women currently wait until the early or mid-twenties and men often until the latter twenties to first marry. Finally, Table 2.5 presents two indicators of health service availability, namely the ratio of the total population to the number of practicing physicians and to the number of hospital beds available. In all countries except the Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Table 2.5. Key Social and Economic Indicators, 1970-95 Annual percent growth in GNP, 1970-95GNP per capita 1995 (US$)Purchasing power parity 1995 (US$)Percent of labor force hiagriculture19701990Percent living hi urban areas 1995Percent of females (15 andover) active in labor force19701980 1995Gross enrollment ratio forsecondary school enrollment (as % of eligible age group) 197019801995Mean age at marriage Males19701990Females19701990Health indicatorsNumber of people per physician 19801993Number of people per hospital bed 19801993Philippines0.61,0502,850584533 543548C 4946647925.426.822.823.87,8488,273589780Thailand5.22,7407,540806413 69d17295524.726.022.023.56,8034,416651765Taiwan8.412,3%14,700"371338b 58"3639 4554809628.829.822.826.51,516892446208Singapore5.726,73022,77030100 50465876e27.830.024.227.01,111714239275 Sources: World Bank, World Development Indicators 1997; World Bank, World Development Report, various years; United Nations, World Urbanization Prospects: The 1996 Revision; International Labour Office, Yearbook of Labour Statistics; United Nations, Demographic Yearbook, various years; Statistical Yearbook of the Republic of China, various years; Central Intelligence Agency, World Fact Book, 1997; Yearbook of Labor Statistics, Republic of China, various years; Taiwan-Fukien Demographic Fact Book, various years; Taiwan Statistical Data Book, 1996. 1 Figure for 1996. b % living in localities of 100,000 or more. c Interpolated between 1978 and 1983. d Adjusted from data on 13+ population. c 1997 (data from World Development Report, 2000-2001). Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 52 The Well-Being of the Elderly in Asia Philippines, the availability of physician services improved between 1980 and 1993, as evidenced by a decline in the ratio of people per physician. Availability of physician services is highest in Singapore and Taiwan, and considerably lower in Thailand and the Philippines. Availability of hospital beds shows a similar pattern across countries, with availability in the latest period being highest for Taiwan (ratio of 208) and Singapore, and lowest in Thailand and the Philippines. Availability of hospital beds deteriorated somewhat in all countries except Taiwan between 1980 and 1993. Changing Characteristics of the Elderly As a result of the demographic, social, and economic changes that have taken place in the four countries, the elderly themselves have changed in important ways. Table 2.6 illustrates some of these changes. The improvements in health, which are reflected in increases in life expectancy at birth and increasing proportions surviving to reach the elderly age span, have also meant that the elderly themselves are living longer. Life expectancy at age 60 has increased between 1970 and 1990 in all four countries for both men and women. Im provements in old age mortality during this period are particularly pronounced in Singapore and Taiwan. In Singapore, in just two decades, at least five years have been added to the length of time a 60 year-old can be expected to sur vive. These improvements in old age mortality are contributing to the aging of the elderly themselves, as evidenced by the increases in the proportion of the elderly aged 70 or above, as was shown in Table 2.4. Improving mortality has also reduced the proportion of elderly who are widowed, although large gender differences persist. The declines in widowhood have been greatest in Taiwan and among women in Singapore. In Thailand and the Philippines, where mortality improvements at old age have been more modest, so have the reductions in widowhood. The spread of mass education earlier in the century is evident in the recent decreases in the proportions of elderly who are illiterate. In 1970, ap proximately half of men aged 60 and over in all four of the countries were unable to read or write. By 1990, illiteracy among elderly men had dropped to about one fifth in all of the countries. In contrast, in 1970 more than 90 per cent of women aged 60 and over were illiterate in all but the Philippines (where 70 percent were illiterate). By 1990, illiteracy among elderly women was also sharply lower in the Philippines and Thailand, but the declines were more modest in Taiwan and Singapore where about two thirds remained illit erate. Since literacy is generally acquired early in life and remains relatively fixed, these reductions primarily reflect a process over time whereby older Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 o ON 0l T3 •o NO e8 e. •o _u Q. E2.ct/5|•aH•OCR)HiQ.—a,10|>cus10B2co*o2c£:11IndisatorvO ON^i/-, oo — • CO r*- "t _ , — r r-. o oo r*-— * ^^ rs vO in i0 ^^ oo vOM^^ C^) ^^rM iO C** r^ •* vO ov lO ^^ CO ^^ ^^ON..r~ONO -^ oo tN — ON m -0r-O— '— iCN NOTtTt ONr^NO -^^^ ^^ ^^•^ vo r^ e^ vo ^ ts in r^ Tt* o vo — — _, r, _ _ inmts ttTtmrNjeoo feONo6^^ mtsoo ^•*on^OO O— itN^ ^^ tN1 lO lO ^ Qv r^ tf1 CO CO CO^ oo t*-•n in r** NO i*^ vO in ^^ oo vO r^ ro^ _ _ „, „ Tt Tt - ...*? B Pr*- g ON . "i rs o ^ oo oo oo r*-o; . q-HC— rt-H— .5tN .NO.s 1| _ fic®BsSi-H-H—" U -^ — * — e U -^ -^ -^ U — ^ -^ -— 24* V 4* a. a. a. CO CO CO oo ON 06 S oo •S « ~ t, H ON NO o NO s i §1 2 S rt Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 54 The Well-Being of the Elderly in Asia and less literate elderly people die off and are replaced by younger, more literate elderly. The gains in literacy shown in Table 2.6 reflect the expansion of educational opportunities, and the gender differentials in opportunity, dur ing the times that current cohorts of elderly were in their childhood and youth. This expansion of educational opportunities has continued (as shown by the secondary enrollment figures in Table 2.5) and will greatly affect the educa tional level of future cohorts of elderly, as discussed by Christenson and Hermalin, 1991, and in Chapter 12. Labor force participation among the elderly has also changed in the four countries. As noted by Raymo and Cornman (1999), Singapore and Tai wan tend to resemble one another in terms of trends in labor force participa tion, as do the Philippines and Thailand. In Singapore and Taiwan, modest but clear-cut declines in economic activity among elderly men are evident between 1970 and 1990. Given the non-agricultural nature of the economies in Singapore and Taiwan, these declines may resemble the almost universal trend toward earlier retirement observed throughout the West (Quinn and Burkhauser, 1994). In the far more agriculturally oriented Philippines and Thailand, the trends for men are less clear although some reduction between 1980 and 1990 is apparent. Among older women, economic activity rates are substantially lower than for men in all four countries, and the trends over time are mixed. (For more detail, see Chapter 7.) Demographic and Social Characteristics of the Sample Elderly Table 2.7 presents basic demographic characteristics of the population aged 60 and over based on four recent surveys of the older populations described in Chapter 1. Overall the age structure of the samples are relatively similar. In all samples the younger elderly predominate. Close to or just over 60 percent of the elderly in each sample are in the 60-69 age range and only about a tenth are 80 or older. Also women in each sample tend to be somewhat older than men, reflecting their higher life expectancies at older ages. All four countries have substantial differences in the marital status dis tributions between men and women, reflecting the combined influence of higher male than female mortality, the tendency for men to marry women younger than themselves, and the greater likelihood for older men than women to remarry following widowhood or divorce. The proportions of men cur rently married differ only modestly across the four countries: approximately three-fourths of men aged 60-plus in Philippines, Singapore, and Taiwan are married, compared to over four-fifths in Thailand. Greater contrasts are evi dent in the proportion of women who are currently married, which ranges Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Ia 71 £I £s^^1n•^• og_op*0 ec S.-3 ccu•0(2C/IB a.D. U B- ia o£1BUvq •* p in m —r- in cc Tf CN CNin m —•1SSr- i; c^ob ri oo irl r^im r^ Ooo — Ov^ o — •ON oi 06 v^ f^O wn irlts t--^ ovo ts —o CO f~Sob ^ (S —m Vi —S tsin oo r-vd ri din ro i— •vo r- t~ Tf rS —in ro — .p irl inin r^S^ r- j.,.___ ** « r- oo<O OO OO -t" CO O •* ^*in *rCO Ov NO COCN o in -^oo oo ^^ inCO CO O —-r- CN•^ Tt ^^ CNCO P-^ NO CN NO CNNO •* 00 CNo vd — — m •*oo r- on^ ^- iri vd Tt m r- —CN CN ON r-CN CN — CONO COin ^^ in —m W vd int-: p m pO CO on^ CM00 —in p r-: oqCN on^ O CNm ^vq ^^ p inCO Ov CO COm inOv vO NO OvO CO CO ^r- CN•oT3 4iu y-a.E §3 -8 1 1S 1 >,T3 ^s !?ss3 S t€ 1. •g « 3 i * SoN •* r^NO CN°°. TJ! ^ — **S ^oo —r- ov TJ'Ov Tt int^ ^ i^iO co C•^ CO Mf% CN inS — oo mCN ^ mo ^ ^— r* i Om vo-T r i ^f^ IS ^NO iO Ov00 — d— r-e* P ^ON —* ON- NO -NO NO OO Ov CN iS— .^ —Oi OO CN— m CsI 1 s •§^ -5•8 £r§EducationNo formal Any prima Secondary888Ov CN OvOv Ov Oin CN —•* •» CNOv O Oin m —•m tN inO 06 -^ NO CN —m o r-00 CN OvooTt vo —Ov O O oo ^-— vo COS CO Ovoom vo —•* vd Ovm inm o r-vd m ovo in ^^mS*>^Urban^*ruralUrban Rural Capital sit;8oCOCN1S00in00!NO^f00t^fg--^,C^l-oCNSample size Si Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 56 The Well-Being of the Elderly in Asia from a low of only slightly above 30 percent in Singapore to over half in Taiwan. In each of the countries only a small percentage of older persons never married. The highest proportion never married is found among men in Taiwan, owing largely to the migration of the nationalist soldiers in the late 1940s, many of whom never married. This migration is also responsible for an aberrant gender distribution in the Taiwanese elderly population, for which males outnumber females. Few elderly are separated or divorced in any of the four countries; the highest occurrence (5 percent) is found among women in Thailand. In all but the Philippines, substantial differences in the educational distributions of older men and women are evident, reflecting historical differ ences, which favored schooling for sons over schooling for daughters. The proportion of the older population with no formal education is highest in Singapore, where half of the men and over four-fifths of the women aged 60 and over have had no schooling. In Taiwan, the lack of formal schooling is also very high for women, among whom three-fifths have had no education. The Philippines stands out both because only relatively minor gender differ ences characterize educational distributions and because only a modest mi nority of older men or women received no schooling. Among those who re ceived schooling, it has generally been limited to the primary level. In all countries the proportion of older persons receiving secondary or higher edu cation is modest, ranging from about 20 percent in the Philippines and Tai wan to less than 10 percent in Thailand and Singapore, but substantially lower for women than men everywhere but the Philippines. Indeed in all countries but the Philippines, less than 10 percent of the current cohort of older women received secondary or higher education. The distribution of elderly with respect to urban and rural residence largely mirrors that of the population as a whole. In Thailand, the most rural of the four countries, over four-fifths of the elderly population are living in rural areas. (It should be noted that the definition of urban in Thailand is relatively restricted and to some extent the high rural proportions are an arti fact of this definition.) In the Philippines, the majority of the elderly popula tion also live in rural areas, while in Taiwan less than a third do so. (Singapore's entire population is urban.) Despite these differences among the three coun tries, the proportion of elderly living in capital cities is relatively similar. Thus the main differences reflect the proportions living in provincial urban areas. Table 2.8 indicates the distribution of elderly on the same characteristics used in Table 2.7, but by age rather than gender. In all four countries substantial differences in the marital status distributions are evident between the population in their 60s and those who are 70 or older. Reflecting the increasing toll that mortality takes on marriages, the proportions currently married are substantially Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 I •5. Q. to *o 1HV§. +1°00vo8-£a 2 +a Ssis30HI +•ee• Op 'ONvo8311 --\s r*-1 -OvNO8O 00 00 Ttri d 3 -^•r^ ^- 00 P^m vie•J* OO ON C^m s s —•* ** ON CNm r-^ io riiO *NO w^ ^ C^iCO ^ •"• *"*oo in 01 ONm vb vb r^ir- —ts rs ^^ t~ri *N -^ mNO COvp to p O•~* oo r*- fi•* on^-io f^ vq —CN O rs on^r- CNm o r- oooi •*t d ts vrt Ttvo ^- o fi•*t ino r- co ^^ -* — CN>o m_1 1Marital statusNevrt married Crtrently marr Widowed Separated/divoov Tt r-MOv — Ovp O; vooo in r-;Ov vo md oo dTt CO CN00 O tNr-i ^^ oo in CN —in Tt — •ts in tsm Tt tsO NO Tt— r i vCm voi-. vp r-;d fi inm Tto; m ooOO Tt NO-^ I-••Tt ONO23255 —ts in —oo ro O— NO CNs >1 1Ed.sationNo formal edu( Any primary Semendary and888Ov CN OvoJ ov d in CN — *Tt Tt M00 ON tNin CN —Ov — Od ov d""CO O t**OO CN ON00t*l NO -—06 d —oo —fl 00 ON00 tN OO 00m vp —as°*- vp r^m mCN o inri t- d d m —Urban^*ruralUrban Rrtal Capital sityiTt°31Tt00"1flOv2m3,»TtTtgvi~*£vqCN—***PinTtr-Sample size Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 58 The Well-Being of the Elderly in Asia lower and the proportions widowed substantially higher for the oldest-old (70-plus years) in all four countries. Only in Taiwan is slightly more than half of the population age 70 and older still currently married. We also find age-based differences in educational distribution. Reflect ing the trend toward increasing education, those age 70 years or older are more likely to have no formal education and less likely to have completed secondary schooling than the elderly in their 60s in all four countries. Differ ences are particularly large in Thailand, where rapid changes in the provision of formal schooling corresponded to the time when the current elderly were reaching school ages. Thus, more than half of those age 70 and above had no education in Thailand, compared to less than 20 percent of those in their 60s. Table 2.9 presents information on the availability of various types of living kin relative to the elderly respondents in the survey. These kinds of data are important because exchanges between elderly and kin of different generations are of course constrained by the availability of such relatives. Between half and two-thirds of respondents have a living spouse. Whether or not an elderly person has a spouse has important implications for their well- being, in terms of both potential care giving and care receiving. This in turn affects the extent of dependency of the elderly upon other persons within the household or within the kin network. Only a minority of respondents have a living parent, which is not sur prising given the sample composition (aged 60 and above). The proportion of respondents from the survey in the Philippines that have a parent is distinctly larger than in Taiwan or Thailand; information on Singapore is not available. Over 95 percent of the elderly in all four countries have a living child and about 90 percent have at least one son and one daughter. Given the age of the respondents, most of their children are already adults (defined as ever- married or age 18 or over) and therefore the percentages with an adult child, adult son, or adult daughter, are only slightly lower than the percentages who have any living child, son, or daughter. Because the children are usually adults, most of them are also married, which means that the percentages of elderly with a married child tend to be almost as high as the percentages with any child. At the same time, in each of the three countries for which data are available, two-fifths or more of the elderly still have a single adult child. Given that most of the respondents' children are already married, most of the respondents have grandchildren. In the three countries with available data, approximately 90 percent have at least one grandchild. Reflecting the fairly large families from which many elderly respondents came, many still have a living sibling. In the Philippines almost all have a living sibling, while in Taiwan and Thailand more than 80 percent do (data are not available for Singapore). Additional details of family structure are presented in Chapter 5. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Demographk, Sodoeconomic. and Cultural Context 59 Table 2.9. Percent of Respondents with \ arious Tvpes of Lrung Kin Note: Adult children include any ever-married child of any age and any single child age 18 plus - = not available. Conclusions PhilippinesSpouseParentChildSonDaughterA dult childAdult sonAdult daughterSingle child 18+Single son 1 8+Single daughter 18+Married childMarried sonMarried daughterGrandchildSibling54.113.696.189.991.794.988.189.839.127.532.892.682.284.992.597.7Thailand62.68395.688s989.4Taiwan67.47.795.890.887.195.5 94.388.789.247.831.928.893.282.783.892.383.289.484.439.530.418.091.181.080.489.882.8Singapore50.8-95.789.086.7__-———mm——-1 This chapter provides context for the analyses that follow by providing dcmo graphic, socioeconomic and cultural detail about each of the four study coun tries and their regions. The initial section compares the major demographic trends and projections for East and Southeast Asia, as well as the four coun tries, with Europe and the United States so that the pace of population aging and its distinctive features may be placed in wider perspective. Over the next 30 years East Asia and some of the countries in Southeast Asia will converge to the aging profiles developing in Europe and the United States. The chapter also highlights some aspects of the cultural and political setting of each country included in this study as well as of the major social and economic changes that have occurred over the last 40 to 50 years. The fact that all countries are situated in Southeast and East Asia narrows to somc extent the range of cultural differences that characterize them, but considerable variation still exists. All share a strong sense of filial piety stemming from a socially instilled sense of reciprocity children feel toward their parents (Asis et al., 1995). Yet differences exist in terms of the views of the appropriate Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 60 The Well-Being of the Elderly in Asia roles of sons and daughters in carrying out these morally grounded obligations towards parents as explored in Chapters 5 and 6 (see also Ofstedal et al., 1999). Such differences reflect broader cultural differences that have long and deep historical roots and that correlate with major ethnic boundaries and religious beliefs. Christianity prevails in the Philippines where the vast majority profess Roman Catholicism. In contrast, 95 percent of the Thai population are Buddhists. Singapore is unusual in being an explicit tripartite multi-cultural society with religion and ethnicity closely intertwined: Buddhism predominates among the majority ethnic Chinese, Islam among the ethnic Malays, and Hinduism among the ethnic Indians. In Taiwan, over 90 percent of the population are Buddhist, Confucian, or Taoist. The circumstances for the elderly respondents at the time of the sur veys were strongly influenced by the major demographic and socioeconomic changes that occurred in all countries during their lifetimes, and particularly over the preceding several decades. Fertility declined, the population aged, economies grew, labor markets shifted, and educational opportunities in creased. Although in all four countries under study these changes have been rapid and profound, they have differed to some extent in terms of their pace and origin. As a result, the circumstances for the elderly of the Philippines, Thailand, Taiwan, and Singapore have common and dissimilar features. This combination provides an intriguing context for the analyses that follow. While this chapter has focused primarily on past changes and their impact on the current cohorts of older persons in each country, it is important to note that future cohorts of elderly will look much different. For example, as was touched on here and will be elaborated in Chapter 12, the educational distributions for the elderly will change precipitously, with the proportion of elderly who have no education declining to a negligible level and the propor tion having completed secondary or higher schooling increasing substantially. In addition, recent declines in fertility mean that future cohorts of elderly will have substantially fewer children than the current elderly. These as well as other anticipated socioeconomic and demographic changes will influence the social and economic roles of future elderly, their family lives, and the types of policies and programs best suited to their needs. REFERENCES Agoncillo, Teodoro A. 1990. History of the Filipino People, 8th edition Quezon City: Garotech Publishing. Asis, Maruja Milagros B., Lita Domingo, John Knodel, and Kalyani Mehta. 1995. "Living Arrangements in Four Asian Countries: A Comparative Perspective," Journal ofCross-Cu Itural Gerontology 10:145-162. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Demographic, Sodoeconomic, and Cultural Context 61 Barclay, George W. 1954. Colonial Development and Population in Taiwan. Princeton: Princeton University Press. Bloom, David E., and Jeffrey G Williamson. 1998. "Demographic Transitions and Economic Miracles in Emerging Asia." World Bank Economic Review 12(3): 419-456. Caldwell, John C. 1982. The Theory of Fertility Decline. New York: Academic Press. Castillo, Gelia. 1977. Beyond Manila: Philippine Rural Problems in Perspective. Ottawa, Canada: International Development Research Center. Christenson, Bruce, and Albert I. Hermalin. 1991. "Comparative Analysis of the Changing Educational Composition of the Elderly Populations in Five Asian Countries: A preliminary Report." Research Report No. 91-11, Population Studies Center, University of Michigan. Coale, Ansley J. and Edgar Hoover. 1958. Population Growth and Economic Development in Low-Income Countries. Princeton, NJ: Princeton University Press. Cohen, Myron L. 1976. House United, House Divided. New York: Columbia University Press. Cornman, Jennifer C. 1999. "Understanding the Ties That Bind: Intergenerational Value Agreement in Taiwan." Unpublished Ph.D. dissertation. Ann Arbor, Michigan: Population Studies Center, University of Michigan. Freedman, Maurice. 1958. Lineage Organization in Southeastern China. Lon don: University of Athlone Press. Fricke, Thomas, Jui-Shan Chang, and L. S. Yang. 1994. "Historical and Ethno graphic Perspectives on the Chinese Family." In Arland Thornton and Hui- Sheng Lin, eds. Social Change and the Family in Taiwan, 22-48. Chicago: University of Chicago Press. Hermalin, Albert I. 1976. "Empirical Research in Taiwan on Factors Underlying Difference in Fertility." In Ansley J. Coale, ed., Economic Factors in Popula tion Growth, 243-266. New York: John Wiley & Sons. Hollnsteiner, Mary. 1973. "Reciprocity in the Lowland Philippines." In Frank Lynch and Alfonso de Guzman, eds., Four Readings on Philippine Values. Quezon City: Ateneo de Manila University Press. Johnston, Mark. 1999. "Population Ageing and Development: Implications for Asia and the Pacific." In United Nations, Promoting a Society for All Ages in Asia and the Pacific. New York: United Nations. Jones, Gavin W. 1993. "Consequences of Rapid Fertility Decline for Old Age Security in Asia." In Richard Leete and Iqbal Alam, eds., Revolution in Asian Fertility: Dimensions, Causes and Implications. Oxford: Clarendon Press. Kaplan, Irving. 1981. "The Society and Its Environment." In Frederico Bunge, ed., Thailand: A Country Study. Washington, D.C.: U.S. Government Print ing Office. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 62 The Well-Being of the Elderly in Asia Knodel, John, Napaporn Chayovan and Siriwan Siriboon. 1992. "The Familial Support System of Thai Elderly: An Overview." Asia-Pacific Population Journal 1 (3): 105-126. Knodel, John, and Chanpen Saengtienchai. Forthcoming. "Studying Living Ar rangements of the Elderly: Lessons from a Quasi Qualitative Case Study Ap proach in Thailand." Journal ofCross-Cultural Gerontology. Knodel, John, Chanpen Saengtienchai, and Werasit Sittitrai. 1995. "The Living Arrangements of Elderly in Thailand: Views of the Populace." Journal of Cross-Cultural Gerontology 10:79-111. Knodel, John, Rossarin Soottipong Gray, Porntip Sriwatcharin and Sara Peracca. 1999. "Religion and Reproduction: Muslims in Buddhist Thailand." Popula tion Studies 53(2): 149-165. Lee, Li-Ju. 1999. "Law and Social Norms in a Changing Society: A Case Study of Taiwanese Family Law." Southern California Review of Law and Women s Studies Spring 1999. Lopez, Maria E. 1991. "The Filipino Family as Home for the Aged." Research Report No. 91-7, Population Studies Center, University of Michigan. Lynch, Frank. 1973. "Social Acceptance Reconsidered." In Frank Lynch and Alfonso de Guzman, eds., Four Readings on Philippine Values. Quezon City: Ateneo de Manila University Press. Martin, Linda. 1989. "Living Arrangements of the Elderly in Fiji, Korea, Malay sia, and the Philippines." Demography 26:627-643. . 1990. "Changing Intergenerational Family Relations in East Asia." The Annals 510:102-114. Mason, Andrew. 1987. "National Saving Rates and Population Growth: A New Model and New Evidence." In D. Gale Johnson and Ronald D. Lee, eds., Population Growth and Economic Development: Issues and Evidence, pp. 523-560. Madison, WI: University of Wisconsin Press. Mason, Andrew, Sang-Hyop Lee, and Gerard Russo. 2001. "Population Momen tum and Population Aging in Asia and Near-East Countries." East-West Cen ter working papers. Population Series No. 1076. Honolulu: East-West Cen ter. Mason, Karen. 1992. "Family Change and Support of the Elderly in Asia: What Do We Know?" Asia-Pacific Population Journal 1 (2): 13-32. Medina, Bclen. 1991. The Filipino Family: A Text with Selected Readings. Diliman, Quezon City: University of the Philippines Press. Ofstedal, Mary Beth, John Knodel. and Napaporn Chayovan. 1999. "Intergenera tional Support and Gender: A Comparison of Four Asian Countries." Research Report No. 99-54, Population Studies Center, University of Michigan. Podhisita, C. 1985. "Buddhism and the Thai World View." In Amara Pongsapich, etaLeds., Traditional and Changing Thai World Hew. Bangkok: Social Sci ence Research Institute, Chulalongkorn University. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Demographic, Socioeconomic, and Cultural Context 63 Prachuabmoh, Visid, John Knodel, Suchart Prasithrathsin, and Nibhon Debavalya. 1972. "The Rural and Urban Population of Thailand: Comparative Profiles." Research Report No. 8, Institute of Population Studies, Chulalongkorn Uni versity, Bangkok. Quinn, Joseph F., and Richard V Burkhauser. 1994. "Retirement and Labor Force Behavior of the Elderly." In Linda G Martin and Samuel H. Preston, eds., Demography of Aging. Washington, D.C.: National Academy Press. Rabibhadana, Akin. 1984. "Kinship, Marriage and the Thai Social System." In Aphichat Chamratrithirong, ed., Perspective on the Thai Marriage. IPSR Publication No. 81. Bangkok: Institute for Population and Social Research, Mahidol University. Raymo, James M., and Jennifer C. Cornman. 1999. "Labor Force Status Transi tions at Older Ages in the Philippines, Singapore, Taiwan, and Thailand, 1970- 1990." Journal of Cross Cultural Gerontology 14:221-244. Schulz, James H., Allan Burowski, and William H. Crown. 1991. Economics of Population Aging: The "Graying" of Australia, Japan, and the United States. New York: Auburn House. United Nations. 1999. World Population Prospects: The 1998 Revision. Volumes I and II. New York: United Nations. Wolf, Margery. 1968. The House ofLim: A Study of a Chinese Farm Family. Englewood Cliffs, N.J.: Prentice-Hall. World Bank. 1994. Averting the Old Age Crisis. New York: Oxford University Press. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Chapter 3 POLICIES AND PROGRAMS IN PLACE AND UNDER DEVELOPMENT Mary Beth Ofstedal, Angelique Chan, Napaporn Chayovan, Yi-Li Chuang, Aurora Perez, Kalyani Mehta, and Albert I. Hermalin The purpose of this chapter is to provide an overview of public programs designed to assist the older population in the four countries that are the focus of this volume: the Philippines, Thailand, Taiwan, and Singapore. Emphasis is placed on social security and health care programs, though examples of other programs will be touched on as well. There are several reasons for monitoring the development of policies and programs in Asia and other developing regions. From the relatively nar row perspective of this volume, knowing the nature of the policies and pro grams in place is a valuable context for the specific analyses to be presented on health, work and retirement, economic well-being, living arrangements, and intergenerational support. In turn, the situation of the older population on these dimensions is relevant input for further program development. Accord ingly, to the extent possible, we will discuss both programs currently in place as well as those under consideration. From a broader perspective, attention to program development related to population aging in these countries provides a rich opportunity to study an important component of social change in the making. Policymakers in Asia and elsewhere are very aware of the potential consequences of population aging in terms of the pressures on traditional family arrangements for support Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 66 The Well-Being of the Elderly in Asia of the elderly, and of the high costs of many social welfare programs in more industrialized countries. They are faced with several conflicting goals: they want to provide for the needs of their older population and anticipate the social and economic changes arising, but they wish to do so in a manner that does not undermine the traditional family arrangements currently providing substantial support and that is not so costly as to jeopardize future economic development. In a sense they are testing whether "modernization" inevitably means "Westernization" in terms of social welfare programs and their costs, or whether they can develop novel arrangements that better suit their cul tures, resources, and needs. Studying this process is not only of broad social science interest; it is possible that the programs that evolve will have implica tions and benefits for social welfare programs in more industrialized coun tries. Several examples of intriguing features built into the programs of these countries will be touched on below. An important backdrop to current policy development is to keep in mind that in most cases these countries are not starting from scratch. Although family support arrangements are key elements of the social fabric, these coun tries have had social welfare programs of various types and breadth for some time, and the structure of these programs and their perceived effectiveness often influence the programs under development. Many of the countries first developed programs for civil servants, military personnel, and even employ ees of state-owned enterprises. Over time new programs were developed for other classes of workers, often in a piecemeal fashion, with various consoli dations and amendments taking place as economies changed and programs matured. Few sources have attempted to summarize these multi-layered pro grams in a comparable manner, making it difficult to delineate them succinctly and with a standard set of criteria. Accordingly the material to be presented should be regarded as a brief and preliminary introduction to a complex topic. It is of interest to note that, in addition to specific programs, broad provisions or concerns about the older population have been included in a number of these countries' five-year plans, constitutional provisions, or gen eral legislation — despite the fact that rapid growth of this age segment is a very recent phenomenon. Often these considerations are coupled with reaffir- mations of the family's responsibility for care and support of the elderly. (For additional discussion of national policies on aging for countries in East and Southeast Asia see Phillips, 2000; and for a larger group of countries see Randel et al., 1999.) For example. Section 4, Article XV (15) of the 1987 Philippine Constitution states: "the family has the duty to care for its elderly members, but the State may also do so throughyw^/ programs of Social Security." And Section 11, Article XIII (13) states: "The State shall adopt an integrated and Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Policies and Programs in Place 67 comprehensive approach to health development which shall endeavor to make essential goods, health and other social services available to all the people at affordable cost. There shall be priority for the needs of the underprivileged, sick, elderly, disabled, women, and children." In Thailand the elderly are explicitly mentioned in the 1997 Constitu tion. Article 54 states that "persons who are 60 years old and over and who have insufficient income to maintain their living are entitled to receive assis tance from the State"; while Article 80 adds that "the State must provide wel fare for the elderly, the poor, the handicapped or disabled, and destitute per sons so they can have a good quality of life and be self-reliant." Chayovan (private communication) points out that elderly issues per se were not of high priority in most of Thailand's Five Year Plans that guide national development, usually including the elderly within a broader set of disadvantagcd, as suggested by the language above. But the current Five Year Plan (1997-2001) does include a section on the "isolated indigent elderly," as well as provisions for universal free health services for the elderly, discounted fares, and for encouraging and assisting families to understand and care for elderly members, among other measures. Thailand did draw up a "Long-Term Plan for the Elderly in Thailand" (1986-2001), following the 1982 World As sembly on Aging in Vienna, and revised it in 1992 to include some explicit provisions regarding welfare policies. Taiwan at an early date developed programs that subsidize organiza tions providing services as well as direct assistance to the elderly. It passed the Senior Citizen Welfare Law in 1980, which subsidizes programs that as sist elderly leisure activities and provides health examinations. The program, which covers those aged 70 or older, also provides or assists privately funded shelters for elderly who have no means for independent living. In 1993, new legislation provided monthly subsidies of US$120 to elderly in poverty (de fined as those 65 or older with monthly income below US$180 and who have no financial support from children). Policies concerning the older population in Singapore must be viewed in the context of a general philosophy that stresses individual responsibility, followed by reliance on family and the community. The government promotes the idea that respect for elders is an integral part of Asian culture and way of life, and the Maintenance of Parents Act passed in 1995 requiring children to support their parents by providing them with a monthly allowance has re ceived considerable attention within Singapore and elsewhere. By the end of 1998, 424 cases were filed and 328 elderly were successful in getting orders compelling their children to support them (Straits Times, 4th April 1999). At the same time, since the 1980s the government has set up a number of inter- ministerial committees to ascertain the social and economic consequences of Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 68 The Well-Being of the Elderly in Asia population aging for Singapore, and many policies and programs have been instituted to cater to current and future needs of the elderly. Based on the recommendations of one inter-ministerial committee established in October 1998, six working groups have been instituted with each assigned to deal with a particular aspect of aging, ranging from financial security, to housing and land use, to cohesion and conflict. (See Vasoo, Ngiam and Cheung, 2000, for a review of policies.) Social Security Retirement Programs Differences in the history and structure of social welfare programs for in come maintenance and health across countries, as well as different account ing and reporting provisions, make it difficult to capture the key elements of each country's programs in tabular form with comparable measures. Table 3.1 attempts to sketch the broad contours of the retirement provisions of the social security programs. Many of these programs also encompass provisions for sickness and injuries (to be discussed below), disability, unemployment, death, and other risks, which may be noted but not discussed in much detail. To supplement the information in Table 3.1, we provide a brief narrative about each country below. The Philippines* The Philippines has one of the oldest social security programs in Asia, with very wide coverage and a broad range of benefits. It is also beset with a num ber of financial and administrative problems that must be addressed to assure solvency and compliance. The social security aspect is two-tiered, with the first tier providing mandatory basic coverage of the defined benefit type, and the second tier providing voluntary supplementary coverage through private occupational pension plans. Coverage under the first tier comes from two sources. The Govern ment Security Insurance System (GSIS), created in 1936 and administered by the Department of Budget and Management, is a retirement benefit plan for selected government employees. Benefits under GSIS cover the following contingencies: retirement, separation, unemployment, disability, and death (through both a compulsory and optional life insurance feature). Retirement benefits are available to those who have 15 years of service and are at least 60 years of age. Compulsory age at retirement is 65 years. Those reaching age 65 with 35 years of service can obtain a pension of close to 80 percent of their last salary, up to a special wage ceiling. Retirees can claim benefits under two Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Policies and Programs in Place 69 basic options, each of which combines a lump sum plus a lifetime annuity. In 1997, about 94,000 people were receiving pensions from GSIS. The Philippines' broader Social Security System (SSS), created in 1954, covers those in the private sector and is of the defined benefit type, with contributions by the employee and employer, but with the government re sponsible for the solvency of the funds and as the guarantor of the mandated benefit levels. The SSS program also provides benefits for death, disability, sickness, and maternity. According to Asher (1998), although the SSS reported 16.6 million members in 1995, two-thirds of these are believed to be non- contributing members, i.e., workers who had not contributed during the prior year but had contributed for at least one month in the past. The World Bank's (1994) estimate that only 19 percent of the 1990 Philippine labor force con tributed to the SSS in the prior year supports this claim. In the same report, the World Bank estimates that only about 7 percent of those 60 or older in 1990 were pensioners. (Data from the 1996 survey, presented in Chapter 8, indicate that only 9 percent of older Philippine couples or individuals rely on pensions or retirement funds as their major source of income, supporting the World Bank's findings of low coverage.) Asher (1996) describes several problems facing the social security pro grams in the Philippines. Major among these is lack of compliance, as it ap pears that only 40 to 60 percent of the 500,000 employers in the SSS program collect and remit the mandated contributions. Asher asserts: The present social security arrangements in the Philippines are unlikely to be sustainable, and steps would need to be taken to adjust the benefit level (including rationalization and drastic reduction in loans to members), improve com pliance, enhance the administrative efficiency, enhance the rate of return on portfolio, and undertake organizational and institutional changes, including greater insulation of social security arrangements from short-term political consider ations. These would pose a formidable challenge to policy makers, (p. 83) Thailand2 Thailand has been developing its social security programs over many years, with legislation dating back to the 1950s and a number of substantial changes and additions since then. With regard to pension provisions, there are three major mechanisms: a well-developed program for government employees and officials; a program of employer-sponsored provident funds under government Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 1t*- 4l O CJiCJM05^ .S5s gQL,-^- CE|*7 U UCO B gS O US CQType of Plan§|JiS ooauto Q£Name of Basis o Mandatory Pro-ra>,EB§;1J—^^t -§voXaO00i Defimd benefioeo.uB EO2£. C/lGovernment Sesrtity Insrtanrt System (G:uaIsio||2 llfIPvo o— o/^ON i^e2^§t5^m E-oo ^*Or-Defimd benefi—IPrivate sector employes; me employed"BSosial Secrtity Syste (SSS)1£2O ^18v<) ^,EC 3v& Iflo o.° EJ3me. Defimd bemfit2. Blended*311Government wi: r-— "•*Civil Servant Plani11— <Provident F.n<> *•e °ex uEmployes in sector and stat enterprisesooProvident Fund Aste o0*uu•ADefimd bemfiPrivate sector employes^ o2 2Sosial Sesrtity Asti Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Cu uj S = . rtntinuedK?-C.2Ii3v WHf.S£|ii||on§1•oQeCCivil servants|1Ov ON— N —c1at§ si °*IllO 05•c^ ONILump sum (4_ months of wa-s)S11Employees in private sestor othrt than finanses servirts and me_rtrt|^^J1I|S<||g 8110.=Balamee in ordinary acme.nt'111All workrts exsept fme4ns partstimes sasual. Voluntary partisipation by self- employed.§e-*ea|•o|1gXSiKs1§1 «•8|1gide m e % of avera-e wa1S.aU1 1and years of servirt, fit and defined mentribution el fits to be paid in 2_14; srtrenta z s ill*V ^ "« •§ ic$ * 85 ^w A C£ a s•? .3 Si891 t a•a § 1 I 1 8..Sc u — Fills1 1 1|U C ^eu u K "^B. -o .5 -a, C u —• A U Ji S E « -o Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 72 The Well-Being of the Elderly in Asia regulation; and, most recently, a government organized retirement program for private sector employees. In addition there is a means tested social pension program for the elderly poor which is described below under programs providing economic assistance. The original program for civil servants, initiated in 1951, was com pletely financed by the government and provided a replacement rate of about 70 percent of last month's salary for those with at least 25 years senice at age 60. (Lump sum options were also available and government workers who retired with 10 to 25 years of service received only a lump sum.) The rela tively high proportion of the labor force in government service (approximately 7 percent) and the projected high cost for future retirees led to a revised pro gram in 1997 that is a blend of a defined benefit and a defined contribution plan. Under the new plan, employees contribute 3 percent of their monthly salary with a matching contribution by government. In addition, the govern ment makes a contribution of 2 percent to the fund (to compensate for the lower pensions) and those in the new plan will receive both a defined benefit pension based on years of service and a pension based on the defined contri bution accumulations and earnings. Under the new plan, those who worked in government senice for 35 years would receive 70 percent of average sal ary over the last five years of service, 15 years of senice are required to receive benefits, and lump sums only are available to those who have 15 to 25 years of service (Ramesh and Asher, 2000, Chapter 3; Asher, 1998). In 1987, Thailand directed private firms meeting certain criteria to es tablish provident funds for their employees. Under this program employees can voluntarily agree to contribute 3 to 5 percent of their salary, and employ ers must match or exceed the workers' contributions. In 1997, state-run enter prises were ordered to develop similar programs (Ramesh and Asher, 2000, Chapter 3). Although in theory this program covers a high proportion of the non-civil servant work force, its optional nature has limited coverage. As of 1999, about one million employees in 4,000 firms were participating, repre senting a small percentage of the labor force or registered firms. The most recent and ambitious component of Thailand's social secu rity program was undertaken with the passage of the Social Security Act of 1990. This act spelled out the intention to phase in a broad program of ben efits to all private sector workers in enterprises of 20 or more employees (reduced to 10 or more in 1993), including coverage for injuries, sickness, disability, death, unemployment, family allowances and old-age pensions, with the latter three benefits to be phased in at a later date and the others initiated in 1991. As of 1997, these provisions covered about six million workers, representing approximately 18 percent of the labor force but about 45 percent of wage and salary workers (Ramesh and Asher, p. 60). Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Policies and Programs in Place 73 The initiation of old-age benefits was delayed until 1999 but has now been instituted under the Social Security Office. Workers have to be at least 55 years of age with 15 years of covered service to collect pensions, so that the first pensions will not be paid until 2014. The amount of benefit is 15 percent of average monthly salary over the five years immediately prior to retirement plus 1 percent of salary for each additional year of service over 15. This means that a retiree with 35 years of service would receive a pension equivalent to 35 percent of average salary over the last five years of employment (Ramesh and Asher, 2000, p. 61). Funding for the program is provided by contributions from the employee (2%), employer (2%), and government (1%). These contribution rates are lower than originally planned and reflect in part the uncertainties arising from the economic crisis. A planned two-phase expansion of the program—first to employers with five to nine employees and then to firms with one to five employees—was accelerated in 2001, with coverage expanded to businesses with only one employee scheduled to become effective as of April 2002. This change would bring the number of covered workers to 9.6 million (Bangkok Post, November 21, 2001). Despite relatively frequent attention to social welfare issues for many years, it would appear that an adequate pension program for most of Thailand's workers is still somewhat in the future. Aside from Thailand's relatively large civil service sector, and those employees in state-owned enterprises actively involved in the provident funds authorized in 1997, the remainder of the la bor force has relatively little coverage at present. This is certainly true of the large number of employed in agriculture, the self-employed, and those in the informal sector, who currently have no coverage, but it also applies to the private workers in the formal sector. Only a relatively small number are par ticipating in the provident funds set-up by employers under government regu lation, and no retirees are scheduled to receive pension benefits under the Social Security Act until 2014, with the projection of wages covered quite low except for those with very long periods of service. Survey results confirm that old-age retirement income is still highly limited in Thailand, with only 2.3 percent of older couples or individuals reporting pension or retirement funds as their major source of income in 1995. Taiwan3 Taiwan, like a number of the other countries, has a well-developed old-age retirement program for civil servants and is in the process of updating and expanding its program for other workers. Under the Civil Servant Retirement Regulations (CSRR), first imple mented in 1943 and amended in 1992, employees can retire when they reach Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 74 The Well-Being of the Elderly in Asia age 60 or complete 25 years of service. The mandatory age of retirement is 65. The amount of the pension, based on a fairly complex formula involving length of service and wage equivalent, provides those with less than 15 years of service a lump sum, while those with more than 15 years have a choice of lump sum, a monthly pension, or some combination. Before 1992 the maxi mum monthly pension was 90 percent of the final wage-equivalent, but the 1992 amendment reduced that maximum to 70 percent. It also introduced an inducement for early retirement at age 55. The amended version also calls for employees to pay 35 percent of the premium, which is set at 8 percent of the "wage equivalent." Before 1992, there was no co-payment from employees. The nature of the calculation of the maximum pension means that workers with long tenure may be near the maximum at relatively early ages (around age 50) and have relatively little economic inducement to remain with the government thereafter. The influx of Mainlanders in 1949-50, their strong role in government service, and the rather generous CSRR pension system stirred interest on the part of other workers for retirement benefits. Shih (1997) reports that more than one-third of labor disputes in the early 1980s concerned pensions and workers' welfare. This led to the passage of the Labor Standards Law (LSL) in 1984, which governs labor contracts and regulates retirement practices in the private sector. It was built on the earlier Provincial Factory Workers Re tirement Act (PFWRA, 1957-1984), which covered production workers in factories owned by the Taiwan provincial government. The LSL covers manu facturing and non-manufacturing workers in the private sector, but excludes employees in the finance, banking, commerce, and service industries. Work ers can retire after 25 years of service, or at age 55 with 15 years of service. Pensions are awarded as a lump sum with a 45-month wage equivalent maxi mum. The pension is funded by employers who are required to set aside a fraction of their expenditures on salaries into the company's pension fund. The government of Taiwan is in the process of starting a new national annuity program, originally to be implemented at the end of 2000, but post poned due to uncertainty as to the appropriate funding mechanism. Accord ing to press reports as of November 2001, a number of different plans have been developed with major differences in the means of financing and the nature of the benefit structure, and no consensus has emerged as to how to proceed (China Post, 2001). Singapore Singapore is the only country in this group that relies on a broad-based compulsory retirement provident fund mechanism for income and related Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Policies and Programs in Place 75 benefits. The Central Provident Fund (CPF) covers about two-thirds of the labor force, excluding foreign, part-time, and casual workers, and self- employed workers who elect not to contribute. (The self-employed are covered under Medisave, the health insurance portion of the program). CPF is financed by mandatory wage-based contributions from employers and employees, which are currently 12 percent and 20 percent, respectively. These high contributory rates make the CPF an important factor in Singapore's high level of savings. The contributions and earnings are exempt from taxes, providing members with a generous tax shelter. A distinguishing feature of Singapore's system is the wide range of objectives and programs that have been incorporated under it. These include home ownership financing, investment in specified outlets, payments for higher education, home and life insurance, and special accounts for hospital and medical expenses. The number of options for use of accumulated funds has, somewhat ironically, weakened the CPF's ability to function as a source of older age income. Many members and contributors arrived at age 55 (the minimum age for withdrawing accumulations from the "ordinary" fund) with insufficient funds to provide adequate monthly income. As a result, the gov ernment mandated minimum amounts that must be kept in this fund, but ac cording to As her (1996), many members are not likely to attain this goal. In 1995 the government set up a program to add funds from general revenues to certain classes of older retirees to assist with their medical or other accounts, despite the government's general philosophy of avoiding entitlements. As Asher notes, this step shows that social insurance principles can be incorporated into a provident fund system with moderate changes. The surprisingly low reliance on the CPF for retirement income is con firmed from the 1995 survey (reported on later), which shows that only 2.2 percent of older Singaporeans list pension or CPF funds as their major source of income, and the proportion relying on children or other relatives is the highest among the four countries compared. In April 2001, the government announced a new supplementary voluntary defined contribution plan (Supple mentary Retirement Scheme) to allow higher rates of retirement savings in order to enhance the role of CPF as a source of retirement funds (Ministry of Finance, 2001). Health Care Programs All of the countries have at least some public health programs for older persons, although the coverage and benefits vary greatly. General features of these programs are summarized in Table 3.2, and will be discussed in more detail Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 76 The Well-Being of the Elderly in Asia below. As with the retirement programs, many of the health programs are restructured or integrated versions of programs in existence for some time. An interesting feature of these programs is that, because many are aimed at employees, older persons are often regarded as dependents on their children's or grandchildren's policies. In addition, as will become apparent in the discussion below, the programs can be complex with regard to what services are included and excluded from coverage. The Philippines In 1995 the Philippine Congress passed the National Health Insurance Act to establish a National Health Insurance Program (NHIP). This program, which became operational in 1998, aims to provide health insurance coverage to the entire Philippine population by the year 2010. NHIP replaces the Medicare Program (which had been in place since 1972) and is being managed by Phil ippines Health Insurance Corporation (referred to as PhilHealth). Under NHIP, individuals and their employers contribute according to the employee's in come to a Health Insurance Fund. This fund is then used to provide benefits equivalent to a certain proportion of the cost of a member's hospital care and selected outpatient procedures. The NHIP currently covers employees in both the private sector (SSS members) and the government (GSIS). In 1998, SSS members comprised about 72 percent of the NHIP's membership, or a total of 19.9 million persons, more than half of whom reside in the National Capital Region (NCR). About 2.0 million (10 percent) of the SSS members covered by the NHIP were self- employed, which includes the "extended self-employed," or the informal sec tor of the business community such as mobile peddlers and sidewalk vendors. Public sector employees comprised 22 percent of the total NHIP membership in 1998. This group includes government employees, retirees, pensioners of the national government including the Armed Forces of the Philippines (AFP) and the Philippine National Police (PNP), local government units, and gov ernment-owned and controlled corporations. As mandated by the Social Security Act of 1997, the Overseas Workers Welfare Administration (OWWA) takes charge of the membership of the Over seas Filipino Workers (OFW). In 1998, some 375,860 OFWs were registered members of the NHIP, representing about 5 percent of the overall member ship. (This number represents only the OFWs in destination countries where the Philippine government forged bilateral agreement for social security pro visions of the Filipino workers.) The contributions of members to the NHIP's Health Insurance Fund vary according to income level, ranging from a monthly premium of 75 pesos Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Policies and Programs in Place 77 (equivalent to approximately US $1.50) for those earning less than 3,000 pe sos per month to a maximum of 187.50 pesos for those earning 7,500 or more pesos per month. For the employed sector, the premium is shared equally by the employer and the employee, whereas for the self-employed sector, the member pays 100 percent of the premium. For indigent members, premium payments are subsidized by the local and national government. The main focus of NHIP is on coverage for hospitalization. NHIP cov ers in-patient care plus some selected procedures such as chemotherapy, ra diotherapy, hemodialysis, and cataract surgery. The program does not cover outpatient doctor visits, preventive care or dental care. Additional exclusions in the benefit package include such things as nonprescription drugs and de vices; out-patient psychotherapy and counseling for mental disorders; drug and alcohol abuse; cosmetic surgery; home and rehabilitation services; opto- metric services; and normal obstetrical delivery (except the first). A careful scrutiny of the exclusions indicates a bias against the elderly in terms of ac cess to home and rehabilitation services and optometric services, for which they have the greater need. In-patient services may be obtained in accredited hospitals and from accredited health care professionals. Access to NHIP benefits requires at least three monthly contributions within the six-month period immediately prior to confinement for both private and government sector members, provided that the 45-day room and board allowance for the member and another 45 days to be shared by the dependents for the calendar year has not been used. The costs of hospitalization chargeable against the NHIP is, for the most part, lower for government than for private sector members to accommodate the differential between hospitalization costs in government and private health institutions. In addition to the government program, several non-governmental or ganizations are attempting to address the health needs of the elderly. The Inte grated Geriatric Care (IGC) project, piloted in Davao province in southern Philippines is a good example. It is a consortium of the Medical Mission Group Hospital and Health Services Cooperative, the PILIPINA Legal Re source Center and the Cooperative Rural Bank of Davao. The program pro vides persons age 55 and over with a comprehensive package of services, including health care, legal services, and financial management. Thailandv Thailand has been expanding health insurance coverage for its population through a series of programs over the last 12 years, and has initiated a number of programs with a special focus on the elderly. Estimates from the Ministry Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 i s, s; <3 CovrtagePaymentInsrteds -a a£ o§615a OU•osares emrt-ensy and ervirtss othrt seleste teemed to be "sosts1 £ 8 !t|.E|J.bin tiiu I/^ O >"»s • e* =1 °lljllo'? .SEmployees in both private sector (SSS members) and -overnment (GSIS memberCovrts meme ovrtseas Filipoo§5?E E K_t= 2 <U 2 ^•S M Q Eb-oSS •— P s £ g.£ Mc ^ i ^Illiic1i£§. ** U c/j._, .. VI 4ia-e for nonspressript 1 devises; outspadent •alih sare; optometris dru- and almehol abu 1 rehabilitadon servis1 *lliliHillQworkersEventually to be expanded _ mevrt entire population:es at -overnment and health rtnters; refrtral system• — Vt ei 3 °wi O, 2<llii^s PIt B 2 33 (d 1 I O u oocu c S Z S E£Inidally disadvanta-ed persons a-e _0 and ovrt; nc all prtmens a-e __ and ovrtia3eg|E:s at -overnment and health rtnters; referral system>-i § S-f-8<ll1booStin eoil— j: J3t*!*Farmerss wmkers in small firmss seasonal wmkrtss sel employed and up to _ membersii;g•2S O1 1M ww ~and outpatient servis facilities; inpadent s: fasilities subjest to .-merit rtilin-sill Hi!S a S 21 c£.83 fe4* 2u O.x 3.t mGovrtnment and state employes and retireert thei spousess parentss and up to shildreniI-s]C 1? uf | E6 -1 ^.« c uo w Ea•&I5 udisabih.ty and death medisal benefits thro private providrt selera a * -s I'S c ° .a §•§ >,C/3 XI D. J3£> IIIfl•s|l,» -2 oo *tl— u aIo.1iygli11, Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 i &•o S v g S u Sa 1 ° PS•S 8 1 s Ert -Q td t o,fsj it£ Z '£ oo•§1 s 2 &-S11 if !!2 B > -3 ° .•SIII! ilM C|l t!l|^ |as|•I i In!.*5 * §>"'§••=C B 0 o B n O QLN CXi 2, u O i*-! D.B11 .1.u•a 1S£1.> ^eS wD JS1•aHtA.§Ji1Approved mediAssumulated CPF mentribution*„18%£=i1Su° ^ § I B°M *o B ^ a*. ^ O B CIHH,III illAnnual premium; amount varies with age"Jli1^ tPI|2|SB —III1—111S SUnspesifiedaGovernment pays; assistanse based on strin-ent meansstes basisIs11llS •a o R V QQM 2 n11SB°| 0sl 0 g.lit*.u S u V uUl J3 -p M -^S-5S|1 " ° MO ™§ S -1 « .•*" " 1 a" B*M 9j B h -a5 1 1 1 -aJ ^ rt u U,_Annual premium; amount varies dependin- on a-e o entry into pro-ram"i3. li eSfl 1^f 1•- Si8- a ^G wi . . 4)g C B £g1 § :§,-§ c&ai11aM •8 ^3 Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 80 The Well-Being of the Elderly in Asia of Public Health and Kamnuansilpa (2001) indicate that the proportion of the population without health insurance declined from about 67 percent in 1991 to 29 percent in 2000. In recognition of the special health care needs of the elderly, the Thai Ministry of Public Health (MOPH) officially established an Institute of Geri atric Medicine in 1994. The government has initiated a program of free medi cal care for persons aged 60 and over. In addition, some Thai elderly have coverage through government programs such as voluntary government health insurance (through the Health Card Scheme), the Social Security Act, and the Civil Servant Medical Benefits Scheme (CSMBS). Those elderly who qualify may be covered by more than one program. Starting in 1989, the MOPH initiated a free medical care program for disadvantaged elderly that was subsequently extended to cover all elderly (for a description of this program, see Knodel et al. 2000). Under this pro gram, all government hospitals and health centers provide free services na tionwide to those 60 years and older with an "elderly card."5 The program operates through a "gatekeeper" and referral system. The first contact in the rural area is at the sub-district (tambol) health center, while in urban areas initial contact is made through local community health facilities. Complicated cases are referred to higher level facilities such as district, provincial, or gen eral hospitals. Through 1997 the elderly card program was non-means tested, and thus all Thai elderly were potentially covered by this form of government health insurance. In 1993, 3.5 million elderly persons were covered by this scheme. Given the scale of the program, it has been a major payer for health costs for the elderly, accounting for about 5 percent of the total MOPH budget during 1995-97 (communication from the Institute of Geriatric Medicine). The program is currently being modified; it will be integrated with the Medi cal Welfare Card program and will be means tested. The government has also established a subsidized prepaid voluntary health insurance program known as the Health Card Scheme. It started as a pilot project in 1983 and was subsequently extended to all provinces. The target groups are farmers in rural areas, workers in small firms (with less than 10 workers), seasonal workers, and self-employed persons. As of 1997, about 2 million health cards were issued. A card costs 500 baht per household per year and covers up to 5 members including the elderly. The government has been subsidizing this project, initially spending 500 baht for each card sold and doubling this amount in FY 1998. The benefits provided by the health card are similar to those of the elderly card, and cases are handled using a similar referral system from lower to higher level facilities. Health care has been included as a fringe benefit offered by the Thai government to all government and state enterprise employees for many decades. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Policies and Programs in Place 81 The health care scheme, which is also extended to employees' spouses, parents, and up to three children, reimburses for outpatient care and either pays directly or reimburses for inpatient care. Retirees, active employees, and their dependents can freely seek inpatient care at public or private facilities with some reimbursement ceiling applied to private outlets.6 Both as a result of the economic crisis that started in 1997 and the rapid increase in expenditures (from 4.3 billion baht in 1990 to 13.6 billion baht in 1996), the Ministry of Finance has instituted new regulations to contain costs by means of co-payment and reimbursement ceilings. The major program for providing health care to the population at large is the Social Security Act, which took effect in 1991 and was described in the previous section on retirement programs. This program provides coverage for injuries, sickness, and disability to private sector workers in firms with 10 or more employees. Workers can select a public or private hospital as their medical provider, with the selected provider receiving a capitation rate for each work ers. The program is financed by a contribution of 1.5 percent of wages (up to a specified maximum) paid each by the employer, employee, and govern ment. The Social Security Office, which administers the program, is also now responsible for the Workmen's Compensation Fund, which began operating in 1974, and covers job-related injuries and sickness. Private health insurance does not play a major role in financing health care for Thai elderly as it tends to be uncommon (Ron, Abel-Smith and Tamburi 1990). It can be obtained, however, for groups or individuals, either attached to a life insurance policy or as a stand-alone insurance policy. In either case, health services are normally provided at private facilities on a fee-for-service basis. The extent of coverage depends on the insurance company and the amount of the premium. Unfortunately, most insurance companies do not of fer policies to people aged 60 years and older because of the high health risks of older persons and potentially large amount of expenditures that could be involved. Taiwan The Universal Health Insurance (UHI) program was implemented in Taiwan in April 1995. UHI replaced previously existing health insurance programs, such as the Laborers' Insurance, Government Employees' Insurance, and Farm ers' Insurance programs, and expanded on their coverage to provide benefits for the workers' dependents and others. UHI provides health insurance cover age for the entire population with the exception of military officers and ser vicemen (who are covered under a separate program), as well as incarcerated and missing persons. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 82 The Well-Being of the Elderly in Asia Funds for the UHI program come in roughly equal proportions from the insured, employers, and the government. A monthly premium based on salary is assessed and is shared between the employee (30-40 percent) and the employer (60-70 percent). For certain categories of workers (e.g., private school teachers, farmers, fisherman, members of occupational associations, retired servicemen and their dependents) premiums are subsidized by the government either in part or in full. Those not working are covered as dependents of their employed family members, who must pay their premiums. Generally parents pay for their dependent children and adult children pay for their elderly or retired parents. Low-income individuals who have no family members to share the costs of health care are covered free of charge by the government. In addition to the premium, the insured is required to make a co-pay ment for each office visit or hospital stay, ranging from NTS50 -150 (US$2- 6) for office visits, up to NT$420 (US$17) for emergency care. Co-pays for hospital stays are charged as a percentage (10 percent) of the total medical costs. The ceiling for co-payments is NT$15,000 (US$600) per person for each inpatient service, and NT$39,000 (US$1,600) per person for an entire calendar year. The UHI program covers dental care as well as most inpatient, outpa tient, in-home nursing care, and preventive care. However, the program re quires co-payments for medications, and no coverage is provided for long- term institutional care or medical aids or equipment, including such items as eyeglasses, hearing aids, dentures, or ambulatory aids. Since the UHI was initiated, utilization of both outpatient and inpa tient care has increased dramatically in Taiwan (Cheng and Chiang, 1997) with a concomitant increase in health care costs (Chiang, 1997). The increase in utilization occurred across the board, but was especially pronounced for those who were uninsured prior to the introduction of UHI. As a result of the deficits experienced, the Bureau of National Health Insurance has made sev eral changes in benefits and premiums and is contemplating others. Singapore Singapore has several government programs in place to take care of medical needs in old age; currently these programs only cover medical care up to age 75. Recently, the government introduced a new program to provide lifetime insurance coverage for long-term care services to elderly Singaporeans with severe disabilities. The details of each of these programs are provided below. In general, these programs do not pay the full cost of care and the insured has to make a co-payment. This is in line with the government's health care financing philosophy that emphasizes personal responsibility. As Health Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Policies and Programs in Place 83 Minister Yeo Cheow said in a 1999 speech: "The co-payment principle, coupled with community support and government subsidies has to date kept acute care services affordable to all Singaporeans." Since 1984, every Singaporean with a Central Provident Fund account has a Medisave account. Contributions to Medisave are shared equally be tween the employer and employee, are tax exempt, and earn interest. At present, between 6 and 8 percent of an individual's monthly contribution to the CPF account is put into the Medisave account, which is used to pay for costs of hospitalization and certain outpatient treatments of the individual or a family member until the age of 75 years. Once an individual has withdrawn his/her lump sum from the CPF account at the age of 55, he/she must maintain S$ 17,000 in the Medisave account to meet health care needs during retire ment. This amount is being increased incrementally to S$25,000 in the year 2003 (Inter-Ministerial Committee on an Ageing Population, 1999). In June 1990 the Medishield and Medishield Plus Schemes were introduced when the government realized that Medisave was insufficient to cover most individual's medical costs. Medishield is a low-cost medical insurance plan that covers hospitalization expenses, catastrophic illnesses, and certain outpatient treatments such as kidney dialysis, chemotherapy, and radiotherapy for cancer. The annual premium is deducted automatically from the member's Medisave Account, unless he or she decides not to be insured. Medishield Plus is similar, but both the premiums and benefits are higher. Both of the plans are voluntary. As of January 2001, about 90 percent of elderly Singaporeans between the ages of 61 and 69 were insured under the Medishield Scheme. This is the result of a government plan instituted at the end of 2000, which helps defray hospitalization costs for the years 2001 and 2002 if elderly signed up for Medishield or a Medisave-approved insurance plan by December 31, 2000. Medishield and Medishield Plus cover health expenses only until the age of 75. For the most part, coverage is partial, with the patient paying some percentage of the cost, and each type of treatment covered carries a claim maximum. Before any claim can be made, the member has to pay a fixed annual deductible (ranging from S$500 to S$4,000 depending on the type of plan). Of the remaining claim amount, Medishield/Medishield Plus pays 80 percent while the member pays the balance from his Medisave account or out of pocket. For outpatient treatment, Medishield pays 80 percent of the actual expenditure up to the maximum amount. Medishield allows the member to claim up to S$20,000 a year with a lifetime limit of S$80,000. Under Medishield Plus Plan A, a member can claim up to S$70,000 a year, with a lifetime limit of S$200,000. For Medishield Plus Plan B, the annual claim limit is S$50,000, and the lifetime limit is S$l50,000. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 84 The Well-Being of the Elderly in Asia In 1993, the government began a Modi fund program for needy Singaporeans who don't have Medisave or Medishield accounts or for whom the savings in these accounts are insufficient. It is not an entitlement and is funded by State budgetary surpluses. Also, in January 1996 the Singapore government began a four-year Top-Up Scheme. Every Singaporean bom on or before April 1, 1934 was given a yearly addition to his or her Medisave account, if a family member or the individual him/herself put a yearly co- payment of S$50 into the Medisave account beginning in 1995. Depending on age, the individual received a top-up of between S$100 and S$350. Recently the government introduced a new insurance plan to provide insurance coverage to elderly Singaporeans who require long-term care. The program, Eldershield, is scheduled to begin in June 2002 and will provide older Singaporeans with basic financial protection and help defray out-of- pocket expenses in the event of severe disability. Eldershield is a voluntary program that is funded primarily by premiums paid by program members. Premiums are structured to be payable annually starting at age 40 and continuing through age 65. The maximum age of entry into the program is 70, and persons with pre-existing disabilities are not eligible to join. Individuals joining after age 40 are required to pay larger premiums over a shorter period of time; those joining at age 65 or later must pay the full premium in a single lump-sum payment. Unlike the Medisave and Medishield programs, which provide coverage only through age 75, Eldershield provides lifetime coverage to its members. In keeping with the government's philosophy of promoting family support and community-based services for care of the elderly, Eldershield payouts will be made in the form of cash benefits, and not tied to the reimbursement of institutional care. This allows individuals requiring long- term care and their families the flexibility to choose where the care will be provided—at home or in a nursing home or other institutional setting. Payouts are made when a person cannot perform three of the following activities of daily living (ADL) unaided: walking, eating, getting out of bed, dressing, bathing, or going to the toilet. Initially payouts will be set at SS300 per month, up to a maximum of 60 months. Once the program is underway the government will consider enhancing Eldershield to provide higher coverage and payouts. Private insurance companies provide a number of health insurance plans available in Singapore. Most of these plans include insurance for prevalent diseases such as heart attack, cancer; and stroke; some also cover hospitalization, surgery, intensive care, and permanent disability. However, most elderly Singaporeans currently do not rely on health insurance to pay their medical costs. As shown in Table 10.2 of Chapter 10, less than 3 percent of the respondents indicated supplemental (private) medical insurance as their Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Policies and Programs in Place 85 main type of support for health and medical costs. Instead, almost half of them relied on their children or other family members to pay the bills through the Medisave accounts of these individuals. Other Programs and Policies Targeted to the Older Population The foregoing sections have focused on the broad retirement and health programs in effect in each country that cover all or major segments of the working population and provide specified benefits before retirement ages and/ or subsequent to retirement. In addition, each country has a number of policies and programs designed for the elderly that address many of the special needs that often arise for that segment of the population. For discussion purposes, these programs may be considered under four broad categories: • Programs Providing or Subsidizing Regular or Special Medical Services. These include free or subsidized medical care or medicines, home nursing services, and centers for rehabilitative or custodial day care. • Programs Addressing Residential or Long-Term Care Needs. These include special residential structures, as well as improvements to current residences, and nursing homes. • Programs to Enhance Daily Life Services to the Elderly. These include day care facilities, senior citizen recreational centers, home visits for doing chores and providing transportation, special educational and counseling services, cultural events and various subsidies to make attendance more affordable. • Programs to Provide Economic Assistance. These include special tax benefits and/or special payments to elderly with low incomes. It should be recognized that a number of these categories overlap to a certain extent. For example, programs that provide day care for elderly still residing at home may be physically part of nursing homes or similar structures; subsidies to the elderly for reduced transportation fares or costs of cultural events contribute directly to enhanced daily life and also provide an economic benefit. Although some of these programs have been operating for years, they represent an area of increasing public and legislative attention as countries seek to adjust their policies to the rapidly growing older population and the changing social and economic environment. In developing new programs and adapting older ones, central governments have been using a range of strategies, often involving local governmental authorities, voluntary organizations, and the private sector, as well as creating new incentive programs for the public. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 86 The Well-Being of the Elderly in Asia Although providing an exhaustive list is not possible, the following section describes several programs and policies in each category that illustrate this range of strategies. In addition, Chapter 11 presents data from the surveys about the knowledge and utilization of several of these programs. In conjunction with the Chapter 11 data, the information presented here will help clarify effective future directions in the provision of senices. Programs for Providing or Subsidizing Regular or Special Medical Services Some of the programs discussed in the Health Care Programs section are mainstays in the provision of medical care to the elderly and will be touched on briefly here along with more specialized programs to round out the picture. In the Philippines, in addition to the benefits the elderly may accrue directly or as dependents under the National Health Insurance Program, current elderly (aged 60 or older) can avail themselves of several special programs. Those earning less than 60,000 pesos a year can register and obtain a senior citizen ID card which entitles them to a 20 percent discount on medicines needed for their own use. In addition those registered are entitled to "free medical and dental services in government establishments anywhere in the country subject to guidelines issued by the Department of Health, the Government Senice Insurance System and the Social Security System" (regulation quoted in Natividad, 2000, p. 272). However, Natividad notes that many institutions are excluded from this regulation because they technically do not fall under Department of Health control. In Thailand, as described in Health Care Programs section, the special "elderly cards" and the family "health cards" provide a range of medical benefits at government hospitals or health centers to a high proportion of the elderly at little or no cost. A1998 survey by Kamnuansilpa et al. (2000) found that 80 percent of the elderly people surveyed possessed the elderly card and another 9 percent had other coverage. (The survey did not cover Bangkok.) The authors also note that, although the Ministry of Public Health has a "home health care" policy, under which public hospitals are to develop outreach teams that visit people at home and sub-district health centers are also to engage in home visits, less than a third of the surveyed elderly report such contact. Older people may also receive physical and occupational therapy at one of the Social Service Centers described further below. The Universal Health Insurance program in Taiwan now provides the elderly with basic in- and outpatient senices. In addition, home health services for the elderly, initiated in 1989 in an experimental program, have been expanding rapidly. As of March 1999, 214 of these freestanding or hospital- Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Policies and Programs in Place 87 based agencies were providing service (Bartlett and Wu, 2000, p. 219). This growth has been greatly facilitated by the fact that home visits by nurses and physicians are now paid for by the Universal Health Insurance System. Basic medical services to the elderly in Singapore are primarily covered through the various health care provisions of the Central Provident Fund, as described above. In addition, the Home Nursing Foundation, a joint public- private sector organization established in 1958, is a major provider of community-based elderly health services. The Foundation provides home nursing service as well as rehabilitation service at its health care centers. Programs Addressing Residential or Long-Term Care Needs Given the widespread expectation in most of Asia that older parents will reside with their children and a demographic profile that until recently was characterized by large families, a young age structure, and modest expectations of life above 60, the development of special residential facilities for the elderly, either for those in relatively good health or for those requiring extensive care, has only recently gained broad significance. Nevertheless each of the four countries has operated a few such facilities for some time and each is currently responding to the growing demand. As with other programs, these efforts often involve the public, private, and voluntary organization sectors within each country. In the Philippines the Department of Social Welfare and Development manages three residential centers for the elderly, mainly for those abandoned and destitute. Civic and religious organizations accredited by the Department operate homes as well—Natividad (2000) cites reports of 19 such homes— although these tend to be smaller in scale. The number of privately run nursing homes is not known, as these are not identified in government documents. Incentives incorporated into the Senior Citizens Act passed in 1991 include real estate tax relief for the first five years of operation of residential communities or retirement villages solely for older persons, and priority for building the roads leading to such communities (Natividad, 2000, p. 276). Although like the other countries Thailand fosters familial care for the elderly, the Department of Public Welfare has established a small number of government residences, mainly intended for the poor, homeless, and deserted elderly. The first such home was established in 1953, and as of 1997,16 were in operation, serving about 2,000 residents. Several charitable organizations, in particular several associated with ethnic Chinese, also operate homes for the aged. In addition, there are private for-profit nursing homes, although systematic data on these homes are unavailable as they are managed free of government regulations. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 88 The Well-Being of the Elderly in Asia As Thailand is a primarily Buddhist country, some elderly live in temples as monks and nuns. According to the 1990 census, about 3 percent of men aged 60 and over live in temples, 98 percent of whom are monks, but only 0.2 percent of elderly women lived in temples. The census figures thus suggest that only a small number of homeless elderly live in temples. A national unpublished study conducted in 1994-95 by the Department of Public Welfare in the 32 provinces reported that there were 1,115 homeless or destitute elderly residing in 362 temples (personal communication from the Department of Public Welfare). As a result of this study, the DPW started a program to promote the use of temples as residences and service centers for needy elderly. As of mid-1998, almost 200 temples located in two-thirds of the provinces had agreed to participate and some have already initiated activities. (The above from Knodel et al., 2000, pp. 251-252.) The government in Taiwan has mounted recent efforts to develop an adequate long-term care system for the elderly with chronic diseases and disabilities that combines community-based services with various residential facilities. Both the Department of Health and the Department of Social Welfare are involved in this system. In 1991 the Department of Health began encouraging hospitals to establish nursing homes on a pilot basis. By 1995 there were 9 such homes with 470 beds. In 1999, Bartlett and Wu (2000, p. 218) reported 101 nursing homes with 4,308 beds registered with the Department of Health, and 123 homes for older people with 7,615 beds under the Department of Social Welfare. In addition they cite a survey identifying 710 uncertified homes with more than 22,000 beds, many of which are in the process of becoming registered under recent legislation designed to involve the private sector in meeting the growing demand. This demand has been stimulated by changes made in 1996 to the Universal Health Insurance Program that allow the cost of medical and nursing care services in nursing homes to be reimbursed. Prior to the expansion of nursing home facilities, a significant portion of long-term care took place through extended stays in hospitals. In addition to long-term housing for elderly with health needs, the Ministry of Social Welfare provides some residential housing for poorer older citizens at no or low cost, as do a few voluntary and private welfare agencies. Housing arrangements for healthy older retirees who wish to live independently are a recent phenomenon with a few large complexes of senior citizen apartments developed in or near several major cities (Bartlett and Wu; 2000). Singapore's program for residential facilities geared to the older population includes a highly developed set of incentives to maintain the elderly in the community and with or near their families, and a program of partnerships with various organizations for maintaining special residential and nursing home facilities. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Policies and Programs in Place 89 The government provides several incentives such as tax relief and priority housing arrangements for families that take care of an aged parent. For example, a co-resident adult child could obtain tax relief of up to S$4,500 per parent in 1998. In addition, children who choose to live within 2 km of elderly parents receive tax rebates and priority housing. A second program, the Joint Selection (JS) scheme, allows married children and parents to submit separate applications to purchase flats close to each other in the same housing estate. Five percent of the flats offered for sale in these estates are set aside for allocation to such applications. These applicants are also entitled to other benefits such as a lower down payment and a larger mortgage. Similar benefits are offered under the Multi-Tier Housing Scheme, which encourages married couples to live in the same flat with their parents or grandparents for care and support. A related program, the Conversion Scheme, was implemented in 1993 to help maintain established community and family ties. It enables lessees of 3-room or smaller flats to purchase an adjoining or smaller flat from the open market for conversion into a larger flat to accommodate married children and parents. More recently, programs have been developed for Singapore's elderly who choose to live alone or with a spouse. The government is building more housing for elderly who choose not to live in an extended family setting and also upgrading existing housing facilities for the elderly. For example, the government started the Studio Apartment Scheme in November 1997 to cater to the special needs of the elderly. The studio apartments, which provide sufficient living space for one or two persons, come with pre-installed fittings and special safety features for the elderly (e.g., pull cords and heat detectors linked to an alarm system; support hand bars, non-slip flooring). In addition, the studio apartments are located near neighborhood centers for easy access to communal and commercial facilities, as well as the transportation network. In addition to these incentive programs, provisions have been made to house Singapore's elderly who are destitute or do not have families. These include a mix of government and organization-sponsored facilities, as well as volunteer and private nursing homes. Homes run by community service organizations and religious organizations are often subsidized by the sponsoring organizations, although residents are required to make a minimum payment, and those who are able to contribute more are encouraged to do so. Last, residential homes for the aged run by volunteer welfare organizations have increased in the last decade to meet increasing demands. At present, there are about 45 such homes (Directory of Social Services, 1997), most of them affiliated with the National Council of Social Services (NCSS), the umbrella body for all voluntary welfare organizations in Singapore. NCSS provides partial funding for these establishments. In addition, there are nursing Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 90 The Well-Being of the Elderly in Asia homes that cater to the disabled and frail elderly, some of which offer temporary respite stays so that family caregivers can have a break from their duties. Also several private nursing homes run by individuals are available to ambulant, semi-ambulant, and non-ambulant persons. The standard of care and facilities available vary according to prices charged. Some homes have qualified nurses on their staff, while others rely on non-nursing personnel. Under a five-year master plan for elder care, the Ministry of Community Development and Sports (MCDS) will revamp its funding scheme for Voluntary Welfare Organizations (VWOs), fund a series of programs for the elderly, and spend S$15 million on a five-year public education program. Under the new funding scheme from April 1st 2001, MCDS will give its 50 percent share for VWOs operational costs, even if the organizations have raised more than their 50 percent share. MCDS will also spend more than S$30 million over the next five years on programs for healthy elderly. These include mutual-help schemes in 142 Senior Citizens' Clubs, enhancing befriender services, and supporting programs developed by senior citizens themselves. MCDS will also spend approximately S$ 15 million to fund support services for frail elderly. These support services involve setting up three centers whose twelve case managers will plan the services and support needed for the individual elderly. In addition, MCDS will be financing the development of six neighborhood service centers to cater to elderly social service needs. The type of information the centers will provide include teaching simple first aid and information on where elderly can go to learn computer skills. MCDS also plans to build three multi-service centers as one-stop facilities for the whole family, providing day-care services for the very young and elderly. The overall plan is to integrate health and social service needs of the elderly. "The objective is to eventually put all community-based services under one roof so that centres will become a one-stop service point for residents of all ages and for different needs" (Abdullah Tarmugi, Minister for Community Development and Sports). Programs to Enhance Daily Life Services This category of services includes a variety of programs designed to assist the older population with problems encountered because of limited mobility, other types of limitations, or the lack of social contact and stimulating activities. We include here discount policies designed to encourage participation in recreational and cultural activities, although these are also programs with economic benefits. In the Philippines, registered seniors with ID cards are entitled to 20 percent discount on costs related to using transportation, hotels, restaurants, and recreational facilities. They also receive a discount of 20 percent or more Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Policies and Programs in Place 91 on admission fees at theaters, cinemas, concert halls, and other places of culture, leisure, and amusement. The Senior Citizen Center Act passed in 1994 provided for the creation of senior citizen centers in each municipality. These centers, envisioned as the hub of activities oriented to older persons, are to be run by the elderly themselves. In the first two years, the Department of Social Welfare and Development was charged with establishing demonstration centers in each province to serve as models for each municipality. It is of interest that the passage of this Act was partly the result of strong lobbying efforts by senior citizen groups under an umbrella organization called the Federation of Senior Citizens Association of the Philippines (Natividad 2000, pp. 276-277). In Thailand, the Department of Public Welfare operates centers designed to provide a variety of social services for elderly. The first center was opened in 1979; by 1998, 13 centers were operating throughout every region of the country. Health care, exercise, counseling, training, and welfare services are provided at the centers; home visits by mobile units are arranged through the centers. In addition, emergency shelter is available for short-term stays of up to 15 days for elderly in urgent need of temporary lodging and assistance (Department of Public Welfare, 1997). In Thailand, elderly clubs or senior citizen associations were first promoted by the government during the 1980s based on the idea that informal group gatherings among older persons is a tradition in Thailand. The goal was to facilitate the formation of organizations that could provide as well as receive assistance or services (Siripanich et al., 1996). Thus, with support from the Ministry of Public Health and the Department of Public Welfare, the number of elderly clubs grew rapidly in Thailand during the 1980s. Although some clubs were formed from genuine local initiatives, most were the result of government prompting and support. Club offices tended to be located either at a government health outlet or a temple. A1994 study identified about 3,500 senior citizen clubs of various sizes and with various activities throughout Thailand (Siripanich et al., 1996). In the absence of continuing follow-up and support by the governmental agencies involved, substantial numbers of clubs were either dissolved or became inactive. The main reasons for the failure of so many clubs included a lack of clear initial aims at the time of formation, inadequate sizes, and inconvenient locations. In their effort to promote the importance of elderly clubs and strengthen networking among them, Thailand's Department of Public Welfare organized a series of annual national conferences of club representatives starting in 1983. One outcome was an agreement to set up an autonomous organization to coordinate activities of elderly clubs all over the country. The Senior Citizen Council of Thailand, registered as a legal entity in 1989, is a strong voice for Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 92 The Well-Being of the Elderly in Asia the needs and rights of elderly. At present, about 300 senior citizen clubs are registered as members of the Council. To qualify for membership, the club must have at least 50 members and have existed for a year or more with continuous activities. Because many of the clubs initiated by the government promotion effort had fewer than 50 members or did not have continuous or organized activities (i.e., had been established in name only), the number of current member clubs is far below the number formed at the height of the government's promotional campaign (Knodel et al. 2000, p. 253.) In Taiwan, several types of centers and programs accommodate the needs and interests of the elderly residing at home. Aside from in-home medical and nursing care provided by the home care offices described above, there were 893,000 person-visits in the year 2000 for assistance with household chores, counseling and related services through the social welfare system and 988,000 person-visits in the year 2000 for "meals-on-wheels" service. Day care centers have also been expanding in Taiwan. Under the Ministry of Health there are approximately 20 day care centers that provide nursing and other health services for those participating. In addition, for the healthy elderly there are a growing number of day care facilities provided through the social welfare system and they report 278,000 person-visits to these facilities during the year 2000. Services provided at these centers include cultural and recreational activities, health maintenance, afternoon rest periods, and snacks. Rounding out the picture are a large number of senior citizen centers and Evergreen Academies operated by local governments and other auspices which provide a wide variety of leisure activities, free or discounted tickets for various events, and a large number of courses, with fees subsidized for low income elderly (Bartlett and Wu, 2000). The Evergreen Academies were established in 1987 to provide the older population with instruction and training in a wide variety of subjects and crafts. Local governments are responsible for developing their own programs, assessing the attitudes and needs of their elderly residents, selecting appropriate locations for the academies, designing teaching plans, and purchasing materials. The academies are open to persons 55 years of age and older, and enrollment fees for low-income elderly or residents of publicly supported old-age homes are subsidized. In 1997 there were 302 Evergreen Academies in Taiwan, and some 188,000 people were enrolled in classes (Bartlett and Wu 2000, p. 218). Other cultural, artistic, social, and training activities are provided for elderly Taiwanese over age 60. These are delivered through the 12 district senior citizen centers in Taipei as well as through various other government, community, and religious organizations in other areas. Discounted tickets and/ or complimentary passes are provided to older persons for transportation on Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Policies and Programs in Place 93 public buses, entrance to exhibitions, and other educational and recreational activities. As in the other countries, Singapore has several different structures— both governmental and private—that address the daily needs of the older population. For those requiring a program of day care, the Ministry of Community Development operates six social day care centers that cater to about 200 elderly. Some of the centers have special transportation for the disabled. The Salvation Army operates a similar center. Charges are usually levied for the services provided. The Ministry plans to open four or five new day care centers in the coming years. In addition, Singapore has 31 government-operated senior citizen's centers that offer recreational activities, health screenings, and opportunities for community services via the senior citizen clubs that meet there. Volunteer programs have become more popular and have been encouraged by the government as a means of promoting active, healthy life styles among the elderly. Some non-governmental centers, like the Singapore Action Group of Elders (SAGE), provide personal and telephone counseling, as well as a variety of cultural and recreational activities. Finally, public transportation facilities in Singapore encourage local travel among the elderly by providing discounts and special equipment such as low steps and nonskid floors. Programs Providing Economic Assistance In addition to the array of services and programs outlined above, each of the countries recognizes that the elderly population as a whole, as well as particular subgroups, may require direct economic assistance. In the Philippines this assistance takes the form of tax exemption for registered elderly who earn less than 60,000 pesos per year, and tax abatements for those who coreside with and claim such elderly persons as dependents (Natividad 2000, p. 276). The elderly may also benefit economically from income assistance projects that provide seed capital to families below a given poverty threshold. Some community-based organizations such as the Coalition of Services to the Elderly (COSE) have set up group homes for highly vulnerable single elderly. In Thailand, the Department of Public Welfare assists extremely poor elderly men and women in rural areas with a minimum pension of 200 baht ($6 U.S.) monthly, under a program initiated in 1993 (Knodel et al., 2000, p. 252). Eligible older persons are identified by the Department's network of Village Welfare Assistance Centers, and must be approved by the provincial welfare office. Starting with 20,000 elderly in 1993, the program was expanded to over 300,000 rural elderly in 1997. Though the amounts provided are too Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 94 The Well-Being of the Elderly in Asia low to meet minimum subsistence requirements, and the program may not have reached all its target population, it is significant as a separately budgeted poverty relief plan not tied to the Social Security Fund. In addition, several non-governmental and religious organizations are working with the Help Age International regional office on income generating and support projects for the elderly. Taiwan provides both direct and indirect assistance to low-income elderly through a cost-of-living assistance program to needy elderly 65 years or older (defined as monthly income of less than US$180 and no support from children); and by paying the full Universal Health Insurance premiums for older low-income insurees. The Ministry of Community Development in Singapore has a social assistance program available to the poor, sick, and disabled elderly based on a means test. The criteria are quite stringent; only 1,789 elderly received payments under this program in 1999 (Singapore Department of Statistics, 2000). Financial assistance may also be provided by Family Service Centers, which are partially funded by the government. In addition, a variety of non governmental organizations and ethnic and religious groups provide financial aid and aid in kind to poor older persons, especially those living alone. Indirect support is also provided through free or low-cost medical clinics and services operated by these organizations, and the government has set up a Medifund program to assist poor patients who cannot meet their hospitalization expenses. Conclusions This chapter provides an overview of policies and programs that address a range of needs often faced by older people. These include the broader social security plans that provide retirement income and the health insurance programs that help defray the costs of medical services to working and retired citizens. In addition, each country has a variety of other programs focused on potential needs such as in-home medical and household services, day care, residential and nursing facilities, leisure and recreational activities, and special economic assistance. The cultural tradition in these countries of relying on families for the physical care, economic security, and emotional support of the elderly, as well as the relative recentness of the demographic shift to older age structures, did not inhibit the early establishment of these programs for the elderly. The countries were influenced in their decisions by several factors: the influence of policies developed by more industrialized countries, the far-sightedness of policymakers in recognizing changing socioeconomic and demographic trends, Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Policies and Programs in Place 95 as well as recognition by governments that regardless of social norms, certain individuals and groups fall outside the prescribed arrangements and require special assistance. It is also clear that both the major programs and the more specialized ones are in a state of development in each country. In the Philippines and Singapore, both of which have broad coverage on paper and long-established programs, benefits are inadequate for many of those targeted for service. The Philippines suffers from a lack of compliance by employers and the need for greater administrative and political action to make the programs work properly. In Singapore, the compulsory coverage and high contribution rates that should provide adequate old-age support have been undermined by having the CPF serve many other goals, from home ownership to investments to meeting the costs of higher education. Government officials are working out ways to reassert the primacy of retirement income within the program and strengthen the provisions for covering medical services at older ages. In Thailand and Taiwan, new pension programs are being developed to broaden coverage from a narrow base of civil servants and select others (such as military personnel or employees of large firms) to most of the labor force, but they are still at a very early stage. The final details and date of implementation for Taiwan are still not set and the social security program in Thailand will not pay its first retirement benefits until 2014. It will be important for these countries to set up these extensions and new programs in as efficient and cost-effective a manner as possible if they are to achieve the twin goals of good coverage with minimum strain on individual and government budgets. In addition, because any program is likely to cover only those in the formal sector, approaches to the informal sector (which can be quite large) also need to be developed. A review by Asher (1998) of the retirement programs in place and under development in Southeast Asia expresses several cautions with regard to the adequacy of financing, the replacement rates (i.e., the portion of wages covered in retirement), and for those relying on provident fund approaches, the availability of appropriate higher-yielding investments and the impact of the plans to increase overall savings and investments. These concerns suggest that these programs will be receiving continued scrutiny in an attempt to improve their effectiveness and efficiency. In contrast to the retirement programs, the programs for providing health coverage within each of the countries seem closer to achieving their aims. Each country, through some combination of financing through public provision of benefits, social insurance, or other insurance and savings plans, is extending health care coverage to a broader share of its general and older populations. But significant pockets of non-coverage remain in some countries, and the portion of health care costs paid privately also remains high in a number Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 96 The Well-Being of the Elderly in Asia (Ramesh and Asher, 2000, Table 4.4). As with pension programs, the health care system and methods of financing health care costs are likely to receive continued scrutiny. Thailand, for example, concerned that its level of health expenditures as a share of gross domestic product is above average for a country of its income level, undertook a major health care reform project between 1996 and 1999 in cooperation with the European Union, and a National Health System Reform Committee was established in July 2000. Countries in Asia will also need to give attention to emerging needs as the proportions at older ages increase, particularly those at the most advanced ages (80 and over), as described in Chapter 2. This development will increase demand for mechanisms to provide and finance long-term care for the elderly. In this regard the steps taken by Singapore under its new Eldershield program described above, and Japan's new mandatory long-term care program initiated in April 2000, after several years of discussion and preparation (Campbell and Ikegami, 2000), are likely to prove influential throughout the region. Also of relevance in the case of Japan, is the intense political debate that occurred in the course of crafting the legislation involving several interest groups and local officials as well as national parties (Eto, 2001; Campbell and Ikegami, 2000). Though many issues remain, it is noteworthy that each country currently has most of the infrastructure needed for initiating new programs and strengthening existing ones. Many of the specialized programs described above are under the jurisdiction of well-established ministries, and in several cases, under joint jurisdiction. In addition, each country has non-governmental and ethnic and religious organizations involved in meeting the needs of the elderly, and quite often these organizations have active partnerships with governmental units in providing services. Also worth noting is that each country has grassroots organizations of older people that are helping shape policies and programs and, occasionally, manage the resulting facilities and services. Although lobbying by age-based groups runs the risk of fomenting intergenerational conflict, it also helps insure that the needs of the elderly will not be overlooked in governmental deliberations. Governments will also face pressure to adjust and enhance these programs from the ups and downs of economic cycles and from competing social demands. The recent economic crisis in Asia is likely to make all countries in the region more cautious about the structure of potentially costly programs, and we can expect countries to balance their desire for strengthening a number of social welfare programs with avoiding steps that may hinder their economic growth. The existence of a range of programs does not in itself speak to the strength and adequacy of their provisions. Although some shortcomings noted in the literature were mentioned, overall we have not touched on the degree to Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Policies and Programs in Place 97 which the various programs, particularly those that address more specialized needs, reach their target populations and fulfill their intended service. To do so requires careful evaluation of each program and many of these have not been undertaken. Chapters 7 and 8 touch on the use of pensions in retirement and as a source of income, and Chapter 10 discusses sources of financing health care. In Chapter 11 we present data from the surveys on respondents' awareness of some of the programs and services reviewed above. These results, together with the overview presented in this chapter, will be used in the concluding chapter to shape several policy recommendations. ENDNOTES 1. The following account borrows heavily from Asher (1996). 2. Based in part on Asher (1996). 3. Follows the discussion by Shih (1997). 4. The following account draws heavily from Knodel et al. (2000). 5. The elderly card entitles the elderly to free medical services at public health facilities for the following items: (a) drugs, operations, intravenous solution, blood or blood component, parenteral nutrition, oxygen, (b) prosthesis and medical device including repairing cost, (c) medical service, medical examination, diagnosis, laboratory investigation and other expense for hospitalization, except special nurse, or professional fee, (d) dental service including acrylic prosthesis, (e) public accommodation and food. 6. The medical benefits covered by the Civil Servant Plan are: (a) drugs on National Essential Drug List, blood and blood components, intravenous solutions, oxygen and others for treatment of diseases, (b) prosthesis and medical devices including cost of repair, (c) medical expenditure, cost of investigation, not including special allowance for special nurses, (d) accommodation and food, and (e) annual personal health examination. REFERENCES Asher, Mukul G 1998. "The Future of Retirement Protection in Southeast Asia." International Social Security Review 51 (1): 3-30. . 1996. "Financing Old Age in Southeast Asia: An Overview." In Daljit Singh and T. K. Liak, &&., Southeast Asian Affairs, 1996, 72-98. Singapore: Institute of Southeast Asian Studies. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 98 The Well-Being of the Elderly in Asia Bartlett, H. and S. C. Wu. 2000. "Ageing and Aged Care in Taiwan." In David R. Phillips, ed., Ageing in the Asia-Pacific Region: Issues, Policies and Future Trends. London: Routledge. Campbell, John C., and Naoki Ikegami. 2000. "Long-term Care Insurance Conies to Japan." Health Affairs 19(3):26-39. Cheng, Shou Shia, and Tung Liang Chiang. 1997. "The Effect of Universal Health Insurance on Health Care Utilization in Taiwan: Results from a Natural Experiment." Journal of the American Medical Association 278:89-93. Chiang, Tung Liang. 1997. "Taiwan's 1995 Health Care Reform." Health Policy 39:225-239. The China Post. November 6, 2001. "CLA Criticized over Retirement Proposal." China Post, Internet Edition: www.chinapost.com. Department of Public Welfare (Thailand). 1997. The 1996 Annual Report (in Thai). Eto, Mikiko. 2001. "Public Involvement in Social Policy Reform: Seen from the Perspective of Japan's Elderly-Care Insurance Scheme." Journal of Social Policy 30:17-36. Interministerial Committee on the Ageing Population. 1999. Report of the Interministerial Committee on the Ageing Population. Singapore: Ministry of Community Development. Kamnuansilpa, Peerasit, Supawatanakom Wongthuanavasu, John Bryant, andAree Prohmmo. 2000. "An Assessment of the Thai Government's Health Services for the Aged." Asia-Pacific Population Journal 15(1):3-18. Kamnuansilpa, Peerasit, et al. 2001. Evaluating the Implementation of Strategy for Health for All by the Year 2000 (in Thai). Khon Kaen: Prathawakan Press. Knodel, John, Napaponi Chayovan, S. Graisurapong, and C. Suraratdecha. 2000. "Ageing in Thailand: An Overview of Formal and Informal Support." In David R. Phillips, ed., Ageing in the Asia-Pacific Region: Issues, Policies and Future Trends. London: Routledge. Ministry of Finance (Singapore). 2001. SRS Booklet www.mof.gov.sg. Natividad, Josefina N. 2000. "Ageing in the Philippines: An Overview." In David R. Phillips, ed.. Ageing in the Asia-Pacific Region: Issues, Policies and Future Trends. London: Routledge. Phillips, David R. 2000. "Ageing in the Asia-Pacific Region: Issues, Policies and Contexts." In David R. Phillips, ed., Ageing in the Asia-Pacific Region: Issues, Policies and Future Trends. London: Routledge. Ramcsh, M.. and Mukul G Asher. 2000. Welfare Capitalism in Southeast Asia: Social Security, Health, and Education Policies. New York: St. Martin's Press. Randel, Judith. Tony German, and Deborah Ewing, eds. 1999. The Ageing and Development Report: Poverty, Independence and the World's Older People. HelpAge International. London: Earthscan Publications Ltd. Ron, Aviva, Brian Abel-Smith, and Giovanni Tamburi. 1990. Health Insurance in Developing Countries. Geneva: International Labor Office. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Policies and Programs in Place 99 Shih, Shiau-ping. 1997. "Private Lives within Public Constraints: Retirement Processes in Contemporary Taiwan." Ann Arbor: Unpublished Ph.D. Dissertation, University of Michigan. Singapore Department of Statistics. 2000. Yearbook of Statistics, 2000. Singapore: Department of Statistics, Ministry of Trade and Industry. Siripanich, B., C. Tirapat, M. Singhakachin, P. Panichacheewa, and P. Pradabmuk. 1996. A Research Report on the Senior Citizen Clubs: A Case Study of the Appropriate Model. Bangkok: Vinyan Printing (in Thai). Straits Times (Singapore), 4th April 1999. Vasoo, S., T.L. Ngiam and P. Cheung. 2000. "Singapore's Ageing Population: Social Challenges and Responses." In David R. Phillips, ed., Ageing in the Asia-Pacific Region: Issues, Policies and Future Trends. London: Routledge. World Bank. 1994. Averting the Old Age Crisis. New York: Oxford University Press. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Chapter 4 THEORETICAL PERSPECTIVES, MEASUREMENT ISSUES, AND RELATED RESEARCH Albert I. Hermalin This chapter establishes the groundwork for the analytic chapters that follow by developing the theoretical framework that guides these analyses, addressing several key measurement issues, and briefly reviewing relevant prior research. Given the breadth of subject matters associated with the study of aging and the number of disciplines involved, the first two sec tions identify the disciplinary stance—the demography of aging—that guides this volume and its relationship to other disciplines that also are actively in volved with this topic. The Demography of Aging The study of aging, or gerontology, is a diffuse enterprise, with theoretical, substantive, and methodological contributions from many disciplines. It is a rapidly expanding field in many parts of the world, but also one with long antecedents. The multi-disciplinary breadth is not surprising. From one perspective, the field focuses on an age-defined segment of the human Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 102 The Well-Being of the Elderly in Asia population with all the related behavioral and social science issues associated with human action and organization. These are treated with differential emphases by researchers from anthropology, economics, history, political science, and sociology. From another perspective, the study of aging is concerned with the basic processes of aging and the special conditions that occur toward the end of life, which highlight the involvement of the biological sciences, medicine, public health, social work, and epidemiology. An important recent development touched on below is the increasing amount of research conducted across the disciplines in each broad grouping. The growing interest in aging within most of these disciplines arises from recent demographic trends in the developed and much of the develop ing world, described in Chapters 1 and 2. According to United Nations pro jections, every region of the world will experience an increasing older popu lation age structure over the next 50 years, and in the less-developed regions the increases will be particularly rapid in contrast to past trends (United Nations, 2001). At the same time, attention to aging and older age as a dis tinct stage of life can be found in very early writings, and even the percep tion of old age as a "problem" in Western civilization appeared in early 20th century literatures (if not before), according to Achenbaum (1996). He notes that the term "geriatrics" was coined in 1914 and the term "gerontology," meaning the scientific study of aspects of growing older, in 1908. We position our work within the demography of aging, an emerging field that is itself eclectic—distinguished both by the traditions, theory, and methods of demography and by the set of problems it has addressed and the methods and data it shares with other social sciences. The field is more de fined by the demographic training of its practitioners and the perspectives that flow from that orientation than by any special problems, theories, or techniques. Indeed, in their introduction to the influential National Acad emy of Science volume, Demography of Aging, Martin and Preston (1994) describe the field as "a rather capacious umbrella for a variety of studies addressed to the causes and consequences of population aging," before at tempting to identify several distinguishing features (p. 3). Part of the vagueness concerning the definition and range of the de mography of aging stems from ambiguity in the parent field of demography, or population studies. In a classic formulation by Hauser and Duncan (1959), the following definition was offered: Demography is the study of the size, territorial distribution, and com position of population, changes therein, and the components of such changes, which may be identified as natality, mortality, territorial movement (migration), and social mobility (change of status), (p. 2) Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Theoretical Perspectives, Measurement Issues 103 As the authors note, this definition casts a wide net for demographic endeavors when one reflects that "composition" can go beyond measuring variation in such basic characteristics as age, gender, or marital status to in clude many other qualities such as education, occupation, and health status; and that changes in composition arise from individuals moving from one sta tus to another as well as through natality, mortality, and migration. This inher ent breadth of the field brings demographers into contact with many other disciplines that address some of the same statuses and changes, leading to interdisciplinary exchanges of measures, methods, and frameworks. The authors' distinction between "demographic analysis" and "popu lation studies" helps further identify where the interdisciplinary exchanges are most likely to occur. They define "demographic analysis" as the study of the components of population variation and change, and reserve the term "population studies" for investigations concerned with the relationships be tween population factors and other variables—social, economic, political, biological, etc. In short, "population studies" is as broad as the range of the determinants and consequences of population trends. From this standpoint, demography may be considered quite narrowly as "demographic analysis" or more broadly as also encompassing "population studies" as well as "demo graphic analysis." It is in the arena of "population studies" that demography overlaps with many other fields, and where multiple frames of reference and strategies come into play alongside the perspectives of demographic analysis (Mauser and Duncan, 1959, pp. 2-3). As a result, "population studies" cannot be thought of as a single theoretical discipline. In recent years several additional reviews of demography have appeared that reaffirm the complexity of the field and its multidimensionality, while stressing the perspectives, concepts, and methods that maintain it as a distinct discipline (Caldwell, 1996; Preston, 1993; Keyfitz, 1993; Stycos, 1987). These dual aspects are clearly in evidence in the study of population aging. The more formal aspects of demography are seen in the measurement of mortality levels, trends, and differentials and through models, such as stable population analysis, that relate age structure to levels of fertility and mortality, and in the analysis of population movements that affect population size and composi tion. The traditional tools and perspectives within demographic analysis— such as the centrality of age, the life table as a device for understanding time- varying risks, the concept of the cohort, the emphasis on decomposition, the attention to quality of data, and expertise with complex forms of data collec tion via registration systems, censuses, and population surveys—have con tributed to elucidating many dimensions of population aging. Increasingly however demographers are combining these tools and perspectives with those of other disciplines to address aging from the standpoint of population studies Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 104 The Well-Being of the Elderly in Asia and a wider array of issues that include studies of health, work and retire ment, living arrangements, and intergenerational exchanges. Related Disciplines and Their Perspectives As noted at the outset, the field of gerontology has a long history involving many disciplines. For each discipline, several issues within the field of aging are particularly consonant with their perspectives and/or methods. For example, four editions of the Handbook of Aging and the Social Sciences (Binstock and Shanas, 1976; 1985; Binstock and George, 1990; 1996), among other compendia (e.g., Birren, 1996), have attempted to trace the theoretical, substantive, and methodological connections within the social science disciplines to various dimensions of research on aging. For sociologists, examinations of age stratification, cohorts, the life course, family and social support systems, and social change figure prominently in aging research (Riley et al., 1972; Uhlenberg and Miner, 1996). Anthropologists have offered insights into the effects of culture on the definition and experience of old age; the number and nature of age grades; and the characteristics of age-set societies in which age becomes a major axis of social organization (Fry, 1996; Foner and Kertzer, 1978; Stewart, 1977). Historians have shared with anthropologists the tasks of tracing the definitions of aging across space and time, as well as investigating particular issues such as the impact of science and technology on perceptions of aging, or variations in the development of private and public measures to address the problem of the elderly over time (Achenbaum, 1985, 1996). A number of theoretical and substantive issues in aging have engaged economists. On the theoretical side, the relevance of rational choice models and the role of expectations are under discussion (Myers, 1996), and the mo tivations associated with intergenerational support patterns have been increas ingly elaborated (Lillard and Willis, 1997). Economists have developed use ful models for investigating the interrelationships between population growth and structure and economic development (Lee, 1994a; Easterlin, 1991; Ma son et al., 2001); they have been active in studying the economic status of the elderly, patterns of labor force activity and retirement, and the implications of different types of public and private retirement programs (Smeeding, 1990; Quinn and Burkhauser, 1994; Gruber and Wise, 1999; Smith, 1995; World Bank, 1994). Political scientists have investigated the political attitudes and behaviors of older people, the politics associated with old-age policies, and the interest groups focused on specific age strata or issues of generational equity, among other topics (Binstock and Day, 1996; Quadagno, 1989). Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Theoretical Perspectives, Measurement Issues 105 Many of these concepts and issues have become common currency within studies of aging, permeating all the fields involved. Thus they will appear frequently under the rubric of "demography of aging" as well as the other disciplines cited. Accordingly, several will be addressed at greater length below and within specific chapters. In addition to the social sciences, the study of aging has strong roots in psychology, biology, and the health sciences. Handbooks on the psychology of aging (Birren and Schaie, 1996) and the biology of aging (Schneider and Rowe, 1996) have appeared in concert with those in social science, and sev eral of the concepts and issues from these disciplines intersect a broad defini tion of the demography of aging as well. For example, psychologists have been prominent in developing scales of depression, cognition, and other im portant mental states, and a number of these have been used in large-scale population surveys to measure variations among the older population. In ad dition, concerns with the life course, stages of development, and disassociat ing effects of aging from cohort effects parallel similar issues in demographic analysis and other social sciences. The biology of aging is a broad field covering human and animal models and molecular and cellular research, as well as analysis of the relation of aging to specific systems (immune response, circulation, neurological, etc.). Although many of these endeavors remain discipline specific, in recent years demographers and biologists have contributed to an emerging interdisciplinary field of "biodemography" that meshes perspectives from each discipline on a number of methodological and substantive issues, especially those centering on longevity. A National Academy of Science committee attempted to codify some of the emerging interrelationships, and their results appear in Wachter and Finch (1997). The topics addressed include establishing the trajectory of mortality at advanced ages in humans as well as other species, evolutionary theories of senescence and their applicability for demographic modeling, assessment of intergenerational relations and the role of the elderly in different societies, and the potential for population surveys to assist with genetic studies. This latter topic has received additional emphasis in a newer National Academy of Science committee report focusing on whether biological measures should be included in social science research (Finch et al, 2001). Of particular relevance is the description of an ongoing project to collect bioindicator data from a subsample of the Taiwanese respondents to the surveys employed in this monograph (Weinstein and Willis, 2001). Researchers in the health sciences have a prominent role in studies of the elderly given their strong interest in various aspects of physical, mental, and emotional health, and the utilization of health services and their costs. In the past, research studies on health and on income and retirement tended to Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 106 The Well-Being of the Elderly in Asia be discrete, limiting analyses of how health interacts with income, assets, and retirement. This limitation, as pointed out in an influential committee report (described in Juster and Suzman, 1995), served to stimulate development of surveys of the elderly that included detailed demographic, socioeconomic, and health sections, such as the Health and Retirement Study (see Juster and Suzman, 1995, for an overview). Demographers, economists, and health sci entists have had frequent contact in both the development and analytic stages of these studies. A major area of intersection for social scientists and public health specialists is the study of relationships between social factors—such as socio- economic status, social integration, and stress—and health status (George, 1996). Studies of the apparently strong influence of social support on health outcomes have received considerable attention (Bruhn et al., 1987; Marmot and Wilkinson, 1999), as have investigations into the relationship of socioeconomic status to levels of mortality and morbidity (Preston and Taubman, 1994). Interdisciplinary contact is also fostered by the overlapping interests of demographers and epidemiologists in the strategies and measures used for assessing factors responsible for differential outcomes in morbidity and mor tality. A particularly fruitful intersection has been in the application of life table techniques to generate measures of active life expectancy that reflect the years a cohort may expect to live without disability under given age-spe cific prevalence levels of chronic disease (Manton and Soldo, 1992). The advent of these measures provides insights as to whether increases in life expectancy at the older ages are accompanied by improving or worsening health (Manton et al., 1997; Manton and Gu, 2001; Crimmins et al., 1997; Verbrugge, 1984; Freedman and Martin, 2001). Aconference in February 2001 was organized to further strengthen connections between demography and epidemiology in aging research and other fields (Weinstein et al., 2001). The breadth of topics of relevance to the demography of aging and these other disciplines, and the multiple perspectives from which they may be approached, rule out the possibility of using a single overarching conceptual framework from which to analyze age-related dimensions and issues. For that matter, even the individual disciplines have multiple theoretical perspectives and alternate ways of classifying these. The Handbook of Theories of Aging (Bengtson and Schaie, 1999) and earlier reviews (Birren and Bengtson, 1988) describe many of the perspectives within each discipline, and Marshall (1996, 1999) discusses the nature of theory within the social sciences and the challenges associated with codification. The strategy adopted here is to provide a broad heuristic framework that can serve to organize and locate many of the topics of interest and to tie in some of the specific related theories. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Theoretical Perspectives, Measurement Issues 107 In the remainder of this chapter we set forth this broad framework and several of the theoretical perspectives associated with the empirical analyses to be presented in subsequent chapters, and review select dimensions of past research on population aging in Asia. The Conceptual Framework The organizing framework for this volume, which expands on the model pre sented in Chapter 1, is set forth in Figure 4.1. The framework attempts to reflect both the major societal-level interrelationships and policy concerns usually addressed via aggregate level or macro-analysis, and the large num ber of meritorious individual or micro-level questions centering on the cir cumstances of the elderly and the determinants of their well-being. The well-being of the elderly, shown in the center, is viewed as con sisting of three broad dimensions: economic well-being; physical, mental, and emotional health; and activity levels, which incorporates work, retire ment, and leisure activities. To the left of these dimensions are two sets of determinants differentiated by their proximity to and influence on the elderly. At the far left are the more remote, broad societal factors that set the bound aries within which the more proximate policies, programs, and influences operate (Intermediate Institutions and Influences). These exogenous social, economic, demographic, and cultural factors influence personal characteris tics of the elderly (e.g., level of education), their living arrangements and support/exchange systems, and the formal arrangements through employers or government programs that are available to them. At the two extremes of age, humans cannot produce enough from their labor to sustain their needed consumption, so that each society must develop a set of mechanisms, including formal and informal elements, that define the timing and content of support, the appropriate participants, and their mutual obligations. In terms of Figure 4.1, living arrangements and other aspects of the family exchange system, along with governmental and employer programs that impact family transfers, are key components of the intergenerational transfer system that societies depend upon to smooth out consumption needs and productive abilities across the life cycle. In periods of relative stability these components are maintained by a well-developed normative and value structure. However, rapid demographic, social, or economic changes of the sort experienced by many countries in Asia are likely to produce strains that compromise these arrangements. Because the elements of the transfer system have such wide-reaching social and economic consequences and are so closely entwined with policy decisions, they are increasingly the subjects of research Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 I k|» tCurrent WellsBein- of the ElderlyAstivity Levelsii/1 /T /Work. Re-iremen-, LeinureEmenomis Weil-BeingInmeme, A.e.s Sonone,, Living ConditionnPhysisals Mental and Emotional HealthNew Program and Polisy DevelopmentXVC.rrent WellsBein- of Other A-e Strata'aitEson_is WellsBcing' ' *InstitutionsIntermediateii1J5jtik 1-- £tfi P U11 *i If- U ea o ^. «!! U 11• a g ^ <^B is a I-B 1 • * ° 1 M iJ II. iilt i II .13 S K- 1 8 * 1 £ I " 1 1 s 1 1 1 1*1 §ls<s £ 3 If Jills US fflf IfflpHjg ^"zJo: [J^uj S HUUt.c s?1(/.If . JII III, S !i 11 i^li 1 i!^3 I«ll ! i1Politisal StrustureHealth Infrastrusture Edusational SystemJ Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Theoretical Perspectives, Measurement Issues 109 in rapidly changing societies. As discussed below, certain elements have received more attention than others, and an important goal of this volume is to achieve a greater balance in analyzing the components of the system and their consequences for the dimensions of well-being. Figure 4.1 also attempts to represent several other important processes. The development of new programs and policies that affect the elderly (far right), as well as those in other age strata, is viewed as guided by perceived needs and preferences. These have been added to draw attention to the logical distinction between measuring the objective status of the elderly and infer ring what this means for policies and programs. In some cases the connec tions may be quite direct, as in developing policies or programs to fill the needs of those with specific physical limitations or who lack any assistance. But in others, the objective status measures must be complemented by assess ments of needs and preferences to adequately inform policy development. For example, do retirees prefer the availability of leisure activities or some level of gainful employment? And how does one appraise the adequacy of income levels or housing quality among the elderly? In this case, it must be said that new policies or programs will be guided not only by the needs and preferences of the elderly, but by the degree of economic and political influ ence they possess, especially in relation to the needs and influence of "com peting" age groups. Including other age groups in this figure also emphasizes that the well-being of the elderly on various dimensions can influence the well-being of other age strata and vice versa. Of course, the development of new policies and programs is also a product of existing programs and institu tions, and the broader socioeconomic and cultural dimensions that character ize a society at any given time, as portrayed in the figure. The process is clearly dynamic, as new policies and programs, along with ongoing social change, affect the well-being of each age group and the structures of the influential institutions. This feedback process is suggested by the dotted lines in the diagram. As described in Chapter 3, the assessment of current policies and programs and the development of new ones are an active process in much of East and Southeast Asia given the recent demo graphic and socioeconomic changes there. A number of the countries in the region have introduced major new welfare programs, or are in the process of reviewing new initiatives and modifying existing arrangements. These activi ties impinge directly on the elderly and non-elderly alike, and—as shown in Figure 4.1—have the potential to alter existing social and economic arrange ments. The framework in Figure 4.1 serves as a heuristic and organizational device for identifying a number of specific theories and models of promi nence in population aging research, and the topics that will be the focus of Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 110 The Well-Being of the Elderly in Asia this volume. In the following sections we review major components of this framework and their relation to specific chapters. Societal Changes and Their Impact on the Elderly The plan of action adopted at the World Assembly on Aging held in Vienna in 1982 recognized two major concerns associated with population aging: the humanitarian issues related to the specific needs of the elderly, and the devel opmental issues related to the societal implications, in particular the effects of population aging on production, consumption, savings, and the associated socioeconomic conditions and policies (United Nations, 1983). Considerable theory and research bearing on both these issues have developed over the years. One major axis of theory and research has been the effects of eco nomic development on the status of the elderly and their well-being along various dimensions. An influential theory by Cowgill (1974; Cowgill and Holmes, 1972) was developed in conjunction with a review contrasting the role of the elderly in primitive societies (Simmons, 1945) with their role in several Western industrialized countries. He found that many of the factors associated with "modernization"—industrialization, urbanization, modern health technology, and increased education—tended to reduce the status of the elderly by limiting their job opportunities, separating them from their fami lies, and lowering their social status relative to the young. Subsequent tests and reviews of this posited relationship have been more agnostic as to the effects of development. Palmore and Manton (1974) used national level cor relations to test Cowgill's thesis and found little direct support, and also drew a distinction between the potential effects on the elderly of early levels of development versus more advanced levels, suggesting the possibility of a U- shaped relationship between elderly status and modernization. Treas and Logue (1986) conclude that the relationship between economic development and the welfare of the elderly may depend on many situational factors. For ex ample, wage labor may make the young more independent, but increasing land values may give the elderly considerable economic power. The thesis that modernization impinges adversely on the older population has also come under attack from historians and comparative social analysts who show that the status of the elderly within societies has varied considerably over time and across societies. They identify times and places throughout history in which the elderly were not well off, despite various levels of economic development or technological progress, as well as instances in which the well-being of the elderly improved along with rapid socioeconomic Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Theoretical Perspectives, Measurement Issues 111 development (Achenbaum and Stearns, 1978; Achenbaum, 1982; Cooper, 1999). A related line of thought points to deleterious ell eeels of industrializa tion and urbanization on the elderly arising from their presumed effects on family structure and relationships. For many years, social science research found evidence of Western industrialization's negative influence in the reduc tion of obligatory family responsibilities, the growth of nuclear living arrange ments, and the decrease in arranged marriages, among other factors. But more recent studies by social historians have established that many of the presumed aspects of family life in the Western past did not accord with the historical record. As summarized by Thornton and Lin (1994, p.4) These historical studies showed that the social and economic changes that had occurred in Western societies had not produced many of the family changes that several generations of earlier scholars had written about. In fact, the data showed that several key features of Western family life previously believed to have been the result of important social and economic changes were widespread in many Western soci eties before the other changes occurred. So revolutionary was this new historical research that many of the earlier conclusions have been la beled as myths in the recent literature. These revisions have led to more careful empirical studies of family relationships and the role of the elderly across historical periods and societies. Hareven (1982) and Quadagno (1982) illustrate this line of research for early industrial periods in the West. In a seminal work, Goode (1963) traced the relationship between in dustrialization and what he termed "the conjugal family" across societies. While recognizing the existence of nuclear families early in Western history and the lack of a simple relationship between modernizing forces and family structures world wide, he notes a broad shift across cultures from a familial focus on cross-generational relationships to an emphasis on the husband-wife relationship. This shift has been manifested by a greater choice of marriage partners among young adults, greater emotional bonds between husbands and wives, and more independent living among young couples, among other out comes. Goode is careful to note, however, that because these changes do not necessarily imply a weakening of other facets of family structure, they may coexist with many traditional family arrangements. Goode's wide-ranging research has been the source of a number of studies, some addressing the broader thesis and others offering detailed analyses of specific regions or societies. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 112 The Well-Being of the Elderly in Asia McDonald (1992), for example, attacks the idea of worldwide conver gence to the conjugal family form in the form set forth by Goode, but recog nizes that social change can lead to reconstitution of family systems via "com promise between ideology and economic structure" (p.23). Many analysts seem in accord with what Freedman (1993) termed a more general conver gence hypothesis, which recognizes movement from more to less dependence on kinship-based institutions in developing societies as a result of socioeco- nomic change, but with significant variation in the forms taken by this trans formation across societies. Wilson and Dyson's (1992) analysis of India and Caldwell and Caldwell's (1972) examination of agricultural peasant societies are in accord with this formulation. Other studies have explored family changes in East Asia over the past several decades. Thornton and Lin (1994), using data on many aspects of family structure gathered in Taiwan during the rapidly changing period fol lowing World War II, examine in some detail which aspects of marriage and family have changed and which have remained relatively intact. Mason et al. (1998) examine the impact of industrialization on several dimension of fam ily life in an analysis comparing several Asian countries to the United States. Whyte and Parish (1984) examine urban life in China in the 1970s to trace the nature of the changes arising from political and economic forces and their similarity to changes elsewhere. Mason (1992) develops a model for tracing how the effects of changes in industrialization, urbanization, and migration on the family in Asia are likely to impact on the care of the elderly. Whyte et al. (forthcoming) utilize contrasts between urban Taiwan and the city of Baoding, PRC, to illustrate that the concepts of modernization and development need to be carefully specified in order to trace potential effects on the family. Although China is less economically advanced than Taiwan, on a number of measures the older people and their children in Taiwan appear more traditional in terms of family living arrangements and exchanges than do their counterparts in Baoding. This comes about, in the view of the authors, in part from the actions of the Chinese government in the 1950s and 1960s to promote a "proletarian culture" distinct from Confucian, Western and, after 1960, Soviet cultural influences. At the same time, the rapid economic growth in Taiwan was distinctive in the prominence of family controlled firms and investments in the process, which maintained to a high degree family controlled work settings as against non-familial enterprises, and served to maintain several aspects of traditional family relationships (Whyte et al., forthcoming). These contrasts highlight the importance of paying close attention to the "micro- institutional" context within which individual and family behavior operates, as well as the broad path of economic development. A related caution is expressed by Thornton and Fricke (1987) with their attention to the importance Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Theoretical Perspectives, Measurement Issues 113 of the "familial mode of social organization," and the need for carefully specifying the contexts was noted by Goode in his original formulation. In terms of Figure 4.1, the foregoing theories and analyses examine the potentially deleterious effects on the well-being of the older population from modernizing forces either directly, or indirectly through their effect on family organization, and the intergenerational support system. It is well to stress, however, that careful empirically testing these hypotheses requires studies over time in which societies of different types are traced to study the impact of these forces, as recognized by Goode (1963) and Thornton (1991) among others. Nevertheless the theoretical underpinnings have considerable relevance for current and future research on population aging in Asia. As previously established, many countries in the region have undergone and are undergoing rapid demographic, social, and economic changes that are transforming the industrial base, the level of urbanization, and the educational level of younger adults. Family size is shrinking, life expectation is increasing, more women are entering the labor force, and new ideas and life styles are widely commu nicated and rapidly adopted. Given these circumstances, a prime area for re search is the impact of these changes on the well-being of the elderly in terms of their health, their labor force participation and economic status, and their family involvement. A second, related, major research objective is to ascer tain the current state of key familial dynamics in terms of living arrange ments, support and exchange arrangements, and levels of social contact, au thority, and respect. The goals here are twofold: to trace how these impact on the well-being of the elderly and to judge the extent to which these basic societal systems are themselves changing. Although the latter cannot be ac complished from a cross-sectional survey, it may be possible to glean on the basis of knowledge of past family arrangements the degree to which various elements may be changing in response to ongoing societal transformations, and giving rise to new modes of accommodation. And of course well ex ecuted current studies serve as benchmarks for monitoring future trends in these basic systems. In focusing on the broad question of the effects of population aging on the welfare of the elderly we should not lose sight of the importance of individual characteristics and lifetime experiences on well-being. As a result of their common history, older people in any given age cohort have been influenced by elements that distinguish them from older and younger groups, ranging from the relative size of their cohort, to the social, economic, and political forces predominant during their different life stages. However, older people within any cohort have considerable variation in their experiences with family support systems and non-familial institutions, which greatly influence Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 114 The Well-Being of the Elderly in Asia their well-being. Thus, the life course perspective (see Thornton and Lin, 1994; Elder, 1987) presents both an important framework for understanding the current situation of the older population and a measurement challenge, given the long and varied histories of older individuals (Hermalin, 1993). Consideration of the life course also reminds us that well-being is a multi dimensional concept and that advantages on certain facets can be associated with disadvantages on others. For example, an analysis by Easterlin et al. (1993) addresses the question of whether the large baby-boom cohort in the United States (roughly those bor n between 1946 and 1964) are likely to be worse off economically than earlier smaller cohorts. Using cohort analysis, which compares the baby-boom cohort at various life cycle stages with earlier cohorts, they conclude that the baby-boomers are not likely to end up at an economic disadvantage. At the same time, the authors look at several family- related characteristics of this cohort—marriage and divorce rates and number of children—and conclude that on these aspects of well-being the baby- boomers may be at a disadvantage in comparison with earlier cohorts. In the analytic chapters that are to follow we incorporate aspects of the life course perspective and the variation across individuals by paying close attention to the way that the several dimensions of well-being vary by sociodemographic characteristics that reflect family dynamics, residence, educational attainment, and occupational history (as well as age and gender). We realize that these are only broad surrogates for the myriad biological and social influences each individual faces. In sum, the research objectives set forth have several purposes. On one hand, they provide snapshots—point-in-time analyses—of the well-being of older populations living in societies going through considerable demographic, social, and economic change. As such, they serve as detailed benchmarks for future studies that will help establish the impact of such change on key insti tutions and the welfare of the older population. More immediately, carefully gauging levels of well-being on several dimensions and understanding how these levels vary with key sociodemographic characteristics are important inputs for policymakers who must make decisions on the kinds of programs to develop to address the needs of the older population. The Effect of Population Aging on Economic Development Population aging has important implications for many facets of an economy, including economic growth, labor force dimensions, saving rates, and income inequality among other dimensions (Johnson and Lee, 1987; Kelley, 1988). Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Theoretical Perspectives, Measurement Issues 115 The connections involve a number of basic theoretical perspectives and have been tested with a variety of models, data, and methods. Some relationships appear well established, while others remain inconclusive, varying with the data and analytic techniques employed. On some topics several key questions are unresolved. We discuss and illustrate here only a few of the connections between population aging and economic outcomes, focusing on those most relevant to the situation in Asia and which intersect with important policy issues. These include the relationships between aging and labor force rates, economic growth, savings, and transfer systems. The most direct effect of changes in fertility and mortality on eco nomic parameters is through their effect on age structure and hence on the proportion of the population that is of working age. As shown in Chapter 2, the dependency ratio which measures the working age population to the non- working older and younger segments is often used to capture this relation ships. The analysis by Bloom and Williamson (1998) of 78 countries finds a strong relationship between the growth rate of the economically active popu lation (a related measure) and the economic growth rate between 1965 and 1990. They argue that East Asia's so-called economic miracle was fueled in part by its demographic transition during this period in which the working age population grew much faster than its dependent population, helping to expand the per capita productive capacity of these countries. The authors caution, however, that this relationship between demographic structure and economic growth is not inevitable but dependent on having appropriate so cial, economic, and political institutions in place that allow the potential af forded by the demographic changes to be realized. (See also Bloom et al., 2000, for related analysis and discussion). It is worth noting that these recent analyses help resolve a long-standing debate about the effect of population growth per se and economic growth, which has resulted in contradictory find ings. By utilizing the components of population change and the resulting age structure more explicitly in modeling and analysis, the manner in which de mographic change interacts with economic change has been clarified. (See Bloom and Williamson, 1998; Bloom et al, 2000; and A. Mason, 1987, for further discussion of the issues.) A related type of analysis of the potential economic impact of population aging is provided by Easterlin (1991), who uses historical data for ten European countries and the United States to trace the relationship between past population growth rates and growth in per capita gross domestic product. Examining these past trends and past and future dependency ratios (of the type discussed in Chapter 2), Easterlin finds little empirical evidence that declining population growth has slowed the rate of economic growth and concludes that projected Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 116 The Well-Being of the Elderly in Asia total dependency ratios are in line with past experience (given the trade-off between fewer young dependents and more older dependents). Beyond the general question of the relation of population change to economic growth, there is particular interest on how demographic trends and structure impinge on savings. A formulation by A. Mason et al. (2001, p. 116) views per capita growth in income as a function of the growth rate of the number of workers per capita (the economic support ratio) and the rate of growth of output per worker. (More precisely, the formulation uses "equivalent adults" and "effective workers" to take into account differences in consumption and productivity with age.) The component representing workers per capita is reflected in the dependency ratio as discussed above and in Chapter 2. The component representing output per worker is affected by technological changes and by physical and human capital per worker. The relation of demographic change to technological change is ambiguous according to the authors but under certain conditions population aging can lead to substantial increases in capital per worker. Several recent analyses indicate that savings rates will increase during the portion of the demographic transition when the ratio of workers to population is increasing (or the dependency ratios are decreasing) but the relationships are complex and not all the research points in the same direction. In the main, a number of aggregate level studies of international time series found a close relationship between demographic change and savings (see reviews in Lee et al., 2000; Bloom and Williamson, 1998), while analyses at the household level are not generally supportive (see for example the findings for Taiwan and the related discussion in Deaton and Paxson, 2000b). In an effort to reconcile these differences, Lee et al. (2000) undertook microdata- based macro simulations of the relation of life cycle savings to the demographic transition in Taiwan and found that "under the assumption of pure life cycle savings, aggregate savings rates would decline modestly during the early stages of the demographic transition, then rise quite substantially during a long middle period, and then decline again as the population aged rapidly in the last stage of the transition." (Lee et al., 2000, p. 195) As described earlier, the need to smooth consumption over the life cycle gives rise to various private and public mechanisms for achieving this goal. In many developing societies, transfers from children to parents are a major source for ensuring adequate consumption in later life. In many indus trialized countries, pay-as-you-go social security systems, which represent another form of transfers across generations, have become a dominant form of retirement income. The importance of transfer systems of different types within all societies and their potential implications on savings behavior, and hence economic growth, have led to extensive modeling of the transfer sys tems in operation within an economy and their relation to population change Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Theoretical Perspectives, Measurement Issues 117 and structure. Formulations by Lee in a series of papers (1994a, 1994b, 2000) provide the basis of an accounting framework that uses age-specific rates of consumption and income, and of the relevant benefits, taxes, or exchanges that go into each transfer system, along with relevant demographic fertility and mortality schedules. These schedules and the underlying demographic and economic assumptions enable one to calculate the average age of receiv ing versus giving for each type of transfer, the average amount of the transfer, and the lifetime transfer wealth. (For a description of this accounting frame work, see National Research Council [2001, Chapter 5]). The consequences of population aging on levels of consumption arise from both the slower rate of population growth and the changing age compo sition associated with decreasing fertility and mortality. In the absence of any other change, slower population growth perse enhances potential consump tion because a smaller share of output is needed as investment to maintain a fixed amount of capital per workers. But when the changing age structure is taken into account, potential consumption may decrease because of the "in- tergenerational transfer effect." This factor arises whenever net aggregate trans fers take place between members of different age groups, as in the case of support of older parents by young children, pay-as-you-go social security pro grams, and other societal mechanisms for smoothing out the need for con sumption across the life cycle, as noted above. In the models developed by Lee, the intergenerational transfer effect is captured by the difference be tween the average age of consumers in a society and the average age of pro ducers. When this difference is negative (i.e., the average age of consumers is less than that of producers) this signifies a net transfer from the older to the younger population and a concomitant enhancement of overall societal con sumption with population aging. A positive difference represents a net trans fer from the younger to the older population and a reduction of overall soci etal consumption with population aging. Two important insights arise from this economic accounting framework. Empirical measures of net transfers across different types of societies show that they flow from the older to the younger generations in primitive societies and some in early stages of development. In the more industrialized societies such as the U.S., Japan, and England, however, net transfers flow from the younger to the older generations (National Research Council, 2001, Figure 5.1). But these models also demonstrate that the flow of transfers may be in opposite directions for different components of the overall transfer system. In the United States, for example, there is a strong upward transfer across ages arising from the public sector social security and health programs, at the same time that there are strong downward net transfers arising from the family sector, through bequests, gifts, child costs, and support for higher education. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 118 The Well-Being of the Elderly in Asia Intergenerational transfers are closely related to the process of wealth accumulation and thus have several important policy implications. In terms of the relation of population growth to economic growth as formulated by A. Mason, et al. (2001) as described above, the potential impact of population aging on increases in capital per worker will depend to a great extent on how the needs of the older population are met. Where transfer systems dominate, the demographic transition to an older population will not necessarily contribute to greater accumulation of capital or real wealth, and thus not enhance the growth of income per capita. As stated by A. Mason et al. (2001, p. 121-122): From the perspective of meeting retirement needs, systems based on transfers and systems based on the accumulation of real wealth are in one sense identical. Either system produces a stream of income during the retirement years. Likewise either flow represents wealth to that in dividual—real wealth in one case, transfer wealth in the other. From the perspective of the economy, however, real wealth and transfer wealth are very different. Real wealth is productive and contributes to im proved standards of living.... In contrast, transfer wealth contributes nothing to economic growth. (Mason et al., 2001, pp. 121-122) There are also a number of important research questions associated with these economic insights. Will families reduce their support for older parents or their own savings behavior after social security programs are insti tuted, or will older parents increase their private transfers to children, realiz ing the tax burdens they may be facing under a pay-as-you-go system? These and related questions cannot be answered from the economic accounting models alone; they require additional research on the behavioral and motiva tional aspects of familial transfers. The uncertainties associated with savings behavior and trends in the national savings rate have led to some caution that a fully funded pension program may not generate an increase of investments for development, and does not address questions of equity (Johnston, 1999). These policy questions are taken up in the concluding chapter. Some of the challenges involved in carrying out research on intergenerational exchanges are developed below where measurement and related issues pertaining to the analytic chapters are discussed. We turn first to a short review of past research on aging in Asia that touches on some of these aspects. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Theoretical Perspectives, Measurement Issues 119 Asian Interests in Research on Aging Questions about the interrelationships among economic growth, aging popu lation structures, and the welfare of the elderly that have gained prominence within industrialized countries have also caught the attention of many less industrialized nations. Researchers and policymakers have interests in both the humanitarian and developmental implications of population aging, par ticularly in East and Southeast Asia where extended living arrangements and support for older parents by their children are traditional social systems. Inso far as forces associated with industrialization and modernization undermine these social systems and otherwise adversely affect the welfare of the elderly, they raise the need for social welfare programs to take their place. Policymakers are also concerned that the adverse age structures associated with population aging will limit their potential for continued economic development and thus the resources required to pay for these programs, as well as other social and economic goals on their planning agendas. Asian policymakers and researchers have displayed prescient attention to these implications of population aging. Although only two Asian societies, Japan and Hong Kong, had more than 10 percent of their populations in the 60 or older age group as of 1990, there has been an outpouring of confer ences, research projects, books, and papers about aging in Asia since the early 1980s. Overviews of these research efforts have been provided by Andrews (1992), Hermalin (1997a), Martin and Kinsella (1994, who also report on other developing regions) and Andrews and Hermalin (2000). These reviews describe the nature of many of the salient studies in the region, discuss re search issues, and provide references to the cross-national or multi-country studies and a number of the single-country efforts. Many of the multi-country studies have been sponsored by international agencies like the WHO Regional Office for the Western Pacific (Andrews et al., 1986; Andrews and Henrick, 1992); the Association of Southeastern Asian Nations (ASEAN) (Chen and Jones, 1989); and the United Nations Economic and Social Commission for Asia and Pacific (ESCAP). The latter organization has been particularly ac tive in promoting research and workshops on population aging throughout the Asian and Pacific regions and a list of their publications on this topic can be found in the Asia-Pacific Population Journal (1997, p. 6), as well as in Andrews and Hermalin (2000). Although it is difficult to summarize the wide variety of studies under taken in the region, it is probably fair to say that they tended to concentrate on measuring the welfare of the elderly on select dimensions, like health, while Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 120 The Well-Being of the Elderly in Asia assessing the state of the underlying familial system by measuring the living arrangements of older persons and the nature of the support they are receiving from children and other kin. In this sense these studies address the first of the broad theoretical issues discussed above—the effect of social and economic change on key family institutions that have traditionally served to support the older population. This motivation is consonant with policymaker concerns about population aging and the desire to discern at an early stage whether rapid changes in key societal institutions were underway. The emphasis on analyzing living arrangements is also consonant with the relative abundance of data on household structures, routinely gathered on censuses and surveys. Analyses of trends in coresidence and differentials in living arrangements by characteristic provide partial indicators of changes underway and can serve as harbingers of what might evolve over time. These research efforts were valuable in establishing the high levels of coresidence among older parents and their adult children that could be used as benchmarks for tracing changes over time (Martin, 1989). And in fact, although these findings demonstrated no imminent crisis in key familial arrangements supporting the elderly that required reactive policy and program implementation (Hermalin, 1997), the longer term trends now available for several countries indicate various de grees of change in the level of coresidence, raising questions about key deter minants and the future, as discussed in Chapter 5. It should also be said that the value of the data provided by these ear lier studies is compromised somewhat by conceptual and measurement prob lems. We address some of the broader issues here and take up more specific problems in discussing the analytic chapters below. In terms of Figure 4.1, a focus on living and support arrangements places undue emphasis on interme diate variables that affect the well-being of the elderly rather than the dimen sions of well-being. Though some of the earlier studies did give special atten tion to health, relatively little attention was paid to the income and wealth levels of the elderly, their work and leisure activities and preferences, or their level of their involvement with families and friends. Even within the realm of health, analyses were often limited to measuring rates of chronic diseases and degrees of disability rather than patterns of health behaviors or health care utilization and its associated economic factors. The focus on living arrangements and the receipt of support from chil dren also tends to slight the other intermediate variables operating and im plicitly assume that those not coresiding or receiving support will suffer vari ous disadvantages. However, even at earlier stages of their history, significant fractions of the older population in many of these countries lived indepen dently, or did not rely on their children or other kin for major support. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Theoretical Perspectives, Measurement Issues 121 Nevertheless, it remains true that living and support arrangements re main important intermediate institutions and their differences over time and across societies can be highly revealing. However, in pursuing these topics many of the earlier studies had insufficient empirical breadth and limitations in their measurement strategies. Perhaps most important, a formal structural definition of living arrangement fails to provide vital information about the content of the relationships and interactions. For example, older parents liv ing with married children may be recipients of considerable financial and emotional support, or they may be mainly aiding their children and grandchil dren with child care, shopping, and meal preparation. Older parents who do not reside with their children may maintain frequent contact but prefer inde pendence, and have the economic means to achieve this (or receive support from children to facilitate independent living). In the U.S., though support and coresidence patterns differ, contact with children remains high (Bumpass, 1990; Eggebeen and Hogan, 1990; Crimmins and Ingegneri, 1990).). At a more general level, despite the rapid socioeconomic and demo graphic changes ongoing in many developing countries, it is unlikely in the near term that existing social arrangements for the support of the elderly will be precipitously abandoned and replaced by completely new structures. Rather one can expect and can already discern a series of accommodations, as exist ing forms adjust to new pressures and roles, and relationships become rede fined accordingly. A number of family scholars have emphasized the impor tance of identifying which (and to what degree) aspects of familial life change in response to broader societal changes and which elements prove more en during (Goode, 1963; Hareven, 1977; Thornton and Lin, 1994; McDonald, 1992). In many parts of Asia where key familial arrangements center on the economic and moral authority of elders, it is likely that these accommoda tions will take the form of subtle redefinition of relationships and that changes in the forms of family structure will proceed more slowly. These considerations suggest that current and future research should devote greater attention to distinguishing the forms of living arrangements from the actual functions they serve, and not try to infer the content from the structure. One way to do this is to measure the functions and duties carried out by those living together, or near one another. There are two challenges here. One is to accurately measure the division of labor within the household (or network) and obtain a reasonable accounting of the mini-household allocations and contributions. For those living apart, more careful accounts of levels of contact, patterns of exchange, and degrees of indirect support are also required. The possibilities here are discussed further below and in Chapters 6 and 8. The second challenge is to gain a better understanding of the meaning Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 122 The Well-Being of the Elderly in Asia of these accommodations and shifts in relationships, and what these may presage for broader normative changes. One way to approach the latter is through focus group discussions and other qualitative data collection efforts that provide opportunities to explore the meanings behind activities that are difficult to capture in the usual survey instrument. As part of the research project in the four countries studied here, 18 to 24 focus group sessions were conducted in each country. Although data from these sessions are used only limitedly in the chapters that follow, they inform a number of the perspectives. (See Knodel, 1995 for a more detailed analysis of these data.) In related research, Ngin and DaVanzo (1999) used focus groups and ethnographic interviews to explore ideal living arrangements in Malaysia; Knodel and Saengtienchai (1999) used a quasi-qualitative case study approach in Thailand to study living arrangements of the elderly; and Hashimoto (1996) used a mix of participant observation, focused interviews, and surveys in a U.S. and Japanese community to study the cultural meaning of exchanges and support in the two cultures. The analytic chapters are organized in keeping with the framework discussed here and presented in Figure 4.1. Chapters 5 and 6 examine the two key intermediate structural arrangements, types of living arrangements and levels of support and exchange, providing a snapshot of current patterns, evi dence of recent change, and variation by sociodemographic characteristics. Chapters 7 through 10 are devoted to establishing levels and differentials of well-being among the elderly in the realms of work and leisure, income and assets, health status, and health care utilization. Each of these chapters dis cusses the overall level of well-being as revealed by specified measures, the variation in levels by individual characteristics (and which characteristics seem most salient as revealed by multivariate analysis), and the change over time in selected indicators. Chapter 11 reverses this style of analysis to some extent, addressing which sociodemographic groups in each country are disadvan- taged on one or more aspects of well-being, and how disadvantaged they are relative to other groups. In the sections below we address key conceptual and measurement issues associated with each of these chapters so that the chap ters themselves can focus on the empirical results. Measuring and Understanding Living Arrangements Living arrangements are often a form of intergenerational transfer and support, as when older parents coreside with children or children provide living expenses for separate quarters. These aspects have distinct measurement challenges, as described below. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Theoretical Perspectives, Measurement Issues 123 An important first step in understanding living arrangements and other forms of support is to "map" the kin available to the elderly respondents— that is, to identify the number of children and other relevant kin, their key characteristics, and their location vis-a-vis the elderly respondent. The num ber and location of kin help establish potential for coresidence and other forms of support, and serve, so to speak, as the denominators for observed levels of support. Chapter 5, for example, reveals that only 4 to 9 percent of the elderly across the four countries are in a situation where they are not living with a child, or without a child nearby, or not in at least weekly contact with a child. A second important step is to refine the definitions of dwelling and household used in many censuses and surveys. These instruments often de fine household as a single dwelling, which excludes children living in mul tiple-family compounds or in the same neighborhoods as their elderly par ents, a common arrangement in much of the developing world. For example, our survey in Taiwan revealed that while native Taiwanese over 60 tend to coreside with only one child, about 1.5 other children live near them in the same township or city. Moreover, while married sons are more likely to coreside than married daughters, daughters are more likely to live nearby, so that over all access to sons and daughters is more balanced than the coresidence data alone imply (Hermalin, Ofstedal and Chi, 1992, Table 4). In Thailand, Knodel and Chayovan (1997) find that a very high proportion of older persons who do not coreside with their children live adjacent to them or see them daily. Another measurement challenge can arise when survey data do not distinguish between married and unmarried coresiding children, although this often has implications for the direction of support, especially among rela tively young coresiding children. Moreover, demographic trends in age at marriage and childbearing can lead to increased levels of coresidence that have no implication for levels of parental support. Chapter 5 makes use of these more nuanced measures of living ar rangements in describing current levels and the factors associated with differ entials by sociodemographic group. Identifying and Measuring Intergenerational Support Dimensions Intergenerational support arrangements, along with other key aspects of the social fabric, are culturally defined. As noted above, a considerable impetus to the growth of aging research in East and Southeast Asia, as well as other regions, stems from concerns that the rapid social changes occurring on many fronts may place in jeopardy existing social welfare and familial arrangements for the elderly. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 124 The Well-Being of the Elderly in Asia Soldo and Hill (1993) review the theoretical orientations to intergen- erational transfers from sociological, economic, and psychological perspec tives, and Lillard and Willis (1997) provide a more detailed overview of the economic hypotheses on the motivations for transfer. The theoretical under pinnings for family transfer systems range from altruism, to shorter or longer term exchange arrangements, to security considerations (Frankenberg et al., n.d.; Schoeni, 1997). As modeled by Becker (1974), an altruistic head of house hold would allocate resources to members of the family in such a way that no one could be made better off without making someone else worse off. Altruis tic transfers to parents may be further encouraged by norms that stress filial responsibility. Andreoni (1989) extended the altruism model to include "warm glow" giving, indicating that providers may be motivated to do more than recipients require or expect. Other models of motivations for intra-family trans fers identify quid pro quo arrangements, where family members assist each other in times of need and transfers serve as insurance against risk. Exchange models include short- and long-term trade-offs among fam ily members. Shorter term transfers would include older parents taking care of grandchildren in exchange for financial or other tangible support from chil dren. Longer term exchanges include adult children supporting their parents in old age in return for the care and assistance they received while growing up. Parents may attempt to insure this support by retaining control over land or a family business. A related long-term exchange model focuses on support for older parents as repayment for the parents' provision of expenses for edu cation. As returns to education increase, so do the potential benefits to both children and parents of extending the children's years of education. Testing these various hypotheses has proven difficult because they re quire detailed data on the amount and nature of various transfers across many family members, often over a long period. Without such information it may be difficult to distinguish which motivating force is the prime determinant of an observed pattern of transfers. The collection of data about family transfers has grown more sophisticated in recent years but formidable measurement challenges remain, stemming from the inherent complexity and multidimen- sionality of the intergenerational system. Further progress is important both to facilitate testing of competing models and to gain an adequate description of the overall system, allowing changes that occur in response to demographic, socioeconomic, and cultural forces to be monitored. The discussion below illustrates various facets of such a system and some of the characteristics associated with each dimension. As Table 4.1 suggests, an adequate description of an intergenerational support system requires three broad elements: a "map" of the network in terms Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Theoretical Perspectives, Measurement Issues 125 of potential providers and recipients, their characteristics, and their locations in relation to one another; a designation of the potential types of exchanges, such as time or money; and a delineation of the content and qualities of all exchange types (e.g., how much and how often). In studying the elderly, intergenerational transfers are conceived as a series of exchanges between an indexed elderly person or couple (i.e., the focal point of the investigation) and his and/or her family network. The definition of network is not straightforward, as the relationships likely to be salient may vary across societies and over time within societies. In some societies for example, siblings may be important sources of exchange, while in others they are not. "Mapping" the number, location, and characteristics of kin, as described above, serves as the denominator against which the frequency and nature of exchanges can be judged. It is important to distinguish those elderly who do not receive certain kinds of support (e.g., physical assistance) because they lack potential or nearby caregivers from elderly who do not receive support despite the availability of kin. In addition, the size and characteristics of the network are important for testing competing theories about the motivations and dynamics of exchange, such as the differences in support provided by sons versus daughters in certain settings (see for example Lee et al., 1994). Soldo and Hill (1993) distinguish the three major transfer "currencies" as space, money, and time, displayed in Table 4.1 as the broad categories of exchange. Space, of course, refers mainly to living arrangements discussed in some detail above. In addition, in terms of exchange, it is important to ascer tain the home owner (and if it's the child, whether the parent assisted in the purchase or transferred title) and the purpose of coresidence (whether it's mainly to provide assistance to the adult child or to the parent). These dimen sions will help establish the direction and significance of the exchange. A high proportion of the elderly in developing countries report receiving support in the form of money or its equivalents in material goods. To understand the nature and significance of exchanges involving money and material goods requires attention to several dimensions. Because the magnitudes involved are clearly salient, it is often helpful to know how regularly or frequently transfers are made and in what amounts. The significance of material exchanges can be gauged by the purpose of the transfer (to help meet basic living needs, supplementary needs, or other purposes) and the degree of impact the transfer has on both the recipient's needs and the provider's resources. Identifying past transfers can be important in understanding current transfers, and for testing theories of altruism versus reciprocity. Whether or not parents have given children gifts or loans in the past for education, travel, business, or other purposes can greatly affect the pattern of exchanges currently observed. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 126 The Well-Being of the Elderly in Asia It is also important to be sensitive to the possibility that material support can be provided indirectly, by paying the rent, taxes, or other costs for the recipient rather than by a direct transfer of funds. The third currency of exchange, time, also requires elaboration along several dimensions. An important consideration is the purpose. Are the elderly receiving help with physical functioning (often referred to as ADLs—Activities of Daily Living) or with basic household activities (the so-called lADLs— Instrumental Activities of Daily Living), or is the contact mainly for emotional support, companionship, or participation in leisure activities? Here also the magnitude of the exchange in terms of frequency and amount of the time spent may be salient. Providers may also make indirect transfers of time by purchasing third-party assistance for recipients, especially in societies where household help is common. Other indirect transfers of time may arise in the division of labor for major household tasks. For example, an elderly parent may be a net provider of time to others through participation in household cooking and childcare, or a net recipient because of limited involvement. It is not surprising in the light of all this complexity that no standard strategy has as yet evolved for capturing the essential elements of an intergenerational support system. There are almost as many approaches as there are questionnaires, as each team of researchers achieves a unique com promise among the constraints and challenges. (For a review of varied ap proaches, see Hermalin, 1997b, and the National Research Council, 2001, Chapter 5.). This lack of uniformity limits the potential for comparative analy ses. One can conceive the end result of collecting all the transfer informa tion implied by Table 4.1 as a large matrix defining the nature and magnitude of various types of exchanges between a focal elderly person and the mem bers of his or her network. These matrices would be the input for constructing measures that capture the social and economic aspects of the elderly person's support system, such as the size and structure of the network, the number of supports received and provided, the sufficiency of support, who the major providers of support are, and whether the support provided is shared across members of the network or concentrated in a few providers. Although the data collected in the four study countries were not suffi cient to produce all the measures enumerated above, they provided enough detail and comparability across surveys to tap a number of key dimensions. Accordingly, Chapter 6 reports on the proportion and characteristics of the elderly receiving each major type of support, the providers of these supports, and the degree of diffusion of support for those elderly with multiple needs. Other aspects of exchange, such as the support provided by the elderly and the flows of direct and indirect support will also be covered in this chapter. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Table 4.1. Outline of Intergenerational Support System SPACE Elderly Unit Characteristics of couple or individual MONEY TIME Dimensions of Intergenerational Transfers Participants: Direction: Purpose: Magnitude: Timing: Form: Impact: Kin Network Relationships Size Location Dyads Involved Underlying Network Characteristics of individual members Provider and Recipient of Each Exchange Basic or Supplementary Support Physical Care; Household Assistance; Child Care; Companionship Amount of Money or Value of Goods Amount of Time Exchanged Regularity and Frequency of Exchange Chronological Time: Current or Past Persistence of Exchange over Time Direct vs. Indirect Gift or Loan Needs of Recipient Resources of Provider Effect on Recipient and Provider Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 128 T/ie Well-Being of the Elderly in Asia Measuring Work and Leisure The nature and duration of the work that people perform over their lives is related to their economic position and health status at older ages. Chapter 7 looks at the main occupations of the elderly in the four countries, the degree to which they remain in the labor force, and the nature of withdrawal from the labor force for those who are not working. The degree to which the elderly are subject to mandatory retirement practices and their coverage by pension pro grams are also addressed. Work is one component of the broader set of activi ties older people may engage in, and to fill in the picture, Chapter 7 also analyzes data collected on the nature of their leisure activities. (Time spent in looking after others and household duties are covered primarily in Chapter 6.) Conceptually, assessing the well-being of the elderly in terms of work and leisure activity levels presents a different challenge than looking at eco nomics or health, where absolute levels can be readily defined. In this case, comparisons between the current and preferred statuses of the elderly will help gauge their well-being. For instance, do those out of the labor force wish they could work, or do those working wish they had less involvement? For those not working, an assessment of their well-being on this dimension re quires knowing what activities they take part in, how involved they are, and how much companionship they experience, among other aspects. In addition to characterizing elderly well-being, measures of the level and nature of labor force participation at the older ages are crucial for coun tries with aging populations because they figure in economic growth and pro ductivity, the cost of social security programs, and the need for community- based centers and programs geared to the elderly. As noted in Chapter 3, those in the formal sectors of the Asian economy—employees of the government and of large corporations—generally must officially retire at a relatively early age, but the prevalence of family farms and businesses has allowed many older individuals to sustain economic activity into advanced years. In addi tion, several of the developing and newly industrialized countries in Asia have followed a pattern observed in developed countries and raised mandatory retirement ages to help reduce pressure on pension and social security pro grams and sustain economic growth. Other forces, however, such as the en hanced resources of the elderly's children and the changing potential for smaller family enterprises, may encourage earlier withdrawals from the labor force. The similarities and differences between the more industrialized countries and the developing world with regard to both the formal and informal dynam ics of retirement and leisure offer rich analytic potential. Chapter 7 sets forth the basic patterns of work, retirement, and leisure in the four countries and traces their implications for the policy challenges these countries are facing. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Theoretical Perspectives, Measurement Issues 129 The Economic Status of the Elderly The framework outlined in Figure 4.1 shows economic status as another di mension of well-being that is key to policy development. The rapid social and economic development experienced by many countries in East and Southeast Asia might be expected, a priori, to increase the economic vulnerability of many elderly. This expectation arises from the potential obsolescence of their traditional skills with industrialization, the absence of wide-spread public- or private-sector retirement programs, the generally rising price levels, and their potential isolation as children migrate from rural areas to jobs and opportuni ties elsewhere. Countering these negative influences is the potential for the current elderly, through their education or businesses/land holdings, to pros per in the course of their countries' economic development, and for them to enjoy increased levels of support from their more affluent children. How these opposing trends have played out for the current elderly is the major focus of Chapter 8, where we investigate a variety of economic well-being measures for different groups of elderly. Although the measurement challenges for economic well-being may be less complex conceptually and operationally than those for intergenerational transfers, they are far from absent. Although income, current and past, is used as a prime indicator of economic well-being in nearly all settings, eliciting reliable estimates in a survey is problematic. Non-response tends to be high and non-random. In less developed economies with large proportions of the population in farming or small business, those so-engaged may have diffi culty providing net income accurately, as might those with multiple sources of income. Among the elderly these basic issues may be compounded by memory problems, lower involvement in economic affairs, and the multiplic ity and irregular flow of income sources. The asset and wealth positions of older individuals, which constitute important components of their economic status, are even more difficult to ascertain. In addition to the limitations noted for income, older people may hold property jointly with children or others, and current market values may not be known. Extended living arrangements, which characterize much of the region under study, further complicate assessing the economic status of an elderly respondent or couple because one needs to distinguish the income of the elderly from the income of the household, the direct and indirect flows that can take place, and the relevant items of consumption and savings. Indeed, as Deaton and Paxson (1990) note, the analysis of savings behavior prominent in more industrialized countries, which centers on permanent income and life-cycle models, may be less applicable to countries where extended households are common, since the economic goals of such households may be protection of Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 130 The Well-Being of the Elderly in Asia living standards from short-term risks rather than transference of resources between generations. To help address this problem, they devise a method of estimating individual age-saving profiles from household data (Deaton and Paxson, 2000a). Chapter 8 addresses the level of economic well-being in each of the four countries using a multiplicity of measures reflecting absolute, relative, and attitudinal dimensions of income, assets, and wealth. Although the num ber and nature of the measures are not identical across countries, there is sufficient overlap to permit detailed comparisons on several dimensions. Health Status and Health Care Utilization A great deal of concern about the effects of population aging stems from the rising disability levels of an increasing older population and the attendant costs. Changes in life expectancy at older ages can be tracked quite directly (although large-scale deficiencies in registration systems exist in many coun tries), but tracking levels and trends in disability is more difficult. A number of important debates in the biodemography of aging center on patterns of mortality at very old ages as over-all life expectancies advance (Wilmoth, 1997; Olshansky et al., 1990). A related strand focuses on the lev els of disability associated with increasing longevity. Will increased life ex pectancy be accompanied by more years of active life or by more years of disability or restricted functioning? The compression of morbidity hypothesis set forth by Fries (1980) argues for the former scenario, as does some recent U.S. data that point to lower levels of disability at older ages in the past few years (Manton and Gu, 2001). The issue, however, is far from resolved. As Chapter 2 indicates, the four countries under study still have a rela tively "young" older population, with approximately 60 percent between the ages 60 and 70. But this population structure will yield rapid increases over the coming years in the "oldest-old" (those over age 75 or 80) in these and other developing and newly industrialized countries in Asia, which in turn will greatly increase the numbers with chronic diseases and various types of disability. Because the health systems in many developing and newly indus trialized countries have been oriented more to fighting infectious diseases and addressing the needs of mothers and children than caring for the elderly, new resources will be needed to address these changes. Policymakers should also keep in mind that once a country achieves low fertility, future mortality levels become increasingly decisive in determining the overall proportion of elderly as well as the age distribution within the older ages. Given the uncer tainty of future levels of mortality at older ages in developing—as well as Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Theoretical Perspectives, Measurement Issues 131 industrialized—countries, policymakers must anticipate that the number and proportion at the oldest ages may be even higher than now projected. In addition, developing and newly industrialized countries need to learn more about the health conditions and health care needs of their older popula tions. Many have rather limited information on the prevalence of chronic con ditions, the levels of physical and mental functioning, and the rates of health care utilization among the elderly. These data are important components for estimating future demand for services and costs, and hence the types of health care programs and policies that should be developed. Data on conditions and utilization come from the providers of health services (hospital and doctors), epidemiological studies of special groups or catchment areas, and population surveys on health conditions and behavior, health care utilization and costs, and self-perceptions of health status (or self- reported health). In some cases it has been possible to combine self-reports with a small number of anthropometric tests and/or to incorporate elements of existing cognitive tests (e.g., the mini-mental exam). Sufficiently detailed questions on health and related dimensions in population surveys greatly enhance the potential to examine differentials in health and health practices by salient socioeconomic variables and to test models about their interrelationships. Of the survey data available in the four countries in this study, all but Singapore's included a considerable amount of detail on health. For this reason, two chapters are devoted to the subject. Chapter 9 focuses on the health status of the elderly, including levels of self- reported health by characteristic, the prevalence of select chronic conditions, the proportions experiencing difficulties with eating, dressing, and other activities of daily living, measures of depression and cognitive scores, and information on current and former smoking behavior. This information by characteristic of the elderly for nationally representative samples provides a backdrop for over-all assessments of the health of the current older population and guidance for future needs and costs. Chapter 10 looks more directly at levels of health care utilization and perceived needs by examining the proportions that have used various ser vices, the coverage associated with these services, and the levels of unmet need. The variation in health insurance coverage across the countries, de scribed in Chapter 3, makes the current levels of utilization important bench marks for assessing future trends as new programs are introduced. Vulnerable Groups and Their Levels of Disadvantage A theme in the social change literature that is especially prominent in discus sions of population aging is the adverse effect of transformations on groups Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 132 The Well-Being of the Elderly in Asia ill-equipped to adjust to the underlying changes. As Chapter 3 points out, many of the broad social welfare programs and programs oriented toward the elderly arose to mitigate the effect of family changes and industrial organiza tion on the older age groups. Chapter 11 uses the survey data from the four countries to identify the sociodemographic groups among the elderly that may be particularly vulner able to adverse outcomes. These include women, the least educated, those residing in rural areas, and several others. Then assessments are made of each group's relative risk of experiencing different types of adverse outcomes on the well-being measures developed, in comparison to those not in these par ticular groups. Chapter 11 also makes use of data from each country on the extent to which older respondents are aware of and make use of various pro grams directed toward the elderly. This knowledge, combined with the find ings on the degree of disadvantage of the most vulnerable groups, can assist policymakers in assessing how well current programs are functioning and what changes might improve their effectiveness. Cross-Cutting Issues, Transitions, and Policy Recommendations Despite the amount of detail to be presented in the next seven chapters, sev eral important questions remain to be discussed. One might be termed cross- cutting issues—the analyses that involve interrelationships among the mea sures of well-being or detailed testing of various hypotheses about factors contributing to levels of well-being. For example, what is the relationship between health status and economic status? Does poor health lead to poor economic status or does low income produce poor health? Although the level of association between these two outcomes at any given point may be mea sured, establishing the causal connections is much more difficult. Other inter relationships that are difficult to sort out in a cause-effect manner include: • The relationship between coresidence and health. To what extent does coresidence with children affect health or health status lead to coresidence? • The relationship between health and retirement. To what degree does health status prompt retirement and withdrawal from the labor force affect health? • The relationship between assets and retirement. To what extent do asset levels affect retirement or retirement affect the level of assets? Although the forthcoming chapters analyze these interrelationships and dis cuss some of the causal issues, it must be recognized that only limited progress on these issues can be made from cross-sectional surveys. Unraveling the Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Theoretical Perspectives, Measurement Issues 133 cause-and-effect questions requires longitudinal data best obtained from panel studies in which representative samples are reinterviewed over time. Panel studies have the advantage of capturing many of the key transitions older people make after they reach age 60 and, increasingly, survive to age 80 and beyond. The very concept of aging is associated with certain key transitions such as the movement from work to retirement, from independent living to coresidence with children or other forms of assisted living, and from good health to lower levels of functioning. In addition to sorting out the cause-and- effect issues illustrated above, panel study data will help establish the timing of these changes, their magnitude, and their differentials across sociodemographic groups. Transition rates can be very useful to policymakers in identifying what kinds of programs are most important to what groups at what ages. Of the four countries studied here, only Taiwan started with a panel survey, with multiple interviews of the original 1989 sample as described in Chapter 1 and Appendix A. However, a special effort was made in Singapore in 1999 to follow up the 1995 sample, and this produced useful panel data for a large subset of respondents. In the Philippines, follow-up efforts are under way in two provinces for the 1996 respondents, both as a test for a larger effort and to gather changes on a few key topics. Although the insights emerg ing from these longitudinal data are incorporated here only to a limited ex tent, several examples are included in the relevant substantive chapters and Chapter 12 devotes a section to transitions. The aforementioned transitions refer to changes that occur to a cohort of older people as they age. It is also important from a research and policy viewpoint to take into account the changes that occur as a result of new co horts coming along who may have very different characteristics than their predecessors. Particularly in the light of the rapid demographic and socioeco- nomic changes that have taken place in East and Southeast Asia, it is critical for policymakers to realize that future cohorts of elderly will be very different from current cohorts in terms of their characteristics, preferences, and needs, and that policy development must look beyond current perspectives. Although future cohorts will have fewer children on average, they will be more urban and much better educated than current cohorts, which is likely to translate into different health needs, different preferences for living arrangements and life styles, and different tastes for work and leisure. Chapter 12 demonstrates these differences through a series of cohort analyses and discusses their im plications for policy formation. The final chapter, Chapter 13, also expands on the policy implications of the substantive analyses presented, and discusses future research direc tions that can help guide policy development in the years ahead. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 134 The Well-Being of the Elderly in Asia REFERENCES Achenbaum, W. Andrew. 1982. "Further Perspectives on Modernization and Ag ing: A (P)review of the Historical Literature." Social Science History 6 (3):347- . 1985. "Societal Perceptions of Aging and the Aged." In Robert H. Binstock and E. Shanas, eds., Handbook of Aging and the Social Sciences, 2nd edition, 129-148. New York: Van Nostrand Reinhold. . 1996. "Historical Perspectives on Aging." In Robert H. Binstock and Linda K. George, eds., Handbook of Aging and the Social Sciences, 4lh edi tion. San Diego: Academic Press. Achenbaum, W. Andrew and Peter N. Stearns. 1978. "Essay: Old Age and Mod ernization." The Gerontologist 18 (3):307-313. Andreoni, James. 1989. "Giving with Impure Altruism: Applications to Charity andRicardian Equivalence." Journal of Political Economy 97(6): 1447-1458. Andrews, Gary. 1992. "Research Directions in the Region: Past, Present, and Fu ture." In David R. Phillips, ed., Aging in East and South-east Asia. London: Edward Arnold. Andrews, Gary, Adrian J. Esterman, Annette J. Braunack-Mayer, and Cam M. Rungie. 1986. "Aging in the Western Pacific: A Four Country Study." West ern Pacific Region World Health Organization. Research Report No. 1. Ma nila, Philippines: WHO. Andrews, Gary, and M. M. Hennink. 1992. "Ageing and Support in Three South East Asian Countries," Asia-Pacific Journal of Population Studies 7:127. Andrews, Gary, and Albert I. Hermalin. 2000. "Research Directions in Ageing in the Asia-Pacific Region: Past, Present and Future." In David R. Phillips, ed., Aging in the Asia-Pacific Region, 51-81. London: Routledge. Asia-Pacific Population Journal 1997. "List of Publications Issued by the ESCAP Secretariat Related to Ageing Issues in ESCAP Countries" 12(4): 16. Becker, Gary. 1974. "A Theory of Social Interactions." The Journal of Political Economy 82(6): 1063-1093. Bengtson, Vern, and K. Warner Schaie. 1999. Handbook of Theories of Aging. New York: Springer Publishing. Binstock, Robert H., and Christine L. Day. 1996. "Aging and Politics." In Robert H. Binstock and Linda K. George, eds., Handbook of Aging and the Social Sciences, 4th edition, 362-387. San Diego: Academic Press. Binstock, Robert H., and Linda K. George, eds. 1990. Handbook of Aging and the Social Sciences, 3rd edition. Dan Diego, CA: Academic Press. . 1996. Handbook of Aging and the Social Sciences, 4th edition. San Diego: Academic Press. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Theoretical Perspectives, Measurement Issues I35 Binstock, Robert H., and Ethel Shanas, eds. 1976. Handbook of Aging and the Social Sciences. New York: Van Nostrand Reinhold Co. . 1985. Handbook of Aging and the Social Sciences, 2nd edition. New York: Van Nostrand Reinhold Co. Birren, James E., ed. 1996. Encyclopedia of Gerontology: Age, Aging, and the Aged. San Diego, CA: Academic Press. Birren, James E., and Vern L. Bengtson, eds. 1988. Emergent Theories of Aging. New York: Springer Publishing Co. Birren, James E., and K. Warner Schaie. 1996. Handbook of the Psychology of Aging, 4lh edition. San Diego, CA: Academic Press. Bloom, David E., David Canning, and Pia N. Mclaney. 2000. "Population Dy namics and Economic Growth in Asia." In C.Y. Cyrus Chu and Ronald D. Lee, eds., Population and Economic Change in East Asia, pp. 1257-1290. New York: Population Council. Bloom, David E., and Jeffrey G. Williamson. 1998. "Demographic Transitions and Economic Miracles in Emerging Asia." World Bank Economic Review 12(3):419-456. Bruhn, John G., Billy U. Philips, Paula L. Levine, and Carlos F. Mendes de Leon, eds. 1987. Social Support and Health: An Annotated Bibliography. New York: Garland Publishing. Bumpass, Larry L. 1990. "A Comparative Analysis of Coresidence and Contact with Parents in Japan and the United States." NSFH Working Paper Series, 41. Madison: Center for Demography and Ecology, University of Wisconsin. Caldwell, John C. 1996. "Demography and Social Science." Population Studies 50:3051. Caldwell, John C., and Pat Caldwell. 1992. "Family Systems: Their Viability and Vulnerability." In Elza Berquo and Peter Xenos, eds., Family Systems and Cultural Change, 46-66. Oxford: Clarendon Press. Chen.AiJu, and Gavin Jones. 1989. "Ageing in ASEAN: Its Socioeconomic Con sequences." Final Inter-Country Report of the ASEAN Ageing Project, Phase 3 ASEAN Population Programme, Inst. of Southeast Asian Studies, Singapore. Cooper, Sheila Mclssac. 1999. "Historical Analysis of the Family." In M. B. Sussman, S. K. Steinmetz, and G. W. Peterson, eds., Handbook of Marriage and the Family, 13-37. New York: Plenum Press. Cowgill, Donald O. 1974. "The Aging of Populations and Societies." The Annals of the American Academy of Political and Social Science, Vol. 415. Cowgill, Donald O., and Lowell D. Holmes. 1972. Aging and Modernization. New York: Appleton-Century-Crafts. Crimmins, Eileen M., and Dominique Ingegneri. 1990. "Interaction and Living Arrangements of Older Parents and Their Children." Research on Aging 12(10):3-35. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 136 The Well-Being of the Elderly in Asia Crimmins, Eileen M., Yasuhiko Saito and Sandra L. Reynolds. 1997. "Further Evidence on Recent Trends in the Prevalence and Incidence of Disability among Older Americans from Two Sources: The LSOA and the NHIS."7owr- nal of Gerontology: Social Sciences 52B(2):59-71. Deaton, Angus, and Christina H. Paxson. 1990. "Patterns of Aging in Thailand and Cote d'lvoire." NBER Working Papers Series No. 3436. Cambridge, MA: National Bureau of Economic Research. . 2000a. "Growth and Saving among Individuals and Households." The Review of Economics and Statistics 82(2):212-225. . 2000b. "Growth, Demographic Structure and National Saving in Tai wan." In C. Y. Cyrus Chu and Ronald D. Lee, eds. Population and Economic Change in East Asia, pp. 141-173. New York: Population Council. Easterlin, Richard A. 1991. "The Economic Impact of Prospective Population Changes in Advanced Industrial Countries: An Historical Perspective." Jour nal of Gerontology: Social Sciences 46(6):S299-S309. Easterlin, Richard A. Christine M. Schaeffer, and Diane J. Macunovich. 1993. "Will the Baby Boomers Be Less Well Off Than Their Parents? Income, Wealth, and Family Circumstances Over the Life Cycle in the United States." Population and Development Review 19(3):497-522. Eggebeen, David J., and Dennis P. Hogan. 1990. "Giving between Generations in American Families." Human Nature 1(3):211-232. Elder, Glen H., Jr. 1987. "Family and Lives: Some Development in Life Course Studies." Journal of Family History 12(1-3): 179-199. Finch, Caleb E., James W. Vaupel, and Kevin Kinsella, eds. 2001. Cells and Sur veys. Washington, DC: National Academy Press. Foner, Anne, and David I. Kertzer. 1978. "Transitions over the Life Course: Les sons from Age-Set Societies." American Journal of Sociology 83:1081 -1104. Frankenberg, Elizabeth, Lee Lillard, and Robert J. Willis. N.D. "Money for Noth ing? Altruism, Exchange, and Old Age Security in Southeast Asia." Unpub lished Manuscript. Santa Monica, CA: RAND. Freedman, Ronald. 1993. "Book Reviews: Family Systems and Cultural Change." Population and Development Review 19 (1):621-623. Freedman, Vicki A., and Linda G. Martin. 20001. "Contribution of Chronic Condi tions to Aggregate Changes in Old-Age Functioning." American Journal of Public Health 90(11): 1755-1760. Fries, James F. 1980. "Aging, Natural Death and the Compression of Morbidity." New England Journal of Medicine 303:130-135. Fry, Christine L. 1996. "Age, Aging, and Culture." In Robert H. Binstock and Linda K. George, eds., Handbook of Aging and the Social Sciences, 4th edi tion, 117-136. San Diego: Academic Press. George, Linda K. 1996. "Social Factors and Illness." In Robert H. Binstock and Linda K. George, eds., Handbook of Aging and the Social Sciences, 4th edi tion, 229-252. San Diego: Academic Press. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Theoretical Perspectives, Measurement Issues 137 Goode, William J. 1963. World Revolution and Family Patterns. Glencoe: The Free Press. Gruber, Jonathan, and David A. Wise, eds. 1999. Social Security and Retirement Around the World. Chicago: University of Chicago Press. Hareven, Tamara K. 1977. "Family Time and Historical Time." Daedulus 106(Spring):57-70. . 1982. Family Time and Industrial Time: The Relationship between the Family and Work in a New England Industrial Community. Cambridge: Cam bridge University Press. Hashimoto, Akiko. 1996. The Gift of Generations: Japanese and American Per spectives on Aging and the Social Contract. Cambridge: Cambridge Univer sity Press. Hauser, Philip M., and Otis D. Duncan. 1959. "Overview and Conclusions." In Philip M. Hauser and O. D. Duncan, eds. The Study of Population: An Inven tory and Appraisal, 1-16. Chicago: University of Chicago Press. Hermalin, Albert I. 1993. Fertility and Family Planning among the Elderly in Tai wan, or Integrating the Demography of Aging into Population Studies." De mography 30(4):507-18. . 1997a. "The Current State of Aging Data in East and Southeast Asia." Paper prepared for the symposium: Work, Retirement, and Wealth: Current Data and Future Needs -An International Perspective, 1997 World Congress of Gerontology, Adelaide, Australia. 1997b. "Challenges to Comparative Research in Inlcrgencrational Transfers." Paper prepared for the symposium: Cross-National Aspects of Intervene rational Transfers, 1997 World Congress of Gerontology, Adelaide, Australia. Also appears as Comparative Study of the Elderly in Asia Research Report tio. 00-56, Population Studies Center, Univ. of Michigan, August 2000. Hermalin, Albert I., Mary Beth Ofstedal and Li Chi. 1992. "Kin Availability of the Elderly in Taiwan: Who is Available and Where Are They?" Comparative Study of the Elderly in Asia Research Report No. 92-18, Population Studies Center, University of Michigan. Johnson, D. Gale, and Ronald D. Lee, eds. 1987. Population Growth and Eco nomic Development: Issues and Evidence. Madison, WI: University of Wis consin Press. Johnston, Mark. 1999. "Population Ageing and Development: Implications for Asia and the Pacific." In United Nations, Promoting a Society for All Ages in Asia and the Pacific. New York: United Nations. Juster, F. Thomas, and Richard M. Suzman. 1995. "An Overview of the Health and Retirement Study." The Journal of Human Resources 30:S7-S56 Supple ment Kelley, Allen C. 1988. "Economic Consequences of Population Change in the Third World." Journal of Economic Literature 26(4): 1687-1728. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 138 The Well-Being of the Elderly in Asia Keyfitz, Nathan. 1993. "Thirty Years of Demography and Demography." Demog raphy 30 (4):533-550. Knodel, John. 1995. "Introduction" (to special issue on focus group research on living arrangement of Asian elderly). Journal of Cross-Cultural Gerontology 10:1-6. Knodel, John, and Napaporn Chayovan. 1997. "Family Support and Living Ar rangements of Thai Elderly." Asia-Pacific Population Journal 12(4):51-68. Knodel, John, and Chanpen Saengtienchai. 1999. "Studying Living Arrangements of the Elderly: Lessons from a Quasi-Qualitative Case Study Approach in Thailand." Journal ofCross-Cultural Gerontology 14(3): 197-220. Lee, Ronald D. 1994a. "The Formal Demography of Population Aging, Transfers, and the Economic Life Cycle." In Linda G. Martin and Samuel H. Preston, eds., Demography of Aging, 8-49. Washington, DC: National Academy Press. . 1994b. "Population Age Structure, Intergenerational Transfers and Wealth: A New Approach with Applications to the U.S." Journal of Human Resources 29(4): 1027-1063. . 2000. "Intergenerational Transfers and the Economic Life Cycle: A Cross-Cultural Perspective." In Andrew Mason and George Tapinos, eds. Sharing the Wealth: Demographic Change and Economic Transfers between Generations. Oxford: Oxford University Press. Lee, Ronald D., Andrew Mason, and Timothy Miller. 2000. "Live Cycle Saving and the Demographic Transition: The Case of Taiwan." In C. Y. Cyrus Chu and Ronald D. Lee, eds. Population and Economic Change in East Asia, pp. 194-223. New York: Population Council. Lee, Yean-Ju, William L. Parish, and Robert J. Willis. 1994. "Sons, Daughters and Intergenerational Support in Taiwan. American Journal of Sociology 94:1010- Lillard, Lee A., and Robert J. Willis. 1997. "Motives for Intergenerational Trans fers: Evidence from Malaysia." Demography 34(1): 115-134. McDonald, Peter. 1992. "Convergence or Compromise in Historical Family Change?" In Elza Berquo and Peter Xenos, eds., Family Systems and Cul tural Change, 15-30. Oxford: Clarendon Press. Manton, Kenneth G, Larry Corder, and Eric Stallard. 1997. "Chronic Disability Trends in Elderly United States Populations: 1982-1994." Proceedings of the National Academy of Sciences of the United States of America 94(6):2593- Manton, Kenneth G, and XiLiang Gu. 2001. "Changes in the Prevalence of Chronic Disability in the United States Black and Nonblack Population above Age 65 from 1982 to 1999." Proceedings of the National Academy of Sciences of the United States of America 98:(11), May 22, 2001. Manton, Kenneth G., and Beth Soldo. 1992. "Disability and Mortality among the Oldest Old: Implications for Current and Future Health and Long-Term-Care Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Theoretical Perspectives, Measurement Issues 139 Service Needs." In Richard M. Suzman, David P. Willis, and Kenneth G. Manton, eds., The Oldest Old. New York: Oxford University Press. Marmot, Michael, and Richard G. Wilkinson, eds. 1999. Social Determinants of Health. Oxford: Oxford University Press. Marshall, Victor W. 1996. "The State of Theory in Aging and the Social Sci ences." In Robert H. Binstock and Linda K. George, eds., Handbook of Ag ing and the Social Sciences, 4lh edition, 12-30. San Diego: Academic Press. . 1999. "Analyzing Social Theories of Aging." In Ver n L. Bengtson and K. W. Schaie, eds., Handbook of Theories of Aging, 434-455. New York: Springer Publishing. Martin, Linda G. 1989. "Living Arrangements of the Elderly in Fiji, Korea, Ma laysia, and the Philippines." Demography 26(4):627-643. Martin, Linda G, and Kevin Kinsella. 1994. "Research on the Demography of Aging in Developing Countries." In Linda G. Martin and Samuel H. Preston, eds. Demography of Aging, 356-404. Washington, DC: National Academy Press. Martin, Linda G., and Samuel H. Preston, eds. 1994. Demography of Aging. Wash ington, DC: National Academy Press. Mason, Andrew. 1987. "National Saving Rates and Population Growth: A New Model and New Evidence." In D. Gale Johnson and Ronald D. Lee, eds. Population Growth and Economic Development: Issues and Evidence, pp. 523-560. Madison, WI: University of Wisconsin Press. Mason, Andrew, Sang-Hyop Lee, and Gerard Russo. 2001. "Population Momen tum and Population Aging in Asia and Near-East Countries." East-West Cen ter Population Series No. 1076. Honolulu: East-West Center. Mason, Karen Oppenheim. 1992. "Family Change and Support of the Elderly in Asia: What Do We Know?" Asia-Pacific Population Journal 7(3): 13-32. Mason, Karen Oppenheim, Noriko O. Tsuya, and Minja Kim Choe, eds. 1998. The Changing Family in Comparative Perspective: Asia and the United States. Honolulu: East-West Center. Myers, George C. 1996. "Aging and the Social Sciences: Research Directions and Unresolved Issues." In Robert H. Binstock and Linda K. George, eds., Hand book of Aging and the Social Sciences, 1-11. San Diego, CA: Academic Press. National Research Council. 2001. "Preparing for an Aging World: The Case for Cross-National Research." Panel on a Research Agenda and New Data for an Aging World. Committee on Population and Committee on National Statis tics. Washington, DC: National Academy Press. Ngin, ChorSwang, and Julie DaVanzo. 1999. "Parent-Child Coresidence and Quasi- Coresidence in Peninsular Malaysia." Southeast Asian Journal of Social Sci ence 27(2):43-64. Olshansky, S. Jay, Bruce A. Carnes, and Christine Cassel. 1990. "In Search of Methuselah: Estimating the Upper Limits to Human Longevity." Science 250:634-640. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 140 The Well-Being of the Elderly in Asia Palmore, Erdman, and Kenneth Manton. 1974. "Modernization and the Status of the Aged: International Correlations." Journal of Gerontology 29:205-210. Preston, Samuel H. 1993. "The Contours of Demography: Estimates and Projec tions." Demography 30(4):593-606. Preston, Samuel H., and Paul Taubman. 1994. "Socioeconomic Differences in Adult Mortality and Health Status." In Linda G. Martin and Samuel H. Preston, eds., Demography of Aging, 279-318. Washington, DC: National Academy Press. Quadagno, Jill. 1982. Aging in Early Industrial Society. NY: Academic Press. . 1989. "Generational Equity and the Politics of the Welfare State." Politics and Society 17:353-376. Quinn, Joseph F., and Richard V. Burkhauser. 1994. "Retirement and Labor Force Behavior of the Elderly." In Linda G. Martin and Samuel H. Preston, eds., Demography of Aging, 50-101. Washington, DC: National Academy Press. Riley, Matilda W., Marilyn E. Johnson, and Anne Foner. \972.Aging andSociety: A Sociology of Age Stratification. New York: Russell Sage. Samuelson, Paul. 1958. "An Exact Consumption-Loan Model of Interest with or without the Social Contrivance of Money." Journal of Political Economy 66(6):467-482. Schneider, Edward L., and John W. Rowe, eds. 1996. Handbook of the Biology of Aging. San Diego, CA: Academic Press. Schoeni, Robert F. 1997. "Private Interhousehold Transfers of Money and Time: New Empirical Evidence." Review of Income and Wealth 43(4):423-448. Simmons, Leo W. 1945. The Role of the Aged in Primitive Society. London: Ox ford University Press. Smeeding, Timothy. 1990. "Economic Status of the Elderly." In Robert H. Binstock and Linda K. George, eds., Handbook of Aging and the Social Sciences, 3rd edition, 362-382. San Diego, CA: Academic Press. Smith, James P. 1995. "Racial and Ethnic Differences in Wealth in the Health and Retirement Study." Journal of Human Resources 30(Supplement):S 158-S183. Soldo, Beth J., and Martha Hill. 1993. "Intergenerational Transfers: Economic, Demographic, and Social Perspectives." Annual Review of Gerontology and Geriatrics 13:187-216. Stewart, Frank H. 1977. Fundamentals of the Group Systems. New York: Aca demic Press. Stycos, Joseph M. 1987. "Demography as an Interdiscipline." Sociological Fo- rwm2(4):616. Thornton, Arland. 1991. "Reading History Sideways: A Critical Evaluation of Western Family History and Theory." Unpublished book manuscript. Thornton, Arland, and Thomas E. Fricke. 1987. "Social Change and the Family: Comparative Perspectives from the West, China, and South Asia." Sociologi cal Forum 2(4):746-779. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Theoretical Perspectives, Measurement Issues 141 Thornton, Arland, and Hui-Sheng Lin. 1994. Social Change and the Family in Taiwan. Chicago: University of Chicago Press. Treas, Judith, and Barbara Logue. 1986. "Economic Development and the Older Population." Population and Development Review 12(4):645-73. Uhlenberg, Peter, and Sonia Miner. 1996. "Life Course and Aging: A Cohort Per spective." In Robert H. Binstock and Linda K. George, eds., Handbook of Aging and the Social Sciences, 208-228. San Diego: Academic Press. United Nations. 1983. World Assembly on Aging, 1982. Vienna International Plan of Action on Aging. New York: United Nations. United Nations. 2001. World Population Prospects, The 2000 Revision: High lights, ESA/P/WP. 165. New York: United Nations. Verbrugge, Lois M. 1984. "Longer Life but Worsening Health? Trends in Health and Mortality of Middle-Aged and Older Persons." Milbank Memorial Fund Quarterly/Health and Society 62:475-519. Wachter, Kenneth W., and Caleb E. Finch, eds. 1997. Between Zeus and the Salmon. Washington, DC: National Academy Press. Weinstein, Maxine, Albert Hermalin, and Michael Stoto, eds. 2001. Population, Health, and Aging: Strengthening the Dialogue between Epidemiology and Demography. Annals of the New York Academy of Sciences, Volume 954. New York: Academy of Sciences. Conference Proceedings of "Demography and Epidemiology: Frontiers in Population Health and Aging," Georgetown University, February 2001.. Weinstein, Maxine, and Robert J. Willis. 2001. "Stretching Social Surveys to In clude Bioindicators: Possibilities for the Health and Retirement Study, Expe rience from the Taiwan Study of the Elderly." In Caleb E. Finch, James W. Vaupel, and Kevin Kinsella, eds., Cells and Surveys, 350-275. Washington, DC: National Academy Press. Whyte, Martin K., Albert I. Hermalin, and Mary Beth Ofstedal. Forthcoming. "Intergenerational Relations in Two Chinese Societies." In Martin K. Whyte, ed. China's Revolutions and Intergenerational Relations. Ann Arbor, MI: Center for Chinese Studies, University of Michigan. Whyte, Martin K., and William L. Parish. 1984. Urban Life in Contemporary China. Chicago: University of Chicago Press. Wilmoth, John R. 1997. "In Search of Limits." In Kenneth W. Wachter and Caleb E. Finch, eds., Between Zeus and the Salmon, 38-64. Washington, DC: Na tional Academy Press. Wilson, Chris, and Tim Dyson. 1992. "Family Systems and Cultural Change: Per spectives from Past and Present." In Elza Berquo and Peter Xenos, eds.. Family Systems and Cultural Change, 31-45. Oxford: Clarendon Press. World Bank. 1994. Averting the Old Age Crisis: Policies to Protect the Old and Promote Growth. A World Bank Policy Research Report. New York: Oxford. Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Generated for anonymous (University of Michigan) on 2014-05-27 12:29 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Chapter 5 PATTERNS AND DETERMINANTS OF LIVING ARRANGEMENTS John Knodel and Mary Beth Ofstedal Living arrangements, and in particular coresidence with an adult child, lie at the heart of the traditional system of familial support for elders in all four of the countries encompassed by this study. For many elders, support and care from adult children living within the same household or residential compound is crucial for their psychological, physical, and economic well- being. Thus it is not surprising that much of the recent sociodemographic research on Asian elderly, including that featured in theme issues of several journals, has focused on this topic (DaVanzo and Chan, 1994; Knodel, 1992, 1995, 1997; Logan, Bian and Bian, 1998; Martin, 1989). Previous studies, however, were often hampered by the lack of data with sufficient detail to analyze important aspects of living arrangements and their place within the broader network of intergenerational exchanges. The recent rounds of surveys undertaken in conjunction with our collaborative project in the Philippines, Taiwan, and Thailand, and to a lesser extent in Singapore, go well beyond most previous data collection efforts in providing detailed information, permitting a fuller description of living arrangements of the older population and a more comprehensive examination of their determinants and consequences. In all of these countries there is a long tradition of extended family households, and expectations concerning coresidence in old age remain high Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 144 The Well-Being of the Elderly in Asia among current cohorts of elderly and their adult children. Views of both elders and their adult children, as expressed in focus groups in all four countries, confirm living arrangements as a crucial part of family support (Asis et al., 1995). Nevertheless, their significance for the well-being of elders should not be exaggerated. As Figure 4.1 in the previous chapter reminds us, although living arrangements may be a means to achieve various dimensions of well- being for both generations, they are not an end in themselves. Focus group data from all four countries underline the costs and benefits of coresidence, from both the elderly parents' and the adult children's perspectives (Asis et al., 1995). For instance, while both generations report that lack of privacy is a disadvantage of coresidence, older family members who require personal assistance and younger members who need help with child-care or finances benefit greatly from coresidence. These costs and benefits are complex in that they involve many dimensions, are embedded in normative expectations, and tend to shift over the life course of both generations and perhaps over historical time. Moreover, intergenerational exchanges of material support and services between households, as well as social visits, are all aspects of the broader interconnected familial support system that influence well-being. These latter types of exchanges will be taken up in Chapter 6. Government planners in all the countries under study are concerned about the potential erosion of family support systems, including coresidence, in response to the socioeconomic changes engulfing Asia and most of the developing world. However, the extent to which living arrangements have changed and will continue to change, the nature of that change, and its impli cations for the well-being of the elderly and their families are all issues that need to be addressed with empirical data. The present chapter casts a broad net on the topic of living arrangements, covering such aspects as trends in living arrangements over time, generational composition of households, and differentials related to the marital status and gender of children. The chapter concludes with a multivariate examination of the determinants of living ar rangements. Measurement Issues Research on living arrangements looks at both household composition and the relative location of kin not residing in the household. One particularly key aspect concerns coresidence with children. For most purposes, the interest in coresidence with children stems from its greater potential for meaningful and frequent exchanges of services between elderly parents and their children. Generally coresidence is defined as living in the same dwelling or household Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Patterns and Determinants of Living Arrangements 145 with the focal elderly person. For some purposes, however, this definition can be considered as either too narrow or too broad. For example, if our goal is to capture all cases where the daily lives of elderly parents and their children are intimately intertwined as a result of living arrangements, the typical concept of coresidence captures only part of this and thus can be considered too nar row. For this reason, we introduce the concept of quasi-coresidence in some analyses to allow for cases in which the focal elderly person either lives adja cent to a child or near enough to have daily contact. In addition, when the research interest in coresidence centers on the ability of living arrangements to facilitate care of elderly, a definition of coresidence that does not incorporate the ages of coresident children may be too broad. In this context, the coresidence of adult children is likely to have quite different implications than the coresidence of dependent minor chil dren. For this reason, we define all unmarried children at least 18 years of age and all married children, regardless of age, as adults. We assume that by age 18 most children are capable of making significant economic contributions to the household and providing various types of physical and emotional assis tance to their parents. No doubt some coresident children we define as adults remain dependent on support from their parents. This is especially likely to be the case for those who are still in school or not working, a situation that varies among the four countries being studied. For example, among coresident chil dren aged 18 to 21 in Thailand, 18 percent were still studying and 28 percent had not worked in the previous year. In contrast, in Taiwan 53 percent were still studying and 62 percent were not currently working. Nevertheless, for convenience, we have imposed a uniform definition of an adult child in the tables that follow. One unusual aspect of the Taiwanese data set in relation to determin ing coresidence should be noted. In the Taiwan survey, children who were attending school away from home, serving in the military, or working tempo rarily away from home were counted as members of their parent's household although they were not physically present. Because most such children are unmarried, this classification has virtually no effect on calculations of the percent of elderly who live with a married child, but it does slightly increase the calculated percent who live with an unmarried child. However, even in the case of unmarried adult sons, for which the impact is largest, the propor tion coresiding increases only by one percentage point, from 20 to 21 percent. Tabulations in this and other chapters refer to elderly living in private households and thus exclude the small minority who live in institutional set tings. The reason for this is that, with the exception of the Taiwan survey, the sampling universes were defined in this manner. For the sake of comparabil ity, we exclude the 0.9 percent of Taiwanese respondents age 60 years and Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 146 The Well-Being of the Elderly in Asia over who were residing in institutions. We note that recent census figures indicate that in the other countries, the fraction of elderly living in institu tional settings is also very small (1.8 percent in Thailand, the vast majority of whom were monks living in Buddhist temples; 1.5 percent in Singapore, and 0.2 percent in the Philippines). Despite their small numbers, however, elderly in institutions obviously differ in their living arrangements from those in pri vate households, and this exclusion should be kept in mind when interpreting results. Living arrangements at any particular point depend not only on the costs and benefits of specific types but also on the availability of kin as poten tial household members. Thus another issue related to the measurement of living arrangements is whether or not tabulations should be conditioned on the availability of children or other relatives. As Table 2.9 indicates, well over 90 percent of elders in each of the four countries have at least one living adult child, a much smaller majority have a living spouse (ranging from one-half in Singapore to two-thirds in Taiwan), and only a very small minority have a living parent with whom they could live. In the analyses in the present chap ter, we typically examine coresidence with adult children only among respon dents who in fact have at least one living adult child. Other analyses are con ditioned on the availability of more specific types of children—for instance, sons, daughters, or married children. When analyses are conditioned on avail ability of particular types of children or kin, this is noted both in the text and table headings. Finally, two important qualifications need to be kept in mind when interpreting our findings. First, because living arrangements are dynamic, changing over the life course of the elderly, they are only imperfectly cap tured in cross-sectional surveys. Second, our measurements are directed at the forms of living arrangements rather than their functions. As qualitative research using both focus groups and quasi-case studies has revealed, similar living arrangements can have quite different meanings for the well-being of the elderly (e.g., Knodel, 1995; Knodel and Saengtienchai, 1999). For ex ample, the same household composition can be associated with very different support and care arrangements across families, as can wider configurations that take account of children and other kin outside the immediate household. Moreover, changes in forms and functions may only be loosely linked. Thus to judge the implications of trends in living arrangements for the well-being of elders requires evidence about possible changes in functions as well. Like wise, even if the form of living arrangements remains relatively unchanged, it should not be presumed that the nature of the relations among household members have also remained static (Hermalin, 2000). Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Patterns and Determinants of Living Arrangements 147 Trends in Coresidence Before examining results in detail from the most recent round of surveys, it is useful to compare basic indicators of living arrangements with those from previous surveys to determine current trends. In all four countries, results from earlier surveys permit some over-time comparisons. As Table 5.1 shows, despite the very substantial social and economic changes over the last several decades described in Chapter 2, coresidence remains the predominant living arrangement for older persons in each country, and only modest changes are evident. As a result, the percentage of elders who live alone or who live with a spouse remains small. The clearest signs of declining coresidence with children are found for Taiwan, although the particularly sharp reduction between 1985 and 1989 might partially be an artifact of differences in the surveys that are the sources of the trend data. Taken at face value, however, it appears that coresidence with a child declined by almost 15 percentage points during the two-decade period for which data are presented. Reductions in coresidence appear to be more modest or even absent in the other countries. In the case of the Philippines, problems of comparability between the sources raise some uncertainty as to the trend. The only prior survey specifically directed at the older population was based on a subnational sample in 1986 (as part of the ASEAN Population Program). If the results of this earlier survey are compared with the 1996 survey, which also targets the older population, there appears to be a modest decline in the percent living with a child. However, the earlier ASEAN survey is probably unrepresenta tive of the national situation. This is suggested by the fact that analysis of households with persons aged 60 and over from the 1988 National Demo graphic Survey, which covered a nationally representative general sample of households, indicated a somewhat lower level of coresidence and one that is similar to that found in the new 1996 survey. Moreover, analysis of the 1993 round of the National Demographic Survey indicates virtually the same level of coresidence with children as the 1988 survey. Thus all surveys except the sub-national ASEAN survey are consistent and point to no change in this aspect of living arrangements. In Singapore, the 1995 survey indicates a slightly lower percent of the elderly living with a child than was true in 1986. Never theless, the level remains extremely high. Thailand shows a modest decline from 77 to 71 percent of elderly coresiding with a child over a nine-year interval, but the level even in 1995 is still very substantial. Although not included in Table 5.1, it is possible for Thailand and Taiwan to determine the percentage of elderly who see a child daily (including Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 148 The Well-Being of the Elderly in Asia Table 5.1. Trends in Household Composition among Those Age 60 or Over Country and YearPhilippinesNat. DemographicSurvey19881993Elderly surveys1984 (subnational)1996ThailandElderly surveys198619941995TaiwanSurvey of Income andExpenditures197619801985Elderly surveys19891996SingaporeElderly surveys19881995Household Composition (Percent Distribution)Lives Alone43.325.54.33.64.3, .tLives w/ Spouse Only109.137.96.711.611.9* N8.812.817.3^i • •>10.29.623.3Y—12.616.135.9Lives with Children6869.37470.476.272.870.983.781.678.370.769.08885.0Other Composi tion1818.32116.112.812.012.97.55.64.46.65.275.8Total Percent100100100100100100100100100100100100100100 Figures for Taiwan refer to population age 65+; figures for other countries are for population age 60+. those who are coresident) for at least two time points. This figure captures situations in which the elderly respondent may not live with a child but lives close enough, often next door, to permit daily interaction. In Taiwan, the percent who see a child daily declined from 82 to 79 percent between 1989 and 1996, which is similar in magnitude to the decline in coresidence. However, in Thailand this percentage declined only one percentage point, from 91 to 90 Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Patterns and Determinants of Living Arrangements 149 percent, between the 1986 and 1995 surveys (Knodel and Chayovan, 1997). Thus the apparent decline in coresidence does not necessarily reflect a trend toward disengagement between elders and their adult children. It may instead reflect only a limited modification of living arrangements. Possibly rising standards of living during the intervening period allowed married adult children to occupy separate dwelling units next to or near their elderly parents. Such arrangements might provide greater privacy for both parties while retaining close interaction consistent with a family system of support and care of elderly. Focus group discussions in Thailand have revealed that some elderly prefer this type of arrangement, especially if they are in good health (Knodel, Saengtienchai, and Sittitrai, 1995). Survey data in the Philippines has also revealed a preference for this arrangement, particularly among married elderly (Natividad and Cruz, 1997). Evidence from other parts of East Asia shows more pronounced de clines in coresidence. For example, in Japan, the percentage of persons age 65 years and over living with any child fell from 77 to 52 percent between 1970 and 1997 (Hermalin, 1995; Japan Aging Research Center, 2000, Table 25). Even more dramatic is the decline that occurred in the Republic of Ko rea, where the percent of older adults coresiding with a child fell about the same magnitude but over a much shorter period (from 78 percent in 1984 to 54 percent in 1994) (Kim, 1999). The pace of these declines is quite remark able when compared to the experience in the U.S. during the late 19th and early 20th centuries. Between 1880 and 1940 the percent of elderly (age 65 years or over) living with children declined from 64 percent to 49 percent (Ruggles, 1994). This decline is somewhat smaller in magnitude than that experienced in Japan and Korea, and it occurred over a much longer period and under very different social and economic circumstances. Interestingly, the decline in the U.S. continued at an accelerated pace after 1940, with the percent coresiding dropping to 30 percent in 1960 and further to 18 percent in 1980 (Ruggles, 1994). It is unclear at this point whether Asia will follow the U.S. pattern or whether declines in coresidence in Asia will level of! or even reverse. General Patterns of Household Composition Most studies of living arrangements of the older population focus on house hold composition. Typically distinctions are made among household mem bers that are likely to reflect their potential for being a source of either mate rial or non-material assistance or, conversely, a dependent of the elderly. Table 5.2 indicates the prevalence of several types of members that are of interest Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 150 The Well-Being of the Elderly in Asia from this perspective. The figures presented are not mutually exclusive; that is, respondents who coreside with one particular type of household member may also live with the other types of members shown. Given that a spouse can play an essential role in providing care for a frail or ill older person, or conversely can require such care, the presence or absence of a spouse is likely to have important implications for the well-being of the referent individual. The proportion of elderly who live with a spouse ranges from slightly under half in Singapore to almost two-thirds in Taiwan. It is important to note that the marital status distribution for the 1995 Singapore sample is rather different than for the population 60 and over enumerated in the 1990 census (Shantakumar, 1994). Thus the low proportion married among Singaporean elderly may reflect a sample bias due to the low response rate. Besides spouses, the presence or absence of children is likely to have important implications. As shown in Table 5.2, the proportion of elderly coresiding with a child is unusually high in Singapore (85 percent) compared to the other countries (69 to 71 percent). In part, the high proportion coresiding with a child in Singapore may be related to the fact that Singapore is a city- state with no rural sector while the others are countries with small town and rural populations. When compared only to the levels of coresidence in the capital cities of the other three countries, the level in Singapore no longer appears unusual. Among elderly with at least one living child, the percent who coreside with a child reaches 92 percent in Manila, 79 percent in Taipei, and 86 percent in Bangkok, compared to 89 percent in Singapore (results not shown). However, the factors contributing to high metropolitan levels of coresidence may be different in Singapore than in the other three countries. One possible contributor to the high coresidence levels in the capital cities of the other countries is the migration of elderly parents from rural areas to join adult children who migrated earlier. Since Singapore is a city-state with no rural sector, this factor could not contribute to high coresidence there, although rural areas in neighboring Malaysia might play a role comparable to that of the rural sectors of the other three countries. Also we do not have direct evidence that rural elderly migrate to join children in the cities to any significant extent. Focus group discussions among rural elderly in the Philip pines and Thailand revealed a strong attachment on the part of elders to their own home and probably a strong preference to "age in place" rather than leave to join children who migrated to cities (Domingo and Asis, 1995; Knodel, Saengtienchai and Sittitrai, 1995). Another factor that likely contributes to high coresidence levels in urban areas in all of the countries is the high cost of housing in such areas. In rural areas, land is more plentiful and building mate rials and labor are less costly, making it affordable for families to build and maintain separate dwellings. However, in urban areas, space is scarce and the Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Patterns and Determinants of Living Arrangements 151 Table 52. Percent of Elderly Living with Specific Types of Household Members Percent living with:SpouseAny childAdult childMarried childMinor child (< 18)GrandchildSiblingParent/parent-in-lawNon-relativePhilippines52.070.467.450.19.258.32.50.94.3Thailand60.470.969.350.75.255.03.32.21.2Taiwan64.069.067.741.92.050.30.71.70.8Singapore48.285.0--40.8--36.41.72.06.5 Notes: Results refer to elderly in private households. Children include own, adopted and stepchildren. -- = not available. cost of buying or renting a home can be exorbitant. In Singapore the govern ment offers housing subsidies to young adults who coreside with their parents and, particularly in light of Singapore's high cost of living, these subsidies are likely to further contribute to the high rates of coresidence there. Finally, it is possible that Singapore's high rate of coresidence is partly an artifact of the sample. As noted earlier, the marital status distribution for the 1995 Singapore survey sample is somewhat different than that enumerated in the 1990 census for the population 60 and over, with the sample containing a higher propor tion of unmarried individuals. Since unmarried elders are more likely to coreside with children, the rate of coresidence observed in the survey may be somewhat inflated. Since most children of elderly are grown, the percentage living with an adult child differs little from the percentage living with any child. The per centage living with a married child, however, is substantially lower and varies among the four countries. The lower proportions in Singapore and Taiwan probably reflect the later ages of marriage that prevailed there in recent de cades compared to the Philippines and Thailand (see Table 2.5). Some elderly still have children who are minors and live with them. Elderly in the Philip pines are most likely to have their minor-aged children living with them. This undoubtedly reflects the higher fertility in the Philippines over the last few decades compared to the other three countries. Far more common than living with their minor-aged children is for elderly to live with grandchildren (not Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 152 The Well-Being of the Elderly in Asia all of whom are minors). In most cases, the grandchildren are the offspring of coresident adult children. Very few elderly live with either their parents or siblings. In the case of parents, a major influence is 'low availability' in that only a small minority still have a living parent. However, even among those respondents who have a living parent, most do not live with the parent. This is not surprising in that respondents are 'in competition' with their siblings for their parents' pres ence. As discussed below, in all four countries, a stem family norm prevails such that an elderly parent eventually lives with only one adult child. The low presence of parents and siblings in the respondent's household reflects this pattern. Coresidence with non-relatives is also rare. In Singapore where it is most common, the non-relatives are mostly domestic helpers in the service of the household. Indeed, the percent of elderly who coreside with a non-relative in Manila and Bangkok (but not Taipei) is similar to that in Singapore, reflect ing the association between the presence of domestic help and living in large cities in some settings. Table 5.3 presents a simple typology of living arrangements based on household membership. The categories shown are mutually exclusive and exhaustive. They are based on combining information about the presence or absence of a spouse, a child or child-in-law, and anyone else. Prominence is given to spouses and children (together with their spouses) since they are likely to have the most general relevance for support and care of the referent elderly household member. In total, eight combinations are possible (includ ing living alone as one category). In each of the four countries, very few elderly live alone and only modest proportions live with only a spouse. Most households include at least one of the respondent's children. Several clear differences are apparent among the countries, however. For example, Singapore is unusual in that a high proportion of elderly live with children/children-in- law only or in combination with a spouse (20 percent and 26 percent, respec tively), but without grandchildren or others present. In addition, in Taiwan one-fourth of all elderly either live alone or reside only with a spouse, a pro portion that is far higher than in any of the other countries. Table 5.4 provides a more detailed examination of the generational depth of households by showing the percent of respondents residing in one, two, and three or higher generation households, with further distinctions of composition for households with two or more generations. The underlined percents in Table 5.4 represent the total percent residing in a two or three-plus generation household, and the figures listed underneath represent the percent in each subgroup. Here again, several differences are apparent across the four countries. Consistent with patterns in previous tables, one-generation house holds are most common in Taiwan (27 percent) and least common in Singapore Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Patterns and Determinants of Living Arrangements 153 Table 5.3. Percent Distribution According to Household Composition Household CompositionAloneSpouse onlyChildren/children-in-law onlyOthers onlySpouse & children/children-in-lawSpouse & othersChildren/children-in-law &othersSpouse, children/children-in-law& othersTotalPhilippines5.57.97.48.915.45.926.122.8100.0Thailand4.311.97.96.615.15.521.627.9100.0Taiwan9.716.15.72.016.02.218.729.6100.0Singapore3.35.920.04.726.31.023.715.1100.0 Notes: Results refer to elderly in private households. Children include own, adopted and stepchildren. Children-in-law are grouped with children for this tabulation. The category "others" includes grandchildren. (13 percent). Two-generation households are considerably more common and three-generation households less common in Singapore compared to the other countries. The proportion of households with three generations is very similar in the Philippines, Taiwan, and Thailand, constituting slightly less than half of all households with elderly members. The extent to which elderly live with a grandchild but no child, a configuration sometimes referred to as a "skip generation" household, is specifically indicated under the two-generation households. This situation typically arises when adult children migrate to take jobs in urban environments and either send their own children back or leave them behind to be cared for by the grandparents. While such households are very rare in Singapore and Taiwan, they represent a discernible if still small minority of cases in the Philippines and Thailand. Table 5.4 also indicates the mean size of households of different gen erational depths. Overall the average size of the households in which elders live in these four countries is between four and five members and thus rather moderate. Of course, mean household size increases with generational depth. Thus the substantial minority of elders who live in three generational house holds are together with far more members than those in one or two genera tional households. Three-generation households in the Philippines are the larg est on average, probably reflecting the persistence of higher fertility among adult coresident children than in the other countries. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 154 The Well-Being of the Elderly in Asia Table 5.4. Percent Distribution by Number of Generations and Generational Composition Number of Generations Generational CompositionOne generationTwo generations- w/ parent- w/ child- w/ grandchildThree or more generations- 3 gen. w/ child & grandchild- 3 gen. w/ parent & child- other 3 gen. & 4+ gen.Mean household size- one generation- two generations- three generationsPhilippines17.236.80.224.512.146.045.40.10.55.092.184.057.01Thailand19.632.80.424.18.247.645.90.90.94.472.113.705.98Taiwan26.725,00.422.12.548.247.00.60.64M1.673.576.78Singapore13.35030.648.91.136.435.01.00.4£312.023.385.88 Note: For Singapore, grandchildren and great-grandchildren can not be distinguished and thus are combined. Coresidence in Relation to Children's Gender and Marital Status The importance of children's gender, sometimes in connection with sibling order, has received considerable attention in the literature on support of elderly parents, particularly in connection with coresidence. As Karen Mason (1992) stressed in a review article about family change and support of the elderly in Asia, two major types of family systems prevail. Patrilineal systems are found in East Asia and the northern tier of South Asia and stress the responsibility of sons for caring and supporting parents. More flexible bilateral systems are found in Southeast Asia and the southern tier of South Asia in which daughters play an equally or more important role than sons. The countries included in the present study, in addition to the ethnic groups within Singapore, fall on both sides of this divide. With the exception of Singapore, the new surveys provide detailed information on characteristics of both coresident and non- coresident children. It is thus possible to examine in some depth the extent to which coresidence is conditioned on gender (as well as other characteristics of adult children). For Singapore, the more limited information available still permits equivalent tabulations to be approximated. As the following analyses Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Patterns and Determinants of Living Arrangements 155 Table 5.5. Coresidence with Sons and Daughters Conditioned on Availability Among elderly who havechild of specified type,% living with any:SonDaughterSingle adult sonSingle adult daughterMarried sonMarried daughterRatio of living with:Single adult son/singleadult daughterMarried son/marrieddaughterPhilip pines51.647.270.260.824.328.61.15.85Thailand40.855.258.263.323.740.4.92.59Taiwan68.320.569.270.157.45.9.999.73SingaporeChinese77.146.646.534.734.68.21.344.22Malay | Indian67.5 66.963.2 48.041.7 36.435.4 32.727.2 28.027.8 12.51.18 1.11.98 2.24 Notes: Single adult children are age 18 and over. For Singapore, results referring to single and married children are based on the 95% of cases for which a determination could be made (see text); in addition they are conditional only on having a child of the particular sex and not on adult status or marital status. reveal, very striking differences are observed in terms of the gender of the children with whom the elderly coreside. These differences, however, are also clearly conditioned on the marital status of the child. Table 5.5 shows the percentage of elderly living with children of dif ferent genders and marital statuses. For all countries except Singapore the results are conditioned on availability of having a child of the respective com bination of these characteristics. For Singapore, because of data limitations, they are only conditioned on having a child of the particular sex (regardless of age and marital status).1 In addition, it was not possible to identify the spe cific gender and marital status combination of coresident children in Singapore for 5 percent of the cases, so these respondents are excluded from the analy sis. Because there are pronounced differences among the three main ethnic groups in Singapore with respect to gender preferences related to coresidence with married children (Mehta, Osman and Lee, 1995), results are shown sepa rately for these three groups rather than for the country as a whole. Elderly parents in general are considerably more likely to live with a son than a daughter in Taiwan as well as in Singapore among the Chinese and Indians ethnic groups. In contrast, Filipino and Singaporean Malay elderly parents are only slightly more likely to live with a son than a daughter, while Thais are more likely to live with daughters than sons. The marital status of Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 156 The Well-Being of the Elderly in Asia the child, however, strongly affects the gender patterning of coresidence.2 The pronounced tendency to live with sons rather than daughters is largely absent with respect to single adult children in Taiwan and very much moder ated among Chinese and Indians in Singapore. Likewise the strong tendency to live with daughters rather than sons in Thailand is almost absent with re spect to single adult children. A very clear pattern of living with married sons rather than married daughters is evident for Taiwan and Singaporean Chinese and Indians. In contrast, Singaporean Malays show an almost equal tendency to coreside with a married son or daughter. Filipino elderly are somewhat more likely to coreside with a married daughter than son and Thai elderly show a strong tendency to live with a married daughter rather than a married son. In order to highlight the extent of any tendency to live with children of one gender over the other, the bottom panel of Table 5.5 presents the ratio of the percentage of elderly who live with sons to the percentage who live with daughters (conditioned on availability). A ratio of unity indicates equal ten dencies of elderly to live with sons as with daughters; ratios above unity re flect greater tendencies to live with sons, while those below unity show greater tendencies to live with daughters. With respect to coresidence with single children, the ratios are close to unity in all countries, indicating the absence of preference for living with children of either sex prior to their marriage. In contrast, pronounced differences in the ratios referring to coresidence with married children are apparent among the countries and ethnic groups in Singapore. The strongest tendency towards patrilocality is evident for Taiwan where elderly are almost ten times more likely to live with a married son than a married daughter. Only six percent of Taiwanese elderly who have a married daughter live with one. Likewise strong son preference is evident in Singapore for the Chinese, for whom the ratio is four times unity, and for Indian elderly, who are more than twice as likely to live with a married son than a married daughter. The reverse situation characterizes Thailand, where elderly are substantially less likely to live with a married son than a married daughter. Both in the Philippines and among Malays in Singapore, elderly are also more likely to live with married daughters than with married sons, although the imbalance is much more muted than in Thailand.daughter. Only 6 percent of Taiwanese elderly who have a married daughter live with one. Likewise strong son preference is evident in Singapore for the Chinese, for whom the ratio is four times unity, and for Indian elderly, who are more than twice as likely to live with a married son than a married daughter. The reverse is true in Thailand, where elderly are substantially less likely to live with a married son than a married daughter. Both in the Philippines and among Malays in Singapore, elderly are also Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Patterns and Determinants of Living Arrangements 157 Table 5.6. Percent Coresident among Elderly Whose Children Are All Married, by Sex Composition of Children Sex Composition of ChildrenAUUnstandardized Standardized1Those with onlyMarried sons Married daughtersPhilippines% | N53.0 550 57.2 55043.7 49 62.3 71Thailand% N65.6 2,116 63.8 2,11957.4 207 67.1 233Taiwan% N60.0 1,010 53.6 1,01056.6 157 32.4 133 "Standardized on the distribution of number of married children for all families with only married children, all of whom are the same sex. more likely to live with married daughters than with married sons, although the imbalance is less pronounced than in Thailand. Table 5.6 assesses the flexibility of elders in terms of the general preference to reside with a child of a particular gender. Since preferences with respect to the gender of a coresident child is largely limited to married children, as the previous table revealed, to facilitate interpretation of results, we limit our analysis to elderly whose children are all married. We then compare the percent coresident among elderly whose married children are all sons or are all daughters with the percent coresident among all elderly whose children are all married. The subgroup of elderly whose children are all of the same gender is of interest because they obviously have no choice with respect to the gender of a coresident child if they wish to coreside with a married child. If gender preferences are flexible, the lack of a child of the preferred gender should have little affect on the likelihood of coresidence. In contrast, if the gender preferences are relatively inflexible, elderly who do not have a child of the preferred gender will forego coresidence and thus exhibit lower levels of coresidence than if they had a child of the preferred gender. In making the comparison between the levels of coresidence among elderly whose children are all of one gender and elderly in general whose children are all married, it is useful to control for the number of children the elderly respondent has. Since the fewer children a respondent has, the greater the chance that they are all of the same gender, elderly whose children are all the same gender tend to have fewer children on average than elderly in general. Given the likely positive association between coresidence and having larger numbers of children, we present the standardized percent coresident among all elderly, using as the standard the distribution for number of children among elderly whose children are all of the same gender. Although Table 5.6 shows both the unstandardized and standardized percentages, the more appropriate comparison is between the standardized percents Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 158 The Well-Being of the Elderly in Asia coresident for all elderly and the percents coresident among those whose children are all of the same gender. Note that given the expected positive association between large numbers of children and coresidence, we would anticipate that the standardized percentages are lower than non-standardized ones. This is true for Taiwan and Thailand but not for the Philippines, which reflects that in the Philippines sample the relationship between number of children and likelihood of coresidence is somewhat irregular (Natividad and Cruz, 1997), which in turn probably reflects fluctuations due to small sample size. Indeed, elderly whose children are all married and of the same gender are a rather small subset of each survey (as indicated by the unweighted number of cases shown in Table 5.6) and thus some caution is necessary in interpreting the results. The results for Thailand indicate considerable flexibility in whatever gender preferences exist regarding coresidence. Elderly whose children are all married daughters are more likely to coreside than elderly whose children are all married sons, but even in the latter situation, the chance of coresidence is only slightly lower than the standardized percentage for all elderly parents (whose children are all married). In the Philippines there is some suggestion of a daughter preference in relation to coresidence with married children. Among elderly Filipinos who have only married sons, the percent coresiding is somewhat lower than the standardized percent (44 percent versus 57 percent). Likewise, there is also an apparent gender preference in Taiwan. In that country, which has a strong patrilocal tradition, having only married daughters substantially reduces the chance of coresidence compared to having all married sons. Nevertheless, in both countries there is some flexibility as evident from the fact that more than two-fifths of Filipino elderly who have only married sons and almost one-third of Taiwanese elderly who only have married daughters still live with one of them. It is not possible to conduct the same analysis for Singapore because the survey does not provide information on the number of a respondent's chil dren who have been married. However, a rough approximation of the degree of flexibility can be obtained by examining coresidence with a married child among respondents whose children are all of the same gender (regardless of whether they are married or single). The results (not shown) indicate that Chinese elderly are 1.4 times as likely to live with a married child when all the children are sons than when all the children are daughters (42.3 versus 29.4 percent, respectively). This suggests somewhat greater flexibility than among the Taiwanese elderly, but still indicates that having only daughters results in a substantial reduction in the likelihood of coresiding with a mar ried child for Singaporean Chinese elderly. Indian elderly are also substan tially more likely to live with a married child if all children are sons than if all are daughters, while for Malays the opposite is true. However, given the rela tively small numbers of cases on which the Indian and Malay results must be based, these findings can only be taken as suggestive.3 Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Patterns and Determinants of Living Arrangements 159 Table 5.7 indicates a hierarchical classification combining living ar rangements and contact with children in order to capture the extent to which elders who do not live in the same household with one of their children may still live in circumstances that permit close links in their daily lives. In both the Philippines and Thailand, more than half of elderly parents who do not coreside with a child are in a quasi-coresident situation as defined above, that is they either live next door to a child or see a child daily. Moreover, substan tial shares of elders who neither live in the same household or next door nor see a child daily, still have fairly frequent contact, either on a weekly or monthly basis. Only a tiny fraction of elderly parents do not see at least one child monthly. In Taiwan, although the combined percentage of elderly parents who do not live with or next to children or close enough to see one daily is higher than elsewhere, weekly visits are quite common. The embeddedness of Asian families is further demonstrated in Table 5.8, which shows the mean number of children living in the household and nearby, according to the child's gender and marital status and the older re spondents' place of residence (urban versus rural). The figures in the top panel simply show the mean number of living children in total (top row) and mean numbers of married and unmarried sons and daughters. Overall, the number of living children ranges from about 4.5 in Taiwan and Singapore to just under 6 children in the Philippines. Thailand falls in between with 5.3 living chil dren on average. A majority of the children in each country are married, but the number of unmarried children is not negligible. In the Philippines, the elderly have about 1.5 unmarried children on average, while the figures for Thailand and Taiwan are approximately 1.2 and 0.8, respectively. The second panel focuses on elderly residing in urban areas and the third panel on those in rural areas; both panels further break down the num bers of children according to whether they live in the household or nearby (children living further away are not included in the second or third panels). Perhaps the most striking finding from this table is the density of the family network in terms of their proximity. In all countries and in both urban and rural areas, the elderly have at least two children on average who either live in the household or nearby. This figure reaches as high as three children in both urban and rural Philippines, and over four children in Singapore. There is interesting variation in the balance of children living in the household and nearby across the countries and in rural versus urban areas. In urban Thailand, the average number of children living in the household is higher than the average number living nearby. In contrast, in urban areas in each of the other countries, the number of children living nearby exceeds that living in the household by some margin. This is particularly true for Singapore, where the number of children living nearby is more than twice the number Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 160 The Well-Being of the Elderly in Asia Table 5.7. Percent Distribution of Elderly Parents According to Location and Contact with Children, among Those with Any Living Children Living with a child in samehouseholdLiving adjacent to a childSees a child dailySees a child at least weeklySees a child at least monthlySees a child, but less thanmonthlyDoes not see a childTotal percentPhilippines73.46.48.34.92.64.40.0100.0Thailand74.29.96.43.23.32.70.4100.0Taiwan72.73.64.79.94.53.80.8100.0Singapore88.8-1.65.72.20.51.2100.0 Notes: Tabulations include both minor and adult children. The classification in this table is hierarchical in that inclusion in a higher category in the list removes the respondent eligibility for lower categories. - = not available. living in the household. In rural Thailand, the pattern is quite distinct from that in urban Thailand—the average number of children living nearby is larger than the average number coresident. This is also consistent with the pattern observed for rural Philippines and Taiwan. Another finding of interest concerns the marital status distribution of children in each location. In general, the number of unmarried children living in the household exceeds the number of coresident married children on aver age (except in rural Taiwan, where married sons account for the largest share of coresident children. In contrast, however, living nearby appears to be ex clusively an arrangement for married children. The average number of un married children living nearby is very low in both urban and rural areas in all countries. In the first part of the chapter we have focused on a descriptive over view of the living arrangements of the elderly and their structure with respect to the marital status and gender composition of coresident children. Before moving on to the analysis of living arrangement determinants, it is useful to summarize the findings thus far. First, we find that intergenerational coresidence is quite common, with the percent living with one or more child ranging from 69 percent in Taiwan to 85 percent in Singapore. Furthermore, close to half of elders in all the countries except Singapore live in three generational households. However, in none of the countries are elders likely to live with more than one married Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Table 5.8. Mean Number of Living Children by Sex and Marital Status of Children and Respondents Place of Residence Residence of Parent Location of Children Sex/Marital Status of ChildrenTotalMarried sonsUnmarried sonsMarried daughters Unmarried daughtersUrbanChildren living in household Married sonsUnmarried sonsMarried daughters Unmarried daughtersChildren living nearby Married sonsUnmarried sonsMarried daughters Unmarried daughtersRuralChildren living in household Married sonsUnmarried sonsMarried daughters Unmarried daughtersChildren living nearby Married sonsUnmarried sonsMarried daughters Unmarried daughtersCountryPhilippines5.912.150.792.16 0.761.40 0.260.470.24 0.411.57 0.830.060.63 0.051.30 0.160.550.19 0.392.02 1.010.060.88 0.07Thailand5.332.040.582.07 0.641.57 0.300.430.25 0.590.63 0.290.020.28 0.041.01 0.160.260.31 0.281.69 0.750.040.84 0.07Taiwan4.401.720.481.86 0.321.09 0.490.340.04 0.221.32 0.570.040.67 0.040.96 0.550.270.02 0.121.27 0.470.030.73 0.04Singapore4.592.34"2.26"1.410.87"0.54"3.04 1.41'1.63"inap. * Means are for all sons and all daughters, regardless of marital status, because the Singapore survey data does not contain marital status information for children. Notes: Means are based on the full elderly sample and thus include the small percentage childless in each country. Means may not sum exactly to the overall category mean because of missing data on children's sex, marital status and place of residence. Residing "nearby" is defined as: same compound, barangay or city (Philippines); next door or same village (Thailand); and next door, same neighborhood or same/neighboring city/town/village (Taiwan); same city/country (Singapore). Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 162 The Well-Being of the Elderly in Asia child at a time reflecting the prevailing stem family structure common to all. On the other hand, living with more than one unmarried child is more com mon for those who have the option of doing so. Relatives other than direct descents (children and grandchildren) and their spouses are only infrequently found in households with elderly. Second, the gender patterns of coresident children vary considerably. These differences are mainly limited, however, to married children. None of the surveys revealed pronounced gender preferences with respect to coresidence with unmarried adult children. In a very general sense, the findings for married children conform to characteristics of the two broad family systems prevailing in Asia referred to above (Mason, 1992). Thais, Filipinos and ethnic Malays are generally considered to have bilateral family systems, while Chinese and Indians (especially from Northern India) are generally considered to have pair ilinca I/patriarchal systems. Our results indicate Thai elderly and, to a lesser extent, Filipino elderly are more likely to coreside with a married daughter than with a married son while in Taiwan and among the Chinese and Indians in Singapore, there is a pronounced tendency to coreside with married sons over married daughters. Finally, most elderly parents who do not live with children either live nearby or have reasonably frequent contact with them. Thus there is little evidence of wide-spread desertion of parents by adult children in any of the four countries. Even the living arrangements of childless elders, with the ex ception of Taiwan, typically include residing with a spouse or relatives and thus appear to reflect a functioning family support system. Determinants of Living Arrangements The foregoing analyses have described the household composition of older Asians in detail, with particular focus on coresidence with children and the gender and marital status composition of coresident children. In this final section of the chapter, we investigate the determinants of living arrangements by examining how different types of living arrangements vary according to characteristics of older adults and their families. The conceptual model that guides this analysis posits that living ar rangements are influenced by three sets of factors: 1) the availability of cer tain types of kin with whom to coreside, 2) the preferences of the individuals involved with regard to shared versus independent living arrangements, and 3) the needs and resources of these various individuals. This model has been elaborated elsewhere both in our own work (Casterline et al., 1991; Hermalin, Ofstedal and Chang, 1991; Ofstedal, 1995) and that of others (Burch and Matthews, 1987; DaVanzo and Chan, 1994; Wolf and Soldo, 1988); here we provide a brief synopsis of the major premises. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Patterns and Determinants of Living Arrangements 163 In terms of availability, an older person's options for living arrange ments are to a large extent a function of having kin, particularly children. Those with larger numbers of children have more opportunities for coresiding with one or more of them, and indeed previous research in Asia (Casterline et al., 1991; Martin, 1989; Ofstedal, 1995) as well as the U.S. (Aquilino, 1990; Wolf and Soldo, 1988) has shown number of living children to be a strong determinant of coresidence with children. Beyond the mere existence of kin, it may also be important to take into account their attributes, such as their gender, age, marital status, and childrearing status. In addition to availability, living arrangements are likely to be influ enced by the preferences of those involved. These preferences may be shaped to a large degree by cultural norms and expectations, which as noted previ ously have traditionally promoted the value of intergenerational coresidence. However, preferences are also influenced by exposure to new ideas and val ues that occurs through the mass media, in classrooms and work places, and in social exchanges with family members and friends. Finally, to act on their preferences, families must have the physical and economic resources to do so. Hence, a last part of the model relates to the needs and resources of various family members in terms of economic, physi cal, and social support. Elderly parents who need daily physical care (e.g., for bathing, dressing, feeding, etc.) may require the presence of another family member in the same household. With regard to economic support, coresidence may benefit older parents as well as adult children who do not have the finan cial resources to maintain separate residences. As alluded to in the above discussion, an important feature of the con ceptual model is that it views coresidence decisions as involving multiple actors (Casterline et al., 1991; Ofstedal, 1995). Thus, it is important to take into account not only the needs, resources and preferences of the elderly indi vidual or couple, but also those of other family members to the extent pos sible. In keeping with this perspective, we expect different factors to be at play when considering coresidence with unmarried versus married children. To a large extent, this will reflect the varying availability of each type of child over the life course. When parents are young, still employed, and in good health, their children are also likely to be young and unmarried; however, as parents age, retire, and become increasingly frail, their children are likely to be predominantly married. In addition to availability, other factors may also come into play. Coresidence with unmarried children, at least during the early stages of old-age, may be in large part a response either to cultural or norma tive patterns that promote parent-child coresidence until the time a child mar ries, or to the needs of the unmarried children who lack the financial and/or social resources to establish independent living quarters. In contrast coresidence Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 164 The Well-Being of the Elderly in Asia with married children, which generally occurs later in old-age, may be more a response to the financial, health, and social needs of older parents. This is not to say that coresidence with married children is primarily beneficial to elderly parents as opposed to the younger generation; indeed, research presented in other chapters in this volume (see especially Chapter 6) and elsewhere (Hermalin, Roan and Perez, 1998) indicates that elderly parents are often very engaged in the care of grandchildren and household maintenance, and this is likely viewed as a major benefit to the younger couple, both members of whom may be working full-time. Whereas the previous analyses focused mainly on coresidence with children, this analysis also examines those who live independently from their families, either alone or with a spouse only. The analysis begins by examining bivariate associations between key life course, socioeconomic and family char acteristics that are indicative of needs/resources, preferences and availability on one hand, and alternate types of living arrangements on the other. These descriptive analyses are presented separately for each country. We then present results from a series of multivariate logistic regression analyses to examine the net effects of these characteristics on living arrangements controlling for the other factors. In the course of the multivariate analysis, we first estimated country- specific models for each outcome to examine the extent to which covariates differed across countries, and then a corresponding set of models in which the data were pooled across countries and relevant country interactions were tested. Tables presenting the pooled results are shown and discussed in the chapter. Those presenting the country-specific models are not shown here but are avail able from the chapter authors. Descriptive Results Table 5.9 presents the percent of respondents in each country who live alone (among those who are unmarried) and who live with a spouse only (among those who are married), by selected sociodemographic characteristics. We focus first on patterns of living alone among the unmarried, and then note similarities and differences in patterns of living with spouse only among the married. The overall percentage of unmarried elderly living alone (shown in the top row of figures) ranges from a low of 6 percent in Singapore to a high of 26 percent in Taiwan, with roughly one-tenth living alone in both the Philippines and Thailand. In all countries, unmarried men are more likely than unmarried women to live alone, although the gender difference is not statistically sig nificant for the Philippines and only marginally so for Thailand. In Taiwan the difference is particularly marked, with men being about three times more Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 3;5&>rtL._cVJe8ts_0o_^-^rt1£flj>,5o1GLv"f1rt_1tTcBC SAlom (am<11^be-i1-01s>.5C/5•sf-1F=(/11ii—I-s18to§^13ooc^_>i&wcoS^<,^§in3CO1utfli15^ ^•*ots!5^^r-8 o-— — CM N— ^ ON -^ O ' cO 00^^ oeCN r-;OOP •* — ? I -^ "> S"" ^ "" ^ vrt2 o tsP t'Oo —oqP d.^ ts vd o — • ON ^^ 'CN CNCNy-^ CSCSON .«•* —O^ '- ^ SO ts^- vo^^p in*j-^ $-i=i. mtst^ 'r:^^o ~ t?-^ mvo8 °°.o:^ ..it^ON t^ •••* e^*/ OO ON ^— • . ' ' e-H ^- CM ^-f- e-* »— * y^ '•.S ofi "•* vO O ^v r*- ^ Vi oo 55O t*^ ON O O Ov O ' ON ^ O— — r- ^" o "" oof^ O NOin c oo cJ ^3 r^* 2i o*Tt ri S O r*S S O :— •— • y- — — * J^ ._-Tt CNTr CO f-i — vo ON 25 ^^ ^ 25""e .^*-'C4 lOO -H^OO ri ON oo^N ^ ^fill**e" ' ^e^ ra N— '?3 vO i t5 eS S ^•S g- S R f 1 1 2 I*? Siilu <<o !S^6uH. s-.1: i g-Cts.itN t**t r-*25^_CN3*§•^ O O tN m oo» 1— * O e^^— e^ -^COsal"™ —wSTt m oo r- o*-e . . ^'NOv-vvO r^ o~e 00iiv^^^3c S-3 f^ ••e 3 sD K 0^vv^ oo rs C^l— ON ^ •—-•pi^^CN t*^ ON Pm oo r*ei v-— . vp2* I^t — S oCS CS CN oovd^ON — Tt 8d od — ^CN CN Tt Omm ts — ; —CN ^^ CN r^rri^ °°. °5 ts~ ootsts*ON f^ CTv C3CN Tt ON O.N — ooodoo ts oo o in — ro O CNS s .1 fIII2s^30 e — .tf3fIlls_oC*lts,ststsH-_tsr*vitsts'_vS11°I*_^N115 Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 166 The Well-Being of the Elderly in Asia likely than women to live alone. As will be noted throughout the volume, the gender pattern in Taiwan reflects the distinctive nature of older unmarried men in Taiwan, many of whom are Mainlanders who never married and, as a result, have limited if any family members with whom to live. In contrast, the vast majority of unmarried older women in Taiwan are widows with children. In the Philippines, those in the younger age group are less likely than those 70 years of age or over to live alone, whereas in Taiwan the reverse is true. There are no differences in the propensity to live alone by age in Thai land and Singapore. Among unmarried respondents, those who are either di vorced, separated or never-married (referred to as 'other unmarried') are con siderably more likely than those who are widowed to live alone. This pattern holds for all countries, but is especially pronounced in Taiwan and Singapore. Living alone is more common in rural than urban areas in all three countries containing rural areas. This is a finding that is observed consistently throughout the remainder of the chapter: rural elderly are more likely to live apart from their children and urban elderly are more likely to coreside with children. This pattern reflects in part out-migration of children from rural areas to seek employment, as well as selective migration of older parents to join their children in urban areas. In addition, the high cost of housing in large cities in all of the study countries places a serious constraint on the ability of urban families to maintain separate dwellings (Lopez, 1991). Finally, there are interesting differences across countries with respect to the association between education and living alone. In both the Philippines and Thailand the percent living alone declines with increasing education, whereas in Taiwan and Singapore the percent living alone increases dramati cally with increasing education. It may be that the greater economic resources associated with higher education allow older Taiwanese and Singaporeans to "buy" privacy, whereas in the Philippines and Thailand the more difficult economic circumstances in these countries require older parents to share their resources with other family members who are not as well-off. Furthermore, at least in Taiwan, those with higher education are disproportionately Mainland ers who have fewer children with whom to coreside and a stronger tendency to prefer independent living arrangements (Cornman, 1999). Turning now to patterns of living with a spouse only, the overall per centages range from 11 percent in Singapore to 24 percent in Taiwan. The propensity for living with a spouse only is higher than that for living alone in all countries except Taiwan. In contrast to living alone, which was more com mon for men than women, the percent living with a spouse only is higher among women than men in all countries, though not statistically significant for Singapore. Those in the oldest age group are more likely than their younger counterparts to live with a spouse only; for Taiwan and Singapore this is also Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Patterns and Determinants of Living Arrangements 167 a departure from the pattern observed for living alone. The urban/rural pat tern is consistent for both types of living arrangements. Rural elderly are more likely than those in urban areas to live independently from their children and other family members, regardless of whether this results in living alone or living with a spouse only. Finally, the pattern by education mirrors that for living alone in the Philippines, with the less educated being more likely to live with a spouse only. In Taiwan, those with the highest level of education have the highest propensity for living with a spouse only, and this finding is also similar to that for living alone. In Thailand and Singapore there are no significant differences by education. In Table 5.10 we return to the most common alternative to independent living and, indeed, the most prevalent type of living arrangement in each of the four study countries, coresidence with children. Here we present the per centage of all respondents (both married and unmarried combined) who coreside with various types of children for each country, according to the same sociodemographic characteristics as in the previous table. As noted pre viously, for several factors we expect their associations with intergenerational coresidence to differ depending on the marital status of the coresident child. Accordingly, we examine separately coresidence with any adult children (first column for each country), with unmarried adult children (middle column), and with married children (last column). It should be noted that respondents can be living with both types of children; in that case they will be represented in all three columns in this table. For the Philippines, Thailand and Taiwan, the percentages shown in the table are conditional on the availability of the child in question. As noted previously, the Singapore survey collected infor mation only on the number of living sons and daughters and not on their characteristics; as a result, the Singapore sample is conditioned only on avail ability of one or more children for each outcome (hence the sample size is the same for each column). As seen in the previous section, among older persons who have at least one living adult child, the percent coresiding with a child ranges from just under three-quarters in the Philippines, Thailand and Taiwan, to nearly nine- tenths in Singapore. Conditional on availability, the propensity to live with an unmarried child is higher than that for a married child in each country, with the differential being especially pronounced in the Philippines. Among those with one or more married children, slightly over half of older parents in Thai land and Taiwan live with a married child, and less than half of those in the Philippines and Singapore do so. The age patterns observed likely reflect life course influences of children and their parents. The propensity for living with any adult child is slightly higher for the younger than the older age group, with the exception of Thailand, Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 D. *2 •o •s 1/i J- 2- il BK1•o1Pei.o.Ei•IsIs73|B61.MIss w?,fs3is o1Is•g| .jr1 uS•0B>^1.,P•g•g ^1°s3B£harac-erin-icO|5 3?g|o •* iri o06 ro t*i o00 W-l xO *j.Si SSSg•» m oo 0sgm M I-; POOa ,,!f4 — ro ._ r- r- r- "1O|'* — q oooe3 £55w^,in n oo ^CN ri rJ -jr- r- r- ™ort(N 0 0 § JO Q * *j.*£g|ri no^""& tndj1i iu • f1 USHI iasg ^=iI iIn wert, r i * OO vC Mo *t ^ v ^ in VM0J .„,!O r*- ^ iri ri *o pm 2m 8^^5 --m£g i33! SSSain ts 2. Ov O r- ^>» *° 2i sm ""- ~*o —— 00~ 2oor-^ ^moo 1— «*i "^ ON CTv W ""^ - *" 5- gm js ONO O ^j, O O O Oo " *" g~ 82 ^ o5 g ^ 2 ^ ^ o(^ 12^ | ^ |r- « ^ r^ r- in ^ — m1 1I •"!1*S -slN<<& s S o 6&, ,1•ii 'iOO* **i ^v^ m »5??agts in oss^Sp "^ "e* vigt= p o?o*0§"* — eT^^SgO f* on^ iN Q —2^00(5 S °^ o^PSS|IHi5J5|P **! - ->|RSS^.X; C O ~SSSffiSgS'o — —« •» •«i0Ks P P i•iei_ §^e i*i O t^ r- r- 35 ^|*rv r-. vi ^~fi"S2^00A0i_mir ( r J — .e^=rJI5l»cig._S 2 S "mis?p^fiJiz £ $ ur-3r;^ni^m*-|e*{OOOv__ci*•5m213101 •s VI — o 4* n & § Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Patterns and Determinants of Living Arrangements 169 for which there is no difference by age. This pattern also generally holds for coresidence with unmarried children, however, the reverse pattern is found for coresidence with married children. In Thailand, Taiwan and Singapore, those in the older age group are more likely than their younger aged counterparts to live with a married child. Although we have conditioned the samples on availability of at least one child of the specified type, these different age patterns may be partly a function of availability in that older parents are likely to have more married children and fewer unmarried children, whereas younger parents may have more children who have not yet married. However, it may also be a function of the different life stages of members of both generations and of the needs and resources associated with these stages. Individuals age 60-69 have fewer needs of their own and greater resources to share than those age 70 or over—they are more likely to be in good health and to be generating income from employment. In addition, their unmarried children are younger and may not yet be established in careers of their own, and a shared living arrangement with their parents may be beneficial if not a necessary way to save on expenses. On the other hand, the higher propensity for coresidence with married children in the older age group may reflect the increased frailty of the older parents, their dependence on children for some level of financial support, and the tendency to turn to married over unmarried children to satisfy these needs. These considerations are also reflected in the patterns by marital status. With the exception of Thailand, those who are married (and also younger) are more likely than their unmarried counterparts to coreside with unmarried children, whereas those who are widowed are more likely to coreside with married children. The finding that divorced and separated individuals tend to have the lowest propensities for coresidence may have important implications for old-age living and support arrangements for future cohorts, for whom divorce and separation are more common experiences. With regard to urban/rural differentials, the patterns mirror those for independent living observed in Table 5.9. Those in urban areas are more likely than those in rural areas to coreside with any child, and this pattern generally holds regardless of the marital status of the child. One exception is that there is no urban/rural difference in the propensity to coreside with unmarried chil dren in the Philippines. Also, in Taiwan the percent coresiding with married children is only slightly higher in urban than rural areas and the difference is not statistically significant. Again, the findings for education reveal some expected and unexpected associations. As noted earlier, our hypothesis is that education is negatively associated with coresidence, particularly with married children. In this regard, the findings suggest strong and consistent support in three of the countries: Thailand, Taiwan and Singapore. With regard to coresidence with unmarried Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 170 The Well-Being of the Elderly in Asia children, the patterns are more varied. In Thailand we find the same negative association as for married children; however, in Taiwan there is no association between education and coresidence with unmarried children, and in Singapore the association is reversed: those with higher education are more likely to coreside with unmarried children. The Philippines shows a very different association between education and intergenerational coresidence. The percent coresiding with any child increases sharply with education, especially when comparing those with primary education to those with no formal schooling. The patterns for coresidence with unmarried and married children are more muted and not statistically significant, but in the same general direction. Multivariate Results Because a number of the characteristics examined in the preceding tables are highly correlated, it is difficult to attribute effects to individual factors based on the bivariate analyses alone. There are several other factors not examined in Tables 5.9-5.10 that we have hypothesized will influence living arrange ments as well. To address these limitations we turn to a set of multivariate logistic regression analyses in which we simultaneously estimate the effects of a number of factors on different types of living arrangements. The living arrangement outcomes examined here are all dichotomous and correspond with those just presented using the same sample restrictions: a) living alone versus with others, among unmarried respondents; b) living with spouse only versus with others, among married respondents; c) living with unmarried chil dren versus not living with unmarried children, among those with one or more such children; and d) living with married children versus not living with mar ried children, among those with one or more such children. As noted previously, in the course of this analysis we first estimated separate regression models for each country for each outcome of interest (re sults not presented). While the country-specific models allow us to examine the extent to which effects are similar or different across countries (e.g., whether the effect of education on living arrangements differs across countries), it is not possible to formally test whether the differences that are observed are statistically significant. In addition, we are also interested in evaluating any remaining differences across countries in the propensity for various living arrangements, once other factors are controlled. In order to address both is sues we pooled or combined the countries into a single data file and estimated one model for each living arrangement outcome. Residual country effects are then assessed by including country indicators for three of the four countries, with Taiwan serving as the omitted group. For factors that exhibited different Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Patterns and Determinants of Living Arrangements 171 associations across countries in the country-specific models, we conducted formal statistical tests for interaction effects. For example, the association between education and living arrangements appeared to be quite distinct for the Philippines in both the country-specific regressions, as well as the bivari- ate results in Tables 5.9-5.10. To test whether the education difference is sta tistically significant, we included interactions between the country indicator for the Philippines and the dummy variables for each level of education. Only those interactions that were found to be statistically significant were retained in the final models and they are presented along with the main effects from the pooled models in Tables 5.11-5.12. As is the convention throughout this volume, results from the logistic regression models are presented in the form of odds-ratios. Odds-ratios above 1.0 indicate that the group in question has a higher likelihood of experiencing the outcome relative to the comparison (omitted) group, and odds-ratios be low 1.0 indicate that the group in question has a lower likelihood of experi encing the outcome. For continuous independent variables, the odds-ratio in dicates the amount of increase or decrease in the log-odds of experiencing the outcome that is associated with a one-point increase in the independent vari able. The first two models focus on the two independent living arrangements, namely living alone and living with a spouse only. Results from these models are shown in Table 5.11. The sociodemographic covariates are listed first in the table, followed by the residual country effects and the country interactions. The only country that exhibited significant interaction effects in relation to these outcomes is the Philippines; interpretation of these results will be given below. Focusing first on the model predicting living alone among unmarried respondents, the results suggest that women are less likely than men to live alone, whereas divorced and separated individuals, rural residents, and those who are currently employed are all more likely than their respective counter parts to live alone. With regard to income there appears to be a curvilinear effect, such that those with moderate levels of income are more likely than those with low or high levels to live alone..t In addition, the likelihood of living alone is negatively associated with both the availability of any children and the number of living children. Specifically, the odds of living alone among those with one or more living child are .4 times the odds for those with no living children. In addition, there is approximately a 10 percent reduction in the odds of living alone with every increase of one in the number of living children. The effects of all of these covariates just described are generally consistent across countries in that no significant country interactions were observed. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Table 5.11. Odds-Ratios for the Effects of Sociodemographic and Health Characteristics on Living Alone or with Spouse Only: Results from Pooled Logistic Regression Models *p<.05 **p<.01 ***p<.001 Independent VariablesAge of respondent60-64 65-6970-74 75-79 80+Female (vs. male)Divor/separ/never marr (vs. widowed)Rural (vs. urban)EducationNo formal schooling Any to completed primary Any secondary and aboveRespondent/spouse's incomeLow income (-lowest quintile) Moderate income High income (-highest quintile) Missing incomeWorking (vs. not working)Self-rated healthHigh/good Moderate Low/poor1+ ADL difficulty (vs. no difficulty)1+ living children (vs. none) Number of living childrenCountry (vs. Taiwan) Philippines Thailand SingaporeCountry interactionsPhilippines*age It 75 Philippines*primary education Philippines Secondary education Philippines*ADL difficulty Philippines'low incomePseudo-R2Live Alone vs. Live w/Others (unmarried respondents)0.98 0.95 1.201.120.53***1.98***1.74***0.97 1.140.61***0.53*** 0.48***1.42**0.91 0.901.100.39*** 0.90***0.96 0.34*** 0.22***0.49*** 0.65* 0.16*** 0.29***.168Live w/Spouse Only vs. Live w/Others (married respondents)0.43*** 0.68*** 0.75*1.121.54***inap.2.04***1.08 1.49***0.62***0.92 0.71**1.050.941.050.810.24*** 0.92***0.69 0.53*** 0.45***0.55** 0.40***2.43***.067 Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Patterns and Determinants of Living Arrangements 173 There were a number of factors, however, for which the effects were significantly different for the Philippines than for the other countries. These are shown in the "country interactions" entries toward the bottom of the table. The first of these factors is age. Whereas age is not associated with living alone net of other factors in Thailand, Taiwan or Singapore, age is a signifi cant factor in the Philippines. Specifically, Filipinos under age 75 are less likely than those age 75-79 to live alone. (This effect was consistent for each of the detailed age groups [60-64,65-69, 70-74] in the Philippines, but due to small sample sizes not all of the individual age interactions were significant. As a result, the groups were collapsed in the country interaction.) Education also exhibited a differential effect in the Philippines, reiterating the bivariate associations observed in Table 5.9. Unlike their counterparts in the other coun tries for whom education is not associated with living alone once other fac tors are controlled, older Filipinos with primary, and especially secondary level education are substantially less likely to live alone than their uneducated peers. This finding is counter to the hypothesis that higher education is asso ciated with a greater preference for independent living, as well as the eco nomic resources to achieve it. Finally, health is also an important factor in the Philippines. Specifically, those who have difficulty performing one or more ADL (for definition, see Chapter 9 on Health Status) are less likely to live alone than those not so impaired. No health effects were observed with re spect to living alone for any of the other countries. Turning lastly to the country effects, the significant odds-ratios for Thai land and Singapore indicate that, other factors being equal, the likelihood of living alone is lower in both of these countries than in Taiwan (the compari son country). These results are consistent with those presented in Table 5.9, which shows substantially lower percentages living alone in Thailand and Singapore compared to Taiwan. Table 5.9 also shows a much lower percent age living alone in the Philippines than in Taiwan. The lack of a residual country effect for the Philippines in the multivariate model indicates that this differential is fully explained by other variables in the model, most impor tantly the interaction effects that are included. The results for the second model predicting living with a spouse only among married respondents show some interesting similarities and differences with the patterns for living alone. Consistent with the previous model, those residing in rural areas are more likely than those in urban areas to live with a spouse only. In addition, low income individuals (in all countries except the Philippines) and those with missing values on income are less likely to live with a spouse only than those with moderate incomes. The propensity for living with a spouse only is also reduced for those with any living children and also as the number of living children increases. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 174 The Well-Being of the Elderly in Asia Differences in associations emerge in relation to age, gender, and edu cation. In the previous model, neither age nor education was associated with living alone, except for older Filipinos. Age has a strong effect on living with a spouse only, however, with those under age 75 less likely to live with a spouse only than those age 75-79. This effect is consistent across all coun tries. In addition, those with secondary education are more likely than those with no formal schooling to live with a spouse only in all countries except for the Philippines. Finally, whereas women were less likely than men to live alone, women are significantly more likely to live with a spouse only. With regard to country interactions, the Philippines again exhibits some deviations as alluded to above. In contrast to the other countries, for which education is positively associated with living with a spouse only, older Filipi nos with primary and secondary education are less likely to live with a spouse only compared to those with no education. In addition, whereas low income is associated with a lower propensity for independent living in the other three countries, in the Philippines those with low income are substantially more likely to live with a spouse only than those with moderate incomes. Finally, the country effects are similar to those observed in the previ ous model. Other things being equal, older Thais and Singaporeans are less likely than their Taiwanese counterparts to live with a spouse only. The lower propensity for living with a spouse only for older Filipinos compared to Tai wanese (as shown in Table 5.9) is to a large extent explained by the interac tions in the model. Some residual country effect remains for the Philippines (odds-ratio of .69) but it is not statistically significant. Although a number of factors were found to be significantly associ ated with independent living, the overall variance explained by the models is fairly small. The model explains approximately 17 percent of the variance in living alone and only 7 percent of the variance in living with a spouse only. Table 5.12 presents logistic regression results for models predicting coresidence with unmarried and married children, respectively. For three of the countries, the samples are restricted to respondents who have at least one child of the specified type; because this information is not available for Singapore, the Singapore sample is conditioned only on availability of one or more living children in both models. As with Table 5.11, results are presented as odds-ratios. Focusing first on the results for coresidence with unmarried children, there is a strong inverse relationship with parent's age, such that the propen sity for living with an unmarried child declines as parent's age increases. Women are less likely than men to live with an unmarried child, as are those who are widowed or otherwise unmarried compared to the currently married. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Patterns and Determinants of Living Arrangements 175 An exception to the pattern for marital status is in Thailand, where widowed individuals are more likely than married persons to coreside with an unmar ried child (note interaction effect in lower panel of table). Other things being equal, there is a sharp reduction in the likelihood of coresiding with unmar ried children in rural versus urban areas. Education and income exhibit strong effects on coresidence with unmarried children, although the effects vary somewhat across countries. In both Taiwan and Thailand, elderly with primary or higher education are less likely than those with no formal schooling to coreside with an unmarried child. This pattern is reversed in the Philippines and Singapore, however, as indicated by the odds-ratios associated with the corresponding interaction effects shown in the lower panel of the table. With regard to income, there is a strong positive association in that the likelihood of coresidence increases with increasing income. This pattern holds in all countries except Thailand, where the association is partially reversed. Multiplication of the income odds-ratios with the odds-ratios for their corresponding interaction variables indicates that low income Thais are more likely to coreside with unmarried children than those with either moderate or high incomes (odds ratios of 1.26 versus 1.01 and 1.00, respectively). Neither work status nor health status are associated with coresidence with unmarried children net of other factors. However, the likelihood of coresidence increases significantly as the number of unmarried children in creases. As discussed above, education, income, and widowhood show varied patterns of association among the countries. Older Filipinos and Singaporeans who received some education and widowed and low income Thais all have a higher than average propensity for living with unmarried children. On the other hand, high income Thais have a lower than average propensity. Net of these differences and of compositional differences on the covariates, how ever, important country differences remain. Thais and particularly Singaporeans are less likely to coreside with unmarried children than Taiwanese. Filipinos also have a slightly reduced likelihood of coresidence compared to Taiwan ese, although the difference is not statistically significant. To some extent the residual country effect for Singapore is likely due to our inability to adequately control for availability of unmarried children. For all other countries, the sample is restricted to those with one or more unmarried child, but for Singapore it is restricted to those with one or more living child. As a result, the Singapore sample may contain a high proportion of respondents who have no unmarried children and for whom coresiding with an unmarried child is, thus, not an option. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Table 5.12. Odds-Ratios for the Effects of Sociodemographic and Health Characteristics on Living with Unmarried and Married Children: Results from Pooled Logistic Regression Models Independent VariablesAge of respondent60-6465-6970-7475-7980+Female (vs. male)Marital statusMarriedWidowedDivorced/separated/never marriedRural (vs. urban)EducationNo formal schoolingAny to completed primaryAny secondary and aboveRespondent/spouse's incomeLow income (~lowest quintile)Moderate incomeHigh income (-highest quintile)Missing incomeWorking (vs. not working)Self-rated healthHigh/goodModerateLow/poor1+ ADL difficulty (vs. nodifficulty)Live with Unmarried Child1.88***1.35***1.25*—0.74**0.81***—0.70***0.55***0.69***-0.63***0.54***0.81**—1.36***1.44**1.071.05—1.021.01Live with Married Child1.071.010.95—0.82*1.04-1.71***1.090.72***-1.041.061.65***—1.051.71***0.67***1.06—0.961.05 Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Patterns and Determinants of Living Arrangements 177 Table 5.12. continued Number of unmarried childreni* Number of married children"Country (vs. Taiwan) Philippines ThailandSingaporeCountry interactionsPhilippines*primary education Philippines*secondary education Singapore*primary education Singapore*secondary education Thailand*widowedThailand*low incomeThailand*high income Singapore*female Taiwan*secondary educationPseudo-R21.23***0.75 0.61***0.19***1.91*** 1.66* 1.85*** 2.83*** 1.99***1.55*0.74*.0971.07***0.68*** 1.070.43***0.54***•1.29** 0.58***.060 aFor the Philippines, Thailand, and Taiwan, samples for each model are restricted to respondents with 1+ child of the specified type. For Singapore, sample in both models is conditioned only on having 1+ living children, since information on availability and number of unmarried versus married children is unknown. ror Singapore, variable represents total number of living children, since information on number of unmarried versus married children is unknown. * p < .05 p<.001 As noted previously, we expect coresidence with married versus unmarried children to respond to a somewhat different set of considerations. In particular, we hypothesize that coresidence with married children will be more responsive to the financial, health and social needs of older parents than is the case for coresidence with unmarried children. The results presented in the far right column of Table 5.12 lend partial support to this hypothesis. With regard to financial needs, the results show that elderly parents with a low level of income are more likely to coreside with a married child than those with moderate or high levels of income. This is in contrast to the pattern for coresidence with unmarried children, for which there is a strong positive association with income. In addition, elderly parents who are retired Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 178 The Well-Being of the Elderly in Asia (or not working) are more likely than those currently working to coreside with a married child. Work status is unrelated to coresidence with unmarried children. These results suggest that coresidence with married children may be responsive to the financial constraints of elderly parents. Another interesting difference in results for married versus unmarried children pertains to the marital status of the elderly parent. Widowed elderly have a substantially higher likelihood than their married counterparts of coresiding with a married child. Other unmarried persons (primarily divorced) are neither more nor less likely than married persons to live with a married child. In contrast, married elderly are much more likely than either widowed or otherwise unmarried elderly to coreside with an unmarried child. Other things being equal, health status does not appear to be associated with coresidence with either married or unmarried children. Age, gender and education show little association with coresidence with married children with a few exceptions. The oldest-old (age 80+) are slightly less likely than those age 75-79 to coreside with a married child, other things being equal, but there is no difference in propensity for those under age 80. The only country for which there is a gender difference is Singapore, where women are more likely than men to coreside with a married child. Finally, in both Singapore and Taiwan, education comes into play in that those with secondary or higher levels of education are less likely than those with no formal education to coreside with a married child. This is consistent with the hypothesis that more educated individuals tend to prefer independent living arrangements. The fact that this education effect is not obsened in Thailand or the Philippines suggests that either such preferences do not operate there or that elderly are less able to act on these preferences due to financial and other constraints. One factor that does show a similar association for coresidence with unmarried and married children is place of residence. Rural elderly are sig nificantly less likely than their urban counterparts to live with a child, regard less of the child's marital status. Also, the likelihood of coresidence with a married child increases with the number of married children, a finding paral lel to that for unmarried children in the previous model. Finally, with regard to country differences, the results mirror the de scriptive findings in Table 5.10, suggesting that the underlying propensity for coresidence with married children is quite similar in Thailand and Taiwan, but that it is somewhat lower in both the Philippines and Singapore. Once again, although a number of factors exhibit strong effects on coresidence with children in both models, the amount of variance in living arrangements that is explained is under 10 percent for each. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Patterns and Determinants of Living Arrangements 179 Conclusions The very substantial social and economic change during the last several de cades in the countries under study does not yet appear to have led to or been accompanied by major shifts in the predominant form of living arrangement among elderly population members. Coresidence with adult children still pre vails widely in the Philippines, Singapore, Taiwan and Thailand. The most notable shift away from coresidence with children is evident in Taiwan where living only with a spouse has risen to levels above those found in the other three countries. But this transformation has been occurring at a rather slow pace compared to the rapid economic and social development that has trans formed Taiwan into one of the most advanced newly industrialized countries in the region. In the case of Thailand, increases in arrangements in which children live very nearby appears to have largely compensated for the modest reduction in coresidence observed there. In the Philippines and Singapore changes appear to be even less pronounced than in Taiwan or Thailand. Despite the remarkable persistence in coresidence in these countries suggested by available data, it is possible that more dramatic change could characterize them in the coming decades. As noted earlier in the chapter, re cent evidence indicates pronounced declines in coresidence in at least two Asian countries, Japan and Korea. In addition, attitudinal data available for Taiwan among young adults suggest further declines in coresidence there in the future. For example, surveys of women aged 20-39 indicate that those who expect to coreside with a married son during their old age declined from 56 percent in 1973 to 45 percent in 1986 (Weinstein et al., 1994). In both instances, these percentages are distinctly lower than the percent of persons aged 65 and over who were living with a married child (almost always a mar ried son rather than daughter) at the time (Hermalin, 1995). Furthermore, fo cus group discussions among the adult children of elderly parents in Thailand elicited frequent views that in the future, when they themselves are elderly, they would be less likely to receive the same care from their own children that they give to their parents and that coresidence would be less common (Knodel, Saengtienchai and Sittitrai, 1995). The patterns of association observed in relation to several socioeco- nomic and demographic factors also suggest the possibility of further declines in coresidence in the coming years. For example, higher levels of education and income tend to be positively associated with independent living arrange ments and negatively associated with coresidence with children. (A key ex ception is the positive association between income and coresidence with un married children, perhaps reflecting a prolonged dependency of children in Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 180 The Well-Being of the Elderly in Asia families who have sufficient financial resources to accommodate such an ar rangement.) In addition, the results show a consistently strong effect of num ber of children on living arrangements, whereby those with larger numbers of children are less likely to live independently and more likely to coreside with unmarried or married children. As the standard of living rises in these coun tries, and as future cohorts of elderly become increasingly educated and have fewer children, we might expect to see continued increases in independent living and declines in intergenerational coresidence. Thus, the modest changes evident to date for the four countries in our project could possibly presage more dramatic future changes in living arrangements. The analysis of determinants of living arrangements revealed a num ber of other key findings, as well. The following table lists the factors that are statistically significant predictors of the various living arrangement outcomes examined in the chapter. Both place of residence (urban versus rural) and number of living children show pervasive effects across all outcomes exam ined, highlighting the importance of availability of kin, as well as more mate rial constraints relating to the high cost of living and relative shortage of land and housing in urban compared to rural areas. Significant Factors Affecting AlternateLiving AloneMale (+)Divorced (+)Rural (+)Income (+/-)Work status (+)# children (-)CountryLiving with Spouse OnlyAge (+)Male (-)Rural (+)Education (+)Income (+)# children (-)CountryLiving ArrangementsLiving with Unmarried ChildAge (-)Male (+)Married (+)Rural (-)Education (-)Income (+)# children (+)CountryLiving with Married ChildWidowed (+)Rural (-)Income (-)Work status (-)# children (+)Country Also of interest is that the dynamics of living alone and living with a spouse only are not always the same. For example, among the unmarried, living alone is much more prevalent for men and for those currently working, and income shows a curvilinear effect, whereby those with moderate levels of income show the highest propensity for living alone. In contrast, living with a spouse only is much more common among women and those of higher socioeconomic status, as reflected by education and income. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Patterns and Determinants of Living Arrangements 181 The dynamics of coresidence with unmarried versus married children are also somewhat different. Whereas coresidence with unmarried children appears to be primarily responsive to the children's needs and life stage, coresidence with married children appears to be more strongly influenced by the physical, social and financial needs of the parents. This may in large part be a function of the way in which the life courses of parents and their children move in parallel, and the opportunities that are thus available at various stages. When children enter adulthood and secure their first jobs or obtain advanced education prior to marrying, their parents are likely to still be fairly young, married, in good health, and possibly working. Likewise, when parents are older, retired, experiencing declining health, and possibly widowed, their chil dren are likely to have become married and possibly even raised families of their own. Hence, the opportunities for living with married versus unmarried children change over time and, at specific points in the life course, may be somewhat constrained. A key point that we have tried to emphasize in this chapter is that an individual's living arrangement at any given point in time reflects the outcome of a complex decision involving multiple family members, each of whom has a unique set of needs and resources to be balanced. Furthermore, living arrangements are fluid and are likely to change during later life, perhaps many times, in response to the changing needs, resources and availability of family members. The present analyses, which are based on cross-sectional data, capture the situation at only a single point in time and in the lives of the survey respondents. To date, little is known about the transitions in living arrangements that occur during later life in these countries. Panel data that follow individuals over time are just becoming available for several of the study countries, and these data will be useful for describing transitions and conducting more refined investigations of their determinants. An example of the potential that longitudinal data offers for studying living arrangements is provided in Chapter 12. Finally, the relation between forms of living arrangements and functions and how this relation may be changing over time is another critical issue to consider. It is indeed these functions that directly determine the implications of living arrangement structures for the well-being of the elderly. As suggested in the framework presented in Figure 4.1, we view living arrangements as an intervening factor that influences various dimensions of well-being, rather than an indicator of well-being in and of itself. In keeping with this perspective, subsequent chapters examine living arrangements as an independent variable to determine what impact they have on specific dimensions of well-being and whether this impact is the same or different across countries. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 182 The Well-Being of the Elderly in Asia ENDNOTES 1. The 1995 Singapore survey includes the following information: total number of sons, total number of daughters; the number of coresident sons; the number of coresident daughters; whether or not the respondent lived with any single children; and whether or not the respondent lived with any ever married children. Based on this information, it is possible to determine for 92% of elderly parents whether or not they are living with a single son, single daughter, married son, or married daughter. The remaining 8% of elderly parents coreside with at least two children of opposite gender and opposite marital statuses. Information on household headship allowed us to sort out the gender and marital status for an additional 3% of the cases. For the remaining 5% of cases, it was not possible to identify the specific gender and marital status combination of coresident children, so these respondents are excluded from all analyses of coresidence. 2. Here and in all other tables in this chapter references to married children include both those who are currently and formerly married (i.e., widowed, divorced or separated). 3. The results for the Chinese elderly are based on 299 elderly whose children are all sons and 224 whose children are all daughters. However, among Malays there are only 30 cases whose children are all sons and 35 cases in which all are daughters and among Indians only 14 have only sons and the same number only daughters. Given that we do not know if any of the children are married, we can not condition the results on having at least one married child as we in effect do for the other three countries. Thus it could be misleading to report more precise results for the Malays and Indians. 4. A dummy variable representing missing income was included in the model as a control. Most of the missing data on income is due to don't knows, although some respondents refused to provide their income. The finding that those with missing income were less likely to live alone or with a spouse only most likely reflects the less complicated financial situation of elderly living alone or with spouse only compared to living in multigenerational households. 5. The rinding that married women are more likely than married men to live with a spouse only, other things being equal, is likely due to the age difference between spouses. Women tend to be several years younger than their spouses, on average. Therefore, a 65 year-old woman is likely to be married to a man in his late sixties, whereas a 65 year-old man is likely to be married to a woman in her early sixties. The former couple is further along in their family life-cycle than the latter couple, and their children are more likely to be married and to have moved away from home. Thus, holding age constant, a married woman is more likely than a married man to be living apart from children and in a couple-only household with her spouse. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Patterns and Determinants of Living Arrangements 183 REFERENCES Aquilino, William S. 1990. "The Likelihood of Parent-Adult Child Coresidence: Effects of Family Structure and Parental Characteristics." Journal of Mar riage and the Family 52(2):405-419. Asis, Maruja Milagros B., Lita Domingo, John Knodel, and Kalyani Mehta. 1995. "Living Arrangements in Four Asian Countries: A Comparative Perspective." Journal of Cross-Cultural Gerontology 10(1/2): 145-162. Burch, Thomas K., and Beverly J. Matthews. 1987. "Household Formation in Developed Societies." Population and Development Review 13(3):495-511. Casterline, John B., Lindy Williams, Albert I Hermalin, Ming-Cheng Chang, et al. 1991. "Differences in the Living Arrangements of the Elderly in Four Asian Countries: The Interplay of Constraints and Preferences." Comparative Study of the Elderly in Asia, Research Report, No. 91-10, Population Studies Cen ter, University of Michigan, Ann Arbor, Michigan. Cornman, Jennifer. 1999. "Understanding the Ties That Bind: Intergenerational Value Agreement in Taiwan." Ph.D. Dissertation. Department of Sociology and Population Studies Center, University of Michigan. DaVanzo, Julie, and Angelique Chan. 1994. "Living Arrangements of Older Ma laysians: Who Coresides with Their Adult Children?" Demography 31(1):95- Domingo, Lita, and Maruja Milagros Asis. 1995. "Living Arrangements and the Flow of Support between Generations in the Philippines." Journal of Cross- Cultural Gerontology 10(1/2):21-51. Hermalin, Albert I. 1995. "Aging in Asia: Setting the Research Foundation." Asia- Pacific Research Reports, No. 4, East-West Center, Program on Population. . 2000. "Ageing in Asia: Facing the Crossroads." Comparative Study of the Elderly in Asia, Research Report, No. 00-55. Population Studies Center, University of Michigan, Ann Arbor, Michigan. Hermalin, Albert I., Mary Beth Ofstedal, and Ming-Cheng Chang. 1991. "Types of Supports for the Aged and Their Providers in Taiwan." In T. Hareven (Ed.) Aging and Generational Relations Over the Life Course: A Historical and Cross-Cultural Perspective. Berlin: Walter de Gruyter. Hermalin, Albert I., Carol Roan, and Aurora Perez. 1998. "The Emerging Role of Grandparents in Asia." Comparative Study of the Elderly in Asia, Research Report, No. 98-52. Population Studies Center, University of Michigan, Ann Arbor, Michigan. Japan Aging Research Center. 2000. Statistical Abstracts of Aging in Japan, 1999/ 2000. Tokyo: Japan Aging Research Center. Kim, Ik Ki. 1999. "Population Aging in Korea: Social Problems and Solutions." Journal of Sociology and Social Welfare 26(1): 107-123. Knodel, John. (Guest Editor). 1992. Asia Pacific Population Journal 7(3) (Theme issue on Population Aging and Support of the Elderly). Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 184 The Well-Being of the Elderly in Asia Knodel, John. (Guest Editor) 1995. Journal ofCross-Cultural Gerontology 10(17 2) (Theme issue on Focus Group Research on the Living Arrangements of Elderly in Asia). Knodel, John. (Guest Editor). 1997. Asia Pacific Population Journal 12(4) (Theme issue on Population Aging and Support of the Elderly). Knodel, John, and Napaporn Chayovan. 1997. "Family Support and Living Ar rangements of Thai Elderly." Asia-Pacific Population Journal 12(4):51-68. Knodel, John, and Chanpen Saengtienchai. 1999. "Studying Living Arrangements of the Elderly: Lessons from a Quasi Qualitative Case Study Approach in Thailand." Journal ofCross-Cultural Gerontology 14(3): 197-220. Knodel, John, Chanpen Saengtienchai, and Werasit Sittitrai. 1995. "The Living Arrangements of Elderly in Thailand: Views of the Populace." Journal of Cross-Cultural Gerontology 10(l/2):79- 111. Logan, John, Fuqin Bian, and Yanjie Bian. 1998. "Tradition and Change in the Urban Chinese Family: The Case of Living Arrangements." Social Forces 76(3):851-82. Lopez, Maria E. 1991. "The Filipino Family as Home for the Aged." Compara tive Study of the Elderly in Asia, Research Report, No. 91-7. Population Studies Center, University of Michigan, Ann Arbor, Michigan. Martin, Linda G. 1989. "Living Arrangements of the Elderly in Fiji, Korea, Ma laysia, and the Philippines." Demography 26(4):627-643. Mason, Karen O. 1992. "Family Change and Support of the Elderly in Asia: What Do We Know?" Asia Pacific Population Journal 7(3): 13-32. Merita, Kalyani, Mohd. Maliki Osman, and Alexander E.Y. Lee. 1995. "Living Arrangements of the Elderly: Cultural Norms in Transition." Journal of Cross- Cultural Gerontology 10:113-43. Natividad, Josefina N., and Grace T. Cruz. 1997. "Patterns in Living Arrange ments and Familial Support for the Elderly in the Philippines." Asia-Pacific Population Journal 12(4):17-34. Ofstedal, Mary Beth. 1995. "Coresidence Choices of Elderly Parents and Their Adult Children in Taiwan." Ph.D. Dissertation. Department of Sociology and Population Studies Center, University of Michigan. Ruggles, Steven. 1994. "The Transformation of American Family Structure." American Historical Review 99( 1): 103-28. Shantakumar, G. 1994. The Aged Population of Singapore. Census of Population, 1990: Monograph No. 1, Singapore. Weinstein, Maxine, Te-Hsiung Sun, Ming-Cheng Chang, and Ronald Freedman. 1994. "Co-Residence and Other Ties Linking Couples and Their Parents." In Arland Thornton and Hui-Sheng Lin, eds., Social Change and the Family in Taiwan. Chicago: University of Chicago Press. Wolf, Douglas A., and Beth J. Soldo. 1988. "Household Composition Choices in Older Unmarried Women." Demography 25(3):387-403. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Chapter 6 INTERGENERATIONAL SUPPORT AND TRANSFERS Ann Biddlecom, Napaporn Chayovan, and Mary Beth Ofstedal Intergenerational support and transfers—which commonly take the forms of coresidence, time, and money—are fundamental aspects of the social fabric in society. Intergenerational support systems play a key mediating role between broad societal-level changes, such as declining fertility rates and rising GNP, and the individual welfare outcomes of the elderly, such as physi cal health and economic status (see Figure 4.1 in Chapter 4). Moreover, fam ily transfers have social and economic consequences that can directly im pinge on the design and effectiveness of public support programs and policies (Mason, Lee and Russo, 2001; National Research Council, 2001; Schoeni, 1997). Both family and public transfers are often intended to smooth the dif fering consumption needs and productive contributions that occur across the life course, and when changes occur in one source of support, there are likely to be accommodations in the other source of support. Family support and transfer systems are particularly critical to exam ine in Asia because of the historically important role of family care for the elderly, especially elderly parents, and the recent broad-level changes that are sweeping the region (see Chapter 2). For example, survey data show that transfers from family members remain central to the economic well-being of most older persons in Asia. Income from family members, primarily adult children, was the primary source of income for the majority of the elderly in Fiji, Korea, Malaysia, and the Philippines in the mid-1980s (Andrews et al., Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 186 The Well-Being of the Elderly in Asia 1986: 72); Taiwan, Singapore, Thailand, and the Philippines in the mid-1990s (see Chapter 8); major urban areas of Vietnam in 1996-1997 (Anh, Cuong, Goodkind and Knodel, 1997); Cambodia in 1998 (Kato, 2000); and South Korea in 1998 (Yoon and Cha, 1999). In Malaysia, the majority of elderly receive money transfers from children, regardless of whether this is a primary source of income or not (Lillard and Willis, 1997). Recent studies of urban areas in China find that from one-third to one-half of Chinese elderly receive financial assistance from family (Chen and Silverstein, 2000; Hermalin, Ofstedal and Shih, forthcoming). Intergenerational support systems are not static institutions in the face of broad demographic, economic and social changes. Rather, these systems of family support come under strain and are forced to adapt in both function and form (Hermalin, 1998; Martin and Kinsella, 1994; Mason, 1992). In this sense, Asia is at a crossroads. Strong norms of obligation for intergenerational support, still very much in force, currently coexist with massive changes on key demographic and socioeconomic dimensions such as lowered fertility, longer life expectancies, increased female labor force participation, and a switch from agriculture-based economies to manufacturing and service-based economies. Given current and projected increases in the older population in both proportionate and absolute terms, governments face questions about the degree to which family support is or will continue to be sufficient for the needs of the elderly (Holden, 1996; Lee and Mason, 2000; Martin, 1990; Mason, Lee and Russo, 2001; Ogawa and Retherford, 1997; Phillips, 2000; Silverman, 1995). Concern is also growing in a number of political and scien tific quarters that the traditional system of familial support for the elderly might be eroding as couples bear fewer children, as more adult children live away from their parents, and as women enter the labor force in larger num bers (Holden, 1996; Westley, Lee and Mason, 2000; Silverman, 1995). It thus becomes crucial to understand intergenerational support because of the key intermediate role that it plays in the well-being of the elderly and the relevance it has for public policy, as countries seek to attend to the emerg ing needs of the elderly without undermining the strong norms for family support of the elderly. Family transfers are quite difficult to measure (Hermalin, 1999), but governments are increasingly pressed to monitor sources of pri vate support as they respond to societal-level changes and public programs to meet the needs of the elderly. As governments continue to develop and imple ment policies and programs in Asia, there is a concomitant need to study how private transfer patterns adapt to new public sources of support and what implications these changes have for future state initiatives. This chapter describes changes to the family support system in the Philippines, Thailand, Taiwan, and Singapore; provides a detailed snapshot Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Intergenerational Support and Transfers 187 of current intergenerational support systems and covariates of transfer pat terns; and draws on this evidence to speculate on likely changes in family support in the 21st century. Much of the existing research on informal support in old-age in Asia has focused exclusively on support received by the elderly and overlooked support provided by the elderly to others. In the countries examined here, we have the opportunity to document support in both direc tions, yielding a full picture of the role of older persons in their support net works. The chapter begins with an overview of measurement issues and re sults from prior studies on intergenerational transfers in Asia. Descriptive assessments of attitudes about support, kin availability, and support behavior follow. We specifically examine what expectations exist for intergenerational care and support, what kind of support each generation actually provides, and who provides it (particularly with respect to gender). The chapter concludes with multivariate models of support flows and a discussion of likely scenarios for future intergenerational support in Asia. Measurement Issues The measurement of support is complex. Important dimensions of measure ment are currencies, timing, participants, and magnitude (see Chapter 4; Hermalin, 1999; National Research Council, 2001; and Soldo and Hill, 1993, for reviews). The currencies, or forms of support, are space (living arrange ments), time (social contact or visits, childcare or help with daily activities), money (cash) and material goods (food, clothing and other goods). Quantify ing the amount of any type of transfer can be difficult. Respondents may not know the fair market value of property or material goods or recall accurately financial transfers made over a period of time. Time transfers can be mea sured in terms of the frequency of contact and/or duration, and by the nature and intensity of services provided. Because the currencies and forms of support are so varied, comparative analyses of support across countries or even within the same country across time are challenging. Even how one particular currency is measured can vary, thus confounding comparisons in support across time or space. For example, survey data on the reported level of material support received by older persons in Taiwan showed a decrease from 1989 to 1996. However, the question about material support—which is often transferred in the form of gifts on special occasions as well as an ongoing form of support—changed between surveys such that gifts for holidays or birthdays, included in the 1989 survey, were explicitly excluded in the 1996 survey. The four country surveys used in the analyses presented here contain information on whether respondents received Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 188 The Well-Being of the Elderly in Asia or provided money or material goods from/to others. However, the questions about money and material goods differ between countries in both wording and about whom the question concerns. For example, questions about money transfers in the Philippines exclude "small gifts"; in Thailand the questions are about transfers of "1,000 Baht or more"; there are no restrictions in the Taiwan data; and in the Singapore survey the question about older people providing money refers only to coresident children. Material goods in the four country surveys refer to household necessities, food, clothing and the like, but again the question wording and referent groups vary among the surveys. Specific questions used in these surveys to measure transfer flows are described in Notes for Tables (Table 6.2) at the end of this chapter. Transfers also occur over time, usually forcing researchers to impose arbitrary windows of time in which to collect information. Current transfers could mean support given in the prior week, month, or year. Or researchers may look at transfers over the entire life course to explore family support issues such as whether parents invest in their children early in life in ex change for support from their adult children later in life (Henretta et al. 1997). Transfers can also be measured into the future, as is the case for bequests, where a usually large transfer of money or assets such as a house or land is made upon the death of the parent. The survey data used in this study contain information about current transfers using varying time periods. In Singapore, the time frame is 6 months prior to the interview for receipt of money and material goods. In Thailand, the Philippines, and Taiwan, most questions re fer to transfers made within the past year (prior to the interview). Survey questions about more nuanced dimensions of transfers, such as the magni tude of transfers or the identification of specific others in the exchange net work, are even more varied across countries. It is worth stressing the importance of defining time boundaries for the measurement of money transfers. Parents may make relatively large money investments in their children's future at certain points in the life course: money to a son or daughter to buy a home, money for higher education, or money to start a new business. These large investments are significant but can be missed if questions about parent-child transfers focus only on current transfers. Tai wan was one of the few countries that attempted to measure large lifetime transfers of money between parents and children as well as current money transfers. The Thailand and Philippines surveys included questions about past transfers concerning the division of property (home, land, or other assets) among children and other family members. For the sake of comparison across all four countries, however, we restrict our analyses mainly to current money transfers. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Intergenerational Support and Transfers 189 Measurement of transfers can also be sensitive to who is doing the reporting. In a comparison of Taiwanese parents' and children's reports of the transfers children made to their parents, Roan, Hermalin and Ofstedal (1996) found a relatively high degree of concordance in reports (upwards of 65 per cent) for the following types of support: money, help with household chores, and visits. When parents and children gave differing reports of transfers, the pattern tended to be children reporting greater support than did parents. In this chapter, intergenerational transfers are measured from the reports of the older respondents. In all countries the surveys collected information about exchanges between older persons and their children. Except in Singapore the children were individually identified and a number of their characteristics were gathered from the older respondents. Though we use the term "intergenerational" to describe transfers in this chapter—and certainly children are important participants in the support systems of the elderly—the measures we use for transfers of money and ma terial goods include exchanges with siblings (intragenerational exchanges), other family members, and non-relatives, such as friends. The countries var ied somewhat in how they defined the pool of potential providers and recipi ents. In the Thailand survey questions about material goods are with respect to children who do not live with the respondent; in other countries the ques tion refers to all children, coresident or not. In Singapore, the question about the provision of monetary support is with respect to coresident children only; in the other three country surveys all children are included. In examining time transfers, the focus is on children when measuring the frequency of social contact and visits, and in looking at the degree to which older persons with grandchildren assist with childcare. Older people may also require time transfers to assist them with activities of daily living (ADl.s). such as dressing or bathing, or with instrumental activities of daily living (lADLs), such as shopping or meal preparation. Older people may also provide assistance to others with household chores or personal assistance. On these dimensions the potential pool of providers and recipients encompasses a wider array of family members and others, including paid providers and domestic helpers. However, comparable data from the four country surveys are limited in this regard. Questions about help with ADLs and lADLs were asked in different sequences, different tasks were listed, and different forms of evaluation were used. For example, in Taiwan respondents are asked whether they have difficulty doing any daily activities (e.g., eating or moving about the house), and if they do, whether there is someone who helps them. In contrast, interviewers in the Singapore survey determined through observation and checking with respondents or family members the need for physical Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 190 The Well-Being of the Elderly in Asia assistance with ADLs. The difficulty in "standardizing" across countries the measure of time transfers limits the scope of our analysis on this dimension of support. In general, data on intergenerational transfers identifies exchanges be tween an elderly person or couple and any other person in a specified class, and the analysis utilizes the dyads to measure the frequency and nature of the exchanges in which older persons are engaged. When the data are sufficiently detailed in terms of the characteristics of those with whom the exchanges take place and the nature of these exchanges, as is the case with the data in the Philippines, Taiwan, and Thailand, it is possible to develop more global net work measures of exchange that take into account the interrelationships across dyad transfers (Walker, Wasserman, and Wellman, 1993). A focus on the whole family support network enriches our understanding of intergenerational sup port by connecting the transfer activities of family members to each other and enabling analysis of the tradeoffs and strategies that occur within families (Hermalin, 1999). Indeed, recent anthropological studies suggest that mul tiple exchanges are often used as part of a family investment strategy (Cicirelli, 1994; Peterson, 1990, 1993). For example, in the Philippines there is evi dence that older, presumably more independent children contribute resources to their parents for investment in younger siblings, such as financing their education or small business ventures (Cicirelli, 1994; Peterson, 1990, 1993). In the analyses presented here, we focus mainly on dyadic transfers, but re sults from other research that examines transfer flows within broader kin net works will be described. Background Prior research on intergenerational support and transfers in Asia has tended to be country or setting specific, often with an emphasis on the relationship between family support and family size or composition (Frankenberg, Beard and Saputra, 1999; Knodel et al., 2000; Lee, Parish and Willis, 1994; Peterson, 1993; Rahman, 2000). As multiple data sets on aging have become available, comparative research on transfers in Asia has pointed to the variability in family support systems across countries, and the differential pathways the systems may take in response to ongoing socioeconomic and demographic change (Andrews et al., 1986; Chen and Jones, 1989; Hermalin, 1995; Martin and Kinsella, 1994; Mason, 1992; Phillips, 2000). By its very nature, comparative research also highlights cultural differences that can lead to distinct family support systems, such as studies on gender differences in transfer Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Intergenerational Support and Transfers 191 behavior across countries, and particularly the importance in some countries of having sons for support in old age (Ofstedal, Knodel and Chayovan, 1999). A useful conceptual framework underlying the study of intergenerational transfers is that support is determined by the needs of the elderly, the avail ability of potential providers and the ability and willingness of the providers to meet the perceived needs (see Chapter 4 and Chapter 5; Hennalin, Ofstedal and Chang, 1996; Soldo and Hill, 1993). Needs are difficult to measure but assessments of income sufficiency and physical and emotional health can provide indicators. Chapter 11 on the vulnerable and disadvantaged elderly addresses measurement issues that bear on the question of need. In the analy ses below we utilize several measures that speak to needs as well as to the size of the kin network. One does not expect a one-to-one correspondence between needs and the receipt of support. A study using an earlier wave of the Taiwan data found, for example, that while the likelihood of receiving ADL assistance increased as the level of functioning decreased (measured as an index of the number of activities the respondent reported having difficulty doing), about 80 percent of older adults who experienced difficulty with 3 or more ADLs said they neither received nor needed assistance (Hermalin, Ofstedal and Chang, 1992). Moreover, there was a substantial discrepancy between the proportion of older adults who received help with ADLs (6 percent) and the proportion reporting a high level of difficulty with ADLs (27 percent). In short, although intergenerational transfers may be made in response to actual or perceived need, there will be cases where the elderly have needs that are not being met. In addition, transfers may be motivated by factors other than need, such as individual preferences or culturally defined obliga tions. Preferences may be expressed on a variety of dimensions, such as a preference for privacy that might lead older adults to live separately from their children. An older adult might also wish to honor the cultural obliga tions of the younger generation to support the older generation, leading an older adult to accept support from children despite the absence of a strong need for that support. A growing area of research on intergenerational transfers examines motivations for transfers, the results of which can help shed light on how the family support system will respond to public programs targeted to the welfare of the elderly. Four economic models of private transfer behavior predominate in the literature (see Chapter 4 for more details), altruism, warm glow, exchange, and insurance (Lillard and Willis, 1997; Schoeni, 1997; Soldo and Hill, 1993), and the evidence—primarily from the United States—does not conclusively support one over another. The altruism model assumes that private Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 192 The Well-Being of the Elderly in Asia transfers are made to those in relative need. The warm glow model assumes that the people who give money derive personal satisfaction from doing so regardless of the level of need of the recipient. The exchange model, as its name implies, assumes that people provide money (or time) in return for something else. Lastly, the insurance model of private transfers posits that such transfers are made when recipients experience a drop in financial resources or are unable to finance a substantial investment (e.g., buying a home). Testing among competing hypotheses is often difficult because of the measurement challenges for distinguishing the effects of transfer motivations. The analyses in this chapter are not structured to test the different hypotheses about motivations for transfers, but the arguments about motivations are important given their implications for how private transfers might respond to public forms of support for the elderly. The underlying question for many policymakers and social scientists is how are family systems of support responding to rapid demographic and economic change? Only longitudinal data can fully address this question. One major strategy for monitoring transfer systems is to conduct detailed cross- sectional studies over time to obtain trend data on which dimensions of fam ily systems change and which remain relatively stable. Panel studies (i.e. fol lowing the same people over time) are also needed because there are likely to be several transitions in family support as individuals age through the life course. Understanding when these individual transitions take place and their nature is important for developing effective public policies that address the well-being of the elderly. We present some figures below from a series of surveys over time to describe changes in family support of the elderly and research findings based on panel data that highlight support changes over the individual life course. Repeated cross-sectional suneys and censuses often allow examina tion of one important form of intergenerational transfer—coresidence—be cause it is relatively easy to measure household structure in a consistent man ner. As Chapter 5 indicates, the majority of older adults studied here coreside with at least one adult child, ranging from 69 percent in Taiwan to 85 percent in Singapore, and any decline in coresidence has been modest. The largest decline among the study countries occurred in Taiwan, where parent-child coresidence decreased by about 15 percentage points over two decades. De clines in extended family coresidence have been more extreme in South Ko rea and Japan, however, where only about half of elderly adults live with their children (Kono, 1998; Yoon and Cha, 1999). Longitudinal data on money and other transfer currencies are harder to find. With caveats about the comparability of data and measures, we can still make a few observations about family transfer behavior in Asia. The most Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Intergenerational Support and Transfers 193 important point is that there is little evidence of a large decline in older adults' participation in family exchanges, especially with receiving money support from family and friends (see table below). Surveys of the aged carried out in the Philippines, Thailand, Taiwan, Indonesia, and Singapore in the 1980s show that a large proportion of elderly women and men received money (or material support) from family at that time (Chen and Jones, 1989: Table 5.7; Knodel, Chayovan and Siriboon, 1992). Approximately a decade later, the percentage of older persons receiving money support from family continues to remain high. In the Philippines in 1984, 60 percent of elderly women and 43 percent of elderly men received such family support (though the survey was not nationally representative). By 1996, more than 4 out of 5 elderly Filipinos received financial transfers from their children or other family members and friends (Natividad and Cruz, 1997: 29). A similar pattern occurred in Thailand. Percentage of Adults Age 60 Years and Older Receiving Money from Family or Friends Country (survey year)Philippines(1984)"(1996)Thailand(1986)" (1995)Singapore (1986)" (1995)Taiwan(1989) (1996)Indonesia(1986)aI.(1993)"Women608859 96867783Percentage Men43885787868460 6062Total~8858 68cc55 Sources: Derived from Chen and Jones (1989, Table 5.7), Frankenberg, Lilian! and Willis (2000), Knodel, Chayovan and Siriboon (1992, Table 7), and authors' tabulations of 1989 Taiwan survey and 1995-1996 Taiwan, Philippines, Thailand and Singapore surveys. 3 Measured as receiving money or material support from family or friends. Measured as receiving money from non-coresident children. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 194 The Well-Being of the Elderly in Asia While interpreting this change is difficult because of the different measures and samples used in each survey, it seems clear that private money transfers have not declined and that a very high proportion of elderly Filipinos and Thais are financially supported to some degree by their families. In Singapore in 1986, 97 percent of elderly women and 86 percent of elderly men reported receiving money or material support from family or friends, with the vast majority of support coming from children or grandchildren (Chen and Jones, 1989: 59). Though the questions were asked differently, a similarly high proportion of Singaporean elderly reported receiving money support from family members in the 1995 survey (Chan, 1997). In Taiwan, money support from family has dipped just slightly over time, from 71 percent in 1989 to 68 percent in 1996, with most of the decline attributed to fewer elderly women receiving money from family. While there appears to be a decline in family support in Indonesia between 1986 and 1993 (Chen and Jones, 1989; Frankenberg, Lillard and Willis, 2000), the measurement differences qualify any firm interpretation and the majority of elderly Indonesians still receive monetary support as of the early 1990s. The ASEAN surveys in Thailand and Indonesia in the mid-1980s also found that most elderly reported family members as their main source of money or material support (Chen and Jones, 1989: 58), a finding that still holds for the elderly one decade later (see Table 8.3 on the economic status of the elderly). Norms and Attitudes about Family Support The still relatively high levels of intergenerational support in Asia reflect the strong norms of filial support characterizing all four Asian countries (Domingo and Asis, 1995; Lopez, 1991; Mason, 1992; Thornton and Lin, 1994). In Tai wan and Singapore, responsibility for parental support traditionally has been placed on a married son (often the oldest), while in the Philippines and Thai land roles are more gender-neutral. Studies of the Philippines and Thailand show no consistent gender or birth-order preferences in intergenerational sup port such as coresidence, school investments, and land and non-land (such as a house or material goods) transfers (Ofstedal, Knodel and Chayovan, 1999; Natividad and Cruz, 1997; Quisumbing, 1994). Focus group data indicate that older persons actively participate in family decisions in Taiwan and that the larger the family, the more likely this is the case (Williams, Lin, and Mehta, 1994). In the Philippines, participation is more directly tied to the role the older person plays in the family exchange system. Older parents who provide support to their children are more likely to control household decisions than those who do not provide support. The Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Intergenerational Support and Transfers 195 qualitative data also indicate that Filipino elderly who redistribute money they receive from children are viewed as being "in charge" of the household (Williams and Domingo, 1993). Several questions on attitudes about intergenerational exchange were included in the country surveys (with the exception of Thailand). In the Phil ippines, more than four out of every five older adults agreed that parents should help out their adult children financially, if they were capable of doing so, and that they are obligated to take care of grandchildren. A similar ques tion in Taiwan about helping out adult children by caring for grandchildren also found overwhelming support among older adults. Concerning flows in the opposite direction, a majority of older adults in Singapore and Taiwan felt that children should support their older parents. In Singapore, the question was phrased with respect to a proposed law "to make it compulsory for children to support their parents" (now officially on the books as the Maintenance of Parent's Act [see Chapter 3]). More than three-quarters of Singaporean elderly were in favor of such a law. Among the 20 percent who said they did not favor such a law, the main reasons were that they did not want to force children to provide support and that they thought providing support should be up to the individual. Interestingly, among those who favored such a law, 87 percent said they would not bring their child to court—the teeth of the law—if indeed the child refused to support the parent. It appears that while most Singaporeans favor a public mandate to protect traditional obligations of support, few are willing to have private conflicts played out in a public sphere. In Taiwan, 84 percent of older adults agreed with the statement, "When parents get old, children should give them money for their living expenses," and only 7 percent disagreed. However, this sentiment may not mean the elderly expect full support: 65 percent of Taiwanese older adults agreed that "elderly people should be economically independent, rather than dependent on their children" (24 percent disagreed and 11 percent were neutral). In a separate analysis using 1989-1999 panel data in Taiwan, elderly respondents generally held positive views on the availability of social support and their views were relatively stable across the ten-year period (Cornman et al., 2001). Of course, these norms and attitudes about filial support are facing challenges. As younger cohorts with greater access to market sources of support (such as private pensions) and with higher levels of education themselves become older, they may place a higher value on independence, and family support may decline in favor of state and market substitutes. For example, in Taiwan between 1973 and 1985, expectations of future support from children among reproductive-age women declined (Chang and Ofstedal, 1991). In the Philippines, the cross-national Value of Children survey of 1975-76 showed Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 196 The Well-Being of the Elderly in Asia that more than 80 percent of married women expected support in old age from both sons and daughters (Chen and Jones, 1989: 50). In contrast, a 1993 survey of reproductive-aged women found that slightly more than half said they expected to live with their adult children in old age and only about one- third expected to receive money or material goods from family (Biddlecom and Domingo, 1996). Similar changes have occurred in Japan, where the proportion of middle-aged women who indicated their expectation to be supported by their children in old age declined from 70 percent in 1950 to just 27 percent in 1986 (Kendig, 1989: 22). However, the degree to which expectations translate into actual behavioral outcomes is a still a matter of debate. For example, a study in Taiwan showed that between 1963 and 1991 norms for obligations to extended kin weakened while social contact with kin increased (Marsh and Hsu, 1995). In short, although norms and attitudes favoring kinship transfers both to and from the elderly remain strong in these Asian countries, there are signs that future older generations will expect to be less dependent on family members, especially adult children, for support. Whether these changes in attitudes translate directly into declines in actual resource transfers is likely to be influenced by future social and economic developments. Availability of Children and Social Contact The availability of children with whom to exchange support is not a con straint in these four Asian countries as it might be for countries that have experienced low fertility rates for many decades. Fertility, while quite low among present-day adults, was much higher for today's older generation (see Chapter 2). Based on survey data, the family size of elderly adults ranges from an average of 4.4 living children in Taiwan to 5.9 living children in the Philippines. The vast majority of older persons in these countries either reside with or adjacent to a child, or see a child daily (see Table 5.9). Because of relatively large family sizes and the socioeconomic changes of the past couple decades, some of the children of today's older generation also live far away from them. In the Philippines, older people report about two children on aver age who live outside their province, and in Thailand the figure is 1.5 children. In Taiwan, older persons report two children on average who live outside their neighboring village, town, or city. Also, nearly one in five older Filipi nos has a child living overseas, illustrating the importance of international labor migration for the Philippines (Rodriguez, 1998). This migration is a predominantly female phenomenon: 14 percent of elderly Filipinos have at least one daughter residing abroad and 8 percent have at least one son abroad. In Taiwan and Singapore this phenomenon is weaker but still impressive: Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Intergenerational Support and Transfers 197 7 percent of older adults in Taiwan and 8 percent in Singapore have at least one child living abroad. The majority of older adults in these four Asian countries visit with their non-coresident children on at least a weekly basis (see Table 6.1). Among older adults with at least one non-coresident child, over half in the Philip pines, two-thirds in Taiwan and Singapore, and three-quarters in Thailand have weekly contact with a child. Interestingly, visits are not "one-way streets" where children come to see parents, though that is the more common form. In the Philippines and Thailand, older parents are almost equally likely to make weekly visits to their children (47 percent in the Philippines and 55 percent in Thailand said they visit a child at least weekly). A study in Singapore found that children most often initiate visits, but a non-trivial proportion of older adults also initiate visits with their children (Chan, 1999: 102). In short, most older adults are involved in their children's lives either by virtue of sharing the same household (see Chapter 5 on living arrangements) or by frequently visiting with children. The nature of this involvement is another issue alto gether. Are older parents sought out for advice or companionship or are they serving as free childcarc providers, or perhaps both? Or are children provid ing a service to their parents through these visits, either by checking to make sure all is well or by providing more tangible support, such as preparing a meal? The next few sections address more concrete types of transfers be tween older adults and others. Current Flows of Intergenerational Support Table 6.2 shows the distribution of exchanges between older adults and oth ers of money (first panel) and material goods (second panel). In Thailand and the Philippines, almost all older adults—90 percent or more—reported that they either receive or provide money or material goods in some form. Nearly half or more of older adults in these countries provide and receive money or material goods. Thus, many older persons in the Philippines and Thailand are not so much dependent on family support as they are active participants in family exchange networks, both giving and receiving resources. This pattern is very distinct from Taiwan where less than 6 percent of older adults give and receive money and less than 1 percent give and receive material goods. Instead, older Taiwanese predominately receive money from others (62 percent). Overall exchange patterns in Singapore are qualified somewhat by the different data measurement used; the survey asked solely about monetary exchanges with coresident children. Even with this restriction, however, 25 percent of older adults in Singapore reported both providing and receiving Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 198 The Well-Being of the Elderly in Asia Table 6.1. Percent Distribution of Weekly Visits between Parents and Children among Elderly Adults with at Least One Non-Coresident Child Note: Totals do not all sum to 100.0 due to rounding error. Parent has at least weekly visit with a childYesNoOf those who visit at least weeklyChild visits parentYesNoParent visits childYesNoPhilippines (1996)584254464753Thailand (1995)762574265545Taiwan (1996)6634-----Singapore (1995)6734------ money; and only 7 percent reported not being involved in any money ex changes with children. Elderly Who Receive Support We now focus on each direction of resource transfers and examine some of the important sociodemographic characteristics associated with receiving and providing support (see Table 6.3). In general, the vast majority of older adults in all four countries receive money support (ranging from 68 percent in Tai wan to 91 percent in Singapore), similarly high percentages of older adults also receive material assistance, the exception being Taiwan (15 percent). The demographic characteristics in the gerontological literature usually associated with lower levels of well-being—being older, female, less educated or unmarried—are also associated here with a higher likelihood of receiving support. However, no characteristic is significantly associated with receiving money and material goods in all four countries. Receiving both types of support from others is more common at older ages and among women in Taiwan and Singapore. The elderly in rural areas more often report receiving money and material support: the differentials are significant for money transfers in each Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Intergenerational Support and Transfers 199 Table 62. Percent Distribution of Money and Material Transfers Involving Older Adults Nate: Totals do not all sum to 100.0 due to rounding error. See Notes for Table 6.2 at the end of this chapter. Money TransfersReceived and providedReceived, did not provideProvided, did not receiveNeither received nor providedTotalTotal receivingTotal providingMaterial TransfersReceived and providedReceived, did not provideProvided, did not receiveNeither received nor providedTotalTotal receivingTotal providingPhilippines59.028.98.23.9100.087.967.269.421.26.72.7100.090.676.1Thailand49.338.85.46.5100.088.154.744.842.72.310.2100.087.547.1Taiwan5.662.24.028.2100.067.89.60.813.73.781.9100.114.54.5Singapore25.465.13.06.5100.090.528.4—~————_ country and for material support only in Thailand. Education is strongly associated with the receipt of assistance but the patterns differ across countries. In the Philippines and Thailand those with primary level education are more likely to receive support than either those with no education or those with a secondary or higher education. In Taiwan and Singapore, by contrast, the pattern is linear, with those in the lowest educational category most often receiving money or material goods. Overall, older adults who live with children are much more likely to receive financial or material assistance. Differences by marital status tend to be small; those that are statistically significant tend to show a higher level of support for the unmarried, except for material goods transfers in Thailand. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 rt r" •O (J if eo Is II 8 tf §3 -S |•a•co1T3U>1u§sT3U.CKgaaCC/5I•a^3grt•aing5.a.~£g1.5§.St-1.3i.g11 Charasteristicm NOr- odt- 00I- OOO ONON mO 00OO 00r- ooON — •00 ONo —00 TtOO ON00 O3ts sr- r- r- oo00 00ON-.NO ONOO OOs +g§^oo$ie ON0?^OOCJ1^ l^t'eJ^~13|N— '—_CN^^^^~"e«sH)i0.IUexror--i0rip06oopf •-,OOi1ts00^00I- u•H ^r^ON00Oton0p00r Im0iNO000600000 00—s8iX•282—I/-JQvCN-_^^CN~S1i|oottogfpv^omm^00O OI1N-^UNOr-p'-;0000mON00ONOO00NOOOm00Marital statwMarriedg_ g06 ^OO "NOOr-is!2— —^^p?Tj* ??vS00 Z_ oo ~ oots OON Oo-r i^^O o ts PON *~-00r- "NO O0o0?oo Cs00 tMoo -—in^^NO *ood 5 ^.2|^1 133z u1ivi-+ON00C)—Ot:o r-0 0000tssRsidenseRrtal1i—O}0000ON00i*00oo^00u.II2p. '^_^^J,r^g^f^,r-o1pNOCO/ — . 00p^e-^00e*g§,0011v.^1u Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 -S .a Ed.cationoot~:mTfOOr-g0Orn r-a-00ooNo fmmal sshoolin-odr-in00 5!an ooooCNON ooo9Primary ed.sationNO OCN0000oPivotsOOoo, Semendary ed.sationooJs11oo81~-Chissq.are (psvalue)Livin- Arran-ementooC4OOoosO|utoovi0000Alone^*Sp_se onlyemin wS000 inTfp vit-O oooow^*Married shildboo**loo00oooo oo§pooonio00 00w^*Unmarried shild*i00 Sints vvOO --.00 ^^CN Qod P00 ^-/0s»^"i o^ ^v0^ 8 s «*0 t^r~ O oo o•o1D. 1llrs Soo ^in O2 Sn,oo •* Ov O133£ |i |O Tt00 ts^ 43te n H v} Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 202 The Well-Being of the Elderly in Asia Who primarily provides support to older persons? Table 6.4 shows that children are the most common sources of money and material support to the elderly, with substantial gender differences by country. In the Philippines, Thailand, and Singapore at least 94 percent of older persons who receive money receive it from children (from about 4 children on average). The trans fers in the Philippines and Thailand are slightly more likely to come from a daughter than a son. In Taiwan the elderly are more likely to receive financial support from sons (the Singapore data do not allow for separate determina tions by gender). Although children are the providers of money for a simi larly high fraction of older persons in Taiwan, fewer children tend to be in volved in the support networks (2 children on average). Similar patterns are observed for material assistance, except that both Singapore and Taiwan have far fewer children involved in providing such support compared to the Phil ippines and Thailand. Also, more daughters than sons tend to be involved in providing material assistance in Taiwan (a reversal of the pattern for finan cial support). The involvement of other family members and friends in the support networks of older persons is relatively small in Taiwan and Singapore. Eight percent or less of the elderly in these countries receive money from people other than children and 18 percent or less rely on others for material goods. In Thailand, 12 percent identify people other than children for money trans fers and 41 percent receive material goods from others, mainly grandchildren (29 percent versus 9 percent who receive from siblings and 3 percent who receive from other kin or friends). The Philippines stands out among the four countries for the breadth of the support network. More than one-third of older Filipinos receiving money identify people other than their children (20 per cent from grandchildren, 16 percent from siblings, and 19 percent from other family members and friends), and more than half receiving material assis tance obtain some from people other than their children. Thus, if one were to focus only on parent-child transfers, all four countries appear similar in the importance of children as providers. More children on average are involved in providing money support in the Philippines, Thailand, and Singapore than in Taiwan, however, and sons tend to be the primary providers of monetary support in Taiwan, whereas daughters tend to be primary providers in the Philippines and Thailand. More children on average are also involved in providing material support in the Philippines and Thailand than in Taiwan and Singapore. The gender patterns in the Philippines and Thailand still hold for material assistance, and daughters become more involved on average than sons in Taiwan. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 3 NO rt e50ito"£o1tdeocCO"S•|H^T3JC/lUC.11Uata00^c*s*(2•gtd.53.1o;=o.rsrs00Vi2iri00NOi00NO^0000ONC^00Perrtnt rertivin- supportNO1^!ON^£ootstsmft00 ONSNOwox-^'ONOf those receiving, pe_ent who rertive support from: 'Children) I1 iOO ~*^ ^iON 0Tt ONON CNt- 001 1— C~oo •**00 •*o oNO —r- oo•* TtVi Or- ooSon(s) Da.-hter(s)rsNO^£;t-§qVIrs06mC4NO«0000mNon-shildrent*-moq•*ONOOtsVJts*eVlets^^r*-ViONGrandchildrenCN—NOr4*rOtev] ts^;mOOrsr-Vi*~*|2 ooVi—f}1IOrsm£-•q-Ot--rsI"-00eOther relatives^*friends^*othe— CS ONr- -*t rsTt CN CNo ts ooTt CN —•TJf rs tsm V) NOVI CN CNoo r- oNO m mNO m m— CN ONt- • * CNTt ts tso rs ooTt CN —•* rs rsm v, NOwS rs rsoo r- oNO m CONO m mN.mber of shildren availab (mean)# children in family # mens in family # dau-hters in familyoios"~r- •*mts•*00fiVItsNOmNONm'S cOf those rertivin-, number providers (mean)# children providin- s.ppoi II 'NO Oe ~NO 00g 2ts rsi ir- •*r- r-— oPS— —m oo— —i# mens providin- support # da.-hters providin- supf I - &^ v O' a ri 8 1 I I 1 U M M £ 8- g •o « 5! >v •a s >, H •> o .« c "1 1 1 G. J2 Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 204 The Well-Being of the Elderly in Asia Gender differences are also reflected in who the elderly respondent reports as being the most important provider of money or material support. In results not presented here, slightly more Filipinos said a daughter was the most important provider of money and material goods, and slightly more Tai wanese said a daughter was the most important provider of material goods. However, an overwhelmingly proportion (83 percent) of Taiwanese elderly named a son as the most important provider of money support. Although large proportions of all respondents report receiving support from others, it is im portant to ascertain to what degree the money and material goods they receive are adequate for their needs or expectations, and to what degree they rely on these transfers as sources of income and consumption. Chapter 8 on eco nomic well-being addresses these and other related issues. The focus thus far has been on money and material goods. Assistance with activities of daily living (ADLs), such as eating, dressing, or going to the toilet, and instrumental activities of daily living (lADLs), such as shopping, preparing meals, or doing light housework, is also an important form of sup port. This type of assistance is often referred to as a transfer of time or ser vice (i.e., informal care). Those who assist with ADLs and lADLs usually live with or near the older person since, unlike money or material goods, one cannot "send" this type of support. It is difficult to draw interpretations about country and subgroup differences in ADL and IADL support with these data, however, because related questions were asked quite differently among the four surveys. Overall, fewer than one in ten older persons report receiving help with an ADL: 3 percent in Thailand, 4 percent in Singapore, 6 percent in Taiwan, and 8 percent in the Philippines. Levels of help received with lADLs are much higher, ranging from 17 percent in Thailand, 18 percent in Singapore, and 23 percent in the Philippines to 33 percent in Taiwan. As with money and material support, children are reported as the most common or important non- spousal providers of ADL and IADL support in the Philippines, Taiwan, and Singapore (the Thailand survey did not distinguish between children and other household members among those who provided ADL assistance). A separate study analyzing 1989 survey data from Taiwan found that older adults often receive a combination of types of support provided by more than one person (Hermalin, Ofstedal and Chang, 1996: 428). For example, 49 percent of older adults in Taiwan reported receiving both financial and IADL assistance, but only 9 percent of these indicated that a single person provided both types of support. Although the data on ADL and IADL support are limited here, Chap ter 9 contains more discussion of the prevalence of functional limitations among the elderly and how these limitations are related to their sociodemographic characteristics. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Intergenerational Support and Transfers 205 Elderly Who Provide Support Although the literature usually focuses on support given to older adults, an extensive set of questions was also asked in these surveys about the assis tance that older persons provide to others. Results from Table 6.5 indicate that 67 and 55 percent of older adults provide money to others in the Philip pines and Thailand, respectively, and 76 and 47 percent of older adults in the Philippines and Thailand provide material goods (such as food and clothing). In Singapore, 28 percent of older persons provide money to others, and in Taiwan, about 10 percent provide money to others (and only 5 percent pro vide material goods). Across all four countries, providing support to others is more common among those at younger ages, among men, the married, the highly educated and older adults who live with children (especially unmar ried children). These subgroup differences reveal a sort of mirror image to those presented earlier for receiving support, and the effects are more consis tent across countries. Women and those at older ages and lower education levels are more likely to be on the receiving end of support and men and those at younger ages and higher education levels are more likely to be on the pro viding end of support. The same type of pattern is not evident, though, by type of living arrangements: those living with children are more likely to be engaged in both receiving and providing support. The relatively strong link between older parents and their children with respect to receiving support is weaker when it comes to older persons provid ing support. Table 6.6 shows that of older persons providing money, less than two-thirds give to children (65,61, and 52 percent, respectively, in the Philip pines, Thailand, and Taiwan) compared to the overwhelming majority of older persons who receive money or material goods from children (Table 6.4). Among the elderly who provide money to children, fewer children on aver age receive money from their older parents (e.g., 1.99 children in the Philip pines) than the average number of children who provide money to their older parents (e.g., 3.92 children in the Philippines, Table 6.4). A higher number of children on average receive money or material goods from older parents in the Philippines and Thailand than in Taiwan, and more children on average receive material support from their parents in the Philippines and Thailand than money support. In short, the network of children involved in family sup port is larger for material goods transfers than money transfers, and network size is smaller with respect to receiving support from elderly parents than providing support to elderly parents. Similar gender patterns exist as with receiving support: sons and daughters are about equally likely to get money from parents in the Philippines and Thailand, but sons are more likely than daughters to get money from parents in Taiwan. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 J2 •e OJ !i 1 = (v <^ ^ r5 § l> t)| C < 1 rf i? 0 5 1 ™- a! •-= 2 -5 b i o S o S ^ £ ga ai -o a. oe in v0 _a; J3 aa1K"i 5J_#H•a•c13<S ^•o c 11Iss(XPhilippims1aCtocN^i Joc ^™*O•o.e? -a£|i1£Characteristici i i: : iGro oi Qin m 2-r^r^^S8 S N00inS S ^00^^ 00 g co d oCOdr-gC'l CO ^e^_ vq~~ . o300 ^ CO ~*«it~ m ^^1I1vrt ^<i i i0°°. ~ 3^* ^ or-oPin oo SOS ^ §PP«tinin ts ^. m t s o;NO*~«v0 CO 8— v0vdgasl0 Sr~ o SFTS.SCN00u"rtGenderMale Female Chissq.are (ps. , ,i i :«^ P §in m 2-OrcgC ) OOri00m ts § ^ \O O00 * ts 00C>i •* -^00«•* o 8« so N—'^H 00JUtsg^5 ^ **^S~«iP s j:ai?2Marital statusMarried Not married Chissq.are (psi i i• i i©ov r^ voCO Tt ^,ooci^^gOO OvCNTt1 1 i1CO •* 8l ^^i^;CNr7C>^^ mS5S Sjo1?1RsidenceRural Urban Chissquare (ps Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 1 ue1 vb Educationq qqNOtsq^00CONO00moacaNo fmmal°) r-iCot-qfl00^Svb*0usationi Primary ed5m inCO••*0inin1^r**edusation1 SecondaryI—g00inCl|°v.ft|**)m~ ^|OvviM|^r JO¥1 6in oo — * ts ^- m wS r^ro t^ ON NOTt p oq ooO NO ^-* CNr^ r- oo r^O m (: — vi i^ ^^ O ONo ov Tf r**•^ v~i vO mCTv fel O **1vO vO I*- NO= -ac 1I § 2-6E. a 2 -SM 3 0 «jHill•? < * * o3r?NOin(n PmqI^0«^|te*eCNS1C4S|00mTt^^,^^rs.u.1Ii 1U5-dt*-CNonp00,7.Sip.3_^-oomO1*,t;^-| Sample size Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 •o tQ S. .5. .S D£ ™- T3 oo 1rt9ca.-;*O .31e e2t5 rt^li.Ss161eaC•£"rtoce5 e^"" c9—sucg11If^— tp^5-^r-Tt00tsNOON00— ti0tsNOPe_ent providin- support: ! 'O^r- vO o o t*- oo— f c'c 0^ *O O *— * O^ Vi-. oo r-r^, r,. „ tsqoots-^r~ vo in i0 m *- •^ .^-" •* *n .^ ^^ ~— oo^ v-i in oo t^ CN feiIII Millrotsv-^ i-^^ooooOvCM OO NO O CO -H ^^ •*i0 CO CN v^ Tt— ^ f^ ^^ r^oTft^rri"•fSo wS O CN ^e ••NO <* TTON Tf C3 ON NO *O ^^ Tf355 — — ^c^'o »u s.• ^eC Sgs. £s | S 15 w ^^ c ±3lss ? 11 1J| c^J l| 3.« 2II. Jit vlllife-§5^° |os5i£S— CN ON•*t rj rs^ ts -^Tt CN CN— ir) 3iri rJ i tsoo r- ovo ro mNO to rot-^ Tt CN Tt CNj CNo ts oo-^ ri tsv—l I—I ~1— m NO«S ts oioo r~ oNO ro mNO m ro'=.Number of recipients (mean# shildren in family # mens in family # da.-hters in familyi i iNO O Oo o o00 CN *Om ts —rs —I —•* 00 NONO ro tsts — • — •1 i iON OO —in ro ts0eein iri o•* r- r--^ o dON CN OO-: ^ o"a er c oOf those providm-, n.mber rertivers (mean)# children reseivin- suppo # mens rertivin- s.pport # da.ghters reseiving s.pp § E c a, JJ 2 £1 ^ e II &-8 •o U II nr ./-. ** 8 Se g o § E 2 •s K g- >, I £ | 1 1 | i* JO vs c_ ,on- C **- eO U 1'1 2 1 v *li Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Intergenerational Support and Transfers 209 In general, the types of people who are provided some support by the elderly are more diverse than the types of people from whom the elderly re ceive support. The elderly in the Philippines and Taiwan are more likely to give material support to relatives or friends other than children than to chil dren. The elderly in Thailand and Taiwan are more likely to give financial support to people other than their children than is the proportion of older adults who receive financial support from that source, as shown in Table 6.4. One of the most common forms of support provided by older persons is giving time, especially in caring for grandchildren. Table 6.7 shows that 38 percent of all older adults in the Philippines and 23 percent in Taiwan say they care for grandchildren. In Thailand, 32 percent of the elderly say they provide childcare for the grandchildren who live with them (the survey ques tion was about care for coresident grandchildren less than 10 years of age). In Singapore, the survey questions asked about care for either coresident grand children or non-coresident grandchildren who visited with the adult children. Seventy percent of Singaporeans who lived with grandchildren said they pro vided care for them, and 57 percent who had adult children visit said they cared for grandchildren during those visits. Although survey questions were phrased differently in each country, the evidence suggests that a sizeable mi nority of older adults are providing assistance to their children and grandchil dren via child care. When the performance of household chores such as cook ing, shopping, and doing laundry is counted, an even higher proportion are providing indirect time support to children and grandchildren (see Hermalin, Roan and Perez, 1998). Data from earlier surveys in the mid-1980s in Fiji, Table 6.7. Percent Distribution (of those with grandchildren) Providing Child Care for Grandchildren Cared for any grand childrenCared for any coresident grandchildrenDid not provide childcare for grandchildrenPhilippines38-62Thailand-3268Taiwan23--77Singapore--7030 Note: Thailand and Singapore measures of childcare are for grandchildren who lived with the elderly respondent (with the further restriction in Thailand to grandchildren under 10 years of age). Childcare measures in the Philippines and Taiwan surveys are with respect to all grandchildren, regardless of where they live. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 210 The Well-Being of the Elderly in Asia South Korea, Malaysia, and the Philippines showed that over half (between 54 and 71 percent) of older adults said they helped care for their grandchil dren (Andrews et al., 1986: 67). While not directly comparable to the figures presented above, 1992-1994 survey data from the United States indicate that 60 percent of adults with grandchildren provided some care for their grand children (i.e., caring for grandchildren at least 1 hour in an average week or have had a grandchild stay overnight at least once without his/her parents in the past year) (Fuller-Thomson and Minkler, 2001). Providing support to others has costs, especially because older persons are more likely to be in poorer health and a more constrained economic situation than their working-age counterparts. In the Philippines and Taiwan, questions were posed about the extent and types of burden experienced in providing for others. Of those providing support in the Philippines, 35 percent said they experienced moderate to considerable financial difficulty, 25 percent experienced moderate to considerable emotional difficulty, and 19 percent experienced moderate to considerable time difficulty. In Taiwan, 24 percent of older adults said they had encountered difficulties or worries in their efforts to provide assistance to others. However, among those who experienced difficulties, only 11 percent said they were bothered very much by these difficulties, and most said these difficulties were financial (72 percent) and fewer identified the difficulties as being time related (16 percent) or emotional (32 percent). In general, these findings on the burden of providing support parallel the overall patterns of providing support by the elderly. In the Philippines, far more older adults give support to others, mainly family members and especially adult children, than in Taiwan, where very few give to others. For at least one in three elderly Filipinos who provide assistance, these transfers are made with difficulty. In Taiwan, although about the same proportion of elderly said they had experienced difficulties, far fewer said they were bothered or burdened providing that support. Transfer Amounts Survey data from the Philippines, Thailand, and Taiwan on the amount of money received and provided in the past year reveal interesting country dif ferences (see Table 6.8). Although the majority of older persons in each coun try reported receiving money in the prior year from family or friends, almost half of older Filipinos said they received small amounts of money (less than $40 (US) total) compared to only 19 and 29 percent of older Thais and Tai wanese, respectively. In terms of providing money support, almost all older Filipinos reported giving relatively large amounts of money ($40 (US) or more) to others, followed by older Thais (74 percent) and Taiwanese (54 percent). Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 ^- 35 •" o <-^ di o*~.0*rck§>OJ™c.iH£•otQ•o.1 0l1aEi/1conEQ.UV0"i/1r£IMtQr1->-I EIIo •£— ^§ fr P *t 3 00V2IQ1-02>•gH1J2.3 Hi.a.d.£1>20^1u§'A•oU•o.>i•8>8p£e8T3•>1-o>e5384voONoot-^«OOTt— ^00oorsvOovC**oo08noc2 ra«ss iM f)a -fag &Ma. £qS^oq06vol/r^*^*t**00•HIt— i^Qr^*^-|eg V)ributionthan $4s<a t/i•— -r*3*ja^eJ*ts— .t-Tf•*'-C*}00ov06ovpCOin§sy^N3N-^§«iJ yliPI - « e8 S -9 -a t. ^, -q i« « p-o JS - ^ Z se Mii Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 212 The Well-Being of the Elderly in Asia These data suggest that the money flows between older persons and their support networks are mainly cash out-flows for older Filipinos, cash- intensive flows for older Thais (relatively large money amounts both to and from others that suggest a redistributive pattern), and cash in-flows for older Taiwanese. These descriptive patterns may be limited by a tendency of people to overreport what they do for others and underreport what they receive from others (Roan, Hermalin and Ofstedal, 1992), and by the limited categories available across countries to capture the magnitude of transfers. As noted in the footnote to Table 6.8, more detailed information on the amount of the exchanges was obtained in Taiwan. An analysis of this information, not shown here, indicates that significant proportions of older Taiwanese are involved in the lower and upper ends of the distribution—i.e., the giving and receiving of both small and large amounts of money is quite common. Taiwan also asked their older respondents whether they had ever spent more than $100,000 NT ($3,777 (US)) on the education or living ex penses of their children, and 22 percent reported affirmatively. Multiple Transfers The focus thus far has been on overall levels of family support and on the common recipients and providers of support. Analyses are based on transfers that occur in dyads, between the older adult and another person, and the data are cumulated to describe general transfer patterns that involve the elderly. A different focus that sheds some light on the family network is to examine transfer activity that involves more than two people. For example, earlier figures in Table 6.2 showed sharp country differences in the degree to which older persons were involved in two-way transfers of support (both giving and receiving), ranging from the majority of older Filipinos engaged in this pattern to a small minority of Taiwanese doing so. A prior study of parent-child transfers in the Philippines and Taiwan found that most of these two-way transfers are parts of multiple transfers; that is, transfers (including all types of currencies) involving the elderly parent and two or more children (Agree, Biddlecom, Chang and Perez, 1998). Among the elderly who are engaged in multiple transfers—providing and receiving any type of resource with more than one person (and not restricted to parent-child transfers)—those in Taiwan transfer mostly between vertical generations, with almost all multiple transfers following the pattern of adult children giving to the older parent who in turn gives to grandchildren (Agree, Biddlecom and Valente, 1999). For example, an adult child may give her elderly parent some material goods (food or clothing) and the elderly parent gives a grandchild some money (perhaps for educational purposes). In contrast, multiple transfers in the Philippines follow Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Intergenerational Support and Transfers 213 a more broadly distributed exchange pattern, with greater involvement of "lateral" relatives, most importantly siblings. The most common pathway in the Philippines for multiple transfers is from adult children to their parent and then to the parents' siblings (34 percent of all multiple transfers) and secondarily from adult children through their parents to grandchildren (27 percent of all multiple transfers). Multivariate Analysis of Transfer Patterns Tables 6.9 and 6.10 show multinomial logistic models of money and material transfer patterns for each country. The outcome variable has three categories: giving and receiving support, only receiving support, and not receiving sup port at all. The latter group includes older adults who are not involved in any transfers at all and those who are only giving support, a small percentage in each country (see Table 6.2). The reference category is only receiving sup port, a common pattern in all four countries (see Table 6.2). We estimate two sets of effects of the independent variables. The first set of effects is the likelihood of making two-way transfers (giving and re ceiving) compared to receiving only (the first column for each country), and the second set of effects is the likelihood of not receiving support compared to receiving only (the second column). In each case, the estimates are net of the other outcome category. For ease of interpretation, we present odds ratios rather than regres sion coefficients. Odds ratios are the exponentiated values of the regression coefficients. In the case of a categorical variable, the odds ratio indicates the proportional change in the odds of making one type of transfer (e.g., both giving and receiving money) versus only receiving money for someone in the specified category relative to someone in the reference category. For an inter val variable, the exponentiated value is the proportional change in the odds ratio for a unit increase in the independent variable. With respect to money transfers (see Table 6.9), the few significant and consistent effects across all countries are that older adults with a secondary level of education or higher and those who are employed are much more likely to be making two-way transfers of money versus receiving money only (net of not receiving). For example, older adults with secondary schooling or higher are about 1.5 times more likely in all four countries to both give and receive monetary support than they are to receive it only (ranging from 1.4 times more likely in Taiwan and the Philippines, to 1.8 times more likely in Singapore). Older adults who are currently working are 1.3 times more likely in the Philippines to both give and receive resources than receive only, and those living in Singapore are 3.1 times more likely to be making such two- Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 rt tJ ti rt LL. •o rt Q. rt • i -!— tQ u c L- H is ^- *i c o o •o •o ai ^y^MHo u 200.£ cut/3 T31ai(A0ea*aH ft,•oi05j^W*i*_iO^2 ^*•='3H^ ^^T3105MMi/i 0c.5..£••= a.-— ^^MH eOsa;sovariates***dq^H3i-i8-^***ONd**ONd**d*ONdec33Gender/marital si i•ii •ii ii ii i•i•iiMarried male***t^-d#od3—,s^^***ood*#oodoo00doooodMarried female2^*1N— .iONde^^*ood**PdNO ONd^q^'- i! Unmarried mal<*ON OOd8*—dCOONd$ONd**ood5*oo••0doONdCOoodc«iUnmarried fem;ii* *3§d3—ONOS!dP—00ood'e«Rural (versus url EducationiiiiiiIi Ii iiic.2ca•ou*^*r^^^*NO ts—q*~***2-HONdP— ,r>-,ONd1* ** »* *CN ON NOq ts -^—I -^ CN* * * * * ** * *r- OO TtT P ^-H m -H*r^ ON oo<-; P P— — — ,* ** 1-4 ^ d** ** * »r— v-i O^ — i-H* ** *OO Vi CNin NO O^™, ^H ^H* *. »^E *^) P-; -H — ;* *in ON NO^* CN ON-; — du •**41secondary or hi Employed (versu selfsrated healthGood***• v1d: NOI ^H^^' NOd: 2-H&*i —-^i 2^^.j;i 00di ^: ood•- 2 Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 8 O; vO i i1 ti its id3 i^*** *d***r^ 1d2 idts iq i— HLiving arrangementAlone/spo.se only w/Married shildi •ii2#*^^=sdSidSd_HO—w/Unmarried childii IiI^3ds4Sd8— *OOd6i,i#**r- esde—***00do-^**^dHas 1+ living childq^^** *1f1dovOvd***dS***00 00d3-HNOONdNumber of living sons3—*** iridOOd***OOOS***00d*&-H_^e•—Number of living daughtersNumber of others in* **o—** *IT!OOd*q&dIpq^^8^^00^^dnetwork1^o— H£of^r.•Oets•*"roTts•-o•.^Ovov-HModel chissquare*outil1c•?.Ci0110600c:«Q.MCiBtse2 oBOCUvi•gaoSiJUa Dependent variable has whether providin- momy (itj* * 3f §a- S- c. |>|tj^ •"* j«•^ Go 0-1 a gv °J M *||gei i s9 CX Q^ C/l ^eS* J5 e*5B •••fl E *oII1Jui73 13 ^ 9KB||||Js (3 ^ ooc M M So ^ •— ^o2 eo. es .f?|;||II l|i- e -o .S2 •« i eSB M C j£C W1 etjo 5 'c* .U£ « , u gJ!§ *~eO r-Jt^ e3 oo .SS e5 in PS ec .a EB 7^ -^.2 M.= o•^ ** o e5§ 1 3 gNumber of others in mt (Philippins); the sum of sj s.m of household member; (Taiwan); and the s.m of h§£•-*p<.s_ **p<.0m e **« # parameter is not stimat Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 216 The Well-Being of the Elderly in Asia way transfers of money (Thailand and Taiwan fall in the middle). The elderly with these characteristics probably have greater access to cash income and other resources and are in a position to give money to kin. However, this is not simply a matter of being able to provide money as the positive effect is for both giving and receiving monetary support versus receiving only. Higher education and current employment also make one more likely not to receive money at all compared to receiving only. Other notable effects are the negative association between age and the likelihood of making two-way money transfers in the Philippines and Thai land; and the negative relationship between age and not receiving money in the Philippines, Thailand, and Singapore. The effect of age appears to be small, but the unit change is in single years. Being female and unmarried (versus a married male) makes one much more likely to be on the receiving end only of money transfers—the exception is the Philippines, where gender and marital status differences among the elderly do not matter much at all for money transfers. Significant urban-rural differences disappear once other effects are controlled for in the model (the exception being Taiwan). Interestingly, living arrangement effects on transfer behavior are inconsistent and less strong across countries, suggesting that coresidence with a child is not a singularly potent explanation for the monetary exchanges that older adults report. It is worth noting though that in Taiwan coresidence with an unmarried child increases the odds that an older parent engages in giving and receiving money transfers versus receiving only. The last set of variables in Table 6.9 represents availability of people with whom to exchange resources. When the older person has one or more living children (an indicator of having any children available), she or he has about half the odds in the Philippines and Thailand of simultaneously providing and receiving financial support versus receiving only. In Taiwan, older adults with any living children have .27 times the odds of providing and receiving money compared to receiving only. There is little effect of the availability of children on the likelihood of not receiving financial support versus receiving only in any of the countries. Of course, the availability of children may affect transfer behavior depending on the gender of the children available. The next two variables, number of living sons and number of living daughters, are included in the model to identify gender-specific effects of intergenerational exchange. Interestingly, the more sons or daughters an older parent has, the less likely they are in Thailand, Taiwan and Singapore to both provide and receive money compared to receiving only, and the magnitude of the effects is similar by gender of the child. In the Philippines there is no statistically significant effect of the number of sons or daughters on two-way transfers of money. In sum, Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Intergenerational Support and Transfers 217 the availability of children, regardless of gender, makes an older adult more likely to receive money support only versus being more actively involved in financial exchanges. In contrast, the availability of children, measured as either having any living children at all or as the number of sons or daughters an older person has, makes little difference in the likelihood that an older adult does not re ceive money support (versus receiving only). The exception is the Philip pines, where the more daughters one has, the higher the odds of not receiving money transfers at all compared to receiving only. It could be that given all the other characteristics controlled for in the model, this effect represents cases where the older parent is providing money only (a small category that is combined with no money transfers at all into "not receiving"). However, the evidence across all four countries suggests that having fewer children avail able (sons or daughters) does not make one less likely to receive monetary support. The last variable included in the model is the number of other people in the older person's network. The number of others in the network is defined differently for each country given data limitations (see the note at the bottom of Table 6.9), but in all four countries the variable is closely identified with household size (minus any coresident children). A small but interesting result is that the larger the number of other people (apart from children) in the elderly person's network, the more likely he or she is not to receive money transfers versus only receiving money, a finding that holds for Thailand, Taiwan, and Singapore. We would expect to see the opposite effect of network size on transfers given that a larger number of people with whom to exchange resources increases the opportunities for such transfers. Since the size of network as defined in each country will be greatly influenced by the number of grandchildren, this small contrary effect may represent in part cases where the older person is providing money to grandchildren (which as noted above is combined with those "not receiving") and households where the large number of younger children precludes cash transfers by the children's parents to their own older parents. Table 6.10 shows results from the multinomial logistic model of mate rial transfers (Singapore is omitted because questions about providing mate rial goods to others were not asked). In general, the effects of the sociodemo- graphic variables on transfers of material goods do not differ significantly from those found for money transfers. Higher education and employment have the same positive effects on giving and receiving material goods, except that the effects are not as consistent across countries as they are for money trans fers. Age continues to have a negative influence on the likelihood of making two-way material transfers or not receiving material support at all, relative to Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Table 6.10. Odds-Ratios for the Effects of Soc iodemographic, Health, and Family Structure Factors on Directional Flow of Material Transfers" CovariatesAgeGender/marital statusMarried maleMarried femaleUnmarried maleUnmarried femaleRural (versus urban)EducationNo formal educationPrimarySecondary or higherEmployed (versus not)Self-rated healthGoodFairPoorLiving arrangementAlone/Spouse onlyw/Married childw/Unmarried childOtherHas 1+ living childNumber of living sonsNumber of livingdaughtersNumber of others innetworkModel chi-squarePhilippines R and P Not R0.95* 0.94***_0.56** 0.960.66 0.770.57*** 0.77*1.00 1.10-.1.18 1.181.73** 1.64***1.96*** 1.72***0.80 1.03..0.76 0.76**1.23 0.80-1.38 0.981.37 1.150.71 1.65*1.02 1.031.02 1.10*0.98 0.991655.81Thailand R and P Not R0.95*** 0.%***_1.12 1.23***0.95 0.920.98 0.13*0.85* 1.01-0.90 1.19***1.14 1.60***1.11 1.35***1.15* 1.02_0.86 1.22"*0.58*** 0.87**_0.91 0.940.45*** 1.010.46*** 1.30*0.94 1.05*0.75*** 1.040.94"* 1.07***7530.00Taiwan R and P Not R0.96*** 0.97_ _0.97 1.130.96 #0.89 0.641.03 1.12„1.25*** 0.831.37*** 1.631.08 0.620.95 0.94-0.98 #0.90 0.53..1.06 0.700.88 0.670.89 #0.94 0.790.97 0.990.99 1.06*2461.76 Dependent variable has thrte categories: rtceiving only (omitted), both rtceiving and providing (R and P) and not receiving, rtgardless of whether providing money (Not R). Number of others in network is measurtd as the sum of household members (not including childrtn), grandchildren and living siblings (Philippines); the sum of spouse, grandchildrtn, siblings, nieces, nephews, and other rtlatives whom the rtspondent sees rtgularly (Thailand); and the sum of household members (not including childrtn), grandchildrtn, siblings, and other rtlatives and neighbors with whom R has rtgular contact (Taiwan). *p<.05 **p<.01 ***p<.001; # pa_meter is not estimable Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Intergenerational Support and Transfers 219 only receiving. The interactive effect of gender and marital status is less con sistent across countries for the exchange of material goods compared to money. Filipinas. whether married or unmarried, are less likely than their married male counterparts to be involved in two-way transfers of material goods rela tive to receiving only. This was not the case for money transfers in the Philip pines. In contrast, while gender and marital status affected whether older per sons in Thailand and Taiwan engaged in two-way transfers of money, these characteristics have no effect on two-way transfers of material goods. The rather weak effects of living arrangements apply also to the exchange of ma terial goods. The exception continues to be Thailand, where older adults who live alone or with a spouse only are much more likely to be on the receiv ing end of material support than either giving and receiving or not receiving at all (the same occurred for the exchange of monetary support). The availability of any living children and the number of living daugh ters continue to be negatively associated with the propensity to engage in two-way transfers of resources in Thailand. No such effects are evident for either the Philippines or Taiwan. For the Philippines and Thailand, having any living children and having more sons (Thailand) or daughters (Philip pines) is associated with a higher likelihood of not receiving material support at all compared to receiving only. Again, this effect may indicate a pattern of the older adult being more likely to provide material support only (rather than not being engaged in any material transfers at all). As we did with money transfers, we encounter the small but interesting finding that in Thailand and Taiwan a larger "other" kin network makes an elderly person more likely not to receive material goods at all than to receive them only, and we interpret this finding in a similar manner. Conclusions With countries and families facing changing demographic and social condi tions, the nature of intergenerational support will certainly be transformed. The support network for today's elderly in Asia is broad and centered on the home, as indicated by both the large numbers of living children and the high levels of coresidence. Yet all but a few Asian countries are expected to have fertility rates below replacement level by 2015 (Lee and Mason, 2000), and this is already the case in Singapore, Taiwan and Thailand (see Table 2.4). Fertility below replacement ensures that future cohorts of elderly parents will have smaller kin networks upon which to depend. Smaller and perhaps more dispersed support networks may lead to a decline in intensive transfers of space or time, such as coresidence or help with activities of daily living, but Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 220 The Well-Being of the Elderly in Asia transfers of other resources, such as money (which has no geographic limita tions) or social contact (via visits or telephone calls) may persist at a high level. The results from the multivariate analyses of transfer patterns certainly indicate that among the older generation today, childlessness or having a small number of sons or daughters is not associated with a greater likelihood of not receiving money or material support at all. At the same time, declining fertility and mortality rates mean there will be fewer children for older parents to support but perhaps a higher likelihood that their own elderly parents will still be alive and need support. The empiri cal evidence presented earlier suggests that already older adults provide re sources to a broader array of kin (e.g., parents, grandchildren, siblings) than the network from which they themselves receive support (i.e., primarily adult children). For example, 15 percent of older Thais providing financial support to others are giving money to their own elderly parents. With rises in life expectancy, this situation is likely to become more common. The active role that the elderly play in providing support is particularly important in Asia as industry shifts increasingly from family farming to wage labor, and as the labor force participation of women rises. In Japan, for ex ample, Morgan and Hirosima (1983) found that older parents were more likely to provide childcare and housework if their daughters or daughters-in-law were working outside the home. Current levels of childcare provided by grand parents to coresident and non-coresident grandchildren are relatively high in the four countries examined in this chapter. These levels may rise higher in conjunction with increased demand (as female labor force participation rates continue to rise) and an increased supply of elderly caregivers (as life expect ancy increases and if age at retirement decreases). Rising levels of education are also likely to alter the qualitative and behavioral aspects of family transfers among older persons. Today's older generation is relatively poorly educated and the gender differences in educa tional attainment are quite large (Christenson and Hermalin, 1991). In con trast, those who turn age 60 in the year 2020 will predominantly be literate and the gender gap in educational attainment will have narrowed consider ably. Analyses in this chapter indicated that secondary and higher levels of education are positively associated with both giving and receiving money or material goods. In fact, education had one of the few consistently strong ef fects on transfers across all four countries. These findings suggest that the family support systems of future elderly will be much more varied in terms of patterns of support flows than the current ones tend to be. Rather than an older generation being primarily on the receiving end in the family support system, future cohorts will likely be both independent and active members of family support systems. The current attitudes of middle-aged women in Asia Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Intergenerational Support and Transfers 221 about support when they are old indicate that this independent and active role is already in the minds of many. Limitations of the analyses in this chapter—and challenges for future research—are that the magnitude and frequency of intergenerational trans fers and the life course patterns in making and receiving transfers are still relatively unknown. Resource flows analyzed in this chapter were with re spect to the existence of transfers in the prior six months to one year from the survey interview. Little is known about the frequency and magnitude of the transfers occurring within this time frame. Although the simple dichotomies used in the analyses (e.g., money was or was not received in the past year from children or the amount received was "small" or "large") provide a sense of the "bottom-line" regarding family transfers, the nature and nuances of these transfers remain unclear. Future research needs to address much more detail about these transfers, as suggested by the discussion in Chapter 4. The analyses presented provide a snapshot of how intergenerational transfers vary by individual and family characteristics and across four culturally and economically distinct Asian countries. The longitudinal picture of how and why transfers change across the life course is the next needed step in research. Longitudinal data allow for more confident inferences about the causal connections between individual and family characteristics and transfer outcomes, as well as the effects of family support on dimensions of elderly well-being, such as health status. Longitudinal data also enable one to observe transitions over the life course in family support activity. For example, Chapter 8 (see Table 8.13) describes a large switch in the major income sources of the elderly in Taiwan over time. Of the elderly receiving primary financial support from children in 1989, only 54 percent continued to receive their primary support from children in 1996 and nearly one-third had shifted to a pension or retirement fund as their major source of income. This example of change over the life course also raises the issue of tradeoffs between private and public transfers. Longitudinal data allow for both time sequencing of transfer behavior and testing of the degree to which (and direction that) private family transfers respond to public transfers. Dur ing the same period in Taiwan described above, two subsidy programs were initiated to provide income to older people outside current pension and social welfare programs. A recent study using the Taiwan panel data found that the farmers' pension program, a monthly pension given to farmers age 65 years and older who were not already receiving a pension, had a significant and positive effect on the likelihood that older persons switched from private trans fers (mainly income from children) as their main source of income in 1993 to pension income in 1996 (Biddlecom et al., 2001). In contrast, the other pen sion program initiated in the mid-1990s in selected counties and cities had Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 222 The Well-Being of the Elderly in Asia little effect on the move away from private transfer income as a main source of support. These issues are important to examine given the current discus sions in many Asian countries of new social security program development, the repercussions that public support programs may have on existing patterns of family support, and the resulting effectiveness of those public support pro grams to improve elderly economic well-being (see Chapter 3 and Mason, Lee and Russo, 2001). The future of family support is not likely to be the simple "decline" or "erosion" story often alluded to, wherein the dependent elderly become the destitute elderly. The current snapshot of intergenerational transfers shows that older adults in Asia are also family support providers—in the Philippines and Thailand this was a substantial share of transfer activity—belying their characterization as dependents. The small declines over time in confidence with children, documented in Chapter 5, are paired with substantial levels of social contact (initiated almost equally between parents and children) and monetary and material goods transfers. The anticipated increases in the num ber and percentage of elderly in Asia are occurring along with socioeconomic and allitudinal changes that mean future cohorts of elderly will be belter edu cated, have more experience working for pay outside the home, and have fewer expectations of depending on children for support than the current co hort. Finally, the paucity of evidence in Asia about how private transfers re spond to public support leaves room for much speculation about how the implementation and design of new and amended social security systems and health care plans will reshape family transfer networks. In short, intergenerational transfer systems are likely to respond to the continued de mographic and socioeconomic changes taking place in Asia but there is noth ing to suggest that the emerging patterns will not be responsive to the family needs that accompany these changes. REFERENCES Agree, Emily M., Ann E. Biddlecom, Ming-Cheng Chang, and Aurora E. Perez. 1998. "Generalized Exchange and Intergenerational Transfers in Taiwanese and Filipino Families." Hopkins Population Center Working Paper No. 98- 06, Baltimore, MD: Johns Hopkins University. Agree, Emily M., Ann E. Biddlecom, and Thomas W. Valente. 1999. "Multi-Gen erational Exchanges in Taiwan and the Philippines: A Social Network Ap proach." Hopkins Population Center Working Paper No. 99-06, Baltimore, MD: Johns Hopkins University. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Intergenerational Support and Transfers 223 Andrews, Gary R., Adrian J. Esterman, Annette J. Braunack-Mayer, and Cam R. Rungie. 1986. Aging in the Western Pacific. Manila, Philippines: World Health Organization. Anh, Truong Si, Bui The Cuong, Daniel Goodkind, and John Knodel. 1997. "Liv ing Arrangements, Patrilineality and Sources of Support among Elderly Viet namese." Asia-Pacific Population Journal 12(4):69-88. Biddlecom, Ann E., and Lita J. Domingo. 1996. "AgingTrends: The Philippines." Journal ofCross-Cultural Gerontology 11:109-114. Biddlecom, Ann E., Albert I. Hermalin, Mary Beth Ofstedal, Ming-Cheng Chang, and Yi-Li Chuang. 2001. "Tradeoffs between Public and Private Economic Support of the Elderly: Results from a Natural Experiment in Taiwan." Com parative Study of the Elderly in Asia: Research Reports No. 01-58, Ann Ar bor, MI: University of Michigan. Chan, Angelique. 1997. "An Overview of the Living Arrangements and Social Support Exchanges of Older Singaporeans." Asia-Pacific Population Jour nal 12(4):35-50. . 1999. "The Role of Formal versus Informal Support of the Elderly in Singapore: Is There Substitution?" Southeast Asian Journal of Social Sc/- ence27(2):87-110. Chang, Ming-Cheng, and Mary Beth Ofstedal. 1991. "Changing Attitudes toward Old-Age Support in Taiwan: 1973-1985." Comparative Study of the Elderly in Asia Research ReportsNo. 91-8, Ann Arbor, MI: University of Michigan. Chen, Ai Ju, and Gavin Jones. 1989. Ageing in Asean: Its Socio-Economic Conse quences. Singapore: Institute of Southeast Asian Studies. Chen, Xuan, and Merril Silverstein. 2000. "Intergenerational Social Support and the Psychological Weil-Being of Older Parents in China." Research on Aging 22(l):43-65. Christenson, Bruce A., and Albert I. Hermalin. 1991. "Comparative Analysis of the Changing Educational Composition of the Elderly Population in Five Asian Countries: A Preliminary Report." Comparative Study of the Elderly in Asia Research Reports, No. 91-11, Ann Arbor, MI: University of Michigan. Cicirelli, Victor G. 1994. "Sibling Relationships in Cross-Cultural Perspective." Journal of Marriage and the Family 56:7-20. Cornman, Jennifer C., Noreen Goldman, Maxine Weinstein, and Hui-Sheng Lin. 2001. "Perceptions of Elderly Taiwanese about the Availability of Social Sup port." Office of Population Research Working Paper Series, No. 2001-05, Princeton, NJ: Princeton University. Domingo, Lita J., and M. M. B. Asis. 1995. "Living Arrangements and the Flow of Support between Generations in the Philippines." Journal of Cross-Cul tural Gerontology 10:21-51. Frankenberg, Elizabeth, Victoria Beard, and Muda Saputra. 1999. "The Kindred Spirit: The Ties that Bind Indonesian Children and Their Parents." Southeast Asian Journal of Social Science 27(2):65-86. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 224 The Well-Being of the Elderly in Asia Frankenberg, Elizabeth, Lee Lillard, and Robert J. Willis. 2000. "Money for Noth ing?: Altruism, Exchange, and Old Age Security in Southeast Asia." Unpub lished working paper. Santa Monica, CA: RAND. Fuller-Thomson, Esme, and Meredith Minkler. 2001. "American Grandparents Providing Extensive Child Care to Their Grandchildren: Prevalence and Pro file." The Gerontologist 4\(2):201-209. Henretta, John C., Martha S. Hill, Wei Li, Beth J. Soldo, and Douglas A. Wolf. 1997. "Selection of Children to Provide Care: The Effect of Earlier Parental Transfers." Journals of Gerontology 52B: 110-119. Hermalin, Albert I. 1995. "Aging in Asia: Setting the Research Foundation." Asia- Pacific Population Research Reports,No A, Honolulu, HI: East-West Center Program on Population. . 1998. "Setting the Agenda for Research in Aging in Developing Coun tries." Comparative Study of the Elderly in Asia Research Reports, No. 98- 51, Ann Arbor: University of Michigan. . 1999. "Challenges to Comparative Research on Intergenerational Trans fers." Southeast Asian Journal of Social Science 27(2):9-20. Hermalin, Albert I., Mary Beth Ofstedal, and Ming-Cheng Chang. 1992. "A Com parison of Alternative Approaches to Measuring Needs and Levels of Sup port Received among the Elderly in Taiwan." In P. Krishnan, et al., eds., Read ings in Population Research: A Volume in Honor of Ronald Free dman, 295- 310. Delhi, India: B.K. Publishing. . 1996. "Types of Supports for the Aged and Their Providers in Tai wan." In T. Hareven, ed., Aging and Generational Relations over the Life Course, 400-437. New York: Walter de Gruyter. Hermalin, Albert I., Mary Beth Ofstedal, and Shiauping Shih. Forthcoming. "Pat terns of Intergenerational Support in Urban China and Urban Taiwan." In M.K. Whyte, ed., China's Revolutions and Intergenerational Relations. Ann Arbor, MI: University of Michigan. Hermalin, Albert I., Carol Roan, and Aurora E. Perez. 1998. "The Emerging Role Grandparents in Asia." Comparative Study of the Elderly in Asia Research Reports, No. 98-52, Ann Arbor, MI: University of Michigan. Holden, Constance. 1996. "New Populations of Old Add to Poor Nations' Bur dens." Science 273(5):46-48. Kato, Elizabeth Uphoff. 2000. "Ageing in Cambodia: Tradition, Change and Chal lenges." In D. R. Phillips, ed., Ageing in the Asia-Pacific Region: Issues, Policies and Future Trends, 360-374. New York, NY: Routledge. Kendig, Hal. 1989. "Social Change and Family Dependency in Old Age: Percep tions of Japanese Women in Middle Age." NUPRI Research Paper Series No. 54. Tokyo, Japan: Ninon University. Knodel, John, Napaporn Chayovan, and Siriwan Siriboon. 1992. "The Familial Support System of Thai Elderly: An Overview." Comparative Study of the Elderly in Asia Research Reports, No. 92-13, Ann Arbor: Univ. of Michigan. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Intergenerational Support and Transfers 225 Kiiodel, John, Jed Friedman, Truong Si Anh, and Bui The Cuong. 2000. "Inter- generational Exchanges in Vietnam: Family Size, Sex Composition, and the Location of Children." Population Studies 54(1):89-104. Kono, Shigemi. 1998. "Demographic Aspects of Population Ageing in Japan." In Aging in Japan, 1998, 7-52. Tokyo, Japan: Japan Aging Research Center. Lee, Sang-Hyop, and Andrew Mason. 2000. "Population Aging Raises Questions for Policymakers." Asia-Pacific Population and Policy Reports, No. 53, Honolulu, HI: East-West Center Program on Population. Lee, Yean-Ju, William L. Parish, and Robert J. Willis. 1994. "Sons, Daughters, and Intergenerational Support in Taiwan." American Journal of Sociology 99:1010-1041. Lillard, Lee A., and Robert J. Willis. 1997. "Motives for Intergenerational Trans fers: Evidence from Malaysia." Demography 34(1): 115-134. Lopez, Maria E. 1991. "The Filipino Family as Home for the Aged." Compara tive Study of the Elderly in Asia Research Reports No. 91 -7, Ann Arbor, MI: University of Michigan. Marsh, Robert M., and Cheng-Kuang Hsu. 1995. "Changes in Norms and Behav ior Concerning Extended Kin in Taipei, Taiwan, 1963-1991." Journal of Comparative Family Studies 26(3):349-369. Martin, Linda G 1990. "Changing Intergenerational Family Relations in East Asia." Annals of the American Academy of Political and Social Science 510:102- Martin, Linda G, and Kevin Kinsella. 1994. "Research on the Demography of Aging in Developing Countries." In L.G Martin and S.H. Preston, eds., De mography of Aging, 356-397. Washington, D.C.: National Academy Press. Mason, Andrew, Sang-Hyop Lee, and Gerard Russo. 2000. Population Momen tum and Population Aging in Asia and Near-East Countries, Honolulu, HI: East-West Center. Mason, Karen O. 1992. "Family Change and Support of the Elderly in Asia: What Do We Know? "Asia-Pacific Population Journal 7:13-32. Morgan, S. Philip, and Kiyosi Hirosima. 1983. "The Persistence of Extended Family Residence in Japan: Anachronism or Alternative Strategy?" American Socio logical Review 48:269-281. National Research Council. 2001. "Intergenerational Transfers." In Preparing for an Aging World: The Case for Cross-National Research, 155-199. Washing ton, D.C.: National Academy Press. Natividad, Josefina N., and Grace T. Cruz. 1997. "Patterns in Living Arrange ments and Familial Support for the Elderly in the Philippines." Asia-Pacific Population Journal 12(4): 17-34. Ofstedal, Mary Beth, JohnKnodel, andNapaporn Chayovan. 1999. "Intergenera tional Support and Gender: A Comparison of Four Asian Countries." South east Asian Journal of Social Science 27(2):21-42. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 226 The Well-Being of the Elderly in Asia Ogawa, Naohiro, and Robert D. Retherford. 1997. "Shifting Costs of Caring for the Elderly Back to Families in Japan: Will It Work?" Population and Devel opment Review 23(1):59-94. Peterson, Jean T. 1990. "Sibling Exchanges and Complementarity in the Philip pine Highlands." Journal of Marriage and the Family 52:441-451. . 1993. "Generalized Extended Family Exchange: A Case from the Phil ippines." Journal of Marriage and the Family 55:570-584. Phillips, David R., ed. 2000. Ageing in the Asia-Pacific Region: Issues, Policies and Future Trends. New York, NY: Routledge. Quisumbing, Agnes R. 1994. "Intergenerational Transfers in Philippine Rice Vil lages: Gender Differences in Traditional Inheritance Customs." Journal of Development Economics 43:167-195. Rahman, Omar. 2000. "The Impact of Co-Resident Spouses and Sons on Elderly Mortality in Rural Bangladesh. " Journal ofBiosocial Science 32(1):89-98. Roan, Carol L., Albert I. Hermalin, and Mary Beth Ofstedal. 1996. "Intergenerational Contact and Support in Taiwan: A Comparison of Elderly Parents' and Children's Reports." Comparative Study of the Elderly in Asia Research Reports, No. 96-36, Ann Arbor, MI: University of Michigan. Rodriguez, Edgard R. 1998. "International Migration and Income Distribution in the Philippines." Economic Development and Cultural Change 46(2):329- Schoeni, Robert F. 1997. "Private Interhousehold Transfers of Money and Time: New Empirical Evidence." Review of Income and Wealth 43(4):423-448. Silverman, Gary. 1995. "Ageing Asia: Honour Thy Father." Far Eastern Eco nomic Review March: 50-52. Soldo, Beth J., and Martha S. Hill. 1993. "Intergenerational Transfers: Economic, Demographic, and Social Perspectives." In G L. Maddock and M. P. Lawton, eds., Annual Review of Gerontology and Geriatrics: Focus on Kinship, Ag ing and Social Change Volume 13, 187-218. New York: Springer Publishing Co. Su, Ya-ping, and Kenneth F. Ferraro. 1997. "Social Relations and Health Assess ments among Older People: Do the Effects of Integration and Social Contri butions Vary Cross-Culturally?" Journal of Gerontology: Social Sciences 52B(1):S27-S36. Thornton, Arland, and Hui Sheng Lin, eds. 1994. Social Change and the Family in Taiwan. Chicago: University of Chicago Press. Walker, Michael E., Stanley Wasserman, and Barry Wellman. 1993. "Statistical Models for Social Support Networks." Sociological Methods and Research 22(1):71-98. Westley, Sidney B., Sang-Hyop Lee, and Andrew Mason. 2000. "Policy Options to Support Asia's Growing Elderly Population." Asia-Pacific Population and Policy Reports 54, Honolulu, HI: East-West Center Program on Population. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Intergenerational Support and Transfers 227 Williams, Linda, and Lita J. Domingo. 1993. "The Social Status of Elderly Women and Men within the Filipino Family." Journal of Marriage and the Family 55:415-426. Williams, Linda, Hui Sheng Lin, and Kalyani Mehta. 1994. "Intergenerational Influence in Singapore and Taiwan: The Role of the Elderly in Family Deci sions." Comparative Study of the Elderly in Asia: Research Reports, No. 94- 28, Ann Arbor, MI: University of Michigan. Wu, Z. Helen, and Laura Rudkin. 2000. "Social Contact, Socioeconomic Status, and the Health Status of Older Malaysians." The Gerontologist 40(2):228- Yoon, Hyunsook, and Heung-Bong Cha. 1999. "Future Issues for Family Care of the Elderly in Korea." Hallym International Journal of Aging l(1):78-86. NOTES FOR TABLES Table 62 Variables were constructed from the questions below to form indicators of the respondent receiving any money from other people, providing any money to other people, receiving any material goods from other people, and providing material goods to other people. All constructed indicators had the following values: o - no and 1 = yes. The questions used to ascertain money and material goods transfers for each country are listed below. When there were several questions that differed only by the person or people referred to, we condensed the questions into one question below and used "[PERSON(S)]" to indicate that the question refers to specific people or types of people (e.g., father/mother or brother/sister). Philippines: Receive money: Outside of small gifts, did [PERSON(S)] give money either regularly or from time to time within the past year? (Y) yes, (N) no. Provide money: Outside of small gifts, did you give money to [PERSON(S)] either regularly or from time to time during the past year? (Y) yes, (N) no. Receive material goods: Outside of small gifts, did [PERSON(S)] give any food, clothing, any personal belonging either regularly or from time to time within the past year? (Y) yes, (N) no. Provide material goods: Outside of small gifts, did you give any food, clothing, or any personal belonging to [PERSON(S)] either regularly or from time to time within the past year? (Y) yes, (N) no. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 228 The Well-Being of the Elderly in Asia Thailand: Receive money: Within this past year did [PERSON(S)] give you a total of 1,000 Baht or more? (1) yes, (2) no. Provide money: Within this past year did you give a total of 1,000 Baht or more to [PERSON(S)]? (1) yes, (2) no. Receive material goods: Within this past year did [PERSON(S)] give you food, clothes or personal belongings? (1) yes, (2) no. This question was asked only about material goods from non-coresident children and relatives. Provide material goods: Within this past year did you give any food, clothing or personal belongings to [PERSON(S)]? (1) yes, (2) no. This question was asked only about material goods provided to non-coresident children and relatives. Taiwan: Receive money: Did any of your children or relatives give you or your spouse money within the past year? (1) yes, (2) no. Provide money: Have you or your spouse spent more than NTS 100,000 on your children's educational expenses or their living expenses? (1) yes, (2) no. Have you or your spouse given money to your children or anyone else for expenses other than educational expenses? (1) yes, (2) no. Receive material goods: Does anyone send you personal belongings such as food, clothes, or other material supports to accommodate the needs of your daily life? [Presents for holidays, birthdays, or socialization are not included] (1) yes, (0) no. Provide material goods: Do you regularly send personal belongings such as food, clothing, and material support to your family, relatives, or others to accommodate their needs for daily life? [Presents for holidays, birthdays, or socialization are not included] (1) yes, (0) no. Singapore: Receive money: Number of children providing cash allowance in the last 6 months (on a regular basis): number. Cash contribution from related persons (grandchildren, brothers/sisters, other relatives): (1) yes, (2) no, (3) not applicable. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Intergenerational Support and Transfers 229 Provide money: In the household, are you involved in any of the following? Provide financial assistance to children (1) yes, (2) no. Receive material goods: Number of children providing accommodation (including food and lodging, payment of rent, utilities, conservancy charges) on a regular basis in the last 6 months: number. Number of children providing rations (e.g., food, household necessities) on a regular basis in the last 6 months: number. Maintenance in kind (accommodation/rations) received from other related persons/non-relatives (grandchildren, brothers/sisters, others) on a regular basis in the last 6 months: (1) yes, (2) no, (3) not applicable. Provide material goods: No questions available in the survey on this type and direction of transfer. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Chapter 7 WORK, RETIREMENT, AND LEISURE Albert 1. Hermalin, Angelique Chan, Ann Biddlecom, and Mary Beth Ofstedal We begin our analysis of the dimensions of well-being with an explora tion of the labor force activities of the older respondents, the factors associated with work and retirement, and the nature and extent of their lei sure. The focus on work marks a logical starting point as it represents for many a major defining characteristic over much of their lives and has strong influence on their assets and income as well as their health. The study of labor force participation at older ages and the process of retirement and subsequent "leisure" activities do not have a long tradition in developing countries. Until relatively recently, most people in developing coun tries worked in the agricultural sector or in other types of family-owned or self-employed enterprises. This, coupled with moderate life expectancies, meant that a high proportion of the population worked until late in life, adjust ing as necessary for diminishing health or functioning. These arrangements have led some observers to call into question the relevance for developing countries of the concepts of work, retirement, and leisure used in more indus trialized settings, as well as the measures typically used to classify activities and people (Cain, 1991; Dharmalingam, 1994; Chen and Jones, 1989, Chap ter 5). Aware of these economic and cultural factors, investigators designed Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 232 The Well-Being of the Elderly in Asia the surveys to capture the variety of work arrangements that might exist in the countries under study. As described below, the detailed questions posed in the surveys permit a reasonable distinction between those working, working without pay and/or in a limited way, and not working. They also assessed the leisure activities of the elderly in a direct manner and elicited a number of attitudinal and preference items about work and leisure, which appear mean ingful. The four countries under study differ considerably in their reliance on agriculture, with the proportion of the overall labor force so engaged varying from zero in Singapore, to 13 percent in Taiwan, 45 percent in the Philip pines, and 64 percent in Thailand as of 1990 (see Table 2.5 in Chapter 2). It should also be noted that, within each country, those employed by the gov ernment or by large-scale private enterprises experience retirement as a for mal process of cessation from employment, with provision for continuing income through specified arrangements involving the individual, the employer, and/or the government. In contrast with the more open-ended work cessa tion ages for farmers and the self-employed, mandatory retirement ages from civil service employment were quite young throughout the region, often stipu lated at age 60 for men and 55 for women (see Chapter 3). Despite the absence of a sharp demarcation between work and retire ment in agricultural communities, some settings have well-specified cultural arrangements for the transfer of duties from elders to children and the provi sion of assistance to parents. For example, Gallin (1966, 215-16) describes how men and women in the Taiwanese village he studied gradually retired from farm work and household tasks as they got older, often after age 50, and observes that old age was generally a period of freedom and leisure. However, continued study of the village revealed that with economic de velopment and more off-farm employment, mothers-in-law were placed in a less advantageous position vis-a-vis their daughters-in-law and sons, and lost much of the support they previously enjoyed (Gallin, 1994). These findings suggest that the nature of the local economy, in addition to cul tural patterns, can influence support arrangements and work patterns at older ages. Nevertheless, the recency of the economic transformation in the region means that a high proportion of the current elderly are now or were previously in agriculture (except for Singapore) or in the self-em ployed or family-owned business sector. A cross-national study of labor force participation rates at older ages by Clark et al. (1999) confirms a strong positive relationship between the percentage of workers in agricul ture and the percentage of older people in the labor force. This distinctive aspect must be kept in mind when analyzing work and retirement pat terns. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Work, Retirement, and Leisure 233 Table 7.1 shows the type of current or last job by age and gender in the four countries. Overall, the proportion of those over 60 whose current or past employment was in agriculture or in a non-agricultural family business was a majority in every country except Singapore, ranging from 55 percent in Taiwan to 76 percent in the Philippines and 91 percent in Thailand. As expected, these proportions are even larger in the rural areas. But even in the urban areas, the proportions in non-agricultural, non-family occupations com prise only a third or less in Thailand and the Philippines, a half in Taiwan, and three quarters in Singapore. At the same time it is worth noting that those aged 60-69 exhibit some what less attachment to agriculture than those 70 and older do, and this is particularly true of women. Recent advances in educational attainment and economic development in these countries mean that future cohorts of elderly will present a very different occupational structure than that shown in Table 7.1. As a result, future retirement ages will be the product of both this sharply changing distribution of sectoral employment and of the changes in retire ment within sectors. It is possible therefore to see a secular decline in the age at retirement overall as more workers enter the private industrial and service sectors, while the age at retirement within sectors advances. Failure to take into account both the compositional shift and the sectoral dynamics can lead to misleading inferences about the reasons for the secular decline and for observed differences cross-culturally. The past labor force patterns and the resulting low saliency of retire ment for many as an expected phase of life also mean that there are relatively few studies in the region of attitudes toward work and leisure. We know little, for example, about preferred ages for retirement, the proportions of elderly who would like to stop work (or return to work), and the range of leisure activities undertaken by those currently working and retired. These dimen sions, described further below, are central in assessing the well-being of the elderly in terms of their activity level—a central component of our conceptual framework. An older person who is employed may prefer to be retired but is constrained to work by economic and family considerations. Conversely, one who is retired may prefer to work but poor health or employer policies or incentives may have dictated otherwise. If well-being is gauged in part by acting in accord with one's preferences, the well-being of older persons with regard to work and leisure cannot be determined from observing their current status alone. This contrasts with measuring well-being in terms of economic or health status, where there is a strong presumption that greater income and assets and better health are desirable outcomes. In a broader context, there may be tension between retirement behav ior and the societal costs and benefits of that behavior as embedded in public Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 -C •ob >^ —. —. O. —. 5 « ^^ w O 4l Q..3i o o C 4l •2 % II </1 ro (5 «- 0l QC C 4l 3£_u auU-_uA-e Rsidemee00c i/5£.t5,—_c£Zljt/5*rt._j=£00t/5^£,—_L:££i>j4• 3£r-dr*-TtCNr-voooinCNs eoovoCNm3r-inONdI/1mCNCNCN NOOOvOCOA-e rt0_9A-ris.lt.rals family farm m b.sinssioom•*r-wNOi tO00fNOO00sivOTti/1vOroCN tNA-ric.lt.ral, nonsfamilyvqCNt^-oos*O C*oro"5r-^NOr-O•*ts^^oo 4ooCN CNroONmNONO^^_>.^2 *c00 CdOvOH Kr-r^t00vOroO CsiriCNooOOr- t^lOinrooCNtNvd r-^•*oo tsr^oot^1s M*uS> Q.00aOi__.ONNOm NONOiooovt*l^^•*•*i-TtCN_NOOOvu1l1O1.C 00a iOA-e _e+ A-ric.lt.rals familyNOd•t-- rn«n00NO 'OvO«sC4citsoo •*«~ r^00mi iNOrsCOvO3mON rofarm m b.sinssi__r^_•* r^oCNi_O_in_vOiOvmootsCfvvO tNA-ric.lt.rals nonsfamily— i t*•;ts r^ CN r-— Oo, — i— i CNr- •*ON •*oo mCN r^im -^CN vO in CNCN r-ON O•** CN— i CNin voON CNm mTt OOTt — '— . ON — . OOCN r-ON vOs e CN CNO voO vo CNw1 COoo r->, £~ a11.C .uDO 00ta ta c c O Oz zI IvoinrOm r-vOiooTfCN~mmI ie8ONv-i^too!• U-=Oo oBD*u. AO Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 •oVecou^35—r^ooo•^^4Om_mroOmTtrs^oooONooooo^— -i^.OinCfttsE t3 £.E- tn eg •«3 o ellCd o^l (s-•e <Di r^• _*tsr- e— r mrsON r-— roin— • tOO ts— •*i r-:I ooO ONvooNO P-rs —NOOO — 'NO vG'ni NOi tsts— i ON-1- f^.ts— ' r^^-* CO^r- ON^t ts— • ts^.cc >v§11*1.5u ?.C c00 0< ZNO it^: ir-rs •*pr. _r^i CNrs tsoo —^-t -^in Ors Ors -^o; ie 'ooON OOoo ONoo ts-i0oo ro— i NOtsNO INO 't*-NO ^5O NOCO CN— r-NO OOCN -^t*- «sm inrs — •u.5u S a »j > >.E.a•c .d oo ooc c o o Z Z.23y•g &3 <i ii i— CNr-- Ots o-f I-^t^NO •*Tt Tt. .* •O •*in —ts om ooON ON-H OOCN — •i ii iON OOoo r-in— ONro mr-— i ONr- ONTt CNO.s gV1•° *2||s .uis- 00I iI irs ri-iON -^-Tt Ttm t*i^^NO f^Cr) r^,ts — ii ,I i— i ON00 00—ON inr^i tsr- tsoo r-m —, ,i ioo r-ON -^oo ints TtON TtOO ON>* Hrs aI.E.Oi) &I)1 1 z zIitsONON—OTt.ivo**^.epriCO,iOO_;•^NOCOt**^b-5c >eDOi11•33 1a.!o ^-pr- ' r-o ' —tsiO CO ^^ON NO OOCN — •m rs rsTf NO ^^NO CN~^ es ooNO OO — ^CN s^ C*ln : r-:O ' 06^Hrs oo r--^ oo NOTJ- O tsm r- tsNO CNoo — * oin 2 r-00 i NOO ' rsrsCN CN ONr- m ONts —O ts ooro in r^vOON s Ttr- — f -tm CN —>ve5a g V ^I -3 &s| -Ji .c ci*- ^ 00 O< ZNO 'I-^ '*-NO r-r- r-rs — •ON inin rom NOr- NO^^ON is 'oom — its oo ro-H CN•* —•* OOON TtNO iNO 'r**CN Ttm tsrs rsO OOO NOO ooin oo-"ui* ° w oa ^> >.C O a. oo.C .Ci i§ §Z Zu£C.CCuo5.Euecd5—3•coof§c—•5-c uEoa.|o.e2e3uco= oo•£ £H.oc ;^n >v••a&2 H3 33 ou j;M ~k i£»! ^^Se s||M OII00 0c - _o c•S c C0M Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 236 The Well-Being of the Elderly in Asia policies. Trends in retirement age, as well as the underlying dynamics and the potential consequences of these trends, have received considerable attention within more industrialized societies. One of the more striking aspects of work ing patterns in recent decades has been the strong movement toward retire ment at earlier ages in most industrialized countries (well documented by Gruber and Wise, 1998;Kohlietal., 1991;KinsellaandTaeuber, 1993;Quinn and Burkhauser, 1994). These trends, triggered in part by financial incentives contained in existing pension programs, levels of individual wealth, and chang ing values toward work and leisure, have created concerns about the eco nomic viability of existing social security programs, and have led to efforts in some places to delay retirement by regulation or altered incentives (World Bank, 1994). Among our focal countries, Singapore advanced the age of mandatory retirement from 55 to 60 in 1993 and to 62 in January 1999 (see Chapter 3). Changes to delay retirement or the age of full entitlement have also occurred in the United States and Japan, among other countries (Ogawa, 1994). These changes are often contentious within a society, signaling com peting interests across generations and raising tensions between individual preferences and broader societal goals. They also serve as cautionary tales to newly industrialized and less developed countries establishing new social se curity programs, or modifying existing ones, about the difficulties of design ing programs that can balance competing interests and operate effectively over the long term. Trends in retirement age and their potential consequences are closely followed by researchers and policy leaders in many developing countries as they try to anticipate labor force shifts and the changes they may necessitate in social security programs. On one hand there is the expectation that as more workers move into the modern sectors of the economy, familial arrange ments for old-age support will not suffice and more formal systems will be needed. At the same time there is a desire to sustain the existing familial system and minimize the potential costs associated with some social security plans, which could hinder the pace of economic development. Complicating the situation in some developing countries has been the pressure to create new jobs for the large cohorts entering the labor force, leading to a desire for earlier retirement among older workers (Agree and Clark, 1991). The remainder of this chapter pursues these several themes. The next section presents a broad picture of labor force participation rates over the last 20 years in the region, as well as a closer look at the work patterns of the older population. This is followed by the development and estimation of mod els for assessing the factors associated with continued work versus retire ment, and a look at the associated reasons and preferences given by respon- Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Work, Retirement, and Leisure 237 dents. These attitudes serve as a backdrop to an examination of the leisure activities among older persons. Trend and Levels of Labor Force Participation Figure 7.1 displays the labor force participation rates by age from 1970 to 1990 for each of the four countries, as derived from census data or Interna tional Labor Office estimates (Raymo and Cornman, 1999). In broad terms, Taiwan and Singapore are quite similar in their overall patterns, while Thai land and the Philippines present quite a different configuration. In the first pair, men traditionally show much higher labor force rates than women but this differential is reduced by 1990, with Singapore showing initial increases for women by 1980. Women in Thailand and the Philippines, by contrast, show a pattern of sustained high labor force participation rates dating to 1970 with the high rates in Thailand particularly notable. Among the men in Taiwan and Singapore, participation rates drop quite sharply after age 50, with the drop-off occurring somewhat later in Thailand and the Philippines. In all cases participation rates by men at older ages seem to have gone down slightly over time, a trend examined in more detail below. In Taiwan and Singapore, where there have been sizable increases in female labor force involvement over time, women show higher participation rates in 1990 than earlier years, and this appears to be true to a modest extent in the Philippines as well. In Thailand, on the other hand, where female involvement has been very high for some time, there is an indication of earlier withdrawal by 1990, judging by the lower participation rates at ages 65 and over in that year. Raymo and Cornman (1999), who also present labor force transition rates based on these data and census survival rates, have estimated trends in exit rates that are similar to the patterns described here. Table 7.2 shows the level of labor force participation of older persons by age and gender, according to the survey data from the four countries. The notes to Table 7.2 provide the questions asked in each country and the deci sion rules used to define labor force activity. In each country it was possible to identify those who were working without pay and in some countries to identify explicitly those who regarded themselves as helping out on a family farm or business but not working for a living, a category often noted in ethnographic accounts. We opted for a conservative definition of labor force activity by excluding those working without pay or claiming to be only "help ing out." The excluded categories were deemed to represent a higher degree of volition in the use of time and therefore to be more appropriately combined Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 £ •J i/) 0l eEL Q. M> IT oooooooo oo OOr-vOi/^^COCN-^ linijiuh.ilJJ«,| 30JOJ J(K|l! I 0i Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 s uonBdppjea aaioj joqai •^ . ~ § O 2 a 3 I 1° — E 1 1 Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 240 The Well-Being of the Elderly in Asia Table 7.2. Percent Currently in Labor Force,* by Age and Gender Gender/ AgeMale60-6465-6970-7475-7980+TotalFemale60-6465-6970-7475-7980+TotalTotal60-6465-6970-7475-7980+TotalPhilippines70.956.643.228.726.052.732.835.928.720.78.728.250.444.334.723.715.138.5Thailand66.253.433.819.811.648.048.130.418.814.22.529.257.040.625.716.56.137.7Taiwan56.131.620.513.15.432.020.810.16.23.11.411.238.423.014.48.33.322.5Singapore55.131.217.510.94.732.516.98.95.23.90.69.035.819.710.86.92.119.9 *The Labor Force Participation Rate is defined as currently working full or part time or not working but looking for work. It excludes those who are working without pay. See Notes for Table 7.2 at the end of this chapter for questions and precise definitions. with those retired. The magnitudes involved are small; the proportions desig nated as working without pay vary from 0.1 percent to 6.2 percent across the four countries (see Table 7.3 below), inclusive of those who said they were "helping out" but not formally employed, and those who said they are working but receiving only material goods in compensation. These percent ages reinforce the notion that the distinction between work and leisure is meaningful among the groups examined. Although in all four countries respondents in the "currently working" category comprise a minority, labor force participation varies considerably Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Work, Retirement, and Leisure 241 across countries by age and gender. As suggested by Figure 7.1, Thailand and the Philippines tend to resemble each other, as do Taiwan and Singapore. In comparison to Taiwan and Singapore, participation rates in Thailand and the Philippines for both men and women are higher; more of the men aged 60 to 64 are in the labor force; and the decline of participation with age is not as steep. As expected, participation rates for women at aged 60 to 64 are much higher in Thailand (48 percent) and the Philippines (33 percent) than in Tai wan (21 percent) and Singapore (17 percent). Except for the Philippines, participation rates for women decline faster with age than for men within each country. The very gradual decline with age in labor force activity in the Philippines for both men and women is quite distinctive in these data. To examine in more depth the labor force activity in these countries at the older ages, we contrast these survey data with a set of earlier surveys from the mid-to late 1980s, as shown in Figure 7.2. (Because the earlier definition included working without pay, our figures have been adjusted ac cordingly.) The impression conveyed is one of stability in the levels of labor force participation across this period. Indeed, among the men aged 60-64 in the Philippines, Singapore, and Taiwan, the proportion in the labor force may have increased somewhat. For Singapore at least this may reflect the change in the age of retirement that took effect in the early 1990s (see Chapter 3). For women, slight increases and decreases characterize the shifts, with the increases occurring most often in the two younger age groups. To provide a useful point of comparison, we contrast the level and trends for the four countries with the picture among the more industrialized nations. In the United States, for example, the 1992 labor force participation rate was 56 percent for men aged 60-64, which is similar to 1995-96 rates in Singapore and Taiwan for men at those ages, as shown in Table 7.2. How ever, in 1964 the U.S. rate was 80 percent for men aged 60-64, illustrating the sharp secular decline noted above. For U.S. women aged 60-64, the labor force participation rate was about 40 percent in 1992, closer to comparable levels in the Philippines and Thailand. For the past 25 years, the trend line for U.S. women at these ages has been flat, with the increasing participation of women in the labor force offsetting the tendency toward earlier retirement (Quinn and Burkhauser, 1994). Other industrialized countries also show trends toward earlier retire ment for men, although the levels vary considerably. Japan is notable among OECD countries in its high labor force participation rates, with over 70 per cent of men aged 60-64, and a third of those 65 or older still in the labor force in the mid-1990s, though both these levels represent declines since 1970 (Quinn and Burkhauser, 1994; Kinsella and Velkoff, 2001). For women, both levels and trends vary across industrialized countries. In some, as noted for Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Figure 7.2. Percent of Older Males and Females in Labor Force, by Age and Country: 1980s and 1995-1996 Males 80% - 70% - 60% - 50% - 40% - 20% 10% - 60-64 65-74 75+ 60-64 65-74 75+ 60^4 65-74 75+ 60-64 65-74 75+ Philippines Thailand Taiwan Singapore D 1980s Data D 1995-1996 Data Females 80% -70% •60% •50% -40% -30% -20% -10% -ilk_——_ ._m r-, I] Irru-60-64 65-74 75+ 60-64 65-74 75+ 60-64 65-74 75+ 60-64 65-74 75+Philippines Thailand Taiwan SingaporeD 1980s Data d 1995-1996 Data Source: 1980s: Chen and Jones, 1989, Tables 5.1-5.3, 1989 Survey of Health and Living Status of the Elderly in Taiwan; 1990s: Table 1. Notes: For the 1990s, participation is defined as currently working, looking for work or working without pay to make the definition accord with that used in the ASEAN surveys. The 1984 Philippines survey was not a nationally representative survey of the older population. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Work, Retirement, and Leisure 243 the United States, there has been little change over the last 20 years, while in others there have been declines at the older ages commensurate with the changes among men. In a few countries such as Sweden and Norway the trend for women aged 60-64 has been toward greater labor force activity. In terms of level, these Scandinavian countries. Japan, and the United States show the highest rates of labor force participation among women 60-64, in the range of 40 to 50 percent. Although labor force participation patterns are similar among the cur rently industrialized countries, there is also enough variation to preclude pre dicting the future path for other countries experiencing economic develop ment. The case of Japan, for example, illustrates the potential for a distinct trajectory despite a high level of industrialization. In addition, several observ ers have noted that the trend toward earlier retirement among men appears to be slowing in a number of countries (Quinn and Burkhauser, 1994). Although the current levels of labor force activity observed in the four Asian countries in this study find their counterparts in one or another of the industrialized nations, these developing countries do not share similar histories. It is also possible that the policies adopted in these countries and others at similar levels of development will lead to different trends than those observed re cently among the more industrialized nations, as discussed further in the Conclusion. Table 7.3 presents a more detailed picture of the labor force activity of older men and women in the four countries by broad age group and urban- rural residence. As suggested earlier, agricultural activities are an important source of employment for the elderly in developing countries, and the pro portion of older men and women currently working is higher in rural than urban areas for each comparison except Singapore, which does not have a rural sector. The proportion working without pay also tends to be higher in rural than urban areas, with a large differential in Taiwan. The proportion working full time, however, is not very different in the urban and rural sec tors. The proportion working part-time is higher in the Philippines than else where, for the comparisons available. As a proportion of those working, part- time employment is more common at age 70-plus than at age 60-69. A higher proportion of women than men who are working are working part-time, though the difference in the Philippines is negligible. Though the higher labor force activity rates for the Philippines and Thailand compared to Taiwan and Singapore are partially accounted for by the differentials within the rural sec tors, the former two countries also show higher rates in urban areas, particu larly for women. The role of cultural effects, in addition to economic and policy dimen sions, is glimpsed in the differential labor force behavior of women in the Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 I 0l ee '«— -5 •o 0J •o 0l LU 73Ba•a&u-a2i u •; cui •< 0ODC••siz•a•g£00c C/)$3H£00B CO?3.<3a.Wmk Activity00r- CNoo^^ CN•*5CNvo•*mCNeNONOt- mmmmr-fi •tro5m00mmtsNOooc -^I2t*i&0 <? ?O r-ci-/—v /~sON Ovm roCN N~'^^^ri nB BS* voeCN PIGO87 vO*vo in^^ N/-^. ^-^ve em mN- 'a aB B^ 0?ts 0Ci O/--. /~'m TtTt ONro ^/ V ^NO NONO •*c. ^^-i ciB BP f-T^Bs full time s part timerof»0CNTtOt*^oNO?4r-oNo0CNr-NtsoiIn LFs not wmkin-— 00o o r-in Tt00 Ofi r^ ^^t ots om— i •*O NO 90Tt COr-£m voCN r- in^^ ON— i CO «— NO0 •* '/^NO —r- ON•*00 —m m mO mm CN coWmkin- w^*o pay Not wmkin-NO.CN.m r- SCi/— ^ /-^oo ono3Sr?oo3 ~^-^ /~vts oo^^ O C- OCo —•* CNm m'*^ ^— /W^— insc.is-i ~z~m* Win oo 2- CS-T p-Tm —m — 'ooe mr- —•S ^o? foe-* 0m %—r? ^fei ^—^s retired s mver wmked•*NONOo-.t~)£C4vO W00CNmt*1tsr4mo tsf1Sev•*CN CN1inmA-e_(M-Crtrently wmkin-ro tsSCJ.r-^ r- ve m00 — ^-x v^.d dC B?T &7m —^^ v.^m inCCt^ OCN — ,d dc ee-— oom ~00 ro5T ots rj^^- -3C C^-^ ^-^vO r^vo t-1s full time - part timeONOr-eoV)or-or-oirto00onr**evo00In LFs not wmkin-ts m0£tN in•»2t- NOO — oom ONr-— **•0§— t--sTf 00O NOGOO ^^— '- r-CN to0£Ov NO^F:— vo-?ts voCN tsNOWmkin- w^*o pay Not wmkin-/•* N /— NNO ONvo m m ro^•v ^—Nr- ootSCN* ?oo m r- e-^-00vo m in —Pe Pem om ve? 5e5 0s«^^ ^^ CN vo— in oo •— '£•"•v ^e^r- tsw -it in tNN Ooo in^StiP o7•* CNr- e-^r? ??so? ooe0N CNm e"s retired s mver wmked Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 oqints-^-^p—<^^ odC^iin-"OO^^O •o _c —I "i JO1 -0 ts — —• r- in ts in d d 1; vo c c — oo vri m — moo r- vo ^ 1 1 1 1 1 1 •"tcsts— tsOtsod r- B B r-i ts r~ vo -^ m M — voin— •*e•^ — Ov —" Ov ON C2 • ^^ m m Oodr-iOmvo.cifi ••• O d ro C C — OwStsnS CN r- vO — ' i i i i i - ^^ ^ . — 0x — — ooootsvo oo c c— • — r-vo — CN NO NO ^ r-in_ ts — _. vOTtts .5 >v -a a •3 3 u • ^ §. .S g -s M -a fist z — oots — vo— ' — O CN — ^ r- in ts tN — in Tt — c c — — ^^ in r- d os NO ** ' vo vo v-^ — Or-in— • in c crsts^^ooo Tt •* ^ a. j= -a Table _ See No Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 246 The Well-Being of the Elderly in Asia four countries. Until quite recently it was uncommon in Chinese societies for married women to work outside the home, an injunction reflected in the high proportion of women in Singapore and Taiwan who report never having worked. At the opposite end of the spectrum, the greater gender equity on this dimension in Thailand, and to a somewhat lesser extent in the Philippines, is revealed by the very low percentage that never worked. Table 7.4 presents some additional background on the current or last job held by the older population in these countries and on their occupational history. Among men who are currently working, farming or other agricultural pursuits dominate except in Singapore, where sales and craftsmen and la borer categories are the norm. The rather low educational level of older per sons is reflected in the small proportions in professional/technical or adminis trative/executive occupations, particularly in the Philippines and Thailand. (The relative high percentage of men reporting administrative or executive positions in Taiwan reflects both the large number of older Mainlanders in government positions and the large number of mid-size, family-owned busi nesses). Commensurate with these occupational distributions, a very high proportion report being in their current job more than 30 years—around 70 percent in the Philippines and 50 percent in Taiwan and Thailand. And al though the countries under study are cash economies, it is not unusual for men to report that some of their wages were partially or totally in goods rather than cash, another indication of the importance of agriculture for these older men. The job characteristics of currently working women are not very dif ferent from those of men, although a slightly higher proportion of women within each country are less involved in agriculture and more involved in sales, seniees, and related occupations. Accordingly, women also display somewhat shorter tenure in their jobs and they are somewhat more likely to receive their wages in cash only. Overall, the pattern for men and women suggests some division of labor, with wives and other older women more actively involved in the market place and complementing farm earnings and production. The nature of the last job held by those now retired is shown in the lower half of Table 7.4 for each country but Singapore, where the data are not available. Contrasts with those working are influenced by the differences in age of the retired versus the currently employed as well as by the influence of the job on the probability and timing of retirement. Thus, among retired men in the Philippines and Taiwan, a smaller proportion report their last job as farming or agricultural than their currently working counterparts, and those in the Philippines show somewhat shorter tenure. In Thailand, by contrast, Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Work, Retirement, and Leisure 247 both retired men and women show similar levels of concentration in agricul ture and job tenure distributions as their currently working counterparts. The influence of several job characteristics on retirement will be pursued more fully below in the multivariate analysis. Modeling the Retirement Process A rapidly growing literature examines the process of retirement within the more industrialized economies, especially during the past few decades as that decision has become more voluntary. As Quinn and Burkhauser (1994) note, the earlier literature viewed the transition to retirement as largely involuntary, driven by poor health, layoffs, and mandatory retirement policies (see also Quinn et al., 1990). As the growth of Social Security and pension benefits began to influence the timing of retirement, however, researchers saw that retirement was increasingly a labor supply choice involving a trade-off be tween work and leisure. In addition to incorporating variables reflecting the availability and/or value of pension and Social Security benefits, models of retirement often included measures of health, asset income, job characteris tics, local labor market conditions, and several family-related variables. These models have been enriched by a life-course perspective that stresses the rela tionship between employment and family life dynamics on such dimensions as health care needs of family members, marital status, and reproductive histories, and the employment status of relevant others (Szinovacz et al., 1992; Shih, 1997, p. 35). In addition to the nature and content of models of retirement, a num ber of measurement and estimation issues must also be addressed. The con cept of retirement has been defined in a variety of ways (Gustman et al., 1995). In some analyses retirement is defined in terms of collecting pension or related benefits regardless of work status; in others it is defined as leaving any full-time job after a certain age. Many analyses that focus on current labor force status include with the retired those working less than full time. In this chapter we consider an older person retired if he or she is neither cur rently working full or part time for pay, nor looking for work. In modeling labor force behavior we are typically examining the oc currence of an event over time, and the preferred mode of analysis is event history or survival analysis, in which the hazard rate of the event occurring during the period of risk is determined (Hayward et al., 1998; Hayward and Grady, 1990). To determine these rates in a multivariate analysis requires Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 cOrjuSi00B (a.*H.ac 111ta00cB1H•gja_cH*Ac.5.Q.ISeuCA5aQ*s*5V1Q...,£«jQ^,jS2Q#5£?*_M00E12 £_>,cuUOro— ifi— iooiTt ntantatasa tatdra . BBBBBB BCB— fiOCNvO^m— iriTtvOr^O^^ OOOCSm— — iCN—i tN — i — • CN MRoo=,2lm 5«==K- RKgoOTtCNO^"OOTtC»l NOr-OOvOOO CNOOO CN Tt — i CN— iTt I^__-H^^vo"^2'^ tdtantdtan nn*aB B B B B B C E Cin-^oointsmvor- 1 irsovoTt -.00 —to- _^_ ^___^. r-tNON — i ts — O f^ i ^ t^-OvtsoomO ror--..JvO — — _ _ vi t^tN.mo)mor~-.CNts oor-r-r-rso oNmooC^- — i I ^ Tf r*^ — ^g j, ^ |•3 e» 3 '3 ^ 1 °°•H -g u i« £. 8. mjihi i !- -- i HiJJ^-^ la 3ifi«&& -5 a & -' •°Be5Bg 1 » u § 1 1, ^ >i>,E -^1.mS^.213, iS ^^oS^u-S «S'eow^.p.S-H O O ^^ ® _T! JC ft3 fa— < ~ qj — k« ^ . , ^^ ^^ 0*^s^ + 6 ^52^ j^ t^*^ U, 0, < U 00 U Co J ^ vo - ts m ^.U U 0 Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 T3 0lec u*2mK,_5g0*C 0Suu0c.cecC/5•aci.aHT3Crt 5—uistc.£L a.IS a.fjuCa.(3it,c^ocH'•0J2PMOC*E. c.—r-a.c^5.^"5t-&c^5*vi5*•$tn.•^t/i5^jO^t^Q^•^D^A33-.^^Cartatdrtc3rtrt rtrtrtrtrjrtcccccccc ccccccoomoovovovooom ovr^^^ — ooootN — tN — — TrvO — o o m m i r- in-itOtst~tseoTt-M-oo^^ S-S-S"0AsAAc3tdAtd AtdtQtdtatadddddddc ddccddtN CN~ — — _-_in.g^r^m-OOiO ^ "i - Z S ~^^CNTtCNOO — CNCN r^i vO r^. O CN NOI > 8133 * |•c .£ g £2 .2, S-MJ= S Ji g JJfe^e> -^ .» !? C£ c ^3 5 e3 •a — £5s2c5n -^an$o*^ u •" r 4> >v >i M oeSc^c^ Co ^^Or-jroaj-^3 I— O ^^ U -^r 2 b- T 6C ^v 2u [L uZ ^ CJ (^ CJ c/^ O Svo^^r^lf^QO JVi -^ *-^4 SO < ^e> Ofl r^.— ^ vCo ^J u'^ ^ a3 *w t/1Vi 0. *jq ."*™ eO (— yy c oO vi eN rt!i|o0 0 —'^> ^ w 2j1l n) J^ ^^ eT . c C-5 £ 0 m_O o £ ^^T3 § tN 2" eu — inII 8 9 9Po ^ Q- c2 u — o « ^ .£ c SeO C ~ ."* ,ft§ 0 ^ U Mao= o J£ Sr= 3 O o c.2 -a .0 > oS £ — x ^S f1- ec ^ 5vS e | ^ i-° l«.^-5•5^2 = 0CO rt U o3 E - -ao «ec eS =g)| £ J Cer> u u ^* ^-* X .— CJ t/l^ .£ > w to;5 .« « + 3u u -o O e^— *- r- vDO ca u *—% ~ e^ -y: su cj K " sI|l2|m £ ^ O •—-•' CJ ^ ¥ —T3 - eO .— •=: ^ ^, 3 a.l^1!^'« 3 Il •- &O ~ ,J y *J*s- ~ « .c • = S J c1 *•• ^i— r-* .7 s^ L.r3 c c -^ •?•a ^ "^ o ^ Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 250 The Well-Being of the Elderly in Asia knowing not only the timing of the event in terms of age or calendar time, but the status of any time-varying covariates. Many studies lack sufficient detail and use instead a measure of current activity along with the hypothesized determinants in a logistic or multinomial regression. Our study data fall be tween these two structures. For the Philippines and Taiwan, retrospective and/or panel data allow for some survival analysis, but the requisite detail is lacking for Thailand and Singapore. Accordingly, in carrying out the com parative analysis, we use cross-sectional data and logistic regression models to examine the determinants of respondents' current retirement status. In addition we supplement these analyses by providing several highlights of sur vival analysis carried out from the Taiwanese data. The determinants included in our model have been suggested above, and are designed to capture the major factors reflecting preferences, needs, and opportunities for employment. Gender is likely to be particularly impor tant given men's and women's very different historical involvement in the labor force and the types of occupations open to each. For this reason we analyze the models separately for men and women. Age, health, education, and urban-rural residence are included as indications of ability to continue work, the wage rate and implicit benefits derived from employment, and employment opportunities. In addition to these individual characteristics, we use factors reflecting family structure, which is thought to be highly influen tial in the decision to work or retire. We include marital status and work status of spouse, which relate to needs and resources as well as the often joint-decision process of retirement (Blau, 1998; Henretta et al., 1993). The number and ages of children and living arrangements speak to the potential for receiving support as well as the need for continued provision of income. The class of worker and other dimensions of employment, as noted above, are also likely to be influential. Farmers and those involved in family busi nesses can withdraw gradually as age and health dictate and may choose to do so because of the absence of a formal retirement plan; workers in other sectors may be forced to retire because of mandatory age rules or may be encouraged to retire early through the structure of the retirement program provided by the employer or the government. In general, coverage under a pension plan is likely to reduce incentives for continued employment and to facilitate retirement. The methods used for measurement and estimation are critical to esti mating the effects of these factors. In many cross-sectional surveys, the timing of key events is not determined, which means cause-and-effect rela tionships cannot be distinguished clearly. Changes in living arrangements may be jointly determined with changes in labor force activity, for example, or poor health status may have arisen after retirement. These shortcomings may Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Work, Retirement, and Leisure 251 be addressed through detailed retrospective probing or appropriately designed panel studies. In the absence of such data, analysis is sometimes restricted to clearly exogenous factors or fuller models are interpreted with due caution as to possible confounding effects. We lean to the latter approach here, omitting some variables that are clearly jointly determined with retirement but includ ing some where causation remains problematic. We believe that on balance these facilitate comparisons across countries and set the stage for more re fined analyses as the data become available. (See Agree and Clark 1991 and McCallum 1992 for related but alternate models for developing countries.) The next section presents both bivariate and multivariate analyses for each country using the determinants discussed above, as well as pooled re gressions for studying the extent to which country effects exist in addition to the included factors. Correlates of Labor Force Participation Bivariate Analysis Table 7.5 defines the measures used for determinants in the model and dis plays their means and distributions, by gender and work/retirement status, for each of the countries. The bivariate strength of each variable can be gauged by contrasting the distribution or value according to work status. As shown for each country, men and women who are retired have an older age distribution and are more likely to be in poor health than those working. Reflecting differential opportunities for continued work, those retired are also more likely to be urban residents and to have been in a non-agricultural, non- family employment in their last job. In every country but Singapore, those retired are more likely to have pensions or severance pay arrangements than those working (discussed further below). Family variables also appear to be influential. To capture both marital status and the employment status of the spouse we use a trichotomy that interacts these two concepts. This reduces the strong co-linearity between two separate variables where all those "not married" on the marital status variable would be "not applicable" in the spouse working variable. At the bivariate level, those retired are less likely to have a working spouse than those currently working, for each of the contrasts, but they are also less likely to be married (mainly because of their older age). Among those mar ried, the proportions of those working and retired who have a working spouse are as follows: Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 252 The Well-Being of the Elderly in Asia PhilippinesThailandTaiwanSingaporeMarried MenCurrently Working.31.79.25.24Retired.31.35.14.09Married WomenCurrently Working.64.75.50.45Retired.44.41.26.15 Except among men in the Philippines, married respondents who are still working are much more likely than those who are retired to have a spouse who is also working. Given the differentials in labor force participation of men and women, older women who are still working are especially likely to have a husband who is also working, but in Thailand working men are as likely as working women to have a spouse who also works. Among the other family variables, those retired are more likely to be living with one or more married children, although, as noted, the cause-and- effect implications of this variable cannot be fully determined from a cross- sectional analysis because a move to join children may follow retirement. Working men are more likely than retired men to have one or more children still in school, reflecting their age as well as the need for continued support. The pattern among women in this respect varies across countries. The po tential for support from children is partially captured by the number of living children and the average numbers suggest that those who are retired tend to have larger families, though this does not occur in every country and some differences are small. Education is a personal characteristic expected to affect an individual's preferences, needs, and opportunities for work, but the influences often op erate in countenailing ways. Those with higher education are more likely to be covered by a pension program and have greater economic resources, promoting earlier retirement; but they are also more likely to have a higher wage rate and derive more implicit benefits from work, leading them to post pone retirement. From an opportunity standpoint, those in rural areas where continued involvement in agriculture is possible often have lower education than urban dwellers. These conflicting influences are revealed in Table 7.5 by the general absence of any strong differentials in the educational distributions of those working and retired, at least in the bivariate situation. Accordingly Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Work, Retirement, and Leisure 253 we turn to the multivariate analysis for a more detailed examination of this covariate and the other determinants. Multivariate Analyses A clearer picture of the effects of each of the independent variables is found in Tables 7.6 and 7.7, which present, for men and women respectively, the results of logistic multivariate analyses for each country. In these tables the dependent variable is retirement, coded 1 for not working, not looking for work, or working without pay, and coded 0 for working for pay or looking for work. The dependent variable in a logistic regression is the natural loga rithm of the odds of a specific outcome—in this case retirement. For ease of interpretation, rather than show the regression coefficients, we present their exponentiated values. These can be understood, for a categorical variable, as the ratio of the odds of being retired for someone in the specified category relative to someone in the reference category. For an interval variable, the exponentiated value is the proportional change in the odds ratio for a unit increase in the variable. The relative odds presented in Tables 7.6 and 7.7 confirm in a broad sense the general patterns observed in the bivariate rela tionships, but offer a more precise picture of the strengths of relationships and their variation across gender and country. In each country we see a significant monotonic increase in the likeli hood of retirement as one moves to the older ages, relative to those 75 or older, with the slope for men somewhat steeper than for women. The effects of education are minimal, rarely significant, and not consistent from country to country. In Singapore men with a secondary education are more likely than men with a primary education to be retired; in Thailand men with secondary education are less likely to be retired. Women with no formal schooling in Thailand are more likely to be retired than women with a primary education, but their counterparts in the Philippines are more likely to be working. In the bivariate analysis, being in poor health was associated with being out of the labor force. This relationship holds in the multivariate analysis as well for each country and for both genders, but the effect is somewhat more consistent and stronger for males than females. For example, the odds of being retired among men in very good health in the Philippines, Taiwan, or Thailand are 11 percent or less of the odds among those with fair or poor health. Among women, this low a proportion is observed only in Taiwan. As expected, the characteristics of the job exert a significant influence on work status. Male elderly in non-agricultural/non-family jobs are more likely to be retired than are those involved in agricultural or non-agricultural family Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 i/?I5cuE.£iuJ_>-I/1uja.2tO>ecD•oVo.0l1i--o'3TV2> cn0E•ocJO,c ^o —.•c £3 3o i"iSBH^uaeOSs•i00c\^o^'^"w U3U00>3H.3s.^£opK.^Sji2H—£u 1"l—S15CO V^ CN ^* vO ^t O v ' ' ^ ^ TtrM Cv1 oJ Csi O ^D C^ in CO ^^r*e) CM ro ^^ ^^ ro NO rM oo •* r^ &voo-^iniri Of^r** ^in r^rs^^— ^ C^^ CN C^) -•* vjO Cs4 r^ CN) tN) T^* tN.NOCN^OO r^Om OO ^OOx roCNCN^^ ^3ON*^ CNOO r*-i i/1 •— ir.r;,*^ ,,^ . ,,r, ,,0-— i ro m ts O2i0 •* in ovinintsCNtsCO CNTJ;^ -*tin -* "} C*1S§S§ 2§^ :| 5^—* ...iO vO e-H oo ^^ ^ v^ ^^ ON O t**- r*^ON O *onT iO ^^ CM O vO ro ^ ON ^H** t*} -1 P CNvq— vq m -: •* no ^^ o ov o o 5? ^* m ^ o® Ou^ CN — ; p ^-^^! e-?°9 "I^PvQ ^^ •* ON -H «N r** ^3 T^ ^^ ^^ ^iir1 r^* ^s ^j ^3 NO r^* ro ^w oo o^ ^5Tf C^ *-H ^e^ 0^ IO ^^ fi NO ^^ NO O^bOc60?.5 oill cI 1 1 li§ I Jl 1_j>33« H U fe-i" T- S. 8. eS S w a -^H T^ ^^ Tf O tfi V1 .Si ^ ^j e~3 ^%NO vO r^* i i ^ ^ V" ^ 3? *—n.ooooooeo'C.ro.^o Se£3 5Jo>:4J^ < . s ^ "e. " . s ". "' _ Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 able 7.5. continuedi-—i — OO ;vQ 1-*- vQ .-•NO —NO ON IO ONO r- -H m m in -^ CN8— ' •* r- r- CN —ro i0 -^ NO^ CN r-O IO Tt OO— i NO CN CN— CO NOO — oo iCN r- oSNO Tt r- oo — i roI0 Tt q mS-*t CN m sO ro "ti0 on^ o NOON CN ON TtCN r- q TfHealth ratingVery -ood Good Fair/pomLast/surrent occupationA-ric.lt.rals family farmIiQqON mqoCNi[NOq300TrriA-ric.lt.rals nonsfamily1 iON I**oo Om — i— i CNCN r-— i r-,i i-.voCN •*— i CNCN r- CN q— ' r-Nonsa-ric.ltural, family Nonsa-ric.lt.rals nonsfamilyHas pension/severance pay from current/prior jobsoo00mC^CNc*ONCN-t•^r-;—CNi_fr^,M0O — • ONI0 oo —IO CO NOO CN 00CN IO ••*vq •* ioOO NO Ttr- ON c5oo Tt i0oo m r^O IO xt.r- Tt i0NO CN OOin O ^^CN CN r-;ON CO p*^I0 — i oor-; Tj; IOf~i ' ^ ro ON Tt v-iNO esvO ^^ ^^oo ro O1•o .CL.09f.5O 4* -J CZ. .> > ZooPONPrsONCNTriOCNON-^r-OCN1 5a SCNON_ro ONOO vOOOr^oovoTtr-;ooooCN OOoONr-;OCNOOr^i5voIOONCO"^wooONIO'nevdNumber of living shildrenoo..r-00 CNsONONONCNOi0CNNOO.ooooCNNumber of cases (using normalized weights)e83CC8 Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 00 O — r I — 0l U3 LO •3 rt sa t/5 •3 ts — oooo CO CN — i CN ro oo vO m CN CN -H m tstsrsC^ — i CN — r- r-tso — * rs vq Oc^O v£>CN o — roO vO tN — r- — ts Troot** oooCN ts m O m -^ ro c* o -<: S CTv oN es r-m V-i •*? _. i li «J OOCTyC*! tsovo iX>— *O ON — i oo Orot-- mtsin OOONCN —i ts r- ts o r- Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 able 7.5. continuedi-ts to i0O oo ^^ — i NO CNoo i0 r-CN NO OCN vq — 'ro ON 00CN iO -Hto oo oo-; NO —ON IO NO— 1 — . NOCN O ~O r- roNO CN — * ro NO poo in r-vo ** ooro I0 OSm oo 00 O—•. ^ —•.m•s eB u S ° 2 00 X.iill S£CN NO CN O ON -^ ^3 OO NO ON ^^ ^j ro iO ^^ r^ei ^^ ^^— i oo •*t r^ oo CNo NO r- i0 — • oo r- o ^ ^ o p ONO OO OO Tt iO IO-!-inn .»iiii ON ^^liti TJe IOi0 •*ON IO r^ ON Tr NO i0 ON OO i0 iO Tt CN p — i Tt p ONro NO 3 NOi/1 r^ Cs ^ 5? ®^'09 ^ . . °i— i ON Tf NO CN OO r^- — i ON — ' 00 — i— — ,>,_ p ONe .E a" 1B C ^^-^2 * S o.2 .2 .§ .2 i S o TJIf!!! si 1 I1.!!! u s 1 1 1 1 1 1 1 2C -" " t. r S 1- r;re3e3ooooa3 -^ Syuf^Su *.Sisbobee «)EtSo $I«zz J|>-z .>U -C iS Jin i0 oN -H oo i0t^- CN CN r- r** mTt iO p ON ^ vOOO CN NO Tf CN CNI0 ^ O ON .. r**m i^ — i ON i0 JOoo — * T* I0 -^ S i0 Tj; p ON ^ |Q.-^ ON CN OO ON NO ON O OO -— i IO OO^t IO O ON 1X mi0 I0 CN oo m ^HCN r- m NO -H NOCN r-; p ON ^: — •in i0 ON -^ m rn CN r^* i0 ^ CN oo ^t i0 p ^^ ^ — ioo CN ON ^H m ^^CN t ^ ON O O I0•^ i0 P °5 i0. ^ro r^ •— * ON ON I0^ IO Tf IO NO -^**} ^? H ^ tn ^— , e1 * >u = -= M J= 2 «• s « 3 j§a c° 8"u -r w «t> > CO *•o s w *r: t_ i_ -a•S ° ° N u h U .S, aj 4, — ;» ,5 „ * * gUO « u O S PC>sz S8>sz § § g X Z Z c Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Table 7.6. Odds-Ratios from Logistic Regression Coefficients for the Effects of Demographic, Socioetonomic, Health and Family Characteristics on Retirement, among Males Who Ever Worked + p<.10 *p<.05 **p<.01 ***p<.001 CharacteristicAge60-6465-6970-7475+Married, spouse workingMarried, spouse not workingNot marriedUrbanEducationNo formal schoolingPrimarySecondary or higherHealth ratingVery good GoodFair/poorLast/current occupationAgricultural, family farm Agricultural, non-familyNon-agricultural, family businessNon-agricultural, non-familyHas pension fromlast/current jobLives with married childHas 1+ children still inschoolNumber of living childrenModel Chi-square(degrees of freedom)Philippines.12***.31**.70—1.22.63—1.421.21~1.08.06***.12***".13*** .16***.09***1.482.07**1.06.98175.5(17)***Thailand.15***.17***.48***~.18***1.02"3.24***1.31—.73.11*** .30***~.56* .43*.18***2.73**.851.211.02753.2(17)***Taiwan.10***.27***.47***-.49***.73*~1.46*1.08—.80.07*** .16***~.55** 1.91 +.61**2.44***1.20.62*1.09**474.9(17)***Singapore.15***.17***.48***-.18***1.02~3.24***1.31~.73.11*** .30***~.56* .43*.18***2.73**.851.211.02753.2(17)*** Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Work, Retirement, and Leisure 259 businesses. For females, this relationship was clearest in the Philippines. Non- agricultural/non-family jobs are most likely to be associated with mandatory retirement ages and to be inflexible regarding work at older ages. In association with type of job, we found that males and females in urban areas in Thailand and Taiwan were more likely to be retired than their rural counterparts. The urban-rural differential did not appear significant in the Philippines after taking account of the other factors. Having a pension or severance pay is clearly associated with a greater likelihood of retirement among Taiwanese and Thai men and women, but not, in terms of significance level, among older Filipinos. In Singapore, the surprising negative effect of pension coverage on retirement, observed in the bivariate relationships, emerges in the multivariate as well. The influence of pension coverage on labor force participation is com plex because it is likely to reflect economic needs as well as skills and oppor tunities for continued employment or reemployment after retiring at an older age. In the Philippines, Taiwan, and Thailand many of those with pensions— especially those receiving more than the minimal amount—are likely to be governmental employees whose overall economic resources may be greater than others, facilitating their withdrawal from the labor force. In Singapore, the pattern is reversed: those with pension income or coverage show higher labor force involvement than their counterparts. Although the reasons for this are not entirely clear, they probably center around the fact that a high propor tion of the population covered by the Central Provident Fund (see Chapter 3) have regular attachment to the labor force and marketable skills, in contrast to those without coverage, many of whom were casual, part-time, and/or contract workers. The early age of mandatory retirement under the Fund allows those receiving benefits to find relevant remunerative work or self- employment. Hayward and Wang (1993), who found a similar effect of pen sion income on employment in Shanghai, attributed this effect to the early age of retirement and the potential to use the pension income in entrepreneurial second careers. Because including measures of both marital status and spousal work status in the regression may introduce high levels of col mean tv we created the following trichotomy to minimize this potential: married, spouse working; married, spouse not working; and not married. As shown in Table 7.6, mar ried men with working wives are much less likely to be retired than non- married men in each country but the Philippines. Married men whose wives do not work show little difference in work status from the non-married, except in Taiwan, where they display a lower likelihood of retirement. The pattern for married women with a working spouse is similar to that for men. In each country but the Philippines they have a lower likelihood Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Table 7.7. Odds-Ratios from Logistic Regression Coefficients for the Effects of Demographic, Socioeconomic, Health and Family Characteristics on Retirement, among Females Who Ever Worked p<.10 *p<.05 **p<.01 'p<.001 CharacteristicAge60-6465-6970-7475+Married, spouse working Married, spouse not working Not marriedUrbanEducationNo formal schooling Primary Secondary or higherHealth ratingVery good GoodFair/poorLast/current occupationAgricultural, family farm Agricultural, non-family Non-agricultural, family businessNon-agricultural, non-familyHas pension from last/current jobLives with married childHas 1+ children still inschoolNumber of living childrenModel Chi-square (degrees of freedom)Philippines.25***.28*** .37**--.93 1.90*1.26.61* 1.00.82 .60".36** .44*.21***1.242.27***2.86*1.01101.4(17)***Thailand.13***.26*** .40***~.39*** 1.41*2.39***1.32* 1.73+.32*** .54***"1.34 1.28.38***3.42+1.45***.66*1.06**645.2(17)***Taiwan.11***.25** .30*~.57* 1.291.63*.90 1.31.06*** .12**"1.69* .77.993.61**2.15***.45*1.18**134.3(17)***Singapore.17***.24*** .46*-.52* 1.56*inap.1.04 1.20.42** .98"n.a..54**2.29***1.211.09*121.2 (13)*** Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Work, Retirement, and Leisure 261 of retirement than the comparison category of unmarrieds. Married women whose husbands do not work, however, are more likely to be retired than their unmarried counterparts. As hypothesized, living with a married child is positively associated with being retired. This effect is strong and significant for females across each of the four countries. For males, living with a married child has a signifi cant positive effect on retirement only in the Philippines and Singapore. Liv ing with a married child is assumed to reflect a larger social support structure than living alone, living with a spouse only, or living with an unmarried child. As noted, however, the direction of causality is not unambiguous in a cross- sectional analysis. Another measure of the family structure effect, having at least one child in school, is negatively associated with retirement for men in Singapore and Taiwan, perhaps an indication that they are working to help finance their children's education. Children attending school are likely to be younger and more dependent on their parents for financial support than those not in school. Women in Taiwan and Thailand are also more likely to work if they have children in school. On the other hand, women in the Philippines are less likely to work if they have children in school, which may indicate they stay home to take care of their children. The greater the number of living children, the more likely are older women to be retired in Singapore, Taiwan, and Thailand. A similar story can be told for men in Singapore and Taiwan. This supports the argument that a greater number of children provides a broader support network and reduces the need to work. In the Philippines, however, number of children has no effect on work status for men or women, nor does it seem to influence the retirement behavior of Thai men. Country Effects A persistent theme in our comparative analyses is the discovery of any coun try-based effects unrelated to either differences in the distribution of key determinants or the way these determinants operate. In Tables 7.2 and 7.3 we noted that the labor force rates were quite distinctive across countries, with the Philippines and Thailand generally showing higher profiles of labor force activity than Singapore or Taiwan. The multivariate analyses in Tables 7.6 and 7.7 indicate that, despite these differences in magnitude, the pattern of effects is generally similar across the countries, judging by the magnitudes of the odds ratios. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 262 The Well-Being of the Elderly in Asia To examine whether other factors distinctive to each country might account for the observed differences in overall labor force participation rates, we pool the data from each country and replicate the model from Tables 7.6 and 7.7, with the addition of dummy variables representing country of resi dence for those living in the Philippines, Taiwan, and Thailand. For those residing in Singapore, we use an interaction term that includes pension status to capture its particular effect in Singapore vis a vis the other countries. If no other effects distinctive to each country are operating—that is, the countries differ only in the distribution of the independent variables—the country-level variables should not be significant. If these conditions do not hold, then we would expect to find country effects. Table 7.8 presents the results separately for men and women and, to reflect the lack of a rural sector in Singapore, separately for urban and rural residence. For urban residents, aside from the pension variable and the inter action term, the other coefficients remain very much in line with their levels in Tables 7.6 and 7.7. The country effects are quite prominent. Relative to urban males in Thailand, those in the Philippines and Taiwan are less likely to be retired. The country variable for Singapore cannot be interpreted without reference to pension status and this will be pursued below. Urban women in Thailand are more likely to be retired than those in the Philippines, but less likely than those in Taiwan. The overall pattern in the rural sector is similar to the urban among the three relevant countries. Philippine men and women are more likely to still be working than their Thai counterparts; Taiwanese men show little difference, but Taiwanese women show a higher likelihood than Thai women to be re tired, after taking into account the individual, family, and job characteristics included in the model. To make clearer the country effects and the special nature of pension status in Singapore in relation to the other countries, Table 7.9 presents the predicted probabilities of being retired for urban and rural men and women by pension status, derived from Table 7.8. (These probabilities are calculated by setting all the other independent variables at their mean level.) In each country but Singapore, those with pensions are more likely to be retired than those without pensions, though the differences are small among rural women. In Singapore, men without pensions are 27 percent more likely to be retired and women without pensions about 12 percent more likely to be retired than their counterparts with pensions. The predicted probabilities also indicate the differences across coun tries in the propensity to be working or retired, taking into account all the variables in the model. To illustrate the relative magnitudes across countries, Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Table 7.8. Odds-Ratios from Pooled Logistic Regressions for the Effects of Country and Demographic, Socioeconomic, Health and Family Characteristics on Retirement, by Urban-Rural and Gender + p<.10 *p<.05 **p<.01 ***p<.001 CharacteristicAge60-6465-6970-7475+Married, spouse workingMarried, spouse not workingNot marriedEducationNo formal schoolingPrimarySecondary or higherHealth ratingVery goodGoodFair/poorHas pension from last/current jobLives with married childHas 1+ children still in schoolNumber of living childrenCountry effectsPhilippinesThailandTaiwanSingaporeInteractionsSingapore * has pensionModel Chi-square (degrees offreedom)UrbanMale.14***.30***.43***--.39***.83—1.12—1.16.22***.41***~2.55***1.60***.66**1.07***.31***—.69*1.52*.14***865.7(17)***Female.17***.23***.34***-.70*1.45**~.90—1.25.48***.65*~2.18***1.94***1.66*1.02.62**~1.72**2.49***.21***336.5(17)***RuralMale.15***.27***.63**~.28***.65***—1.19~1.12.08***.18***~3.30***1.171.27.99.35***—1.16inap.inap.803.8(15)***Female.19***.32***.51***~.59***1.59***—1.03~1.43.39***.55***"1.041.73***1.071.05**.66***—1.91***inap.inap.397.2(15)*** Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 264 The Well-Being of the Elderly in Asia the data in Table 7.9 are set forth as indexes, with the Philippines taken as the reference point: With pensionPhilippines ThailandTaiwanSingaporeWithout pensionPhilippines ThailandTaiwanSingaporeUrbanMale1.001.221.17.901.00 1.501.371.62Female1.00 1.071.13.971.00 1.141.261.32RuralMale1.00 1.401.45na1.001.882.01naFemale1.001.141.32na1.00 1.151.33na Relative to the Philippines all the comparisons except Singaporeans with pensions indicate a higher probability of retirement, confirming the high- level of labor force activity in the Philippines observed in Tables 7.2 and 7.3. In contrast, Singaporean men and women without a pension show the high est probability of retirement among the urban elderly in the four countries. Contrary to the impression from the gross measures of labor force participa tion, Taiwanese urban men are somewhat less likely to be retired than urban Thai men, whether or not they have a pension. Taiwanese women and rural Taiwanese men on the other hand are more likely to be retired than their Thai counterparts. In general, controlling for the other factors results in relatively small differences across countries in the probabilities of retirement for women in comparison to men. It should be kept in mind that the country differences shown above and in Table 7.9 are conditional on the average values and distributions de rived from the pooled data. For individuals with a specific health, education, or family status, the differentials across countries might shift. The reasons for these distinct country effects cannot be determined without more re search. They may reflect in part individual-level variables omitted from the model or aspects of each country's level of economic development, policies, or culture. With only four countries, it is not possible to test for specific national characteristics that might be accounting for country differences. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Work, Retirement, and Leisure 265 Table 7.9. Predicted Probabilities of Retirement by Urban-Rural Residence, Gender, Pension Status and Country na - not applicable With pension/severancepay from current orprior jobsPhilippinesThailandTaiwanSingaporeWithout pensionPhilippinesThailandTaiwanSingaporeUrbanMen.736.901.864.666.522.782.713.845Women.819.880.927.796.675.771.853.894RuralMen.558.782.807na.277.522.559naWomen.628.718.829na.620.711.824na Note: Probabilities are calculated after setting the other independent variables at their mean values derived from the pooled sample. Attitudes and Preferences Related to Work and Retirement Reasons Given for Retirement In the preceding sections we have inferred some of the dynamics for work and retirement by examining the relative role of a number of individual, fam ily, and job-related characteristics. Another approach to understanding the factors at work is through direct questioning of respondents about their main reason for stopping work. Questions of this type were posed to those who had worked but were not currently working in each of the four countries. Although some country-level differences in the form of question and proce dures exist, the over-all patterns shown in Table 7.10 are quite revealing. The reasons given fall primarily into four broad categories: reaching mandatory retirement age (and secondarily, job-related aspects such as layoffs or plant closings), health and/or the ability to perform the job (too old, or job too demanding), economic sufficiency, and need to take care of family members. These reasons are not mutually exclusive; for example, those who report that they stopped working mainly because they reached the mandatory retirement age may also have had sufficient economic resources at that point to retire. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 M I/) QC *Q. Q. i/1 w ^^ cx>, c S (Q Nj 1/1 t I 0l — OS O <* «s* °fl<1=3 ^SON1- oi.S i& 3£1001•Y-tGenderReamen stoppeio oo o -H Co in I C^Tt Tt ON — ' o fS ' r^— i Tf tsm ro ro r- in oo IOON NO r*S o O — * '06Tt flNO m O oo — ' o i OO tN ro — O NO 'NOrs •* tsoo in CO rO o in r^ ONOe^TNOfiOOOO.n— oor-CNOvOfiroON r^* if^ ro O in O oo''OON i0 O iO O s-* ^^ OtsQr^COots-^COsl fc , o oo<3 "i_ }J E c1 S 1 1 1 1 1^8e£!i ^B^^H-1^2 1 in e § S 2^ E -g i e in 13ilhtiii!§.' S SlslllsfSlfS |llsli!llt*^ CN O — • NO OO i O— NO NO — i O tN ' CNTt rO ~O Tt ON — r^t tN i *-i-^ -H -^ — H o fi ' —eNO r- Tt — ON •* i ON-H -H ON — ' O tN ' CN in ts —iino-^r^-HtsooNOTt — omtsr^^^NO -^ ^ CO oo ON ^ iOTt ON OO ^5 CJ NO Tt fel•* — i — •Tt — ovO^m-^NO-t. rr.. — — — r-^ in '0 CO — * tN — •V **c „ u 2 &Qe-I w — >•* ^ eO ^-.fa c o E -- S oitjlifliUr-OSuS^1" JJ eS 2 2 u u &= S3 Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 - 1nooi/5'9*.a1 i1^ g5 M^ <5r^ON1- 8^* ^<pvO100cReamen stopped WmkCNmooin00m00OininSov9rrt-9M5ucMale Reached mandatory retr- :'0 'm iin 'O 'vo 'l ,oo 'inP im- i5 'T>C 0sealth problem, scide Job too demandin-/tooCOooomO^t-**002^"u£pTake care of family meoo Ooo — •oo mvO — 'oo t^O Oi °°.I intNI init:oTJ3^3uu °U T5r fljsave eno._ money to Crtpany/b.siness clos: Pi ^^i COi r-I e, ^^i 1^,i m^»i ^.—c2O05i2 =S. e£!.§1 £6NOON^vOininrn^t*ies,_— ioSoncucFemale Reshed mandatory retP i00 'ON iin '00 iTt •— im§ 'o :ON 'w —C Osealth problem, scide Job too demandin-/tooinvOO•*•««00o>-pin ro"wJ^pTake sare of family mem r^ •CN e imNO CN iO O 'OO CN iJo ~i ^ ii o iCN: P ii 00 'IBo-QJ2T3>O 00± E ,c4> -rt^ *^f^" 0 osave eno.-h money to Crtpany/b.sinss clos Family opposed to R wONCNtsm^00CNON5Other reason •o X. t5 tf) C/3 Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 268 The Well-Being of the Elderly in Asia The patterns of response differ sharply across countries and by gender. For men, health-related reasons (health problems or the inability to fulfill the de mands of the job) account for about 74 percent of the responses in the Phil ippines and Thailand, with almost 18 percent mentioning mandatory retire ment ages, lay-offs, or plant closings. In Taiwan, only 58 percent of the men cite a health-related reason, while about half cite mandatory retirement or another job- related reason. (In Taiwan multiple responses were possible.) In Singapore by contrast, more than 83 percent of the men cited mandatory retirement age as the reason for stopping work and almost 10 percent more mentioned having sufficient funds for retirement, most likely stemming from the accumulation in the Central Provident Fund. For women the mix of reasons is quite different and some new ele ments appear. In the Philippines, Taiwan, and Thailand, health-related rea sons (health problems or job too demanding) remain substantial, particularly in Thailand where this is reported by over 80 percent of the women. Manda tory retirement age is not cited very often in any country, but the need to take care of family members is a significant factor for about a fifth of the Philip pine women and a third of the Taiwanese women aged 60 to 69. For Singapore, the reasons given by women resemble those of men in that mandatory retire ment age and sufficiency of economic resources predominate, with women more often citing the latter reason than the men. The differences by gender and by country are more distinctive than those by age. Generally, the reasons for retirement given by those aged 70- plus follow the pattern of those aged 60 to 69. In Taiwan, where multiple responses were permitted, men and women over 70 mentioned both manda tory retirement age and health problems more frequently than their younger counterparts. The need to take care of family members is more frequently given by women 60-69 in the Philippines and Taiwan than by those who are older. Table 7.11 compares the reasons for ceasing work given in the three Southeast Asian countries with the reasons given in the earlier ASEAN round of surveys (Chen and Jones, 1989). In the roughly ten-year period, the re ported reasons have remained largely stable, with some notable shifts. In the Philippines, a health-related constraint (including poor health and the related dimensions of too old/job too demanding) is the most prevalent reason in both surveys, but somewhat surprisingly the proportion of men who cite reaching retirement age as the reason drops substantially during the period. In Singapore, the movement is in the other direction, with sharp increases in the proportion citing retirement age or sufficient financial means—as might be expected with the greater penasiveness of the Central Provident Fund (see Chapter 3)—and concomitant declines in the proportion citing health-related reasons Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Table 7.11. Percent Distribution of Reasons Given for Retirement by Those Not Working, Mid-1980s vs. Mid-1990s Note: na (not available) included with other reasons. PhilippinesReasonsReached Retirement AgePoor HealthToo Old / Job Too Demanding /Worked EnoughDomestic Reasons /Family, Children Oppose /Take Care of FamilySufficient Financial MeansOther ReasonsTotalThailandReasonsReached Retirement AgePoor HealthToo Old / Job Too Demanding /Worked EnoughDomestic Reasons /Family, Children Oppose /Take Care of FamilySufficient Financial MeansOther ReasonsTotalSingaporeReasonsReached Retirement AgePoor HealthToo Old / Job Too Demanding /Worked EnoughDomestic Reasons /Family, Children Oppose /Take Care of FamilySufficient Financial MeansOther ReasonsTotalMale19844246na5na71001996116564015100Male19861074nanana1610019951445291011100Male1986452361011510019958360092100Female1984463na14na1910019965411021221100Female1986175nanana241001995146361115100Female198614131032823100199550807323100 Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 270 The Well-Being of the Elderly in Asia or domestic and family considerations. In Thailand the earlier survey had relatively few comparable categories but the proportions citing health-related reasons are similar for men and women in both years, and dominate all other factors. Overall, Tables 7.10 and 7.11 reveal that financial incentives and retirement age regulations are the main factors among Singaporean elderly, while health and to a lesser extent family considerations are the main driving forces in the Philippines and Thailand, with Taiwan occupying an intermedi ate position. It should be kept in mind that reasons given for retirement often involve a certain degree of rationalization and the responses must be weighed against other evidence. Attitudes and Preferences Related to Retirement At the outset we noted that the well-being of older persons with regard to work and leisure can be gauged in part by the degree of concordance between their preferences and their actual activities. To this end it is relevant to measure the older population's attitudes about the conditions under which it is appropriate to stop working. Because the questions tapping attitudes about work versus retirement varied across countries to capture local ways of thinking, we present the actual questions asked as well as the responses in Table 7.12. In Thailand and Taiwan the question of continued work is conditional on good health, which is not explicitly mentioned in the Philippine version. In Singapore, the question is phrased in terms of a set age because regulating retirement age had been common there for some years. For all countries, the results point strongly toward the attitude of working as long as one is able. The distributions for Thailand, Taiwan, and the Philippines are very similar, with 68 to 85 percent expressing this preference across the three countries. These preferences seem in accord "with the experience of the currently retired in these countries as indicated in Tables 7.10 and 7.11, where health and related reasons for stopping work predominate. But, even in Singapore, where the Central Provident Fund and mandatory retirement ages have been in place for some time, over 80 percent disagree or are not sure that the cessation of work should be mandated by age. It should be stressed that these are generalized attitudes that may be at considerable variance from behavior, and which may grow discordant over time. In the United States, for example, preferences for working at older ages have continued to be widely expressed even as the retirement age declines. The General Social Survey (National Opinion Research Center, 1996) asked a sample of Americans currently working or temporarily not working for the period 1973 to 1996: "If you were to get enough money to live as comfortably Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Work, Retirement, and Leisure 271 as you would like for the rest of your life, would you continue to work or would you stop working?" From two-thirds to three-quarters of those under age 60 say they would continue working, and about half to two-thirds of those over age 60 respond the same way. (For further discussion of subjective factors in retirement, see Inkeles and Usui, 1988.) Plans for Retirement or for Returning to Work Additional insight into older respondents' attitudes and preferences can be gleaned from their expressed plans to stop working (if currently working) or return to work if currently retired. Respondents who were currently working in the Philippines, Taiwan, and Thailand were asked if they planned to stop working, with those in the first two countries allowed to qualify their answer. Four-fifths of those currently working in the Philippines and two-thirds in Taiwan said that stopping work depended on circumstances, and about 90 percent of the time the contingency cited was health. The remaining respon dents in each country were about equally divided between an unequivocal "yes" and an unequivocal "no." In Thailand, where a "depends" option was not available, 44 percent said they planned to stop working and 53 said they did not, with the remainder unsure or not replying. Overall these responses indicate that a clear plan to stop working at a certain age is not salient among this generation of older workers. A majority see themselves working indefi nitely, stopping only if their health status requires it. Those who were currently retired in these three countries were asked about plans or interest in returning to work. In Taiwan, where the question was in terms of planning to go back to work, 96 percent said "no." In the Philippines, where they were asked if they had any "interest" in returning to work, 91 percent said "no." In Thailand, respondents who said they were "fully retired"—a smaller base than used in Table 7.11 above—were asked if they wanted to work. If they responded positively, they were asked if they were physically able. The responses were: Yes, want to and able 3 percent Yes, want to but not able 54 percent Do not want to 43 percent This additional detail from Thailand suggests that many of the "no" responses in the Philippines and Taiwan were prompted by health conditions and the ability to perform the often-demanding duties of the jobs to which the elderly might aspire. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Table 7.12. Percent Distribution of Opinions about Stopping Work in Old Age among Adults 64 Years and Older' Question Posed in Survey, by Country and Year Philippines (1996): "Some people feel that an elderly person should continue to work as long as they can. Others feel that an elderly person should stop working completely at a certain age. What do you think?" Should stop working completely at a certain age Should continue to work for as long as s/he can Depends on health No opinion/ NI Thailand (1995): "In your opinion, if an elderly is healthy and strong, should s/he stop working completely or continue working during old age?" Should stop Should continue Depends No answer/Missing Taiwan (1993): "Some people feel that as long as an elderly person is healthy, they can continue to work. People should 'live until an old age, and work until an old age', doesn't matter the age. Some people feel that an elderly person should stop working completely, and to enjoy family life at home. What do you think?" Should stop working completely Should continue as long as health condition permits Other Missing Singapore (1995): "In your opinion, do you think there should be an age set for a person to stop working (i.e., retire)?" Yes No Don't know/Not sure a Age 64 years and above selected for comparability with 1993 Taiwan survey. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Work, Retirement, and Leisure 273 To what extent are these attitudes and preferences consonant with observed behaviors and what do they reveal about the well-being of the elderly in terms of work and leisure? Although labor force participation rates at the older ages tend to be high in comparison with industrialized countries, they are not unduly different, and collectively they appear in accord with expressed desires to continue working as long as health permits. This pattern is also confirmed both by the emphasis placed on health and related dimensions as reasons for retirement in all countries except Singapore, and by the regression results in each country. Consequently, retirees show little inclination or ability to return to work. Similarly, responses to plans to stop working are explicitly contingent on health and related factors in the Philippines and Taiwan, and are likely to be implicitly considered as well in Thailand. The Process of Retirement To this point we have examined the current status of elderly respondents with regard to various facets of work. In this section we take advantage of the panel data in Taiwan and the strong retrospective data there and in the Philip pines to review several dimensions of retirement as a process over time. Quinn and Burkhauser (1994) note that in the U.S. part-time work rises with age and is a significant factor in the employment picture of older men and especially older women. The proportion of older men and women working part-time has been rising in the U.S. and in a number of other developed economies. In these countries most of the part-time work at the older ages is voluntary; in many cases it is part of an exit pattern from a full-time career job to withdrawal from the labor force. This pattern has been the subject of several studies (see Quinn and Burkhauser, 1994, p. 71-72). Although less is known about the process of retirement in developing countries, our retro spective data on work histories plus the panel information in Taiwan allow us to sketch a few broad dimensions. Table 7.3 showed the percentage currently working part-time. Ex pressed as a proportion of those currently working, the figures for those aged 60 and older are as follows: PhilippinesSingaporeTaiwanPercentage Working Part TimeMale33.822.211.5Female34.945.720.4 Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 274 The Well-Being of the Elderly in Asia Although strictly comparable data does not exist for Thailand, related information in the survey suggests that the proportions working part-time there might resemble the levels in the Philippines. These percentages show that part-time work is a substantial aspect of older age employment in these countries and should be considered in tracing trends and developing policy. In the Philippines those working part-time were asked if they ever worked full-time, and more than one-fifth of the men and one-ninth of the women replied affirmatively. Schoenbaum (1995, Table 2.2), who analyzed the job history data from the 1989 Taiwan survey, found that among respondents who currently worked part-time or seasonally, more than three-quarters of men and about three-fifths of women reported working full-time in the past. Whether these patterns and changes reflect voluntary labor supply decisions or constraints on the availability of jobs cannot be determined from these data and must await further investigation. We can also trace in part the retirement process of those already re tired in the Philippines and Taiwan. In the Philippines those retired within the past year were asked the number of hours worked on their last job, and if less than 40 hours, whether they had previously worked full-time. For this select group of retirees, 37 percent of both men and women reported working part- time at their last job, similar to the proportion reported above for those cur rently working. Of those who had been working part-time, only 14 percent of the men and 4 percent of the women said they had worked full-time in the past. It would thus appear that for many Philippine workers, as well as retir ees, part-time employment, perhaps in more than one job, has been a way of life under recent economic conditions rather than an exit strategy in the course of retirement. More detailed data, of a somewhat different nature, exist about the retirement process in Taiwan. Schoenbaum (1995, Tables 2.3a and 2.3b) used the somewhat limited retrospective histories in the 1989 Taiwan panel to examine transitions from a worker's major sector (class of employment) or occupation to current sector or occupation, including in each case being out of the labor force at the latter date. He found that a very high proportion of respondents exited the labor force, and only a small proportion had moved from their major sector or occupation to a different one. For example, of the entire sample of those who ever worked, only 9 percent of men and 5 percent of women moved to a job in a different sector. Of those currently retired, 11 percent of men and 8 percent of women changed sector of employment between their major and last job, and 22 percent of men and 9 percent of women changed occupational category. Though these data cannot fully ac count for intermediate jobs within the same occupation, they do suggest limited work mobility before retirement. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Work, Retirement, and Leisure 275 Shih (1997) provides other insights into the process of retirement via her analyses of the 1993 panel data from Taiwan, which contain additional information on occupational history, including job change within occupa tions. In tracing the employment trajectories of respondents after age 50, Shih found that 26 percent of men and 8 percent of women held more than one job after age 50. Job change was even more common for Main lander men than their Taiwanese counterparts: about half of them reported at least one job change after age 50, a differential that may be explained by their much greater likelihood of government employment and pension eligibility. Defining retirement as being out of the labor force for 12 months or more, Shih found that 11 percent of men and 3 percent of women returned to work after retiring, again with a sharp contrast between Mainlander and Taiwanese men (21 percent versus 5 percent). When these analyses are further restricted to look at the non-farm sector only, Shih concludes that the job-change rates of older Taiwanese approach those found in the United States. Though the data available across the four countries are limited, they do suggest that part-time employment is non-trivial among older workers, and that the countries vary in the degree to which job changes and re-entry after retirement occur, with Taiwan at least displaying a number of the job- mobility features found in the more industrialized countries. Leisure Activities of the Elderly At the outset of the chapter we noted that the concepts of work, retirement, and leisure, and the distinctions among them, are likely to be influenced by the area's underlying economic structure, level of economic development, and numerous cultural elements. Studies of leisure often point to the differ ences in its nature and organization across societies according to cultural, historical, and economic dimensions. (See, for example, Kelly, 1982, Chap ters 6 and 7, for an overview and Shaw, 1994; Mercado, 1974; and Linhart, 1988, for specific countries in Asia.) The strong cultural base associated with leisure may become embodied in language so that terms signifying work in one culture may signify leisure in another context, complicating cross-cul tural comparisons. (See Blair, 1991, for a terminological and language ex ample.) Although common aspects of mass media and mass culture have permeated most countries, leading to greater uniformity in the concepts and range of activities constituting leisure, considerable variation persists. As noted, we found generally strong preferences for work among older respondents, but we also found substantial proportions retired, primarily Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 276 The Well-Being of the Elderly in Asia because of poor health or factors arising from their job (mandatory retirement, lay-offs, etc.). Among those retired, we found little inclination to return to work, in part out of awareness of their limited ability to do so. Assessment of well-being requires some attention to the leisure activities of this older population, both those working and those not working. Leisure has been conceptualized and measured in a number of ways. Veblen's simple definition of leisure as the "nonproductive consumption of time" (Veblen, 1899, cited in Kelly, 1982) placed the emphasis on time spent away from subsistence activities, or more positively, the discretionary use of time (Kelly, 1982; Murphy, 1974). Others have emphasized the activities them selves, identifying the sports, hobbies, events, and social interactions that seem to capture leisure in a particular culture. Still others have focused on the function of leisure to the individual as purposive behavior designed for self- enhancement. This is captured by Dumazedier's definition Leisure is activity—apart from the obligations of work, family, and society—to which the individual turns at will, for either relaxation, diversion, or broadening his knowledge and his spontaneous social participation, the free exercise of his creative capacity. (Dumazedier, 1967, quoted in Kelly, 1982.) From this perspective, leisure is distinct not only from remunerative work, but from activities arising from obligations to family, friends, or church. The questionnaires used in the four countries measure leisure among the elderly mainly by assessing their activities, ranging from the more passive to the more demanding in terms of physical activity or social involvement. The activities fall into five major categories: reading newspapers and books; watching television or listening to the radio; involvement in sports, regular physical exercise, or hobbies—like gardening—with a high physical activity level; participation in religious services or prayer; and involvement in a variety of social clubs and community organizations. Here we do not include visiting with family as a leisure activity per se, rather including it in Chapter 6 with intergenerational relations because it is such an important dimension for many elderly in terms of both time spent and emotional support. Table 7.13 summarizes participation in leisure activities for those age 60 and older, defining in the notes the activities included for each country. The most pervasive form of leisure reported by all elderly is regular television viewing or radio listening, reported by over 80 percent of respondents in Thailand and well over 90 percent in Taiwan and the Philippines. (This activ ity was not asked about in Singapore.) It is of interest to note, given the fairly low educational levels among the current elderly, that from a fifth to a third also report reading newspapers and books at least once a week. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Table 7.13. Percentage of Adults Age 60 Years and Older Engaged in Leisure Activities by Work Status, Gender, Age, and Self- Reported Health Status Type of Leisure ActivityTotalReading newspapers/books"Radio/TV3Physical activity IPhysical activity IICReligious activityClubseCurrently workingReading newspapers/booksRadio/TVPhysical activity IPhysical activity IIReligious activityClubsRetiredReading newspapers/booksRadio/TVPhysical activity IPhysical activity IIReligious activityClubsNever workedReading newspapers/booksRadioA-VPhysical activity IPhysical activity IIReligious activityClubsPhilippines289730309528279632*32*9531*319731319528229721219820Thailand20838858915923**90**89**67**92*66**187887529155338674468533Taiwan35945811334544**97**49**13**29**55**37946213334616925664031Singaporen.a.n.a.n.a.n.a.6627n.a.n.a.n.a.n.a.6525**n.a.n.a.n.a.n.a.6530n.a.n.a.n.a.n.a.6822 continued Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Table 7.13. continued Type of Leisure ActivityMenReading newspapers/booksRadio/TVPhysical activity IPhysical activity IIReligious activityClubsWomenReading newspapers/booksRadio/TVPhysical activity IPhysical activity IIReligious activityClubsAge 60-64Reading newspapers/booksRadio/TVPhysical activity IPhysical activity IIReligious activityClubsAge 65-74Reading newspapers/booksRadio/TVPhysical activity IPhysical activity IIReligious activityClubsPhilippines3196**34**34**92**2926982727982733*9734**34**9528309732329630Thailand30**88**90**62**90**63**12798555935523**88**89**62**92**65**218490609362Taiwan53**96**63**16**27**53**1392526403533**96**50**13**3543*409563133347Singaporen.a.a a.n.a.n.a.63**30**n.a.n.a.n.a.n.a.6824n.a.n.a.n.a.n.a.71**26n.a.n.a.n.a.n.a.6827 Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Table 7.13. continued X2: * .01 < p < .05 ** p < .01 Type of Leisure ActivityAge 75+Reading newspapers/booksRadio/TVPhysical activity IPhysical activity IIReligious activityClubsVery good healthReading newspapers/booksRadio/TVPhysical activity IPhysical activity IIReligious activityClubsGood healthReading newspapers/booksRadio/TVPhysical activity IPhysical activity IIReligious activityClubsPoor healthReading newspapers/booksRadio/TVPhysical activity IPhysical activity IIReligious activityClubsPhilippines18962121942636**95**37*37*94*35299830309727169524249328Thailand13718246874328**88**91**67**93**62188289579360117871428662Taiwan2489576304250**98**68**18**33**52**2996579364515793231928Singaporen.a.n.a.n.a.n.a.5527n.a.n.a.n.a.n.a.75**36**n.a.n.a.n.a.n.a.6926n.a.n.a.n.a.n.a.4423 Note: Chi-square results represent differences across categories. Asterisks, where applicable, only show first category listed. See Notes for Table 7.13 at the end of this chapter. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 280 The Well-Being of the Elderly in Asia Various forms of religious activity are reported by over 90 percent of the elderly in Thailand and the Philippines, equaling or exceeding radio/TV involvement, while two-thirds of respondents in Singapore and one-third in Taiwan indicate religious activities. Two measures of physical activity were developed, a broad definition that includes "taking walks at least once a week" in the Taiwan and Thailand survey, and a narrower definition that excludes this activity for comparability to the Philippines survey, which did not explic itly ask about "walking." Under either definition there is considerable varia tion across countries. For the narrower definition (Physical activity II in Table 7.13), the proportions range from 11 percent in Taiwan, to 30 percent in the Philippines, to 58 percent in Thailand. When "taking walks at least once a week" is added as a physical activity in Taiwan and Thailand, the overall percentages rise to 58 percent and 88 percent respectively, putting physical activity on a par with radio/TV and religious activities in Thailand. Participation in social clubs and community organizations is also preva lent among the elderly, and variable across countries, as might be expected given differences in local infrastructures and cultural traditions. High partici pation is recorded in Thailand (59 percent) and Taiwan (45 percent), but only about a quarter of the elderly are so involved in the Philippines and Singapore. Overall, the elderly in all the countries report engagement in a variety of leisure and recreational activities, with each country displaying a some what distinct profile. In Thailand, relatively high levels are reported for all types of activities except reading; in the Philippines, radio and television and religious activities predominate; and in Taiwan, radio and television and social clubs are most popular, if a narrower definition of physical activity is used. Table 7.13 also displays these levels of participation by several factors most likely to affect them: work and retirement status, gender, age, and health. These characteristics clearly overlap one another to a great extent. For ex ample, retirees tend to be older and in poorer health than those currently working, making any bivariate differences likely the result of multiple fac tors. At a first approximation, the overall country profiles of activities de scribed above hold within each of the sub-groups shown in Table 7.13, with no sharp differentials across groups within countries. Whether retirement leads to heightened recreational activity is highly relevant to this chapter's focus. Although those retired will have more time, they may be less able or less inclined to participate in more activities. These competing influences cannot be sorted out from cross-sectional data, since cohort differences and self-selection intervene. We use our panel data below to study changes in behavior. Table 7.13 reveals generally little difference in level of leisure activities between those currently working and those retired, with the significant changes mainly in the direction of lower levels of activity Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Work, Retirement, and Leisure 281 among the retired. Those who never worked, a relatively small group, and mostly females, show a somewhat distinct pattern relative to the other two Gender differences, though not large, follow expected social and cul tural traditions. Older women show lower levels of reading because of their higher levels of illiteracy, and they are less likely to engage in physical activi ties or social clubs, but show higher participation rates in religious activities. The differences between those aged 60-64 and 65-74 are minimal, but the level of leisure activities does diminish moderately among those 75 and older. Leisure activities also tend to decline with declining health status, with the drop-off most noticeable among those categorized as in "poor health." The items shown in Table 7.13 do not exhaust the range of activities often subsumed under leisure, and in Taiwan and the Philippines attention was also given to social activities such as attending cultural events or going out to eat with friends and family. The surveys in Thailand and Taiwan also inquired about travel, with Taiwan identifying overnight and overseas trips. The range of supplementary items covered and the proportions in volved are shown below: Percent of Those Age 60+ Participating in Various Social Activities Measurement definitions: Social activity ActivitySocial activitiesCharity workChat with relatives, friends or neighbors/drink tea sociallyTravelPhilippines23%--„—Thailand—--_27%Taiwan15%4%60%41% Philippines: Attended social activities such as getting together with friends, family, neighbors, going out to eat, walking for pleasure, went to movies, or gambled with others at least once per week. Taiwan: Participated in group activities (e.g., chess, cards, singing groups, dancing, tai-chi, karaoke), attended concerts, or went on local tours or day- trips at least once per week. Travel: Thailand: Traveled or went sightseeing in past year. Taiwan: Went to places outside local areas for tours, on overnight tnps, to Mainland China or overseas in past year. Note: -- means not available. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 282 The Well-Being of the Elderly in Asia These data indicate fairly frequent involvement in the Philippines (23 percent) and Taiwan (15 percent) with what might be termed semi-struc tured social activities. In addition, a high proportion of Taiwanese (60 per cent) report more informal types of interaction with friends and neighbors. Only 4 percent in Taiwan are explicitly involved in charitable work, although some of the club membership and activities may have this as a component. It will be of interest to monitor trends in this kind of activity as countries give more emphasis to involving the elderly (and others) in volunteer activities. Travel is an important activity with 41 percent in Taiwan reporting travel outside their local area over the course of the year and over one-quarter of those in Thailand also reporting either local or longer distance travel. The foregoing picture of the older population's involvement in leisure activities, especially when coupled with the material in Chapter 6 on level of family visits, suggests that the elderly are quite actively engaged with others and in the world around them. This observation, however, comes with sev eral cautions. First, our analyses do not provide insights on the degree to which these various activities occupy the "free" time of the respondents, or, more important, on the extent to which respondents derive meaning or plea sure from participation in them. Are these activities undertaken as substitutes for other preferred ways of spending time, or do they represent the older persons' priorities? Chapters 9 and 11 touch on the level of loneliness and isolation experienced by the elderly and indicate that these are far from trivial, despite the activity levels displayed in Table 7.13. Further analyses on the relation between activity and emotional well-being on several dimensions re main to be undertaken. Also, it should also be noted that the analysis represented by Table 7.13 is cross-sectional in nature and tells us relatively little about how the level of leisure activities changes as a cohort moves from work to retirement, or as older people experience some physical limitations. The panel data from Taiwan permit some insights into these important transition issues and we illustrate the potential by investigating whether those who retire change their level of activity. A priori one might expect an increase in leisure activity with retirement, as a result of more free time, but this might be countered by increasing age and the fact that retirement is often prompted by ill health, as shown in Table 7.10. We compared the level of activity reported by the panel in 1989 (when respondents were 60 or older) to that reported in 1996 (when they were 67 or older) according to their work/retirement status at each point. Those who were working in 1989 but retired in 1996 reported modest increases in their involvement with clubs and associations, religious activity, and physical activity, Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Work, Retirement, and Leisure 283 and a slight decline in reading—changes that were similar to those reported by respondents who were working in both 1989 and 1996. This latter group showed a particularly large increase in the proportion involved in social and community organizations over the seven-year period. On balance little evidence indicates that retirement per se led to increased leisure activity, though this initial analysis must be tested further with additional controls. Those who were retired both in 1989 and 1996 reported slight declines in reading and engaging in physical activity (using the physical activity measure from Table 7.13), but no changes in the other types of activities. Overall, the use of the panel reveals little change in levels of leisure activity over the seven-year period as a result either of shifting from work to retirement or of aging per se (as observed in those who were retired at both points). It is possible that period effects, such as the creation of more com munity associations or more elderly outreach organizations, promoted greater involvement in social clubs and associations—a surmise that could be ex plored using independent information on the density of such organizations. Conclusions This chapter has used both aggregate and individual level data to develop a comparative picture of the labor force patterns of the older population in each country under study, and the factors affecting their labor force participation. In addition, recent changes that might signal future trends are assessed. Be cause the majority of those 60 and older in each country are no longer work ing, we also look at the level and type of leisure activity they are engaged in, and what is known about their preferences for work versus retirement. At the aggregate level, the older population in the four countries show two rather distinct profiles. The Philippines and Thailand have somewhat higher labor force participation rates for men, and much higher rates for women, than Singapore and Taiwan (Table 7.2). In the latter two countries, substantial involvement of married women in the labor force is a relatively recent phenomenon, as shown in Figure 7.1, which traces activity over the life course from 1970 to 1990. These data, as well as comparisons with similar data from the 1980s (Figure 7.2), indicate very little tendency toward earlier retirement—which is contrary to the strong trends observed in most industrialized countries over the past two or more decades. The current lev els of labor force activity in the four countries under study overlap the levels observed among the more industrialized countries, though they do not display similar trends. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 284 The Well-Being of the Elderly in Asia The survey data from each country were used to investigate the effect of a set of factors representing preferences, needs, and opportunities for employment, paralleling models used in more industrialized settings, as well as the developing world (Tables 7.6 and 7.7). Beyond the expected strong relationship with age, the logistic regressions in each country revealed that poor health was a very important factor leading to retirement in each country for both men and women (except Philippine women). Urban residence and involvement in the non-agricultural, non-family sector also tended to promote retirement; education had generally small and inconsistent effects. Pension income or coverage had strong effects, but not in a similar direction across all four countries. In Taiwan and Thailand, having pension income or coverage was a strong positive factor in the likelihood of being retired among both men and women. It was not a significant factor in the Philippines, and it was a negative factor in Singapore—that is, Singaporeans with pension coverage or income were more likely to be working than those without such benefits. We traced these differences to the breadth of the Singapore pension program, the skills of those covered, and the relatively early age of retirement, although further investigation of the precise dynam ics is clearly warranted. We found that family factors also played a role in retirement. Men and women who had a spouse working were much less likely to be retired (ex cept in the Philippines) than the non-married, suggesting that joint decisions often come into play. Generally, those living with a married child were more likely to be retired, though the cause-and-effect relationship here is some what ambiguous, as were those with greater numbers of living children. To more formally pursue the comparative analysis, the survey data were pooled to examine the presence of distinct country effects, after con trolling for the other determinants, with an additional interaction term added to represent the distinct effects of pension coverage in Singapore (Table 7.8). The logistic regressions indicated significant country effects, which are cap tured in Table 7.9 for those with and without pension coverage. Among urban residents with pension coverage, Philippine and Singaporean men and women are less likely to be retired than those in Taiwan and Thailand. But among urban residents without pension coverage, Singaporean men and women show the highest probability of retirement, while Philippine men and women show the least, after taking into account other factors. The higher involvement of Philippine elderly in the labor force also appears clearly in the rural areas. The pooled multivariate analysis confirms the distinct labor force pro file of the Philippines and the very sharp effects of pension coverage in Singapore; but it diminishes, and in some cases reverses, the differences between Taiwan and Thailand. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Work, Retirement, and Leisure 285 The underlying theoretical focus on multiple dimensions of well-being requires that emphasis be given to the older population's preferences about work and leisure in studying labor force activity. Relevant to this goal is studying the reasons workers give for retirement, as well as their expressed desires about retirement. Tables 7.10-7.12 touch on these issues, and reveal that among the current elderly, retirement as a voluntary and desirable stage of life is far from normative. In the Philippines, Thailand, and Taiwan the major reasons for retirement center around health problems or the job being too demanding for the respondent's age or capacity. In Singapore, and to a certain extent in Taiwan, the emphasis is on reaching a mandatory retirement age, though the trend data for Singapore (Table 7.11) reveal an increase in the proportion who retire because they are financially able to do so. These re sponses coupled with the prominence of the attitude that people should work as long as they are able (Thailand, Philippines, and Taiwan) and doubt about the desirability of a mandatory retirement age (Singapore), suggest that pref erences among the current elderly lean strongly toward continued employ ment to the extent possible. As Quinn and Burkhauser (1994) note, the literature of the 1940s and 50s described retirement largely as an involuntary process driven by poor health, layoffs, and mandatory retirement policies. In the United States, a trend toward more voluntary retirement appeared in the early 1960s, but it was not until 1982 that more retirements were voluntary than involuntary, according to a survey of Social Security beneficiaries. From this perspective, the current older generation in the four countries appears to be at the stage that characterized the U.S. and other industrialized nations 30 to 40 years ago when retirement was less a matter of choice and more a matter of involuntary factors centering around health and conditions of employment. In the coming years, it is reasonable to expect a rapid transformation of retirement toward a more voluntary process in these countries for a num ber of reasons. One set of factors stems from the expansions and rationaliza tions of the social security systems ongoing in each country (see Chapter 3), which should increase the proportion of workers with coverage and the amount of the benefits. The second broad force for change arises from the new cohorts of elderly in coming years who will likely be better educated, more often employed in the modern sectors of the economy, and in better health, and will likely have different expectations and preferences about how to spend the later years of life. The move toward retirement as a more voluntary decision that bal ances financial and other incentives to withdraw from employment against incentives and benefits of continued work does not necessarily presage a trend toward earlier ages at retirement in the four countries under study, or Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 286 The Well-Being of the Elderly in Asia others similarly situated. The trend data available for the four countries (Fig ures 7.1 and 7.2) show very little change in labor force participation rates at the older ages, in contrast to those observed in OECD countries. In addition, the Asian countries which have historically mandated relatively early-age of retirement for government employees and those in the modern sector, are taking steps to increase the mandatory age of retirement and to develop pro grams that induce firms to retain older workers and help older workers main tain marketable skills. Policymakers in many of the newly industrialized and developing countries are aware of the financial incentives in the pension and social security systems of the more industrialized countries that contributed to the sharp trend toward early retirement. They can benefit from this expe rience in designing or modifying their own systems. Although many older workers in the four countries reported leaving the labor force because of poor health or job factors, they do not report inactivity. The overview of leisure activities in Table 7.13 shows that a high percentage of the older population is engaged in one or another form of lei sure including more physical and social activities as well as more passive and solitary pursuits. Surprisingly little variation was found in the level of involve ment across basic categories of gender, health, and labor force activity. The longitudinal analysis from Taiwan shows that leisure activity does not change greatly with retirement, suggesting that participation in these activities tends to be established earlier and maintained to the extent possible as a person ages. These conjectures however need to be tested further with more refined data that account for the amount of time devoted to various activities as well as the degree to which leisure activity supports the emotional health of the elder ry. Several policy challenges flow from the current and emerging pattern of labor force participation in tandem with the other aspects of well-being. The high proportion of the current elderly who report they stopped working because of poor health or their inability to meet job demands indicates the need for policymakers to assure that health and related support facilities are adequate and accessible and that recreational facilities are sufficient for those with special needs. To the extent that current retirees have stopped work without adequate financial resources, policymakers should assure that the coverage and benefit levels under social security and pension programs are adequate in conjunction with familial and other sources of support. At the same time it is necessary to look ahead at those in their prime working years. It will be necessary to fine tune retirement programs so that they provide adequate income in the years after work but not inducements to stop work too early, thereby generating higher costs and lost productivity Policymakers should also promote programs that enhance the health and Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Work, Retirement, and Leisure 287 marketable skills of adult workers, increasing the number who will be able and willing to work at older ages. Concomitantly, policymakers should rec ognize that the older population's recreation and leisure interests, particularly in combination with greater incomes and longer life expectancies, can be a strong stimulus for the economy if the activities developed are accessible and appropriate to this age group. REFERENCES Agree, Emily M, and Robert L. Clark. 1991. "Labor Force Participation at Older Ages in the Western Pacific: A Microeconomic Analysis." Journal of Cross Cultural Gerontology 6:413-29. Blair, Karin. 1991. "Leisure in China. Problems of Defining the Invisible: An Attempt to See Beyond the Mirror of One's Cultural Assumptions." Loisir et societe/ Society and Leisure 14(2):343-55. Blau, David M. 1998. "Labor Force Dynamics of Older Married Couple." Journal of Labor Economics 16(3):595-629. Cain, Mead T. 1991. "The Activities of the Elderly in Rural Bangladesh." Population Studies 45:189-202. Chen, Ai Ju, and Gavin Jones. 1989. Ageing in ASEAN: Its Socio-Economic Consequences. Singapore: Institute of Southeast Asian Studies. Clark, Robert L., E. Anne York, and Richard Anker. 1999. "Economic Development and Labor Force Participation of Older Persons." Population Research and Policy Review 18:411 -432. Dharmalingam, A. 1994. "Old Age Support: Expectations and Experiences in a South Indian Village." Population Studies 48:5-19. Dumazedier, Joffre. 1967. Toward a Society of Leisure. New York: Free Press. Gallin, Bernard. 1966. Hsin Hsing, Taiwan: A Chinese Village in Change. Berkeley, CA: University of California Press. Gallin, Rita S. 1994. "The Intersection of Class and Age: Mother-in-Law/Daughter- in-Law Relations in Rural Taiwan." Journal of Cross-Cultural Gerontology 9:127-140. Gruber, Jonathan, and David Wise. 1998. "Social Security and Retirement: An International Comparison." American Economic Review, May. Gustman, Alan L., Olivia S. Mitchell, and Thomas L. Steinmeier. 1995. "Retirement Measures in the Health and Retirement Study." The Journal of Human Resources, XXX, Supplement 1995. Hayward, Mark D., Samantha Friedman, and Hsinmu Chen. 1998. "Career Trajectories and Older Men's Retirement." Journals of Gerontology, Series B: Social Sciences 53(2): S91 -S 103. Hayward, Mark D., and William R. Grady. 1990. "Work and Retirement among a Cohort of Older Men in the United States, 1966-1983." Demography27(3):337- Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 288 The Well-Being of the Elderly in Asia Hay ward, Mark D., and Wei Wang. 1993. "Retirement in Shanghai." Research on /lg//7gl5(l):3-32. Henretta, John C., Angela M. O'Rand, and Christopher G. Chan. 1993. "Joint Role Investments and Synchronization of Retirement: A Sequential Approach to Couples' Retirement Timing," Social Forces 11(4):981-1000. Holtz-Eakin, Douglas, and Timothy M. Smeeding. 1994. "Income, Wealth, and Intergenerational Economic Relations of the Aged." In Linda G. Martin and Samuel H. Preston, eds., Demography of Aging, 102-145. Washington: National Academy Press. Inkeles, Alex and Chikako Usui. 1988. "The Retirement Decision in Cross-National Perspective." In R. Ricardo-Campbell and E. P. Lazear, eds., Issues in Contemporary Retirement, pp. 273-311. Stanford, CA: Hoover Institution Press. Kelly, John R. 1982. Leisure. Englewood Cliffs, NJ: Prentice Hall. Kinsella, Kevin, and Cynthia M. Taeuber. 1993. An Aging WorldII. International Population Report Series, P-95. Washington, D.C.: Bureau of the Census. Kinsella, Kevin, and Victoria A. Velkoff. 2001. An Aging World: 2001. U.S. Census Bureau, Series P95/01-1. Washington, DC: U.S. Government Printing Office. Kohli, Martin, Martin Rein, Anne-Marie Guillemard, and Herman Von Gunsteren. 1991. Time for Retirement: Comparative Studies of Early Ex it from the Labor Force. New York: Cambridge University Press. Linhart, Sepp. 1988. "From Industrial to Postindustrial Society: Changes in Japanese Leisure-Related Values and Behavior." Journal of Japanese Studies 14(2):271- McCallum, John. 1992. "Asia Pacific Retirement: Models for Australia, Fiji, Malaysia, Philippines and Republic of Korea." Journal of Cross Cultural Gerontology 7(l):25-43. Mercado, Leonardo N. 1974. "Notes on the Filipino Philosophy of Work and Leisure." Philippine Studies 22:71 -80. Murphy, James. 1974. Concepts of Leisure: Philosophical Implications. Englewood Cliffs, N.J.: Prentice Hail. National Opinion Research Center. 1996. General Social Survey. Cumulative Codebook, 1972-1996. Ogawa, Naohiro. 1994. "Economic and Social Implications of Population Aging in Japan." In Aging in Japan, Part II, 55-87. Japan Aging Research Center, Tokyo, Japan. Quinn, Joseph F., and Richard V. Burkhauser. 1994. "Retirement and Labor Force Behavior of the Elderly." In Linda G. Martin and Samuel H. Preston, eds.,DemographyofAging,pp. 50-101. Washington, D.C.: National Academy Press. Quinn, Joseph F., Richard V. Burkhauser, and Daniel A. Myers. 1990. Passing the Torch: The Influence of Financial Incentives on Work and Retirement. Kalamazoo, MI: W. E. Upjohn Institute on Employment Research. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Work, Retirement, and Leisure 289 Raymo, James M., and Jennifer C. Cornman. 1999. "Labor Force Status Transitions at Older Ages in the Philippines, Singapore, Taiwan, and Thailand, 1970-1990." Journal ofCross-Cultural Gerontology 14:221-244. Schoenbaum, Michael L. 1995. "The Health Status and Labor Force Behavior of the Elderly in Taiwan." In Health, Longevity and Labor Force Behavior among Older Workers. Unpublished dissertation, University of Michigan. Shaw, Thomas A. 1994.'" We Like to Have Fun:' Leisure and the Discovery of the Self in Taiwan's 'New' Middle Class." Modern China 20(4):416-445. Shih, Shiau-ping R. 1997. "Private Lives within Public Constraints: Retirement Processes in Contemporary Taiwan." Unpublished dissertation, University of Michigan. Szinovacz, Maximiliane, David J. Ekerdt, and Barbara H. Vinick, eds. 1992. Families and Retirement. Newbury Park, CA: SAGE Publications, Inc. U.S. Bureau of the Census. Forthcoming. An Aging World, 1999. . Forthcoming. An Aging World, 2001. Veblen, Thorstein. 1899. The Theory of the Leisure Class. New York: Macmillan. World Bank. 1994. Averting the Old Age Crisis: Policies to Protect the Old and Promote Growth. World Bank Policy Research Report. New York: Oxford University Press. NOTES FOR TABLES Table 7.2 and Table 7.3 The questions used to ascertain current work activity for each country are as follows. Philippines We would like to know about your current work situation. At present (during the last month), would you say you are: (1) working, (2) have a job but are temporarily laid off or absent, (3) not working but looking for work, (4) helping out with family business, but not working for a living, (5) helping out in farm, but not working for a living, (6) keeping house, (7) not doing any of the above. The following recedes were used for the classification in Tables 7.2 and 7.3: 1 = currently working 2,3 = in labor force, not currently working 4,5 = working without pay 6,7 = not working Additionally, those currently working less than 40 hours/week were coded as working part-time, and those currently working 40+ hours per week were coded as working full-time. Respondents who were not currently working but had worked sometime in the past \vere identified as not working/retired, and those who had never worked as not working/never worked. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 290 The Well-Being of the Elderly in Asia Thailand E1. Within this past week did you work to support yourself or family? (1) yes, (2) no E2. (If didn't work in past week) Why did you not work within this past week? (1) retired, (2) old, (3) poor health/sick, (4) physical problem, (5) had an accident, (6) other. E3. (If didn't work in past week) Within the past year did you work to support yourself or your family? (1) yes, (2) no E13. Did/do you receive any compensation in terms of money or materials? (1) money only, (2) material only, (3) money and material, (4) did not receive compensation. E4.1. (If didn't work in past week) Have you ever worked to support yourself or family? (1) yes, (2) no E44. Do you want to work? If so, do you think you are able to? (1) want to and able to, (2) want to but not able to, (3) don't want to. The following recedes were used for the classification in Tables 7.2 and 7.3: IfEl=l and KH = 1 or 3, then R is currently working If El = OandE3 = 1 andE2 = 3,4, or 5 andE44 = 1, thenRisinLF, not currently working If E1 = 1 and E13 = 2 or 4, then R is working without pay If El = 0 and E4.1 = 1 and R is not in labor force, then R is not working/retired. If El = 0 and E4.1 = 2, then R is not working/never worked. The definition of currently working in Thailand does not include those who say they are working and yet only receive matehal goods. This is a different definition that that used in the Philippines, Singapore, and Taiwan. We opted for this distinction because the other countries began the series of questions about employment with an overall question about current work status that included categories similar to "unpaid family worker". The questionnaire in Thailand used a different strategy of asking sequential questions first about work status in general (related to "supporting oneself or one's family") and then questions about forms of compensation. The slightly different definition for current work status that we use for Thailand is an effort to identify those unpaid family workers who might have been screened out of current work status if they had been asked the more global work status question that was used in the other countries. Note: The Thai questionnaire does not include questions to ascertain whether R is currently working full or part-time. Taiwan Next we'd like to ask about your work. Are you presently working? (1) yes (includes primary or secondary profession), (2) have a job but haven't worked in a Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Work, Retirement, and Leisure 291 while, (3) help out with family farm or business, not a formal job, (4) unemployed, looking for work, (5) do housework (cooking, cleaning, shopping, childcare), (6) none of the above. The following recedes were used for the classification in Tables 7.2 and 7.3: 1 = currently working 2,4 = in labor force, not currently working 3 = working without pay 5,6 = not working Additionally, currently working Rs were asked whether they were working full or part-time, and this designation is reflected in Table 1. Respondents who were not currently working but had worked sometime in the past were identified as not working/retired, and those who had never worked as not working/never worked. Singapore Activity/occupational status: (1) working, self-employed; (2) working employee/odd job; (3) working, unpaid family worker; (4) not working but looking for work; (5) not working and not looking for work. The following recedes were used for the classification in Tables 7.2 and 7.3: 1,2 = currently working 4 = in labor force, not currently working 3 = working without pay 5 = not working Additionally, currently working Rs were asked whether they were working full or part-time, and this designation is reflected in Table 1. Respondents who were not currently working but had worked sometime in the past were identified as not working/retired, and those who had never worked as not working/never worked. Table 7.10 Tor the Philippines, combinations of responses were coded for 17 respondents. These were assigned to the above response categories as follows: 1. If both health problems and family care were reported, health problems were coded. 2. If both health problems and job demands were reported, health problems were coded. 3. If both job demands and family care were reported, family care was coded. 4. If both reached mandatory age and health problems were reported, mandatory age was coded. bFor Taiwan, respondents were allowed to give multiple responses, so percentages will not add up to 100. Tor Taiwan, this category includes all reasons related to family. Dashed (—) entries indicate that the category was not coded as one of the response categories for that country. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 292 The Well-Being of the Elderly in Asia Table 7.13 The following survey questions were used to ascertain types of leisure activities in each country. a. Mass media Philippines: 1) Read newspapers, magazines or books at least once a week 2) listen to radio or watch TV at least once a week. Thailand: 1) Within the past week, read any newspaper? 2) listen to the radio or watch any television? Taiwan: 1) Read any newspaper, books, magazines, or novels at least once a week 2) listen to radio/ tapes or watch TV/videos at least once per week. Singapore: Relevant questions not asked. b. Physical activity (1st measure) Philippines: 1) Do physical exercises such as jogging and outdoor recreational activities like golf or fishing at least once per week Thailand: 1) Walk at least once per week 2) Exercise (e.g., run, taichi, gardening, planting trees) at least once per week. Taiwan: 1) Take walks at least once a week 2) jogging, mountain climbing, playing ball or other physical exercise outside the home. Singapore: Relevant questions not asked. c. Physical activity (2nd measure) Philippines: 1) Do physical exercises such as jogging and outdoor recreational activities like golf or fishing at least once per week Thailand: 1) Exercise (e.g., run, taichi, gardening, planting trees) at least once per week. Taiwan: 1) jogging, mountain climbing, playing ball, or other physical exercise outside the home. Singapore: Relevant questions not asked. d. Religious activity Philippines: 1) Meditate or pray at home at least once a week 2) attend religious services. Thailand: 1) In the past month, offer alms to Buddhist monks (or asked someone else to do on own behalf) 2) pray or meditate at least once per week. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Work, Retirement, and Leisure 293 Taiwan: 1) Chant sutras, pray and bur n incense, go to temple to offer incense or study sutras, pray, or go to church at least once per week. Singapore: 1) Visit a church/temple/mosque in the past month? e. Social clubs Philippines: 1) Member of religious groups 2) business, professional or farm association 3) voluntary, welfare or aid group 4) political group 5) community center or social/recreational club 6) clan association 7) organization of retired elderly persons 8) other association/club. Thailand: 1) Member of a cooperative group 2) housewife group 3) elderly group 4) funeral group. Taiwan: 1) Belong to neighborhood association 2) religious association 3) farmers' association 4) political association 5) social services groups 6) village or lineage association or 7) elderly clubs. Singapore: 1) Ever used Senior Citizen's Club before? 2) Ever used Social Day Centre before? 3) visited a community centre in the past month 4) visited a Senior Citizen's Comer in the past month. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Chapter 8 ECONOMIC WELL-BEING: INSIGHTS FROM MULTIPLE MEASURES OF INCOME AND ASSETS Albert I. Hermalin, Ming-Cheng Chang, and Carol Roan As discussed in greater length in Chapter 4, a major concern in the study of population aging is the effect of demographic and socioeconomic trends on the economic well-being of the elderly. Many observers, though by no means all, have concluded that factors of development would adversely affect the status of the elderly by trapping them in more traditional and less rewarding jobs, and separating them from their families as children migrate to urban areas in pursuit of better jobs (see Cowgill, 1974 vs. Treas and Logue, 1986). Many of these speculations have their basis in the history of currently industrialized societies; relatively little direct empirical experience has as yet accumulated about the economic well-being of the elderly in developing and newly industrialized countries. Even when attention has been given to the material well-being of the older population in developing countries, the mea sures employed have centered on their living arrangements or continued re ceipt of some support from children or other family members. As noted in Chapter 4 and elsewhere (Hermalin, 1995; and Hermalin, 1997), undue reli ance on simple measures of coresidence may confuse "form" and "function" by failing to describe the actual directions and magnitudes of support across Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 296 The Well-Being of the Elderly in Asia generations. Indeed Chapter 6, which examines the intergenerational transfer system within each country, demonstrates that the older persons provide considerable assistance to their children and others in a number of forms at the same time that they are frequently recipients of money, material goods, and time from others. Nevertheless, one cannot infer the economic well- being of the elderly from their coresidence with children or their patterns of exchanges. For this purpose one needs more objective measures of income and assets, complemented by the elderly's own perception of sufficiency of income and satisfaction with their economic status. Direct measurement of economic status is also important from a policy perspective. As a way of minimizing the high costs of "Western" systems of social security and old-age care, policymakers in many developing and newly industrialized countries would like to maintain to the degree possible tradi tional family arrangements in which children provide considerable support for their older parents. To this end, policymakers need to go beyond simple measures of living arrangements and support to assess the economic well- being of the older population in some detail. Otherwise they run the risk of failing to identify (a) problems with traditional family arrangements, (b) groups of elderly who are most vulnerable, and (c) additional programs that might bolster existing arrangements. Detailed economic data for the older population in developing coun tries have not been frequently collected. However, the 1994-1996 round of surveys reported on here did collect a variety of income and asset informa tion. This chapter uses this data to describe the economic situation of the elderly from a variety of perspectives, within each country and compara tively, and to analyze characteristics associated with levels of economic well- being. The next section describes the measures available and the plan of analysis for the remainder of the chapter. Measurement Issues Related to Economic Well-Being The measurement of economic well-being of the elderly is complicated by a number of conceptual and operational issues. Some of these are of a rather general nature, common to most countries and age groups; others impinge more directly on older age groups and countries with more complex living arrangements. Income is often used as a prime indicator of economic well- being, and in nearly all settings eliciting reliable estimates from a survey is problematic. Non-response tends to be high and non-random; farmers and the self-employed may have difficulty providing net income accurately, as might those with multiple sources of income. Among the elderly these basic Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Economic Well-Being 297 issues may be compounded by memory problems, lower involvement in eco nomic affairs, a multiplicity of sources, and perhaps the irregular flow of some of them. Household arrangements may present additional complications. Be cause a high proportion of elderly individuals and couples in the industrialized countries of the West maintain their own households, correspondence is gen erally high between their income and that of the household. As a result, fairly direct comparisons between their income level and those of other households can be made, after adjusting for household size and perhaps other factors. In many developing and newly industrialized countries (as well as Japan), how ever, a high proportion of the elderly live with children, and many, whether or not they coreside, receive financial support from their children (see Chapter 6; Hermalin, 1995; Knodel and Debavalya, 1997). As a result one needs to distinguish between the income of the elderly respondent or couple and the income of the household, and take account of allocations within the house hold both direct and indirect (for example, one elderly couple coresiding with children may have all their food and other expenses covered, while another contributes substantially to the running of the household). Even the elderly living alone may have certain expenses paid for by others, instead of or in addition to receiving money income from children. Beyond these considerations, a number of distinct dimensions of in come can be operationalized and used in description and analysis. In addition to the absolute level of income of the elderly unit or household, interest often centers on relative measures that permit comparison with other families in general or those in the same age range through attention to an appropriate aggregate distribution (for example, identifying whether a family is in the top or bottom 10 or 20 percent of a distribution). And, as suggested, the sources of income among the elderly and the relationship between source and level might also be gauged. Another measurement possibility is to assess the suffi ciency of income in relation to expenditures or to some poverty level. At a more perceptual level, inquiry can be made about the degree of satisfaction with income or economic well-being. Income, regardless of its complexity, is only one dimension of economic well-being. The asset and debt situation of an elderly unit or household is also a major determinant of the elderly's ability to obtain vital resources for living and becomes more critical after retirement. Assets may serve as a current or future source of income and may also be part of an exchange arrangement with children. Obtaining accurate asset information in a survey can prove even more difficult than income given problems of valuation, joint ownership, and privacy, and questions on the topic often result in high levels of non- response and misreporting. In recognition of these problems, the U.S. Health Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 298 The Well-Being of the Elderly in Asia and Retirement Survey, for example, has initiated several approaches for reducing non-response and identifying the proper asset range through a series of "unfolding" questions (Juster and Suzman, 1995; Juster and Smith, 1997; Hurd, 1999). Beyond the measurement difficulties, cultural practices must be taken into account. In Taiwan and other Chinese societies, it is not uncommon for parents to subdivide their assets while alive, sometimes in anticipation of the support they will be receiving from children. For example, Li et al. (1993) found that 25 percent of older parents surveyed in 1989 in Taiwan reported dividing all or part of their property. The variety of measures implied by the foregoing can be collected at some point in time, but interest also centers on changes in income and assets over time. Although evidence of change can be approached through retro spective questions, panel designs that obtain repeated measures are preferred for their accuracy. The four surveys used here collected a variety of mea sures of economic well-being that will be exploited in this analysis. In addi tion, the panel data for Taiwan permits an examination of how income and assets change for a cohort as it ages, and some of these dynamics will be presented as well, with additional analysis in Chapter 12. Collectively, the measures of income, expenditures, and assets de rived from the surveys include the following, though not all the question naires covered all the dimensions: • Total income of respondent (and spouse) • Total income of household from all income earners • Household pooling arrangements for meeting household expenses • Payment of household expenses by those not residing in household • Major source(s) of income for respondent (and spouse) • Ability of respondent (and spouse) to meet household expenses • Satisfaction with current economic status • Proportion owning their home and value of home (and land) • Proportion owning other assets. In obtaining the amount of individual (or couple) and household in come, respondents were asked to report a specific amount, and if they could not (or would not) do that, they were asked to select from a list of income intervals. In each country, except Taiwan, this served to significantly reduce non-response rates for individual/couple income, but not non-response for household income. The response rates for the other measures of economic well-being were quite high. Because the sampling design called for interviewing only one elderly respondent per household, it is important to be clear about the unit for which the older person was responding. Given that married couples often pool income, Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Economic Well-Being 299 respondents in the Philippines, Thailand, and Taiwan who were currently married and living with their spouse were explicitly asked to report their joint income, their major source of income as a couple (although sources received by husband and wife were recorded separately), and their joint perception of income adequacy and satisfaction. In addition to potentially having two sources of income, the elderly married often differ on other characteristics from the elderly who are not married (i.e., widowed, divorced, or never married). They are likely to be at a somewhat earlier life cycle stage, to be working, and to have younger children. In addition, non-married older men will often be quite distinct in their work histories, ages, and other characteristics from older non-married females. Accordingly, in the tables that follow we often stratify or control the results by marriage status and gender, creating three categories of "economic units": married couples, unmarried males, and unmarried males. It is important to note that these units have a variety of living arrangements, ranging from residing alone, to living with children, or living with others. Any reference to household income refers to the total income of all household members as reported by the respondent. Although the various dimensions of income and economic well-being described above are all interrelated and do not form a distinct system, many can be ordered in a heuristic and logical manner as suggested in Figure 8.1. The figure helps make clear the interconnections among the income-related measures and serves as a broad guide to the analyses presented. According to the figure the direct income of the elderly "unit" is deter mined by the characteristics of the respondent or couple and the sources of income. These, along with the amount of income, determine whether there are also forms of indirect support. The levels of direct and indirect support determine the degree to which the elderly "unit" has sufficient income to meet expenses, and this level of sufficiency, along with the amounts of in come, determine the level of satisfaction the elderly have with their economic well-being. The amount and nature of the assets owned will also influence the sources and amount of income, and are likely to affect expressed levels of satisfaction as well, but the dynamics between income and assets are not reflected in Figure 8.1. The following sections present analyses based on the economic mea sures listed above and the logic of Figure 8.1. The next section presents an overall picture of income levels of the elderly in absolute and relative terms. This is followed by an examination of the major sources of support for the elderly economic "units" and how the amount of income varies by source and sociodemographic characteristics. We then turn to the level and impor tance of indirect sources of support by examining household income and pooling arrangements. This leads to an assessment of economic well-being in Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 oc e33 I/) .3? oc i/i J2 tu 0l oe Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Economic Well-Being 301 terms of reported adequacy and satisfaction. The final sections report on assets and provide some illustrative panel analyses for Taiwan. Levels of Income in Absolute and Relative Terms In this section we focus on the amount and the distributions of income for the elderly economic units as reported for 1995 or 1996. We report on a variety of indicators of both absolute and relative income in both local curren cies and U.S. dollars to facilitate comparisons across the countries. Table 8.1 presents the basic data collected in each country showing the mean and me dian of annual money income received by the elderly respondent or couple. (The lack of clarity in Singapore as to whether married respondents were reporting for themselves or for the couple should be kept in mind—see foot notes to Table 8.1). In each country, not surprisingly, married couples report higher in come than unmarried respondents; but the difference between unmarried males and females varies by country. In Taiwan and Thailand, males who are widowed, divorced, or never married have substantially more income than similar females, but in the Philippines and Singapore, men and women in these marital statuses are not far apart. The proportion of all the units report ing that they received no money income varied from 3 percent in Taiwan, to 8 percent in the Philippines, to 10 percent in Thailand (not asked in Singapore). In each country the lowest category of income (or the lowest two in cases with those reporting no income) contain one-fifth to one-third of all eco nomic units. In Taiwan and Singapore, the distributions are quite highly skewed, with a sharp drop off in proportions in the upper categories, but the Philippines and Thailand show a more regular distribution across the whole range. (Conversions into U.S. dollars are given in subsequent tables but it is worth noting that by coincidence the Philippines, Thailand, and Taiwan have very similar conversion rates of their basic unit—peso, baht, and New Tai wan dollar, respectively- with each worth about .04 U.S. dollars in the period covered.) Table 8.2A presents more direct comparisons of elderly income across countries by giving the averages in U.S. dollars, and sets the stage for assessing its adequacy by presenting measures of average income for the population in terms of GNP per capita or GDP purchasing parity per capita. The latter attempts to incorporate differences in cost of living and related factors to achieve more comparability in assessing relative welfare across countries (see footnote to Table 8.2 and CIA, 1997. In comparison to GNP, the GDP Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Table 8.1. Mean, Median, and Percentage Distribution of Yearly Income by Economic Unit: Philippines, Thailand, Taiwan, and Singapore PhilippinesIncomeMean (in Peso)MedianNo incomeLess than 2,5002,500 to 7,4997,500 to 17,49917,500 to 47 ,499More than 47,500Valid NMissing incomeMarried Couples36,35115,0001.98.017.326.126.719.96235.3Unmarr. Males1847610,00014.713.514.022.128.17.61422.9Unmarr. Females186526,00014.620.318.619.617.49.64646.9Total28,21410.2187.712.917.423.423.714.91,2295.6TaiwanIncomeMean ($NT10,000)MedianNo incomeLess than $100,000$100,000 to $300,000$300,000 to $600,000$600,000 to $1 million$1 million and aboveValid NMissing incomeMarried Couples27.820.51.726.841.620.65.73.61,80921.5Unmarr. Males19.215.43.238.042.312.23.11.339519.3Unmarr. Females13.810.37.855.426.18.51.6.558325.2Total24.016.83.133.738.717.24.62.82,78821.9 Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Table 8.1. continued See Notes for Table 8.1 at the end of this chapter. ThailandIncomeMean (in Baht)MedianNo incomeLess than 5,0005,000 to 9,99910,000 to 19,99920,000 to 49,999More than 50,000Valid NMissing incomeMarried Couples37,62215,0005.519.212.917.725.519.22,0696.7Unmarr. Males16,5095,00013.534.212.814.017.08.637013.1Unmarr. Females12,9713,75019.434.614.312.513.55.81,55413.9Total29,01610,00010.124.713.316.021.414.63,9939.4SingaporeIncomeMean (Singapore $)MedianLess than $3,000$3,000-5,999$6,000-8,999$9,000-11,999$12,000-14,999$ 15,000 & OverValid NMissing incomeMarried Couples8,4496,78115.729.020.112.410.312.41,691.88Unmarr. Males5,8984,70928.736.915.49.34.74.85410.5Unmarr. Females5,2394,55629.838.919.65.24.12.51,7691.2Total6,9505,45122.533.719.49.47.37.73,9651.0 Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 304 The Well-Being of the Elderly in Asia reduces the range of incomes across the four countries, and has a particularly notable effect in raising the income levels of Thailand and the Philippines. In terms of yearly averages in U.S. dollars, older economic units have money incomes of about $1,100 in the Philippines and Thailand, $4,900 in Singapore, and $8,800 in Taiwan. Income distribution patterns produce me dian annual incomes that are less than one-half their means in Thailand and the Philippines, and about three-quarters their means in Singapore and Tai wan. The GDP and GNP per capita measures permit a comparison of the general living standards in each country against which to appraise the income of the elderly. In terms of purchasing parity, the per capita GDP for Singaporeans is the highest at $21,200; the Taiwanese have about 70 percent of this amount, Thais have about one-third, and Filipinos have about one- eighth. The GNP per capita figures assume the same sequence among the four countries, with an even a stronger advantage for those in Singapore. The first notable surprise comes from comparing the income of the elderly with these national averages. Despite the overall economic advantage for Singapore, the elderly there receive less absolute income than those in Taiwan, and de spite much lower national income averages in the Philippines, the elderly there have incomes equal to those in Thailand. Clearly a number of local factors must be considered in comparing money incomes across countries. Table 8.2B presents some relative measures based on these data. The first four columns express the averages of the elderly and the population as index numbers with the Philippines as 100. These reveal, as noted above, that the Philippine elderly have about the same income as their Thai counterparts although the country has only about a third of the GDP or GNP per capita. Relative to the Philippines, Singapore shows much higher multiples of GDP and GNP than it does for elderly income, while for Taiwan the differences in elderly income relative to the Philippines are basically similar in magnitude to the differences in overall GDP and GNP. The last four columns express the money incomes of the elderly (both means and medians) as a proportion of the two overall income measures in their countries. The incomes of the elderly in Taiwan and the Philippines assume a similar percentage of national average per capita income—roughly 40 to 50 percent in terms of the ratio of the median to average GNP. The incomes of the elderly in Singapore and Thailand are also comparable in terms of their proportion to the national average, but are significantly lower at only about 15 percent of the per capita GNP. If we compare the mean elderly income to GDP per capita, perhaps the most appropriate measure, the picture is similar. The Philippine and Taiwanese elderly have proportions at the high end, 42 and 60 percent of the overall average respectively, and Singapore and Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 s(/itQ9^rVI^—c_octQ3c 1Q wueoF D-P tOt OX)3jie5 ^w c— J.E|s12.— e.||eS £5•o CJC C10 .5.c a.10 —^Z 2C00QJ-C31"C/3SSuEeucec3.0e^3zaCe5T3U^Qcrtecuoo21^<^S™eQt^bfe)e^^^5i&ooD^^^VsvhaU113s§ell|O•.3 1on|g|d, ^-, "7.Cu ^Q u0 eiISa•o1c.2e8ScCOUCN NO O OOin o ov -^t~] oo CO r^8888ts" r^•e Tt" ^^-H tsoo CN in ON^ ** 2 moo in ^^ ^* Sve in ov — ^ r^ oo^— » — T oo" Tt"OO Q Q — i— O O ino" O oo inTt vO Q OCN O O Ovoo" eC o vO"CN CN •*tts:= -3 .-5 gl>F tS i2 •—0, H H oo Q. tQ DC -o tQ JO .tQ •o to 0) C0 tN 11O^OeCOCa1s•ocCOB3si*-iO8 •a/— S.8ii1DH8Xu•o^s/— V^^^^^^,^^§s^~H8* Hp*ovOPL,QO/— V§.2^0s0?IuOO •* O vO"! - "i -s 5 ?; s^^in in CN oo -^ p ^| ^ts m O t*^ •^ — ' vS CN8r- oo o vo i ^ in— i CN — • fi— ' ts"^5 Qv ^3 ^^-H rN in oo8— i — * iO o ^ vo^8 8 S 1-H -H r** ^~ 1 1 e£ H H i/5 -a CN Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 306 The Well-Being of the Elderly in Asia Thailand are at the low end, with 23 and 15 percent respectively. The rela tively poor position of the Thai elderly on this measure occurs in part because of the low cost-of-living assumptions for Thailand implicit in the conversion from GNP to GDP. Though one cannot extrapolate from the experience of these four countries, it would appear that there is no simple relationship be tween level of development and the well-being of the elderly as expressed in these measures. It is true that Singapore, the most developed of the coun tries, displays low absolute and relative income of the elderly, but Taiwan, the second most well-developed economy, does very well in these respects. Major Sources of Income In keeping with the framework described in Figure 8.1 and the interest in these countries in maintaining traditional support patterns, Table 8.3 intro duces the question of the sources of support for the older economic units. It provides the distribution of the major sources of income reported by the respondents for themselves (and for themselves and their spouses). Across all countries and types of economic unit, money from children or other relatives is reported most frequently as the major source of income, varying from a low of 31 percent in the Philippines to a high of 77 percent in Singapore. Taiwan resembles the Philippines in this regard, with Thailand midway at 55 percent. Singapore and Thailand evidence the greatest percent ages on this source of income for each type of economic unit, with the proportion among unmarried females reaching 90 percent for Singapore and almost 80 percent for Thailand. In each of the four countries, unmarried females are more likely to rely on children or other relatives than are married couples or unmarried males, reflecting their generally lower attachment to the labor force and lesser control over other potential sources of income. In the Philippines and Thailand, unmarried males are more likely to rely on support from children than married couples, but in Singapore and Taiwan, they are somewhat less likely. In each country but Taiwan, income from work is the second most frequently reported major source of support among all economic units. If one includes agricultural pursuits (shown separately for the Philippines and Taiwan) with work, this combination is reported as the major source of income by more than half of all the economic units in the Philippines, by over a third in Thailand, about a quarter in Taiwan, and close to a fifth in Singapore. This ordering of the prominence of work is consistent for each type of economic unit, though a higher proportion of unmarried Singaporean than Taiwanese males report work as their major source of income. Among unmarried females, Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Table 8.3. Major Source of Income: Percentage Reporting Each Source by Type of Economic Unit and Country 1 Includes: rentals, savings, real estate, stock, annuity, severance pay. b Includes: fanning, timber, fishing, animal husbandry. Entire sampleWorkPension, retirement, CPF Income from investments"Family business Agriculture1* Children or other relativesOtherMarried couplesWorkPension, retirement, CPF Income from investments1Family business Agriculturei' Children or other relativesOtherUnmarried malesWorkPension, retirement, CPF Income from investments*Family business Agricultureh Children or other relativesOtherUnmarried femalesWorkPension, retirement, CPFIncome from investments"Family business Agriculturei' Children or other relativesOtherPhil Thai Taiw30.0 37.9 20.49.1 1.92.3 29.2 4.4 5.94.6 -- 1.6 23.3 - 4.0 31.1 55.4 36.5~2.432.4 49.8 25.99.4 1.93.2 28.1 4.3 5.95.2 -- 2.0 29.8 -- 4.9 21.3 42.8 31.7-1.528.4 23.1 11.911.1 0.51.5 44.5 5.8 4.84.7 - 1.1 18.7 - 4.0 36.6 69.6 27.4-26.37.92.46.317.2 6.90.7 23.14.3 6.83.5 -- 0.5 12.8 - 0.8 47.2 77.8 59.0--2.9Sing17.62.2 1.876.61.824.82.7 1.569.51.421.74.4 3.566.34.16.30.81.789.71.6 Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 308 The Well-Being of the Elderly in Asia only those in the Philippines and Thailand rely in significant proportions on work as a major source of income. This accords with the labor force participation rates given in Chapter 7. Aside from income from children and work, the only other substantial category is income from pensions or other retirement arrangements, and this source is a major factor among the current older population only in Taiwan, where it is reported by 29 percent of all economic units. As discussed in more detail elsewhere (Shih, 1997), the prominence of retirement funds there can be traced to the high proportion of Mainlanders among the elderly—Chinese who migrated to Taiwan in 1949-50 following the civil war and who were predominantly military personnel and officials eligible for governmental pensions. Also, because many of the Mainlander men never married, the proportion of the unmarried Taiwanese males reporting pensions as the most important source of income is especially large. Moreover, starting in 1994, all elderly in several cities and counties in Taiwan received special pension benefits as a result of promises made in local elections, which may also increase the proportion of those reporting pension as the major source of income. (More detail on these special payments and their effect is given in Chapter 6 and discussed further in Chapter 12.) The infrequent reliance on pensions as the major source of support in Singapore is surprising, and merits some comment. The Central Provident Fund (CPF) in Singapore dates to 1955 and is among the highest in popula tion coverage of any plan in Asia. According to Chan's analysis (1997), al though it began mainly as a retirement program, the CPF soon evolved into an umbrella program for multiple purposes—education, medical insurance, and housing—which often meant that the amount available at retirement was in sufficient or unavailable for regular living expenses. Of those with CPF ac counts (about a third of the older population), over 60 percent report the amounts as inadequate for old age, and only 7.5 percent used it for 25 percent or more of their monthly living expenses (Chan, 1997, Table 2). Changes in the minimum amount that must be set aside for retirement were recently instituted, and studies now underway may show a change in the role of CPF funds as a source of income for retirees. (See the Asset section below and Chapters 3 and 7.) Factors Affecting Levels of Income This section investigates the degree to which levels of income vary by the sociodemographic characteristics of the economic units and their major sources of income. Tables 8.4 and 8.5 use three measures of income that facilitate Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Economic Well-Being 309 comparisons across characteristics and countries. Table 8.4 reports the ratio of the mean and the median incomes of those in each sociodemographic category to the overall mean and median incomes for each country. Table 8.5 reports the proportion of those in each sociodemographic category that fall in the lowest tier of the income distribution, highlighting those who are particu larly vulnerable as a result of low income. The characteristics investigated include age, marital status/gender (type of economic unit), education, urban- rural residence, living arrangements, self-reported health, and major source of income. Where the economic unit is a married couple, the characteristics of the husband are used for the age and education variables. The characteristics analyzed in Tables 8.4 and 8.5 have been shown to be closely associated with income over many studies, but the relationship in a number of instances are complex and caution must be exercised in attributing causality, particularly from cross-sectional data. Age, education, and, to a large extent for these older groups, marital status/gender and location can be considered determinants of income. However, for living arrangements, health, and major source of income, influences run in both directions. For example, if a widowed woman with insufficient resources moves into the home of her married child, her income has affected her living ar rangements. Similarly health and income have relationships that can vary throughout the life course and by time and place. For those still of working ages, income can exert a strong influence on health through the level of resources provided for medical care. At older ages, health often affects labor force participation (as shown in Chapter 7) and thus income. (See Smith, 1999 and Preston and Taubman, 1994, for additional discussion.) Level of income and major source of income are also characterized by bidirectional causality. Level of current and past earnings can affect whether or when older people retire and whether they receive major support from their chil dren. For these and other reasons, the differentials in income shown for these characteristics should be viewed as reflecting the level of association, rather than indications of cause. Our interest centers on how strong these associa tions are, and the degree of similarity across countries. The differentials in the indexes of mean and median income shown in Table 8.4 are highly consistent from country to country. This shows up graphically in Figure 8.2 in the similarity of profiles across countries. Also, the differentials by characteristic offer few surprises. Income is higher among the younger respondents or husbands (in the case of married couples, the age and education of the husband was used to represent the couple and contrast with the unmarried males and females), and declines quite steeply with age. For example, the mean income of those 60-65 years old in the Philippines is 2.4 times the mean income of those 80 or older. (The ratio of one index to Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 o rt 0l rt o. Q. be 0l 0l O 3 V5 rt oe 00ct/5t*t3c B-.2•5usJ3H1I^.3ci.t315O^sarsteristico o r** v^ -^ r^ ^^ — ' OO NO CN OO OOoo ON ro f^* ^^ ^ r^-CN — ' 00 Tt CS OO Ttf— ^ ^H ^-;O O CN CN O O oo in rs r*- ^ in in ro— i —i . e — iro r**- oo oo r^* oo &v IO ^^ O^ NO Tt ON iO^H — * *-*ro r^ — * Tt CN Vi v^CO CD OO vG CN OO t^*O oo *-* ON NO O r^ ro o oo vO ^H oo in— — — -••H ON oo ON o r^- v^Tt oo oo 55 rs vo xo— i e e ° — ;1 1 JJ| ^^^ 1J a g-e ^ fe 1 8-8-g « - e8 .1 1^ ^ s ^ , 1 i 1 1S^^ocJ e5|J| ecVvot-oo «2D^>•^ ^*^ CO CO o l r°i . f*v .. — i -H e e cKS^ 8SS_; ts • _; _;00 O O O O Oro o rs ts vi in. -J t^ — i CN CNO oo ON oo ^H rO vO ON vo ON ^ OCN ^^ ^^CO CN CN X o ri"-^ fitro in in r** ^- tN vo ON o r^ o Ttrs . _; _;CN CO e^ NO CO CNr-; oq ro r- m ONro . CNj _,in oo i0 Tt i^ •* r-; r-; p ov p rsCN —* -Hh006 iw b + 0 -a § >•• .^ 1 1 .i S p. G| 1 tl §_ g aA c<-' ^^ e .onS ^ ™ *S1 o fs gs^8-•|Z<c/o SeSDU Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 •o0lco u00_euooo-He^HOO ON2-HeON ONoo I";••2B _>,Living arrangen:Alone/spe.se 01«soots£oo ON°.oON8^•oW/ married chil— i m ts qON ON CO ON~HO ein q— ' — 'r- ooTt ON—28— i — i££28— iing^W/ .nmarried W/ etherNO CO r^- NO OO O ON OOr- r- oo m in oo r^i in^HO in O oo in r-; in m^HNO oo ro oo NO — in oo. -H -HOO — i OO OOoo o o r--H ^rs o oo ^NO ON C^ OOr- ON r- oNO OO Cvl Tt. ^SNO ON O ON in l~;ae «i I1 §1 1 a 11J3 -» T3 r r eO — ^ eH M « ^|^SJ£ |u-.Sin q-H -^8C<1 NOts ONoo oo *^) -^-H CN~°^ — i—• rsOO ON•*t oo— H Ttq ONcWmk/b.simss Pensien/retiremi InvstmentsON— '-*_3— iON OOm— '^r^NO8i— ^oo-H31wT oo.>ft t/^§8~^s— ,8_;8— -88— Je^Overall index£Mean er medianm8ooNO0oooo«HtsosmONNOs1osNOOONs2rsooCNlecal currency .'u2se5.a.a1..Q.£ri' See Tabls m e and Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 w, vo r- o 1" 61gs•£s•ssHet3sT3.t1C3a 11gBPB.2|U- T3 £ + iv 2 2 5s , C Q^ H H ill I III5 1 oQcN*.09_e3BS1a.2nSRelativs-cs-§eiucw Oetirement1.0 'ViCu Ou}1E8c •oThailan^_^1— 1TlFee- ——[1 — 1r—"' —- ro. in- CNv}. ci in rs i/)0J.a. a.Q.—]„—-f]^^—ir—n—- tsi/^ e Oo Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 ri a. tQ v0 v0 V wi r- o o °° r- l-Son3 *?I •= § •« •o -r •o 2 D I in •auz>vs^<+ >*1o&>v§cn 3CuE8C2<CJ]C£<ied Childi5zied Childh.CQEC^W^*Other* a3Ii355jEc30 •a II U M > g u •* •5 o Q£ __flir—[_^—— •5 M if .2 -s a g i S 'o I/e) ts Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 314 The Well-Being of the Elderly in Asia another is also the ratio of one mean to another, given the definition of the index.) As noted above, economic units consisting of a married couple have consistently higher incomes than do unmarried males or females. Education shows a sharp income gradient in each country, with the relatively small number of those with secondary or higher education particularly advantaged in comparison to their less educated counterparts. (The number of elderly in each category can be gauged by the table in Appendix 8.A, which gives the sample sizes.) Income tends to be lower among the rural elderly, but the differences between those in capital cities versus other urban areas tend to be smaller than the urban-rural differentials, and are not consistent from country to country. The last three characteristics address the relationships of living ar rangements, health, and major source of income to income level. Given the concern often expressed about the welfare of the elderly who are not inte grated into families and the vulnerabilities of the "abandoned" elderly, it is important to note that on every income comparison but one shown in Table 8.4, those living alone or as a couple do as well or better in terms of money income than those living with married children. This is not surprising given that more income is generally required to maintain a separate household, an issue treated at greater length below. Those living with unmarried children tend to report the highest income, almost certainly because they tend to be younger and still in the labor force. The smaller numbers living with other relatives or in other arrangements show no clear pattern, often lying interme diate to the incomes reported in the other living arrangements. Overall, there is rather little variation in relative income level across living arrangements or in the pattern across countries. In contrast, income varies consistently and strongly with self-reported health status in every country except Singapore. In the Philippines, Taiwan, and Thailand, those who report the highest health status (excellent or very good) have twice or more the income of those in the lowest category, and those at the middle level (who report their health as good or fair) have about 50 percent more income than those who report their health as poor. (See Chapter 9 for further details on self-assessed health measures and distribu tions.) As discussed above there are complex interelationships between health and income, with income levels strongly affecting health over much of the life span and health affecting income later in life. For Singapore, the lack of strong differentials probably reflects the very high percentage of elderly who live with their children, the tendency for nearly all elderly there to report their health as good, and perhaps the wider availability of health services. The last characteristic presented, major source of income, also speaks to the adequacy of family support, given the predominance of children and Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Economic Well-Being 315 others as the major source of support reported in Table 8.3. (In Tables 8.4, 8.5, and 8.6 those reporting work, farm, or business as their major source were combined, since income from labor could not be easily distinguished from income as entrepreneur.) Older economic units who report children/ other relatives as their major source of economic support consistently report lower total income than those with any other major source. It is important to emphasize that total income is being analyzed in relation to the reported major source. Those who report work as their major source, for example, may also receive some of their income from children. Nevertheless, it remains the case that the elderly who rely on family as the main provider of income are disad- vantaged in the total amounts they receive in comparison to those whose major source of income is another source. Among the three nonfamilial sources of income, no clear patterns in income levels appear, but as shown in Table 8.3, major reliance on pensions is rare except in Taiwan and relatively few of these elderly derive most of their income from investment or savings in any of the countries. Table 8.5 displays for the same sociodemographic characteristics pro portions that fall in the lowest tier of the elderly income distribution—ranging from the 22nd to the 37th percentile. (The use of bracketed income levels in some countries made it impossible to adopt a uniform definition—like the lowest third—for all four countries.) Given the interrelationships among the income measures, the differentials and patterns in Table 8.5 largely follow those in Table 8.4, but they present an alternate way of characterizing the income position of a specific group. Among those economic units where the respondent or husband is 80 years of age or older, about four in ten in Singapore, almost half in the Philippines, and about two-thirds in Taiwan and Thailand are in the lowest income tier of their country. Among unmarried females, about half or more are in the lowest tier in the Philippines, Taiwan, and Thai land. And among those relying on children or other relatives for major sup port, over half in Taiwan and Thailand, 42 percent in the Philippines, and 27 percent in Singapore are in the lowest tier of income. For those with the poorest self-reported health, the proportion in the lowest tier of income ranges from a third in Singapore to more than a half in Taiwan. By contrast, among those with the highest self-reported health, less than a quarter of the elderly in each country are found in the lowest tier of income. The characteristics shown in Tables 8.4 and 8.5 are closely interre lated in many ways: those younger are more likely to be working and married, for example. To estimate the effects of each characteristic on the income position of the elderly unit, we performed logistic regressions in which the dependent variable is the log-odds of being in the lowest income tier. Table 8.6 presents the odds-ratios derived from the regression coefficients, which Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Table 8.5. Percent of Respondents in Bottom Income Tier by Sociodemographic Characteristics and Country CharacteristicAge of R or husband<6565-6970-7980+Marital status/genderMarried coupleUnmarried maleUnmarried femaleEducation of R or husbandNo formalAny primarySecondary or moreLocationRuralUrbanCapital cityLiving arrangementsAlone/spouse onlyW/ married childW/ unmarriedW/ otherHealth StatusLowMiddleHighMajor source of incomeChildren/other relat.Work/businessPension/retirementInvestment, savings, otherTotalBottom Income TierPhil23.129.639.448.122.638.348.444.734.215.834.331.235.337.240.126.629.244.832.824.141.923.49.713.133.3Thai20.428.142.767.423.645.353.251.929.613.437.220.017.028.643.124.231.446.338.224.253.613.22.014.334.1Taiw27.629.941.962.628.541.263.255.933.08.550.033.023.430.846.422.235.851.941.024.754.316.533.929.536.8Sing11.816.927.443.615.728.729.826.019.78.1Inap.Inap.Inap.17.630.416.023.834.620.119.927.01.412.232.022.5 Notes: Bottom income tier is defined as follows: Philippines less than or equal to 5,000 Pesos per year. Thailand less than or equal to 4,000 Baht per year. Taiwan less than 100,000 New Taiwan Dollars per year. Singapore less than 3,000 Singapore Dollars per year. Inap. = not applicable. Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Table 8.6. Odds-Ratios for the Effects of Sociodemographic Characteristics on Being in the Bottom Tier of the Income Distribution, by Country *p<.10 *p<.05 **p<.01 ***p<.001 Inap. = not applicable. CharacteristicAge of R or husband<6565-6970-7980+Marital status/genderMarried coupleUnmarried maleUnmarried femaleEducation of R or husbandNo formalAny primarySecondary or moreLocationRuralUrbanCapital cityLiving arrangementsAlone/spouse onlyW/ married childW/ unmarriedW/ otherMajor source of incomeChildren/other relativesWork/businessPension/retirementInvestment, savings, otherModel Chi-Square(degrees of freedom)Odds-RatioPhil—1.161.361.67*-1.161.93***—.73*27***-.86.811.14—.78.66+—.58***.17***.29*121.4(15)***Thai-1.24*1.85***3.21***-1.85***2.26***—.70***,49***—.35***.24***.79*—.57***.67**—.21***.04***.16***1,231.3(15)***Taiw-1.131.36*2.39***-1.79***2.30***—.59***.17***—.39***.26***.62***—.54***.65*—.22***.51***.50***594.4(15)***Sing~1.181.79***3.07***-1.83***1.46***—1.02.37***Inap.Inap.Inap..72*—.62***1.04—.06***.691.84**477.4(13)*** Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 318 The Well-Being of the Elderly in Asia can be interpreted as the odds of being in the lowest tier for the category shown relative to the omitted category and controlling for the other indepen dent variables. The independent variables used are those included in Tables 8.3 and 8.4 with the exception of health. Although, as noted, the cause and effect directions are ambiguous for a number of the variables, this is particu larly true of health. A striking feature of the results is that in all countries but the Philip pines, each of the variables continues to exert a strong and significant influ ence on income position as observed in the bivariate relationships. That is, those above age 70, unmarried, with lower education, residing in rural areas, living with married children, and who receive their major source of income from children are more likely than their counterparts to be in the lowest tier of income. In the Philippines, however, this increased likelihood applies only to those above 80, those with little education, and those with children as major source of income. To study the relationship of health to income and the effect of including health, living arrangements, and major source of income on the more exogenous variables, Appendix 8. B presents a set of multivariate analyses in which different combinations of variables are regressed against income position. Panel A of the Appendix repeats the multivariate analyses shown in Table 8.6 but adds health status as an additional variable. A very consistent picture emerges. In each country, the odds ratios for the variables shown in Table 8.6 remain virtually unchanged in magnitude or significance level with the inclusion of the health status variables. At the same time, the health status variable displays a strong and consistent effect on income position. In each country those with the highest self-reported health status are much less likely than those with the lowest self-reported health to be in the lowest income tier, with the proportions for the former group running from one-half to two-thirds, compared to the latter. The results are significant at the 5 percent level in all the countries but Taiwan. The persistence of the magnitude of the other variables and the generally strong effect of health, suggest that health status is only weakly associated with the other variables (as described in more detail in Chapter 9) so that the relationship of health to income position is not dependent on the other characteristics, and each maintains an independent effect on income tier. Panel B of Appendix Table 8.B introduces two reduced models to re move the other variables—living arrangements, and major source of income— which are likely to be endogenous with income level and position. Model 1 omits both health and living arrangements, while Model 2 omits health, living arrangements and major source of income. Comparing Table 8.6 with Model 1 indicates a very consistent pattern: in each of the countries the odds ratios Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Economic Well-Being 319 for the variables in common remain virtually unchanged. Omitting living ar rangements from the Table 8.6 specification has no effect on the remaining variables. When Table 8.6 is compared with Model 2, which omits both living ar rangements and major source of income, the results are more complex. In Model 2, the effect of age on income position becomes more pronounced in each country as does the effect of being unmarried (but only for females in Singapore and Taiwan). In addition for the Philippines, the contrast between those residing in rural areas and the greater Manila area is heightened in Model 2 (though still not significant), and the rural-urban contrast in Taiwan is also heightened. Interestingly, there is little effect from omitting major source of income on the gradients with education. Overall, the fully reduced Model 2 strengthens the effects of the four exogenous variables—age, marital status, education and place of residence—on the current position of the elderly in terms of income distribution. The results in this section suggest that a number of sociodemographic factors each exert a distinct effect on the money income of the elderly. Be fore we can adequately characterize the economic situation of the elderly, however, we must examine the prevalence and importance of their indirect forms of support—those not captured by money income—and their asset and debt positions. Household Income and Indirect Sources of Support At the outset it was noted that one source of complexity in assessing the economic well-being of the elderly is their tendency to be embedded in other households. Because a high proportion of the elderly live with children or other relatives, they may benefit from the provision of housing, food, and other material goods by virtue of shared arrangements. But sharing per se does not guarantee enhanced economic welfare, because the elderly in such arrangements may be providing resources to others from their income and assets. Measuring indirect sources of economic support requires a fairly com plete accounting of household income and expenditures that identifies all the income earners and their incomes, as well as the pooling and transfer pro cesses within the household. Although the questionnaires in the four coun tries fall short of that goal, a number did identify key components of indirect sources of support within the household, which will be reported on here. In addition to complexities arising from intra-household exchanges, the elderly who live alone as couples or individuals may receive economic support from their children or others in the form of payment for major or ongoing expenses Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 320 The Well-Being of the Elderly in Asia such as housing or food. Some data on support of this nature were also collected and will be included below. A priori, we would expect to find a higher prevalence of indirect sup port among those coresiding with children and others both because coresidence facilitates this type of transfer and because lack of sufficient funds can be a precipitating factor for the elderly to give up independent living. This expectation is supported by Table 8.7A, which shows the propor tion of older economic units receiving some type of indirect support (as defined in the notes to the table) for those living with others and those living alone. The table reveals the high level of indirect economic support for the elderly in each type of economic unit and in each country—ranging from 44 percent for married couples in Thailand to 93 percent for unmarried females in Singapore. For each type of unit and within each country, those living with others are more likely to report indirect economic support than their counter parts living alone, with the proportions receiving such assistance ranging from 50 percent to over 90 percent among those coresiding. By type of economic unit, unmarried females more frequently report receiving indirect economic support than others, but they are less distinctive from unmarried males in the Philippines and Thailand than in Singapore and Taiwan. To what extent is this high level of indirect support targeted to assist those older economic units whose money income proves insufficient? Table 8.7B gives a partial response to this by measuring the degree to which eco nomic units in the lowest tier of income receive indirect assistance relative to all the units in similar living arrangements. In the Philippines, Taiwan, and Thailand there is evidence of an increased likelihood that units at the lower end of the income distribution will receive indirect economic assistance, and in most cases this tendency is stronger for those living alone. In Singapore, indirect assistance does not appear to be targeted to those with lower in comes, but it is here that 85 percent of all units at all levels of income receive help of this kind. (In analyses not shown here that relate income position to amount of indirect support received, we found for each country but Singapore a clear gradient of increased support as one moved to lower income levels.) Although the indirect economic support received was not valued mon etarily, its high prevalence and its association with income levels indicate that it is a significant factor in the economic well-being of the elderly. Additional insights into how pooling and sharing arrangements enter into this assess ment can be obtained by using the total household income when older eco nomic units live with others, and by direct questioning concerning pooling arrangements. Tables 8.8A and 8.8B develop various relevant measures. Table 8.8A shows for each type of economic unit the median income of those living Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Table 8.7A. Percent Who Have Others Pay for AH or Some of Their Household Expenses (indirect support), by Type of Economic Unit and Living Arrangement Economic UnitMarried couplesLiving with spouse onlyLiving with othersAllUnmarried malesLiving aloneLiving with othersAllUnmarried femalesLiving aloneLiving with othersAllTotalPhil24534815696123676255Thai28494430837633837860Taiw30524613764844857953Sing45837832868066959385 Table 8.7B. Ratios of Those in Lowest Income Tier Receiving Indirect Support to All Economic Units in Same Living Arrangement See Notes for Table 8.7A and 8.7B at the end of this chapter. Economic UnitMarried couplesLiving with spouse onlyLiving with othersAllUnmarried malesLiving aloneLiving with othersAllUnmarried femalesLiving aloneLiving with othersAllPhil1.161.271.241.891.201.221.571.161.16Thai1.581.401.431.221.121.121.121.081.09Taiw1.271.221.242.021.081.351.091.061.07Sing1.181.121.14.851.051.03.971.001.00 Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 322 The Well-Being of the Elderly in Asia alone versus those living with others. In the Philippines, those living alone have lower median money incomes than those living with others, but in the other countries the reverse is the case, with those living alone generally showing higher incomes (except among unmarried females in Taiwan). A priori, one might expect countervailing forces to be at work. Supporting a separate house hold is generally more expensive than coresiding, leading to an expectation of higher income among the independent elderly; but living alone may result from the unavailability of children or other adverse circumstances, generat ing an expectation of lower income.1 In the Philippines the adverse factors appear to predominate, but in the other countries there is generally little dif ference in income by living arrangement, suggesting a trade-off in these countervailing forces.2 Table 8.8B looks at some of the household parameters for those living with others. The median income of the household given in the first three columns shows in every case that the income of the household is much higher than just the income of the elderly unit residing there (as reported in Panel A). Whether this extra income translates into direct benefits to the elderly coresidents depends on a number of factors, including the size of the household, the pooling and sharing arrangements, and the relation of the elderly unit's income to the household income. Are older economic units with low income also resident in households with relatively low income or is there a reasonable degree of independence in these two measures of welfare? Column (4) of Table 8.8B addresses this question in part by showing the proportion of lowest-tier elderly economic units in the lowest tier of household income. The figures in parentheses are the percentages of all households in the lowest tier of household income. If these measures have a weak or nonexistent association, one would expect about the same proportion in the lowest income tier by economic unit to also be in the lowest category of household income.3 The results indicate that in Singapore the two proportions are in close agreement, pointing to a weak association between couple or individual income and the income of the household in which they reside. In other words, many Singaporean older couples and individuals who appear to have low money income are benefiting from being in households which are relatively well off. This is less often the case in Taiwan and Thailand where over 40 percent of economic units in the lowest tier of "unit" income are also in the lowest tier of household income, and in the Philippines, the association between the two levels of income appears very strong, with two-thirds of units in the lowest tier of income also in the lowest tier of household income. Assessing the potential benefits to older couples or individuals of coresiding with others also requires attention to the direction of flows within the households. How often do older economic units share their income with Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 ecou£>J2*.E^J^v3^c0l1ctQ<*UrtcIT•oceQt/)•o010'BIocco 5_Q$0OaiXIaa|oIO(JHHi^sSgJ35•jHBOB^_e.S3B3aIB1BSeS• «cBDM1eU"aSe8 J•p o.iiue3£BOU1•Q M.Ji J-t §•168NO ro CO§8 in ON^ c5 ^*j oo2 v$ m8fm-°°. °5. o oots 00 CNOv ^ -^ oo^t" wi Tf§CN °00 CN8 - vo °°P. ts o SSCN NO CN yQ- -a^ BOCU H H co I" f I S^- 4J i r » II gJ=0*rJocc.>JSoU•o oj;0O ~u—Oi Structurew tQIncomef^iCUce0OV.cf>1CI exe,1ie0 g^*>1^ pU -^^S 2e _§C V1Z^3SCAuB *5s3B .5.(__nnjg.° ^a1USeac*e1e £B.2S.gEsBMarriedr^Female:1M Ue5.e U^B3OUm ts •* r-uo o -H voO ON vO ~^ CO CN -H C11v~' ^-e ^^ ^^m ^ in r^i/l m vO vO^^Q Tf CO OOfsj O CO f1^! Tf tsvo O ? OO ro ^ ts— ' itLc *c .t3 BCU H H Kts00 T3C00C/5 _0f2 uu u13e5g—*—i ^ S - £ o. IS 2 S .E ! « ; u Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 324 The Well-Being of the Elderly in Asia others? The last column of Table 8.8B addresses this issue by examining the proportion of coresiding households in which the elderly unit is the only income recipient, strongly pointing to support of others within the household. This measure also reveals the relative disadvantage of Filipino elderly: in a third of all coresiding economic units, the respondents are the only income recipients in the household, indicating they are providing for others as well as themselves. In the other countries this is much more rare, occurring in only 7 to 14 percent of the cases. Table 8.8C completes this overview of household income dynamics by providing additional detail for the Philippines and Taiwan. When economic units in the Philippines coreside in multiple-income households, there is an average 1.0 to 1.3 income recipients in addition to the older couple or indi vidual, and about 40 percent of the time some type of pooling of income occurs. (In the remaining households, residents take care of assigned ex penses or engage in some combination of assignment and pooling.) In Tai wan, the average number of additional income recipients ranges from 1.3 to 1.5, and the pooling varies sharply by type of economic unit. Older married couples coresiding with others earning income pay most of the household expenses in about half the cases, while unmarried males or females pay most of the household expenses only 18 and 13 percent of the time, respectively, suggesting their much greater economic dependency. Subjective Measures of Economic Well-Being In addition to the array of income measures presented in the previous sec tions, the surveys also obtained subjective assessments of their economic situation from the respondents. Collectively the countries investigated the perceived adequacy of the older couple's or individual's income relative to their needs and expenditures, or inquired more generally about their satisfac tion with their economic circumstances, though not all countries used both lines of questioning. Discrepancies between more objective measures of status and subjective assessments have been observed in many behavioral studies and have been a stimulus for theory and empirical research. For overviews of subjective measures of well-being see Diener (2000) and Cummins (2000). For attempts to reconcile objective and subjective economic measures or economic-oriented models see Douthitt, MacDonald and Mullis (1992) and Kushman and Ranney (1990).) Here discrepancies between objective and subjective measures of economic well-being may arise in part from differences between the Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Table 8.8C. Additional Details on Pooling Arrangements for Economic Units Living with Others Philippines Of Households with Multiple Income Recipients Average Number of Income Recipients in Household besides Elderly Married couples 1.0 Unmarried males 1.3 Unmarried females 1.2 Percent of Those with Other Income Recipients Pooling Income Married couples 43.8% Unmarried males 46.1% Taiwan Unmarried females 41.5% Of Households with Multiple Income Recipients Average Number of Income Recipients in Households Besides Elderly* Married couples 1.5 Unmarried males 1.3 Unmarried females 1.5 •Based on identifying up to 3 other members of household with income Percentage Distribution of Pooling Arrangements for Those Residing with Other Income Recipients Married Couple 1. Respondent (and spouse) pay most of expenses with no or some assistance from others 2. Others pay most of expenses, respondent (and spouse) give no or some assistance 3. Respondent (and spouse) and others in household share expenses Total 47.2 48.3 4.5 100.0 Unmarried Male Female 18.3 13.1 79.6 85.2 2.2 1.7 100.0 100.0 Generated for anonymous (University of Michigan) on 2014-05-27 12:30 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 326 The Well-Being of the Elderly in Asia respondents' current income levels and those experienced in the past, comparisons with how well other elderly are doing, prior expectations of how well-off they would be, or the degree of support expected from children or other sources. The direction of the difference between the more objective and subjective measures may be positive or negative, leading to expressed dissatisfaction despite apparently comfortable circumstances, or expressed contentment despite apparent shortfalls in the objective indicators. Given the strong economic growth in the region, one might expect the latter contrast to arise with some frequency if the older population has shared in the growth and experienced improved living conditions. Table 8.9A presents the distribution of responses in the Philippines, Singapore, and Taiwan to questions assessing the adequacy of income related to needs and expenditures. Sharp differences across the three countries emerge: in the Philippines, almost a quarter of the respondents report considerable difficulty in contrast to less than 5 percent in Singapore and Taiwan. Those expressing some or considerable difficulty exceed 60 percent in the Philip pines, compared to about 25 percent in Taiwan and 10 percent in Singapore. The relative economic well-being of older units as conveyed by Table 8.9A contrasts quite sharply with that portrayed in Tables 8.2A and 8.2B, where the Philippines showed a much higher proportion of GNP per capita going to the elderly than Singapore; and Taiwan elderly units had higher absolute in comes (in U.S. dollars) than those in Singapore, despite lower GNP per capita. These reversals reflect in part the effect of indirect sources of support as well as differences in the overall standards of living, among other factors, and demonstrate the importance of using multiple indicators of economic well-being. Table 8.9B investigates how the provision of indirect support inter sects with the perception of inadequate income. Among those economic units living alone in the Philippines and Taiwan, those receiving indirect support more often express difficulty with insufficient income than those not receiv ing this type of assistance, suggesting that indirect financial help at a distance often arises from a perceived need but tends to fall short of alleviating that need. In Singapore, there appears to be little differential in expressed diffi culty between those with and without indirect support, except among unmar ried females where those without support express more difficulty. Among those who live with others there is generally less difference in expressed difficulty with income shortfalls between those who do and do not receive indirect support. The differences that do emerge—mainly in Singapore (particularly for unmarried females) and among unmarried females in Tai wan—reveal that those without indirect support more often report difficulty, suggesting that even with coresidence, the needs of the elderly may be over- Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Economic Well-Being 327 Table 8.9A. Distribution of Sufficiency of Income: Philippines, Taiwan, and Singapore Enough money with some left overJust enough money, no difficultySome difficultyConsiderable difficultyPhilippines8.929.938.223.0100.0Taiwan7.966.521.64.1100.1Singapore8.481.58.02.1100.0 looked by other household members. Regardless of circumstances, the strong negative assessment of the Philippine elderly about the adequacy of their income stands out. Table 8.9C displays findings for a broader measure—satisfaction with economic situation or status—for Taiwanese and Thai elderly, the only two countries using this item. (As noted in the footnote, the questions posed were somewhat different in the two countries, making direct comparison between them somewhat tentative.) Overall 17 percent of older Taiwanese respon dents express some degree of dissatisfaction with their economic situation and, as the top tier of the table shows, this percentage is closely related to their relative income position. In Thailand, almost 30 percent of the units report dissatisfaction, with a slight decrease in dissatisfaction reported among those in the highest tier of income. Differentials in levels of satisfaction ap pear not to be greatly affected by type of economic unit or living arrange ment. In Taiwan, unmarried males and females living alone show somewhat higher levels of dissatisfaction than those living with others, and this is true for unmarried males in Thailand as well. Though the subjective measures reviewed must not be taken out of context of the other economic indicators, they do signal that perhaps 60 percent of Philippine's older couples and individuals find their economic situ ation at least somewhat difficult, as do a fifth to a quarter of Thai and Tai wanese. In Singapore, by contrast, only one in ten older respondents express some degree of difficulty with the adequacy of their income. Assets and Home Ownership In assessing the economic status of the elderly, attention must also be given to the nature and magnitude of their asset holdings and their debt position. The level of assets held by older people is an indicator of their economic Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 1 a. 0l b oc •n •jj « >^ c* Slfl ._ UJ Q.S ai a. I? SI I/) UJ is. •8? ^ "= S.1 ee 1I3^)O<>313Hi 8 1 •-I 1all*3«t13H••— . *— . *— .D 0 feIII;> S ^5o|•^ c 3HH C/5._eS^3o.i110 Wu1*~ oo -^ ^ fei O ro O ONON ^3 ^3 ro Tt O oo O N ON'0 i0 NO CN tN roO) vO oo oo o^ f^- oo ^5 roTt Tt lO Tt ^ tv^ CN Tf iOiOiOvO CNC^Co i-H^Hro'OONCN OCO^l e^ *• ^ ^*e^ r^ ^3 ro .^ oo r^ oo oovO iO vO CN CN CNvo Vi oo r*-* e^ vo C*~ ^H Ttto v^ r^* ro ^H f*** ON vi ^H 'of^Tt CNroro ^HCsiCNOl ON Tt ON ro r*- oO ON CONOOCN -HO•*t O^.i/S i0 r^ Tt CN ro ro -H rN rovo ^5 r ^ ON e^ ON r^ v^ roTf eO ON r^* ON ^^ ON IO Ttr^ovo CNroTt ^HCN^HBB 98 VOal a> wCfl P— * M -— - w5 —JecL§§ '&,§§ PDsC§g G *^- 3 C ^^ 3 S **^O-O-Q o-o-c o-o-owjCJqjO UOJOl O^^* — J3 c c ^^Scc Sb^cc5 S D P -a 2 D P H^PPa, E- &n 5 K .2 5 - Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 u II •^ 3 J2 10 E Ei! ^ 00 SI U 00 tQ U wn-= u- C o o ^ s_ IS J'o ON B.1H•oI.3j=H1eSieD.-^tu-oJ•f>cD*-.^e^^,u>*e8en•otd00e8c100T3!Cusvti£e8t/:.aOOe8CO00•o1CCO.a00U1 sharsteristir- NO ooOO CN NOoo ON r-ON NO —f"i r*- OCN — — ir*- oo ON— • r^- 06r- r- r-in NO oTJ Tt -^NO Co CNC^1 Of &CN — "oq oq — ;ON ro CNtN to CN— i oo Ofi oo CNNO in NO—i in ONr*- r^* In-HInrtme tierBettem Middle TepON ^ C*1oo in Tt in in —-*' CNOO CN NO^ r^ NO•^ Tt r-;f- NO NOp» f^ l^r- 't r-t^ NO NOin r- CNr^ 00 NOTt NO —— ' CNNO O mO CO CNf^ fei roCN tN OO— OO OONO in mCN oo r-00 OO OOC U M O e£fl C Q-1 Cu.5 & =s -5 ^2 DC i !E e .E ^lllj J"tS13Unmarried0ONfjCN CNCNOinr^OCNOfiSvoNOce00Joo r-** SCN ^*ON r-;- SO -tO inoo r^— • oo06 t-^JO NOCO TTO t-in mCN CNin mvQ NONO Oo oeu0•S00inet^ro CNovf^0•^ooSCNCOCNCN~4OO NOvoONCA13IsUnmarried Livin- aliIQ00in^[—CNf*r--t"-:ON CNin—^ONCNesesNOoooNL.cO•S00JOS^00ooin°ii/Sfi_^ONCNCNC*lNOOOON<o00 CNfeiq—•**•NOr-NOo•^f-HONCNC-NO-HON3 '3 [2 Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 330 The Well-Being of the Elderly in Asia success over their lifetime, may be influential in shaping the nature of their intergenerational exchanges, and may be crucial in generating needed income after withdrawal from the labor force. As noted earlier in the chapter, the accurate measurement of assets poses considerable difficulties and the data collected in the four-country sur veys must be regarded as a first approximation to a complex situation. The strategy generally adopted among the countries was to give most attention to home ownership, the asset most widely owned by the elderly, ascertaining both the details of ownership and the underlying values. For other assets, the general approach was to determine ownership within particular categories, the definition of which varied somewhat across the countries. Respondents were asked to place values on these other assets in two countries, but these data will not be used here. Table 8.10 presents an overview of the asset situation of the older population in the four countries, describing the proportion owning each cat egory of assets as well as the proportion with debt (where available) and the proportion without any assets. In each of the four countries, the house the elderly reside in represents the most widely held asset. Alone, or jointly with children, home ownership by the older respondents ranges from 50 percent in Taiwan to over 80 percent in the Philippines and Thailand. (The lower figures in Taiwan and perhaps Singapore are discussed further below in connection with Table 8.11.) The next five rows of the table provide the percentages owning specific categories of assets. Ownership of real estate (house and/or land) other than place of residence is reported by about 40 percent of Philippine respondents and 20 percent of Taiwanese, but by less than 10 percent of the Thai elderly. Philippine and Taiwanese elderly also report significant interests in farms, fish ponds, or ranches other than the one on which they reside. Together with the other real estate interests, these figures suggest that assets associated with land or agriculture-related enterprises are a major form of holding in both countries. This is not surprising given the significant proportion of each country's older population who are rural residents, but in Thailand, which has the highest concentration of rural elderly, other real estate (which may also reflect interests in agricultural enterprises) does not appear significant. In Thailand about a third of the older economic units report cash savings or ownership of stocks or bonds, a proportion that is slightly higher in Taiwan, but much lower (10.5 percent) in the Philippines. In Singapore, the profile of asset ownership is quite different. Given that Singapore is a city, the survey did not specifically ask about ownership of agricultural enterprises, and ownership of other real estate is also likely to be insignificant. By contrast, more than two-thirds of Singaporean elderly reported ownership of savings accounts of various kinds Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 £"c3OO•81'ouQ. f*30iM)cO.E3•iT-oLLJ — . 1£o"x,01J3i21ta00£5g.31&oc.5..£•1c^ ^ ^ NOS Z Z -ON m in •*O — i — i in m rs ts00 CN •< <ts ON z zOO S. S.r-^-^Gv I^i r**-O ON oooo e0 ^Hi y 'srteewnership (land and/m he.se)Other real state Farms f_sh pends rameh Other b.sinss interests-H CN 00 Ttfi -; ts p <| mvO *•* CO ••HON t*1 ^ —H ^ ^ts tN vd 3 Zm ts -n-^_^•^ t^* oo t^i ^,CO j-J CN Tt %r^—T—'^—^in f^ m r^ OO CO Ov OO ^^-H »-H i^l r^^ ^V- y 'I•o «a e ^ « .p •2 i «s -g•_> o " % -a* -5 *! i JB1 1 22 1 11 "- 1 1 1 ! 1 1 II > 1U on O O 3C Om NO r- r- oo ONvo060OS'nov^se.sehold pessessienso-; vqi§ 2ro viTt -^oo —Iro pO 06CNReceivin- m expectin- pensien imerte Men Wrten— -^ od Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 332 The Well-Being of the Elderly in Asia and another 6 percent owned stocks or bonds—a difference highly attributable to their participation in the Central Provident Fund (CPF) (described above and in Chapters 3 and 7). A high proportion of Singaporeans who had accumulations in the CPF reported withdrawing them, with many opting to place all or some of their funds in savings accounts of various kinds. It is worth noting that among the current elderly in all four countries, investment in stocks or bonds is not very common, reflecting both the relative recency of capital markets in these countries and the modest asset position of many elderly. The proportions owning assets in lines 1 through 6 are not of course mutually exclusive and many elderly will own more than one type. Line 7a presents the proportion of elderly who own none of the assets listed above as a way of assessing the proportion who may have no resources to fall back on. This proportion is 12 to 13 percent in the Philippines, Thailand, and Singapore but rises to 26 percent in Taiwan. This latter figure is somewhat misleading as it arises mainly from the lower proportion of home ownership in Taiwan, which is associated with the practice of division of property be fore death. It will be recalled from Tables 8.2A and 8.2B, that the relative income of the elderly in Taiwan compares well with over-all per capita in come levels, reinforcing the point that multiple measures of economic well- being are needed to obtain a balanced picture. Line 7b shows the percentage of older respondents or couples who do not own any of the assets shown in lines 2 to 6 (including line 9 for Singapore). These respondents may or may not own their home but they do not own any other financial assets. In Singapore, one third of respondents are in this situation, more than 40 percent in Taiwan, and in the Philippines and Thailand about 60 percent of the respondents or couples have no financial asset, except possibly their homes. Further discus sion of the characteristics of those without any assets is reserved for Chapter 11, which focuses on identifying the disadvantaged elderly from the perspec tives of economics, health, and social relations. Line 8 indicates the proportion of elderly units with household debt in the Philippines and Thailand, the only two countries that measured this. About two out of five respondents in the Philippines and one out of four in Thailand reported debt, and in Thailand 10 percent had debt exceeding 25,000 baht (equivalent to US$1,000)." In addition to the issue of debt, a rounded picture of assets should take into account the degree of control that these older persons have over their assets. In the Philippines, homeowners were asked if they were free to sell or mortgage their home/property, and 72 percent said yes. In Taiwan, respondents who owned assets other than their home were asked if they were free to use these assets in any way they desired, and 77 percent said yes, 9 percent said Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Economic Well-Being 333 no, and the remainder said they fully controlled only part of the assets. This high level of control is somewhat surprising given that almost 40 percent reported inheriting some of these assets.5 These respondents were also asked about the adequacy of their assets to support them throughout their life, and 55 percent of those owning such assets said they were sufficient or more than sufficient. Lines 9 through 11 present indicators of other types of assets reflect ing resources and/or lifestyle levels for the elderly. Line 9, "other valuables," represents a mixed group of assets that were identified in the survey but not in a uniform manner. In all countries with data, these include jewelry, art work, or similar valuables. In the Philippines this category also includes appli ances, which accounts for the relatively high level. In Singapore the figure of 10.5 percent also includes those reporting owning or receiving income from life insurance and annuities, or pensions (other than the Central Provident Fund), or maintaining an accumulation within the Central Provident Fund, items for which comparable information is not available for the other coun tries. To provide greater comparability on the subject of appliances and household possessions, line 10 compares three of the countries on the pres ence within the elderly unit's home (but not necessarily their ownership) of one or more basic appliances such as a refrigerator, television, and air condi tioner. This measure speaks more to the standard of living of the elderly than to their asset position. Line 10 shows that almost all the elderly in Thailand and Taiwan have such appliances in their households, while about a fifth of those in the Philippines do not benefit from these basic items. It should be noted that Table 8.10 does not address the ownership or presence of cars or other vehicles—tangible assets that can be significant components of total household wealth. In the survey, only 6.7 percent of the respondents in the Philippines reported the presence of a car, truck, or jeep in their household; the comparable figures for Thailand and Taiwan were 17 percent and 38 percent, respectively. Given these relatively low levels, which may nor may not reflect ownership, it appears that this class of tangible asset is a minor aspect of household wealth. In assessing the asset position of the older population it is important to go beyond the components of household wealth (home equity, tangible as sets, and financial assets) to include their stake in social security and pension programs to which they may be entitled. These future streams of income can represent important resources, although their total value may not be convert ible into cash or its equivalent. Where the amounts of the future payments are known, it is possible to calculate their present value and combine this with the value of other assets. For the United States, Smith (1999) has shown that the Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 334 The Well-Being of the Elderly in Asia present value of social security and pension payments is about equal to house hold wealth, when considering mean values. For the median household, the present value of social security exceeds that of household wealth and the value of pensions alone are over 40 percent of the value of household wealth, for those 51 to 61 years of age. For the four Asian countries, the detailed amounts received or ex pected are not known. The percentages given in line 11 represent the propor tion of all older men and women who are either retired and receiving benefits from a private or public pension (or have received a lump sum or limited number of payments in the past), or still working and report expecting such benefits upon retirement. As such the proportions are not independent of those appearing in other lines of the table. For example, those in Singapore who have accumulated funds in their CPF accounts and have withdrawn these funds and placed them in a savings account would appear both in line 5 and line 11. The data on line 11 show a wide variation in pension coverage across countries and by gender. Almost half the men in Singapore have or can antici pate such coverage, compared to a third in Taiwan, a fifth in the Philippines, and less than 10 percent in Thailand. Among women, the proportions are very low, except in Singapore where about 14 percent have or expect pension funds or payments. (See Chapter 3 for further discussion of pension plans and Chapter 7 for an analysis of the effect of pensions on retirement.) As presented in Table 8.10, a significant proportion of the elderly in these four countries own some type of financial resources, which tend to consist primarily of their current residence. Given the agricultural background of so many of today's older population in these countries, interests in farms, fish ponds, and other real estate is quite common in the Philippines and Tai wan, but somewhat surprisingly not in Thailand. Among other savings and investment mechanisms used, the emphasis appears to be on savings ac counts and other fixed deposits. This generation of elderly is little involved with the capital markets. Though the impact of their assets cannot be deter mined without an assessment of their value, the high proportion who own only their home, and the significant proportion in the Philippines and Thailand with household debt, suggests that most of the older population have not had an opportunity to accumulate sizable savings or other assets, despite the gen erally strong economic times they have lived in. Even in Singapore, with the broad coverage under the Central Provident Fund and the high savings rate embedded in this plan, the large proportion regarding their accumulations as inadequate for old age and the small percentage using these funds as a major source of income point to a generally weak asset position. In Taiwan, how ever, more than half of the elderly report owning an asset other than their Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Economic Well-Being 335 home (line 7b) and over half of these respondents indicate that these assets were sufficient for life-time support—indicating a strong asset position for more than a quarter of the elderly. Table 8.11 takes a close look at the nature of home ownership and the characteristics of those that do and do not own their own homes. The first row provides percentages of the elderly (a) who solely own their residence, (b) who own it jointly with other household members (ascertained in the Philippines and Singapore) or who contributed to their children's ownership of it (ascertained in Thailand and Taiwan), (c) who do not own any part of it, or (d) who have some other ownership position. The differential across the four countries is significant if one focuses on sole ownership, with variation from a less than 43 percent in Singapore to almost 80 percent in Thailand. But incorporating joint ownership or direct assistance, the variation is smaller— from 66 percent in Singapore to 85 percent in Thailand. In Chinese culture it is not uncommon for parents to divide property with children while still alive or at the death of the patriarch (see Li et al., 1993), and this may account for the lower proportion of homeowners among the elderly in Taiwan and Singapore. These data demonstrate that a high proportion of the elderly, regard less of current living arrangements, maintain an important asset through sole or joint home ownership, or have a claim for continued support from some of their children through the previous transfer of this asset. In Taiwan, for ex ample, of those living in a house owned by their children, 66 percent gave all or part of the cost or the house itself to the children, and this is true of 46 percent of the comparable group in Thailand. Home ownership by sociodemographic characteristics tends to vary in similar ways across countries, decreasing with age, increasing with educa tion, highest among married couples, lowest among unmarried females (largely widows), and highest in rural locations (reflecting lower housing prices and fewer housing alternatives than in urban areas). In addition, home ownership by the elderly tends to be lowest when they are living with married children or others. In each country there is a clear gradient of home ownership by self- reported health status, with those reporting better health having higher levels of ownership. The differentials are quite marked in each of the countries except Thailand. Sharp differentials in the asset situation of the elderly with health status have been reported in the United States as well (Smith, 1999; Smith and Kington, 1997). The interrelationship between home ownership and health is complex. Those with more education and economic opportunities may be in a better position than their counterparts to both accumulate assets and protect their Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 _u5.2obou•o_o>^.0Vuu"O.i%0lQ£_>-b•oUJ*—a.*JJ|JS£Vuoio£c c0 =15-°-aSs«.e£60 «-to « c~S 2u ~aJJ= tu00al-C»21'c3_C99 DCg|^- £UfCOB22JOU^^12 iMj5ea."ua*- TJC 0U C.t3 ew0- ffinl|^1 Ocub2JsUuH_O^c .l||*5> lcut: P c ^ # _e.0_M|iE tflUt^ - •^tsOv^ Ttm^^ -^Tj-NO06 t-^evr^riS "^22 2^°°2 S^g 2§g 2^^in ts ro r** o rs t*- ^^ oo ^ ^ *•*t- — o •* O Os*CN r-mvoOv OO ^ ^3 ^3 O O CN iO iO ^^r-* oo oo t^* wn ON r** v o vo oo oovQ in r*" ^ o oo ^^ oo ro ^^ ^^ ON O r^* ^^ ON NO ro CM ON oo fi— ' TtTtinr- ^^ovO — i Tt tsro NO o od oo r^ r^ ^ i/ei ^ Tje—* — i — . — . — iCN — • — 'm CNrririSTJ: CN-^Tf r"SCNrotf) r** O ^^ ON oo Tf r*^ ^^ oo ^^S g?S55 S5B3 £££0 _u eg o §eo, p u 05 ^ ur ._, 3 C *s• «« C. _*0 § J 1 1 B s0 eI .S S*•^ONONeC *S.r§2 '^wea-S _ ^vor^n yfc££ gja^^g3 uowioo o,i5£^ 3s>5^®0 M NC ^ . > >, ^ s 3 £ Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 able 8.11. continuech=p— f^**pinNO OOOinNO^tspooPlace of residenseR.ralm ONO 06CN roON NOON NOi-HnNO C-Tf CNNO roOO TtON t^m•* t^CN NO— i in•*t CNr^m ror- m.u 5 3| -aLiving arrangementsin r-Tt vdCN CNCN OP ONCN t^f^ fel-^ inin t**-roooWl vQ„NO^O mON CNOO NOCAlene/rt.ple only With married shildreCN TtON OCNfi NO•* -in ro•* CNp I-;ts inoo r-oq CNNO f^CNr-; Csvd r^iNO OOTtON OO— ts oo r-cWith .nmarr. childre Other**!00CN00NO•6ONNONOt^iooNO0tnOtHealth statusLewOO ONCN NONO in— ' OONO t*i•* oON -^r^ ooin r-;ON inCO OOO Ttm CNCN inr- ooa2 u•g-aS £Tt Tt — . inoo NO oo o— ^Tf O oo — i— i fS ts Or- NO -^ oqNO ro Tt r-in — i o NOro r- vi -Ht** 00 OO C-^* in r- r-;riS CN CN 00m r-; O invi i ON iri ciCN CNNO oq Tt pCN t^J CN TtON in -H ooSON ^ NO r- oo NO*)Inrtme tierLewst secend rt-hstMain seurce ef inrtms i Idren/relati vsNO ONO r-^p ONin CN.ON ONON OOOO fe^r^ eNO inr^ NOp ^*CN in.NO OOCN r-^oo r-Earnin-s frrt work Pensienrt sever, pay«NOTtinNOin00CNoo„^oNCNCNONNOoou uoie•~c8 Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 c ce 10003B5e3H&f1o1ol«"g*Cu5C£ I*SOBOBSJ 2Us"O (u-S- K1se8U OuO.K1 §^ ee3u _gSiuRlUC N O -^ ON *3e — ' r^ NO O -^ inin Tt •* in oo Tt^•* voTtin iiimOOTt —-tv o vOOTrCN -^CNCO^ -^CN't mts— ii1ev — iinooo r^vooo •*Ttinr^j ro -i- no 06 ON ON r^* in f i od ' ' 'CN CNCNCN— ' -^eNCN CNts— :.ts invdmin r-Ttin i/^^HS iii2 522S 335 g§3 ^r^i Tf i-H ^3 oo NO CN •* C^ r** ^^ ^^ oo v1-, r. ^, r* — i CN — i ^ — ,— .vo — • m ov es fjtsoo — ; in p oq— ^tsts — i — i ts m — i — • ro Csts— tso4— 'ON OONmoo iOOOCN NOONTt CNOOt^O ^ oo ro i0 CN OO ON ON NO ^ ro in NOm vd ^ Tt rs NO m - m ^ m vo •* •*- ** Sou _U g o _ 3•Q. e J * ^ •- °S.onSSOTS^'S _ H S3 >, -a 2r0 § ^ -C onC g T3 | .1 ^ £ ^^^ 3vor^§ crtEE w^^-r^o c* 2 _S £•*e* ^ ^^ "1 ^^ ^^ ^^ ^3 r^ f^ 3 jj t-Z ^Jt ** •£ _TJ •— ^ .-^H <' ^. ~ ' " ' ^* "". - ' S * . o n s Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Table 8.11. continuedOO O O •*r-^ d d CNCN •*O CN m Od ^ •* din — *in p CN r-;•^ CN — i mCN m -^r** r^- in ON^ ^ ^ r^vo CN m TtCN vq oq r-;^* ON re*- dCN — i •*ro ^* vq inTt r^ in vd-H Ov p pCN ON ^ m_ CN -i —•t CN in oqON m CN ONin Tt vO mLiving arrangementsAlene/rt.ple only With married children With .nmarr. children Otherin ro •*r^ in CNin r-; ONin t~ mm CN CNCN — ; ONd -^ vdCN CN CNr- ON oqvd CN vdm Tt Ttvq oq CN— H in int- 1 _ _^H co ovON — i oo— H Tf ^'r*i -H dCN CN CN-H oq in vd — i inm m mHealth statusLew Mederate rt-heBe « 55•E S 1-1 «^^-.m O^^vqts | £. 8C,8r-inTj mdoovd « fcSJ[5l"H -H " . ^* 5 LL« 2 ^.3 =II 1e^^OOCNTJ- tstsOr^ -S 'g o."^ SmTtvd TtONCN^ ^ a Sa2Ttm~- m ~^~^ - 1 £-3So § u £ * u i_ 2 f-i >vs. i Is5!Omt~ ooon^r*--^ ^ 3 ^Sa^ NO ^^ in ov ON ^ .- *o u « cCNCNCN ts—,-^—. ,= o -=21ac •§ •? ^.T3 a * u 0•3 E .E js g•S • 8 ||pmvq pinr-;p 3i-;.c S>v~in^r- r~doNin 2SW T3~eStsmm mvomin ^Ss .KSu502 £ is, -sit! ^ ^J J3 O 2s°:e E c « «in^in mr-;CNON 2«g ^.^ cvdodr- <^-H— ^O _2£_- iS^-a-H--H --r. r, g_g-3 g|.gBU 1 o |l^COro mTrooO — E-13 uoovdeNin oomoNr-^ j:|^ ^^Slf$? Ill•— >% CO ^ •*-U e2 e° 2 O ^3S a -u ,5 « T3inNOOO C^OvOvO ^c e^ONre-•^ odmodvd °eo# C-CMCN — i — i CN — • — CN !2 2 0 °t^,a1 S .2 « c £Sis 1 1 gC J= C a >,« u «J Sot ONvOin t--oomin 2-eU 2aor^Ttts od — din SeS afiSeminvo mr^inTt c.Ev- SicSt3 ~ e3 c e o.•£ *i *S i_ C/5 u H oo S oc S -g .E •£.E o-.E | S . » e8.Cco.CoMT3C1 -a & E.ga«sg|O UN M U *^ O •*? *S •-— i OJ.MJ «i«ifhe^|> |^|5 fil| r g£«fc" 2-c42.S^.2'SSs-as il&i" III ssili§.t lilli !§H M?|5«a |3|jg -i-. s-ja (o i 2 Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 340 The Well-Being of the Elderly in Asia health. Also, those with poorer health may not be able to sustain independent living, requiring them to live with children and others. These reciprocal ef fects can be sorted out to some extent with panel data that trace the same individuals over time. In the U.S., for example, panel data reveal that those experiencing diminishing health over time tend to have a lower net worth than their counterparts who remain at the same level (Smith and Kington, 1997). The remaining panels in Table 8.11 also reveal rather sharp differences in home ownership by the income position and income sources of the elderly. As with health, home ownership increases with income position, particularly in the Philippines, Taiwan, and Singapore. Also influential is the source of income; home ownership is lowest when the main source of income is from children or relatives versus sources that point to more economic indepen dence. As Table 8.4 and 8.5 showed, the sources of income are also related to the amount of income received. These differentials, though quite marked, should not mask the high level of home ownership that exists in almost every category or the claims that parents may have through joint ownership. The ownership patterns of unmarried females, mostly widows, across the four countries illustrate the level and suggest some of the ongoing dynamics for a group thought to be particularly vulnerable economically. In the Philippines and Thailand over 60 percent of unmarried women have sole ownership of their place of residence (regardless of living arrangement); in Taiwan, although only 19 percent of unmarried females are sole owners, another 37 percent are living in a home of a child for which they have provided full or partial financing; in Singapore 25 percent are sole owners but another 28 percent own the property jointly with children. Many of the transfers by older widows in Taiwan and Singapore probably occurred upon the death of their spouse as part of an arrangement of long-term support. The significance of this broad pattern of home ownership for the economic well-being of the elderly depends in large part on the values of the homes. Table 8.12 presents median home values by key sociodemographic characteristics and country, for the respondents who own their current resi dence for three countries which tried to obtain magnitudes. As the notes to Table 8.12 make clear, these values were obtained with somewhat different questions in each country and these differences should be kept in mind. It is also likely that many older respondents have lived in their homes for a long time and do not have a very accurate assessment of the market value. For these reasons the values given in Table 8.12 should be taken as a first estimate of the underlying values. Appendix 8.C reports on attempts to locate general house and land values for the three countries. Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Economic Well-Being 341 In the Philippines the median value of home and land reported by homeowners was 125,000 pesos, equivalent to about $4,900 U.S. This amount represents about twelve times the median income of all elderly economic units reported in Table 8.1. About two thirds of the reported value was the value of the land itself, suggesting that the housing structures themselves are quite modest. In Thailand, the median value of house and land is 270,000 baht, equivalent to $10,800 U.S. This value represents 27 times the median income of all elderly units, indicating that this asset is an important financial resource. As in the Philippines, more than two-thirds of this overall value in Thailand was due to the value of the land. Taiwan, in keeping with its higher per capita GNP, has an average house (and land) value of over one million New Taiwan dollars, equivalent to nearly $41,000 U.S.—or about ten times the median value in the Philippines and nearly four times the median Thai value. In relation to current income reported in Table 8.1, the median Taiwan house value represents nearly seven times the median income of all elderly units. By sociodemographic group, differences in house and land values gen erally follow the patterns observed for income (Tables 8.4 and 8.5) and for home ownership (Table 8.11), but with a few notable differences. Values are higher in the youngest age group in the Philippines and Taiwan, with a clear gradient by age in Taiwan. In Thailand, however, the older two age groups report somewhat higher values. Also Taiwan and the Philippines show the highest values for owners living with married children, but in Thailand that occurs among those living with unmarried children. In Taiwan, unmarried female homeowners have higher values than the other marital statuses, but for the Philippines and Thailand, married couples report the highest value. In all three countries, there is a sharp gradient of house/land value with educa tion level (similar to the observed gradient of income with education) and between amount of income, as measured by income tier, and home values. Those whose main source of income is from pensions or investments also tend to have more valuable homes, as do those with better health—which parallels the relationship found between income and health and the pattern observed in U.S. studies. A distinct factor for house and land values is location, with those owning a home in the capital cities reporting a median value about 10 times that reported in the rural areas in the Philippines, 6 times the value reported in rural Taiwan, and nearly 18 times the value reported in rural Thailand. It should be noted, however, as shown in Table 8.11, that the proportion of the elderly owning their own homes within the capital cities is quite low in com parison with ownership levels in rural areas. Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 O. NO tN iO —i — O o in •o .u in 0l o. •o I/1 vC = * ^ o e 3 ™3 0.2 §1 e8 eM !•— cd 3 O CN r- oo Tt 22 g" {rt e ec H •< o o oori^ I/X 25 CN CN So o o Q m oo OO O Ov CN CN CN CN CN — ^NONO S'cT — i ON ^. o.tS ov r^ o" o O Tttso 5 oo 25 oooo oovoTf o'oo TJ•O-H OS 1- oo vo in —i orio CNCNC^ OO mts J i a ?^ '» 00 00 OO vo m ^^ m in v cf OCNVT) tstsr- wn i0 e^ Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 ecu 0) .«5 t\i _aj x> ents Living arrang So -H •* CN r^ ro ro •* es to r** ro -H O 00 ^ CJN CO ro r- s o ro i0 CN m o o CN CN "* CN CO O ON —i m m r^ ro —' CN CN -H Alone/rt.ple only With married children e8 NO -H i0 NO CO ov — oo r- i0 O Tt ro oo -^ CO ON —i O CN •* ro -*t r- ro t** O CN ro ro CN r- vo 55 -H ON iO OO NO ON r^ r^* -H CN CN oo ro NO CN ro oo O oo CN ro ro oo oo r^ oo •^ CN r^ in CN OO ON NO CN CN i0 NO CN CN n n -* •H 8 .a jjf i It is •* C u JS .on3 O Tj U 5J* O •o tS C .2 ^ .a o -a « C ._ O — |N s - c s S 3 a 3 lilt*' 2 * ^-i « * i~ « c j= £ — -a u . L fa sis! e 5 J3 s* o 's •g £ = •£ S S u a H 0 ~* a H •= ^ •£ E E S 2 ° = j2 3 8 ° 1 • * •s TJ 3 o « t: gl SielS ts eZ o •>, r -• ill ?ll- ^5 g 8. 1 S • a ^ c i. ^ 5 . _- _^ si « e a Philippins a s were ebtai Notes: In th dependent va .3 ** t of a e of their al. pe r .ni val en the val.e r the stimated :s I I e5-" g I3l S5S. 11 'S J 5-a ?:•= g 1?! u Jo u- _c u >' •c -o .o E <-• « Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 344 The Well-Being of the Elderly in Asia Changes in Economic Well-Being over Time To this point we have presented a cross-sectional picture of the economic well-being of the elderly in four countries on various dimensions as of 1995- 1996. The sources and levels of income as well as the asset mix for the older population are likely to undergo several transitions over their later life cycle, as they move from work to retirement, change dwellings, encounter health problems, and respond to changes among family members. Relatively little is known about the changes that occur to the economic condition of the elderly as few countries have mounted the detailed panel studies that provide the necessary information. Taiwan, however, has reinterviewed the original panel several times since 1989, in addition to replenishing the sample at younger ages, and we illustrate the potential for studying change from these data. Table 8.13 compares the major sources of support in 1989 with 1996 for the approximately 2,400 elderly who were interviewed initially and sur vived to be interviewed seven years later. In terms of sources of income, over 70 percent of those who reported pension and retirement income as their major source in 1989 still did in 1996. For the other sources there is considerable mobility. Only about half of those who gave children as the major source in 1989 report this source in 1996, many of them attributing their major income at the latter date to pensions and retirement funds. It is interesting to note from the marginal distributions that despite the aging of this cohort, there is an overall decrease in the reliance on children over the seven years. This may well be a temporary phenomenon related to the special pension program provided in some cities and counties around 1995 for a short period (see Chapter 6). A preliminary analysis of cohort and aging ef fects on major sources of income shows that both factors are important. Reliance on children as the major source declines with age, within cohorts, and reliance on pension income increases. At the same time, younger co horts, controlling for age, display lower reliance on children and more reli ance on pensions than their older counterparts. Figure 8.3 shows the changes in the distribution of responses on the adequacy of income to meet expenses in 1989 and 1996, for the same cohort of elderly in Taiwan. Treating the data as two separate cross-sections gives the impression of very little change on this subjective measure of economic well-being. In contrast, Figure 8.4, which presents the 1996 distributions as a function of the 1989 response, shows considerable change. About two- thirds of those who said they had enough money with some left over in 1989 report themselves as having "just enough" in 1996. Likewise, there is some net shift to more difficulty for those reporting "just enough" in 1989. At the same time, significant proportions of those who experienced some or much difficulty in 1989 moved up to the "just enough" in 1989. Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 00 ON Ie vO ON ON ONijm se.rss,•• S3 1I.£* ^J•B |c °°i•a4l >. •> S c — opilsCQu .-— ^i c/T111 it.s^lsijs s uu *odj 4iOScU j-Hla001£eon fS°oo(2| m steck1M>ei£itN iO ONrii CJ ts^*r -- r^^ °° ts oo tH roPHSO ON NO ^^ iri •*r- ON ooo m oor^ ON CN f^ CN r-— ; Tt oovo Ti oiCN iO -HNO iT) O^ ^1 iE SiP .at*3 PW*V1 C "II!cfl J3 avBUS"f*^oovo0ONCNCNO 00§l. S^•uS xSi uO 0lQ. 3e3 tSs°P u o Sifc MImerte savin-rtONm§NOCNSC**OOONts tsMNU1CA1f0015s8NONCSt— itoON06 mON ON1 IM E? •o eo oo Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 vD ON ON tQ ON rt vO •C •g 0l •u CO 0l ON oe Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 £ *u 1- -C ^O vO W-t ^o vo S S suioouyo Aouaioiynsjo |3Asq 9551 Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 348 The Well-Being of the Elderly in Asia A similar degree of change in attitudes is found on the more subjective question of satisfaction with current economic situation. The table below shows shifts from satisfaction levels reported in 1989 to those reported in 1996. It is clear that many of those who were satisfied in 1989 became neutral or dissatisfied in 1996, while many who expressed dissatisfaction in 1989 expressed more neutral or optimistic views in 1996. At the same time the overall distributions in the two years were very similar. Taiwan: Percent Distribution of Satisfaction with Economic Situation in 1996, by Attitude in 1989 Attitudes in 1989Attitudes in 1996Very Satisfied or SatisfiedVery satisfied 67.4Satisfied 47.8Neutral 36.9Unsatisfied 20.2Very unsatisfied 6.5Total in 1989 40.9Neutral27.443.046.645.047.843.1Unsatisfied or Very UnsatisfiedTotal in 19965.2 10.59.2 35.416.5 37.634.8 14.545.7 1.915.9 99.9 Additional discussion of these data are presented in Chapter 12. The goal here is to illustrate how economic well-being is likely to fluctuate for many of the elderly. One has a sense that the shifts observed represent a complex adjustment process as some older people lose ground as earnings cease, but are able to call on children or other resources to compensate if needed. Further analysis of these transitions and their causes and consequences are clearly in order if we are to gain deeper understanding of the underlying dynamics, and fashion appropriate policies for assistance. Conclusions Attempts to appraise the economic well-being of the elderly have often de pended on inferences derived from their living arrangements and whether they were receiving money or material assistance from children or others. In an effort to obtain a more direct assessment, this chapter has introduced an array of income and asset data derived from surveys conducted in 1995- Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Economic Well-Being 349 1996 in the Philippines, Singapore, Taiwan, and Thailand. The results suggest that more than one or two measures are needed to effectively analyze the economic well-being of the elderly and fashion appropriate policies. The close connections the elderly have with their children and the high proportion that are embedded in other households require that both individual and household income and asset dimensions be ascertained. In addition, the responsibilities and allocations within households and the flows between households need to be taken into account. The main findings from this initial analysis reveal the following. 1. In terms of money income and proportion of income to GDP per capita, the elderly in the Philippines and Taiwan appear to be better off than those in Singapore and Thailand. But when levels of indirect support (which affect 50 percent or more of the older units in each country) and household income, for those who coreside with others, are taken into account, the relative stand ing of the older population in each country changes considerably. For ex ample, in a third of households in which the Philippine elderly coreside, they are the only income recipients; two-thirds of older Philippine units that are in the lowest income tier of money income are also in the lowest income tier of household income, when they reside with others. 2. The importance of these additional dimensions of economic status appear confirmed by the subjective measures of economic well-being. Over 60 per cent of Philippine elderly economic units express some difficulty with the sufficiency of their income, compared to a quarter in Taiwan and a tenth in Singapore. In terms of satisfaction with their economic situation, almost 30 percent of Thai older couples or individuals expressed dissatisfaction com pared to 17 percent of their counterparts in Taiwan. 3. Examination of sources of income reveal that children are the most impor tant source in all countries and across each type of economic unit, with earnings from work the second most important source. 4. Many factors are related to the level of money income, including age, marital status, gender, education, health, and urban-rural residence. Some what surprisingly in view of the emphasis placed on living arrangements, elderly units living alone generally have higher incomes than those living with married children; and those whose main source of income are children or other family members generally have lower incomes than those who rely on other sources. Although these factors are interrelated, they all tend to retain strong and significant effects in regression analyses (though to a lesser extent in the Philippines compared to the other countries.) Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 350 The Well-Being of the Elderly in Asia 5. A majority of the elderly in each country owns their current residence (alone or jointly) and in Thailand and the Philippines the percentage is around 80 percent. This is the most frequently owned asset and from one-third (Singapore) to three-fifths of the elderly (Philippines and Thailand) own no other assets. In the latter two countries, a significant proportion of the elderly also reports household debt. Involvement in the capital markets is negligible among these elderly in all the countries. 6. Though the overall asset position of the elderly appears limited from the data available, it is important to note that a high proportion, regardless of current living arrangements, maintain an important asset through sole or joint home ownership, or by having a claim for continued support from some of their children through previous transfer of this asset. Home ownership by sociodemographic group varies in similar ways across the four countries. And, although there are differentials in ownership by age, marital status, residence, and education, levels of ownership are quite high in almost every category—even among older unmarried women, a group thought to be par ticularly vulnerable economically. Sharp gradients in home ownership by health status and income level are observed in a number of the countries, paralleling patterns observed elsewhere. 7. In the Philippines, Thailand, and Taiwan, the median value of elderly homeowners' house and land represents from 6 to 18 times the median in come for all elderly, suggesting both that a home is a substantial asset for many elderly and that there is considerable variation in the asset-income rela tionship across countries. 8. The panel data from Taiwan indicate that the sources of income and the economic well-being of the older population may shift considerably as they age. Only by studying these older life-cycle patterns can we hope to obtain a full picture of their economic well-being and the forces affecting it. The multi-dimensionality of the economic well-being of the older popu lation and the many forces at work present a challenge to policymakers seek ing to fashion programs to aid those in need and sustain family sources of support. The use of simple measures may lead to both unnecessary program costs and ineffective targeting of those in need. Developing programs that correctly target assistance and build structures for sustaining economic well- being will require careful planning. Chapter 11 looks at identifying the elderly who are economically disadvantaged, along with policy and program impli cations; chapter 12 reviews the broader policy challenges for maintaining economic well-being among the elderly. Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Economic Well-Being 351 ENDNOTES 1. From a dynamic standpoint, one would expect ongoing selection whereby older economic units who cannot sustain independent households move in with others, leaving behind both those with better resources s well as those without opportunity for coresidence. 2. The much higher income in Taiwan of unmarried males living alone to those living with others most likely arises from Mainlander soldiers and officials who never married but are receiving relatively high pension income. 3. More formally, because the income of the economic unit is one component of total household income, one would not expect complete independence. 4. The nature of the debt was not ascertained and it is possible that some of the debt reflected amounts owing on the respondent's home, since home values in these countries were asked in terms of market value but not explicitly net of mortgages, see notes for Table 8.11. 5. The proportion with full control is lower among those reporting inheriting all or some of their assets than for the total, but it was still 64 percent for this group. REFERENCES Central Intelligence Agency. 1997. The World Factbook, 1997. Washington, DC: Central Intelligence Agency (sold by Superintendent of Documents). Chan, Angelique. 1997. "An Overview of Living Arrangements and Social Support Exchanges of Older Singaporeans." Asia-Pacific Population Journal 12 (4): 35-50. Cowgill, Donald O. 1974. "Aging and Modernization: A Revision of Theory." In Jaber F. Gubrium, ed., Late Life; Communities and Environmental Policy. Springfield, Illinois: Charles C. Thomas. Cummins, Robert A. 2000. "Objective and Subjective Quality of Life: An Interactive Model." Social Indicators Research 52:55-72. Diener, Ed. 2000. "Subjective Well-Being: The Science of Happiness and a Proposal for a National Index," American Psychologist 55 (1):34-43. Directorate-General of Budget, Accounting and Statistics (DGBAS). 1990. Report on the Survey of Personal Income Distribution in Taiwan Area of the Republic of China. Executive Yuan, Republic of China. Douthitt, Robin A., Maurice MacDonald, and Randolph Mullis. 1992. "The Relationship between Measures of Subjective and Economic Well-Being: A New Look." Social Indicators Research 26(4):407-422. Hermalin, Albert I. 1995. "Aging in Asia: Setting the Research Foundation." East- West Center: Asia-Pacific Population Research Reports No. 4, April. Hermalin, Albert I. 1997. "Drawing Policy Lessons for Asia from Research on Ageing," Asia-Pacific Population Journal, December. Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 352 The Well-Being of the Elderly in Asia Hill, Daniel. 1999. "Unfolding Bracket Method in the Measurement of Expenditures and Wealth." In James P. Smith and Robert J. Willis, eds., Wealth, Work, and Health: Innovations in Measurement in the Social Sciences. Ann Arbor: University of Michigan Press. Kurd, Michael. 1999. "Anchoring and Acquiescence Bias in Measuring Assets in Household Surveys." Journal of Risk and Uncertainty 19:111-136. luster, F. Thomas, and James P. Smith. 1997. "Improving the Quality of Economic Data: Lessons from the HRS and AHEAD," Journal of the American Statistical Association 92:1268-1278 Juster, F. Thomas, and Richard Suzman. 1995. "An Overview of the Health and Retirement Study." The Journal of .Human Resources 30(Supplement 1995):S7- S56. Knodel, John, and Nibhon Debavalya. 1997. "Living Arrangements and Support among the Elderly in South-East Asia: An Introduction." Asia-Pacific Population Journal 12(4):December. Kushman, John E., and Christine K. Ranney. 1990. "An Ordered-Response Income Adequacy Model." The Journal of Consumer Affairs 24(2):338-339. Li, Rose M., Yu Xie, and Hui Sheng Lin. 1993. "Division of Family Property in Taiwan." Journal ofCross-Cultural Gerontology 8:49-69. Preston, Samuel H, and Paul Taubman. 1994. "Socioeconomic Differences in Adult Mortality and Health Status." In Linda Martin and Samuel H. Preston, eds., The Demography of Aging. Washington, D.C.: National Academy Press. Shih, Shiau-ping R., 1997. "Private Lives within Public Constraints: Retirement Processes in Contemporary Taiwan." Ann Arbor: Unpublished Ph.D. Dissertation, University of Michigan. Smith, James P. 1999. "Healthy Bodies and Thick Wallets: The Dual Relationship between Health and Economic Status." Journal of Economic Perspectives 13:145-166. Smith, James P., and Raynard Kington. 1997. "Race, Socioeconomic Status and Health in Late Life." In Linda Martin and Beth Soldo, eds., Racial and Ethnic Differences in the Health of Older Americans, 106-162. Washington, onal Academy Press. Treas, Judith, and Barbara Logue. 1986. "Economic Development and the Older Population." Population and Development Review 12(4):645-73. U.S. Bureau of the Census. 1998. Statistical Abstract of the United States, 1998. Washington, DC. NOTES FOR TABLES Table 8.1 Philippines: Respondents were asked to give an exact answer for yearly income. A bracketing technique was used for those who had difficulty with this. Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Economic Well-Being 353 Thailand: Thai respondents were asked about exact yearly income. Those who could not give an exact amount were asked to choose from a given set of brackets income (as shown in the table.).) Mean and median income was calculated by assigning the midpoint of the bracket range to those who only gave bracket information and using the exact amount for those who gave exact information. Taiwan: Taiwan respondents were asked about exact yearly income. Those who could not give an exact amount were asked to choose from a given set of income brackets (as shown in the table.) Mean income was calculated by assigning the midpoint of the bracket range to those who only gave bracket information and using the exact amount for those who gave exact information. Because of the high proportion of respondents who gave bracket answers median income was computed using linear interpolation of the bracket categories. Singapore: Question on level of income refers to only the respondent's income. However analysis of the data suggests that in some cases R is reporting both respondent's and spouse's income. Because the Singapore respondents were asked about monthly income the bracket endpoints were multiplied by 12 to convert to yearly income. This will underestimate (or overestimate) income if income varies greatly throughout the year. Income mean was calculated by assigning midpoints to each of the categories, the median was calculated using linear interpolation techniques. Table 82B Sources: GDP 1996 per capita: Central Intelligence Agency, 1997 GNP1995 per capita: U.S. Bureau of the Census, 1997, Table 1347 for Philippines, Thailand, and Taiwan GNP 1995 for Singapore: Singapore, Department of Statistics, 1998, Table 1.1 (converted per rate below) Conversion of means and medians to $US based on following exchange rates (CIA, 1998): Philippines: 25.7 Pesos = $1 US Thailand: 24.9Baht = $lUS Taiwan: $27.4 New Taiwan = $ 1 US Singapore: $1.42 Singapore = $1 US " GDP per capita based on purchasing power parity calculations utilizing standardized international price weights so that $ 1,000 will buy approximately the same market basket of goods in each country. Tables 8.7A and 8.7B Indirect support is defined as follows: Philippines: Economic unit pays less than 75 percent of household expenses Thailand: Respondent or spouse is not main provider for household and does not pay all the household expenses. Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 354 The Well-Being of the Elderly in Asia Taiwan: For those living with others: Economic unit shares household expenses or reports that someone else pays most or all of expenses. For those living alone: Non-household member pays part or all of household expenses. Singapore: Economic unit reports receiving accommodations or rations from children or maintenance in kind from anyone other than spouse. Table 8.9C Notes: In Thailand, the question was, "Generally, how satisfied are you with your current financial or economic situation?" with the options "very satisfied, satisfied, not satisfied," recorded. In Taiwan, the question posed was, "In general, are you satisfied with your current economic status," with the options "(1) very satisfied, (2) satisfied, (3), neither, (4) unsatisfied, (5) very unsatisfied," recorded. In the table above options (2) and (3) are combined into "satisfied" and options (4) and (5) combined into "not satisfied." Table 8.10 Line 1 - Represents joint or sole ownership of house and/or land where elderly reside; see Table 11 for detailed notes. In the Philippines only 3.6 percent of respondents owned the land only; in Thailand the comparable percentage was 3.2 percent. The separation is not relevant for Taiwan or Singapore. Line 2 - Represents house or land ownership other than that of current residence. Line 3 - Does not include the farm or ranch on which the respondents reside. Line 7a - Represents the proportions who do not own any of the assets listed in lines 1 through 6. For Singapore, also reflects lack of ownership of the assets described in line 9. Line 7b - Represents the proportions who own none of the assets listed in lines 2 through 6 (including line 9 for Singapore). The respondents may or may not own their own home. Line 9 - Represents a mix of other assets. In the Philippines it covers jewelry and appliances owned by the elderly economic unit. In Taiwan, it covers only ownership of jewelry and similar valuables, like artwork; in Singapore it includes ownership of annuities and endowments, current accumulations in the central Provident Fund and coverage by other pensions. Line 10 - Represents the proportion of the elderly living in households with one or more basic appliances such as refrigerators, electrical/gas stove, television, air conditioner. The list varied across countries. The elderly do not necessarily own the appliances, especially in situations where they live with others. Line 11 - Represents the proportion of older male and female respondents who have collected pension payments in the past, are currently collecting payments, or expect to receive pension payments after retiring. Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 ob --i T3 e0 bfi 0l to tu vee . o CO 0l Cv Q. •0 1l!CS00ci/5O uJCOO c•5 ."*S HcIua.1T3^ _1 100 fS.5DMB.5.o.£Charsteristis^ vO OO vo O_; O O -^e •*}• in rs oz S S 5. 2— i —in o ve ofc* O — ' —I ONCN CN CN CNe in vo vo r\1 2 O in r- m— '^o vO in vO in^ NO vo ^^ r^^a oo — i m —ZNO { ^ CN t**) CN O CN Tt— — i — .wg O -H OO CNT^- in vo t^6 ts O • * fl•oM5e Mi^_ ON ON« s ^ ^ +u in o o ao V vo r- ooON in CNCD O -Hin oo rs^** ro in—i —in m CN— i Ov inrs — rsO ve ro — ' ON OO °° m in••*vo in ^^ve rs m•* CN — i ON -t vo^— ' •H•* p ONiO f) NOCN Tt NOeZ UQI a> ec3Marital status/-Married rt.ple Unmarried mal Unmarried femvo CN mo — ts— ' 00 COin r- rof^; 00 COCNr- rs eNNO O -*trs rs —^^ O m r-. vo ^^ ON m ro—^ro ON pro r^- C*^vo ON oo ro ON Tt— ^ CNro Tt — •-^ e^ ™^— 'r- ve ^^ o in CNrs r- rsbOEdusatien ef R husbandNe fermal Any primary sertndary +d. d. d.C C CH- 1 ^^ HHoq r^ vqr*ei O fiCN CN tsr** ov ?T) rSSSS—*rQ — mV) ve CN ro in O p in froCN Tt rots ro inroO ON Q ON ro OvO •* — i& Ue _ c 2 « 1 -fi &• j 3 t- Cfle a! ^ U Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 T3 CO •5 Q. O. 1ci/2i8c*"e*60 '-.£ ="s .»S H o55ZT3e5> ec^ !i.2? HicCOI£Characteristis., t"1 °° 1^ rs^ rs o o ri. CN CN CN ONO Tf TT ^ rojr m oo in tN^^ '^ t-; 00 vq vp•^ 06 in r^ Tt-H CN — I — I. o *O v o ooI £ £. S 2•-<NO r*** ^3 ^^J* ^ ^ O v£J-H ^H^ ! vQ r^ Q i/1-^ -v^O ^^ ^ CO e^* CeJ C*^ v& roO Tt i^i ^^ -H^y oo ro ^^ ^^ «-• _H Tt m CNar- **.Living arrangemeiAlene/spe.se enl) W/ married child W/ .nmarried W/ etherr- i=i. rsd -^ -i-t -V f-£ 3 Inrs"ro Tt NO06 -^ dCN CN CNTf -r inCN NOs oo— ir~ — < ro06 d ooNO ON m 323CN — iNO a; oqTT ^3 ^^oo in — ' OO •* ON— -1 OO ^^Health statusLew Middle rt_r-; ts 00 rod ^ -H to^*T 3 CNtoO; ro -^ inm CN m vdCN CN — i — *O ^ ro ON NO r- Tt m O in 0v CNm m rs ^^^ ts ro ^^^*1 NQ ^^ f^CN -*t — i CNCN -*"ro to O -Ov ro in in— • ts r- TfTt CN -H CNbI 1ei u MCJ M *-• nnMajer seurse ef insildren/ether rel; Wmk/b.sinss Pensien/retiremen Invstments savinjqm^roON__;CN00 00r-CN"T}-o!groNOin§1 'g T3 Z oo £ i g h- Cfl •s Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Appendix 8.B. Alternate Specifications of Factors Affecting Being in the Bottom Tier of the Income Distribution Panel A: Odds-Ratios for the Effects of Demographic, Socioeconomic Characteristics and Health Status •p<.10 *p<.05 **p<.01 ***p<.001 Inap. = not applicable Age of R or husband<6565-6970-7980+Marital status/genderMarried coupleUnmarried maleUnmarried femaleEducation of R or husbandNo formalAny primarySecondary +LocationRuralUrbanCapital cityLiving arrangementsAlone/spouse onlyW/ married childW/ unmarriedW/otherMajor source of incomeChildren/other relativesWork/businessPension/retirementInvestment, savings, otherHealth statusLowModerateHighModel Chi-Square(degrees of freedom)Odds-RatioPhil-1.171.341.58+--1.131.90***—.7629***--.86.831.16—.79.67+—.59***.17***.30+...75.52*126.2(17)***Thai—1.24*1.80***3.16***-2.01***2.30***—.70***.53***--.36***.27***.82+—.57***.67**—.22***.04***.16***-..81 +.48***1,273.3(17)***Taiw-1.131.35*2.49***-1.78***2.23***—.61***.19***—.40***.27***.62***—.55***.66—.22***.51***.47***—1.10.69+596.0(17)***Sing—1.141.69***2.84***-1.84***1.44***—1.02.34***Inap.Inap.Inap..70*—.61***.99—.06***.691.81**...54***.64**506.2(15)*** Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 •o I/) _o "— • tQ •o OQ CO ce a. a. o•s5•311t/i —•8s1B1His•«,i115eS1.5.o.i-1* *# * *# # *r^ o oo ; Tf in oo* ** ** *OO CN CO J ^^ OO fl* » * «+ # *O OO ON*•* *i — Tt NO_ _ ri# » »* * *# # #^ m r^; m oo vo~ «s uS* ** *. * . CN OO tftI CN oq CN— — r'!# ** *CN fI oo • to t~ O— ^ CN+ *: I] 2 -;•oA-e of R or husban<_ ___ ___8_+tMarital status^*-end***i OO**•* *•: 3***i OO' .-_J***: S****: Ss^: °vMarried meuple Unmarried male,*00***in* **fi^**r-r I***r-rt**«•OO CNCN* #*£CN*ON OOUnmarried female***OO ^*— VI: P **)* * * ** *# #* ** *. s s1# ** ** #: ^ °^# *# ** *CN —f ** *: £ £* *"."'!3Edusation of R or bNo formal Any primary Semendary or moreo. o. o.a a ed_c ^ _cfff ,* * * ** *, O —' v) **}* ** ** *, O ^ i . Tj; CN* ** ** *1 vO r-* ** ** *:SS i' ON CN•Mi oo oo iS »ti >0 onsLocationRrtal Urban Capital sity Major sourrt of insChildren/othrt rela**s****^* *%.1 1P***$*1* **r-Pension^*retirement***OO**4^***2h1Investments saving*1n00* *§Tf* *iri**c•SW* **S0000*5—* *&o* *5BModel Chi-Square (degrees of freedo 2? v a. o.g Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Economic Well-Being 359 Appendix 8.C. Notes on House and Land Values An attempt was made to find independent house and land values for each of the three countries shown in Table 8.12, in order to validate the levels reported and if possible compare the housing values of the older people with the general population. This proved very difficult; apparently sound indicators of housing values are not a standard aspect of the accounting framework in these countries. Neither regular surveys nor systematic reports of houses bought and sold could be located. Complicating the situation is the great disparity between housing values in major or capital cities and elsewhere, as reflected in Table 8.12. Partial reports that are location-specific or weighted in non-representative ways can greatly distort the averages and hinder comparisons. In Taiwan, the annual survey of personal income (and expenditures) (DGBAS, 1990) asked land and house values for each household through the early 1990s but this question was discontinued according to a personal communication because the values reported appeared too low in terms of current market values. Many householders apparently reported "registered values" (akin to assessed valuations), which were not reflective of the market. An analysis of the values reported in 1990 found that households in which the head was 60 or older reported median values very close to those reported in Table 8.12 for older respondents who own their home. Assuming this is the most appropriate comparison group and allowing for the increases in overall values since 1990, suggests that values reported in Table 8.12 might be a third lower than those that would be reported by older household heads in the Survey of Personal Income circa 1996. This must be treated as a rough estimate and it should be recalled that the values in the Survey of Personal Income are themselves assumed to be too low. In the Philippines, the only data located were reports from the Housing and Urban Development Coordinating Council on housing loans (mortgages) taken out under the mortgage financing activity of the Home Development Mutual Fund, a government savings institution for housing. In 1996, the overall mortgage loan value for houses, houses plus lot, or lot only was 210,000 pesos, some two-thirds larger than the value shown in Table 8.12. But almost 60 percent of the loans were in the National Capital Region and the average value of these loans was only 225,000 pesos, considerably less than the average for Manila shown in Table 8.12. This suggests that this program has a limited range of loan values that may not be representative of the broader housing market and offers little guidance as to the accuracy of the values shown in Table 8.12. In Thailand, like the Philippines, only a few specialized reports could be located. A table from the Agency for Real Estate Affairs showed for 1997 an average value of 3.3 million baht for new single home units for Bangkok and vicinity. This is somewhat lower than the value reported in Table 8.12. The same table however reported an average of 1.25 million baht for the remaining provinces, much higher than the non-Bangkok housing reported in Table 8.12. As the column of numbers Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 360 The Well-Being of the Elderly in Asia of respondents reporting shows, however, the data in Table 8.12 is heavily weighted by older rural residents; the data from the Agency for Real Estate Affairs, on the other hand, may emphasize urban housing, according to one communication. More fugitive data from Thailand showed an average value of 4.1 million baht for second hand homes sold in 1997-98 (presumably in Bangkok) by one company; and a distribution of "homes for sale" in Bangkok and vicinity for 1996 indicated a median value of about 833,000 baht, but it could not be determined how representative this base was and whether the value of land was included. Overall, the small amount of data located suggests that the values reported in Table 8.12 for Bangkok are not greatly out of line, and that the values for other urban housing may also be reasonable estimates. Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Chapter 9 PHYSICAL AND MENTAL HEALTH OF THE ELDERLY Zachary Zimmer, Josefina N. Natividad, Mary Beth Ofstedal, and Hui-Sheng Lin Central to the study of gerontology and influential in its development have been the changes in health that occur at older ages and the implications of these changes for social and medical services, family relationships, and public programs. It has long been recognized that mortality rates increase at older ages, along with functional limitations and chronic diseases. But the level of the underlying rates are far from fixed and there is considerable variation over time and across groups, prompting intensive study of the em pirical levels as well as underlying mechanisms from the standpoint of biol ogy and biodemography (see Wachter and Finch, 1997). One facet of these studies has been to develop a detailed profile of the health status of the elderly population as the basis for gauging both their physical and mental well-being and the social and economic implications of a growing older population. In Asia, where the proportion of the population at older ages has been low until recently and where older people traditionally resided with their chil dren, fewer detailed studies have been conducted on the state of their health. In addition, the health infrastructure there and in most developing countries has been oriented toward problems of infectious diseases and maternal and child health. An important start toward collecting more systematic data on the health status of the elderly was a WHO-sponsored study carried out in 1984, Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 362 The Well-Being of the Elderly in Asia which surveyed older people in the Philippines, Malaysia, South Korea, and Fiji (Andrews et al., 1986). A second WHO-sponsored study in the region in 1990 included five additional countries (or specified provinces) (Andrews and Hennink, 1992). These studies, while significant in directing attention to the health of the elderly in the Asia and Pacific regions, were often deficient in terms of sample size and representativeness, limiting the inferences that could be drawn at the national level or for cross-cultural comparisons. Several other multi-country surveys of the elderly in Asia have also included some items devoted to health status (for a review of these surveys, see Andrews and Hermalin, 2000). These include the ASEAN-sponsored surveys of 1984- 1986 (Chen and Jones, 1989), which we have drawn on in several chapters for over-time comparisons with the surveys reported here and will use again below for assessing changes in health status. The surveys conducted among the four focal countries of this study in the mid-1990s each contained a wide range of questions about health con ditions. Although the specific measures and amount of detail varied some what across countries, the data permit close comparisons on many facets of health. This chapter exploits these data to examine several health-related be haviors and compare the countries on physical, emotional, and cognitive di mensions of health, as well as on differences in levels of self-reported health. In addition to overall country differences, the chapter examines how key sociodemographic groups vary in their health and health behaviors within and across countries. Methodological Issues Much of the data collected in gerontological research on the health of older adults, especially studies that are representative of the broader population, comes from responses that older adults give to survey questions. Accord ingly, the validity and accuracy of these self-reported data are dependent upon older respondents' understanding and knowledge of their health condi tions and their ability and willingness to report this information accurately. Despite careful attempts at questionnaire wording and at building good rap port with respondents, any health-related data collected via questionnaires have some inherent limitations. This section discusses some of the method ological issues involved in collecting and analyzing this type of data. As noted previously, health is a multidimensional construct and infor mation on several different facets was collected from the respondents. Physical health was measured by asking respondents whether they currently have Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Physical and Mental Health of the Elderly 3 63 specific medical conditions and/or limitations in the ability to undertake sev eral basic activities of daily living (like walking, dressing, eating, and bathing). Emotional health was examined by asking respondents about the presence of several indicators of depression or by their level of loneliness. Respondents' cognitive level was assessed by asking them to perform certain tasks such as word recall, and/or by their ability to correctly orient themselves in terms of reporting their age, the date, day of week, and so on. In addition, each survey obtained an overall assessment of health by asking respondents to rate their health on a four- or five-point scale. The wording of questions used in each country often varied and for this reason we have included a special section that lists all the questions used and the definitions employed in the analysis. Major variations are also noted in the discussion and table footnotes. These several dimensions not only represent different domains of the complex structure of health, but they also intersect to varying degrees with alternate perspectives on health. The literature often distinguishes the medical definition of health, which stresses the absence of illness (approximated here with the information on chronic illnesses), from the social definition of health, which stresses how well individuals function within their social environment. The latter of these perspectives implies that two individuals with the same chronic conditions might be viewed as having different levels of health, de pending on their physical and social environments. This insight is most rel evant when appraising the implications of different reported levels of physical functioning, because the degree of difficulty experienced in carrying out dif ferent tasks will depend on the physical environment (e.g., the ease of bath ing will be affected by shower accessibility) as well as social expectations and traditional practices. The social aspect is also likely to be tapped by ques tions on loneliness and sadness and other items designed to measure emo tional health. The information on health status used here is subjective in the sense that it is reported by the respondents. But the questions on emotional health and overall health are also subjective in the sense that they ask respondents to make assessments or describe feelings rather than to report behaviors, to undertake tasks (as with the word recall questions tapping cognitive health), or to report diagnoses given to them by health practitioners. This subjectivity should be kept in mind when reviewing the dimensions of health reported here. For example, self-assessments of health are likely to depend on re spondents' expectations about what constitutes good health, the comparison groups from which the respondents make their assessments, and inherent cultural and linguistic factors. The small amount of research that has been Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 364 The Well-Being of the Elderly in Asia conducted on these factors suggests that they do affect self-assessments of health. For example, Angel and Cleary (1984) compared reports of self-as sessed health given by Hispanic Americans and found that the results differed depending on whether the questionnaire was answered in English or Spanish. Other researchers have found differences in the ways in which men and women respond to self-assessed health questions (Wolinsky and Johnson, 1992). Also, it has been shown that people tend to rate themselves higher on health with advancing age (Cockerham, Sharp and Wilcox, 1983), presum ably because their reference group is subject to additional health problems. Thus, an 80 year old inteniewed in the rural Philippines whose reference group is comprised of relatively poor farmers may have a different evaluation of his or her health than a 60 year old professional interviewed in Taipei. Even the more "objective" health measures used here are likely to be influenced by broader social and economic forces. Most important, reports on the presence of chronic diseases depend heavily on the degree of access that older individuals have to medical personnel that can carry out the relevant examination and provide a diagnosis. The health infrastructure varies from country to country, as described in Chapter 3, and also tends to vary across regions within each country. The rural elderly, for example, may have fewer health facilities nearby and less access to medical personnel than urban dwell ers. On the other hand, special health initiatives such as Thailand's program of free exams to all elderly or the universal health insurance program started by Taiwan in 1995 may mitigate these differentials to some extent. In addition to these broader environmental aspects, the individual char acteristics of the elderly may greatly influence their access to and use of health facilities. Those with greater resources and knowledge can be ex pected to have higher utilization rates than others, unless offset with special programs of the type noted above (see Chapter 10). Workers in more formal sectors of the economy or working for larger firms may also have relative advantages in terms of health care access than those not so employed. Edu cational differentials also come into play in measuring cognitive level, since performance in word and number recall tests tends to be highly influenced by educational level (Uhlmann and Larson, 1991). The relationship between socioeconomic status and health has been subject to considerable study in the West (particularly in the United States and United Kingdom) and is receiving increasing attention elsewhere. Although a strong association has been demonstrated between socioeconomic status and health, there has been some debate as to the direction of causality. Many analysts have held that socioeconomic status affects health outcomes, particularly at younger ages, but others have suggested that the direction of causality runs the other way at older ages, or that there are underlying individual Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Physical and Mental Health of the Elderly 365 traits that influence both health and socioeconomic status (see discussions in Anderson and Armstead, 1995; House et al., 1994; Smith, 1999; as well as in Chapter 8 in this volume). Models used to evaluate the effect of socioeconomic status on health have identified several underlying mechanisms through which this effect may occur. These include a set of psychosocial factors, including level of social support, feelings of self-efficacy, psychological well-being, and stress; health- related behaviors, such as smoking, drinking, exercise, and nutrition; as well as health-related knowledge and access to and utilization of health services. (See House et al., 1994 and Anderson and Armstead, 1995, for a more de tailed description of these models.) Zimmer (1998) tests aspects of this model for the influence of education on functioning in three of the Asian countries under study here—the Philippines, Taiwan, and Thailand. Although in this chapter we do not carry out a formal test of this model, several of the analyses we present are relevant to these consider ations. The main portion of the chapter examines health differentials on sev eral sociodemographic characteristics—age, gender, urban-rural residence, marital status, and education. These factors may have indirect effects on health status through access and other intermediate factors mentioned above, and they may also exhibit direct effects. In addition we provide evidence on the nature of several risk behaviors and their relationship to sociodemographic characteristics. (Related data on health are given in Chapter 10, which exam ines patterns of health care utilization.) Measures of emotional well-being may also be dependent upon cul ture, language, and expectations, although, again, very little research has been conducted to verify this. For instance, a question such as "Do you feel lonely?" depends on an interpretation of what it means to be lonely, which might be different in a society where older adults tend to have large families and coreside with children than in one where they tend to live alone or with a spouse only. It also depends upon the survey's success in translating the question into different languages in a consistent manner. Along with the measurement issues discussed above, we need to rec ognize that, although the set of surveys being analyzed were constructed to be comparable, each country had a fair degree of independence in finalizing the way questions were asked. As such, many of the questions are similar but they are not always asked in a uniform manner. For instance, the question "How much difficulty do you have walking?" refers to 200 to 300 meters in the Philippine and Taiwan surveys, and 1 kilometer in the Thailand survey. Clearly, one would expect a higher proportion to report difficulties walking the latter distance. Nonetheless, we might expect that an individual who re ports "a lot of difficulty" walking 1 kilometer would also have problems Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 366 The Well-Being of the Elderly in Asia walking a distance somewhat shorter than that. As noted previously, question wording for each of the measures is provided in the technical notes in Appen dix 9. A, and major differences in question wording are also noted in the table footnotes. These various methodological issues should not deter us from the important task of comparative analyses of health status among the elderly and do not undermine the analyses being conducted in the Asian setting. How ever, we must be sensitive to these issues of measurement and perception when interpreting our results. What might appear to be large differences in health between individuals in different settings may, in fact, be the result of differences in question wording, interpretation, or perception. In the discus sion that follows, we strive to provide an appropriate balance between these considerations. Health Risk Behaviors As noted above, health risk behaviors have been advanced as powerful inter mediate variables accounting for differentials in health status among socio- economic groups. We thus begin the analysis by focusing on several risk behaviors that are of particular relevance in the study countries, including consumption of alcohol, tobacco, and betel nut. The surveys used in the analysis included questions to ascertain the prevalence of smoking and drink ing among older adults in the Philippines, Thailand, and Taiwan, and of betel nut chewing in Thailand and Taiwan. For the present analyses we defined drinking as consuming two or more drinks per day on average, to reflect moderate or excessive drinking, and smoking and betel nut chewing as con suming any amount. Examining these risk behaviors of older adults is important for several reasons. First, these types of behaviors can be detrimental to health. Smoking and betel nut chewing are associated with increased risks of heart disease, hypertension, and certain types of cancer, and excessive drinking is associ ated with heart and liver problems. In addition, these risk behaviors may be related to other types of behaviors or risks for which information is not available, such as dieting, exercising, and obesity. Finally, there tend to be strong socioeconomic and demographic patterns to risk behaviors in Western industrialized societies. For example, in the United States adverse risk behav iors are more prevalent among persons with lower education (Pamuk et al., 1998) and men (National Center for Health Statistics, 1999), and in the United States and United Kingdom among in those who live in rural areas (House, Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Physical and Mental Health of the Elderly 367 Lepkowski, Williams et al., 2000; Morgan, Armstrong, Huppert et al., 2000). In this analysis, we examine whether similar patterns exist in Asian countries. Table 9.1 presents the percent of elderly respondents in each country except Singapore who report engaging in each risk behavior, by sociodemo- graphic characteristic. The table provides results first for the total popula tion, then for males and females separately. The data suggest that smoking is most prevalent in the Philippines, followed by Thailand and Taiwan. Drinking is also most prevalent in the Philippines followed by Thailand, and least preva lent in Taiwan. Betel nut chewing is much more prevalent in Thailand than in Taiwan. In fact, chewing is more common than smoking among the current cohort of older Thais. With regard to differentials on characteristics, several interesting pat terns emerge. In general, men have a much greater tendency than women to engage in these risk behaviors, except for betel nut chewing in Thailand, for which the prevalence is about five times higher for women than for men. (This will be examined in more detail below.) These risk behaviors tend to be more prevalent among younger as opposed to older elderly, which may indi cate cohort differences in the likelihood of undertaking these behaviors or may be an artifact of selection due to longer survival of those who abstain from risk behaviors. Again betel nut chewing in Thailand is an exception in that it is more prevalent among those age 70 or older. Risk behaviors also tend to be more prevalent among rural than urban residents. The patterns for smoking and drinking by education level are more complex. In the Philippines and Thailand, higher education is associated with a lower prevalence of smoking, though not necessarily of drinking. In con trast, in Taiwan those with the lowest level of education have the lowest prevalence of smoking. This surprising pattern for Taiwan is largely a com positional effect in that the prevalence of smoking is extremely low among older women in Taiwan, who make up the bulk of those with no formal schooling. In fact, higher education is associated with a lower prevalence of smoking among Taiwanese males, consistent with the overall patterns for the Philippines and Thailand, while no significant association was found among Taiwanese women. Another interesting pattern emerges when comparing betel nut chewing and smoking in Thailand for men and women. As noted, the prevalence of betel nut chewing in Thailand is high, with more than one-third of older adults engaging in such behavior, including nearly half of those age 70 and older and of those with no formal schooling. Closer examination of the data indicates that betel nut chewing is primarily a custom among females in Thailand, and especially among rural women with less education. To some extent, the Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Q. 0l •o .u •8 C* ea _c DD .So to DC g a -Q CO = M on e in 3 00 2s o u _c BQ u •a (A .£ on co ts ts ro vD CN OO NO NO — iwnve eN — • oi o ^^ Tt - - CN —i CN t ^ -^ ON OO O OO ^^r-;—^f-; OO Tt OO OO , - . O in ON ono tsts-H rsts ^^-^fi in^oo od-*tr- CN tsr-mo f^ CN — i OC- — vdr^o. O OO rO O r-; -H OO m CN CN m CN , " OO CN —i CN —i m P-i vo ts oots— iev rs m CN — •i — 2 c a &»! ll! 's Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 § 3 00c ceu *uCQ -gc ooc1 .-IW .5H Q00i003 00C .£2 u1) £03 uT3 00C C.(3 c1 Sooee00oocI^ ^^a.— ooJS Co, 300u•sMUS~s .— s§I*e~ O'"•* Tt O Ooo m m v— ' Tt tN oo *— . — i — • mON ro— ' rN O ONO O O ON m — * r- — tm m O *—' ro CN — * *—.^-v ^oo O ON Qfl IO ^^ f^1 tNl 1O e-^ ^5O^ ON v*i ^"^ ro ^^ r*^ N*-/ Tf ro ^^ •* i^i/— N . — ^viw-i-O rimooSt ON ^ . — . Tt m — * N— ' —•vO — CN^s ^r^oomg tsoooJSTJ•veooo ts — i O ovn ^3 *O (^ •—* •* ON ps0 — i Tt O ve oo i0 Ove r-ON ^^ i/~i O lnmOts CNTtOroi0 ts — i --^ rs in — * ^^— * ~4 — • — *^-^ ^-sCS — i CN QTJ•CS^O ONTtvOO— • o ve •— . r-rso^1O •* CO IO ^^ON JJ Si -*- S3 ^ ^ ^O O « D « 3 ^7^^^ aet3tf.5M W) -l) _r" i u 3 _H i«<<uS DoJuS/ ^O QO m oo Tt or- ^^ — i oo ^-- tsoer- oo— i CN Tt Tt •*rs m m —^^ve o o r- oO oo r- o •— •5,m O rs ^^ i0 oo ovi — • -^ T* *-~-^ts Tt r- oo 2ts ts r~ m oP vO m ° gi0 vei ON ON v_jNO IO CN vO*—*r**^ ^p vO Tt — CN Om oo ts Tt ^— •—• — ^ —t^oo o vO t^* ON O OTt Tt vO OO e-^vo Tf m ^^3 tj= C"u >, ni« C^ T3i 1 I i 3^£i"l° C C -= Q. Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 c a- 1 S3!} 109 e5c M§ s1 ^H i00cB003 ooe ineo3 *1 jgPQ OT3 60S c1 160cae0000. =sD .C •S QQ.e5 0412 eCu ^2cooy SI uedIU/ — ^OOSTt OO.*t* ^o c* ^*™*'COO ooOOO*^i— 'oo r^^O Tf ^^ NOTt Tt O ^^00 8oo ex 22 ^^Tt vO Of^^O NO CN NOm rs O •— 'C^1NO o — i •*— ON -*t ^^ON oovo m -* n-in -t o ^^ONSu-i r-Tt m O v-'CN CN?*!«o o 2. 3^ ^ fe*- t!n *3« D tU .Ji, >B OJ1 OJ) j= 'I « u 3£gOv vO — 8e ON -0 — 'iO^Te-H CO CN OO •-* OO v— '](-^800 •*Tt •* I—' ^^rsS?^ooCN -HO — i in rsCN CO — i *-^oov^ ro in ^^r- ^H i0s^Sf— i O ro CNTt NO -H ^^COiO QO ON ON Ooo r- oo ^— '-" CNIf_D 2 — >eD (S U -S9o in in in om in o oo • — .vO es rs — rs r-O -H ] in ein oo vn oo CNm •*t vo CN ^-'2 £ "* S; wrs00 Qrs o ts m 25Tt CN O CN e-'i0 r- m No v-^oo O CO '•^ roro No m CN ^§' "*oo CN ON ro e— .CO CN CO2_e >,u >, c3M U T3— S c 1 ^S ^ U 3-3.5 o. S S^-^•s• >v >%.j. >O C C JS 'Z «U 3> Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Physical and Mental Health of the Elderly 371 prevalence rates of betel nut chewing among Thai women mirror those of smoking among Thai men, as the following data make clear: Betel Nut Chewing and Smoking in Thailand, by Selected Characteristics SexWomenMenEducationNoneNoneHighHighNoneNoneHighHighResidenceRuralUrbanRuralUrbanRuralUrbanRuralUrbanN1,13036152150357103176260% Chewing702419414372% Smoking1696171324035 In this table, "high" education is defined as secondary or higher (ex cluding primary education). The table shows that in Thailand 70 percent of low-educated rural women chew betel nut while 71 percent of low-educated rural men smoke; and only 16 percent of such women smoke while 14 per cent of such men chew betel nut. Betel nut chewing is almost nonexistent among higher educated urban women (only 4 percent), as is smoking (1 percent). Among men in urban areas, the proportion chewing betel nut drops to 3 percent or less and the proportion smoking drops to about one-third. Also notable in Thailand is the increase in betel nut chewing with advancing age: fully 80 percent of women age 70 or over who have no education and live in rural areas chew betel. (Data not shown.) Thus, there appears to be a strong cultural inclination among rural residents in Thailand for women to chew betel and men to smoke, particularly among those with little education. Physical Health We turn now to an examination of the health status of older adults, beginning with their physical health. Physical health indicators include both chronic conditions and difficulties in conducting functional tasks. As noted previ ously, in examining chronic conditions we are looking at health from the Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 3 72 The Well-Being of the Elderly in Asia medical perspective, while in examining functional tasks we are looking at health from a social perspective. With chronic conditions it is useful to distinguish those that are life threatening, such as heart disease or cancer, from those that are often debili tating but rarely cause death, such as arthritis. Although some older people may be well aware that they have life-threatening health problems, others may have them without clear indications or diagnoses. For instance, some individuals may have life-threatening conditions that are asymptomatic in the early stages; others may have symptoms but lack the health care access necessary for a diagnosis. Also, because in some cultures including the Phil ippines diagnoses for serious illnesses such as cancer are commonly with held from patients, a professional medical diagnosis does not always guaran tee patient awareness. On the other hand, the debilitating conditions that we consider here (arthritis, incontinence, cataracts, vision problems, stomach problems), by definition, are more readily apparent to an older individual through their discomforting effects. Hence, we might consider the following cross-classification of chronic diseases: Life threateningDebilitatingEasy to detecte.g., strokee.g., arthritis, incontinenceDifficult to detecte.g., heart disease, cancer, diabetesnot applicable The importance of such a classification is that conditions more diffi cult to detect are more likely to be underreported by those with limited access to a health care system, a situation common among many older adults in developing countries. On the other hand, we would expect less reporting bias for conditions that are easy to detect. Thus, we must be cautious when interpreting differences in prevalence rates of self-reported chronic condi tions across countries and across subgroups within countries. Table 9.2 reports prevalence rates for 13 chronic conditions across four countries, by gender and age group. It should be noted that, the question asking about chronic conditions was quite different in the Singapore survey than in the Philippines, Taiwan, and Thailand surveys. In the latter three coun tries respondents were asked to indicate whether or not they currently had any illnesses presented in a list. Singaporean respondents were asked whether they were currently receiving treatment from a doctor for any longstanding Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 O r-; Tt ; O 0l 0l •o tQ SL JB ^—m _O Ji 0l oc a. 0l 0l 0J Q. Q. ta ca rt Q, a. td JU _u ta _«J td -5 p ON vri ON CN Tf ^H Ov r** Tf "io.d— i i ; i i i — —; v6 — CN —i tN CN IO O CN NO iO —^ —* ro CN —i CN —' CNOOOoOONvO—iONOO-^fi-H —i —i ro CN CN —i CN — CN Tr — — —. ' r^fSdmodvoCNvd^3Oodod I — m —i CN Tt —i —i —i ts —i CN Tt — . — , — ' — I - - el w — -° Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 •o si JU JQ ga UH A-e/Cenditien iriocJOi i I i i i— '2 ts oo —i ^ . n"* ts CN e . s —i — oo ^s '•o d ts —* d ts vi i/S 0 06 r- i -H CN •* - H 0v O ts r^ ^ n *r, uc ,/c .xj I ^e. .". r ts 5 — —i —' vqTt-^OOtNTf—iCNl0t-;-^^^O CN -t —i —i —i —i Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 O Si sin-apmee ?Hca H1o.Q.In3Femaleaa<°t31aCO5JUtaEVi<130A-e/senditienro ro r- pro Tt oo oor-: -* ci cj ifi r'*! r*^ rot- ^^ ro e ro— * — i rooo— * — • rovO Tt t^ oo ve— i — • CN- Tf r- o vn— i . " ' CNO ON ro CN Or- r^ — • CN ro — • CNe -0 s ON CNco Tt o -* r-- r- ro i0 in— i CNNO s e Tt e— TT CMoo ro r- r- Tfoo ro O ro •^ CN^jt/1 Qo .^SlJuaSvo n -g c 0 Q.iiiooCNtsro. .2ONror-CN^.roin55OMNArthritis/rhe.m.iIi^O«sOO-.^***f^•es0NCN^Cft°rotsT.be_.lesisii1^^^^ovv»—r^e*«eso«^=oo0N^2L.n- disease1Ii00NO«*1^i0vOoi0e•*CN^er~**oo2Kidmy preblems':i iON2CNoo£o—oroI000r^2—-^—satarsts^**!pNOvotsvOe^r-t~^vOoovONO,^ON2m*-~Inrtntimmeerooo«*JSO2O2^.oooONt~sONvO. .CNO•O2Visien preblemsvqroroi0evOON*::•OvONOONi0E a.00*cCO Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 •o =- sin-apm jr ex o. CO _U _u A-e/Cenditien r- oo oo oo -^ — -:: i : i i i -• £* 2 Ov r- rs o ts —i m ts O CN NO CN —i m CN —• —i —< _ "e>on-m ^ ^ ^ rr. ~* 10 __ —ip — r^tsprsONTJ;—;Ooo _ ^ '~" e"* £! 5 —i ^ —' — ~ ts m " -H _ 15 ol ("- _ ?3 CN iO —* CN E| 3 u — •& Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 ro oo Tt OJ iJ sin-aper .5. ex cu JO CO JU u- _e — i o io od d OO tN m OO Ov CN in i* e CN O ^ CN ON *n WI m •••* CN OO ^ OO -H CNCN tN ts) _- _ _« — . . CN CO . —i —i tN ' ooootNOmvpvOvp-^ T|; — V; O t^ O in O oo ^ m rs oo C^ rs ts fi . — _, ,~i ooOOooooOiOOO v^^3—^iO — tsiri —* . CN —i m CN - 8 el (X td CN e< ,s Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 378 The Well-Being of the Elderly in Asia illnesses and, if so, to name them. (See Notes for Tables for exact question wording.) The Singapore question format is likely to result in considerable underreporting of chronic conditions, particularly the debilitating conditions that often do not require ongoing treatment from a doctor. For this reason, we included only life-threatening conditions for Singapore and, even for these, cross-country differences involving Singapore must be interpreted with cau tion. A cross-country comparison of the results in Table 9.2 indicates that Taiwan has the highest reported prevalence rates for more than half of the conditions, the exceptions being arthritis, tuberculosis, lung disease, inconti nence, and vision and hearing problems. In some cases, the prevalence rates for the overall population are substantially higher in Taiwan than in the other countries. For instance, while more than 17 percent report heart disease in Taiwan, the total is about 13 percent in Thailand, 9 percent in the Philippines, and only 6 percent in Singapore. Conversely, rates for arthritis, lung disease, incontinence, and vision problems are highest in the Philippines. (The propor tion of Filipinos reporting arthritis is twice that of the Taiwanese, and the Thai rate is close to that of the Philippines.) These latter conditions often have serious debilitating consequences and are easy to detect, whereas many of the other conditions require diagnosis. Hence, the pattern seems to be that Taiwan has the highest rates for diseases that are more difficult to detect, while the Philippines has the highest rates for some of the more apparent diseases. The higher prevalence rates for most conditions in Taiwan may be attributed to several factors. First, the older Taiwanese may have greater access to health diagnostics because they tend to enjoy a higher standard of living than their counterparts in other countries (which is often associated with more frequent doctor visits), and because their national health care sys tem offers aggressive screening programs, local health care facilities, and more recently, the Universal Health Insurance program. Also, a selection ef fect may increase their prevalence rates: earlier diagnosis and better medical care lead to increased survival, which translates into a larger old-age cohort surviving to report chronic diseases. In terms of gender, the results indicate that women in all countries tend to have higher prevalence than men for most conditions. It is particularly interesting that women report higher rates of heart disease in the Philippines, Thailand, and Taiwan, which is opposite the pattern normally found in the U.S. (Kramarow et al., 1999). This may be because women are diagnosed earlier, manage their disease better, and therefore have longer survival. Women are less likely than men to report strokes in these three countries, a pattern that is consistent with United States findings (Kramarow et al., 1999). Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Physical and Mental Health of the Elderly 379 When we look at patterns within five-year age groups, they remain generally the same: the Taiwanese have the highest prevalence rates for many diseases, with this apparent disadvantage increasing with age. For certain debilitating diseases like arthritis, however, Taiwan has the lowest rates across all age groups. In fact, the reporting of arthritis is much lower in Taiwan, particularly among persons age 70 or over, than would be expected among persons of comparable age in Western industrialized countries. For example, a 1995 health survey in the U.S. found the prevalence of arthritis among people age 70-plus years to be 63 percent for women and 50 percent for men (Kramarow et al., 1999). One interesting observation is that among Thai respondents aged 60 to 64, about 22 percent of women report heart disease, compared to just over 6 percent of men. Differences in heart disease rates between Thai men and women are smaller for other age groups, and one possible explanation for the large gender difference in the youngest age group may be a significant differ ential in health care utilization. As shown in Chapter 10 Thai women are more likely than men to use preventive and outpatient services, although these dif ferences alone do not seem large enough to account for their much higher reporting of heart disease. In Tables 9.3A and 9.3B we collapse the data into categories of life- threatening and debilitating conditions, respectively, and present the percent age of respondents who report having one or more of the examined condi tions in each group. With the exception of those shown by age group in the top panel, the percentages are age-standardized within each country (see Appendix 9.A for details on age-standardization) and shown by selected so- ciodemographic characteristics. Focusing first on life-threatening conditions (Table 9.3 A), we see that Taiwanese report the highest prevalence and Filipi nos the lowest, with Thais and Singaporeans falling in the middle. The high rate among Thais is due mainly to the high prevalence of heart disease among women. With regard to age patterns, the results are mixed. There is a ten dency for older persons to report lower prevalence of life-threatening condi tions in the Philippines, Thailand, and Singapore, but these age differences are significant only for Singapore. This pattern likely reflects greater access to health services and awareness of health problems among younger than older persons. In Taiwan, we see a more expected pattern, whereby the prevalence of life-threatening conditions increases significantly with age, at least through age 70-74. Turning to the age-standardized results, we find a significantly higher prevalence of life-threatening conditions for women in Thailand, but no significant gender differences in the other countries. In addition, we find significantly higher prevalence levels in urban than in rural areas for the three Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Q. •C .i t/t ac iu •s aj rt r-0 o c w nj c e— o s S 2 01 ^ Ct U i_ 2 00in53•fHced1?lcD.£sarsteristis|vO -^ ON vO Tt V)™H|ts O — i ON CN *— .ro ts CN m toropve o o in •* oON — •* ON o t*- o — i rN — i CN — • r-tsoo ts in ve t^ °^Tt in in to T t ^s3Unstandardized __4 rt5_97_74 _5+ sissq.are (psvaNO r^ ^Z'Z™~ mO1oON oo OOO — i ONts ro o1m -H — '2«^ro f>ts— — .2 5 toro3AgesstandardizedMale Female Chissq.are (psvalO ON:I 2 £«lve8O ON mro ts * ro CN ts^ OI 5-^ r^ v-" o rs ooo r- f5 — i ON ots — ^ ts — — ^ ts^§ inoo in O t s ^ •O O oo in ro ^''•.": •—• ooCN e3 3> >i. o.U O U,-llslI 2 I IIIT: 3 js e ,2 js 3 oi U Z S UIin in — *^ ror-i«svo ve O O— * OO -H o ro CN ro t^-tsg-es «s r- O— i ON CN Tt ts — • CN into^Bin o -^ o00 Tt CN TtONOo *-^5 31 b|ie1 E e | ,e a. « §• e c' c*Sz « u -=. i*- •5 •o -3 O J c a. o u — T3 •a a 8£ « 8. Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 I/I ai tQ a. oc jQ II ~ (Q - J2 g * U8.n00CGOBP3>|2J2"CA.5.iharsteristisU• iiii iii~P•* -^ in — i NO *--8 5? 8 Si IS £?ON.— CO — i •* — 1 NO ^^*— i ON O CN 00 2 NO IO ^D NO NOOr^* oo o "^ °^ O<M g®aCO1 I1 _ 11 6^6^=5 c o vO r*- r^ sJ3I I'g.Gi— * ts e-^ERS«Tf.^^TJ- oo *— 'ri ***i nIrI NO CNinooo O O06 ON ONIrJ NO NOro3gesstandardizedMale Female shissq.are (psvalI i(-sp_; r-: NONO iO ^^^/— V^q§33g^NCNCN ON 3NO NO g3Urban R.ral shissq.are (psval;OtCNNOoNNOOt06NO•o wo•^^ONr^'/IrjNOMarriedine8NOR22rijK•-NiO8— -8oo3sissquare (psvali1pv8ooCONO^.NO NO00c1151ei'CNr^oosON«1.EBi.Pi0tN^m"*Any sertndary+gpv — .8CNCNf-^N|ONON^-^OO8231iO.Si- 3tyW5 IU •o _u §1 — eO II £ 2 l.e- ^§« 2 .2 a o J^ i- £< Q-ON •g.a « T3 II oo a. O < .if Q « I = 00 o 5 u Q a Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 382 The Well-Being of the Elderly in Asia Table 9.4. Percent Reporting Difficulty Performing Selected Activities of Daily Living (ADL), by Age and Gender ADL/ Age groupWalking60-6465-6970-7475+All agesEating60-6465-6970-7475+All agesBathing or toileting60-6465-6970-7475+All agesDressing60-6465-6970-7475+All agesMale7.96.211.119.010.46.11.45.64.64.55.14.110.613.27.64.14.75.811.26.1PhilippinesFemale6.66.310.920.710.94.33.34.9554.44.62.94.115.56.84.74.14.49.75.7All7.26.311.020.110.75.12.55.25.24.54.83.46.914.67.14.44.35.010.35.9Male8.78.417.327.413.62.31.14.28.13.42.73.06.212.15.12.20.74.57.43.2ThailandFemale11.112.817.833.417.91.72.21.011.74.03.34.94.917.17.21.0121.010.23.2All9.910.917.631.015.92.01.72.410.33.73.04.15.515.16.21.61.02.69.13.2 countries with an urban/rural distinction. Education is also positively associated with reporting life-threatening conditions in the Philippines and Singapore. Differential patterns of access to and utilization of health care may be responsible for the seemingly increased risk of disease among urban residents and those who are more highly educated, although extended survival of individuals in these groups may also play a role. With regard to debilitating conditions (Table 9.3B), which are easier to recognize in the absence of clinical diagnoses, at least three-fifths of the elderly in the three countries for which data are available report suffering Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Physical and Mental Health of the Elderly 383 Table 9.4. continued See Notes for Table 9.4 at the end of this chapter. ADL/ Age groupWalking60-6465-6970-7475+All agesEating60-6465-6970-7475+All agesBathing or toileting60-6465-6970-7475+All agesDressing60-6465-6970-7475+All agesMale3.33.03.711.54.92.41.32.56.22.83.33.94.812.15.53.33.34.110.94.9TaiwanFemale2.74.16.212.45.91.21.32.67.12.84J5.35.916.77.72.73.75.512.15.6All3.03.44.811.95.31.81.32.56.62.83.94.55.314.46.53.33.54.711.55.2SingaporeMale0.52.35.28.33.50.30.31.92.91.20.81.35.05.72.7—————Female1.53.68.913.56.60.01.31.84.01.70.83.24.510.54.6—————All1.03.07.211.45.10.20.81.93.61.50.82.34.78.53.7————— from one or more of these conditions. The Filipino elderly have the highest prevalence at 65 percent. Within each of the three countries, the prevalence increases monotonically with age, and women experience debilitating condi tions at higher rates than men. Consistent with our expectation that debilitat ing conditions tend to be easier to detect than life-threatening conditions in the absence of medical diagnosis, the more educated show lower prevalence and no urban/rural differences are found for the Philippines or Thailand— patterns that differ sharply from those observed for life-threatening condi tions. (In Taiwan, however, the rural elderly report fewer debilitating condi tions than urban elderly.) Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 384 The Well-Being of the Elderly in Asia The second dimension of physical health that we examine is func tional limitation. To assess functional limitation, we use four indicators of ability to perform activities of daily living (ADLs)—walking around the house, eating, bathing and toileting, and dressing. Table 9.4 presents the percent reporting difficulty in performing each activity. Singapore, which has data for only three of the four measures, is included in the table where applicable. In general, the most commonly reported difficulty was walking around the house, with Thai elderly showing the highest percentage (about 16 per cent) and the Taiwanese and Singaporean elderly reporting much lower preva lence of this problem (about 5 percent). The Filipino elderly have the highest percentages reporting difficulty with eating, bathing, and dressing, while the elderly in Singapore show the lowest levels for difficulty with eating and bathing/dressing. With regard to gender differences, women report more difficulty walk ing around the house in Thailand and Singapore, but the pattern is mixed in Taiwan and the Philippines. For the remaining ADL indicators there are no consistent patterns associated with gender. In terms of age, however, ADL difficulty generally increases with age in all countries, with a substantial jump often observed for the oldest age group. In Table 9.5 we present the percentages reporting difficulty with one or more ADL by selected socioeconomic characteristics. The percentages are age-standardized within countries (with the exception of those shown by age group in the very top panel). Overall, Thais and Filipinos report higher levels of ADL difficulty than Taiwanese and Singaporeans—a differential that may be partly accounted for by environmental factors, particularly housing conditions, which tend to be more favorable in the relatively more developed countries of Taiwan and Singapore. When we examine patterns by the sociodemographic characteristics, we see in all countries strong increases in ADL difficulty by age, particularly above age 70. Women tend to report higher levels of difficulty than men, except in the Philippines, where there is no significant gender difference. Urban-rural residence and marital status are unrelated to ADL difficulty, and the results for education are mixed. No significant association between edu cation and ADL functioning is found in the Philippines or Thailand, while different patterns are found in Taiwan and Singapore. In Taiwan higher edu cation is associated with lower difficulty, a result also found in the United States (Kaplan, Strawbridge, Camacho, and Cohen, 1993; Guralnik, Land, Blazer et al., 1993). However, in Singapore the pattern appears to be curvilin ear, whereby those with no education and with secondary or higher levels both report slightly higher levels of functional impairment than those with primary education. Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Physical and Mental Health of the Elderly 385 Table 9.5. Percent Reporting Difficulty Performing One or More ADL, by Selected Sociodemographic Characteristics (age standardized) CharacteristicUnstandardized60-6465-6970-7475+Age-standardizedMaleFemaleChi-square (p-value)UrbanRuralChi-square (p-value)Not marriedMarriedChi-square (p-value)No formal schoolingAny primaryAny secondary+Chi-square (p-value)Philippines13.58.48.015.324.913.713.40.02 (.881)14.812.61.33 (.248)14.113.50.10 (.751)14.212.514.10.80 (.672)Thailand17.210.712.218.933.215.318.79. 17 (.002)17.017.20.02 (.888)18.116.90.99 (.321)17.317.216.40.1 3 (.938)Taiwan7.34.15.16.416.06.38.55.25 (.022)7.66.41.24 (.266)8.36.53.34 (.068)9.56.34.319. 15 (.000)Singapore*5.91.23.78.112.94.37.111.56 (.001)——6.15.30.94 (.332)6.24.27.16.27 (.044) 'Singapore results are based on three AIM v results for other countries are based on four ADI.s as defined in previous table. See Notes for Table 9.5 at the end of this chapter and Appendix 9.A for a description of age-standardization. Emotional and Cognitive Health Surveys in three of the countries—Thailand, Taiwan and the Philippines— collected data on emotional health, although with some differences in ap proach. They all asked respondents to self-assess their emotional state by reporting whether or not they experienced certain affective feelings within a specified time period. Both Taiwan and the Philippines used a subset of mea sures from the Centers for Epidemiological Studies-Depression Scale (CES- D) (Radloff, 1977). The exact items that were used in each country differed somewhat and are listed in Appendix 9.A. The Thai survey included only a single indicator for emotional health, asking how lonely the respondent felt during the year prior to the survey. Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Table 9.6. Ratio of Mean Scores on Indicators of Depression and Loneliness, by Selected Sociodemographic Characteristics CharacteristicTotalRatioMean/percentAge 60-64Age 65-69Age 70-74Age 75+MaleFemaleUrbanRuralNot marriedMarriedNo formalschoolingAny primaryAny secondary*Philippines(ratio of depressive symptoms)1.002.410.930.981.021.100.871.090.971.021.070.951.161.010.82Taiwan(ratio of depressive symptoms)1.002.010.920.881.121.180.861.170.991.011.240.891.250.880.76Thailand(ratio of % feeling lonely)1.005.720.930.991.051.090.931.060.801.041.090.951.110.980.69 Notes: For the Philippines and Taiwan, the ratio is defined as the mean number of depressive symptoms for the subgroup of interest divided by the mean number of depressive symptoms for the total sample. For Thailand, the ratio is defined as the percent reporting feeling sometimes or very lonely for the subgroup of interest, divided by the percent reporting this feeling for the total sample. See additional Notes for Table 9.6 at the end of this chapter. Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Physical and Mental Health of the Elderly 387 To standardize data presentation given the differences in measures, the results are shown as ratios of the mean for each subgroup to the overall mean for the country. Thus, ratios above one indicate that respondents in the group in question had a higher average level of depressive symptoms (i.e., worse emotional health) than the average of all respondents in the country, and ratios below one indicate that the group in question had lower than aver age depressive symptoms (i.e., better emotional health). Given the lack of common measures, we are unable to compare absolute levels of emotional health across the three countries. However, using the ratio approach, we are able to compare the direction of differences in emotional health on any given indicator across countries. The data in Table 9.6 show very similar patterns of emotional health across the three countries. In all countries emotional health tends to decline as age increases. Hence, not only are the elderly gelling physically weaker with age, but they are also emotionally less healthy. In addition, women, those not currently married (primarily widowed), and those with lower edu cation all tend to report poorer emotional health. Rural residents also appear to be at somewhat of a disadvantage with regard to experiencing loneliness in Thailand, but urban-rural differences are essentially non-existent for the other countries. Levels of depression and loneliness among older Asians have been little studied to date, but the differentials observed in Table 9.6 suggest the need for greater attention to these aspects of emotional health. Data on cognitive functioning were collected in the Philippines, Tai wan, and Singapore. The Taiwan and Singapore surveys contain different subsets of items from the Mini-Mental State Exam (Folstein, Folstein and McHugh, 1975), including basic orientation tests, as well as number and word recall tests. The Philippines survey contains only a word recall test. The exact tests used in each country are listed in Notes for Tables at the end of this chapter. The cognitive scores represent the number of correct re sponses on all of the items for a given country. Because, as with emotional health, cognitive functioning was measured somewhat differently across the countries, these results are also presented in ratio form. As shown in Table 9.7, in all three countries persons over 70 years of age and with less education had average scores below their respective coun try means—differences consistent with those found in the United States (Herzog and Wallace, 1997). These relationships are particularly strong in the Philippines and Taiwan. Adjusted for age, there are no appreciable differences by gender, rural-urban residence, or marital status. Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 388 The Well-Being of the Elderly in Asia Table 9.7. Ratio of Mean Cognition Scores, by Selected Sociodemographic Characteristics (age standardized) CharacteristicTotalRatioMeanAge60-6465-6970-7475+Age-standardizedMaleFemaleUrbanRuralNot marriedMarriedNo formalschooling Any primary Any secondary+Philippines1.003.131.161.030.960.780.991.011.050.960.981.000.821.01 1.16Taiwan1.004.131.101.070.950.761.070.911.000.990.921.040.841.03 1.26Singapore1.008.771.041.020.970.941.030.98_-0.991.020.981.04 1.07 Notes: For all countries, the ratio is defined as the mean number of correct responses for the subgroup of interest, divided by the mean number of correct responses for the total sample. Ratios above one indicate that the subgroup scored higher on average than the overall sample, and ratios below one indicate that the subgroup scored lower. See additional Notes for Table 9.7 at the end of this chapter and Appendix 9.A for description of age-standardization. Self-Assessed Health Measures of self-assessed health that ask respondents to rate their overall health on a scale from excellent to poor have become ubiquitous in surveys of older adults. The reason is clear: self-assessed health is a simple and eco nomical question about health that tends to capture a broad array of health dimensions. Self-assessed health has been found to be sensitive to variation in objective health status (Farraro, 1980). When individuals are asked to assess Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Physical and Mental Health of the Elderly 389 their health on a general scale, they seem to consider a multitude of health criteria, including chronic, functional, emotional, and cognitive disorders. In addition, at least on a subconscious level, they may be taking into account other more sensitive criteria about health that are difficult to detect in surveys or even in medical examinations. Indeed, a number of studies have found self-assessed health to be a powerful predictor of mortality, beyond the ef fects of more objective health indicators such as disease and disability (Idler and Benyamini, 1997). Because of the global nature of the measure and the fact that it can be captured with a single question, self-assessed health is a widely used measure in studies of health around the world and perhaps the most commonly used measure for cross-cultural comparisons. Nevertheless as a subjective mea sure it is vulnerable to influence by cultural norms, and sensitive to the actual phrasing of both the question and the response categories—which even if translated similarly in English, may not mean the same thing in native lan guages across countries. In the four countries examined here, the questions were asked quite similarly. However, the response categories were somewhat different in that Singapore used four categories, whereas the other three coun tries used five categories. For purposes of cross-country comparison, we collapsed categories into three levels, which we label as high, moderate and low health. The exact questions and the mapping of responses into these categories are given in the Notes for Tables. Table 9.8 presents the percent that report high, moderate, and low levels of self-assessed health in each country by age and gender. Older adults in Thailand and Taiwan tend to rate their health more favorably—that is, they are more likely to report a high assessment and less likely to report a low assessment—than do those in Singapore and the Philippines, where moderate self-assessments predominate. It should be noted that for Singapore (where the response categories differed) only the top one of four categories was regarded as high self-assessed health compared to the top two of five catego ries for each of the other countries. Therefore, it is possible that the distribu tion shown for Singapore elders is biased downward to some extent. In addition to these comparisons across countries, comparisons within countries are revealing. For example, despite changing reference groups, in creasing age is associated with lower self-assessed health in all countries. The age pattern is most notable in the Philippines, where the percent who rate their health as low increases from about 10 percent among those age 60 to 64 to about 27 percent among those 75 and older. With respect to gender, men tend to rate their health more favorably than do women. In all but two age groups (70-74 in Singapore and 75+ in the Philippines), high self-assess ments are more prevalent among men than women. Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 390 The Well-Being of the Elderly in Asia Table 9.8. Self-Assessed Health Status, by Age and Gender Age Health RatingAll agesHighModerateLow60-64HighModerateLow65-69HighModerateLow70-74HighModerateLow75+HighModerateLowPhilippinesMale18.567.913.624.666.78.718.672.29.217.071.111.99.561.528.9Female13.071.715.414.475.210.514.171.414.513.675.211.29.664.825.5Total15.270.114.619.071.39.715.971.812.315.073.511.59.663.626.8ThailandMale41.051.97.045.250.54.343.152.44.639.449.511.130.157.112.8Female30.357.112.629.357.813.035.253.910.928.060.611.327.557.415.1Total35.254.710.137.254.18.738.753.28.133.255.511.228.657.314.2 Table 9.9 presents the age-standardized percentage of respondents who report low self-assessed health. Here again we see that women generally report worse health than men, but the gender difference is significant only in Thailand and Singapore. In addition, rural residents tend to report poorer health than their urban counterparts in both the Philippines and Thailand. (The urban/rural differential is in the same direction for Taiwan, but is small and not statistically significant.) With regard to marital status, there is a slight tendency for married people to report better health, but the difference is significant only in Taiwan. Of all the sociodcmographic characteristics considered, education shows the strongest association with poor self-assessment of health, with Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Physical and Mental Health of the Elderly 391 Table 9.8. continued See Notes for Table 9.8 at the end of this chapter. Age Health RatingAll agesHighModerateLow60-64HighModerateLow65-69HighModerateLow70-74HighModerateLow75+HighModerateLowPhilippinesMale36.954.68.538.355.56.240.353.85.936.954.98.229.254.316.5Female25.265.49.431.064.15.024.567.97.520.568.710.821.161.717.3Total31.659.58.934.659.85.634.059.56.629.860.89.325.257.916.9ThailandMale17.568.414.122.071.16.920.565.913.69.267.523.313.167.819.1Female14.666.618.815.670.214.218.264.916.914.065.620.410.864.524.7Total15.967.416.618.870.610.619.365.415.311.866.521.711.765.922.4 significant differences observed in all four countries. In the Philippines, Thai land, and Taiwan, those with no formal schooling are two to three times more likely than those with the highest education to report poor health. For Singapore the education pattern appears to be curvilinear, with those in the lowest and highest education categories reporting less favorable health ratings than those in the middle category. Although this result is somewhat surprising in light of education patterns for the other countries, it is consistent with the education patterns observed for chronic conditions and ADL functioning described ear lier. To the extent that higher educated persons are in fact less healthy or at least perceive themselves to be less healthy than their less educated peers, the reports of poorer self-assessed health are reasonable. Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 392 The Well-Being of the Elderly in Asia Table 9.9. Percent Reporting Low Assessments of Health, by Selected Socio- demographic Characteristics (age standardized) CharacteristicTotalMale Female Chi-square (p-value)Urban Rural Chi-square (p-value)Not married Married Chi-square (p-value)No formal schooling Any primary Any secondary+ Chi-square (p-value)Philippines14.614.1 15.10.23 (.629)11.8 16.7 6. 12 (.013)15.9 13.6 1.30 (.254)18.414.4 8.0 11.60 (.003)Thailand10.17.2 12.6 33.93 (.000)7.4 10.6 7.20 (.007)9.2 10.7 2.30 (.129)12.1 9.6 4.018.23 (.000)Taiwan8.98.6 9.40.49 (.484)8.7 9.4 0.36 (.549)11.3 7.7 10.56 (.001)12.2 7.05.3 3 1.22 (.000)Singapore16.614.5 18.5 9.79 (.002)-17.1 15.9 0.83 (.363)16.714.5 21.7 8.64 (.013) See Notes for Table 9.9 at the end of this chapter and Appendix 9.A for description of age- standardization. Table 9.10 examines trends over time comparing our survey results from 1995 and 1996 with the results from the ASEAN surveys conducted in 1986 in the Philippines, Thailand, and Singapore and the baseline Taiwan elderly survey conducted in 1989. Although they provide a useful point of comparison, the results should be interpreted with caution. Different sam pling techniques and methods of data collection may compromise compara bility of the two sets of studies. In addition, response categories for self- assessed health differed across studies for all countries except Singapore. For example, we define poor health in Taiwan as those reporting fair or poor in 1989 and those reporting not so good or poor in 1996. With these cautions in mind, the results suggest that there have been improvements in self-as sessed health in the Philippines and Singapore, where the percent reporting very good/good health increased and the percent reporting poor health de creased. There is also evidence of improvement in Thailand where, although no change took place in the percent reporting very good/good, there was a decline in the percent reporting poor. Taiwan is the only country to show an Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Physical and Mental Health of the Elderly 393 Table 9.10. Trends over Time in Self-Assessed Health See Notes for Table 9.10 at the end of this chapter. Country and YearPhilippines198619961996 age standardizedThailand198619951995 age standardizedTaiwan198919961996 age standardizedSingapore198619951995 age standardizedSelf-Assessed Health RatingVery Good/Good Poor35%505135353539333367838411%1514413435233232511 unfavorable shift in self-assessed health, with a higher proportion reporting poor and a lower proportion reporting very good/good health in the later period. As noted above, self-assessed health has proven to be a very powerful global measure of health. Individuals tend to assess their health by appraising the aggregate of their objective health conditions, such as whether they have any chronic conditions or functional limitations, as well as their mental health. These health conditions are in turn associated with a number of sociodemo- graphic traits that reflect variations in access to medical care, knowledge about illness, resources, and levels of social support. As a result it is not surprising that a number of these sociodemographic characteristics also dis play a strong association with self-assessed health, as shown in Table 9.9. It is also plausible that cultural factors may influence the way in which individu als assess their health (Angel and Thoits. 1987). To examine these issues in Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 394 The Well-Being of the Elderly in Asia more detail, we turn to a multivariate analysis of the factors influencing self- assessed health. In this last stage of the analysis, we use an ordinal logistic regression technique to simultaneously estimate the effects of sociodemographic char acteristics, health behaviors, and physical and mental health status on self- assessed health, separately for each country. An ordinal logistic regression model is a proportional model that can be used for dependent variables with more than two categories that contain an inherent ordering. The model is similar to a bivariate logistic model in that it estimates the log-odds of obtain ing one response versus another when collapsing the dependent variable into any two categories (Agresti, 1996). (Indeed, a bivariate logistic regression is a special case of an ordinal logistic regression for a two-category dependent variable.) The dependent variables used in the present analysis correspond to the measures of self-assessed health shown in Table 9.8, coded such that l=high, 2=moderate, and 3=low. Table 9.11 presents results from the regression models in the form of odds-ratios. Given the coding on the dependent variable (i.e., that an increase in value implies a less favorable rating), an odds-ratio above one indicates that the group in question rated their health less favorably on average than their respective counterparts, all other things being equal, and an odds-ratio below one indicates that the group rated their health more favor ably. We expect the presence of physical, emotional, and cognitive health problems to be strongly associated with poorer health assessments. Further more, we expect the other factors—sociodemographic characteristics and health behaviors—to be associated primarily through actual health status. For example, as shown in the bivariate analyses for self-assessed health, women tend to rate their health less favorably than men. However, women also had poorer outcomes than men on the physical, emotional, and cognitive out comes in general, and this may account in large part for their poorer self- assessments. Likewise, health risk behaviors are generally associated with negative health outcomes, which in turn are associated with poorer health ratings. Hence, when health status is controlled, as in the following multivari ate analyses, the effects of sociodemographic and health behavior factors should be reduced. Turning to the results, we find somewhat mixed effects of sociode mographic characteristics on self-assessed health across countries. The ef fect of age on self-assessed health is generally not significant, controlling for the other factors, except in Singapore where persons aged 65-69 report more favorable ratings than those age 60-64. The strong age effects that were observed for the other countries in the bivariate analyses in Table 9.8 are Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 Physical and Mental Health of the Elderly 395 likely explained in large part by the health factors included in the model. The finding for Singapore seems somewhat surprising, but it is plausible in light of the argument by Cockerham and colleagues (1983) discussed earlier in the chapter. That is, those who are in reasonable health at older ages may actually report their health more favorably than those in reasonable health at younger ages, because the older respondents' reference group is less healthy. Hence, once underlying health differences are controlled in the multivariate models, we would expect the negative age effect to be substantially reduced if not reversed. The gender effect is also not significant except in Thailand where, other things being equal, females report less favorable health than males. Marital status is not a significant factor in any country. Persons in rural areas report less favorable health ratings than those in urban areas in Thailand and Taiwan. However the rural effect that was observed for the Philippines in the bivariate analysis (Table 9.9) disappears in the multivariate analysis, suggest ing that much of the urban-rural difference among older Filipinos is accounted for by differences in actual health status and the other characteristics shown. With regard to education, both the Philippines and Taiwan show sig nificant effects. In the Philippines, those with no formal schooling have lower self-assessed health than those who had primary schooling, but those with more than primary education do not necessarily fare better. In Taiwan the education effect is in the expected direction for those with the highest educa tion, but there is no significant difference between those with no formal versus primary education. The effect of education is in the expected direction for Thailand, although the differences are small and not statistically signifi cant. For Singapore, the coefficients again suggest a curvilinear effect as observed in the bivariate analysis, although as with Thailand the effects are small and non-significant. We include both living arrangements and existence of children as indi cators of social support. In Asia, where living with a child is the norm, those who have no children or who live in alternative living arrangements may report less favorable health because there may be no ready sources of imme diate help, or because they feel more dejected and lonely. On the other hand, those who have children but are not living with them may be somewhat healthier individuals who are able to live independently and do not require care from children. The results are supportive of both of these hypotheses. In Singapore persons who do not live with a child report worse health, and this is particularly true for those who live with a spouse only or in other arrange ments. The same direction of effects is found in the Philippines, although the differences are not significant. The reverse is observed in Thailand where those who do not live with children report more favorable health, especially Generated for anonymous (University of Michigan) on 2014-05-27 12:31 GMT / http://hdl.handle.net/2027/mdp.39015060636282 Creative Commons Attribution-NonCommercial / http://www.hathitrust.org/access_use#cc-by-nc-3.0 i/) o i/1 S.JJ §-bb t/I QC 3 Oi |0 £ E -o o JJX — -a >. tu •o c O ° p8.00B1.t9H•3'3£waugo.—£Independent Variable*ON f^ m Q o r*^ ^^ O v O ONo _; o — d222 ^ S^-;_;_; -; 6#* *— * vp O r- iO ON O ts m —io — — -; —— i — i oo m o . ON oo CN ^ ^ Oo d _; — i -H