The Economic Status of Rural America in the President Trump Era and beyond

The Economic Status of Rural America in the President Trump Era and beyond Abstract To set the stage for future research aimed at developing public policies that support economic prosperity in rural areas, we review the current economic conditions of rural America and the current literature. Rural America is often characterized as a uniform, distressed place where agriculture dominates. In fact, rural America is diverse, with many regions doing well economically. In some areas, labor-saving technologies have reduced the workforce in manufacturing and resource-dependent industries. However, integration with urban areas has weakened the economic divide between urban and some rural areas, while natural amenities have boosted the fortunes of others. There is also evidence that homegrown enterprises can support growth even in the most remote, distressed regions. To support economic growth, policies should recognize the unique features of each place or region and balance the farm sector with the larger nonfarm rural economy. Economists are well-positioned to provide research-based evidence of what works, as well as rigorous evaluation of new polices. Rural America is an important source of America’s food, water, energy, and other natural resources, and it contains areas of great natural beauty. However, urban Americans rarely give much thought to their rural compatriots. One reason may be that rural America appears to be in decline. Another is that few urbanites have a personal attachment to rural America. While rural-to-urban migration heavily dominated from the 1920s to the 1970s, it has since tapered off, meaning fewer family connections between the two. A wave of recent events, culminating with the 2016 Presidential election, has put rural America under a new spotlight (Goetz et al. 2017). Nonmetropolitan America voted for President Trump in higher percentages than metropolitan America, propelling him into the White House.1 A perceived growing divergence between urban and rural America has renewed interest in understanding rural America. In this paper we provide an overview of rural America and the economic issues relevant to the agricultural economics and related professions, within the context of the new administration and beyond. As the last election illustrated, rural America and the conditions surrounding its residents are important to the national economy, and rural voters matter. Yet even defining what constitutes a rural area (“one knows it when one sees it”) remains a challenge, and rural definitions have changed over time. We start by examining recent population trends and how they have affected rural/urban classifications. Then we review the changing industrial make-up of rural areas and attendant social and economic problems. We also present issues surrounding data availability for analyzing and modeling conditions in rural regions that form the basis for modern regional development research. The paper concludes with a discussion of rural programs and recommendations for future research directions and policy analysis for economists. Rural Population Change and Recent Voting Trends Some historical background about rural areas is helpful in prioritizing future research directions. Labor-saving technological change in agriculture and other primary sectors led to an expansion of output that has benefited all global citizens. Yet it also dramatically reduced the demand for labor in rural America, leading to migration to urban areas. At the same time, improvements in automobiles and construction of modern highways that unlocked rural-to-urban commuting opportunities mitigated these effects and allowed some of the displaced labor to find work while remaining in and supporting their rural communities. This in turn led to the emergence of large city-centered functional economic regions in which the largest urban center typically represents the engine of growth for rural and urban residents alike. During this same period, rural communities located in areas with attractive landscapes, mountains, lakes, oceans, and a pleasant climate benefited from amenity-driven in-migration (Partridge 2010). The net effect was that, beginning in the 1970s, rural areas with access to urban labor markets and high amenities essentially reversed the long-run net out-migration from rural areas. Rural America is now diverse and comprised of three distinct types of areas: (a) high-amenity regions; (b) metro-adjacent rural communities; and (c) remote or extractive-based rural communities; it is the latter that have generally struggled. To illustrate this, figures 1 and 2 show population growth over the 1969–2015 and 2000–2015 periods in nonmetropolitan and metropolitan areas. Shown in the legend in parentheses is the number of counties in each definition, with a declining number of nonmetro counties evident over time. Population growth is measured in several ways: (a) using the 1950 metropolitan area definitions that reflect older core cities with virtually no recent history of a “rural” heritage; (b) using 1973 definitions that reflect the beginning of urban expansion into suburban and exurban areas; and (c) using subsequent redefinitions of metropolitan areas in 1983, 1993, 2003, and 2013. The 2003 and 2013 definitions include more counties that had previously been considered rural, due to a classification change in which the U.S. Census Bureau lowered the cross-commuting threshold required for a county to be part of a metro region to 25%. In many cases, the recent additions are far on the periphery of the urban center, appearing to many observers as more rural than urban—even though there are moderate urban labor linkages. Figure 1 View largeDownload slide Population growth of non-metro area by historical MSA definition: 100 = 1969 Population Data Sources: U.S. Bureau of Economic Analysis for population (2017), and U.S. Census Bureau for metro definitions (2017). Figure 1 View largeDownload slide Population growth of non-metro area by historical MSA definition: 100 = 1969 Population Data Sources: U.S. Bureau of Economic Analysis for population (2017), and U.S. Census Bureau for metro definitions (2017). Figure 2 View largeDownload slide Population growth of metro area by historical MSA definition: 100 = 1969 Population Data Sources: U.S. Bureau of Economic Analysis for population (2017) and U.S. Census Bureau for metro definitions (2017). Figure 2 View largeDownload slide Population growth of metro area by historical MSA definition: 100 = 1969 Population Data Sources: U.S. Bureau of Economic Analysis for population (2017) and U.S. Census Bureau for metro definitions (2017). The metropolitan definitions are split across time to make several points. First, the validity of the notion that rural America is in decline depends on the definition of rural. Using the (virtually) current 2013 definition, the metro population grew by 67% between 1969–2015 (see figure 2), compared to 26% for nonmetro areas (see figure 1). However, much of this apparent lagging performance for rural America relates to reclassification over time of the fastest-growing nonmetropolitan counties as metropolitan (Partridge et al. 2008). The late Andy Isserman argued that this metropolitan area reclassification is like taking the best team out of a sports league every year; eventually the league would become low-quality (Isserman 2005). Using the most urbanized 1950 definitions, figures 1 and 2 show that 1969–2015 and 2000–2015 population growth in nonmetropolitan (rural) counties was double that of metropolitan counties in the earlier period, and also greatly exceeded it in the twenty-first century. In other words, counties considered rural in 1955 have done quite well. As previously nonmetropolitan counties have been redefined as metropolitan, nonmetro growth has been subsequently slower over time and metro county growth has increased; in other words, the reclassified counties were actually growing more rapidly than the original metro counties. In that sense the story of rural malaise is just as much a story of some nonmetro counties being remarkably successful. For instance, as shown in figure 1, using the 1973 nonmetropolitan definition, rural population growth equaled the national (and metro) average up until 2007. Moreover, if the 500 counties that were reclassified as metropolitan between 1973 and 2013 had remained in the nonmetro category, the rural population growth rate would have tripled.2 Figures 1 and 2 also illustrate an inflection point in 2007 when relative rural population growth slowed across all definitions, while relative metropolitan growth either slightly increased or showed no discernable change. This suggests that both current nonmetropolitan areas and reclassified (former) nonmetro counties are experiencing slower growth. Some of the slowdown relates to the housing bust’s effects on reducing exurban residential development (and these more recent metro counties are more remote). Another is that lower-wage rural manufacturing has experienced productivity growth and fierce foreign competition that suppressed job growth (though since 2001, urban and rural manufacturing have performed about equally in terms of negative job growth). It is not surprising that metro job growth exceeds rural job growth, mainly because many rural residents commute to urban jobs. Yet, using 2013 definitions, between 2001–2007 and 2007–2015, total nonmetropolitan job growth slowed from a positive 4.6% to a decrease of 0.3%, whereas metro job growth slowed from 9.7% to 6.6%, further increasing the rural-urban employment gap. Another factor that has hurt rural areas is the decline in industries that previously employed many workers. Even though the circa 2003–2013 super-commodity cycle should have helped rural areas (as was the case with the 1970s super cycle), agriculture continues to shed employment and rural mining job growth has lagged.3 This shows that even traditional rural engines are no longer sufficient to lift local growth (on average). President Trump, to his political credit, exploited this (relatively recent) economic decline in rural America in the 2016 election.4Table 1 divides the 2016 vote into metro/nonmetro areas using the 1950 and 2013 definitions based on when the previously rural county was reclassified as metropolitan. The 1950 metro group only gave President Trump 37% of the vote, but those reclassified between 1950 and 1973 gave him 53%. President Trump’s vote share steadily increased to 63% for the group of counties redefined between 2003 and 2013. The 2003–2013 reclassified counties’ 63% Trump vote share almost equaled his 2013 nonmetropolitan county share of 64%. There are other reasons (including economics) for this pattern, but the amount of time an area has been classified as part of a metro area is a good predictor of how President Trump performed; in other words, how recently they were considered “rural.” Future research could examine the reasons for the divide between long-time urban and recently-integrated urban areas; this includes examining whether as a county becomes urban, what is the dynamic process for losing its rural heritage and assimilating into urban culture. This also has implications for research into the geographical extent of regional governance. Table 1 2016 Election Results for Trump   Metro Definition and Change  Total Metro Votes  Trump Metro Votes  Trump Metro Votes Share  (1)  Metro 1950  68,380,703  25,469,903  37%  (2)  Metro Def. Change 1950–1973  26,522,379  14,147,642  53%  (3)  Metro Def. Change 1973–1983  6,245,681  3,457,467  55%  (4)  Metro Def. Change 1983–1993  3,764,839  2,214,950  59%  (5)  Metro Def. Change 1993–2003  3,548,019  2,075,511  60%  (6)  Metro Def. Change 2003–2013  2,040,345  1,298,349  63%  (7)  Metro 2013  110,501,966  48,663,822  44%  (8)  Nonmetro 2013  19,409,615  12,401,021  64%  (9)  Total United States  129,911,581  61,064,843  47%    Metro Definition and Change  Total Metro Votes  Trump Metro Votes  Trump Metro Votes Share  (1)  Metro 1950  68,380,703  25,469,903  37%  (2)  Metro Def. Change 1950–1973  26,522,379  14,147,642  53%  (3)  Metro Def. Change 1973–1983  6,245,681  3,457,467  55%  (4)  Metro Def. Change 1983–1993  3,764,839  2,214,950  59%  (5)  Metro Def. Change 1993–2003  3,548,019  2,075,511  60%  (6)  Metro Def. Change 2003–2013  2,040,345  1,298,349  63%  (7)  Metro 2013  110,501,966  48,663,822  44%  (8)  Nonmetro 2013  19,409,615  12,401,021  64%  (9)  Total United States  129,911,581  61,064,843  47%  Source: Townhall.com. Missing county-level data for Alaska. Current metropolitan counties are split into those defined in 1950 and those that were subsequently reclassified as metropolitan by time period of reclassification. Rural County Classifications and the Rural-Urban Continuum Policymakers and researchers who are examining “rural” areas need clear definitions of what is rural. While there is no binary rural/urban dichotomy, some sort of strict division is necessary for informing policy and related research. Otherwise, given the numerous definitions and the poor understanding of the tradeoffs involved in using each definition, poor academic research and policy may be the result. Given the changing definition of rural, the National Academies of Sciences recently hosted a workshop defining what is rural (National Academy of Sciences, Engineering and Medicine 2016). In practice, an urbanized continuum stretches out from the urban core, through the suburbs to exurban areas with high urban commuting rates, to rural areas (and very sparse and remote rural could be added as a distinct category). However, there are two key ways of thinking about rural in trying to make a reasonable division. First is the popular notion that is often portrayed by the media of rural being sparsely populated, with a relatively large number of farms or a high concentration in other primary sectors–that is, the “real rural.” In this case, rural definitions based on population density would be useful in understanding rural/urban differences in culture and other beliefs. However, a population density-based notion of rural is simplistic and can produce misleading results when conducting economic analysis. Namely, local economic questions relate to behaviors in functional economic regions. If people live and work as well as conduct their other economic business in the same region, then that is their functional economic region regardless of the density of the resident’s particular “neighborhood.” For example, a researcher examining the effects of job growth on poverty should not separate functional economic regions into heterogeneous groups that are unrelated to the underlying economic processes. Fortunately for those who are conducting spatial economic research, the U.S. Census Bureau’s current metropolitan (micropolitan)/nonmetropolitan (core rural) designations are functional economic areas, which should be useful in most cases of economic research. The census uses a 25% commuting rate threshold and a minimum principal-city size of 50,000 in defining metropolitan area thresholds. Economists interested in rural economic issues should not apologize for using this delineation because it is exactly what theory suggests. Yet this does not mean that the current metropolitan definition is without problems. In 1950, it may have been reasonable to assume that an urban core of 50,000 people represented sufficient agglomeration economies in the United States to be “urban.” Yet agglomeration thresholds appear to be increasing over time as economic functions that once were found in smaller locales are now only found in larger places (Partridge et al. 2008; 2010), and the current definitions may not accurately reflect today’s concept of urban versus rural. Additionally, in the current definition (and beginning with the 2003 definition), the commuting threshold or the percentage of a county that cross-commutes with the principal city was lowered to the current 25% level; the result is that the 50 to 100 most populous metro areas are typically geographically spread out up to a radius of 100 miles or more. This also means that more distant counties from the urban core were reclassified as metropolitan. These are counties with a lower population density and a little higher reliance on agriculture and resource-based industries; that is, the type of reclassified counties that voted in large numbers for President Trump. The relatively low commuting threshold extends metro areas to peripheral rural places that are less economically integrated with the urban center. We suggest that the commuting threshold be raised to 30% or 35% to mitigate this problem of classifying true “rural” counties as metropolitan. In 1950 it may have been reasonable to assume that an urban core of 50,000 people represented sufficient agglomeration economies to be “urban.” Yet, agglomeration thresholds appear to be increasing over time as economic functions that were found in smaller locales are now only found in larger places (Partridge et al. 2008; 2010) and the current definitions may not accurately reflect today’s concept of urban versus rural. Additionally, The U.S. definition is also significantly different from that used in Canada and Europe, especially with regard to the commuting threshold. Compared to the 50,000 urban threshold used in the United States, Statistics Canada uses a 100,000 total population threshold and a 50% commuting threshold to define metropolitan areas.5 The French “urban area” definition (the closest to the U.S. metro definition) is based on a threshold of an urban core of at least 10,000 jobs and a 40% commuting threshold for neighboring municipalities.6 Economists should provide research on the impact of alternative definitions that are based on more updated concepts of urbanization and economic integration and are more in line with those in other countries. Alternative definitions of urban versus rural have been developed by the USDA. Commuting zones (CZ) are based on clustering of commuting patterns; however, the most current definitions are from 1993. While they are increasingly used in many labor economics papers (e.g., Autor, Dorn, and Hanson 2013; Chetti et al. 2014), the CZ definitions are both outdated, and unlike the Census definitions, are not derived based on any economic theory, but cluster analysis (Partridge et al. 2017). For example, there are no commuting thresholds and thus weakly integrated rural areas are connected together as a CZ and even low-integrated rural counties with weak commuting thresholds below 25% are included with metropolitan areas. Likewise, highly integrated metropolitan areas are split into separate CZs (even the same city is split) without any coherent rationale. They also do not account for the fact that counties often send commuters to multiple destination counties, especially in the densely settled Northeastern United States. Recent research seeks to circumvent some of the problems with CZs and MSAs by using a link (or edge) community method from network theory to construct overlapping commuting networks, where counties can belong to multiple Labor Market Areas (LMAs) at the same time (Han and Goetz 2017). The study shows that counties with high employment in primary industries are more likely to belong to a single LMA, while counties with more households of two or more workers more likely belong to multiple such LMAs. The second major USDA categorization includes the urban influence codes (UICs) and Beale codes (BC). For example, the 2013 UICs split metropolitan areas into two groups —those with more and less than one million residents. The nonmetro counties are divided into ten groups, based on a three-way division: (a) micropolitan areas; (b) noncore rural areas with a city of greater than 2,500; and (c) noncore rural areas with a city less than 2,500.7 These codes were first introduced over 30 years ago, and historically showed that growth was much higher in counties adjacent to metro areas and that, conversely, remote locations without a city of at least 2,500 people were generally stagnant or in decline (Wu and Gopinath 2008). However, over time, especially after 2000, the BCs and the UIC have become less valuable and have explained less of the variation in socioeconomic outcomes. This is partly due to the changes in the U.S. Census definitions for metropolitan areas, on which these classification schemes rely. The low 25% commuting threshold means that counties that would have previously been considered adjacent nonmetropolitan are now designated metropolitan. Now rural counties “adjacent” to metropolitan areas can actually be quite remote, meaning that the expected positive spillovers are considerably less than before. Likewise, when these codes were introduced, the economic importance of a town of 2,500 was different from today for the same reasons that a 50,000 urban-core population no longer may be insufficiently large to capture sufficient agglomeration economies to be classified metropolitan. While the value of the BCs and UICs is waning, they were helpful in gaining a better understanding of how many economic outcomes change with remoteness from urban centers (Partridge et al. 2008; 2010). This has led to a more recent body of literature that more finely describes distance effects than simple adjacency (e.g., Tsvetkova, Partridge, and Betz 2017; Weber et al. 2017) and illustrates that the rural remoteness costs are rising over time as technological change favors larger cities (e.g., Partridge et al. 2010).8 The Industrial Composition of Rural Areas Given the last election results, it is important to understand how the rural economy has changed. For the most remote, rural regions, economic integration with urban areas is limited. Historically, the economy in these regions is more likely to have been dominated by agriculture or other resource-based industries. These regions have been hurt by changes in these industries, especially increased mechanization. Focusing on agriculture, figure 3 shows the 1969–2015 farm share of employment using Bureau of Economic Analysis (BEA) data for nonmetropolitan areas, metropolitan areas, and the total United States, as well as Current Population Survey data for the U.S. share of total employment in which agriculture is the primary job. In rural America, the farm employment share declined from 15% in 1969 to 6% in 2015, illustrating a profound restructuring. Likewise, the farm share also declined in urban areas, but agriculture is a very small portion of urban economies. Figure 3 View largeDownload slide Percentage of total jobs in farming: 1969–2015 Data Sources: Farming share of total employment for Metropolitan, Non-Metropolitan, and United States: Bureau of Economic Analysis; Total Full-Time and Part-Time Employment CA25 (1969–2000) CA25N (2001–2015). Farming as primary job as a share of total employment: Bureau of Labor Statistics, Table cpsaat01 (2017). Figure 3 View largeDownload slide Percentage of total jobs in farming: 1969–2015 Data Sources: Farming share of total employment for Metropolitan, Non-Metropolitan, and United States: Bureau of Economic Analysis; Total Full-Time and Part-Time Employment CA25 (1969–2000) CA25N (2001–2015). Farming as primary job as a share of total employment: Bureau of Labor Statistics, Table cpsaat01 (2017). The over reliance on export-based industries (such as mining, manufacturing, and large-scale agriculture) has led to lower growth, persistent unemployment, and a highly-segmented labor market that is not adaptable to change (Weiler 2001; Kilkenny and Partridge 2009). Increasing global trade with low-wage countries has also reduced and often eliminated the traditional wage and land cost advantages for rural manufacturers. Even when industries remain, many do so with much smaller workforces. While many rural economic development leaders and politicians believe the solution is to recruit another large employer, rural areas are increasingly at a disadvantage in attracting large firms because of factors such as limited workforce, low education levels, insufficient infrastructure, or high transport costs (Partridge and Olfert 2011). In addition, rural areas with large natural resource endowments (coal, oil, or natural gas) have long experienced boom and bust cycles. New technologies such as hydraulic fracturing made possible the extraction of natural gas and oil that were previously difficult to access. The resulting post-2007 boom affected traditional oil and gas rural areas in Texas as well as new areas such as North Dakota, and areas covered by the Marcellus and Utica shale formation in the Eastern United States, especially Ohio, Pennsylvania, and West Virginia. However, as natural gas (and oil) prices have declined so have exploration and employment. At the same time, the coal industry experienced large-scale layoffs, in small part due to stricter regulations, but mainly due to rapid productivity growth and more recently due to competition from cheaper natural gas, and this has especially affected eastern coal-mining communities. Combined, these resource economic busts helped drive the support for President Trump in many resource-dependent communities. Natural gas exploration and jobs may rebound, as natural gas replaces coal as a cheaper and more environmentally-friendly source for electricity generation, though promises of vast numbers of energy jobs are generally unrealistic (Kelsey, Partridge, and White 2016). Coal mining employment is unlikely to rebound, even if plans for new environmental regulations are abandoned (Betz et al. 2015). Dating back to the end of World War II, coal mining employment has declined as the industry implemented labor-saving technologies and this is exacerbated by declining foreign markets for U.S. coal (Betz et al. 2015). From 1989 to 2012, data from the Mine Safety and Health Administration shows that as coal output held relatively steady in Kentucky and West Virginia, coal-related employment fell from 79,000 to 41,000. Even if mines do reopen, they are likely to be more automated and the industry has already announced it will pay lower wages.9 Indeed, research suggests that scaling back environmental protections is the opposite of what coal mining regions need for long-term growth (Fan, Goetz, and Liang 2016).10 With the coal industry in decline, it will be nearly impossible to attract the human capital and the other firms needed to replace coal if quality-of-life is degraded in these remote areas (Betz et al. 2015). For some rural areas with natural amenities, there is the potential for future amenity-led growth. This may include an influx of new population or tourists and vacation homeowners seeking better climates and natural surroundings (Wehrwein and Johnson 1943; Deller et al. 2001; Carruthers and Vias 2005; McGranahan 2008; Wu and Gopinath 2008; Partridge 2010; Fan, Goetz, and Liang 2016). Yet, since not all rural areas are endowed with high levels of natural amenities or a favorable climate, this is not a solution for all communities. Even if natural amenities exist, if communities are impacted by energy development and have suffered from a degraded environment, efforts by rural communities to benefit from such growth will be hampered. Despite the challenges facing the most distressed rural communities, there is growing evidence that having more local entrepreneurs and self-employed can help rural and remote communities sustain growth and prosper (Stephens and Partridge 2011; Rupasingha and Goetz 2013; Stephens, Partridge, and Faggian 2013), as well as mitigate trade shocks (Liang and Goetz 2016). Stephens, Partridge, and Faggian (2013) find that even in the most distressed rural areas where there may be a large number of necessity entrepreneurs (Low, Henderson, and Weiler 2005; Acs et al. 2009) or those who start businesses because of a lack of other opportunities, there still are spillovers into the greater region including more wage and salary jobs (e.g., people working for someone other than themselves). Local entrepreneurs further magnify the positive economic impacts of their business activity on the community by favoring procurement from other local businesses (Fleming and Goetz 2011). Local entrepreneurs can be successfully “home-grown,” as rural entrepreneurship often requires relatively low educational levels, while the businesses purchase inputs locally, fulfill local demand and benefit from local assets, are perceptive to the regional milieu, and further build the existing entrepreneurial culture (Akgün et al. 2010; Stephens, Partridge, and Faggian 2013; Fritsch and Wyrwich 2014; Li et al. 2016). Indeed, past research shows that a key reason for a natural resource curse in coal country is that it lays the seeds of decline during a bust because coal mining employment crowds out self-employment and small business development that would cushion the blow (Betz et al. 2015). Yet, there are large gaps in understanding how to foster entrepreneurship and self-employment for rural growth. Poverty and Social Issues in Rural America Increasing economic opportunities is an important concern in rural areas, especially where globalization and automation have caused profound structural change. Rising productivity has displaced blue-collar workers and, in general, while there have been economic winners from trade and automation, many of the economic losers have been in rural areas. Although poverty is an important urban problem in terms of numbers of poor, official nonmetropolitan poverty rates—and especially for children—have been consistently higher than metropolitan poverty rates dating back to 1959, when poverty rates were first measured.11 In particular, rural child poverty rates are especially high relative to urban child poverty rates and over 85% of counties classified as “poor” are rural (USDA 2017).12 Deteriorating social and economic conditions are now reducing life expectancy in many rural communities, partly as a result of rising rates of drug use and suicide; recent studies show that where individuals live influences their likely cause of death (Moy et al. 2017). Despite being the major source of the nation’s food supply, rural areas also face issues related to food insecurity as explained in Gundersen and Ziliak (2018). Clearly, geography and space play roles in deteriorating socioeconomic outcomes, but the precise causal mechanisms and effects on human well-being remain poorly understood (Goetz et al. 2017). While both local poverty and inequality have long been studied (e.g., Levernier, Partridge, and Rickman 1998; Rupasingha and Goetz 2007), they are receiving heightened attention with the work of Chetty et al. (2014) and Picketty (2013). Using big data in the form of individual tax records, Chetty et al. (2014) show that geography affects economic mobility. Interestingly, rural areas perform better at promoting youth up the income ladder than urban areas, other factors held constant. Yet it is not clear whether this occurs because rural youth leave in pursuit of better urban opportunities, or whether rural areas inherently have advantages, perhaps because of better and smaller schools that, among other benefits, provide more leadership opportunities. Select studies are starting to examine this question. Recent research that reevaluates the Moving to Opportunities project, in which families were relocated from low-income housing to wealthier neighborhoods (Chetty, Hendren, and Katz 2016), should also be reexamined for its potential research, practical, and policy implications for rural areas. Of particular importance for research are the questions of (a) how poverty and inequality evolve over time and space, and the rural-urban continuum in particular, and what forces are driving the changes; (b) what role government can play (if any) in mitigating rising poverty and inequality through programs and policy; (c) the effects of globalization and automation, including robots and 3-D printing technology on the changing relationship between poverty and inequality; and (d) the interrelationship between inequality and poverty, or economic growth. Poverty and inequality occur to different degrees among U.S. counties. Since 1989, the number of counties with both high poverty rates and high inequality levels has increased. This has been especially true in counties in small and midsize metropolitan areas, where nearly one-half of all counties now have high rates of both poverty and inequality, up from on 22% in 1989 (Jarosz and Mather 2014). Additionally, 44% (up from 35%) of the most rural counties have high rates of both poverty and inequality, compared with only 21% (up from 11%) of the most urban counties (Jarosz and Mather 2014). Clearly, the forces causing the expanding coexistence of poor people and uneven income distributions are weakest in urban areas, strongest in the next larger category of counties, and somewhat milder in the most rural counties, but still prominent. Understanding why these changes are happening in the different sized communities and across the rural-urban continuum is an important first step towards addressing them. Although some level of inequality is beneficial for growth because it provides incentives for people to work harder, innovate, acquire training and human capital, invest, and promote entrepreneurship, evidence is mounting that the United States has now reached levels that may be suppressing aggregate economic growth (Partridge and Weinstein 2013).13 At the subnational level, early research tended to find a positive linkage with growth, especially in the long-run (Partridge 1997; 2005). However, more recent evidence suggests a negative linkage, notably in the short-run (Atems 2013). This relationship has not been investigated rigorously at the subnational or regional levels, or for rural versus urban areas. There is also evidence, at the level of nations, that higher rates of inequality are associated with shorter periods of economic growth (Berg and Ostrey 2011). This could have important implications for the economic future of the entire United States and its rural communities. Empirical Advances and Data Understanding rural growth and assessing the impact of rural policy requires establishing causal effects (see Athey and Imbens 2017). Recent decades have seen a revolution in empirical identification in order to establish causation, and these methods are increasingly being applied to address regional and rural economic issues.14 For example, empirical advances have been made using natural experiments, propensity-score matching and synthetic controls, instrumental variables, regression discontinuity, quantile regressions, and spatial modeling, including locally-or geographically-weighted regressions. We focus here on advances related to estimating multipliers as they are especially useful in evaluating economic development policy and in informing decision-making to support rural economic growth. Multipliers Estimating the economic impacts of a new business in terms of total net job or income creation is essential in assessing whether community support should be provided to recruitment efforts. For a distressed rural area, providing tax incentives to recruit a new employer may be tempting. It is especially important that with limited resources, the benefits must exceed the costs. The impacts of new development include the new jobs from that firm, the new jobs supported by the activity and salaries of that firm and other induced effects that create employment as well as offsetting displacement or crowding-out effects. The same types of analysis can also help a community evaluate the effect of a major plant closure or shock to an industry. In both cases, knowing the net multiplier effects would be valuable. A local manufacturing multiplier of 2.5, for example, means that 100 new manufacturing jobs would be expected to create a total of 250 jobs, of which 150 are indirectly created through spillovers; a multiplier greater than one indicates positive spillovers from an economic shock, while a multiplier of less than one indicates that, while new jobs are being created, they are displacing other jobs, offsetting the positive effects. Popular software to estimate these economic impacts and multipliers include private vendors IMPLAN and REMI among others, using research that dates to the 1950s. These tools are most effective in estimating the positive economic spillovers, but much less so in estimating displacement effects (see Drucker 2015 for a critical discussion of impact assessments from commercial software.) For instance, an incentivized office park typically just rearranges workers from existing local office parks to the new office park, leading to no net increase in jobs/income even if the new office park is full. Similarly, an incentivized restaurant may simply displace an existing restaurant, creating no net jobs or wealth. A new rural manufacturer could bid up wages and land prices, reducing the competitiveness of existing local firms. Another issue is the “false precision” from commercial software estimates, in which the software produces exact employment or income estimates by detailed sector, ignoring any indication of the potentially large standard errors in those estimates due to the imposition of a strict structure on the local economy. Economists are well-positioned to improve on the multiplier estimates for local economies using econometrically estimated multipliers, providing confidence intervals to the estimates, and avoiding the imposition of a strict structure on the local economy as is done by commercial software; in other words, allowing the data to speak. The recent surge of interest began with Moretti’s (2010) short multiplier paper, though the precise empirical tools were described in Bartik’s (1991) seminal contribution. While Moretti indirectly estimates the multiplier, estimating employment elasticities using the industry mix shift share term as an instrument for the change in employment from a key sector, such as manufacturing.15 However, Moretti’s approach suffers from several disadvantages: (a) it is nonlinear and multipliers are linear by definition; (b) to convert into a multiplier, the regression coefficient needs to be scaled by the relative size of the industry in question, which can be greatly affected by base year issues, especially when there is rapid growth; (c) it indirectly estimates the multiplier using the industry mix shift share term as an instrument for the change in employment from a key sector, such as manufacturing, and using instrumental variable estimation (IV) when more efficient/direct methods are available such as ordinary least squares (OLS) with best, linear, unbiased, estimators (BLUE).16 Bartik and others use a more direct method, replacing (for example) the manufacturing term with its industry mix shift share term and estimating the model with OLS. In this case, the regression coefficient on the manufacturing shift-share term is the multiplier and no further adjustments are needed—see Tsvetkova and Partridge (2016) for an empirical example of local energy-sector multipliers. Thus, as long as a researcher has detailed local data on industry shares (most likely a county or metro area), it is possible to identify an accurate multiplier that is not constrained by the strong assumptions used in commercial modeling software and produces a confidence interval that more realistically conveys the uncertainty in the estimates. Rural Data Good data are needed to evaluate policies and hold politicians, policymakers, and the private sector accountable for performance. The federal government has an important role in producing subnational statistics (and all data in general). These data are also essential for providing the private sector with key information for marketing and firm location and for academic research to identify better ways to achieve prosperity and improve quality-of-life (Partridge, Goetz, and Kilkenny 2013). The costs of federal statistical efforts are a small part of the overall federal budget, suggesting that a benefit-cost analysis would be astronomically in their favor.17 Despite the urgent need, there are long-running congressional efforts to reduce and eliminate federal statistical programs. One reason is that surveys such as the American Community Survey are perceived by some as asking sensitive questions about income or demographics (e.g., Rampell’s 2012’s discussion). However, the loss of federal statistical programs would have serious consequences and it is unlikely that the private sector can or would fill the gap. Private collection of data would suffer from the many problems associated with pure public goods and would be underprovided by the market. Such data are also likely to be expensive in order to allow vendors to recover the high costs. These expenses would reduce transparency in the government and increase costs for business and research. In addition, private sector data collection would inherently be less precise because federal agencies can incorporate administrative records such as tax data in their estimates. It would also likely be more ad hoc and less comparable over time. This would result in important business decisions being based on poor quality data, hurting the economy. Additionally, an important issue for rural America and rural researchers would be that private vendors would have little incentive to fill in the gap for rural data. Because the market for a tiny rural county’s socioeconomic data is so small, rural areas would be especially hurt by such efforts. At the same time that government data collection is in jeopardy, the last decade has seen a surge in potentially new data sources such as Google Analytics, Facebook data, as well as private companies’ customer and sales data. Without considering the ethical and privacy concerns of academics using such data for research, there are clearly some interesting questions that can be analyzed with Big Data that could not be done with conventional data. However, there are reasons to believe the Big Data fad is overhyped. Beyond selection issues, big Data shares with other private data the same problems of reliability and comparability over time, as well as problems for replication, a key standard for scientific research. And due to the fact that these data are collected by private vendors, for rural researchers, the availability of these data may be sparse. Rural Policy and Its Evaluation: Current and Future Research Directions Over the next few years, the federal government is likely to consider many policies that affect rural economies. Given the important differences across counties and the heterogeneity in what is rural, it is clear that a one-size-fits-all rural policy will not work. Federal labor, fiscal, trade, and interest rate policies all have disparate spatial impacts that depend, among other factors, on the export dependence and capital intensity of local industries. Most rural policy continues to be equated with farm policy, yet nutrition assistance programs make up 71% of the $151 billion in the USDA budget and are overwhelmingly paid out to urban consumers. Of the remaining USDA outlays, 16% are for farm and commodity programs, and 7% are for conservation and forestry (compared to agriculture accounting 5% of rural employment). The remainder (6%) is shared across rural development, research, food safety, marketing and regulation, and management functions. The policy-related word network analysis of Reimer et al. (2016) confirms that rural development issues are overwhelmingly framed as agricultural issues, even though a relatively small number of counties remain dependent on agriculture. Even though farm interests tend to be split along commodity lines, farmers overall have been able to articulate their policy choices effectively, largely due to the power of farm-state Senators (Reimer et al. 2016). It remains to be seen whether President Trump’s Administration will change rural policy to focus more on nonfarm rural issues, though early indications are that agriculture will remain the key rural focus. Policy researchers and political scientists concerned about rural economies should find this phenomenon to be an important area of inquiry (O’Brien and Ahearn 2016). Rural development (RD) is funded by the federal government through USDA as well as through the Small Business Administration (SBA), Economic Development Agency (EDA), and Appalachian Regional Commission (ARC), and, overall, it is a very small portion of federal spending. There is some evidence that the Trump Administration may not support current rural development (RD) programs, as indicated by initial proposals to strike key elements of the USDA RD infrastructure and USDA RD and SBA business financing activities, as well as to eliminate agencies such as the EDA and ARC (OPM 2017). On the one hand, it has long been recognized that consolidating federal economic development activities into one agency may be better than having them spread out over dozens of different departments and agencies that could lead to redundancies. Yet the early Trump Administration proposals suggest a relative neglect of rural areas in terms of producing sustainable economic development and aid for the most distressed communities. Hence, there is a role for researchers to provide sound evidence of rural policy effectiveness in order to inform the current administration as well as future policymakers. While place-based policies have a mixed track record—often due to poor implementation—there are examples of place-based policies that have potential, including the ARC, an expanded Delta Regional Authority, and other newer regional initiatives such as the Stronger Economies Together (SET) program hosted by the Southern Rural Development Center that do not pursue one-size-fits-all projects. In particular, given its long history, the ARC has received much attention; it provides bridge loans and seed grants for infrastructure and supports other programs such as workforce training. Yet given its limited resources, the ARC’s main role is as a broker that can foster regional collaborations of businesses, communities, nonprofits, local development districts, and various state and federal agencies while providing small funding matches to ensure projects can be implemented. The ARC also includes multi-county regional development districts (referred to as local development districts), which are composed of functional economic regions that can help facilitate cooperation and collaboration between neighboring communities. This may be especially important for rural communities that lack the capacity to tackle programs on their own. Since functional economic areas are predominantly centered on urban areas, this model also allows rural communities to work with their urban neighbors and benefit from urban-led growth through commuting, for example. The ARC, with a federal allocation of $90 million in FY 2015 and $146 in FY 2016 (with some temporary expenditures), is limited in what it can do for its over 25 million residents spread across 13 states. Despite the modest resources and the persistent underfunding of the agency, there is evidence that the ARC has been effective. Isserman and Rephann (1995) is one of many supportive studies finding that ARC counties had significantly better economic outcomes than observationally similar counties. A more comprehensive and longer-term recent study using propensity scoring and careful matching finds that per capita income in ARC-funded counties grew at a 5.5% higher rate than it did in non-funded control counties (Sayago-Gomez et al. 2017). Given the evidence that the ARC’s model may be one way to help distressed, rural regions, future research should evaluate the impact of similar federal-state regional development organizations in other rural areas. Recent research also suggests that outreach or other technical assistance programs are considerably more cost-effective than direct farm subsidy grants in keeping farmers on the farm (Goetz and Davlsheridze 2016). In other words, helping farmers through technical assistance or human capital training is more cost effective than is providing cash transfers. This is a critical research area that deserves more attention, especially given the importance of traditional federal farm programs, and the evidence that they may, in fact, accelerate the consolidation of farms and rural depopulation (Goetz and Debertin 2001). Future research could help increase our understanding of the interrelated effects of farm programs on farm size distribution, farm productivity, and the aging of farmers. Here the experience of New Zealand may offer fruitful guidance for structuring research questions (Johnston and Frengley 1991; Evans et al. 1996). A significant research void also exists in evaluating the effectiveness of existing USDA Rural Development programs and other rural development programs, including different types of policy designs and mechanisms. While important and carefully executed studies of individual programs exist, they are limited in number and they are piecemeal, focusing on only one program at a time. For example, research on the broadband loan programs (e.g., Whitacre Gallardo, and Strover 2014) shows somewhat mixed results, but finds that, on balance, it has positive effects. To understand the impact of these efforts and any synergies between programs, more holistic and comprehensive evaluations over time, ideally with panel data, of all rural development programs are needed. It would be especially useful to consider natural experiments when feasible, that is, where individuals or geographic areas are randomly assigned to treatments. Carefully designed matched comparisons or new approaches using synthetic control groups are other options, as are new opportunities to merge program data with secondary public records such as the National Employment Times Series data base or NETS. This kind of research is being encouraged by the Commission on Evidence-Based Policymaking. A critical research need lies in examining the benefits and the costs both of individual rural development policies and grant and loan programs, as well as their interactive effects. Related to this are questions of investing funds in people versus places. While economists have long held that investing in people is preferable to investing in places, as place-based investments tend to benefit the owners of fixed assets (e.g., in the form of higher housing prices), there is evidence that carefully-selected and targeted place-based investments can make a positive difference (Busso, Gregory, and Kline 2013), even though these claims are controversial (e.g., Hanson 2009; Hanson and Rohlin 2011). A strong interest has recently emerged in estimating the benefits and cost of local economic development programs and policies (e.g., Neumark and Simpson 2015). While much of this research has focused on more urban areas, extensions and applications to rural areas will likely have high pay-offs. One key lesson from this research is that picking winners and losers by subsidizing individual firms or industries is ineffective, if not counterproductive. Instead, it is important to improve overall economic conditions and workforce quality that benefit all businesses, so that the most competitive ones emerge. Further, economic spillovers in the form of externalities are also essential. For example, if a downtown business attracts consumers who also patronize other businesses on the same Main Street, a positive externality is created. In cases where barriers to entry prevent such public externalities, federal investments may be highly effective. In practice, some rural and remote communities likely lack the critical mass needed to generate the externalities that make such federal investments cost-effective. In that case, it is difficult to make the economic case for intervention. However, researchers can help identify such communities, as well as the thresholds needed to establish conditions for not just survival, but growth, though politically allocating funds on this basis will be difficult. Such research could use criteria such as a minimum set of businesses and services available within a given radius, while also considering population density and distance to major metropolitan areas—one example of which is provided by Stabler and Olfert (2002). Examining why some communities are resilient and able to bounce back from major shocks while others are not will also provide useful future policy insights (Han and Goetz 2015). This research may be especially valuable if it considers the historical local industry mix as well as location of the county on the rural-urban continuum. Closely related to this is the need to better understand the tradeoffs of having multiple layers of local rural (and urban) governments: these are gains from tailoring local policies versus redundancies, fewer spillovers across small jurisdictions, and the loss of economies of scale in service/infrastructure provision. Additional research is needed to understand what policies will support economic growth in rural areas. Evidence suggests that attracting large firms is typically unsuccessful, especially for rural areas. However, all areas have the potential to build a locally-diverse business environment based on higher-quality human capital. In addition, the type of parents who demand high-quality local schools are the types of workers a successful community needs. As mentioned previously, there is also evidence that even distressed areas can benefit from home-grown businesses. More research is needed on what types of entrepreneurs are effective in promoting growth and on what types of infrastructure and technical-assistance support (including business planning and other forms of education) are most effective in helping develop and grow local businesses. Economists can also assess the costs and benefits of improving rural education. For rural areas that have historically depended on resource extraction industries, environmental degradation is a threat to future growth. People are unwilling to remain in or migrate to areas that have been degraded and this impacts future economic prosperity. Regional economists in collaboration with agricultural economists and environmental economists have the expertise to provide evidence of the economic benefits of environmental protection and to evaluate ways to help communities transition to new a new economic paradigm. Our paper also identifies other areas of research that support this transition. Overall, while U.S. policymakers often eagerly support various programs and policies they believe will help their rural constituents, they generally have less patience for formal and systematic evaluation of such programs once implemented. It takes time for the full effects of programs and policies to play out, and rigorous and impartial assessments require additional resources. Above all, policymakers with a genuine interest in understanding the effects of their policies, and who want to make the best possible use of taxpayer dollars should be concerned about any efforts to scale back public data collection efforts. Additionally, a more balanced approach to RD in many parts of the country would place agriculture within the much larger rural nonfarm economy, rather than as the primary or sole economic sector. Researchers in agricultural economics and related professions are well-positioned to provide data-driven research to evaluate the effectiveness of new rural policies and support new solutions to help rural areas. Finally, while rigorous rural policy research and evaluation are critical, it is also essential that results be communicated to policymakers and their staff in formats that are easy to digest. This includes policy briefs, fact sheets, and on-demand webinars that review complex economic topics. One potential model for this is the National Agricultural and Rural Development Policy (NARDEP) Center, in which a research center is adequately funded to be a central point of research and policy outreach on critical rural policy issues and evaluation. While there are many sources of information on farm policy and urban policy, there is little about broad-based rural policies. Overall, colleges of agriculture have the experience in extension and outreach to make a difference in supporting policymaking related to rural areas. Footnotes 1 We use the terms metropolitan/urban and nonmetropolitan/rural interchangeably. 2 There are 2,494 nonmetro counties using the 1973 definition and 1,975 nonmetro counties under the 2013 definition. 3 Using International Monetary Fund data, we use the rapid rise in commodity prices beginning in 2003 and subsequent decline in 2014 to date the commodity price cycle (Index Chart 2000–2017 available at: http://www.imf.org/external/np/res/commod/index.aspx). 4 Rural areas have been swinging towards the GOP since the 1950s when Democratic Party support was about equal in cities and the countryside (Badger and Bui 2016), and this trend continued in the twenty-first century. Scala and Johnson (2017) document the decline in Democratic Party support over the 2000–2016 period in rural areas, with the decline being the most stark in the most remote rural regions. However, support for Democratic Presidential candidates literally collapsed between 2012 and 2016, with the average decline ranging from 7 to 10 percentage points depending on the type of nonmetropolitan county. 5 See https://www.statcan.gc.ca/pub/92-195-x/2011001/geo/cma-rmr/def-eng.htm for more details. 6 See https://en.wikipedia.org/wiki/Urban_area_(France) for more details. 7 The precise UICs are defined at: https://www.ers.usda.gov/data-products/urban-influence-codes/documentation/. 8 In discussions over the last 20 years, a common theme in rural development is that improvements in information communication technologies have made (or will make) it easier for people to live and work in rural areas because they are more connected to larger markets. Yet, such a view misses that the same information technologies are also making it easier for workers to telecommute and not physically commute to work every day, thereby improving their quality of life and generally reducing urban congestion—that is, making urban areas more livable. 9 Available at: http://www.wsj.com/articles/arch-coal-files-for-bankruptcy-1452500976. 10 In 1929, 1948, and 1970 (when the Clean Air Act of 1970 was passed), U.S. Bureau of Economic Analysis (BEA) data suggests that U.S. coal employment totaled 622,000, 533,000, and 146,000 people, respectively, illustrating that productivity growth was greatly reducing employment well before pollution regulations became binding, and showing the tremendous role of productivity growth prior to 1970. Illustrating the important role of coal booms relative to regulation, despite the Clean Air Act of 1970, the late 1970s energy shortage lifted coal mining employment to 254,000 workers. When the Clean Air Act of 1990 was passed, coal mining employment had fallen back to 148,000. Since then, U.S. Bureau of Labor Statistics data shows that coal employment has continued to decline to 50,500 workers in 2016 (between January 2017 and October 2017, coal mining employment increased by 2,000 workers). 11 For more details, see https://www.ers.usda.gov/topics/rural-economy-population/rural-poverty-well-being/poverty-overview/ (accessed November 22, 2017). 12 In 2013, just under 63% of U.S. counties were nonmetropolitan (see https://www.ers.usda.gov/topics/rural-economy-population/rural-classifications/). 13 There is a long literature discussing the tradeoff between equity and efficiency. Some examples include Okun (1975) and Welch (1999) with summaries in Barro (2000) and Partridge (1997; 2005). 14 A recent overview is contained in Patrick et al. 2017. 15 Rather than a linear term such as percentage change in total employment as the dependent variable, Moretti uses the log of employment as the dependent variable. For large employment changes over a decade, as in Moretti’s case, this change could matter. 16 The manufacturing industry mix term %ΔIMMAN equals ∑iShir0%ΔNEMPit in which Sh is area r’s initial period employment share of industry of i and %ΔNEMPit the percent change in national employment in industry i. Summing over all industries i in the manufacturing sector produces the predicted manufacturing growth rate in the area if all of the industries in the manufacturing sector are growing at the national growth rate. Unless there is some labor supply response to the initial industry shares (or one that is not accounted for by a demographic control variable), because industry national growth rates are used, the shift-share term is assumed to be exogenous. If one needs a multiplier for another sector, then the shift-share should be calculated for that industry grouping. 17 The total costs of the three main federal statistical agencies—Bureau of Economic Analysis, Census Bureau, and Bureau of Labor Statistics—is approximately $2.6 billion in fiscal year 2017, or about 0.06% out of a total federal budget of about $4.1 trillion. References Acs Z.J., Braunerhjelm P., Audretsch D., Carlsson B. 2009. The Knowledge Spillover Theory of Entrepreneurship. Small Business Economics  32 ( 1): 15– 30. Google Scholar CrossRef Search ADS   Akgün A.A., Nijkamp P., Baycan T., Brons M. 2010. Embeddedness of Entrepreneurs in Rural Areas: A Comparative Rough Set Data Analysis. Tijdschrift Voor Economische en Sociale Geografie  101: 538– 53. Google Scholar CrossRef Search ADS   Angrist J. D., Pischke J. 2009. Mostly Harmless Economics, An Empiricist’s Companion. Princeton, NJ: Princeton University Press. Atems B. 2013. A Note on the Differential Regional Effects of Income Inequality: Empirical Evidence Using U.S. County Level Data. Journal of Regional Science  53: 656– 71. Google Scholar CrossRef Search ADS   Athey S., Imbens G.W. 2017. The State of Applied Econometrics: Causality and Policy Evaluation. Journal of Economic Perspectives  31 ( 2): 3– 32. http://dx.doi.org/10.1257/jep.31.2.3 Google Scholar CrossRef Search ADS   Autor D.H., Dorn D., Hanson G.H. 2013. The China Syndrome: Local Labor Market Effects of Import Competition in the United States. American Economic Review  103: 2121– 68. Google Scholar CrossRef Search ADS   Badger E., Bui Q. 2017. Why Republicans Don’t Even Try to Win Cities Anymore. New York Times. November 2. Available at: https://www.nytimes.com/2016/11/03/upshot/why-republicans-dont-even-try-to-win-cities-anymore.html?_r=0. Barca F., McCann P., Rodríquez-Pose A. 2012. The Case for Regional Development Intervention: Place–Based versus Place–Neutral Approaches. Journal of Regional Science  52: 134– 52. Google Scholar CrossRef Search ADS   Barro R.J. 2000. Inequality and Growth in a Panel of Countries. Journal of Economic Growth  5: 5– 32. Google Scholar CrossRef Search ADS   Bartik T.J. 1991. Who Benefits from State and Local Economic Development Policies? Kalamazoo, MI: W. E. Upjohn Institute for Employment Research. Google Scholar CrossRef Search ADS   Berg A.G., Ostry J.D. 2011. Inequality and Unsustainable Growth: Two Sides of the Same Coin? IMF Staff Discussion Note, SDN/11/08, April 8, 20. Betz M., Farren M., Lobao L., Partridge M.D. 2015. Coal Mining, Economic Development, and the Natural Resources Curse. Energy Economics  50: 105– 16. Google Scholar CrossRef Search ADS   Busso M., Gregory J., Kline P. 2013. Assessing the Incidence and Efficiency of a Prominent Place Based Policy. The American Economic Review  103: 897– 947. Google Scholar CrossRef Search ADS   Carruthers J., Vias A.C. 2005. Urban, Suburban, and Exurban Sprawl in the Rocky Mountain West: Evidence from Regional Adjustment Models. Journal of Regional Science  45: 21– 48. Google Scholar CrossRef Search ADS   Chetty R., Hendren N., Katz L.F. 2016. The Effects of Exposure to Better Neighborhoods on Children: New Evidence from the Moving to Opportunity Experiment. The American Economic Review  106: 855– 902. Google Scholar CrossRef Search ADS   Chetty R., Hendren N., Kline P., Saez E. 2014. Where is the Land of Opportunity? The Geography of Intergenerational Mobility in the United States. The Quarterly Journal of Economics  129: 1553– 623. Google Scholar CrossRef Search ADS   Deller S.C., Tsai T-H., Marcouiller D.W., English D.B.K. 2001. The Role of Amenities and Quality of Life in Rural Economic Growth. American Journal of Agricultural Economics  83: 352– 65. Google Scholar CrossRef Search ADS   Drucker J. 2015. Economic Impact Analysis amid Rapid Change: Challenges, Strategies, and Examples from Defense Communities. Journal of Planning Education and Research  35: 401– 18. Google Scholar CrossRef Search ADS   Evans L., Grimes A., Wilkinson B., Teece D. 1996. Economic Reform in New Zealand 1984-95: The Pursuit of Efficiency. Journal of Economic Literature  34: 1856– 1902. Fan Q., Goetz S.J., Liang J. 2016. The Interactive Effects of Human Capital and Quality of Life on Economic Growth. Applied Economics  48: 5186– 200. Google Scholar CrossRef Search ADS   Fleming D.A., Goetz S.J. 2011. Does Local Firm Ownership Matter? Economic Development Quarterly  53: 277– 81. Google Scholar CrossRef Search ADS   Fritsch M., Wyrwich M. 2014. The Long Persistence of Regional Levels of Entrepreneurship: Germany, 1925–2005. Regional Studies  48: 955– 73. Google Scholar CrossRef Search ADS   Goetz S.J., Davlasheridze M. 2016. State-Level Cooperative Extension Spending and Farmer Exits. Applied Economic Perspectives and Policy  39( 1): 65– 86. Goetz S.J., Davlasheridze M., Han Y. 2015. County-Level Determinants of Mental Health, 2002–2008.  Social Indicators Research  24: 657– 70. Google Scholar CrossRef Search ADS   Goetz S.J., Davlasheridze M., Han Y., Fleming D. 2017. Explaining the 2016 Vote for President Trump in U.S. Counties. In Review. Goetz S.J., Debertin D.L. 2001. Why Farmers Quit: A County-Level Analysis. American Journal of Agricultural Economics  83 ( 4): 1010– 23. http://dx.doi.org/10.1111/0002-9092.00226 Google Scholar CrossRef Search ADS   Gundersen C., Ziliak J. 2018. Food Insecurity Research in the United States: Where We Have Been and where We Need to Go. Applied Economic Perspectives and Policy  40 ( 1): 119– 35. Han Y., Goetz S.J. 2017. Overlapping Labor Market Areas Based on Link Communities. Papers in Regional Science. DOI: 10.1111/pirs.12326 Han Y., Goetz S.J. 2015. The Economic Resilience of US Counties During the Great Recession. Review of Regional Studies  45: 131– 149. Hanson A. 2009. Local Employment, Poverty, and Property Value Effects of Geographically-Targeted Tax Incentives: An Instrumental Variables Approach. Regional Science and Urban Economics  39: 721– 31. Google Scholar CrossRef Search ADS   Hanson A., Rohlin S. 2011. Do Location-Based Tax Incentives Attract New Business Establishments? Journal of Regional Science  51 ( 3): 427– 49. Google Scholar CrossRef Search ADS   Isserman A. 2005. In the National Interest: Defining Rural Correctly for Research and Policy. International Regional Science Review  28 ( 4): 465– 99. http://dx.doi.org/10.1177/0160017605279000 Google Scholar CrossRef Search ADS   Isserman A.M., Rephann T. 1995. The Economic Effects of the Appalachian Regional Commission. Journal of the American Planning Association  61: 345– 64. Google Scholar CrossRef Search ADS   Johnston W.E., Frengley G.A. 1991. The Deregulation of New Zealand Agriculture: Market Intervention (1964–84) and Free Market Readjustment (1984–90). Western Journal of Agricultural Economics  16: 132– 43. Kelsey T., Partridge M.D., White N. 2016. Unconventional Gas and Oil Development in the United States: Economic Experience and Policy Issues. Applied Economic Perspectives and Policy  38 ( 2): 191– 214. http://dx.doi.org/10.1093/aepp/ppw005 Google Scholar CrossRef Search ADS   Kilkenny M., Partridge M.D. 2009. Export Sectors and Rural Development. American Journal of Agricultural Economics  91 ( 4): 910– 29. http://dx.doi.org/10.1111/j.1467-8276.2009.01320.x Google Scholar CrossRef Search ADS   Levernier W., Partridge M.D., Rickman D.S. 1998. Differences in Metropolitan and Nonmetropolitan U.S. Family Income Inequality: A Cross-Country Comparison. Journal of Urban Economics  44: 272– 90. Google Scholar CrossRef Search ADS   Li M., Goetz S.J., Partridge M., Fleming D.A. 2016. Location Determinants of High-Growth Firms. Entrepreneurship and Regional Development  8: 97– 125. Google Scholar CrossRef Search ADS   Liang J., Goetz S.J. 2016. Self-Employment and Trade Shock Mitigation. Small Business Economics  46: 45– 56. Google Scholar CrossRef Search ADS   Low S., Henderson J., Weiler S. 2005. Gauging a Region’s Entrepreneurial Potential. Economic Review.  Third Quarter 61–89. Kansas City, MO: Federal Reserve Bank of Kansas City. McGranahan D. 2008. Landscape Influence on Recent Rural Migration in the U.S. Landscape and Urban Planning  85: 228– 40. Google Scholar CrossRef Search ADS   Moretti E. 2010. Local Multipliers. American Economic Review  1100: 373– 77. Google Scholar CrossRef Search ADS   Moy E., Garcia M.C., Bastian B., Rossen L.M., Ingram D.D., Faul M., Massetti G.M., et al.   2017. Leading Causes of Death in Nonmetropolitan and Metropolitan Areas—United States, 1999–2014. MMWR. Surveillance Summaries  66: 1– 8. Google Scholar CrossRef Search ADS   National Academies of Sciences, Engineering, and Medicine. 2016. Rationalizing Rural Area Classifications for the Economic Research Service: A Workshop Summary.  Washington, DC: The National Academies Press. Available at: https://doi.org/10.17226/21843. Neumark D., Simpson H. 2015. Place-Based Policies. In Handbook of Regional and Urban Economics , eds. Duranton G., Henderson J.V., Strange W.C., 1197– 287. Amsterdam, Netherlands: Elsevier B.V. Okun A.M. 1975. Equality and Efficiency: The Big Tradeoff . Washington: Brookings Institution. O’Brien D.J., Ahearn M.C. 2016. Rural Voice and Rural Investments: The 2016 Election and the Future of Rural Policy. Choices , 4th Quarter. U.S. Office of Personal Management. 2017. America First: A Budget Blueprint to Make America Great Again.  Available at: https://www.whitehouse.gov/sites/whitehouse.gov/files/omb/budget/fy2018/2018_blueprint.pdf. Patrick C., Ross A., Stephens H. 2017. Designing Policies to Spur Economic Growth How Regional Scientists Can Contribute to Future Policy Development and Evaluation. In Regional Research Frontiers: The Next 50 Years , ed. Jackson R., Schaeffer P. pp 119– 133. New York, New York, USA: Springer. Google Scholar CrossRef Search ADS   Partridge M.D. 2010. The Dueling Models: NEG vs Amenity Migration in Explaining US Engines of Growth. Papers in Regional Science  89: 513– 36. Google Scholar CrossRef Search ADS   Partridge M.D., Weinstein A.L. 2013. Rising Inequality in an Era of Austerity: the Case of the US. European Planning Studies  1: 388– 410. Google Scholar CrossRef Search ADS   Partridge M.D., Rickman D.S., Ali K., Olfert M.R. 2008. Lost in Space: Population Growth in the American Hinterlands and Small Cities. Journal of Economic Geography  8: 727– 57. Google Scholar CrossRef Search ADS   Partridge M.D., Rickman D.S., Ali K., Olfert M.R. 2010. The Spatial Dynamics of Factor Price Differentials: Productivity or Consumer Amenity Driven? Regional Science and Urban Economics  40: 440– 52. Google Scholar CrossRef Search ADS   Partridge M.D., Olfert M.R. 2011. The Winners' Choice: Sustainable Economic Strategies for Successful 21st Century Regions. Applied Economic Perspectives and Policy  33: 143– 78. Google Scholar CrossRef Search ADS   Partridge M.D., Goetz S.J., Kilkenny M. 2013. Why We Need Federal Statistical Data for States and Counties. Choices  28. Available at:http://www.choicesmagazine.org/choices-magazine/submitted-articles/why-we-need-federal-statistical-data-for-states-and-counties. Partridge M.D., Rickman D.S., Olfert M.R., Tang Y. 2017. International Trade and Local Labor Markets: Do Foreign and Domestic Shocks Affect Regions Differently? Journal of Economic Geography  7: 375– 409. Piketty T. 2013. Capital in the Twenty-First Century.  Boston, Massachusetts: Belnap Press/Harvard. Reimer A., Han Y., Goetz S.J., Loveridge S., Albrecht D. 2016. Word Networks in US Rural Policy Discourse. Applied Economic Perspectives and Policy  38: 215– 38. Google Scholar CrossRef Search ADS   Rampell C. 2012. The Beginning of the End of the Census? New York Times. May 19, 2017. Available at:http://www.nytimes.com/2012/05/20/sunday-review/the-debate-over-the-american-community-survey.html. Rupasingha A., Goetz S.J. 2013. Self‐Employment and Local Economic Performance: Evidence from US Counties. Papers in Regional Science  92: 141– 61. Rupasingha A., Goetz S.J. 2007. Social and Political Forces as Determinants of Poverty: A Spatial Analysis. The Journal of Socio-Economics  36: 650– 71. Google Scholar CrossRef Search ADS   Sayago-Gomez J-T., Piras G., Jackson R., Lacombe D. 2017. Impact Evaluation of Investments in the Appalachian Region: A Reappraisal. International Regional Science Review  DOI: 10.1177/0160017617713822. Scala D., Johnson K.M. 2017. Political Polarization along the Rural–Urban Continuum? The Geography of the Presidential Vote 2000–2016. The Annals of Political and Social Science  672: 162– 84. Google Scholar CrossRef Search ADS   Stabler J.C., Olfert M.R. 2002. Saskatchewan’s Communities in the 21st Century: From Places to Regions.  Regina, SK: Canadian Plains Research Center, University of Regina. Stephens H., Partridge M. 2011. Do Entrepreneurs Enhance Economic Growth in Lagging Regions? Growth and Change  42: 431– 65. Google Scholar CrossRef Search ADS   Stephens H., Partridge M., Faggian A. 2013. Innovation Entrepreneurship and Economic Growth in Lagging Regions. Journal of Regional Science  53: 778– 812. Google Scholar CrossRef Search ADS   Tsvetkova A., Partridge M.D. 2016. Economics of Modern Energy Boomtowns: Do Oil and Gas Shocks Differ from Shocks in the Rest of the Economy? Energy Economics  59: 81– 95. Google Scholar CrossRef Search ADS   Tsvetkova A., Partridge M.D., Betz M. 2017. Entrepreneurial and Wage and Salary Employment Across the Rural-Urban Continuum. The ANNALS of the American Academy of Political and Social Science  672: 83– 102. Google Scholar CrossRef Search ADS   U.S. Department of Agriculture. 2016. FY 2017 Budget Summary, Office of Budget and Program Analysis, Washington, DC. Available at: http://www.obpa.usda.gov/budsum/fy17budsum.pdf. Weber B., Fannin M.J., Miller K., Goetz S.J. 2017. Intergenerational Mobility of Low-Income Youth in Metropolitan and Nonmetropolitan America: A Spatial Analysis. Working Paper. Wehrwein G., Johnson H.A. 1943. A Recreation Livelihood Area. Journal of Land and Public Utility Economics  19 ( 2): 193– 206. http://dx.doi.org/10.2307/3159233 Google Scholar CrossRef Search ADS   Weiler S. 2001. Unemployment in Regional Labor Markets: Using Structural Theories to Understand Local Jobless Rates in West Virginia. Industrial and Labor Relations Review  54: 573– 92. Google Scholar CrossRef Search ADS   Welch F. 1999. In Defense of Inequality. American Economic Review  89: 1– 17. Google Scholar CrossRef Search ADS   Whitacre B., Gallardo R., Strover S. 2014. Does Rural Broadband Impact Jobs and Income? Evidence from Spatial and First-Differenced Regressions. The Annals of Regional Science   53: 649– 70. Google Scholar CrossRef Search ADS   Wooldridge J.M. 2010. Econometric Analysis of Cross Section and Panel Data . Cambridge, MA: MIT Press. Wu J., Gopinath M. 2008. What Causes Spatial Variations in Economic Development in the United States? American Journal of Agricultural Economics  90: 392– 408. Google Scholar CrossRef Search ADS   Ziliak J.P. 2015. Income, Program Participation, Poverty, and Financial Vulnerability: Research and Data Needs. Journal of Economic and Social Measurement  40: 27– 68. Google Scholar CrossRef Search ADS   © The Author(s) 2018. Published by Oxford University Press on behalf of the Agricultural and Applied Economics Association. All rights reserved. For Permissions, please e-mail: journals.permissions@oup.com http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Applied Economic Perspectives and Policy Oxford University Press

The Economic Status of Rural America in the President Trump Era and beyond

Loading next page...
 
/lp/ou_press/the-economic-status-of-rural-america-in-the-president-trump-era-and-4dDTr7Rn5v
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of the Agricultural and Applied Economics Association. All rights reserved. For Permissions, please e-mail: journals.permissions@oup.com
ISSN
2040-5790
eISSN
2040-5804
D.O.I.
10.1093/aepp/ppx061
Publisher site
See Article on Publisher Site

Abstract

Abstract To set the stage for future research aimed at developing public policies that support economic prosperity in rural areas, we review the current economic conditions of rural America and the current literature. Rural America is often characterized as a uniform, distressed place where agriculture dominates. In fact, rural America is diverse, with many regions doing well economically. In some areas, labor-saving technologies have reduced the workforce in manufacturing and resource-dependent industries. However, integration with urban areas has weakened the economic divide between urban and some rural areas, while natural amenities have boosted the fortunes of others. There is also evidence that homegrown enterprises can support growth even in the most remote, distressed regions. To support economic growth, policies should recognize the unique features of each place or region and balance the farm sector with the larger nonfarm rural economy. Economists are well-positioned to provide research-based evidence of what works, as well as rigorous evaluation of new polices. Rural America is an important source of America’s food, water, energy, and other natural resources, and it contains areas of great natural beauty. However, urban Americans rarely give much thought to their rural compatriots. One reason may be that rural America appears to be in decline. Another is that few urbanites have a personal attachment to rural America. While rural-to-urban migration heavily dominated from the 1920s to the 1970s, it has since tapered off, meaning fewer family connections between the two. A wave of recent events, culminating with the 2016 Presidential election, has put rural America under a new spotlight (Goetz et al. 2017). Nonmetropolitan America voted for President Trump in higher percentages than metropolitan America, propelling him into the White House.1 A perceived growing divergence between urban and rural America has renewed interest in understanding rural America. In this paper we provide an overview of rural America and the economic issues relevant to the agricultural economics and related professions, within the context of the new administration and beyond. As the last election illustrated, rural America and the conditions surrounding its residents are important to the national economy, and rural voters matter. Yet even defining what constitutes a rural area (“one knows it when one sees it”) remains a challenge, and rural definitions have changed over time. We start by examining recent population trends and how they have affected rural/urban classifications. Then we review the changing industrial make-up of rural areas and attendant social and economic problems. We also present issues surrounding data availability for analyzing and modeling conditions in rural regions that form the basis for modern regional development research. The paper concludes with a discussion of rural programs and recommendations for future research directions and policy analysis for economists. Rural Population Change and Recent Voting Trends Some historical background about rural areas is helpful in prioritizing future research directions. Labor-saving technological change in agriculture and other primary sectors led to an expansion of output that has benefited all global citizens. Yet it also dramatically reduced the demand for labor in rural America, leading to migration to urban areas. At the same time, improvements in automobiles and construction of modern highways that unlocked rural-to-urban commuting opportunities mitigated these effects and allowed some of the displaced labor to find work while remaining in and supporting their rural communities. This in turn led to the emergence of large city-centered functional economic regions in which the largest urban center typically represents the engine of growth for rural and urban residents alike. During this same period, rural communities located in areas with attractive landscapes, mountains, lakes, oceans, and a pleasant climate benefited from amenity-driven in-migration (Partridge 2010). The net effect was that, beginning in the 1970s, rural areas with access to urban labor markets and high amenities essentially reversed the long-run net out-migration from rural areas. Rural America is now diverse and comprised of three distinct types of areas: (a) high-amenity regions; (b) metro-adjacent rural communities; and (c) remote or extractive-based rural communities; it is the latter that have generally struggled. To illustrate this, figures 1 and 2 show population growth over the 1969–2015 and 2000–2015 periods in nonmetropolitan and metropolitan areas. Shown in the legend in parentheses is the number of counties in each definition, with a declining number of nonmetro counties evident over time. Population growth is measured in several ways: (a) using the 1950 metropolitan area definitions that reflect older core cities with virtually no recent history of a “rural” heritage; (b) using 1973 definitions that reflect the beginning of urban expansion into suburban and exurban areas; and (c) using subsequent redefinitions of metropolitan areas in 1983, 1993, 2003, and 2013. The 2003 and 2013 definitions include more counties that had previously been considered rural, due to a classification change in which the U.S. Census Bureau lowered the cross-commuting threshold required for a county to be part of a metro region to 25%. In many cases, the recent additions are far on the periphery of the urban center, appearing to many observers as more rural than urban—even though there are moderate urban labor linkages. Figure 1 View largeDownload slide Population growth of non-metro area by historical MSA definition: 100 = 1969 Population Data Sources: U.S. Bureau of Economic Analysis for population (2017), and U.S. Census Bureau for metro definitions (2017). Figure 1 View largeDownload slide Population growth of non-metro area by historical MSA definition: 100 = 1969 Population Data Sources: U.S. Bureau of Economic Analysis for population (2017), and U.S. Census Bureau for metro definitions (2017). Figure 2 View largeDownload slide Population growth of metro area by historical MSA definition: 100 = 1969 Population Data Sources: U.S. Bureau of Economic Analysis for population (2017) and U.S. Census Bureau for metro definitions (2017). Figure 2 View largeDownload slide Population growth of metro area by historical MSA definition: 100 = 1969 Population Data Sources: U.S. Bureau of Economic Analysis for population (2017) and U.S. Census Bureau for metro definitions (2017). The metropolitan definitions are split across time to make several points. First, the validity of the notion that rural America is in decline depends on the definition of rural. Using the (virtually) current 2013 definition, the metro population grew by 67% between 1969–2015 (see figure 2), compared to 26% for nonmetro areas (see figure 1). However, much of this apparent lagging performance for rural America relates to reclassification over time of the fastest-growing nonmetropolitan counties as metropolitan (Partridge et al. 2008). The late Andy Isserman argued that this metropolitan area reclassification is like taking the best team out of a sports league every year; eventually the league would become low-quality (Isserman 2005). Using the most urbanized 1950 definitions, figures 1 and 2 show that 1969–2015 and 2000–2015 population growth in nonmetropolitan (rural) counties was double that of metropolitan counties in the earlier period, and also greatly exceeded it in the twenty-first century. In other words, counties considered rural in 1955 have done quite well. As previously nonmetropolitan counties have been redefined as metropolitan, nonmetro growth has been subsequently slower over time and metro county growth has increased; in other words, the reclassified counties were actually growing more rapidly than the original metro counties. In that sense the story of rural malaise is just as much a story of some nonmetro counties being remarkably successful. For instance, as shown in figure 1, using the 1973 nonmetropolitan definition, rural population growth equaled the national (and metro) average up until 2007. Moreover, if the 500 counties that were reclassified as metropolitan between 1973 and 2013 had remained in the nonmetro category, the rural population growth rate would have tripled.2 Figures 1 and 2 also illustrate an inflection point in 2007 when relative rural population growth slowed across all definitions, while relative metropolitan growth either slightly increased or showed no discernable change. This suggests that both current nonmetropolitan areas and reclassified (former) nonmetro counties are experiencing slower growth. Some of the slowdown relates to the housing bust’s effects on reducing exurban residential development (and these more recent metro counties are more remote). Another is that lower-wage rural manufacturing has experienced productivity growth and fierce foreign competition that suppressed job growth (though since 2001, urban and rural manufacturing have performed about equally in terms of negative job growth). It is not surprising that metro job growth exceeds rural job growth, mainly because many rural residents commute to urban jobs. Yet, using 2013 definitions, between 2001–2007 and 2007–2015, total nonmetropolitan job growth slowed from a positive 4.6% to a decrease of 0.3%, whereas metro job growth slowed from 9.7% to 6.6%, further increasing the rural-urban employment gap. Another factor that has hurt rural areas is the decline in industries that previously employed many workers. Even though the circa 2003–2013 super-commodity cycle should have helped rural areas (as was the case with the 1970s super cycle), agriculture continues to shed employment and rural mining job growth has lagged.3 This shows that even traditional rural engines are no longer sufficient to lift local growth (on average). President Trump, to his political credit, exploited this (relatively recent) economic decline in rural America in the 2016 election.4Table 1 divides the 2016 vote into metro/nonmetro areas using the 1950 and 2013 definitions based on when the previously rural county was reclassified as metropolitan. The 1950 metro group only gave President Trump 37% of the vote, but those reclassified between 1950 and 1973 gave him 53%. President Trump’s vote share steadily increased to 63% for the group of counties redefined between 2003 and 2013. The 2003–2013 reclassified counties’ 63% Trump vote share almost equaled his 2013 nonmetropolitan county share of 64%. There are other reasons (including economics) for this pattern, but the amount of time an area has been classified as part of a metro area is a good predictor of how President Trump performed; in other words, how recently they were considered “rural.” Future research could examine the reasons for the divide between long-time urban and recently-integrated urban areas; this includes examining whether as a county becomes urban, what is the dynamic process for losing its rural heritage and assimilating into urban culture. This also has implications for research into the geographical extent of regional governance. Table 1 2016 Election Results for Trump   Metro Definition and Change  Total Metro Votes  Trump Metro Votes  Trump Metro Votes Share  (1)  Metro 1950  68,380,703  25,469,903  37%  (2)  Metro Def. Change 1950–1973  26,522,379  14,147,642  53%  (3)  Metro Def. Change 1973–1983  6,245,681  3,457,467  55%  (4)  Metro Def. Change 1983–1993  3,764,839  2,214,950  59%  (5)  Metro Def. Change 1993–2003  3,548,019  2,075,511  60%  (6)  Metro Def. Change 2003–2013  2,040,345  1,298,349  63%  (7)  Metro 2013  110,501,966  48,663,822  44%  (8)  Nonmetro 2013  19,409,615  12,401,021  64%  (9)  Total United States  129,911,581  61,064,843  47%    Metro Definition and Change  Total Metro Votes  Trump Metro Votes  Trump Metro Votes Share  (1)  Metro 1950  68,380,703  25,469,903  37%  (2)  Metro Def. Change 1950–1973  26,522,379  14,147,642  53%  (3)  Metro Def. Change 1973–1983  6,245,681  3,457,467  55%  (4)  Metro Def. Change 1983–1993  3,764,839  2,214,950  59%  (5)  Metro Def. Change 1993–2003  3,548,019  2,075,511  60%  (6)  Metro Def. Change 2003–2013  2,040,345  1,298,349  63%  (7)  Metro 2013  110,501,966  48,663,822  44%  (8)  Nonmetro 2013  19,409,615  12,401,021  64%  (9)  Total United States  129,911,581  61,064,843  47%  Source: Townhall.com. Missing county-level data for Alaska. Current metropolitan counties are split into those defined in 1950 and those that were subsequently reclassified as metropolitan by time period of reclassification. Rural County Classifications and the Rural-Urban Continuum Policymakers and researchers who are examining “rural” areas need clear definitions of what is rural. While there is no binary rural/urban dichotomy, some sort of strict division is necessary for informing policy and related research. Otherwise, given the numerous definitions and the poor understanding of the tradeoffs involved in using each definition, poor academic research and policy may be the result. Given the changing definition of rural, the National Academies of Sciences recently hosted a workshop defining what is rural (National Academy of Sciences, Engineering and Medicine 2016). In practice, an urbanized continuum stretches out from the urban core, through the suburbs to exurban areas with high urban commuting rates, to rural areas (and very sparse and remote rural could be added as a distinct category). However, there are two key ways of thinking about rural in trying to make a reasonable division. First is the popular notion that is often portrayed by the media of rural being sparsely populated, with a relatively large number of farms or a high concentration in other primary sectors–that is, the “real rural.” In this case, rural definitions based on population density would be useful in understanding rural/urban differences in culture and other beliefs. However, a population density-based notion of rural is simplistic and can produce misleading results when conducting economic analysis. Namely, local economic questions relate to behaviors in functional economic regions. If people live and work as well as conduct their other economic business in the same region, then that is their functional economic region regardless of the density of the resident’s particular “neighborhood.” For example, a researcher examining the effects of job growth on poverty should not separate functional economic regions into heterogeneous groups that are unrelated to the underlying economic processes. Fortunately for those who are conducting spatial economic research, the U.S. Census Bureau’s current metropolitan (micropolitan)/nonmetropolitan (core rural) designations are functional economic areas, which should be useful in most cases of economic research. The census uses a 25% commuting rate threshold and a minimum principal-city size of 50,000 in defining metropolitan area thresholds. Economists interested in rural economic issues should not apologize for using this delineation because it is exactly what theory suggests. Yet this does not mean that the current metropolitan definition is without problems. In 1950, it may have been reasonable to assume that an urban core of 50,000 people represented sufficient agglomeration economies in the United States to be “urban.” Yet agglomeration thresholds appear to be increasing over time as economic functions that once were found in smaller locales are now only found in larger places (Partridge et al. 2008; 2010), and the current definitions may not accurately reflect today’s concept of urban versus rural. Additionally, in the current definition (and beginning with the 2003 definition), the commuting threshold or the percentage of a county that cross-commutes with the principal city was lowered to the current 25% level; the result is that the 50 to 100 most populous metro areas are typically geographically spread out up to a radius of 100 miles or more. This also means that more distant counties from the urban core were reclassified as metropolitan. These are counties with a lower population density and a little higher reliance on agriculture and resource-based industries; that is, the type of reclassified counties that voted in large numbers for President Trump. The relatively low commuting threshold extends metro areas to peripheral rural places that are less economically integrated with the urban center. We suggest that the commuting threshold be raised to 30% or 35% to mitigate this problem of classifying true “rural” counties as metropolitan. In 1950 it may have been reasonable to assume that an urban core of 50,000 people represented sufficient agglomeration economies to be “urban.” Yet, agglomeration thresholds appear to be increasing over time as economic functions that were found in smaller locales are now only found in larger places (Partridge et al. 2008; 2010) and the current definitions may not accurately reflect today’s concept of urban versus rural. Additionally, The U.S. definition is also significantly different from that used in Canada and Europe, especially with regard to the commuting threshold. Compared to the 50,000 urban threshold used in the United States, Statistics Canada uses a 100,000 total population threshold and a 50% commuting threshold to define metropolitan areas.5 The French “urban area” definition (the closest to the U.S. metro definition) is based on a threshold of an urban core of at least 10,000 jobs and a 40% commuting threshold for neighboring municipalities.6 Economists should provide research on the impact of alternative definitions that are based on more updated concepts of urbanization and economic integration and are more in line with those in other countries. Alternative definitions of urban versus rural have been developed by the USDA. Commuting zones (CZ) are based on clustering of commuting patterns; however, the most current definitions are from 1993. While they are increasingly used in many labor economics papers (e.g., Autor, Dorn, and Hanson 2013; Chetti et al. 2014), the CZ definitions are both outdated, and unlike the Census definitions, are not derived based on any economic theory, but cluster analysis (Partridge et al. 2017). For example, there are no commuting thresholds and thus weakly integrated rural areas are connected together as a CZ and even low-integrated rural counties with weak commuting thresholds below 25% are included with metropolitan areas. Likewise, highly integrated metropolitan areas are split into separate CZs (even the same city is split) without any coherent rationale. They also do not account for the fact that counties often send commuters to multiple destination counties, especially in the densely settled Northeastern United States. Recent research seeks to circumvent some of the problems with CZs and MSAs by using a link (or edge) community method from network theory to construct overlapping commuting networks, where counties can belong to multiple Labor Market Areas (LMAs) at the same time (Han and Goetz 2017). The study shows that counties with high employment in primary industries are more likely to belong to a single LMA, while counties with more households of two or more workers more likely belong to multiple such LMAs. The second major USDA categorization includes the urban influence codes (UICs) and Beale codes (BC). For example, the 2013 UICs split metropolitan areas into two groups —those with more and less than one million residents. The nonmetro counties are divided into ten groups, based on a three-way division: (a) micropolitan areas; (b) noncore rural areas with a city of greater than 2,500; and (c) noncore rural areas with a city less than 2,500.7 These codes were first introduced over 30 years ago, and historically showed that growth was much higher in counties adjacent to metro areas and that, conversely, remote locations without a city of at least 2,500 people were generally stagnant or in decline (Wu and Gopinath 2008). However, over time, especially after 2000, the BCs and the UIC have become less valuable and have explained less of the variation in socioeconomic outcomes. This is partly due to the changes in the U.S. Census definitions for metropolitan areas, on which these classification schemes rely. The low 25% commuting threshold means that counties that would have previously been considered adjacent nonmetropolitan are now designated metropolitan. Now rural counties “adjacent” to metropolitan areas can actually be quite remote, meaning that the expected positive spillovers are considerably less than before. Likewise, when these codes were introduced, the economic importance of a town of 2,500 was different from today for the same reasons that a 50,000 urban-core population no longer may be insufficiently large to capture sufficient agglomeration economies to be classified metropolitan. While the value of the BCs and UICs is waning, they were helpful in gaining a better understanding of how many economic outcomes change with remoteness from urban centers (Partridge et al. 2008; 2010). This has led to a more recent body of literature that more finely describes distance effects than simple adjacency (e.g., Tsvetkova, Partridge, and Betz 2017; Weber et al. 2017) and illustrates that the rural remoteness costs are rising over time as technological change favors larger cities (e.g., Partridge et al. 2010).8 The Industrial Composition of Rural Areas Given the last election results, it is important to understand how the rural economy has changed. For the most remote, rural regions, economic integration with urban areas is limited. Historically, the economy in these regions is more likely to have been dominated by agriculture or other resource-based industries. These regions have been hurt by changes in these industries, especially increased mechanization. Focusing on agriculture, figure 3 shows the 1969–2015 farm share of employment using Bureau of Economic Analysis (BEA) data for nonmetropolitan areas, metropolitan areas, and the total United States, as well as Current Population Survey data for the U.S. share of total employment in which agriculture is the primary job. In rural America, the farm employment share declined from 15% in 1969 to 6% in 2015, illustrating a profound restructuring. Likewise, the farm share also declined in urban areas, but agriculture is a very small portion of urban economies. Figure 3 View largeDownload slide Percentage of total jobs in farming: 1969–2015 Data Sources: Farming share of total employment for Metropolitan, Non-Metropolitan, and United States: Bureau of Economic Analysis; Total Full-Time and Part-Time Employment CA25 (1969–2000) CA25N (2001–2015). Farming as primary job as a share of total employment: Bureau of Labor Statistics, Table cpsaat01 (2017). Figure 3 View largeDownload slide Percentage of total jobs in farming: 1969–2015 Data Sources: Farming share of total employment for Metropolitan, Non-Metropolitan, and United States: Bureau of Economic Analysis; Total Full-Time and Part-Time Employment CA25 (1969–2000) CA25N (2001–2015). Farming as primary job as a share of total employment: Bureau of Labor Statistics, Table cpsaat01 (2017). The over reliance on export-based industries (such as mining, manufacturing, and large-scale agriculture) has led to lower growth, persistent unemployment, and a highly-segmented labor market that is not adaptable to change (Weiler 2001; Kilkenny and Partridge 2009). Increasing global trade with low-wage countries has also reduced and often eliminated the traditional wage and land cost advantages for rural manufacturers. Even when industries remain, many do so with much smaller workforces. While many rural economic development leaders and politicians believe the solution is to recruit another large employer, rural areas are increasingly at a disadvantage in attracting large firms because of factors such as limited workforce, low education levels, insufficient infrastructure, or high transport costs (Partridge and Olfert 2011). In addition, rural areas with large natural resource endowments (coal, oil, or natural gas) have long experienced boom and bust cycles. New technologies such as hydraulic fracturing made possible the extraction of natural gas and oil that were previously difficult to access. The resulting post-2007 boom affected traditional oil and gas rural areas in Texas as well as new areas such as North Dakota, and areas covered by the Marcellus and Utica shale formation in the Eastern United States, especially Ohio, Pennsylvania, and West Virginia. However, as natural gas (and oil) prices have declined so have exploration and employment. At the same time, the coal industry experienced large-scale layoffs, in small part due to stricter regulations, but mainly due to rapid productivity growth and more recently due to competition from cheaper natural gas, and this has especially affected eastern coal-mining communities. Combined, these resource economic busts helped drive the support for President Trump in many resource-dependent communities. Natural gas exploration and jobs may rebound, as natural gas replaces coal as a cheaper and more environmentally-friendly source for electricity generation, though promises of vast numbers of energy jobs are generally unrealistic (Kelsey, Partridge, and White 2016). Coal mining employment is unlikely to rebound, even if plans for new environmental regulations are abandoned (Betz et al. 2015). Dating back to the end of World War II, coal mining employment has declined as the industry implemented labor-saving technologies and this is exacerbated by declining foreign markets for U.S. coal (Betz et al. 2015). From 1989 to 2012, data from the Mine Safety and Health Administration shows that as coal output held relatively steady in Kentucky and West Virginia, coal-related employment fell from 79,000 to 41,000. Even if mines do reopen, they are likely to be more automated and the industry has already announced it will pay lower wages.9 Indeed, research suggests that scaling back environmental protections is the opposite of what coal mining regions need for long-term growth (Fan, Goetz, and Liang 2016).10 With the coal industry in decline, it will be nearly impossible to attract the human capital and the other firms needed to replace coal if quality-of-life is degraded in these remote areas (Betz et al. 2015). For some rural areas with natural amenities, there is the potential for future amenity-led growth. This may include an influx of new population or tourists and vacation homeowners seeking better climates and natural surroundings (Wehrwein and Johnson 1943; Deller et al. 2001; Carruthers and Vias 2005; McGranahan 2008; Wu and Gopinath 2008; Partridge 2010; Fan, Goetz, and Liang 2016). Yet, since not all rural areas are endowed with high levels of natural amenities or a favorable climate, this is not a solution for all communities. Even if natural amenities exist, if communities are impacted by energy development and have suffered from a degraded environment, efforts by rural communities to benefit from such growth will be hampered. Despite the challenges facing the most distressed rural communities, there is growing evidence that having more local entrepreneurs and self-employed can help rural and remote communities sustain growth and prosper (Stephens and Partridge 2011; Rupasingha and Goetz 2013; Stephens, Partridge, and Faggian 2013), as well as mitigate trade shocks (Liang and Goetz 2016). Stephens, Partridge, and Faggian (2013) find that even in the most distressed rural areas where there may be a large number of necessity entrepreneurs (Low, Henderson, and Weiler 2005; Acs et al. 2009) or those who start businesses because of a lack of other opportunities, there still are spillovers into the greater region including more wage and salary jobs (e.g., people working for someone other than themselves). Local entrepreneurs further magnify the positive economic impacts of their business activity on the community by favoring procurement from other local businesses (Fleming and Goetz 2011). Local entrepreneurs can be successfully “home-grown,” as rural entrepreneurship often requires relatively low educational levels, while the businesses purchase inputs locally, fulfill local demand and benefit from local assets, are perceptive to the regional milieu, and further build the existing entrepreneurial culture (Akgün et al. 2010; Stephens, Partridge, and Faggian 2013; Fritsch and Wyrwich 2014; Li et al. 2016). Indeed, past research shows that a key reason for a natural resource curse in coal country is that it lays the seeds of decline during a bust because coal mining employment crowds out self-employment and small business development that would cushion the blow (Betz et al. 2015). Yet, there are large gaps in understanding how to foster entrepreneurship and self-employment for rural growth. Poverty and Social Issues in Rural America Increasing economic opportunities is an important concern in rural areas, especially where globalization and automation have caused profound structural change. Rising productivity has displaced blue-collar workers and, in general, while there have been economic winners from trade and automation, many of the economic losers have been in rural areas. Although poverty is an important urban problem in terms of numbers of poor, official nonmetropolitan poverty rates—and especially for children—have been consistently higher than metropolitan poverty rates dating back to 1959, when poverty rates were first measured.11 In particular, rural child poverty rates are especially high relative to urban child poverty rates and over 85% of counties classified as “poor” are rural (USDA 2017).12 Deteriorating social and economic conditions are now reducing life expectancy in many rural communities, partly as a result of rising rates of drug use and suicide; recent studies show that where individuals live influences their likely cause of death (Moy et al. 2017). Despite being the major source of the nation’s food supply, rural areas also face issues related to food insecurity as explained in Gundersen and Ziliak (2018). Clearly, geography and space play roles in deteriorating socioeconomic outcomes, but the precise causal mechanisms and effects on human well-being remain poorly understood (Goetz et al. 2017). While both local poverty and inequality have long been studied (e.g., Levernier, Partridge, and Rickman 1998; Rupasingha and Goetz 2007), they are receiving heightened attention with the work of Chetty et al. (2014) and Picketty (2013). Using big data in the form of individual tax records, Chetty et al. (2014) show that geography affects economic mobility. Interestingly, rural areas perform better at promoting youth up the income ladder than urban areas, other factors held constant. Yet it is not clear whether this occurs because rural youth leave in pursuit of better urban opportunities, or whether rural areas inherently have advantages, perhaps because of better and smaller schools that, among other benefits, provide more leadership opportunities. Select studies are starting to examine this question. Recent research that reevaluates the Moving to Opportunities project, in which families were relocated from low-income housing to wealthier neighborhoods (Chetty, Hendren, and Katz 2016), should also be reexamined for its potential research, practical, and policy implications for rural areas. Of particular importance for research are the questions of (a) how poverty and inequality evolve over time and space, and the rural-urban continuum in particular, and what forces are driving the changes; (b) what role government can play (if any) in mitigating rising poverty and inequality through programs and policy; (c) the effects of globalization and automation, including robots and 3-D printing technology on the changing relationship between poverty and inequality; and (d) the interrelationship between inequality and poverty, or economic growth. Poverty and inequality occur to different degrees among U.S. counties. Since 1989, the number of counties with both high poverty rates and high inequality levels has increased. This has been especially true in counties in small and midsize metropolitan areas, where nearly one-half of all counties now have high rates of both poverty and inequality, up from on 22% in 1989 (Jarosz and Mather 2014). Additionally, 44% (up from 35%) of the most rural counties have high rates of both poverty and inequality, compared with only 21% (up from 11%) of the most urban counties (Jarosz and Mather 2014). Clearly, the forces causing the expanding coexistence of poor people and uneven income distributions are weakest in urban areas, strongest in the next larger category of counties, and somewhat milder in the most rural counties, but still prominent. Understanding why these changes are happening in the different sized communities and across the rural-urban continuum is an important first step towards addressing them. Although some level of inequality is beneficial for growth because it provides incentives for people to work harder, innovate, acquire training and human capital, invest, and promote entrepreneurship, evidence is mounting that the United States has now reached levels that may be suppressing aggregate economic growth (Partridge and Weinstein 2013).13 At the subnational level, early research tended to find a positive linkage with growth, especially in the long-run (Partridge 1997; 2005). However, more recent evidence suggests a negative linkage, notably in the short-run (Atems 2013). This relationship has not been investigated rigorously at the subnational or regional levels, or for rural versus urban areas. There is also evidence, at the level of nations, that higher rates of inequality are associated with shorter periods of economic growth (Berg and Ostrey 2011). This could have important implications for the economic future of the entire United States and its rural communities. Empirical Advances and Data Understanding rural growth and assessing the impact of rural policy requires establishing causal effects (see Athey and Imbens 2017). Recent decades have seen a revolution in empirical identification in order to establish causation, and these methods are increasingly being applied to address regional and rural economic issues.14 For example, empirical advances have been made using natural experiments, propensity-score matching and synthetic controls, instrumental variables, regression discontinuity, quantile regressions, and spatial modeling, including locally-or geographically-weighted regressions. We focus here on advances related to estimating multipliers as they are especially useful in evaluating economic development policy and in informing decision-making to support rural economic growth. Multipliers Estimating the economic impacts of a new business in terms of total net job or income creation is essential in assessing whether community support should be provided to recruitment efforts. For a distressed rural area, providing tax incentives to recruit a new employer may be tempting. It is especially important that with limited resources, the benefits must exceed the costs. The impacts of new development include the new jobs from that firm, the new jobs supported by the activity and salaries of that firm and other induced effects that create employment as well as offsetting displacement or crowding-out effects. The same types of analysis can also help a community evaluate the effect of a major plant closure or shock to an industry. In both cases, knowing the net multiplier effects would be valuable. A local manufacturing multiplier of 2.5, for example, means that 100 new manufacturing jobs would be expected to create a total of 250 jobs, of which 150 are indirectly created through spillovers; a multiplier greater than one indicates positive spillovers from an economic shock, while a multiplier of less than one indicates that, while new jobs are being created, they are displacing other jobs, offsetting the positive effects. Popular software to estimate these economic impacts and multipliers include private vendors IMPLAN and REMI among others, using research that dates to the 1950s. These tools are most effective in estimating the positive economic spillovers, but much less so in estimating displacement effects (see Drucker 2015 for a critical discussion of impact assessments from commercial software.) For instance, an incentivized office park typically just rearranges workers from existing local office parks to the new office park, leading to no net increase in jobs/income even if the new office park is full. Similarly, an incentivized restaurant may simply displace an existing restaurant, creating no net jobs or wealth. A new rural manufacturer could bid up wages and land prices, reducing the competitiveness of existing local firms. Another issue is the “false precision” from commercial software estimates, in which the software produces exact employment or income estimates by detailed sector, ignoring any indication of the potentially large standard errors in those estimates due to the imposition of a strict structure on the local economy. Economists are well-positioned to improve on the multiplier estimates for local economies using econometrically estimated multipliers, providing confidence intervals to the estimates, and avoiding the imposition of a strict structure on the local economy as is done by commercial software; in other words, allowing the data to speak. The recent surge of interest began with Moretti’s (2010) short multiplier paper, though the precise empirical tools were described in Bartik’s (1991) seminal contribution. While Moretti indirectly estimates the multiplier, estimating employment elasticities using the industry mix shift share term as an instrument for the change in employment from a key sector, such as manufacturing.15 However, Moretti’s approach suffers from several disadvantages: (a) it is nonlinear and multipliers are linear by definition; (b) to convert into a multiplier, the regression coefficient needs to be scaled by the relative size of the industry in question, which can be greatly affected by base year issues, especially when there is rapid growth; (c) it indirectly estimates the multiplier using the industry mix shift share term as an instrument for the change in employment from a key sector, such as manufacturing, and using instrumental variable estimation (IV) when more efficient/direct methods are available such as ordinary least squares (OLS) with best, linear, unbiased, estimators (BLUE).16 Bartik and others use a more direct method, replacing (for example) the manufacturing term with its industry mix shift share term and estimating the model with OLS. In this case, the regression coefficient on the manufacturing shift-share term is the multiplier and no further adjustments are needed—see Tsvetkova and Partridge (2016) for an empirical example of local energy-sector multipliers. Thus, as long as a researcher has detailed local data on industry shares (most likely a county or metro area), it is possible to identify an accurate multiplier that is not constrained by the strong assumptions used in commercial modeling software and produces a confidence interval that more realistically conveys the uncertainty in the estimates. Rural Data Good data are needed to evaluate policies and hold politicians, policymakers, and the private sector accountable for performance. The federal government has an important role in producing subnational statistics (and all data in general). These data are also essential for providing the private sector with key information for marketing and firm location and for academic research to identify better ways to achieve prosperity and improve quality-of-life (Partridge, Goetz, and Kilkenny 2013). The costs of federal statistical efforts are a small part of the overall federal budget, suggesting that a benefit-cost analysis would be astronomically in their favor.17 Despite the urgent need, there are long-running congressional efforts to reduce and eliminate federal statistical programs. One reason is that surveys such as the American Community Survey are perceived by some as asking sensitive questions about income or demographics (e.g., Rampell’s 2012’s discussion). However, the loss of federal statistical programs would have serious consequences and it is unlikely that the private sector can or would fill the gap. Private collection of data would suffer from the many problems associated with pure public goods and would be underprovided by the market. Such data are also likely to be expensive in order to allow vendors to recover the high costs. These expenses would reduce transparency in the government and increase costs for business and research. In addition, private sector data collection would inherently be less precise because federal agencies can incorporate administrative records such as tax data in their estimates. It would also likely be more ad hoc and less comparable over time. This would result in important business decisions being based on poor quality data, hurting the economy. Additionally, an important issue for rural America and rural researchers would be that private vendors would have little incentive to fill in the gap for rural data. Because the market for a tiny rural county’s socioeconomic data is so small, rural areas would be especially hurt by such efforts. At the same time that government data collection is in jeopardy, the last decade has seen a surge in potentially new data sources such as Google Analytics, Facebook data, as well as private companies’ customer and sales data. Without considering the ethical and privacy concerns of academics using such data for research, there are clearly some interesting questions that can be analyzed with Big Data that could not be done with conventional data. However, there are reasons to believe the Big Data fad is overhyped. Beyond selection issues, big Data shares with other private data the same problems of reliability and comparability over time, as well as problems for replication, a key standard for scientific research. And due to the fact that these data are collected by private vendors, for rural researchers, the availability of these data may be sparse. Rural Policy and Its Evaluation: Current and Future Research Directions Over the next few years, the federal government is likely to consider many policies that affect rural economies. Given the important differences across counties and the heterogeneity in what is rural, it is clear that a one-size-fits-all rural policy will not work. Federal labor, fiscal, trade, and interest rate policies all have disparate spatial impacts that depend, among other factors, on the export dependence and capital intensity of local industries. Most rural policy continues to be equated with farm policy, yet nutrition assistance programs make up 71% of the $151 billion in the USDA budget and are overwhelmingly paid out to urban consumers. Of the remaining USDA outlays, 16% are for farm and commodity programs, and 7% are for conservation and forestry (compared to agriculture accounting 5% of rural employment). The remainder (6%) is shared across rural development, research, food safety, marketing and regulation, and management functions. The policy-related word network analysis of Reimer et al. (2016) confirms that rural development issues are overwhelmingly framed as agricultural issues, even though a relatively small number of counties remain dependent on agriculture. Even though farm interests tend to be split along commodity lines, farmers overall have been able to articulate their policy choices effectively, largely due to the power of farm-state Senators (Reimer et al. 2016). It remains to be seen whether President Trump’s Administration will change rural policy to focus more on nonfarm rural issues, though early indications are that agriculture will remain the key rural focus. Policy researchers and political scientists concerned about rural economies should find this phenomenon to be an important area of inquiry (O’Brien and Ahearn 2016). Rural development (RD) is funded by the federal government through USDA as well as through the Small Business Administration (SBA), Economic Development Agency (EDA), and Appalachian Regional Commission (ARC), and, overall, it is a very small portion of federal spending. There is some evidence that the Trump Administration may not support current rural development (RD) programs, as indicated by initial proposals to strike key elements of the USDA RD infrastructure and USDA RD and SBA business financing activities, as well as to eliminate agencies such as the EDA and ARC (OPM 2017). On the one hand, it has long been recognized that consolidating federal economic development activities into one agency may be better than having them spread out over dozens of different departments and agencies that could lead to redundancies. Yet the early Trump Administration proposals suggest a relative neglect of rural areas in terms of producing sustainable economic development and aid for the most distressed communities. Hence, there is a role for researchers to provide sound evidence of rural policy effectiveness in order to inform the current administration as well as future policymakers. While place-based policies have a mixed track record—often due to poor implementation—there are examples of place-based policies that have potential, including the ARC, an expanded Delta Regional Authority, and other newer regional initiatives such as the Stronger Economies Together (SET) program hosted by the Southern Rural Development Center that do not pursue one-size-fits-all projects. In particular, given its long history, the ARC has received much attention; it provides bridge loans and seed grants for infrastructure and supports other programs such as workforce training. Yet given its limited resources, the ARC’s main role is as a broker that can foster regional collaborations of businesses, communities, nonprofits, local development districts, and various state and federal agencies while providing small funding matches to ensure projects can be implemented. The ARC also includes multi-county regional development districts (referred to as local development districts), which are composed of functional economic regions that can help facilitate cooperation and collaboration between neighboring communities. This may be especially important for rural communities that lack the capacity to tackle programs on their own. Since functional economic areas are predominantly centered on urban areas, this model also allows rural communities to work with their urban neighbors and benefit from urban-led growth through commuting, for example. The ARC, with a federal allocation of $90 million in FY 2015 and $146 in FY 2016 (with some temporary expenditures), is limited in what it can do for its over 25 million residents spread across 13 states. Despite the modest resources and the persistent underfunding of the agency, there is evidence that the ARC has been effective. Isserman and Rephann (1995) is one of many supportive studies finding that ARC counties had significantly better economic outcomes than observationally similar counties. A more comprehensive and longer-term recent study using propensity scoring and careful matching finds that per capita income in ARC-funded counties grew at a 5.5% higher rate than it did in non-funded control counties (Sayago-Gomez et al. 2017). Given the evidence that the ARC’s model may be one way to help distressed, rural regions, future research should evaluate the impact of similar federal-state regional development organizations in other rural areas. Recent research also suggests that outreach or other technical assistance programs are considerably more cost-effective than direct farm subsidy grants in keeping farmers on the farm (Goetz and Davlsheridze 2016). In other words, helping farmers through technical assistance or human capital training is more cost effective than is providing cash transfers. This is a critical research area that deserves more attention, especially given the importance of traditional federal farm programs, and the evidence that they may, in fact, accelerate the consolidation of farms and rural depopulation (Goetz and Debertin 2001). Future research could help increase our understanding of the interrelated effects of farm programs on farm size distribution, farm productivity, and the aging of farmers. Here the experience of New Zealand may offer fruitful guidance for structuring research questions (Johnston and Frengley 1991; Evans et al. 1996). A significant research void also exists in evaluating the effectiveness of existing USDA Rural Development programs and other rural development programs, including different types of policy designs and mechanisms. While important and carefully executed studies of individual programs exist, they are limited in number and they are piecemeal, focusing on only one program at a time. For example, research on the broadband loan programs (e.g., Whitacre Gallardo, and Strover 2014) shows somewhat mixed results, but finds that, on balance, it has positive effects. To understand the impact of these efforts and any synergies between programs, more holistic and comprehensive evaluations over time, ideally with panel data, of all rural development programs are needed. It would be especially useful to consider natural experiments when feasible, that is, where individuals or geographic areas are randomly assigned to treatments. Carefully designed matched comparisons or new approaches using synthetic control groups are other options, as are new opportunities to merge program data with secondary public records such as the National Employment Times Series data base or NETS. This kind of research is being encouraged by the Commission on Evidence-Based Policymaking. A critical research need lies in examining the benefits and the costs both of individual rural development policies and grant and loan programs, as well as their interactive effects. Related to this are questions of investing funds in people versus places. While economists have long held that investing in people is preferable to investing in places, as place-based investments tend to benefit the owners of fixed assets (e.g., in the form of higher housing prices), there is evidence that carefully-selected and targeted place-based investments can make a positive difference (Busso, Gregory, and Kline 2013), even though these claims are controversial (e.g., Hanson 2009; Hanson and Rohlin 2011). A strong interest has recently emerged in estimating the benefits and cost of local economic development programs and policies (e.g., Neumark and Simpson 2015). While much of this research has focused on more urban areas, extensions and applications to rural areas will likely have high pay-offs. One key lesson from this research is that picking winners and losers by subsidizing individual firms or industries is ineffective, if not counterproductive. Instead, it is important to improve overall economic conditions and workforce quality that benefit all businesses, so that the most competitive ones emerge. Further, economic spillovers in the form of externalities are also essential. For example, if a downtown business attracts consumers who also patronize other businesses on the same Main Street, a positive externality is created. In cases where barriers to entry prevent such public externalities, federal investments may be highly effective. In practice, some rural and remote communities likely lack the critical mass needed to generate the externalities that make such federal investments cost-effective. In that case, it is difficult to make the economic case for intervention. However, researchers can help identify such communities, as well as the thresholds needed to establish conditions for not just survival, but growth, though politically allocating funds on this basis will be difficult. Such research could use criteria such as a minimum set of businesses and services available within a given radius, while also considering population density and distance to major metropolitan areas—one example of which is provided by Stabler and Olfert (2002). Examining why some communities are resilient and able to bounce back from major shocks while others are not will also provide useful future policy insights (Han and Goetz 2015). This research may be especially valuable if it considers the historical local industry mix as well as location of the county on the rural-urban continuum. Closely related to this is the need to better understand the tradeoffs of having multiple layers of local rural (and urban) governments: these are gains from tailoring local policies versus redundancies, fewer spillovers across small jurisdictions, and the loss of economies of scale in service/infrastructure provision. Additional research is needed to understand what policies will support economic growth in rural areas. Evidence suggests that attracting large firms is typically unsuccessful, especially for rural areas. However, all areas have the potential to build a locally-diverse business environment based on higher-quality human capital. In addition, the type of parents who demand high-quality local schools are the types of workers a successful community needs. As mentioned previously, there is also evidence that even distressed areas can benefit from home-grown businesses. More research is needed on what types of entrepreneurs are effective in promoting growth and on what types of infrastructure and technical-assistance support (including business planning and other forms of education) are most effective in helping develop and grow local businesses. Economists can also assess the costs and benefits of improving rural education. For rural areas that have historically depended on resource extraction industries, environmental degradation is a threat to future growth. People are unwilling to remain in or migrate to areas that have been degraded and this impacts future economic prosperity. Regional economists in collaboration with agricultural economists and environmental economists have the expertise to provide evidence of the economic benefits of environmental protection and to evaluate ways to help communities transition to new a new economic paradigm. Our paper also identifies other areas of research that support this transition. Overall, while U.S. policymakers often eagerly support various programs and policies they believe will help their rural constituents, they generally have less patience for formal and systematic evaluation of such programs once implemented. It takes time for the full effects of programs and policies to play out, and rigorous and impartial assessments require additional resources. Above all, policymakers with a genuine interest in understanding the effects of their policies, and who want to make the best possible use of taxpayer dollars should be concerned about any efforts to scale back public data collection efforts. Additionally, a more balanced approach to RD in many parts of the country would place agriculture within the much larger rural nonfarm economy, rather than as the primary or sole economic sector. Researchers in agricultural economics and related professions are well-positioned to provide data-driven research to evaluate the effectiveness of new rural policies and support new solutions to help rural areas. Finally, while rigorous rural policy research and evaluation are critical, it is also essential that results be communicated to policymakers and their staff in formats that are easy to digest. This includes policy briefs, fact sheets, and on-demand webinars that review complex economic topics. One potential model for this is the National Agricultural and Rural Development Policy (NARDEP) Center, in which a research center is adequately funded to be a central point of research and policy outreach on critical rural policy issues and evaluation. While there are many sources of information on farm policy and urban policy, there is little about broad-based rural policies. Overall, colleges of agriculture have the experience in extension and outreach to make a difference in supporting policymaking related to rural areas. Footnotes 1 We use the terms metropolitan/urban and nonmetropolitan/rural interchangeably. 2 There are 2,494 nonmetro counties using the 1973 definition and 1,975 nonmetro counties under the 2013 definition. 3 Using International Monetary Fund data, we use the rapid rise in commodity prices beginning in 2003 and subsequent decline in 2014 to date the commodity price cycle (Index Chart 2000–2017 available at: http://www.imf.org/external/np/res/commod/index.aspx). 4 Rural areas have been swinging towards the GOP since the 1950s when Democratic Party support was about equal in cities and the countryside (Badger and Bui 2016), and this trend continued in the twenty-first century. Scala and Johnson (2017) document the decline in Democratic Party support over the 2000–2016 period in rural areas, with the decline being the most stark in the most remote rural regions. However, support for Democratic Presidential candidates literally collapsed between 2012 and 2016, with the average decline ranging from 7 to 10 percentage points depending on the type of nonmetropolitan county. 5 See https://www.statcan.gc.ca/pub/92-195-x/2011001/geo/cma-rmr/def-eng.htm for more details. 6 See https://en.wikipedia.org/wiki/Urban_area_(France) for more details. 7 The precise UICs are defined at: https://www.ers.usda.gov/data-products/urban-influence-codes/documentation/. 8 In discussions over the last 20 years, a common theme in rural development is that improvements in information communication technologies have made (or will make) it easier for people to live and work in rural areas because they are more connected to larger markets. Yet, such a view misses that the same information technologies are also making it easier for workers to telecommute and not physically commute to work every day, thereby improving their quality of life and generally reducing urban congestion—that is, making urban areas more livable. 9 Available at: http://www.wsj.com/articles/arch-coal-files-for-bankruptcy-1452500976. 10 In 1929, 1948, and 1970 (when the Clean Air Act of 1970 was passed), U.S. Bureau of Economic Analysis (BEA) data suggests that U.S. coal employment totaled 622,000, 533,000, and 146,000 people, respectively, illustrating that productivity growth was greatly reducing employment well before pollution regulations became binding, and showing the tremendous role of productivity growth prior to 1970. Illustrating the important role of coal booms relative to regulation, despite the Clean Air Act of 1970, the late 1970s energy shortage lifted coal mining employment to 254,000 workers. When the Clean Air Act of 1990 was passed, coal mining employment had fallen back to 148,000. Since then, U.S. Bureau of Labor Statistics data shows that coal employment has continued to decline to 50,500 workers in 2016 (between January 2017 and October 2017, coal mining employment increased by 2,000 workers). 11 For more details, see https://www.ers.usda.gov/topics/rural-economy-population/rural-poverty-well-being/poverty-overview/ (accessed November 22, 2017). 12 In 2013, just under 63% of U.S. counties were nonmetropolitan (see https://www.ers.usda.gov/topics/rural-economy-population/rural-classifications/). 13 There is a long literature discussing the tradeoff between equity and efficiency. Some examples include Okun (1975) and Welch (1999) with summaries in Barro (2000) and Partridge (1997; 2005). 14 A recent overview is contained in Patrick et al. 2017. 15 Rather than a linear term such as percentage change in total employment as the dependent variable, Moretti uses the log of employment as the dependent variable. For large employment changes over a decade, as in Moretti’s case, this change could matter. 16 The manufacturing industry mix term %ΔIMMAN equals ∑iShir0%ΔNEMPit in which Sh is area r’s initial period employment share of industry of i and %ΔNEMPit the percent change in national employment in industry i. Summing over all industries i in the manufacturing sector produces the predicted manufacturing growth rate in the area if all of the industries in the manufacturing sector are growing at the national growth rate. Unless there is some labor supply response to the initial industry shares (or one that is not accounted for by a demographic control variable), because industry national growth rates are used, the shift-share term is assumed to be exogenous. If one needs a multiplier for another sector, then the shift-share should be calculated for that industry grouping. 17 The total costs of the three main federal statistical agencies—Bureau of Economic Analysis, Census Bureau, and Bureau of Labor Statistics—is approximately $2.6 billion in fiscal year 2017, or about 0.06% out of a total federal budget of about $4.1 trillion. References Acs Z.J., Braunerhjelm P., Audretsch D., Carlsson B. 2009. The Knowledge Spillover Theory of Entrepreneurship. Small Business Economics  32 ( 1): 15– 30. Google Scholar CrossRef Search ADS   Akgün A.A., Nijkamp P., Baycan T., Brons M. 2010. Embeddedness of Entrepreneurs in Rural Areas: A Comparative Rough Set Data Analysis. Tijdschrift Voor Economische en Sociale Geografie  101: 538– 53. Google Scholar CrossRef Search ADS   Angrist J. D., Pischke J. 2009. Mostly Harmless Economics, An Empiricist’s Companion. Princeton, NJ: Princeton University Press. Atems B. 2013. A Note on the Differential Regional Effects of Income Inequality: Empirical Evidence Using U.S. County Level Data. Journal of Regional Science  53: 656– 71. Google Scholar CrossRef Search ADS   Athey S., Imbens G.W. 2017. The State of Applied Econometrics: Causality and Policy Evaluation. Journal of Economic Perspectives  31 ( 2): 3– 32. http://dx.doi.org/10.1257/jep.31.2.3 Google Scholar CrossRef Search ADS   Autor D.H., Dorn D., Hanson G.H. 2013. The China Syndrome: Local Labor Market Effects of Import Competition in the United States. American Economic Review  103: 2121– 68. Google Scholar CrossRef Search ADS   Badger E., Bui Q. 2017. Why Republicans Don’t Even Try to Win Cities Anymore. New York Times. November 2. Available at: https://www.nytimes.com/2016/11/03/upshot/why-republicans-dont-even-try-to-win-cities-anymore.html?_r=0. Barca F., McCann P., Rodríquez-Pose A. 2012. The Case for Regional Development Intervention: Place–Based versus Place–Neutral Approaches. Journal of Regional Science  52: 134– 52. Google Scholar CrossRef Search ADS   Barro R.J. 2000. Inequality and Growth in a Panel of Countries. Journal of Economic Growth  5: 5– 32. Google Scholar CrossRef Search ADS   Bartik T.J. 1991. Who Benefits from State and Local Economic Development Policies? Kalamazoo, MI: W. E. Upjohn Institute for Employment Research. Google Scholar CrossRef Search ADS   Berg A.G., Ostry J.D. 2011. Inequality and Unsustainable Growth: Two Sides of the Same Coin? IMF Staff Discussion Note, SDN/11/08, April 8, 20. Betz M., Farren M., Lobao L., Partridge M.D. 2015. Coal Mining, Economic Development, and the Natural Resources Curse. Energy Economics  50: 105– 16. Google Scholar CrossRef Search ADS   Busso M., Gregory J., Kline P. 2013. Assessing the Incidence and Efficiency of a Prominent Place Based Policy. The American Economic Review  103: 897– 947. Google Scholar CrossRef Search ADS   Carruthers J., Vias A.C. 2005. Urban, Suburban, and Exurban Sprawl in the Rocky Mountain West: Evidence from Regional Adjustment Models. Journal of Regional Science  45: 21– 48. Google Scholar CrossRef Search ADS   Chetty R., Hendren N., Katz L.F. 2016. The Effects of Exposure to Better Neighborhoods on Children: New Evidence from the Moving to Opportunity Experiment. The American Economic Review  106: 855– 902. Google Scholar CrossRef Search ADS   Chetty R., Hendren N., Kline P., Saez E. 2014. Where is the Land of Opportunity? The Geography of Intergenerational Mobility in the United States. The Quarterly Journal of Economics  129: 1553– 623. Google Scholar CrossRef Search ADS   Deller S.C., Tsai T-H., Marcouiller D.W., English D.B.K. 2001. The Role of Amenities and Quality of Life in Rural Economic Growth. American Journal of Agricultural Economics  83: 352– 65. Google Scholar CrossRef Search ADS   Drucker J. 2015. Economic Impact Analysis amid Rapid Change: Challenges, Strategies, and Examples from Defense Communities. Journal of Planning Education and Research  35: 401– 18. Google Scholar CrossRef Search ADS   Evans L., Grimes A., Wilkinson B., Teece D. 1996. Economic Reform in New Zealand 1984-95: The Pursuit of Efficiency. Journal of Economic Literature  34: 1856– 1902. Fan Q., Goetz S.J., Liang J. 2016. The Interactive Effects of Human Capital and Quality of Life on Economic Growth. Applied Economics  48: 5186– 200. Google Scholar CrossRef Search ADS   Fleming D.A., Goetz S.J. 2011. Does Local Firm Ownership Matter? Economic Development Quarterly  53: 277– 81. Google Scholar CrossRef Search ADS   Fritsch M., Wyrwich M. 2014. The Long Persistence of Regional Levels of Entrepreneurship: Germany, 1925–2005. Regional Studies  48: 955– 73. Google Scholar CrossRef Search ADS   Goetz S.J., Davlasheridze M. 2016. State-Level Cooperative Extension Spending and Farmer Exits. Applied Economic Perspectives and Policy  39( 1): 65– 86. Goetz S.J., Davlasheridze M., Han Y. 2015. County-Level Determinants of Mental Health, 2002–2008.  Social Indicators Research  24: 657– 70. Google Scholar CrossRef Search ADS   Goetz S.J., Davlasheridze M., Han Y., Fleming D. 2017. Explaining the 2016 Vote for President Trump in U.S. Counties. In Review. Goetz S.J., Debertin D.L. 2001. Why Farmers Quit: A County-Level Analysis. American Journal of Agricultural Economics  83 ( 4): 1010– 23. http://dx.doi.org/10.1111/0002-9092.00226 Google Scholar CrossRef Search ADS   Gundersen C., Ziliak J. 2018. Food Insecurity Research in the United States: Where We Have Been and where We Need to Go. Applied Economic Perspectives and Policy  40 ( 1): 119– 35. Han Y., Goetz S.J. 2017. Overlapping Labor Market Areas Based on Link Communities. Papers in Regional Science. DOI: 10.1111/pirs.12326 Han Y., Goetz S.J. 2015. The Economic Resilience of US Counties During the Great Recession. Review of Regional Studies  45: 131– 149. Hanson A. 2009. Local Employment, Poverty, and Property Value Effects of Geographically-Targeted Tax Incentives: An Instrumental Variables Approach. Regional Science and Urban Economics  39: 721– 31. Google Scholar CrossRef Search ADS   Hanson A., Rohlin S. 2011. Do Location-Based Tax Incentives Attract New Business Establishments? Journal of Regional Science  51 ( 3): 427– 49. Google Scholar CrossRef Search ADS   Isserman A. 2005. In the National Interest: Defining Rural Correctly for Research and Policy. International Regional Science Review  28 ( 4): 465– 99. http://dx.doi.org/10.1177/0160017605279000 Google Scholar CrossRef Search ADS   Isserman A.M., Rephann T. 1995. The Economic Effects of the Appalachian Regional Commission. Journal of the American Planning Association  61: 345– 64. Google Scholar CrossRef Search ADS   Johnston W.E., Frengley G.A. 1991. The Deregulation of New Zealand Agriculture: Market Intervention (1964–84) and Free Market Readjustment (1984–90). Western Journal of Agricultural Economics  16: 132– 43. Kelsey T., Partridge M.D., White N. 2016. Unconventional Gas and Oil Development in the United States: Economic Experience and Policy Issues. Applied Economic Perspectives and Policy  38 ( 2): 191– 214. http://dx.doi.org/10.1093/aepp/ppw005 Google Scholar CrossRef Search ADS   Kilkenny M., Partridge M.D. 2009. Export Sectors and Rural Development. American Journal of Agricultural Economics  91 ( 4): 910– 29. http://dx.doi.org/10.1111/j.1467-8276.2009.01320.x Google Scholar CrossRef Search ADS   Levernier W., Partridge M.D., Rickman D.S. 1998. Differences in Metropolitan and Nonmetropolitan U.S. Family Income Inequality: A Cross-Country Comparison. Journal of Urban Economics  44: 272– 90. Google Scholar CrossRef Search ADS   Li M., Goetz S.J., Partridge M., Fleming D.A. 2016. Location Determinants of High-Growth Firms. Entrepreneurship and Regional Development  8: 97– 125. Google Scholar CrossRef Search ADS   Liang J., Goetz S.J. 2016. Self-Employment and Trade Shock Mitigation. Small Business Economics  46: 45– 56. Google Scholar CrossRef Search ADS   Low S., Henderson J., Weiler S. 2005. Gauging a Region’s Entrepreneurial Potential. Economic Review.  Third Quarter 61–89. Kansas City, MO: Federal Reserve Bank of Kansas City. McGranahan D. 2008. Landscape Influence on Recent Rural Migration in the U.S. Landscape and Urban Planning  85: 228– 40. Google Scholar CrossRef Search ADS   Moretti E. 2010. Local Multipliers. American Economic Review  1100: 373– 77. Google Scholar CrossRef Search ADS   Moy E., Garcia M.C., Bastian B., Rossen L.M., Ingram D.D., Faul M., Massetti G.M., et al.   2017. Leading Causes of Death in Nonmetropolitan and Metropolitan Areas—United States, 1999–2014. MMWR. Surveillance Summaries  66: 1– 8. Google Scholar CrossRef Search ADS   National Academies of Sciences, Engineering, and Medicine. 2016. Rationalizing Rural Area Classifications for the Economic Research Service: A Workshop Summary.  Washington, DC: The National Academies Press. Available at: https://doi.org/10.17226/21843. Neumark D., Simpson H. 2015. Place-Based Policies. In Handbook of Regional and Urban Economics , eds. Duranton G., Henderson J.V., Strange W.C., 1197– 287. Amsterdam, Netherlands: Elsevier B.V. Okun A.M. 1975. Equality and Efficiency: The Big Tradeoff . Washington: Brookings Institution. O’Brien D.J., Ahearn M.C. 2016. Rural Voice and Rural Investments: The 2016 Election and the Future of Rural Policy. Choices , 4th Quarter. U.S. Office of Personal Management. 2017. America First: A Budget Blueprint to Make America Great Again.  Available at: https://www.whitehouse.gov/sites/whitehouse.gov/files/omb/budget/fy2018/2018_blueprint.pdf. Patrick C., Ross A., Stephens H. 2017. Designing Policies to Spur Economic Growth How Regional Scientists Can Contribute to Future Policy Development and Evaluation. In Regional Research Frontiers: The Next 50 Years , ed. Jackson R., Schaeffer P. pp 119– 133. New York, New York, USA: Springer. Google Scholar CrossRef Search ADS   Partridge M.D. 2010. The Dueling Models: NEG vs Amenity Migration in Explaining US Engines of Growth. Papers in Regional Science  89: 513– 36. Google Scholar CrossRef Search ADS   Partridge M.D., Weinstein A.L. 2013. Rising Inequality in an Era of Austerity: the Case of the US. European Planning Studies  1: 388– 410. Google Scholar CrossRef Search ADS   Partridge M.D., Rickman D.S., Ali K., Olfert M.R. 2008. Lost in Space: Population Growth in the American Hinterlands and Small Cities. Journal of Economic Geography  8: 727– 57. Google Scholar CrossRef Search ADS   Partridge M.D., Rickman D.S., Ali K., Olfert M.R. 2010. The Spatial Dynamics of Factor Price Differentials: Productivity or Consumer Amenity Driven? Regional Science and Urban Economics  40: 440– 52. Google Scholar CrossRef Search ADS   Partridge M.D., Olfert M.R. 2011. The Winners' Choice: Sustainable Economic Strategies for Successful 21st Century Regions. Applied Economic Perspectives and Policy  33: 143– 78. Google Scholar CrossRef Search ADS   Partridge M.D., Goetz S.J., Kilkenny M. 2013. Why We Need Federal Statistical Data for States and Counties. Choices  28. Available at:http://www.choicesmagazine.org/choices-magazine/submitted-articles/why-we-need-federal-statistical-data-for-states-and-counties. Partridge M.D., Rickman D.S., Olfert M.R., Tang Y. 2017. International Trade and Local Labor Markets: Do Foreign and Domestic Shocks Affect Regions Differently? Journal of Economic Geography  7: 375– 409. Piketty T. 2013. Capital in the Twenty-First Century.  Boston, Massachusetts: Belnap Press/Harvard. Reimer A., Han Y., Goetz S.J., Loveridge S., Albrecht D. 2016. Word Networks in US Rural Policy Discourse. Applied Economic Perspectives and Policy  38: 215– 38. Google Scholar CrossRef Search ADS   Rampell C. 2012. The Beginning of the End of the Census? New York Times. May 19, 2017. Available at:http://www.nytimes.com/2012/05/20/sunday-review/the-debate-over-the-american-community-survey.html. Rupasingha A., Goetz S.J. 2013. Self‐Employment and Local Economic Performance: Evidence from US Counties. Papers in Regional Science  92: 141– 61. Rupasingha A., Goetz S.J. 2007. Social and Political Forces as Determinants of Poverty: A Spatial Analysis. The Journal of Socio-Economics  36: 650– 71. Google Scholar CrossRef Search ADS   Sayago-Gomez J-T., Piras G., Jackson R., Lacombe D. 2017. Impact Evaluation of Investments in the Appalachian Region: A Reappraisal. International Regional Science Review  DOI: 10.1177/0160017617713822. Scala D., Johnson K.M. 2017. Political Polarization along the Rural–Urban Continuum? The Geography of the Presidential Vote 2000–2016. The Annals of Political and Social Science  672: 162– 84. Google Scholar CrossRef Search ADS   Stabler J.C., Olfert M.R. 2002. Saskatchewan’s Communities in the 21st Century: From Places to Regions.  Regina, SK: Canadian Plains Research Center, University of Regina. Stephens H., Partridge M. 2011. Do Entrepreneurs Enhance Economic Growth in Lagging Regions? Growth and Change  42: 431– 65. Google Scholar CrossRef Search ADS   Stephens H., Partridge M., Faggian A. 2013. Innovation Entrepreneurship and Economic Growth in Lagging Regions. Journal of Regional Science  53: 778– 812. Google Scholar CrossRef Search ADS   Tsvetkova A., Partridge M.D. 2016. Economics of Modern Energy Boomtowns: Do Oil and Gas Shocks Differ from Shocks in the Rest of the Economy? Energy Economics  59: 81– 95. Google Scholar CrossRef Search ADS   Tsvetkova A., Partridge M.D., Betz M. 2017. Entrepreneurial and Wage and Salary Employment Across the Rural-Urban Continuum. The ANNALS of the American Academy of Political and Social Science  672: 83– 102. Google Scholar CrossRef Search ADS   U.S. Department of Agriculture. 2016. FY 2017 Budget Summary, Office of Budget and Program Analysis, Washington, DC. Available at: http://www.obpa.usda.gov/budsum/fy17budsum.pdf. Weber B., Fannin M.J., Miller K., Goetz S.J. 2017. Intergenerational Mobility of Low-Income Youth in Metropolitan and Nonmetropolitan America: A Spatial Analysis. Working Paper. Wehrwein G., Johnson H.A. 1943. A Recreation Livelihood Area. Journal of Land and Public Utility Economics  19 ( 2): 193– 206. http://dx.doi.org/10.2307/3159233 Google Scholar CrossRef Search ADS   Weiler S. 2001. Unemployment in Regional Labor Markets: Using Structural Theories to Understand Local Jobless Rates in West Virginia. Industrial and Labor Relations Review  54: 573– 92. Google Scholar CrossRef Search ADS   Welch F. 1999. In Defense of Inequality. American Economic Review  89: 1– 17. Google Scholar CrossRef Search ADS   Whitacre B., Gallardo R., Strover S. 2014. Does Rural Broadband Impact Jobs and Income? Evidence from Spatial and First-Differenced Regressions. The Annals of Regional Science   53: 649– 70. Google Scholar CrossRef Search ADS   Wooldridge J.M. 2010. Econometric Analysis of Cross Section and Panel Data . Cambridge, MA: MIT Press. Wu J., Gopinath M. 2008. What Causes Spatial Variations in Economic Development in the United States? American Journal of Agricultural Economics  90: 392– 408. Google Scholar CrossRef Search ADS   Ziliak J.P. 2015. Income, Program Participation, Poverty, and Financial Vulnerability: Research and Data Needs. Journal of Economic and Social Measurement  40: 27– 68. Google Scholar CrossRef Search ADS   © The Author(s) 2018. Published by Oxford University Press on behalf of the Agricultural and Applied Economics Association. All rights reserved. For Permissions, please e-mail: journals.permissions@oup.com

Journal

Applied Economic Perspectives and PolicyOxford University Press

Published: Mar 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 12 million articles from more than
10,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Unlimited reading

Read as many articles as you need. Full articles with original layout, charts and figures. Read online, from anywhere.

Stay up to date

Keep up with your field with Personalized Recommendations and Follow Journals to get automatic updates.

Organize your research

It’s easy to organize your research with our built-in tools.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

Monthly Plan

  • Read unlimited articles
  • Personalized recommendations
  • No expiration
  • Print 20 pages per month
  • 20% off on PDF purchases
  • Organize your research
  • Get updates on your journals and topic searches

$49/month

Start Free Trial

14-day Free Trial

Best Deal — 39% off

Annual Plan

  • All the features of the Professional Plan, but for 39% off!
  • Billed annually
  • No expiration
  • For the normal price of 10 articles elsewhere, you get one full year of unlimited access to articles.

$588

$360/year

billed annually
Start Free Trial

14-day Free Trial