Is The Sky the Limit? Fair Executive Pay as Performance Rises

Is The Sky the Limit? Fair Executive Pay as Performance Rises Abstract Recent research documents the public discontent with high income inequality yet an important limitation in our understanding is why such high pay is problematic according to many Americans. This study juxtaposes two explanations for the discontent that many Americans have shown regarding high Chief Executive Officer (CEO) pay: a) that the American discontent with extremely high CEO pay stems mainly from a belief that CEOs are not contributing highly enough to merit such high pay and b) the American discontent with extremely high pay is due to such pay being objectionable in principle. Using data from a national survey experiment (N = 989) uniquely designed to test these propositions, this study evaluates the relationship between performance and fair CEO pay. Respondents show aversion to high CEO pay while also embracing the principle of proportionality of rewards and contributions. Findings suggest that discontent with extremely high pay may be reconciled with support for pay for performance: the slope of the pay and performance function may be at issue rather than a hard limit on pay. income inequality, distributive justice, executive compensation, attitudes, fair pay Many Americans are aware of the high levels of income inequality and are displeased with it (Bartels 2005; McCall 2003; McCall and Kenworthy 2009; McCall 2013; Norton and Ariely 2011; Osberg and Smeeding 2006; Page and Jacobs 2009). With the rise in income inequality at the top of the income distribution in the past decades and with the Great Recession of 2008, the media and the American public turned their attention to high earners. Some scholars argued that the Great Recession of 2008 may have called into question the responsiveness of the market to effort. On the one hand, heads of financial companies were largely blamed for the recession undermining the belief that their high rewards were well-deserved. On the other hand, many hard working people were laid off, showing that the market does not reward all who work hard (Grusky and Wimer 2010). Among those in the highest-paid occupations, Chief Executive Officers (CEOs) have been under a considerable amount of criticism even long before the Great Recession of 2008 (Crystal 1991). Research on newspaper coverage about CEO compensation shows that between 1990 and 2010 over 26,000 news pieces were published on the subject with only a handful having a positive tone on the issue (Kuhnen and Niessen 2012). While a growing number of studies document the American discontent with extremely high incomes, an important limitation in our understanding of attitudes toward extremely high pay is why many Americans have a strong negative response to high CEO pay. Americans have long been known for their belief in economic individualism and free market and for tolerating income inequality under fair conditions (Bowman 2000; Hochschild 1996; Huber and Form 1973; Kluegel and Smith 1986; Ladd 1994; Ladd and Bowman 1998; Lane 1986; Schlozman and Verba 1979; Verba and Orren 1985). Why do large numbers of Americans find very high CEO pay problematic? Many Americans find the highest paid occupations, among them executives, overpaid (Kiatpongsan and Norton 2014; Kluegel and Smith 1986; McCall 2013). According to polls from a variety of sources between 1976 and 2011, more than 72% of Americans routinely found CEOs overpaid (McCall 2013). Recent research shows that extremely high pay is becoming more contentious (Osberg and Smeeding 2006) and problematic (Bartels 2005; McCall 2003; McCall 2013; Page and Jacobs 2009). Two possible explanations for the discontent with income inequality are especially noteworthy. The first is an account which emphasizes that the public views the current high earners as undeserving of their high pay. People may believe that the high earners are not contributing proportionally to their rewards. For example, some scholars suggest that some form of rent-seeking explains the pay that CEOs receive (Bebchuk and Grinstein 2005; Bebchuk and Fried 2005; DiPrete, Eirich and Pittinsky 2010; Piketty and Saez 2006; Weeden and Grusky 2014). Recent research shows that the pay-setting process is influenced by the CEO’s localized ties, especially board networks (Kim, Kogut and Yang 2015). If CEOs are perceived as overpaid in relation to their contributions (whether through rent-seeking or through any means by which a disconnect occurs between pay and performance), then, it is possible that were they understood as contributing proportionally highly, the American public might be more inclined to accept their extremely high pay. A second possibility is that the American public finds such extremely high pay and extreme inequality objectionable in principle. Research examining the moral economy shows the importance of morality and fairness in evaluating economic activities (Booth 1994; Elster 1989; Etzioni 1988; Fourcade and Healy 2007; Hirschman 1984; Mansbridge 1990; Sayer 2005; Svallfors 2007; Thompson 1971). Based on this proposition, even if high-earners were understood as contributing proportionally to such extreme levels of pay, people would still feel uneasy about these high levels of pay. Using data from a national survey experiment uniquely designed to test these propositions, this study aims to advance our understanding about the reasons for the negative public response to high CEO pay. Do Americans believe that the better a CEO performs, the more he or she should earn, without a limit? Alternatively, is there a limit to acceptable pay at the top? background and conceptual framework Almost 60% of the top 1% of incomes in the U.S. come from various types of employment income including wages, salaries, pensions, exercised stock options and bonuses (Piketty and Saez 2006). A large percentage of top income earners are executives, managers, supervisors, and financial professionals. People in these occupations make up 60 percent of the top 0.1 percent of income earners (Bakija, Cole and Heim 2012). In addition, not only is the pay for executives very high but it also has seen a great increase in the past twenty-five years (Bebchuk and Fried 2005; Hall and Liebman 1998; Jensen, Murphy and Wruck 2004). Median CEO compensation in S&P 500 companies was between $6 million and $7 million throughout the 2000s (Frydman and Jenter 2010). Although variations in its formulations exist, attitudes are understood as tendencies to evaluate an attitude object (Eagly and Chaiken 1993) or the “categorization of a stimulus object along an evaluative dimension” (Zanna and Rempel 1988). Based on expectancy-value models of attitudes (Fishbein and Ajzen 1975; Rosenberg 1960), attitudes are composed of the sum of various relevant beliefs about an object each weighed by an evaluation (for example, favorable or unfavorable) of the relevant belief. Based on these formulations, attitudes toward CEO pay are not based on a single belief, but multiple beliefs and their evaluations. In the next section, I discuss some of the potentially salient considerations in the formation of attitudes about CEO pay. Justice Research and Equity Theory Research in the past decades has shown support for the idea that, in market settings, people prefer to allocate rewards proportionally to relevant inputs such as effort and performance (Hegtvedt 2001; Miller 1992). People weigh factors differently in deciding how goods and rewards should be distributed under different circumstances (Deutsch 1975; Lamm and Schwinger 1980; Leventhal 1976; Miller 1999; Scott et al. 2001; Walzer 1983). It is not always performance that counts heaviest in allocating rewards, however, most people believe that fair pay should be proportional to contribution in the workplace (Evans, Kelley and Peoples 2010; Hegtvedt 2001; Hochschild 1981; Lane 1962; Leung and Park 1986; Leventhal 1976; Marshall et al. 1999; Miller 1992; Miller 1999; Mitchell et al. 1993; Prentice and Crosby 1987; Scott et al. 2001). The importance placed on contribution, and especially performance, is not unique to Americans and is widespread across countries as well as across strata within countries (Evans, Kelley and Peoples 2010; Hochschild 1981; Lane 1962; Marshall et al. 1999). A long research tradition has examined people’s justice perceptions. Major constructs of interest are distributive justice which concerns itself with perceived fairness of outcomes of allocation decisions (Adams 1965; Homans 1974; Walster, Berscheid and Walster 1973), procedural justice which emphasizes the fairness in procedures through which those allocations are made (Thibaut and Walker 1975), and interactional justice which emphasizes the interpersonal playout of that allocation (Bies and Moag 1986). Most recently, justice research has made considerable advances in organizational settings (Folger and Greenberg 1985; Greenberg 1987). The principle of proportionality of rewards and contributions is central to theories of distributive justice. Based on equity theory, perceived fairness depends on workers having equivalent ratios of inputs and outcomes (Adams 1965; Messick and Cook 1983; Walster and Walster 1975; Walster, Walster and Berscheid 1978). To illustrate, “…if two equally talented students work equally hard and do an equally good job, they are entitled to equal salaries… If one of the students works twice as hard as the other and thus does twice a good job, he is entitled to twice the salary” (Walster, Walster and Berscheid 1978). However, William Goode (1978) argues that the rule of proportionality construed as a constant ratio of rewards and contributions seems incorrect, particularly at the lowest and highest levels of performance. Instead, he posits that where along the reward and contributions curve an individual is greatly affects returns. Goode (1978) writes, an “important deficiency in the proportionality rule is that what is viewed as a justly larger or smaller reward in money or prestige actually varies in a curvilinear way with output or contribution.” For example, for high performers, rewards rise faster than contributions (for winner-take-all markets, see for example, Rosen 1981 and Frank and Cook 1995). Other arguments in favor of the nonlinearity of the rule of proportionality have also been advanced. Research in the moral economy shows that economic decisions and actions are often motivated by norms and morality (Booth 1994; Elster 1989; Etzioni 1988; Fourcade and Healy 2007; Hirschman 1984; Mansbridge 1990; Sayer 2005; Svallfors 2007; Thompson 1971). While previous research has focused on homo economicus and later on homo socialis, a new conception of homo moralis is emerging (Skitka, Bauman and Mullen 2008). Some evidence shows that there may be a morally-grounded maximum level of acceptable pay in the workplace. This upper limit to ethically acceptable pay has been called a “social maximum” (Alves and Rossi 1978) or an “ethical ceiling” (Osberg and Smeeding 2006). Based on findings from their vignette study, Wayne Alves and Peter Rossi (1978) have argued that some earnings are too high to be justified by an earner’s education or occupation. If indeed there is an ethical ceiling to pay, it could be the source of the discontent with high CEO pay. In light of this body of work, there is need for empirical research on the rule of proportionality focusing on the top of the income distribution and on current levels of CEO pay. This study offers exactly such a test. Marginal Productivity Theory and CEO Compensation Robert Frank and Philip Cook (1995) write that the theory of marginal productivity which aims to explain how workers are remunerated was not intended as a tool to evaluate fairness in the distribution of incomes and yet people have come to see fairness in pay that reflects marginal product. Whether CEO pay reflects marginal product is a big debate among scholars of CEO compensation. Some scholars argue that CEO pay, however high it may be, is a reflection of the CEO’s performance and a competitive market for talent (Frank and Cook 1995; Gabaix and Landier 2008; Kaplan 2008; Kaplan and Rauh 2010; Mankiw 2010). Other scholars have tried to disabuse the public of the idea that marginal productivity theory explains the dramatic rise in executive pay (Piketty 2014; Reich 2010; Reich 2012; Reich 2015; Stiglitz 2003; Stiglitz 2012). They have argued that it is difficult to estimate the marginal product of a CEO in a large organization due to the team-based nature of their work and that the argument that marginal product determines CEO pay is made further problematic by data showing cross-country differences in CEO pay among industrialized countries (Piketty 2014). Some scholars have argued that CEO pay follows a pay-for-luck model where CEO pay rises with general market trends (Reich 2012; Stiglitz 2012) or that holes in corporate governance drive the high pay (Bebchuk and Grinstein 2005; Bebchuk and Fried 2005; Bebchuk 2010; DiPrete, Eirich and Pittinsky 2010; Piketty and Saez 2006). For example, Thomas Piketty, Emmanuel Saez and Stefanie Stantcheva (2014) put forth a compensation bargaining model showing that CEO pay depends on top marginal tax rates as CEOs presumably invest more in bargaining for higher pay when tax rates are low. While CEO pay is highly contested, the question that receives most attention is how closely CEO pay is tied to performance and how this alignment can be improved. For example, some scholars have argued in favor of a pay schedule that changes linearly with performance without a lower or upper limit to pay (Jensen, Murphy and Wruck 2004; Stewart 1990). The aim of this paper is not to weigh in on the debate regarding whether actual CEO pay reflects marginal product. Instead, the question addressed here is whether public discontent would decrease if CEO pay were indeed tightly tied to performance. Based on previous work, attitudes toward CEO pay are conceptualized as evaluations based on beliefs about the principle of equity, a CEO’s perceived marginal product, and the morality of high inequality. This study uses a unique set of data from a national survey experiment to examine two explanations for the American discontent with extremely high CEO pay: a) the American public shows discontent because CEOs are not seen as contributing proportionally to their pay b) the American discontent with extremely high pay is due to such pay being objectionable in principle. Empirical Expectations The central empirical expectation is that the performance and perceived fair pay function is not a linearly increasing function. Instead, fair pay will rise less than proportionally to performance at very high levels of pay thereby showing diminishing returns to performance. H1 (inequality aversion hypothesis): fair CEO pay will not scale linearly with performance. H0 (rent-seeking aversion hypothesis): fair CEO pay will scale linearly with performance. Beyond the central hypotheses, additional nuances in findings are expected. Considerations about fair pay are influenced by perceived actual pay, whether perceptions reflect reality or not (Osberg and Smeeding 2006). The actual hierarchy of occupations is replicated by most respondents when they are asked about just pay for occupations (Kluegel and Smith 1986). Prior research also shows that in responding to questions people are likely to gravitate towards responses hinted at them by the researchers even if these are arbitrary numbers. This phenomenon called anchoring shows large and consistent effects (Cervone and Peake 1986; Chapman and Johnson 2002; Green et al. 1998; Markovsky 1988; Northcraft and Neale 1987; Schwarz and Bless 1992; Switzer and Sniezek 1991; Tversky and Kahneman 1974; Wilson et al. 1996; Wright and Anderson 1989). I use the term anchoring to refer to a “starting point for adjustment” (Jacowitz and Kahneman 1995). In this type of anchoring, respondents are given a numerical idea of a possible response and are asked their opinion using that same numerical scale. For example, when respondents are told that a reward of a certain dollar amount was given in appreciation for a deed and are later asked how much to reward someone in a similar situation, they are likely to gravitate towards this anchor (Markovsky 1988). CEO pay ranges widely. According to data from the Bureau of Labor Statistics (2012), mean annual CEO pay was approximately $200,000 while it is much higher in Standard & Poor’s 500 companies (Frydman and Jenter 2010). Therefore, a low anchor such as $200,000 and a high anchor such as $5 million are both realistic amounts. Based on previous research showing strong and consistent anchoring effects, I expect that preferences for fair CEO pay will be affected by what respondents are told an average CEO is usually paid in an industry. However, despite the strong anchoring effects found in earlier research, I expect that respondents will not uncritically follow that suggested pay. I hypothesize that people have beliefs about what an appropriate amount of pay would be for the CEO of a large company. While the anchor will have an effect, responses will also gravitate towards this amount. Empirically, I test the proposition that the higher people are told the average actual CEO pay is, the more they will size it down to yield fair pay. H2a: the higher the anchor for average CEO pay, the higher the fair pay reported. H2b: the higher the anchor for average CEO pay, the greater the downward adjustment of fair pay. We should observe diminishing returns especially at higher levels of pay because the higher the level of pay, the closer the pay will be to any one person’s ethical pay ceiling, if there is one. Therefore, responses from respondents in the treatment condition where the anchor for CEO pay is highest should produce the function with the most evident diminishing returns. H2c: the higher the anchor for average CEO pay, the greater the curvilinearity of the fair pay and performance function. Americans are well-known for their strong belief in a free market and in economic individualism (Hochschild 1996; Huber and Form 1973; Kluegel and Smith 1986; Ladd 1994; Ladd and Bowman 1998; Schlozman and Verba 1979). Some scholars find that American public opinion about inequality is characterized by polarization (Osberg and Smeeding 2006). I expect to find differences in attitudes between people who strongly believe in the free market ideology and those who do not. Specifically, I expect to find that those who support the free market strongly believe that pay should increase linearly proportionally with performance without a limit. H3: belief in the free market ideology leads to linear fair pay preferences. Data and Methods Survey experiments combine the strength of surveys in providing a sample that closely resembles the target population and the strength of randomized experiments in internal validity (Mutz 2011; Sniderman and Grob 1996). Data come from the Performance and Fair CEO Pay Study, a national survey experiment conducted in partnership with the YouGov Survey Company in July 2013. YouGov maintains a web-based and opt-in panel with more than one million respondents over the age of 18 in the U.S. and uses a sample matching methodology to construct its samples. This includes matching panelists from the large YouGov panel to a target sample – a probability sample of the target population to be mimicked (Rivers 2006). The objective of this matching process is to approximate the overall distribution of several indicators in the high quality sample to correct for the non-representativeness of YouGov’s opt-in sample. To accomplish this, one or more YouGov panelists are matched to each member of the target sample. For this particular survey, the respondents were matched on gender, age, race, education, party identification, ideology, and political interest. The synthetic sampling frame was constructed using the 2010 American Community Survey, along with data on voter registration status and turnout from the November 2010 Current Population Survey, and data on political interest and party identification from the 2007 Pew Religious Landscape Survey. Data gathered using YouGov’s matched sample method may perform as well as data from high-quality national surveys with probability sampling (Ansolabehere and Schaffner 2011; Ansolabehere and Rivers 2013; Vavreck and Iyengar 2013) accurately predicting voting behavior and election results (Rivers and Bailey 2009; Rivers 2012; Silver 2012). A total of 1,000 YouGov panelists completed the present survey, 11 cases were list-wise deleted due to missing values, yielding a final sample of 989 respondents with each of the treatment conditions having 225 to 277 subjects. Accounting for nonresponse and incompletes, the American Association for Public Opinion Research Response Rate 3 (RR3) for the study is 45.3%. Experiment Design To examine how fair CEO pay changes as a result of changing performance at different levels of pay, I cross two factors in this study: CEO performance (seven levels: a zero-point where there are no profits or losses, 11.6% profit; 23.3% profit; 35% profit and symmetrical losses) and average CEO pay in the industry (four levels: a control group with no mention of average pay, and $200,000, $1 million, or $5 million). This survey experiment features a between-subjects component and a within-subjects component. As a between-subjects manipulation, I provide each respondent with information about the average CEO pay in an industry. The control group is not provided with any information about the average pay in the industry while those in the three treatment conditions read that the industry average is $200,000, $1 million, or $5 million. As a within-subjects manipulation, I ask each respondent to indicate fair pay for a fictive CEO for seven levels of performance. Because the central hypothesis tests the limits of the rule of proportionality, it is important that the performance levels in the experiment reach very high and yet realistic levels for a company of the size evoked in this experiment. I use information compiled by Forbes for the purpose of benchmarking the numbers. Forbes provides data on measures of financial success for a list of 400 Best Big Companies who have a large presence in the U.S. market. Forbes also works with Audit Integrity to ensure that the list is of financially healthy companies. Based on these data, 12-month profit margins of companies in the top ten of this list ranked by this measure were between 30.3% to 46.2% (DeCarlo 2008). The highest level of performance reached in this survey experiment is 35% profit with performance ranging from 35% loss to 35% profit. I space intervals approximately equally to accommodate seven levels of performance with the inclusion of a zero-point. The experiment yields seven pay and performance pairings per respondent. The information is presented to the respondents using a vignette describing information about a fictitious CEO named John Hall of a fictitious company named American Wares Company. Men make up the vast majority of all CEOs in the largest companies (Eagly and Carli 2007; Sandberg 2013). I focus on respondents’ views about the pay of a male CEO to maintain statistical power for the experiment. Future research should focus on addressing differences in perceptions of fair pay for male and female CEOs as well as white and nonwhite CEOs. For example, Guillermina Jasso and Eva Meyersson Milgrom (2008) find that an otherwise identical female CEO is perceived as deserving of 84 to 94 percent of earnings of a male CEO. They also find that respondents do not hold consensus on the importance of various characteristics of the CEO such as years of schooling or characteristics of the firm in determining just pay. In this experiment, I focus on CEOs and their workplace contributions. Variables Dependent Measure: Fair CEO Pay I measure fair pay by asking respondents what John Hall, the CEO of a large multinational company, should be paid. A dropdown menu of response choices ranging from less than $50 thousand to $10 million is provided. The response options are in $10 thousand increments up to $100 thousand and in $100 thousand increments thereafter. As mentioned earlier, I measure fair pay at seven different levels of CEO performance. I begin with the question of how much the CEO should be paid when he performs exactly at the average level in his industry. I then ask respondents about fair pay in ascending order of performance, and following that, in descending order. The decision to use dollars or log dollars in the following regression models is an important one for the results of this experiment. Based on equity theory, rewards are perceived as fair when the ratio of inputs and outcomes is equivalent between two equivalent workers (Adams 1965; Messick and Cook 1983; Walster and Walster 1975; Walster, Walster and Berscheid 1978). Pilot interviews based on the scenarios used in this study show that, typically, when people believe that pay should increase directly proportionally to performance, they tend to increase pay by a fixed dollar amount at each performance interval rather than by calculating a percent increase to maintain proportionality of rewards and contributions. I therefore operationalize upholding the rule of proportionality as increasing fair pay by some dollar amount (rather than by the same percentage). Consequently, I use fair pay in dollars (rather than log dollars) as the dependent variable in the following regression models. Note that if log of fair pay was used in the models, then, the fair pay function for a respondent who increases fair pay by the same dollar amount at each performance interval would show diminishing returns because as pay increases the percent increase in pay from one level of performance to the next will diminish. Performance Performance and performance-squared (to indicate the curvature of the fair pay function) are the main predictors. I provide respondents with information about company profits and losses to indicate the CEO’s performance. I measure fair pay at seven different levels of performance: a zero-point where the company is doing exactly as well as the average in the industry, three levels of profits (11.6% profit; 23.3% profit; 35% profit) and symmetrical losses (11.6% loss; 23.3% loss; 35% loss). I recode this variable from 1 to 7 to indicate the seven levels of performance (where higher numbers indicate better performance). Manipulation checks were included in a pretest of this study conducted with 186 respondents using Amazon’s Mechanical Turk platform. A seven-point scale item to measure how respondents “rate John Hall compared to the average CEO in this industry” in terms of his performance shows that respondents find John Hall equivalent to the average CEO in his industry at the zero-point as intended. Responses to an open-ended item asking respondents to explain how they decided on John Hall’s pay also show that profits and losses are understood as signifying John Hall’s performance and that these are taken into account in deciding pay. Belief in the Free Market I measure respondents’ belief in the free market ideology by replicating an item from the General Social Survey. The item reads, “private enterprise is the best way to solve America’s economic problems.” Respondents are asked to state whether they agree or disagree with this statement. The response choices are strongly agree, agree, neither agree nor disagree, disagree, and strongly disagree. I recode this variable to indicate whether the respondent strongly agrees or agrees with this statement (1) or not (0). Control Variables I include age (in tens of years), sex, race, marital status, education (using indicator variables to indicate the educational categories less than high school, some college, BA degree, and more than BA degree), and family income. I also include a measure of political ideology (very liberal, liberal, conservative, very conservative, and independent). Models I specify the relationship between pay and performance using Ordinary Least Squares (OLS) regression models. This relationship is specified by adding a squared term for performance in the model to indicate the anticipated curvature. Respondents have different ideas about how fair pay and performance should be related: closely or loosely, proportionally or not. Therefore, pay and performance pairings will depend on the respondent. I take this into account with clustering and robust standard errors. ANALYSIS The Effect of Perceived Average CEO Pay on Fair Pay Figure 1 shows the median dollar amount of fair CEO pay chosen by respondents in the four experimental conditions as the CEO’s performance changes. The x-axis shows the CEO’s performance and the y-axis shows fair pay for that level of performance. The zero-point on the x-axis indicates the scenario where the CEO is described as performing equally well as the average CEO in his industry with neither loss nor profit. Positive numbers indicate better performance, and negative numbers indicate worse performance.1 Figure 1. View largeDownload slide Median Fair CEO Pay (by Level of Performance and Mentioned Average CEO Pay in the Industry) Source: Performance and Fair CEO Pay Study, 2013. N = 989. Figure 1. View largeDownload slide Median Fair CEO Pay (by Level of Performance and Mentioned Average CEO Pay in the Industry) Source: Performance and Fair CEO Pay Study, 2013. N = 989. I begin by examining the median fair pay chosen by respondents at the zero-point of performance. The respondents in the control condition were not given any information about how much the average CEO in the industry is paid and median pay in this condition at the zero-point is $400,000. In contrast, the median fair pay for respondents who were told that the average CEO is paid $200,000 is exactly $200,000. The median response among those who are given the information that the average CEO is paid $1 million is $700,000. Finally, median fair pay chosen by those in the $5 million treatment group is $1 million. Hence, the higher or lower the average CEO pay mentioned in the treatment, the higher or lower the median fair pay. To further examine the effect of perceived industry average pay, Table 1 Table 1. Mean Fair Pay for CEO with Average Performance, by Treatment Condition Mentioning Average CEO Pay in Industry Control $200,000 Treatment $1 Million Treatment $5 Million Treatment Mean Fair Pay $1,287,339 (163,403) $305,610*** (51,198) $818,339*** (73,964) $2,007,867*** (134,999) N 233 254 277 225 Control $200,000 Treatment $1 Million Treatment $5 Million Treatment Mean Fair Pay $1,287,339 (163,403) $305,610*** (51,198) $818,339*** (73,964) $2,007,867*** (134,999) N 233 254 277 225 Note: Standard errors are in parentheses. Source: Performance and Fair CEO Pay Study 2013. ***p < .001 (two-tailed) compared with treatment condition that precedes in dollar value. Table 1. Mean Fair Pay for CEO with Average Performance, by Treatment Condition Mentioning Average CEO Pay in Industry Control $200,000 Treatment $1 Million Treatment $5 Million Treatment Mean Fair Pay $1,287,339 (163,403) $305,610*** (51,198) $818,339*** (73,964) $2,007,867*** (134,999) N 233 254 277 225 Control $200,000 Treatment $1 Million Treatment $5 Million Treatment Mean Fair Pay $1,287,339 (163,403) $305,610*** (51,198) $818,339*** (73,964) $2,007,867*** (134,999) N 233 254 277 225 Note: Standard errors are in parentheses. Source: Performance and Fair CEO Pay Study 2013. ***p < .001 (two-tailed) compared with treatment condition that precedes in dollar value. shows mean fair pay by treatment condition and whether the mean pay for each of the conditions is significantly different from the condition that precedes it in dollar value. The mean fair pay chosen by the respondents in the four treatment conditions are ordered in the expected direction. The mean for the $200,000 treatment condition is the lowest: it is $305,610. The $1 million treatment group has the next highest mean fair pay with $818,339. The control group’s mean is higher than the mean for the $1 million treatment group: it is $1,287,339. Using data from the General Social Survey and the International Social Survey Programme, Leslie McCall (2013) shows that mean perceived actual pay for the “chairman of a large national company” was approximately $3 million in 2010. It is therefore not surprising that the mean for the control group is higher than the $1 million treatment group. Pay is also more right-skewed in the control condition than it is in the $1 million condition. In addition, the standard deviation of pay in the control condition is higher. This explains why the control group has a lower median fair pay and a higher mean fair pay compared with the $1 million treatment condition.2 Lastly, mean fair pay is $2,007,867 for the $5 million treatment group. These are statistically significantly different from the control group. More importantly, the mean for each treatment group is statistically significantly different from the group preceding it in dollar value. Therefore, the results show that people pay attention to what they believe to be average CEO pay in the industry in deciding fair CEO pay. Hypothesis 2a which states, “The higher the anchor for average CEO pay, the higher the fair pay reported,” finds support from the data. Next, I consider the hypothesis, “the higher the anchor for average CEO pay, the greater the downward adjustment of fair pay.” Table 2 Table 2. Mean Change in Pay from Industry Average Pay to the Respondent’s Choice, by Treatment Condition Mean SE N $200,000 Treatment 52.8%*** (25.60) 254 $1 Million Treatment −18.2%*** (7.396) 277 $5 Million Treatment −59.8%*** (2.700) 225 Mean SE N $200,000 Treatment 52.8%*** (25.60) 254 $1 Million Treatment −18.2%*** (7.396) 277 $5 Million Treatment −59.8%*** (2.700) 225 Source: Performance and Fair CEO Pay Study 2013. ***p < .001 (two-tailed) compared with treatment condition that precedes in dollar value. Table 2. Mean Change in Pay from Industry Average Pay to the Respondent’s Choice, by Treatment Condition Mean SE N $200,000 Treatment 52.8%*** (25.60) 254 $1 Million Treatment −18.2%*** (7.396) 277 $5 Million Treatment −59.8%*** (2.700) 225 Mean SE N $200,000 Treatment 52.8%*** (25.60) 254 $1 Million Treatment −18.2%*** (7.396) 277 $5 Million Treatment −59.8%*** (2.700) 225 Source: Performance and Fair CEO Pay Study 2013. ***p < .001 (two-tailed) compared with treatment condition that precedes in dollar value. presents the percent change in pay from the industry average provided in each treatment condition to the respondent’s choice of fair pay at the zero-point of performance. The control condition is not included in this table because an industry average is not mentioned in the control group. I calculate the percent change in pay for each respondent and report the mean for all responses by treatment condition. The mean percent change for respondents in the $200,000 treatment condition is 52.8%. Respondents in this condition, on average, adjust the $200,000 suggested pay upwards, increasing it by 52.8%. In contrast, those in the $1 million treatment condition tend to adjust their suggested pay downwards by 18.2%. Lastly, respondents in the $5 million treatment condition, on average, decrease the suggested pay in this condition by 59.8%. The means for all treatment conditions are significantly different from one another. Therefore, the respondents do not uncritically follow the industry average pay in deciding fair pay. The findings suggest that people have opinions about how much CEOs should be paid and that they adjust their response accordingly. Hypothesis 2b finds support from the data. How Does Fair CEO Pay Respond to Performance? Next, I consider the central hypothesis (inequality aversion hypothesis) which states, “fair CEO pay will not scale linearly with performance.” The empirical expectation is that the fair rewards function will show diminishing returns evidenced by a positive and significant performance coefficient in the model followed by a negative and significant performance-squared coefficient. I examine the overall pattern of the data using an OLS regression model with clustering to account for having seven observations (i.e. pay and performance pairings) per respondent and using robust standard errors. Model 1 in Table 3 Table 3. Coefficients from OLS Regression Model Predicting Fair CEO Pay M1 M2 Variables β SE β SE Performance 101,942*** (19,159) 168,193*** (50,721) Performance2 6,117** (2,204) −2,673 (5,698) Treatment Condition (reference = control)  Treatment $200,000 −819,511*** (141,058) −310,344* (135,136)  Treatment $1 Million −349,361* (145,497) −139,018 (138,405)  Treatment $5 Million 942,548*** (188,768) 500,871** (182,154) Interactions  Performance × Treatment $200,000 −182,206*** (54,964)  Performance × Treatment $1 Million −97,148 (57,585)  Performance × Treatment $5 Million 34,081 (71,247)  Performance2 × Treatment $200,000 10,983 (6,365)  Performance2 × Treatment $1 Million 8,912 (6,479)  Performance2 × Treatment $5 Million 15,268 (8,147) Constant 607,542*** (130,607) 518,345*** (127,372) R-squared 0.1514 0.1614 Observations 6,923 6,923 Clusters (standard errors adjusted for clusters) 989 989 M1 M2 Variables β SE β SE Performance 101,942*** (19,159) 168,193*** (50,721) Performance2 6,117** (2,204) −2,673 (5,698) Treatment Condition (reference = control)  Treatment $200,000 −819,511*** (141,058) −310,344* (135,136)  Treatment $1 Million −349,361* (145,497) −139,018 (138,405)  Treatment $5 Million 942,548*** (188,768) 500,871** (182,154) Interactions  Performance × Treatment $200,000 −182,206*** (54,964)  Performance × Treatment $1 Million −97,148 (57,585)  Performance × Treatment $5 Million 34,081 (71,247)  Performance2 × Treatment $200,000 10,983 (6,365)  Performance2 × Treatment $1 Million 8,912 (6,479)  Performance2 × Treatment $5 Million 15,268 (8,147) Constant 607,542*** (130,607) 518,345*** (127,372) R-squared 0.1514 0.1614 Observations 6,923 6,923 Clusters (standard errors adjusted for clusters) 989 989 Source: Performance and Fair CEO Pay Study 2013. Robust standard errors in parentheses *p < 0.05, **p < 0.01, ***p < 0.001 (two-tailed) Table 3. Coefficients from OLS Regression Model Predicting Fair CEO Pay M1 M2 Variables β SE β SE Performance 101,942*** (19,159) 168,193*** (50,721) Performance2 6,117** (2,204) −2,673 (5,698) Treatment Condition (reference = control)  Treatment $200,000 −819,511*** (141,058) −310,344* (135,136)  Treatment $1 Million −349,361* (145,497) −139,018 (138,405)  Treatment $5 Million 942,548*** (188,768) 500,871** (182,154) Interactions  Performance × Treatment $200,000 −182,206*** (54,964)  Performance × Treatment $1 Million −97,148 (57,585)  Performance × Treatment $5 Million 34,081 (71,247)  Performance2 × Treatment $200,000 10,983 (6,365)  Performance2 × Treatment $1 Million 8,912 (6,479)  Performance2 × Treatment $5 Million 15,268 (8,147) Constant 607,542*** (130,607) 518,345*** (127,372) R-squared 0.1514 0.1614 Observations 6,923 6,923 Clusters (standard errors adjusted for clusters) 989 989 M1 M2 Variables β SE β SE Performance 101,942*** (19,159) 168,193*** (50,721) Performance2 6,117** (2,204) −2,673 (5,698) Treatment Condition (reference = control)  Treatment $200,000 −819,511*** (141,058) −310,344* (135,136)  Treatment $1 Million −349,361* (145,497) −139,018 (138,405)  Treatment $5 Million 942,548*** (188,768) 500,871** (182,154) Interactions  Performance × Treatment $200,000 −182,206*** (54,964)  Performance × Treatment $1 Million −97,148 (57,585)  Performance × Treatment $5 Million 34,081 (71,247)  Performance2 × Treatment $200,000 10,983 (6,365)  Performance2 × Treatment $1 Million 8,912 (6,479)  Performance2 × Treatment $5 Million 15,268 (8,147) Constant 607,542*** (130,607) 518,345*** (127,372) R-squared 0.1514 0.1614 Observations 6,923 6,923 Clusters (standard errors adjusted for clusters) 989 989 Source: Performance and Fair CEO Pay Study 2013. Robust standard errors in parentheses *p < 0.05, **p < 0.01, ***p < 0.001 (two-tailed) shows the results using performance and performance-squared as the predictor variables. Interaction effects are added in Model 2. Model 1 shows that the coefficient for performance is positive and significant and that the coefficient for performance-squared is also positive and significant. This indicates that fair pay increases with performance and that the rewards function increases faster at higher pay levels. The coefficients for the treatment conditions confirm the earlier findings: fair pay is lower in the $200,000 treatment condition than it is in the control condition; it is also lower in the $1 million condition and higher in the $5 million condition compared to the control condition. Model 2 adds interaction effects for performance and performance-squared and the treatment conditions. When interactions are included, the performance-squared term is no longer statistically significant. Most importantly, however, I do not find evidence for diminishing returns to performance at higher pay levels under either model specification. While findings suggest that respondents show aversion to high inequality in other ways (as will be further discussed below), Hypothesis 1 (inequality aversion hypothesis) which states that fair CEO pay will not scale linearly with performance does not find support from the data. This finding reveals a commitment to the principle of proportionality of rewards and contributions even at levels of pay reached in this experiment. Examining the interaction effects, I find that the $200,000 treatment condition has an overall flatter fair pay function than other conditions. The interaction between that condition and performance is negative and significant, and it points in the opposite direction of the main effect in the model. Further examination (not shown) reveals that, for this treatment group, the mean percent change from one level of performance to the next is positive (ranging from 27% to 125% increase) but usually lower than in other experimental conditions. Moreover, for this treatment group, median percent change in fair pay from one level of performance to the next is zero. In comparison, median percent change for the $1 million treatment group is approximately 10% at each interval. Data cannot weigh in on the reasons for the reduced sensitivity to performance for the lower treatment group, however, to speculate, respondents may not want to allocate too low a pay for a position as high as the CEO’s. This possibility is congruent with results presented in Table 2 where, on average, respondents increase the suggested pay for this treatment group. An interesting question for future research is whether there could be an occupational floor to fair pay (rather than an ethical ceiling as hypothesized here). Other interaction effects included in the model are not statistically significant. No significant differences are found between the control group and the higher dollar amount conditions in slope or in the degree of curvature for the fair pay functions. Considering the findings, I do not find evidence to support Hypothesis 2c which states that higher anchors for average CEO pay will lead to greater curvilinearity. Belief in the Free Market Ideology and Fair Pay I hypothesized that respondents’ belief in the free market ideology would lead to linear fair pay preferences. I examine the effect of agreeing with the statement on private enterprise as well as how the curvature of the function changes with belief in this statement. Table 4 Table 4. Coefficients from OLS Regression Model Predicting Fair CEO Pay Model 1 Model 2 Variables β SE β SE Performance 168,193*** (50,776) 146,175** (52,487) Performance2 −2,673 (5,704) −6,336 (6,004) Treatment Condition (reference = control)  Treatment $200,000 −321,043* (139,699) −303,550* (139,287)  Treatment $1 Million −140,733 (141,668) −119,712 (141,645)  Treatment $5 Million 469,944** (181,261) 489,601** (181,867) Interactions  Performance × Treatment $200,000 −182,206*** (55,024) −184,593*** (55,048)  Performance × Treatment $1 Million −97,148 (57,648) −100,017 (57,668)  Performance × Treatment $5 Million 34,081 (71,325) 31,398 (71,330)  Performance2 × Treatment $200,000 10,983 (6,372) 10,586 (6,350)  Performance2 × Treatment $1 Million 8,912 (6,486) 8,435 (6,461)  Performance2 × Treatment $5 Million 15,268 (8,156) 14,821 (8,149)  Age (in tens of years) −49,814 (29,132) −49,814 (29,136)  Female −273,389** (98,594) −273,389** (98,608)  Married −202,519* (99,191) −202,519* (99,205)  Race nonwhite 70,821 (109,324) 70,821 (109,340) Education (reference = high school)  <High School 285,826 (277,706) 285,826 (277,746)  Some College 189,040 (112,334) 189,040 (112,350)  BA Degree 64,906 (123,788) 64,906 (123,806)  >BA Degree 193,524 (182,967) 193,524 (182,993) Family Income (ln) 239,771*** (69,242) 239,771*** (69,252) Political Ideology (reference = moderate)  Very Liberal −2,646 (154,560) −2,646 (154,582)  Liberal 96,222 (133,447) 96,222 (133,466)  Conservative 332,231* (130,033) 332,231* (130,052)  Very conservative 324,278* (144,465) 324,278* (144,486)  Independent 44,548 (180,525) 44,548 (180,552) Values  Free Market Ideology (agree) 475,842*** (117,661) 84,299 (128,189)  Free Market Ideology (agree) × Performance 53,439 (38,534)  Free Market Ideology (agree) × Performance2 8,889* (4,428) ConstantR-squared −2026,000**0.222 (743,098) −1865,000*0.227 (746,682) ObservationsClusters (standard errors adjusted for clusters) 6,923989 6,923989 Model 1 Model 2 Variables β SE β SE Performance 168,193*** (50,776) 146,175** (52,487) Performance2 −2,673 (5,704) −6,336 (6,004) Treatment Condition (reference = control)  Treatment $200,000 −321,043* (139,699) −303,550* (139,287)  Treatment $1 Million −140,733 (141,668) −119,712 (141,645)  Treatment $5 Million 469,944** (181,261) 489,601** (181,867) Interactions  Performance × Treatment $200,000 −182,206*** (55,024) −184,593*** (55,048)  Performance × Treatment $1 Million −97,148 (57,648) −100,017 (57,668)  Performance × Treatment $5 Million 34,081 (71,325) 31,398 (71,330)  Performance2 × Treatment $200,000 10,983 (6,372) 10,586 (6,350)  Performance2 × Treatment $1 Million 8,912 (6,486) 8,435 (6,461)  Performance2 × Treatment $5 Million 15,268 (8,156) 14,821 (8,149)  Age (in tens of years) −49,814 (29,132) −49,814 (29,136)  Female −273,389** (98,594) −273,389** (98,608)  Married −202,519* (99,191) −202,519* (99,205)  Race nonwhite 70,821 (109,324) 70,821 (109,340) Education (reference = high school)  <High School 285,826 (277,706) 285,826 (277,746)  Some College 189,040 (112,334) 189,040 (112,350)  BA Degree 64,906 (123,788) 64,906 (123,806)  >BA Degree 193,524 (182,967) 193,524 (182,993) Family Income (ln) 239,771*** (69,242) 239,771*** (69,252) Political Ideology (reference = moderate)  Very Liberal −2,646 (154,560) −2,646 (154,582)  Liberal 96,222 (133,447) 96,222 (133,466)  Conservative 332,231* (130,033) 332,231* (130,052)  Very conservative 324,278* (144,465) 324,278* (144,486)  Independent 44,548 (180,525) 44,548 (180,552) Values  Free Market Ideology (agree) 475,842*** (117,661) 84,299 (128,189)  Free Market Ideology (agree) × Performance 53,439 (38,534)  Free Market Ideology (agree) × Performance2 8,889* (4,428) ConstantR-squared −2026,000**0.222 (743,098) −1865,000*0.227 (746,682) ObservationsClusters (standard errors adjusted for clusters) 6,923989 6,923989 Source: Performance and Fair CEO Pay Study 2013. Robust standard errors in parentheses *p < 0.05, **p < 0.01, ***p < 0.001 (two-tailed) Table 4. Coefficients from OLS Regression Model Predicting Fair CEO Pay Model 1 Model 2 Variables β SE β SE Performance 168,193*** (50,776) 146,175** (52,487) Performance2 −2,673 (5,704) −6,336 (6,004) Treatment Condition (reference = control)  Treatment $200,000 −321,043* (139,699) −303,550* (139,287)  Treatment $1 Million −140,733 (141,668) −119,712 (141,645)  Treatment $5 Million 469,944** (181,261) 489,601** (181,867) Interactions  Performance × Treatment $200,000 −182,206*** (55,024) −184,593*** (55,048)  Performance × Treatment $1 Million −97,148 (57,648) −100,017 (57,668)  Performance × Treatment $5 Million 34,081 (71,325) 31,398 (71,330)  Performance2 × Treatment $200,000 10,983 (6,372) 10,586 (6,350)  Performance2 × Treatment $1 Million 8,912 (6,486) 8,435 (6,461)  Performance2 × Treatment $5 Million 15,268 (8,156) 14,821 (8,149)  Age (in tens of years) −49,814 (29,132) −49,814 (29,136)  Female −273,389** (98,594) −273,389** (98,608)  Married −202,519* (99,191) −202,519* (99,205)  Race nonwhite 70,821 (109,324) 70,821 (109,340) Education (reference = high school)  <High School 285,826 (277,706) 285,826 (277,746)  Some College 189,040 (112,334) 189,040 (112,350)  BA Degree 64,906 (123,788) 64,906 (123,806)  >BA Degree 193,524 (182,967) 193,524 (182,993) Family Income (ln) 239,771*** (69,242) 239,771*** (69,252) Political Ideology (reference = moderate)  Very Liberal −2,646 (154,560) −2,646 (154,582)  Liberal 96,222 (133,447) 96,222 (133,466)  Conservative 332,231* (130,033) 332,231* (130,052)  Very conservative 324,278* (144,465) 324,278* (144,486)  Independent 44,548 (180,525) 44,548 (180,552) Values  Free Market Ideology (agree) 475,842*** (117,661) 84,299 (128,189)  Free Market Ideology (agree) × Performance 53,439 (38,534)  Free Market Ideology (agree) × Performance2 8,889* (4,428) ConstantR-squared −2026,000**0.222 (743,098) −1865,000*0.227 (746,682) ObservationsClusters (standard errors adjusted for clusters) 6,923989 6,923989 Model 1 Model 2 Variables β SE β SE Performance 168,193*** (50,776) 146,175** (52,487) Performance2 −2,673 (5,704) −6,336 (6,004) Treatment Condition (reference = control)  Treatment $200,000 −321,043* (139,699) −303,550* (139,287)  Treatment $1 Million −140,733 (141,668) −119,712 (141,645)  Treatment $5 Million 469,944** (181,261) 489,601** (181,867) Interactions  Performance × Treatment $200,000 −182,206*** (55,024) −184,593*** (55,048)  Performance × Treatment $1 Million −97,148 (57,648) −100,017 (57,668)  Performance × Treatment $5 Million 34,081 (71,325) 31,398 (71,330)  Performance2 × Treatment $200,000 10,983 (6,372) 10,586 (6,350)  Performance2 × Treatment $1 Million 8,912 (6,486) 8,435 (6,461)  Performance2 × Treatment $5 Million 15,268 (8,156) 14,821 (8,149)  Age (in tens of years) −49,814 (29,132) −49,814 (29,136)  Female −273,389** (98,594) −273,389** (98,608)  Married −202,519* (99,191) −202,519* (99,205)  Race nonwhite 70,821 (109,324) 70,821 (109,340) Education (reference = high school)  <High School 285,826 (277,706) 285,826 (277,746)  Some College 189,040 (112,334) 189,040 (112,350)  BA Degree 64,906 (123,788) 64,906 (123,806)  >BA Degree 193,524 (182,967) 193,524 (182,993) Family Income (ln) 239,771*** (69,242) 239,771*** (69,252) Political Ideology (reference = moderate)  Very Liberal −2,646 (154,560) −2,646 (154,582)  Liberal 96,222 (133,447) 96,222 (133,466)  Conservative 332,231* (130,033) 332,231* (130,052)  Very conservative 324,278* (144,465) 324,278* (144,486)  Independent 44,548 (180,525) 44,548 (180,552) Values  Free Market Ideology (agree) 475,842*** (117,661) 84,299 (128,189)  Free Market Ideology (agree) × Performance 53,439 (38,534)  Free Market Ideology (agree) × Performance2 8,889* (4,428) ConstantR-squared −2026,000**0.222 (743,098) −1865,000*0.227 (746,682) ObservationsClusters (standard errors adjusted for clusters) 6,923989 6,923989 Source: Performance and Fair CEO Pay Study 2013. Robust standard errors in parentheses *p < 0.05, **p < 0.01, ***p < 0.001 (two-tailed) presents results from OLS regression models which include this explanatory variable as well as control variables. According to Model 1 in Table 4, agreeing with this statement has a positive and significant (p < 0.001) effect on fair pay. Controlling for other factors, those who agree with this statement allocate higher pay to the CEO described in the vignette. As interactions are added in Model 2, I find that while agreeing with this statement is no longer statistically significant, the quadratic term generated with this variable is significant and positive (p < 0.05). This finding suggests that the curvature of the fair pay function is different for people who agree with this statement compared with people who do not agree with the statement. For those who agree with the free market ideology, performance counts heavier in determining fair pay. Other noteworthy findings are that people with higher family incomes allocate larger rewards to the CEO, as do the respondents who politically identify as conservative. Women and those who are married show the opposite effect. DISCUSSION AND CONCLUSIONS The Great Recession of 2008 revived interest in high executive pay and income inequality. The primary objective of this study has been to juxtapose two propositions: a) that the American discontent with extremely high CEO pay stems mainly from people’s belief that CEOs are not contributing highly enough to merit such high pay and b) people find extremely high pay objectionable in principle and they would oppose extreme levels of pay even if they believed that CEOs were contributing highly. Respondents enter surveys with previously held beliefs, including about whether CEOs are leading their companies to profits or causing economic recessions. Performance and Fair CEO Pay Study, which uses data from a national survey-experiment, brings new information on how perceived actual CEO pay and perceived performance affect fair CEO pay. Data do not only focus on an average case, and are particularly well-suited for examining the general public’s views about the highest levels of CEO pay. The experiment provides data on fair pay for a range of performances, including very poor and very good performances. Findings reveal that attitudes toward CEO pay manifest themselves in complex ways and in ways differently than initially hypothesized. On average, respondents show aversion to extremely high pay but not in the form specified under empirical expectations. Data do not show support for the hypothesis that pay will not scale linearly with performance. As a CEO’s contribution changes, so does the CEO’s fair pay, linearly proportionally to performance and without an observed upper limit. However, respondents find $5 million as well as $1 million too high, and $200,000 too low for an average performing CEO. Only extraordinary performance leads to very high levels of fair pay. Even then, fair pay seldom reaches heights observed in reality. The highest level of performance examined here is rewarded with a median fair pay of $2 million and a mean fair pay of about $3 million (where the average CEO is described as being rewarded with $5 million). In addition, even when the average CEO is described as being paid $5 million, median fair pay increases arguably slowly – from $500,000 to $2 million as performance changes from large losses to large gains. Based on these findings, opposition to extremely high pay may, in fact, be reconciled with support for pay for performance: the slope of the pay and performance function may be at issue rather than a hard limit on pay. The findings have important policy implications. In 2014, the top one percent of earners in the U.S. received about 18% of all income, excluding capital gains (Alvaredo et al. 2011). If its current trajectory continues unchanged, according to Piketty (2014), such concentration may reach new heights especially due to the “super managers” phenomenon examined here. Joseph Stiglitz (2012) argues, “it is not just that they [bankers and CEOs] have become the whipping boys of popular opinion. They are emblematic of what has gone wrong.” (p. liv). Trends in income inequality are heavily shaped by political forces (Hacker and Pierson 2010; Piketty 2014; Reich 2012). Jacob Hacker and Paul Pierson (2010) write that political “drift” or “the deliberate failure to adapt public policies to the shifting realities of a dynamic economy” (p. 52) has led to the new heights in executive pay. Some scholars argue that whether income inequality will be intensified in the future depends largely on how effectively pay is justified and how tolerant people are of high CEO pay (Piketty 2014). CEO pay is determined by a variety of factors and the acceptability of high pay differentials has an important place among them. New policies have been proposed in the U.S. and elsewhere with the aim of reining in executive pay. Among these, a new rule by the Securities and Exchange Commission which will take effect beginning with the fiscal year 2017 and which requires the disclosure of the ratio of the chief executive’s pay to the pay of the median employee for 3,571 registrants is one that is designed to empower such norms. The objective of the disclosure is to provide new data to aid in investment decisions and say on pay voting decisions (Securities and Exchange Commission 2015). A contribution of the new disclosure requirement will be new data on the proportionality of rewards and contributions within-firms. An Associated Press/Equilar study estimates the ratio at 257:1 (Associated Press 2014). While even within-firm changes in ratios will be challenging to assess due to changes in business practices, the new data will be useful in monitoring trends. Since incomes at the middle of the income distribution are growing slower than those at the top (Mishel et al. 2012), the data will likely highlight gains within-firms also disproportionately going to the top. Companies will have some leeway in calculating median pay, but a high estimate of the median, which would reveal a lower ratio of CEO pay to median pay, may not necessarily be advantageous as employees who compare unfavorably to the median may be dissatisfied (Card et al. 2012). Alexandre Mas (2014) writes, “compensation is sensitive to increased transparency” (p. 5) because public opinion on perceived excessive pay matters. The new data may also serve to increase regular employees’ pay through the diffusion of power afforded by the disclosure (Rosenfeld and Denice 2015), or generate support for policies such as inequality taxes (Ayres and Edlin 2011) or millionaire taxes (Young et al. 2016). The findings suggest important avenues for future research. First, we cannot rule out the possibility that through extraordinary contributions, extremely high pay could be justified. By considering performance only, this study cannot weigh in on the effects of other relevant inputs on fair CEO pay. Future research should examine the combined effects of various types of inputs along with performance. For example, an examination of the effects of job creation and layoffs may be especially useful. The literature will benefit from a better understanding of how the general public believes CEO pay is actually decided (for example, which forms of rent-seeking, if any, are partially responsible for determining pay).3 Findings in this study show that attitudes toward CEO pay, in part depend on political ideology and on values such as a strong belief in a free market. Further studies can illuminate differences between stronger and weaker adherents of a free market ideology. What are the mechanisms through which the free market ideology has an effect on attitudes toward CEO pay?4 How do business leaders (who have more power in determining executive compensation) and the general public differ in a host of relevant values and beliefs and in acceptance of high CEO pay? In addition to these inquiries, gaps in perceived fair pay for male and female CEOs as well as for white and nonwhite CEOs should also be part of the future research agenda. APPENDIX Figure A1. View largeDownload slide Mean Fair CEO Pay By Level of Performance and Mentioned Average CEO Pay in the Industry Source: Performance and Fair CEO Pay Study 2013. N = 989. Figure A1. View largeDownload slide Mean Fair CEO Pay By Level of Performance and Mentioned Average CEO Pay in the Industry Source: Performance and Fair CEO Pay Study 2013. N = 989. Figure A2. View largeDownload slide Box Plots for Fair CEO Pay at Average Level of Performance, by Treatment Condition Source: Performance and Fair CEO Pay Study 2013. N = 989. Source: Performance and Fair CEO Pay Study 2013. N = 989. Figure A2. View largeDownload slide Box Plots for Fair CEO Pay at Average Level of Performance, by Treatment Condition Source: Performance and Fair CEO Pay Study 2013. N = 989. Source: Performance and Fair CEO Pay Study 2013. N = 989. Figure A3. View largeDownload slide Standard Deviation of Fair CEO Pay by Level of Performance and Average CEO Pay Mentioned in Treatment Source: Performance and Fair CEO Pay Study 2013. N = 989. Figure A3. View largeDownload slide Standard Deviation of Fair CEO Pay by Level of Performance and Average CEO Pay Mentioned in Treatment Source: Performance and Fair CEO Pay Study 2013. N = 989. The author wishes to thank David B. Grusky, Michael J. Rosenfeld, Cristobal Young, Paul Sniderman, Michael Tomz, Michelle Jackson, Shelley J. Correll, Erin M. Cumberworth, Beth Red Bird, Rachel Wright, members of the Stanford Laboratory for the Study of American Values and the anonymous Social Problems reviewers for their valuable comments on prior drafts. This research was supported by funding from the Stanford Laboratory for the Study of American Values. Footnotes 1 Figure A1 shows results using the mean and confidence intervals around the mean. 2 Figure A2 shows box plots for fair pay at the zero point of performance by treatment group. Figure A3 shows the standard deviation of mean fair pay by treatment group and by performance. Fair pay is right-skewed and especially so for the $5 million treatment condition and the control group. The standard deviation for the $200,000 and the $1 million treatment groups are lower than those for the control group and the $5 million group. An interpretation of these findings is that those in the control group, not provided with guidance, have vastly differing views of how much the average CEO is paid and should be paid. And, to speculate, despite receiving guidance on what the average CEO is paid in the industry, some respondents in the $5 million treatment group find this pay too high and therefore choose not to align their responses with this anchor while others do so. 3 I thank an anonymous reviewer for this point. 4 I thank an anonymous reviewer for this point. REFERENCES Adams J. Stacy 1965 . “ Inequity in Social Exchange .” in Advances in Experimental Social Psychology Vol. 2 , edited by Berkowitz L. New York Academic Press . Google Scholar CrossRef Search ADS Alvaredo Facundo , Atkinson Anthony , Piketty Thomas , Saez Emmanuel 2011 . The World Top Incomes Database . Retrieved December 2015 (http://topincomes.g-mond.parisschoolofeconomics.eu/). Alves Wayne M. , Rossi Peter H. 1978 . “ Who Should Get What? Fairness Judgments of the Distribution of Earnings .” American Journal of Sociology 84 3 : 541 - 64 . Google Scholar CrossRef Search ADS Ansolabehere Stephen , Schaffner Brian F. 2011 . “ Re-Examining the Validity of Different Survey Modes for Measuring Public Opinion in the U.S.: Findings from a 2010 Multi-Mode Comparison .” Pp. 12 - 15 in AAPOR Annual Conference Phoenix, AZ . Ansolabehere Stephen , Rivers Douglas 2013 . “ Cooperative Survey Research .” Annual Review of Political Science 16 : 307 - 29 . Google Scholar CrossRef Search ADS Associated Press 2014 . “ Median Pay for CEOs Rises above $10 Million in 2013 .” in Los Angeles Times. May 27, 2014 . Ayres Ian , Edlin Aaron S. 2011 . “ Don’t Tax the Rich. Tax Inequality Itself’ .” Pp. A29 in New York Times. Dec 19, 2011 . Bakija Jon , Cole Adam , Heim Bradley T. 2012 . “ Jobs and Income Growth of Top Earners and the Causes of Changing Income Inequality: Evidence from U.S. Tax Return Data .” Working Paper. Williams College . Bartels Larry M. 2005 . “ Homer Gets a Tax Cut: Inequality and Public Policy in the American Mind .” Perspectives on Politics 3 01 : 15 - 31 . Google Scholar CrossRef Search ADS Bebchuk Lucian , Grinstein Yaniv 2005 . “ The Growth of Executive Pay .” Oxford Review of Economic Policy 21 2 : 283 - 303 . Google Scholar CrossRef Search ADS Bebchuk Lucian A. , Fried Jesse M. 2005 . “ Pay without Performance: Overview of the Issues .” Journal of Applied Corporate Finance 17 4 : 8 - 23 . Google Scholar CrossRef Search ADS Bebchuk Lucian A. , Fried Jesse M. 2010 . “ Tackling the Managerial Power Problem: The Key to Improving Executive Compensation .” Pathways Magazine ( Summer ): 9 - 13 . Bies Robert J. , Moag Joseph S. 1986 . “ Interactional Justice: Communication Criteria of Fairness .” Pp. 43 - 55 in Research on Negotiation in Organizations Vol. 1 , edited by Lewicki R. J. , Sheppard B. H. , Bazerman M. H. Greenwich, CT JAI Press . Booth William James 1994 . “ On the Idea of the Moral Economy .” The American Political Science Review 88 3 : 653 - 67 . Google Scholar CrossRef Search ADS Bowman Karlyn H. 2000 . “ No Din over Dollar Divide .” in Chicago Sun-Times. July 31, 2000 . Card David , Mas Alexandre , Moretti Enrico , Saez Emmanuel 2012 . “ Inequality at Work: The Effect of Peer Salaries on Job Satisfaction .” The American Economic Review 102 6 : 2981 - 3003 . Google Scholar CrossRef Search ADS Cervone Daniel , Peake Philip K. 1986 . “ Anchoring, Efficacy, and Action: The Influence of Judgmental Heuristics on Self-Efficacy Judgments and Behavior .” Journal of Personality and Social Psychology 50 3 : 492 - 501 . Google Scholar CrossRef Search ADS Chapman Gretchen B. , Johnson Eric J. 2002 . “ Incorporating the Irrelevant: Anchors in Judgments of Belief and Value .” Pp. 120 - 38 in Heuristics and Biases: The Psychology of Intuitive Judgment , edited by Gilovich T. , Griffin Dale , Kahneman Daniel New York Cambridge University Press . Google Scholar CrossRef Search ADS Crystal Graef S. 1991 In Search of Excess: The Overcompensation of American Executives New York Norton . DeCarlo Scott 2008 . “ America’s Best Big Companies .” Forbes . Deutsch Morton 1975 . “ Equity, Equality, and Need: What Determines Which Value Will Be Used as the Basis of Distributive Justice? ”. Journal of Social Issues 31 3 : 137 - 49 . Google Scholar CrossRef Search ADS DiPrete Thomas A. , Eirich Gregory M. , Pittinsky Matthew 2010 . “ Compensation Benchmarking, Leapfrogs, and the Surge in Executive Pay .” American Journal of Sociology 115 6 : 1671 - 712 . Google Scholar CrossRef Search ADS Eagly Alice H. , Chaiken Shelly 1993 The Psychology of Attitudes Fort Worth, TX Harcourt Brace Jovanovich . Eagly Alice H. , Carli Linda L. 2007 Through the Labyrinth: The Truth About How Women Become Leaders Boston, MA Harvard Business School Press . Elster Jon 1989 . “ Social Norms and Economic Theory .” The Journal of Economic Perspectives 3 4 : 99 - 117 . Google Scholar CrossRef Search ADS Etzioni Amitai. 1988 The Moral Dimension: Toward a New Economics New York Free Press . Evans M.D.R. , Kelley Jonathan , Peoples Clayton D. 2010 . “ Justifications of Inequality: The Normative Basis of Pay Differentials in 31 Nations .” Social Science Quarterly 91 5 : 1405 - 31 . Google Scholar CrossRef Search ADS Fishbein Martin , Ajzen Icek 1975 Belief, Attitude, Intention and Behavior: An Introduction to Theory and Research Reading, MA Addison-Wesley . Folger Robert , Greenberg Jerald 1985 . “ Procedural Justice: An Interpretive Analysis of Personnel Systems .” Research in Personnel and Human Resources Management 3 : 141 - 83 . Fourcade Marion , Healy Kieran 2007 . “ Moral Views of Market Society .” Annual Review of Sociology 33 : 285 - 311 . Google Scholar CrossRef Search ADS Frank R. H. , Cook P. J. 1995 The Winner-Take-All Society: Why the Few at the Top Get So Much More Than the Rest of Us New York Free Press . Frydman Carola , Jenter Dirk 2010 . “ CEO Compensation .” Annual Review of Financial Economics 2 1 : 75 - 102 . Google Scholar CrossRef Search ADS Gabaix Xavier , Landier Augustin 2008 . “ Why Has CEO Pay Increased So Much? ”. The Quarterly Journal of Economics 123 1 : 49 - 100 . Google Scholar CrossRef Search ADS Goode William J. 1978 The Celebration of Heroes: Prestige as a Social Control System Berkeley University of California Press . Green Donald , Jacowitz Karen E. , Kahneman Daniel , McFadden Daniel 1998 . “ Referendum Contingent Valuation, Anchoring, and Willingness to Pay for Public Goods .” Resource and Energy Economics 20 2 : 85 - 116 . Google Scholar CrossRef Search ADS Greenberg Jerald. 1987 . “ A Taxonomy of Organizational Justice Theories .” Academy of Management Review 12 1 : 9 - 22 . Google Scholar CrossRef Search ADS Grusky David B. , Wimer Christopher 2010 . “ Is There Too Much Inequality? ” Pp. 3 - 10 in The Inequality Puzzle: European and U.S. Leaders Discuss Rising Income Inequality , edited by Berger R. , Grusky D. B. , Raffel T. , Samuels G. , Wimer C. New York Springer . Google Scholar CrossRef Search ADS Hacker Jacob S. , Pierson Paul 2010 Winner-Take-All Politics: How Washington Made the Rich Richer–and Turned Its Back on the Middle Class New York, NY Simon and Schuster . Hall Brian J. , Liebman Jeffrey B. 1998 . “ Are CEOs Really Paid Like Bureaucrats? ”. The Quarterly Journal of Economics 113 3 : 653 - 91 . Google Scholar CrossRef Search ADS Hegtvedt Karen , Cook Karen 2001 . “ Distributive Justice: Recent Theoretical Developments and Applications .” Pp. 93 - 132 in Handbook of Justice Research in Law , edited by Sanders I. J. , Hamilton V. L. New York Kluwer . Hirschman Albert O. 1984 . “ Against Parsimony: Three Easy Ways of Complicating Some Categories of Economic Discourse .” Bulletin of the American Academy of Arts and Sciences 37 8 : 11 - 28 . Google Scholar CrossRef Search ADS Hochschild Jennifer L. 1981 What’s Fair? Americans’ Attitudes toward Distributive Justice Cambridge, MA Harvard University Press . Hochschild Jennifer L. 1996 Facing up to the American Dream: Race, Class, and the Soul of the Nation Princeton, NJ Princeton University Press . Homans George C. 1974 Social Behavior: Its Elementary Forms New York Harcourt, Brace, Jovanovich . Huber Joan , Humbert Form William 1973 Income and Ideology: An Analysis of the American Political Formula New York Free Press . Jacowitz Karen E. , Kahneman Daniel 1995 . “ Measures of Anchoring in Estimation Tasks .” Personality and Social Psychology Bulletin 21 : 1161 - 66 . Google Scholar CrossRef Search ADS Jasso Guillermina , Meyersson Milgrom Eva M. 2008 . “ Distributive Justice and CEO Compensation .” Acta Sociologica 51 2 : 123 - 43 . Google Scholar CrossRef Search ADS Jensen Michael , Murphy Kevin , Wruck Eric 2004 . “ Remuneration: Where We’ve Been, How We Got to Here, What Are the Problems, and How to Fix Them .” Working Paper, Harvard University . Google Scholar CrossRef Search ADS Kaplan Steven N. 2008 . “ Are U.S. CEOs Overpaid? ”. The Academy of Management Perspectives 22 2 : 5 - 20 . Google Scholar CrossRef Search ADS Kaplan Steven N. , Rauh Joshua 2010 . “ Wall Street and Main Street: What Contributes to the Rise in the Highest Incomes? ”. Review of Financial Studies 23 3 : 1004 - 50 . Google Scholar CrossRef Search ADS Kiatpongsan Sorapop , Norton Michael I. 2014 . “ How Much (More) Should CEOs Make? A Universal Desire for More Equal Pay .” Perspectives on Psychological Science 9 : 587 - 93 . Google Scholar CrossRef Search ADS Kim Jerry W. , Kogut Bruce , Yang Jae-Suk 2015 . “ Executive Compensation, Fat Cats, and Best Athletes .” American Sociological Review 80 2 : 299 - 328 . Google Scholar CrossRef Search ADS Kluegel James R. , Smith Eliot R. 1986 Beliefs About Inequality: Americans’ Views of What Is and What Ought to Be Hawthorne, NY Aldine de Gruyter . Kuhnen Camelia M. , Niessen Alexandra 2012 . “ Public Opinion and Executive Compensation .” Management Science 58 7 : 1249 - 72 . Google Scholar CrossRef Search ADS Ladd Everett Carll 1994 The American Ideology: An Exploration of the Origins, Meaning, and Role of American Political Ideas Storrs, CT Roper Center for Public Opinion Research . Ladd Everett Carll , Bowman Karlyn H. 1998 Attitudes toward Economic Inequality Washington, DC Aei Press . Lamm Helmut , Schwinger Thomas 1980 . “ Norms Concerning Distributive Justice: Are Needs Taken into Consideration in Allocation Decisions? ” Social Psychology Quarterly 43 : 425 - 29 . Google Scholar CrossRef Search ADS Lane Robert E. 1962 Political Ideology: Why the American Common Man Believes What He Does New York The Free Press . Lane Robert E. 1986 . “ Market Justice, Political Justice .” The American Political Science Review 80 : 383 - 402 . Google Scholar CrossRef Search ADS Leung Kwok , Park Hun-Joon 1986 . “ Effects of Interactional Goal on Choice of Allocation Rule: A Cross-National Study .” Organizational Behavior and Human Decision Processes 37 1 : 111 - 20 . Google Scholar CrossRef Search ADS Leventhal Gerald S. 1976 . “ The Distribution of Rewards and Resources in Groups and Organizations .” Advances in Experimental Social Psychology 9 : 91 - 131 . Google Scholar CrossRef Search ADS Mankiw N. Gregory 2010 . “ Spreading the Wealth Around: Reflections Inspired by Joe the Plumber .” Eastern Economic Journal 36 3 : 285 - 98 . Google Scholar CrossRef Search ADS Mansbridge Jane J. 1990 Beyond Self-Interest Chicago University of Chicago Press . Markovsky Barry 1988 . “ Anchoring Justice .” Social Psychology Quarterly 51 3 : 213 - 24 . Google Scholar CrossRef Search ADS Marshall Gordon , Swift Adam , Routh David , Burgoyne Carole 1999 . “ What Is and What Ought to Be: Popular Beliefs About Distributive Justice in Thirteen Countries .” European Sociological Review 15 4 : 349 - 67 . Google Scholar CrossRef Search ADS Mas Alexandre 2014 . “ Does Transparency Lead to Pay Compression? ”. NBER Working Paper . McCall Leslie. 2003 . “ Do They Know and Do They Care? Americans’ Awareness of Rising Inequality .” Working Paper . McCall Leslie , Kenworthy Lane 2009 . “ Americans’ Social Policy Preferences in the Era of Rising Inequality .” Perspectives on Politics 7 3 : 459 - 84 . Google Scholar CrossRef Search ADS McCall Leslie 2013 The Undeserving Rich: American Beliefs About Inequality, Opportunity, and Redistribution New York Cambridge University Press . Google Scholar CrossRef Search ADS Messick David M. , Cook Karen S. 1983 Equity Theory: Psychological and Sociological Perspectives New York Praeger . Miller David 1992 . “ Distributive Justice: What the People Think .” Ethics 102 3 : 555 - 93 . Google Scholar CrossRef Search ADS Miller David. 1999 Principles of Social Justice Cambridge, MA Harvard University Press . Mishel Lawrence , Bivens Josh , Gould Elise , Shierholz Heidi 2012 The State of Working America Cornell University Press . Mitchell Gregory , Tetlock Philip E. , Mellers Barbara A. , Ordonez Lisa D. 1993 . “ Judgments of Social Justice: Compromises between Equality and Efficiency .” Journal of Personality and Social Psychology 65 4 : 629 . Google Scholar CrossRef Search ADS Mutz Diana C. 2011 Population-Based Survey Experiments Princeton, NJ Princeton University Press . Google Scholar CrossRef Search ADS Northcraft Gregory B. , Neale Margaret A. 1987 . “ Experts, Amateurs, and Real Estate: An Anchoring-and-Adjustment Perspective on Property Pricing Decisions .” Organizational Behavior and Human Decision Processes 39 1 : 84 - 97 . Google Scholar CrossRef Search ADS Norton Michael I. , Ariely Dan 2011 . “ Building a Better America—One Wealth Quintile at a Time .” Perspectives on Psychological Science 6 1 : 9 - 12 . Google Scholar CrossRef Search ADS Osberg Lars , Smeeding Timothy 2006 . “ “Fair” Inequality? Attitudes toward Pay Differentials: The United States in Comparative Perspective .” American Sociological Review 71 3 : 450 - 73 . Google Scholar CrossRef Search ADS Page Benjamin I. , Jacobs Lawrence R. 2009 Class War? What Americans Really Think About Economic Inequality Chicago University of Chicago Press . Google Scholar CrossRef Search ADS Piketty Thomas , Saez Emmanuel 2006 . “ The Evolution of Top Incomes: A Historical and International Perspective .” The American Economic Review 96 2 : 200 - 05 . Google Scholar CrossRef Search ADS Piketty Thomas 2014 Capital in the Twenty-First Century Cambridge, MA The Belknap Press . Google Scholar CrossRef Search ADS Piketty Thomas , Saez Emmanuel , Stantcheva Stefanie 2014 . “ Optimal Taxation of Top Labor Incomes: A Tale of Three Elasticities .” American Economic Journal: Economic Policy 6 1 : 230 - 71 . Google Scholar CrossRef Search ADS Prentice Deborah A. , Crosby Faye 1987 . “ The Importance of Context for Assessing Deservingness .” Pp. 165 - 82 in Social Comparison, Social Justice, and Relative Deprivation: Theoretical, Empirical, and Policy Perspectives , edited by Masters J. C. , Smith W. P. Hillsdale, NJ Lawrence Erlbaum . Reich Robert B. 2010 Aftershock: The Next Economy and America’s Future New York Vintage . Reich Robert B. 2012 Beyond Outrage: What Has Gone Wrong with Our Economy and Our Democracy, and How to Fix It New York Vintage . Reich Robert B. 2015 Saving Capitalism: For the Many, Not the Few New York Knopf . Rivers Douglas 2006 . “ Sample Matching: Representative Sampling from Internet Panels .” Polimetrix White Paper Series . Rivers Douglas , Bailey Delia 2009 . “ Inference from Matched Samples in the 2008 U.S. National Elections .” Pp. 627 - 39 in Proceedings of the Joint Statistical Meetings . Rivers Douglas 2012 . “ Yougov Poll Performance in the 2012 U.S. Elections ”. Retrieved December 15, 2013 (http://today.yougov.com/news/2012/11/07/yougov-poll-performance-2012-us-elections/) . Rosen Sherwin. 1981 . “ The Economics of Superstars .” The American Economic Review 71 : 845 - 58 . Rosenberg Milton J. 1960 . “ An Analysis of Affective-Cognitive Consistency .” Pp. 15 - 64 in Attitude Organization and Change , edited by Hovland C. I. , Rosenberg M. J. New Haven, CT Yale University Press . Rosenfeld Jake , Denice Patrick 2015 . “ The Power of Transparency Evidence from a British Workplace Survey .” American Sociological Review 80 : 1045 - 68 . Google Scholar CrossRef Search ADS Sandberg Sheryl 2013 Lean In: Women, Work, and the Will to Lead New York Random House . Sayer Andrew. 2005 The Moral Significance of Class New York Cambridge University Press . Google Scholar CrossRef Search ADS Schlozman Kay Lehman , Verba Sidney 1979 Injury to Insult: Unemployment, Class, and Political Response Cambridge, MA Harvard University Press . Schwarz Norbert , Bless Herbert 1992 . “ Constructing Reality and Its Alternatives: An Inclusion/Exclusion Model of Assimilation and Contrast Effects in Social Judgment .” Pp. 217 - 45 in The Construction of Social Judgments , edited by Martin L. L. , Tresser A. Hillsdale, NJ Lawrence Erlbaum . Scott John T. , Matland Richard E. , Michelbach Philip A. , Bornstein Brian H. 2001 . “ Just Deserts: An Experimental Study of Distributive Justice Norms .” American Journal of Political Science 45 : 749 - 67 . Google Scholar CrossRef Search ADS Securities and Exchange Commission 2015 . “ Pay Ratio Disclosure .” Vol. 17 CFR Parts 229 and 249 . Silver Nate 2012 . “ Which Polls Fared Best (and Worst) in the 2012 Presidential Race .” in International New York Times. November 10th . Skitka Linda J. , Bauman Christopher W. , Mullen Elizabeth 2008 . “ Morality and Justice: An Expanded Theoretical Perspective and Review .” Pp. 1 - 27 in Advances in Group Processes Vol. 25 , edited by Hedgvedt K. A. , Clay-Warner J. Bingley, UK Emerald Group Publishing Limited . Google Scholar CrossRef Search ADS Sniderman Paul M. , Grob Douglas B. 1996 . “ Innovations in Experimental Design in Attitude Surveys .” Annual Review of Sociology 22 : 377 - 99 . Google Scholar CrossRef Search ADS Stewart G. Bennett 1990 . “ Remaking the Corporation from Within .” Harvard Business Review 68 4 : 126 - 37 . Stiglitz Joseph E. 2003 The Roaring Nineties: A New History of the World’s Most Prosperous Decade New York, NY W.W. Norton & Company, Inc . Stiglitz Joseph E. 2012 The Price of Inequality: How Today’s Divided Society Endangers Our Future New York, NY W.W. Norton & Company, Inc. Svallfors Stefan 2007 . “ Class and Attitudes to Market Inequality .” in The Political Sociology of the Welfare State , edited by Svallfors S. Stanford, CA Stanford University Press . Google Scholar CrossRef Search ADS Switzer Fred S. , Sniezek Janet A. 1991 . “ Judgment Processes in Motivation: Anchoring and Adjustment Effects on Judgment and Behavior .” Organizational Behavior and Human Decision Processes 49 2 : 208 - 29 . Google Scholar CrossRef Search ADS Thibaut John W. , Walker Laurens 1975 Procedural Justice: A Psychological Analysis L. Erlbaum Associates. Thompson E. P. 1971 . “ The Moral Economy of the English Crowd in the Eighteenth Century .” Past & Present 50 : 76 - 136 . Tversky Amos , Kahneman Daniel 1974 . “ Judgment under Uncertainty: Heuristics and Biases .” Science 185 4157 : 1124 - 31 . Google Scholar CrossRef Search ADS United States Department of Labor, Bureau of Labor Statistics 2012 . “ Occupational Employment Statistics: Occupational Employment and Wages .” Vavreck Lynn , Iyengar Shanto 2013 . “ Online Panels and Experimentation .” Pp. 156 - 68 in The Oxford Handbook of American Public Opinion and the Media , edited by Shapiro R. Y. , Jacobs L. R. OUP Oxford . Verba Sidney , Orren Gary R. 1985 Equality in America: The View from the Top Cambridge, MA Harvard University Press . Walster Elaine , Berscheid Ellen , William Walster G. 1973 . “ New Directions in Equity Research .” Journal of Personality and Social Psychology 25 2 : 151 - 76 . Google Scholar CrossRef Search ADS Walster Elaine , William Walster G. 1975 . “ Equity and Social Justice .” Journal of Social Issues 31 3 : 21 - 43 . Walster Elaine , William Walster G. , Berscheid Ellen 1978 Equity: Theory and Research Boston Allyn and Bacon Walzer Michael 1983 Spheres of Justice: A Defense of Pluralism and Equality New York Basic Books . Weeden Kim A. , Grusky David B. 2014 . “ Inequality and Market Failure .” American Behavioral Scientist 58 3 : 473 - 91 . Google Scholar CrossRef Search ADS Wilson Timothy D. , Houston Christopher E. , Etling Kathryn M. , Brekke Nancy 1996 . “ A New Look at Anchoring Effects: Basic Anchoring and Its Antecedents .” Journal of Experimental Psychology 125 4 : 387 - 402 . Google Scholar CrossRef Search ADS Wright William F. , Anderson Urton 1989 . “ Effects of Situation Familiarity and Financial Incentives on Use of the Anchoring and Adjustment Heuristic for Probability Assessment .” Organizational Behavior and Human Decision Processes 44 1 : 68 - 82 . Google Scholar CrossRef Search ADS Young Cristobal , Varner Charles , Lurie Ithai Z. , Prisinzano Richard 2016 . “ Millionaire Migration and Taxation of the Elite Evidence from Administrative Data .” American Sociological Review 81 3 : 421 - 46 . Google Scholar CrossRef Search ADS Zanna Mark P. , Rempel John K. 1988 . “ Attitudes: A New Look at an Old Concept .” Pp. 315 - 34 in The Social Psychology of Knowledge , edited by Bar-Tal D. , Kruglanski A. New York Cambridge University Press . © The Author 2017. Published by Oxford University Press on behalf of the Society for the Study of Social Problems.All rights reserved. For permissions, please e-mail: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Social Problems Oxford University Press

Is The Sky the Limit? Fair Executive Pay as Performance Rises

Social Problems , Volume Advance Article (2) – Mar 2, 2017

Loading next page...
 
/lp/ou_press/is-the-sky-the-limit-fair-executive-pay-as-performance-rises-ZfDC90quUR
Publisher
University of California Press
Copyright
© The Author 2017. Published by Oxford University Press on behalf of the Society for the Study of Social Problems.All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
ISSN
0037-7791
eISSN
1533-8533
D.O.I.
10.1093/socpro/spw050
Publisher site
See Article on Publisher Site

Abstract

Abstract Recent research documents the public discontent with high income inequality yet an important limitation in our understanding is why such high pay is problematic according to many Americans. This study juxtaposes two explanations for the discontent that many Americans have shown regarding high Chief Executive Officer (CEO) pay: a) that the American discontent with extremely high CEO pay stems mainly from a belief that CEOs are not contributing highly enough to merit such high pay and b) the American discontent with extremely high pay is due to such pay being objectionable in principle. Using data from a national survey experiment (N = 989) uniquely designed to test these propositions, this study evaluates the relationship between performance and fair CEO pay. Respondents show aversion to high CEO pay while also embracing the principle of proportionality of rewards and contributions. Findings suggest that discontent with extremely high pay may be reconciled with support for pay for performance: the slope of the pay and performance function may be at issue rather than a hard limit on pay. income inequality, distributive justice, executive compensation, attitudes, fair pay Many Americans are aware of the high levels of income inequality and are displeased with it (Bartels 2005; McCall 2003; McCall and Kenworthy 2009; McCall 2013; Norton and Ariely 2011; Osberg and Smeeding 2006; Page and Jacobs 2009). With the rise in income inequality at the top of the income distribution in the past decades and with the Great Recession of 2008, the media and the American public turned their attention to high earners. Some scholars argued that the Great Recession of 2008 may have called into question the responsiveness of the market to effort. On the one hand, heads of financial companies were largely blamed for the recession undermining the belief that their high rewards were well-deserved. On the other hand, many hard working people were laid off, showing that the market does not reward all who work hard (Grusky and Wimer 2010). Among those in the highest-paid occupations, Chief Executive Officers (CEOs) have been under a considerable amount of criticism even long before the Great Recession of 2008 (Crystal 1991). Research on newspaper coverage about CEO compensation shows that between 1990 and 2010 over 26,000 news pieces were published on the subject with only a handful having a positive tone on the issue (Kuhnen and Niessen 2012). While a growing number of studies document the American discontent with extremely high incomes, an important limitation in our understanding of attitudes toward extremely high pay is why many Americans have a strong negative response to high CEO pay. Americans have long been known for their belief in economic individualism and free market and for tolerating income inequality under fair conditions (Bowman 2000; Hochschild 1996; Huber and Form 1973; Kluegel and Smith 1986; Ladd 1994; Ladd and Bowman 1998; Lane 1986; Schlozman and Verba 1979; Verba and Orren 1985). Why do large numbers of Americans find very high CEO pay problematic? Many Americans find the highest paid occupations, among them executives, overpaid (Kiatpongsan and Norton 2014; Kluegel and Smith 1986; McCall 2013). According to polls from a variety of sources between 1976 and 2011, more than 72% of Americans routinely found CEOs overpaid (McCall 2013). Recent research shows that extremely high pay is becoming more contentious (Osberg and Smeeding 2006) and problematic (Bartels 2005; McCall 2003; McCall 2013; Page and Jacobs 2009). Two possible explanations for the discontent with income inequality are especially noteworthy. The first is an account which emphasizes that the public views the current high earners as undeserving of their high pay. People may believe that the high earners are not contributing proportionally to their rewards. For example, some scholars suggest that some form of rent-seeking explains the pay that CEOs receive (Bebchuk and Grinstein 2005; Bebchuk and Fried 2005; DiPrete, Eirich and Pittinsky 2010; Piketty and Saez 2006; Weeden and Grusky 2014). Recent research shows that the pay-setting process is influenced by the CEO’s localized ties, especially board networks (Kim, Kogut and Yang 2015). If CEOs are perceived as overpaid in relation to their contributions (whether through rent-seeking or through any means by which a disconnect occurs between pay and performance), then, it is possible that were they understood as contributing proportionally highly, the American public might be more inclined to accept their extremely high pay. A second possibility is that the American public finds such extremely high pay and extreme inequality objectionable in principle. Research examining the moral economy shows the importance of morality and fairness in evaluating economic activities (Booth 1994; Elster 1989; Etzioni 1988; Fourcade and Healy 2007; Hirschman 1984; Mansbridge 1990; Sayer 2005; Svallfors 2007; Thompson 1971). Based on this proposition, even if high-earners were understood as contributing proportionally to such extreme levels of pay, people would still feel uneasy about these high levels of pay. Using data from a national survey experiment uniquely designed to test these propositions, this study aims to advance our understanding about the reasons for the negative public response to high CEO pay. Do Americans believe that the better a CEO performs, the more he or she should earn, without a limit? Alternatively, is there a limit to acceptable pay at the top? background and conceptual framework Almost 60% of the top 1% of incomes in the U.S. come from various types of employment income including wages, salaries, pensions, exercised stock options and bonuses (Piketty and Saez 2006). A large percentage of top income earners are executives, managers, supervisors, and financial professionals. People in these occupations make up 60 percent of the top 0.1 percent of income earners (Bakija, Cole and Heim 2012). In addition, not only is the pay for executives very high but it also has seen a great increase in the past twenty-five years (Bebchuk and Fried 2005; Hall and Liebman 1998; Jensen, Murphy and Wruck 2004). Median CEO compensation in S&P 500 companies was between $6 million and $7 million throughout the 2000s (Frydman and Jenter 2010). Although variations in its formulations exist, attitudes are understood as tendencies to evaluate an attitude object (Eagly and Chaiken 1993) or the “categorization of a stimulus object along an evaluative dimension” (Zanna and Rempel 1988). Based on expectancy-value models of attitudes (Fishbein and Ajzen 1975; Rosenberg 1960), attitudes are composed of the sum of various relevant beliefs about an object each weighed by an evaluation (for example, favorable or unfavorable) of the relevant belief. Based on these formulations, attitudes toward CEO pay are not based on a single belief, but multiple beliefs and their evaluations. In the next section, I discuss some of the potentially salient considerations in the formation of attitudes about CEO pay. Justice Research and Equity Theory Research in the past decades has shown support for the idea that, in market settings, people prefer to allocate rewards proportionally to relevant inputs such as effort and performance (Hegtvedt 2001; Miller 1992). People weigh factors differently in deciding how goods and rewards should be distributed under different circumstances (Deutsch 1975; Lamm and Schwinger 1980; Leventhal 1976; Miller 1999; Scott et al. 2001; Walzer 1983). It is not always performance that counts heaviest in allocating rewards, however, most people believe that fair pay should be proportional to contribution in the workplace (Evans, Kelley and Peoples 2010; Hegtvedt 2001; Hochschild 1981; Lane 1962; Leung and Park 1986; Leventhal 1976; Marshall et al. 1999; Miller 1992; Miller 1999; Mitchell et al. 1993; Prentice and Crosby 1987; Scott et al. 2001). The importance placed on contribution, and especially performance, is not unique to Americans and is widespread across countries as well as across strata within countries (Evans, Kelley and Peoples 2010; Hochschild 1981; Lane 1962; Marshall et al. 1999). A long research tradition has examined people’s justice perceptions. Major constructs of interest are distributive justice which concerns itself with perceived fairness of outcomes of allocation decisions (Adams 1965; Homans 1974; Walster, Berscheid and Walster 1973), procedural justice which emphasizes the fairness in procedures through which those allocations are made (Thibaut and Walker 1975), and interactional justice which emphasizes the interpersonal playout of that allocation (Bies and Moag 1986). Most recently, justice research has made considerable advances in organizational settings (Folger and Greenberg 1985; Greenberg 1987). The principle of proportionality of rewards and contributions is central to theories of distributive justice. Based on equity theory, perceived fairness depends on workers having equivalent ratios of inputs and outcomes (Adams 1965; Messick and Cook 1983; Walster and Walster 1975; Walster, Walster and Berscheid 1978). To illustrate, “…if two equally talented students work equally hard and do an equally good job, they are entitled to equal salaries… If one of the students works twice as hard as the other and thus does twice a good job, he is entitled to twice the salary” (Walster, Walster and Berscheid 1978). However, William Goode (1978) argues that the rule of proportionality construed as a constant ratio of rewards and contributions seems incorrect, particularly at the lowest and highest levels of performance. Instead, he posits that where along the reward and contributions curve an individual is greatly affects returns. Goode (1978) writes, an “important deficiency in the proportionality rule is that what is viewed as a justly larger or smaller reward in money or prestige actually varies in a curvilinear way with output or contribution.” For example, for high performers, rewards rise faster than contributions (for winner-take-all markets, see for example, Rosen 1981 and Frank and Cook 1995). Other arguments in favor of the nonlinearity of the rule of proportionality have also been advanced. Research in the moral economy shows that economic decisions and actions are often motivated by norms and morality (Booth 1994; Elster 1989; Etzioni 1988; Fourcade and Healy 2007; Hirschman 1984; Mansbridge 1990; Sayer 2005; Svallfors 2007; Thompson 1971). While previous research has focused on homo economicus and later on homo socialis, a new conception of homo moralis is emerging (Skitka, Bauman and Mullen 2008). Some evidence shows that there may be a morally-grounded maximum level of acceptable pay in the workplace. This upper limit to ethically acceptable pay has been called a “social maximum” (Alves and Rossi 1978) or an “ethical ceiling” (Osberg and Smeeding 2006). Based on findings from their vignette study, Wayne Alves and Peter Rossi (1978) have argued that some earnings are too high to be justified by an earner’s education or occupation. If indeed there is an ethical ceiling to pay, it could be the source of the discontent with high CEO pay. In light of this body of work, there is need for empirical research on the rule of proportionality focusing on the top of the income distribution and on current levels of CEO pay. This study offers exactly such a test. Marginal Productivity Theory and CEO Compensation Robert Frank and Philip Cook (1995) write that the theory of marginal productivity which aims to explain how workers are remunerated was not intended as a tool to evaluate fairness in the distribution of incomes and yet people have come to see fairness in pay that reflects marginal product. Whether CEO pay reflects marginal product is a big debate among scholars of CEO compensation. Some scholars argue that CEO pay, however high it may be, is a reflection of the CEO’s performance and a competitive market for talent (Frank and Cook 1995; Gabaix and Landier 2008; Kaplan 2008; Kaplan and Rauh 2010; Mankiw 2010). Other scholars have tried to disabuse the public of the idea that marginal productivity theory explains the dramatic rise in executive pay (Piketty 2014; Reich 2010; Reich 2012; Reich 2015; Stiglitz 2003; Stiglitz 2012). They have argued that it is difficult to estimate the marginal product of a CEO in a large organization due to the team-based nature of their work and that the argument that marginal product determines CEO pay is made further problematic by data showing cross-country differences in CEO pay among industrialized countries (Piketty 2014). Some scholars have argued that CEO pay follows a pay-for-luck model where CEO pay rises with general market trends (Reich 2012; Stiglitz 2012) or that holes in corporate governance drive the high pay (Bebchuk and Grinstein 2005; Bebchuk and Fried 2005; Bebchuk 2010; DiPrete, Eirich and Pittinsky 2010; Piketty and Saez 2006). For example, Thomas Piketty, Emmanuel Saez and Stefanie Stantcheva (2014) put forth a compensation bargaining model showing that CEO pay depends on top marginal tax rates as CEOs presumably invest more in bargaining for higher pay when tax rates are low. While CEO pay is highly contested, the question that receives most attention is how closely CEO pay is tied to performance and how this alignment can be improved. For example, some scholars have argued in favor of a pay schedule that changes linearly with performance without a lower or upper limit to pay (Jensen, Murphy and Wruck 2004; Stewart 1990). The aim of this paper is not to weigh in on the debate regarding whether actual CEO pay reflects marginal product. Instead, the question addressed here is whether public discontent would decrease if CEO pay were indeed tightly tied to performance. Based on previous work, attitudes toward CEO pay are conceptualized as evaluations based on beliefs about the principle of equity, a CEO’s perceived marginal product, and the morality of high inequality. This study uses a unique set of data from a national survey experiment to examine two explanations for the American discontent with extremely high CEO pay: a) the American public shows discontent because CEOs are not seen as contributing proportionally to their pay b) the American discontent with extremely high pay is due to such pay being objectionable in principle. Empirical Expectations The central empirical expectation is that the performance and perceived fair pay function is not a linearly increasing function. Instead, fair pay will rise less than proportionally to performance at very high levels of pay thereby showing diminishing returns to performance. H1 (inequality aversion hypothesis): fair CEO pay will not scale linearly with performance. H0 (rent-seeking aversion hypothesis): fair CEO pay will scale linearly with performance. Beyond the central hypotheses, additional nuances in findings are expected. Considerations about fair pay are influenced by perceived actual pay, whether perceptions reflect reality or not (Osberg and Smeeding 2006). The actual hierarchy of occupations is replicated by most respondents when they are asked about just pay for occupations (Kluegel and Smith 1986). Prior research also shows that in responding to questions people are likely to gravitate towards responses hinted at them by the researchers even if these are arbitrary numbers. This phenomenon called anchoring shows large and consistent effects (Cervone and Peake 1986; Chapman and Johnson 2002; Green et al. 1998; Markovsky 1988; Northcraft and Neale 1987; Schwarz and Bless 1992; Switzer and Sniezek 1991; Tversky and Kahneman 1974; Wilson et al. 1996; Wright and Anderson 1989). I use the term anchoring to refer to a “starting point for adjustment” (Jacowitz and Kahneman 1995). In this type of anchoring, respondents are given a numerical idea of a possible response and are asked their opinion using that same numerical scale. For example, when respondents are told that a reward of a certain dollar amount was given in appreciation for a deed and are later asked how much to reward someone in a similar situation, they are likely to gravitate towards this anchor (Markovsky 1988). CEO pay ranges widely. According to data from the Bureau of Labor Statistics (2012), mean annual CEO pay was approximately $200,000 while it is much higher in Standard & Poor’s 500 companies (Frydman and Jenter 2010). Therefore, a low anchor such as $200,000 and a high anchor such as $5 million are both realistic amounts. Based on previous research showing strong and consistent anchoring effects, I expect that preferences for fair CEO pay will be affected by what respondents are told an average CEO is usually paid in an industry. However, despite the strong anchoring effects found in earlier research, I expect that respondents will not uncritically follow that suggested pay. I hypothesize that people have beliefs about what an appropriate amount of pay would be for the CEO of a large company. While the anchor will have an effect, responses will also gravitate towards this amount. Empirically, I test the proposition that the higher people are told the average actual CEO pay is, the more they will size it down to yield fair pay. H2a: the higher the anchor for average CEO pay, the higher the fair pay reported. H2b: the higher the anchor for average CEO pay, the greater the downward adjustment of fair pay. We should observe diminishing returns especially at higher levels of pay because the higher the level of pay, the closer the pay will be to any one person’s ethical pay ceiling, if there is one. Therefore, responses from respondents in the treatment condition where the anchor for CEO pay is highest should produce the function with the most evident diminishing returns. H2c: the higher the anchor for average CEO pay, the greater the curvilinearity of the fair pay and performance function. Americans are well-known for their strong belief in a free market and in economic individualism (Hochschild 1996; Huber and Form 1973; Kluegel and Smith 1986; Ladd 1994; Ladd and Bowman 1998; Schlozman and Verba 1979). Some scholars find that American public opinion about inequality is characterized by polarization (Osberg and Smeeding 2006). I expect to find differences in attitudes between people who strongly believe in the free market ideology and those who do not. Specifically, I expect to find that those who support the free market strongly believe that pay should increase linearly proportionally with performance without a limit. H3: belief in the free market ideology leads to linear fair pay preferences. Data and Methods Survey experiments combine the strength of surveys in providing a sample that closely resembles the target population and the strength of randomized experiments in internal validity (Mutz 2011; Sniderman and Grob 1996). Data come from the Performance and Fair CEO Pay Study, a national survey experiment conducted in partnership with the YouGov Survey Company in July 2013. YouGov maintains a web-based and opt-in panel with more than one million respondents over the age of 18 in the U.S. and uses a sample matching methodology to construct its samples. This includes matching panelists from the large YouGov panel to a target sample – a probability sample of the target population to be mimicked (Rivers 2006). The objective of this matching process is to approximate the overall distribution of several indicators in the high quality sample to correct for the non-representativeness of YouGov’s opt-in sample. To accomplish this, one or more YouGov panelists are matched to each member of the target sample. For this particular survey, the respondents were matched on gender, age, race, education, party identification, ideology, and political interest. The synthetic sampling frame was constructed using the 2010 American Community Survey, along with data on voter registration status and turnout from the November 2010 Current Population Survey, and data on political interest and party identification from the 2007 Pew Religious Landscape Survey. Data gathered using YouGov’s matched sample method may perform as well as data from high-quality national surveys with probability sampling (Ansolabehere and Schaffner 2011; Ansolabehere and Rivers 2013; Vavreck and Iyengar 2013) accurately predicting voting behavior and election results (Rivers and Bailey 2009; Rivers 2012; Silver 2012). A total of 1,000 YouGov panelists completed the present survey, 11 cases were list-wise deleted due to missing values, yielding a final sample of 989 respondents with each of the treatment conditions having 225 to 277 subjects. Accounting for nonresponse and incompletes, the American Association for Public Opinion Research Response Rate 3 (RR3) for the study is 45.3%. Experiment Design To examine how fair CEO pay changes as a result of changing performance at different levels of pay, I cross two factors in this study: CEO performance (seven levels: a zero-point where there are no profits or losses, 11.6% profit; 23.3% profit; 35% profit and symmetrical losses) and average CEO pay in the industry (four levels: a control group with no mention of average pay, and $200,000, $1 million, or $5 million). This survey experiment features a between-subjects component and a within-subjects component. As a between-subjects manipulation, I provide each respondent with information about the average CEO pay in an industry. The control group is not provided with any information about the average pay in the industry while those in the three treatment conditions read that the industry average is $200,000, $1 million, or $5 million. As a within-subjects manipulation, I ask each respondent to indicate fair pay for a fictive CEO for seven levels of performance. Because the central hypothesis tests the limits of the rule of proportionality, it is important that the performance levels in the experiment reach very high and yet realistic levels for a company of the size evoked in this experiment. I use information compiled by Forbes for the purpose of benchmarking the numbers. Forbes provides data on measures of financial success for a list of 400 Best Big Companies who have a large presence in the U.S. market. Forbes also works with Audit Integrity to ensure that the list is of financially healthy companies. Based on these data, 12-month profit margins of companies in the top ten of this list ranked by this measure were between 30.3% to 46.2% (DeCarlo 2008). The highest level of performance reached in this survey experiment is 35% profit with performance ranging from 35% loss to 35% profit. I space intervals approximately equally to accommodate seven levels of performance with the inclusion of a zero-point. The experiment yields seven pay and performance pairings per respondent. The information is presented to the respondents using a vignette describing information about a fictitious CEO named John Hall of a fictitious company named American Wares Company. Men make up the vast majority of all CEOs in the largest companies (Eagly and Carli 2007; Sandberg 2013). I focus on respondents’ views about the pay of a male CEO to maintain statistical power for the experiment. Future research should focus on addressing differences in perceptions of fair pay for male and female CEOs as well as white and nonwhite CEOs. For example, Guillermina Jasso and Eva Meyersson Milgrom (2008) find that an otherwise identical female CEO is perceived as deserving of 84 to 94 percent of earnings of a male CEO. They also find that respondents do not hold consensus on the importance of various characteristics of the CEO such as years of schooling or characteristics of the firm in determining just pay. In this experiment, I focus on CEOs and their workplace contributions. Variables Dependent Measure: Fair CEO Pay I measure fair pay by asking respondents what John Hall, the CEO of a large multinational company, should be paid. A dropdown menu of response choices ranging from less than $50 thousand to $10 million is provided. The response options are in $10 thousand increments up to $100 thousand and in $100 thousand increments thereafter. As mentioned earlier, I measure fair pay at seven different levels of CEO performance. I begin with the question of how much the CEO should be paid when he performs exactly at the average level in his industry. I then ask respondents about fair pay in ascending order of performance, and following that, in descending order. The decision to use dollars or log dollars in the following regression models is an important one for the results of this experiment. Based on equity theory, rewards are perceived as fair when the ratio of inputs and outcomes is equivalent between two equivalent workers (Adams 1965; Messick and Cook 1983; Walster and Walster 1975; Walster, Walster and Berscheid 1978). Pilot interviews based on the scenarios used in this study show that, typically, when people believe that pay should increase directly proportionally to performance, they tend to increase pay by a fixed dollar amount at each performance interval rather than by calculating a percent increase to maintain proportionality of rewards and contributions. I therefore operationalize upholding the rule of proportionality as increasing fair pay by some dollar amount (rather than by the same percentage). Consequently, I use fair pay in dollars (rather than log dollars) as the dependent variable in the following regression models. Note that if log of fair pay was used in the models, then, the fair pay function for a respondent who increases fair pay by the same dollar amount at each performance interval would show diminishing returns because as pay increases the percent increase in pay from one level of performance to the next will diminish. Performance Performance and performance-squared (to indicate the curvature of the fair pay function) are the main predictors. I provide respondents with information about company profits and losses to indicate the CEO’s performance. I measure fair pay at seven different levels of performance: a zero-point where the company is doing exactly as well as the average in the industry, three levels of profits (11.6% profit; 23.3% profit; 35% profit) and symmetrical losses (11.6% loss; 23.3% loss; 35% loss). I recode this variable from 1 to 7 to indicate the seven levels of performance (where higher numbers indicate better performance). Manipulation checks were included in a pretest of this study conducted with 186 respondents using Amazon’s Mechanical Turk platform. A seven-point scale item to measure how respondents “rate John Hall compared to the average CEO in this industry” in terms of his performance shows that respondents find John Hall equivalent to the average CEO in his industry at the zero-point as intended. Responses to an open-ended item asking respondents to explain how they decided on John Hall’s pay also show that profits and losses are understood as signifying John Hall’s performance and that these are taken into account in deciding pay. Belief in the Free Market I measure respondents’ belief in the free market ideology by replicating an item from the General Social Survey. The item reads, “private enterprise is the best way to solve America’s economic problems.” Respondents are asked to state whether they agree or disagree with this statement. The response choices are strongly agree, agree, neither agree nor disagree, disagree, and strongly disagree. I recode this variable to indicate whether the respondent strongly agrees or agrees with this statement (1) or not (0). Control Variables I include age (in tens of years), sex, race, marital status, education (using indicator variables to indicate the educational categories less than high school, some college, BA degree, and more than BA degree), and family income. I also include a measure of political ideology (very liberal, liberal, conservative, very conservative, and independent). Models I specify the relationship between pay and performance using Ordinary Least Squares (OLS) regression models. This relationship is specified by adding a squared term for performance in the model to indicate the anticipated curvature. Respondents have different ideas about how fair pay and performance should be related: closely or loosely, proportionally or not. Therefore, pay and performance pairings will depend on the respondent. I take this into account with clustering and robust standard errors. ANALYSIS The Effect of Perceived Average CEO Pay on Fair Pay Figure 1 shows the median dollar amount of fair CEO pay chosen by respondents in the four experimental conditions as the CEO’s performance changes. The x-axis shows the CEO’s performance and the y-axis shows fair pay for that level of performance. The zero-point on the x-axis indicates the scenario where the CEO is described as performing equally well as the average CEO in his industry with neither loss nor profit. Positive numbers indicate better performance, and negative numbers indicate worse performance.1 Figure 1. View largeDownload slide Median Fair CEO Pay (by Level of Performance and Mentioned Average CEO Pay in the Industry) Source: Performance and Fair CEO Pay Study, 2013. N = 989. Figure 1. View largeDownload slide Median Fair CEO Pay (by Level of Performance and Mentioned Average CEO Pay in the Industry) Source: Performance and Fair CEO Pay Study, 2013. N = 989. I begin by examining the median fair pay chosen by respondents at the zero-point of performance. The respondents in the control condition were not given any information about how much the average CEO in the industry is paid and median pay in this condition at the zero-point is $400,000. In contrast, the median fair pay for respondents who were told that the average CEO is paid $200,000 is exactly $200,000. The median response among those who are given the information that the average CEO is paid $1 million is $700,000. Finally, median fair pay chosen by those in the $5 million treatment group is $1 million. Hence, the higher or lower the average CEO pay mentioned in the treatment, the higher or lower the median fair pay. To further examine the effect of perceived industry average pay, Table 1 Table 1. Mean Fair Pay for CEO with Average Performance, by Treatment Condition Mentioning Average CEO Pay in Industry Control $200,000 Treatment $1 Million Treatment $5 Million Treatment Mean Fair Pay $1,287,339 (163,403) $305,610*** (51,198) $818,339*** (73,964) $2,007,867*** (134,999) N 233 254 277 225 Control $200,000 Treatment $1 Million Treatment $5 Million Treatment Mean Fair Pay $1,287,339 (163,403) $305,610*** (51,198) $818,339*** (73,964) $2,007,867*** (134,999) N 233 254 277 225 Note: Standard errors are in parentheses. Source: Performance and Fair CEO Pay Study 2013. ***p < .001 (two-tailed) compared with treatment condition that precedes in dollar value. Table 1. Mean Fair Pay for CEO with Average Performance, by Treatment Condition Mentioning Average CEO Pay in Industry Control $200,000 Treatment $1 Million Treatment $5 Million Treatment Mean Fair Pay $1,287,339 (163,403) $305,610*** (51,198) $818,339*** (73,964) $2,007,867*** (134,999) N 233 254 277 225 Control $200,000 Treatment $1 Million Treatment $5 Million Treatment Mean Fair Pay $1,287,339 (163,403) $305,610*** (51,198) $818,339*** (73,964) $2,007,867*** (134,999) N 233 254 277 225 Note: Standard errors are in parentheses. Source: Performance and Fair CEO Pay Study 2013. ***p < .001 (two-tailed) compared with treatment condition that precedes in dollar value. shows mean fair pay by treatment condition and whether the mean pay for each of the conditions is significantly different from the condition that precedes it in dollar value. The mean fair pay chosen by the respondents in the four treatment conditions are ordered in the expected direction. The mean for the $200,000 treatment condition is the lowest: it is $305,610. The $1 million treatment group has the next highest mean fair pay with $818,339. The control group’s mean is higher than the mean for the $1 million treatment group: it is $1,287,339. Using data from the General Social Survey and the International Social Survey Programme, Leslie McCall (2013) shows that mean perceived actual pay for the “chairman of a large national company” was approximately $3 million in 2010. It is therefore not surprising that the mean for the control group is higher than the $1 million treatment group. Pay is also more right-skewed in the control condition than it is in the $1 million condition. In addition, the standard deviation of pay in the control condition is higher. This explains why the control group has a lower median fair pay and a higher mean fair pay compared with the $1 million treatment condition.2 Lastly, mean fair pay is $2,007,867 for the $5 million treatment group. These are statistically significantly different from the control group. More importantly, the mean for each treatment group is statistically significantly different from the group preceding it in dollar value. Therefore, the results show that people pay attention to what they believe to be average CEO pay in the industry in deciding fair CEO pay. Hypothesis 2a which states, “The higher the anchor for average CEO pay, the higher the fair pay reported,” finds support from the data. Next, I consider the hypothesis, “the higher the anchor for average CEO pay, the greater the downward adjustment of fair pay.” Table 2 Table 2. Mean Change in Pay from Industry Average Pay to the Respondent’s Choice, by Treatment Condition Mean SE N $200,000 Treatment 52.8%*** (25.60) 254 $1 Million Treatment −18.2%*** (7.396) 277 $5 Million Treatment −59.8%*** (2.700) 225 Mean SE N $200,000 Treatment 52.8%*** (25.60) 254 $1 Million Treatment −18.2%*** (7.396) 277 $5 Million Treatment −59.8%*** (2.700) 225 Source: Performance and Fair CEO Pay Study 2013. ***p < .001 (two-tailed) compared with treatment condition that precedes in dollar value. Table 2. Mean Change in Pay from Industry Average Pay to the Respondent’s Choice, by Treatment Condition Mean SE N $200,000 Treatment 52.8%*** (25.60) 254 $1 Million Treatment −18.2%*** (7.396) 277 $5 Million Treatment −59.8%*** (2.700) 225 Mean SE N $200,000 Treatment 52.8%*** (25.60) 254 $1 Million Treatment −18.2%*** (7.396) 277 $5 Million Treatment −59.8%*** (2.700) 225 Source: Performance and Fair CEO Pay Study 2013. ***p < .001 (two-tailed) compared with treatment condition that precedes in dollar value. presents the percent change in pay from the industry average provided in each treatment condition to the respondent’s choice of fair pay at the zero-point of performance. The control condition is not included in this table because an industry average is not mentioned in the control group. I calculate the percent change in pay for each respondent and report the mean for all responses by treatment condition. The mean percent change for respondents in the $200,000 treatment condition is 52.8%. Respondents in this condition, on average, adjust the $200,000 suggested pay upwards, increasing it by 52.8%. In contrast, those in the $1 million treatment condition tend to adjust their suggested pay downwards by 18.2%. Lastly, respondents in the $5 million treatment condition, on average, decrease the suggested pay in this condition by 59.8%. The means for all treatment conditions are significantly different from one another. Therefore, the respondents do not uncritically follow the industry average pay in deciding fair pay. The findings suggest that people have opinions about how much CEOs should be paid and that they adjust their response accordingly. Hypothesis 2b finds support from the data. How Does Fair CEO Pay Respond to Performance? Next, I consider the central hypothesis (inequality aversion hypothesis) which states, “fair CEO pay will not scale linearly with performance.” The empirical expectation is that the fair rewards function will show diminishing returns evidenced by a positive and significant performance coefficient in the model followed by a negative and significant performance-squared coefficient. I examine the overall pattern of the data using an OLS regression model with clustering to account for having seven observations (i.e. pay and performance pairings) per respondent and using robust standard errors. Model 1 in Table 3 Table 3. Coefficients from OLS Regression Model Predicting Fair CEO Pay M1 M2 Variables β SE β SE Performance 101,942*** (19,159) 168,193*** (50,721) Performance2 6,117** (2,204) −2,673 (5,698) Treatment Condition (reference = control)  Treatment $200,000 −819,511*** (141,058) −310,344* (135,136)  Treatment $1 Million −349,361* (145,497) −139,018 (138,405)  Treatment $5 Million 942,548*** (188,768) 500,871** (182,154) Interactions  Performance × Treatment $200,000 −182,206*** (54,964)  Performance × Treatment $1 Million −97,148 (57,585)  Performance × Treatment $5 Million 34,081 (71,247)  Performance2 × Treatment $200,000 10,983 (6,365)  Performance2 × Treatment $1 Million 8,912 (6,479)  Performance2 × Treatment $5 Million 15,268 (8,147) Constant 607,542*** (130,607) 518,345*** (127,372) R-squared 0.1514 0.1614 Observations 6,923 6,923 Clusters (standard errors adjusted for clusters) 989 989 M1 M2 Variables β SE β SE Performance 101,942*** (19,159) 168,193*** (50,721) Performance2 6,117** (2,204) −2,673 (5,698) Treatment Condition (reference = control)  Treatment $200,000 −819,511*** (141,058) −310,344* (135,136)  Treatment $1 Million −349,361* (145,497) −139,018 (138,405)  Treatment $5 Million 942,548*** (188,768) 500,871** (182,154) Interactions  Performance × Treatment $200,000 −182,206*** (54,964)  Performance × Treatment $1 Million −97,148 (57,585)  Performance × Treatment $5 Million 34,081 (71,247)  Performance2 × Treatment $200,000 10,983 (6,365)  Performance2 × Treatment $1 Million 8,912 (6,479)  Performance2 × Treatment $5 Million 15,268 (8,147) Constant 607,542*** (130,607) 518,345*** (127,372) R-squared 0.1514 0.1614 Observations 6,923 6,923 Clusters (standard errors adjusted for clusters) 989 989 Source: Performance and Fair CEO Pay Study 2013. Robust standard errors in parentheses *p < 0.05, **p < 0.01, ***p < 0.001 (two-tailed) Table 3. Coefficients from OLS Regression Model Predicting Fair CEO Pay M1 M2 Variables β SE β SE Performance 101,942*** (19,159) 168,193*** (50,721) Performance2 6,117** (2,204) −2,673 (5,698) Treatment Condition (reference = control)  Treatment $200,000 −819,511*** (141,058) −310,344* (135,136)  Treatment $1 Million −349,361* (145,497) −139,018 (138,405)  Treatment $5 Million 942,548*** (188,768) 500,871** (182,154) Interactions  Performance × Treatment $200,000 −182,206*** (54,964)  Performance × Treatment $1 Million −97,148 (57,585)  Performance × Treatment $5 Million 34,081 (71,247)  Performance2 × Treatment $200,000 10,983 (6,365)  Performance2 × Treatment $1 Million 8,912 (6,479)  Performance2 × Treatment $5 Million 15,268 (8,147) Constant 607,542*** (130,607) 518,345*** (127,372) R-squared 0.1514 0.1614 Observations 6,923 6,923 Clusters (standard errors adjusted for clusters) 989 989 M1 M2 Variables β SE β SE Performance 101,942*** (19,159) 168,193*** (50,721) Performance2 6,117** (2,204) −2,673 (5,698) Treatment Condition (reference = control)  Treatment $200,000 −819,511*** (141,058) −310,344* (135,136)  Treatment $1 Million −349,361* (145,497) −139,018 (138,405)  Treatment $5 Million 942,548*** (188,768) 500,871** (182,154) Interactions  Performance × Treatment $200,000 −182,206*** (54,964)  Performance × Treatment $1 Million −97,148 (57,585)  Performance × Treatment $5 Million 34,081 (71,247)  Performance2 × Treatment $200,000 10,983 (6,365)  Performance2 × Treatment $1 Million 8,912 (6,479)  Performance2 × Treatment $5 Million 15,268 (8,147) Constant 607,542*** (130,607) 518,345*** (127,372) R-squared 0.1514 0.1614 Observations 6,923 6,923 Clusters (standard errors adjusted for clusters) 989 989 Source: Performance and Fair CEO Pay Study 2013. Robust standard errors in parentheses *p < 0.05, **p < 0.01, ***p < 0.001 (two-tailed) shows the results using performance and performance-squared as the predictor variables. Interaction effects are added in Model 2. Model 1 shows that the coefficient for performance is positive and significant and that the coefficient for performance-squared is also positive and significant. This indicates that fair pay increases with performance and that the rewards function increases faster at higher pay levels. The coefficients for the treatment conditions confirm the earlier findings: fair pay is lower in the $200,000 treatment condition than it is in the control condition; it is also lower in the $1 million condition and higher in the $5 million condition compared to the control condition. Model 2 adds interaction effects for performance and performance-squared and the treatment conditions. When interactions are included, the performance-squared term is no longer statistically significant. Most importantly, however, I do not find evidence for diminishing returns to performance at higher pay levels under either model specification. While findings suggest that respondents show aversion to high inequality in other ways (as will be further discussed below), Hypothesis 1 (inequality aversion hypothesis) which states that fair CEO pay will not scale linearly with performance does not find support from the data. This finding reveals a commitment to the principle of proportionality of rewards and contributions even at levels of pay reached in this experiment. Examining the interaction effects, I find that the $200,000 treatment condition has an overall flatter fair pay function than other conditions. The interaction between that condition and performance is negative and significant, and it points in the opposite direction of the main effect in the model. Further examination (not shown) reveals that, for this treatment group, the mean percent change from one level of performance to the next is positive (ranging from 27% to 125% increase) but usually lower than in other experimental conditions. Moreover, for this treatment group, median percent change in fair pay from one level of performance to the next is zero. In comparison, median percent change for the $1 million treatment group is approximately 10% at each interval. Data cannot weigh in on the reasons for the reduced sensitivity to performance for the lower treatment group, however, to speculate, respondents may not want to allocate too low a pay for a position as high as the CEO’s. This possibility is congruent with results presented in Table 2 where, on average, respondents increase the suggested pay for this treatment group. An interesting question for future research is whether there could be an occupational floor to fair pay (rather than an ethical ceiling as hypothesized here). Other interaction effects included in the model are not statistically significant. No significant differences are found between the control group and the higher dollar amount conditions in slope or in the degree of curvature for the fair pay functions. Considering the findings, I do not find evidence to support Hypothesis 2c which states that higher anchors for average CEO pay will lead to greater curvilinearity. Belief in the Free Market Ideology and Fair Pay I hypothesized that respondents’ belief in the free market ideology would lead to linear fair pay preferences. I examine the effect of agreeing with the statement on private enterprise as well as how the curvature of the function changes with belief in this statement. Table 4 Table 4. Coefficients from OLS Regression Model Predicting Fair CEO Pay Model 1 Model 2 Variables β SE β SE Performance 168,193*** (50,776) 146,175** (52,487) Performance2 −2,673 (5,704) −6,336 (6,004) Treatment Condition (reference = control)  Treatment $200,000 −321,043* (139,699) −303,550* (139,287)  Treatment $1 Million −140,733 (141,668) −119,712 (141,645)  Treatment $5 Million 469,944** (181,261) 489,601** (181,867) Interactions  Performance × Treatment $200,000 −182,206*** (55,024) −184,593*** (55,048)  Performance × Treatment $1 Million −97,148 (57,648) −100,017 (57,668)  Performance × Treatment $5 Million 34,081 (71,325) 31,398 (71,330)  Performance2 × Treatment $200,000 10,983 (6,372) 10,586 (6,350)  Performance2 × Treatment $1 Million 8,912 (6,486) 8,435 (6,461)  Performance2 × Treatment $5 Million 15,268 (8,156) 14,821 (8,149)  Age (in tens of years) −49,814 (29,132) −49,814 (29,136)  Female −273,389** (98,594) −273,389** (98,608)  Married −202,519* (99,191) −202,519* (99,205)  Race nonwhite 70,821 (109,324) 70,821 (109,340) Education (reference = high school)  <High School 285,826 (277,706) 285,826 (277,746)  Some College 189,040 (112,334) 189,040 (112,350)  BA Degree 64,906 (123,788) 64,906 (123,806)  >BA Degree 193,524 (182,967) 193,524 (182,993) Family Income (ln) 239,771*** (69,242) 239,771*** (69,252) Political Ideology (reference = moderate)  Very Liberal −2,646 (154,560) −2,646 (154,582)  Liberal 96,222 (133,447) 96,222 (133,466)  Conservative 332,231* (130,033) 332,231* (130,052)  Very conservative 324,278* (144,465) 324,278* (144,486)  Independent 44,548 (180,525) 44,548 (180,552) Values  Free Market Ideology (agree) 475,842*** (117,661) 84,299 (128,189)  Free Market Ideology (agree) × Performance 53,439 (38,534)  Free Market Ideology (agree) × Performance2 8,889* (4,428) ConstantR-squared −2026,000**0.222 (743,098) −1865,000*0.227 (746,682) ObservationsClusters (standard errors adjusted for clusters) 6,923989 6,923989 Model 1 Model 2 Variables β SE β SE Performance 168,193*** (50,776) 146,175** (52,487) Performance2 −2,673 (5,704) −6,336 (6,004) Treatment Condition (reference = control)  Treatment $200,000 −321,043* (139,699) −303,550* (139,287)  Treatment $1 Million −140,733 (141,668) −119,712 (141,645)  Treatment $5 Million 469,944** (181,261) 489,601** (181,867) Interactions  Performance × Treatment $200,000 −182,206*** (55,024) −184,593*** (55,048)  Performance × Treatment $1 Million −97,148 (57,648) −100,017 (57,668)  Performance × Treatment $5 Million 34,081 (71,325) 31,398 (71,330)  Performance2 × Treatment $200,000 10,983 (6,372) 10,586 (6,350)  Performance2 × Treatment $1 Million 8,912 (6,486) 8,435 (6,461)  Performance2 × Treatment $5 Million 15,268 (8,156) 14,821 (8,149)  Age (in tens of years) −49,814 (29,132) −49,814 (29,136)  Female −273,389** (98,594) −273,389** (98,608)  Married −202,519* (99,191) −202,519* (99,205)  Race nonwhite 70,821 (109,324) 70,821 (109,340) Education (reference = high school)  <High School 285,826 (277,706) 285,826 (277,746)  Some College 189,040 (112,334) 189,040 (112,350)  BA Degree 64,906 (123,788) 64,906 (123,806)  >BA Degree 193,524 (182,967) 193,524 (182,993) Family Income (ln) 239,771*** (69,242) 239,771*** (69,252) Political Ideology (reference = moderate)  Very Liberal −2,646 (154,560) −2,646 (154,582)  Liberal 96,222 (133,447) 96,222 (133,466)  Conservative 332,231* (130,033) 332,231* (130,052)  Very conservative 324,278* (144,465) 324,278* (144,486)  Independent 44,548 (180,525) 44,548 (180,552) Values  Free Market Ideology (agree) 475,842*** (117,661) 84,299 (128,189)  Free Market Ideology (agree) × Performance 53,439 (38,534)  Free Market Ideology (agree) × Performance2 8,889* (4,428) ConstantR-squared −2026,000**0.222 (743,098) −1865,000*0.227 (746,682) ObservationsClusters (standard errors adjusted for clusters) 6,923989 6,923989 Source: Performance and Fair CEO Pay Study 2013. Robust standard errors in parentheses *p < 0.05, **p < 0.01, ***p < 0.001 (two-tailed) Table 4. Coefficients from OLS Regression Model Predicting Fair CEO Pay Model 1 Model 2 Variables β SE β SE Performance 168,193*** (50,776) 146,175** (52,487) Performance2 −2,673 (5,704) −6,336 (6,004) Treatment Condition (reference = control)  Treatment $200,000 −321,043* (139,699) −303,550* (139,287)  Treatment $1 Million −140,733 (141,668) −119,712 (141,645)  Treatment $5 Million 469,944** (181,261) 489,601** (181,867) Interactions  Performance × Treatment $200,000 −182,206*** (55,024) −184,593*** (55,048)  Performance × Treatment $1 Million −97,148 (57,648) −100,017 (57,668)  Performance × Treatment $5 Million 34,081 (71,325) 31,398 (71,330)  Performance2 × Treatment $200,000 10,983 (6,372) 10,586 (6,350)  Performance2 × Treatment $1 Million 8,912 (6,486) 8,435 (6,461)  Performance2 × Treatment $5 Million 15,268 (8,156) 14,821 (8,149)  Age (in tens of years) −49,814 (29,132) −49,814 (29,136)  Female −273,389** (98,594) −273,389** (98,608)  Married −202,519* (99,191) −202,519* (99,205)  Race nonwhite 70,821 (109,324) 70,821 (109,340) Education (reference = high school)  <High School 285,826 (277,706) 285,826 (277,746)  Some College 189,040 (112,334) 189,040 (112,350)  BA Degree 64,906 (123,788) 64,906 (123,806)  >BA Degree 193,524 (182,967) 193,524 (182,993) Family Income (ln) 239,771*** (69,242) 239,771*** (69,252) Political Ideology (reference = moderate)  Very Liberal −2,646 (154,560) −2,646 (154,582)  Liberal 96,222 (133,447) 96,222 (133,466)  Conservative 332,231* (130,033) 332,231* (130,052)  Very conservative 324,278* (144,465) 324,278* (144,486)  Independent 44,548 (180,525) 44,548 (180,552) Values  Free Market Ideology (agree) 475,842*** (117,661) 84,299 (128,189)  Free Market Ideology (agree) × Performance 53,439 (38,534)  Free Market Ideology (agree) × Performance2 8,889* (4,428) ConstantR-squared −2026,000**0.222 (743,098) −1865,000*0.227 (746,682) ObservationsClusters (standard errors adjusted for clusters) 6,923989 6,923989 Model 1 Model 2 Variables β SE β SE Performance 168,193*** (50,776) 146,175** (52,487) Performance2 −2,673 (5,704) −6,336 (6,004) Treatment Condition (reference = control)  Treatment $200,000 −321,043* (139,699) −303,550* (139,287)  Treatment $1 Million −140,733 (141,668) −119,712 (141,645)  Treatment $5 Million 469,944** (181,261) 489,601** (181,867) Interactions  Performance × Treatment $200,000 −182,206*** (55,024) −184,593*** (55,048)  Performance × Treatment $1 Million −97,148 (57,648) −100,017 (57,668)  Performance × Treatment $5 Million 34,081 (71,325) 31,398 (71,330)  Performance2 × Treatment $200,000 10,983 (6,372) 10,586 (6,350)  Performance2 × Treatment $1 Million 8,912 (6,486) 8,435 (6,461)  Performance2 × Treatment $5 Million 15,268 (8,156) 14,821 (8,149)  Age (in tens of years) −49,814 (29,132) −49,814 (29,136)  Female −273,389** (98,594) −273,389** (98,608)  Married −202,519* (99,191) −202,519* (99,205)  Race nonwhite 70,821 (109,324) 70,821 (109,340) Education (reference = high school)  <High School 285,826 (277,706) 285,826 (277,746)  Some College 189,040 (112,334) 189,040 (112,350)  BA Degree 64,906 (123,788) 64,906 (123,806)  >BA Degree 193,524 (182,967) 193,524 (182,993) Family Income (ln) 239,771*** (69,242) 239,771*** (69,252) Political Ideology (reference = moderate)  Very Liberal −2,646 (154,560) −2,646 (154,582)  Liberal 96,222 (133,447) 96,222 (133,466)  Conservative 332,231* (130,033) 332,231* (130,052)  Very conservative 324,278* (144,465) 324,278* (144,486)  Independent 44,548 (180,525) 44,548 (180,552) Values  Free Market Ideology (agree) 475,842*** (117,661) 84,299 (128,189)  Free Market Ideology (agree) × Performance 53,439 (38,534)  Free Market Ideology (agree) × Performance2 8,889* (4,428) ConstantR-squared −2026,000**0.222 (743,098) −1865,000*0.227 (746,682) ObservationsClusters (standard errors adjusted for clusters) 6,923989 6,923989 Source: Performance and Fair CEO Pay Study 2013. Robust standard errors in parentheses *p < 0.05, **p < 0.01, ***p < 0.001 (two-tailed) presents results from OLS regression models which include this explanatory variable as well as control variables. According to Model 1 in Table 4, agreeing with this statement has a positive and significant (p < 0.001) effect on fair pay. Controlling for other factors, those who agree with this statement allocate higher pay to the CEO described in the vignette. As interactions are added in Model 2, I find that while agreeing with this statement is no longer statistically significant, the quadratic term generated with this variable is significant and positive (p < 0.05). This finding suggests that the curvature of the fair pay function is different for people who agree with this statement compared with people who do not agree with the statement. For those who agree with the free market ideology, performance counts heavier in determining fair pay. Other noteworthy findings are that people with higher family incomes allocate larger rewards to the CEO, as do the respondents who politically identify as conservative. Women and those who are married show the opposite effect. DISCUSSION AND CONCLUSIONS The Great Recession of 2008 revived interest in high executive pay and income inequality. The primary objective of this study has been to juxtapose two propositions: a) that the American discontent with extremely high CEO pay stems mainly from people’s belief that CEOs are not contributing highly enough to merit such high pay and b) people find extremely high pay objectionable in principle and they would oppose extreme levels of pay even if they believed that CEOs were contributing highly. Respondents enter surveys with previously held beliefs, including about whether CEOs are leading their companies to profits or causing economic recessions. Performance and Fair CEO Pay Study, which uses data from a national survey-experiment, brings new information on how perceived actual CEO pay and perceived performance affect fair CEO pay. Data do not only focus on an average case, and are particularly well-suited for examining the general public’s views about the highest levels of CEO pay. The experiment provides data on fair pay for a range of performances, including very poor and very good performances. Findings reveal that attitudes toward CEO pay manifest themselves in complex ways and in ways differently than initially hypothesized. On average, respondents show aversion to extremely high pay but not in the form specified under empirical expectations. Data do not show support for the hypothesis that pay will not scale linearly with performance. As a CEO’s contribution changes, so does the CEO’s fair pay, linearly proportionally to performance and without an observed upper limit. However, respondents find $5 million as well as $1 million too high, and $200,000 too low for an average performing CEO. Only extraordinary performance leads to very high levels of fair pay. Even then, fair pay seldom reaches heights observed in reality. The highest level of performance examined here is rewarded with a median fair pay of $2 million and a mean fair pay of about $3 million (where the average CEO is described as being rewarded with $5 million). In addition, even when the average CEO is described as being paid $5 million, median fair pay increases arguably slowly – from $500,000 to $2 million as performance changes from large losses to large gains. Based on these findings, opposition to extremely high pay may, in fact, be reconciled with support for pay for performance: the slope of the pay and performance function may be at issue rather than a hard limit on pay. The findings have important policy implications. In 2014, the top one percent of earners in the U.S. received about 18% of all income, excluding capital gains (Alvaredo et al. 2011). If its current trajectory continues unchanged, according to Piketty (2014), such concentration may reach new heights especially due to the “super managers” phenomenon examined here. Joseph Stiglitz (2012) argues, “it is not just that they [bankers and CEOs] have become the whipping boys of popular opinion. They are emblematic of what has gone wrong.” (p. liv). Trends in income inequality are heavily shaped by political forces (Hacker and Pierson 2010; Piketty 2014; Reich 2012). Jacob Hacker and Paul Pierson (2010) write that political “drift” or “the deliberate failure to adapt public policies to the shifting realities of a dynamic economy” (p. 52) has led to the new heights in executive pay. Some scholars argue that whether income inequality will be intensified in the future depends largely on how effectively pay is justified and how tolerant people are of high CEO pay (Piketty 2014). CEO pay is determined by a variety of factors and the acceptability of high pay differentials has an important place among them. New policies have been proposed in the U.S. and elsewhere with the aim of reining in executive pay. Among these, a new rule by the Securities and Exchange Commission which will take effect beginning with the fiscal year 2017 and which requires the disclosure of the ratio of the chief executive’s pay to the pay of the median employee for 3,571 registrants is one that is designed to empower such norms. The objective of the disclosure is to provide new data to aid in investment decisions and say on pay voting decisions (Securities and Exchange Commission 2015). A contribution of the new disclosure requirement will be new data on the proportionality of rewards and contributions within-firms. An Associated Press/Equilar study estimates the ratio at 257:1 (Associated Press 2014). While even within-firm changes in ratios will be challenging to assess due to changes in business practices, the new data will be useful in monitoring trends. Since incomes at the middle of the income distribution are growing slower than those at the top (Mishel et al. 2012), the data will likely highlight gains within-firms also disproportionately going to the top. Companies will have some leeway in calculating median pay, but a high estimate of the median, which would reveal a lower ratio of CEO pay to median pay, may not necessarily be advantageous as employees who compare unfavorably to the median may be dissatisfied (Card et al. 2012). Alexandre Mas (2014) writes, “compensation is sensitive to increased transparency” (p. 5) because public opinion on perceived excessive pay matters. The new data may also serve to increase regular employees’ pay through the diffusion of power afforded by the disclosure (Rosenfeld and Denice 2015), or generate support for policies such as inequality taxes (Ayres and Edlin 2011) or millionaire taxes (Young et al. 2016). The findings suggest important avenues for future research. First, we cannot rule out the possibility that through extraordinary contributions, extremely high pay could be justified. By considering performance only, this study cannot weigh in on the effects of other relevant inputs on fair CEO pay. Future research should examine the combined effects of various types of inputs along with performance. For example, an examination of the effects of job creation and layoffs may be especially useful. The literature will benefit from a better understanding of how the general public believes CEO pay is actually decided (for example, which forms of rent-seeking, if any, are partially responsible for determining pay).3 Findings in this study show that attitudes toward CEO pay, in part depend on political ideology and on values such as a strong belief in a free market. Further studies can illuminate differences between stronger and weaker adherents of a free market ideology. What are the mechanisms through which the free market ideology has an effect on attitudes toward CEO pay?4 How do business leaders (who have more power in determining executive compensation) and the general public differ in a host of relevant values and beliefs and in acceptance of high CEO pay? In addition to these inquiries, gaps in perceived fair pay for male and female CEOs as well as for white and nonwhite CEOs should also be part of the future research agenda. APPENDIX Figure A1. View largeDownload slide Mean Fair CEO Pay By Level of Performance and Mentioned Average CEO Pay in the Industry Source: Performance and Fair CEO Pay Study 2013. N = 989. Figure A1. View largeDownload slide Mean Fair CEO Pay By Level of Performance and Mentioned Average CEO Pay in the Industry Source: Performance and Fair CEO Pay Study 2013. N = 989. Figure A2. View largeDownload slide Box Plots for Fair CEO Pay at Average Level of Performance, by Treatment Condition Source: Performance and Fair CEO Pay Study 2013. N = 989. Source: Performance and Fair CEO Pay Study 2013. N = 989. Figure A2. View largeDownload slide Box Plots for Fair CEO Pay at Average Level of Performance, by Treatment Condition Source: Performance and Fair CEO Pay Study 2013. N = 989. Source: Performance and Fair CEO Pay Study 2013. N = 989. Figure A3. View largeDownload slide Standard Deviation of Fair CEO Pay by Level of Performance and Average CEO Pay Mentioned in Treatment Source: Performance and Fair CEO Pay Study 2013. N = 989. Figure A3. View largeDownload slide Standard Deviation of Fair CEO Pay by Level of Performance and Average CEO Pay Mentioned in Treatment Source: Performance and Fair CEO Pay Study 2013. N = 989. The author wishes to thank David B. Grusky, Michael J. Rosenfeld, Cristobal Young, Paul Sniderman, Michael Tomz, Michelle Jackson, Shelley J. Correll, Erin M. Cumberworth, Beth Red Bird, Rachel Wright, members of the Stanford Laboratory for the Study of American Values and the anonymous Social Problems reviewers for their valuable comments on prior drafts. This research was supported by funding from the Stanford Laboratory for the Study of American Values. Footnotes 1 Figure A1 shows results using the mean and confidence intervals around the mean. 2 Figure A2 shows box plots for fair pay at the zero point of performance by treatment group. Figure A3 shows the standard deviation of mean fair pay by treatment group and by performance. Fair pay is right-skewed and especially so for the $5 million treatment condition and the control group. The standard deviation for the $200,000 and the $1 million treatment groups are lower than those for the control group and the $5 million group. An interpretation of these findings is that those in the control group, not provided with guidance, have vastly differing views of how much the average CEO is paid and should be paid. And, to speculate, despite receiving guidance on what the average CEO is paid in the industry, some respondents in the $5 million treatment group find this pay too high and therefore choose not to align their responses with this anchor while others do so. 3 I thank an anonymous reviewer for this point. 4 I thank an anonymous reviewer for this point. REFERENCES Adams J. Stacy 1965 . “ Inequity in Social Exchange .” in Advances in Experimental Social Psychology Vol. 2 , edited by Berkowitz L. New York Academic Press . Google Scholar CrossRef Search ADS Alvaredo Facundo , Atkinson Anthony , Piketty Thomas , Saez Emmanuel 2011 . The World Top Incomes Database . Retrieved December 2015 (http://topincomes.g-mond.parisschoolofeconomics.eu/). Alves Wayne M. , Rossi Peter H. 1978 . “ Who Should Get What? Fairness Judgments of the Distribution of Earnings .” American Journal of Sociology 84 3 : 541 - 64 . Google Scholar CrossRef Search ADS Ansolabehere Stephen , Schaffner Brian F. 2011 . “ Re-Examining the Validity of Different Survey Modes for Measuring Public Opinion in the U.S.: Findings from a 2010 Multi-Mode Comparison .” Pp. 12 - 15 in AAPOR Annual Conference Phoenix, AZ . Ansolabehere Stephen , Rivers Douglas 2013 . “ Cooperative Survey Research .” Annual Review of Political Science 16 : 307 - 29 . Google Scholar CrossRef Search ADS Associated Press 2014 . “ Median Pay for CEOs Rises above $10 Million in 2013 .” in Los Angeles Times. May 27, 2014 . Ayres Ian , Edlin Aaron S. 2011 . “ Don’t Tax the Rich. Tax Inequality Itself’ .” Pp. A29 in New York Times. Dec 19, 2011 . Bakija Jon , Cole Adam , Heim Bradley T. 2012 . “ Jobs and Income Growth of Top Earners and the Causes of Changing Income Inequality: Evidence from U.S. Tax Return Data .” Working Paper. Williams College . Bartels Larry M. 2005 . “ Homer Gets a Tax Cut: Inequality and Public Policy in the American Mind .” Perspectives on Politics 3 01 : 15 - 31 . Google Scholar CrossRef Search ADS Bebchuk Lucian , Grinstein Yaniv 2005 . “ The Growth of Executive Pay .” Oxford Review of Economic Policy 21 2 : 283 - 303 . Google Scholar CrossRef Search ADS Bebchuk Lucian A. , Fried Jesse M. 2005 . “ Pay without Performance: Overview of the Issues .” Journal of Applied Corporate Finance 17 4 : 8 - 23 . Google Scholar CrossRef Search ADS Bebchuk Lucian A. , Fried Jesse M. 2010 . “ Tackling the Managerial Power Problem: The Key to Improving Executive Compensation .” Pathways Magazine ( Summer ): 9 - 13 . Bies Robert J. , Moag Joseph S. 1986 . “ Interactional Justice: Communication Criteria of Fairness .” Pp. 43 - 55 in Research on Negotiation in Organizations Vol. 1 , edited by Lewicki R. J. , Sheppard B. H. , Bazerman M. H. Greenwich, CT JAI Press . Booth William James 1994 . “ On the Idea of the Moral Economy .” The American Political Science Review 88 3 : 653 - 67 . Google Scholar CrossRef Search ADS Bowman Karlyn H. 2000 . “ No Din over Dollar Divide .” in Chicago Sun-Times. July 31, 2000 . Card David , Mas Alexandre , Moretti Enrico , Saez Emmanuel 2012 . “ Inequality at Work: The Effect of Peer Salaries on Job Satisfaction .” The American Economic Review 102 6 : 2981 - 3003 . Google Scholar CrossRef Search ADS Cervone Daniel , Peake Philip K. 1986 . “ Anchoring, Efficacy, and Action: The Influence of Judgmental Heuristics on Self-Efficacy Judgments and Behavior .” Journal of Personality and Social Psychology 50 3 : 492 - 501 . Google Scholar CrossRef Search ADS Chapman Gretchen B. , Johnson Eric J. 2002 . “ Incorporating the Irrelevant: Anchors in Judgments of Belief and Value .” Pp. 120 - 38 in Heuristics and Biases: The Psychology of Intuitive Judgment , edited by Gilovich T. , Griffin Dale , Kahneman Daniel New York Cambridge University Press . Google Scholar CrossRef Search ADS Crystal Graef S. 1991 In Search of Excess: The Overcompensation of American Executives New York Norton . DeCarlo Scott 2008 . “ America’s Best Big Companies .” Forbes . Deutsch Morton 1975 . “ Equity, Equality, and Need: What Determines Which Value Will Be Used as the Basis of Distributive Justice? ”. Journal of Social Issues 31 3 : 137 - 49 . Google Scholar CrossRef Search ADS DiPrete Thomas A. , Eirich Gregory M. , Pittinsky Matthew 2010 . “ Compensation Benchmarking, Leapfrogs, and the Surge in Executive Pay .” American Journal of Sociology 115 6 : 1671 - 712 . Google Scholar CrossRef Search ADS Eagly Alice H. , Chaiken Shelly 1993 The Psychology of Attitudes Fort Worth, TX Harcourt Brace Jovanovich . Eagly Alice H. , Carli Linda L. 2007 Through the Labyrinth: The Truth About How Women Become Leaders Boston, MA Harvard Business School Press . Elster Jon 1989 . “ Social Norms and Economic Theory .” The Journal of Economic Perspectives 3 4 : 99 - 117 . Google Scholar CrossRef Search ADS Etzioni Amitai. 1988 The Moral Dimension: Toward a New Economics New York Free Press . Evans M.D.R. , Kelley Jonathan , Peoples Clayton D. 2010 . “ Justifications of Inequality: The Normative Basis of Pay Differentials in 31 Nations .” Social Science Quarterly 91 5 : 1405 - 31 . Google Scholar CrossRef Search ADS Fishbein Martin , Ajzen Icek 1975 Belief, Attitude, Intention and Behavior: An Introduction to Theory and Research Reading, MA Addison-Wesley . Folger Robert , Greenberg Jerald 1985 . “ Procedural Justice: An Interpretive Analysis of Personnel Systems .” Research in Personnel and Human Resources Management 3 : 141 - 83 . Fourcade Marion , Healy Kieran 2007 . “ Moral Views of Market Society .” Annual Review of Sociology 33 : 285 - 311 . Google Scholar CrossRef Search ADS Frank R. H. , Cook P. J. 1995 The Winner-Take-All Society: Why the Few at the Top Get So Much More Than the Rest of Us New York Free Press . Frydman Carola , Jenter Dirk 2010 . “ CEO Compensation .” Annual Review of Financial Economics 2 1 : 75 - 102 . Google Scholar CrossRef Search ADS Gabaix Xavier , Landier Augustin 2008 . “ Why Has CEO Pay Increased So Much? ”. The Quarterly Journal of Economics 123 1 : 49 - 100 . Google Scholar CrossRef Search ADS Goode William J. 1978 The Celebration of Heroes: Prestige as a Social Control System Berkeley University of California Press . Green Donald , Jacowitz Karen E. , Kahneman Daniel , McFadden Daniel 1998 . “ Referendum Contingent Valuation, Anchoring, and Willingness to Pay for Public Goods .” Resource and Energy Economics 20 2 : 85 - 116 . Google Scholar CrossRef Search ADS Greenberg Jerald. 1987 . “ A Taxonomy of Organizational Justice Theories .” Academy of Management Review 12 1 : 9 - 22 . Google Scholar CrossRef Search ADS Grusky David B. , Wimer Christopher 2010 . “ Is There Too Much Inequality? ” Pp. 3 - 10 in The Inequality Puzzle: European and U.S. Leaders Discuss Rising Income Inequality , edited by Berger R. , Grusky D. B. , Raffel T. , Samuels G. , Wimer C. New York Springer . Google Scholar CrossRef Search ADS Hacker Jacob S. , Pierson Paul 2010 Winner-Take-All Politics: How Washington Made the Rich Richer–and Turned Its Back on the Middle Class New York, NY Simon and Schuster . Hall Brian J. , Liebman Jeffrey B. 1998 . “ Are CEOs Really Paid Like Bureaucrats? ”. The Quarterly Journal of Economics 113 3 : 653 - 91 . Google Scholar CrossRef Search ADS Hegtvedt Karen , Cook Karen 2001 . “ Distributive Justice: Recent Theoretical Developments and Applications .” Pp. 93 - 132 in Handbook of Justice Research in Law , edited by Sanders I. J. , Hamilton V. L. New York Kluwer . Hirschman Albert O. 1984 . “ Against Parsimony: Three Easy Ways of Complicating Some Categories of Economic Discourse .” Bulletin of the American Academy of Arts and Sciences 37 8 : 11 - 28 . Google Scholar CrossRef Search ADS Hochschild Jennifer L. 1981 What’s Fair? Americans’ Attitudes toward Distributive Justice Cambridge, MA Harvard University Press . Hochschild Jennifer L. 1996 Facing up to the American Dream: Race, Class, and the Soul of the Nation Princeton, NJ Princeton University Press . Homans George C. 1974 Social Behavior: Its Elementary Forms New York Harcourt, Brace, Jovanovich . Huber Joan , Humbert Form William 1973 Income and Ideology: An Analysis of the American Political Formula New York Free Press . Jacowitz Karen E. , Kahneman Daniel 1995 . “ Measures of Anchoring in Estimation Tasks .” Personality and Social Psychology Bulletin 21 : 1161 - 66 . Google Scholar CrossRef Search ADS Jasso Guillermina , Meyersson Milgrom Eva M. 2008 . “ Distributive Justice and CEO Compensation .” Acta Sociologica 51 2 : 123 - 43 . Google Scholar CrossRef Search ADS Jensen Michael , Murphy Kevin , Wruck Eric 2004 . “ Remuneration: Where We’ve Been, How We Got to Here, What Are the Problems, and How to Fix Them .” Working Paper, Harvard University . Google Scholar CrossRef Search ADS Kaplan Steven N. 2008 . “ Are U.S. CEOs Overpaid? ”. The Academy of Management Perspectives 22 2 : 5 - 20 . Google Scholar CrossRef Search ADS Kaplan Steven N. , Rauh Joshua 2010 . “ Wall Street and Main Street: What Contributes to the Rise in the Highest Incomes? ”. Review of Financial Studies 23 3 : 1004 - 50 . Google Scholar CrossRef Search ADS Kiatpongsan Sorapop , Norton Michael I. 2014 . “ How Much (More) Should CEOs Make? A Universal Desire for More Equal Pay .” Perspectives on Psychological Science 9 : 587 - 93 . Google Scholar CrossRef Search ADS Kim Jerry W. , Kogut Bruce , Yang Jae-Suk 2015 . “ Executive Compensation, Fat Cats, and Best Athletes .” American Sociological Review 80 2 : 299 - 328 . Google Scholar CrossRef Search ADS Kluegel James R. , Smith Eliot R. 1986 Beliefs About Inequality: Americans’ Views of What Is and What Ought to Be Hawthorne, NY Aldine de Gruyter . Kuhnen Camelia M. , Niessen Alexandra 2012 . “ Public Opinion and Executive Compensation .” Management Science 58 7 : 1249 - 72 . Google Scholar CrossRef Search ADS Ladd Everett Carll 1994 The American Ideology: An Exploration of the Origins, Meaning, and Role of American Political Ideas Storrs, CT Roper Center for Public Opinion Research . Ladd Everett Carll , Bowman Karlyn H. 1998 Attitudes toward Economic Inequality Washington, DC Aei Press . Lamm Helmut , Schwinger Thomas 1980 . “ Norms Concerning Distributive Justice: Are Needs Taken into Consideration in Allocation Decisions? ” Social Psychology Quarterly 43 : 425 - 29 . Google Scholar CrossRef Search ADS Lane Robert E. 1962 Political Ideology: Why the American Common Man Believes What He Does New York The Free Press . Lane Robert E. 1986 . “ Market Justice, Political Justice .” The American Political Science Review 80 : 383 - 402 . Google Scholar CrossRef Search ADS Leung Kwok , Park Hun-Joon 1986 . “ Effects of Interactional Goal on Choice of Allocation Rule: A Cross-National Study .” Organizational Behavior and Human Decision Processes 37 1 : 111 - 20 . Google Scholar CrossRef Search ADS Leventhal Gerald S. 1976 . “ The Distribution of Rewards and Resources in Groups and Organizations .” Advances in Experimental Social Psychology 9 : 91 - 131 . Google Scholar CrossRef Search ADS Mankiw N. Gregory 2010 . “ Spreading the Wealth Around: Reflections Inspired by Joe the Plumber .” Eastern Economic Journal 36 3 : 285 - 98 . Google Scholar CrossRef Search ADS Mansbridge Jane J. 1990 Beyond Self-Interest Chicago University of Chicago Press . Markovsky Barry 1988 . “ Anchoring Justice .” Social Psychology Quarterly 51 3 : 213 - 24 . Google Scholar CrossRef Search ADS Marshall Gordon , Swift Adam , Routh David , Burgoyne Carole 1999 . “ What Is and What Ought to Be: Popular Beliefs About Distributive Justice in Thirteen Countries .” European Sociological Review 15 4 : 349 - 67 . Google Scholar CrossRef Search ADS Mas Alexandre 2014 . “ Does Transparency Lead to Pay Compression? ”. NBER Working Paper . McCall Leslie. 2003 . “ Do They Know and Do They Care? Americans’ Awareness of Rising Inequality .” Working Paper . McCall Leslie , Kenworthy Lane 2009 . “ Americans’ Social Policy Preferences in the Era of Rising Inequality .” Perspectives on Politics 7 3 : 459 - 84 . Google Scholar CrossRef Search ADS McCall Leslie 2013 The Undeserving Rich: American Beliefs About Inequality, Opportunity, and Redistribution New York Cambridge University Press . Google Scholar CrossRef Search ADS Messick David M. , Cook Karen S. 1983 Equity Theory: Psychological and Sociological Perspectives New York Praeger . Miller David 1992 . “ Distributive Justice: What the People Think .” Ethics 102 3 : 555 - 93 . Google Scholar CrossRef Search ADS Miller David. 1999 Principles of Social Justice Cambridge, MA Harvard University Press . Mishel Lawrence , Bivens Josh , Gould Elise , Shierholz Heidi 2012 The State of Working America Cornell University Press . Mitchell Gregory , Tetlock Philip E. , Mellers Barbara A. , Ordonez Lisa D. 1993 . “ Judgments of Social Justice: Compromises between Equality and Efficiency .” Journal of Personality and Social Psychology 65 4 : 629 . Google Scholar CrossRef Search ADS Mutz Diana C. 2011 Population-Based Survey Experiments Princeton, NJ Princeton University Press . Google Scholar CrossRef Search ADS Northcraft Gregory B. , Neale Margaret A. 1987 . “ Experts, Amateurs, and Real Estate: An Anchoring-and-Adjustment Perspective on Property Pricing Decisions .” Organizational Behavior and Human Decision Processes 39 1 : 84 - 97 . Google Scholar CrossRef Search ADS Norton Michael I. , Ariely Dan 2011 . “ Building a Better America—One Wealth Quintile at a Time .” Perspectives on Psychological Science 6 1 : 9 - 12 . Google Scholar CrossRef Search ADS Osberg Lars , Smeeding Timothy 2006 . “ “Fair” Inequality? Attitudes toward Pay Differentials: The United States in Comparative Perspective .” American Sociological Review 71 3 : 450 - 73 . Google Scholar CrossRef Search ADS Page Benjamin I. , Jacobs Lawrence R. 2009 Class War? What Americans Really Think About Economic Inequality Chicago University of Chicago Press . Google Scholar CrossRef Search ADS Piketty Thomas , Saez Emmanuel 2006 . “ The Evolution of Top Incomes: A Historical and International Perspective .” The American Economic Review 96 2 : 200 - 05 . Google Scholar CrossRef Search ADS Piketty Thomas 2014 Capital in the Twenty-First Century Cambridge, MA The Belknap Press . Google Scholar CrossRef Search ADS Piketty Thomas , Saez Emmanuel , Stantcheva Stefanie 2014 . “ Optimal Taxation of Top Labor Incomes: A Tale of Three Elasticities .” American Economic Journal: Economic Policy 6 1 : 230 - 71 . Google Scholar CrossRef Search ADS Prentice Deborah A. , Crosby Faye 1987 . “ The Importance of Context for Assessing Deservingness .” Pp. 165 - 82 in Social Comparison, Social Justice, and Relative Deprivation: Theoretical, Empirical, and Policy Perspectives , edited by Masters J. C. , Smith W. P. Hillsdale, NJ Lawrence Erlbaum . Reich Robert B. 2010 Aftershock: The Next Economy and America’s Future New York Vintage . Reich Robert B. 2012 Beyond Outrage: What Has Gone Wrong with Our Economy and Our Democracy, and How to Fix It New York Vintage . Reich Robert B. 2015 Saving Capitalism: For the Many, Not the Few New York Knopf . Rivers Douglas 2006 . “ Sample Matching: Representative Sampling from Internet Panels .” Polimetrix White Paper Series . Rivers Douglas , Bailey Delia 2009 . “ Inference from Matched Samples in the 2008 U.S. National Elections .” Pp. 627 - 39 in Proceedings of the Joint Statistical Meetings . Rivers Douglas 2012 . “ Yougov Poll Performance in the 2012 U.S. Elections ”. Retrieved December 15, 2013 (http://today.yougov.com/news/2012/11/07/yougov-poll-performance-2012-us-elections/) . Rosen Sherwin. 1981 . “ The Economics of Superstars .” The American Economic Review 71 : 845 - 58 . Rosenberg Milton J. 1960 . “ An Analysis of Affective-Cognitive Consistency .” Pp. 15 - 64 in Attitude Organization and Change , edited by Hovland C. I. , Rosenberg M. J. New Haven, CT Yale University Press . Rosenfeld Jake , Denice Patrick 2015 . “ The Power of Transparency Evidence from a British Workplace Survey .” American Sociological Review 80 : 1045 - 68 . Google Scholar CrossRef Search ADS Sandberg Sheryl 2013 Lean In: Women, Work, and the Will to Lead New York Random House . Sayer Andrew. 2005 The Moral Significance of Class New York Cambridge University Press . Google Scholar CrossRef Search ADS Schlozman Kay Lehman , Verba Sidney 1979 Injury to Insult: Unemployment, Class, and Political Response Cambridge, MA Harvard University Press . Schwarz Norbert , Bless Herbert 1992 . “ Constructing Reality and Its Alternatives: An Inclusion/Exclusion Model of Assimilation and Contrast Effects in Social Judgment .” Pp. 217 - 45 in The Construction of Social Judgments , edited by Martin L. L. , Tresser A. Hillsdale, NJ Lawrence Erlbaum . Scott John T. , Matland Richard E. , Michelbach Philip A. , Bornstein Brian H. 2001 . “ Just Deserts: An Experimental Study of Distributive Justice Norms .” American Journal of Political Science 45 : 749 - 67 . Google Scholar CrossRef Search ADS Securities and Exchange Commission 2015 . “ Pay Ratio Disclosure .” Vol. 17 CFR Parts 229 and 249 . Silver Nate 2012 . “ Which Polls Fared Best (and Worst) in the 2012 Presidential Race .” in International New York Times. November 10th . Skitka Linda J. , Bauman Christopher W. , Mullen Elizabeth 2008 . “ Morality and Justice: An Expanded Theoretical Perspective and Review .” Pp. 1 - 27 in Advances in Group Processes Vol. 25 , edited by Hedgvedt K. A. , Clay-Warner J. Bingley, UK Emerald Group Publishing Limited . Google Scholar CrossRef Search ADS Sniderman Paul M. , Grob Douglas B. 1996 . “ Innovations in Experimental Design in Attitude Surveys .” Annual Review of Sociology 22 : 377 - 99 . Google Scholar CrossRef Search ADS Stewart G. Bennett 1990 . “ Remaking the Corporation from Within .” Harvard Business Review 68 4 : 126 - 37 . Stiglitz Joseph E. 2003 The Roaring Nineties: A New History of the World’s Most Prosperous Decade New York, NY W.W. Norton & Company, Inc . Stiglitz Joseph E. 2012 The Price of Inequality: How Today’s Divided Society Endangers Our Future New York, NY W.W. Norton & Company, Inc. Svallfors Stefan 2007 . “ Class and Attitudes to Market Inequality .” in The Political Sociology of the Welfare State , edited by Svallfors S. Stanford, CA Stanford University Press . Google Scholar CrossRef Search ADS Switzer Fred S. , Sniezek Janet A. 1991 . “ Judgment Processes in Motivation: Anchoring and Adjustment Effects on Judgment and Behavior .” Organizational Behavior and Human Decision Processes 49 2 : 208 - 29 . Google Scholar CrossRef Search ADS Thibaut John W. , Walker Laurens 1975 Procedural Justice: A Psychological Analysis L. Erlbaum Associates. Thompson E. P. 1971 . “ The Moral Economy of the English Crowd in the Eighteenth Century .” Past & Present 50 : 76 - 136 . Tversky Amos , Kahneman Daniel 1974 . “ Judgment under Uncertainty: Heuristics and Biases .” Science 185 4157 : 1124 - 31 . Google Scholar CrossRef Search ADS United States Department of Labor, Bureau of Labor Statistics 2012 . “ Occupational Employment Statistics: Occupational Employment and Wages .” Vavreck Lynn , Iyengar Shanto 2013 . “ Online Panels and Experimentation .” Pp. 156 - 68 in The Oxford Handbook of American Public Opinion and the Media , edited by Shapiro R. Y. , Jacobs L. R. OUP Oxford . Verba Sidney , Orren Gary R. 1985 Equality in America: The View from the Top Cambridge, MA Harvard University Press . Walster Elaine , Berscheid Ellen , William Walster G. 1973 . “ New Directions in Equity Research .” Journal of Personality and Social Psychology 25 2 : 151 - 76 . Google Scholar CrossRef Search ADS Walster Elaine , William Walster G. 1975 . “ Equity and Social Justice .” Journal of Social Issues 31 3 : 21 - 43 . Walster Elaine , William Walster G. , Berscheid Ellen 1978 Equity: Theory and Research Boston Allyn and Bacon Walzer Michael 1983 Spheres of Justice: A Defense of Pluralism and Equality New York Basic Books . Weeden Kim A. , Grusky David B. 2014 . “ Inequality and Market Failure .” American Behavioral Scientist 58 3 : 473 - 91 . Google Scholar CrossRef Search ADS Wilson Timothy D. , Houston Christopher E. , Etling Kathryn M. , Brekke Nancy 1996 . “ A New Look at Anchoring Effects: Basic Anchoring and Its Antecedents .” Journal of Experimental Psychology 125 4 : 387 - 402 . Google Scholar CrossRef Search ADS Wright William F. , Anderson Urton 1989 . “ Effects of Situation Familiarity and Financial Incentives on Use of the Anchoring and Adjustment Heuristic for Probability Assessment .” Organizational Behavior and Human Decision Processes 44 1 : 68 - 82 . Google Scholar CrossRef Search ADS Young Cristobal , Varner Charles , Lurie Ithai Z. , Prisinzano Richard 2016 . “ Millionaire Migration and Taxation of the Elite Evidence from Administrative Data .” American Sociological Review 81 3 : 421 - 46 . Google Scholar CrossRef Search ADS Zanna Mark P. , Rempel John K. 1988 . “ Attitudes: A New Look at an Old Concept .” Pp. 315 - 34 in The Social Psychology of Knowledge , edited by Bar-Tal D. , Kruglanski A. New York Cambridge University Press . © The Author 2017. Published by Oxford University Press on behalf of the Society for the Study of Social Problems.All rights reserved. For permissions, please e-mail: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Social ProblemsOxford University Press

Published: Mar 2, 2017

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 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

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

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off