#2855-ASR 71:3 filename:71305-Osberg
“Fair” Inequality? Attitudes toward Pay Differentials: The United States in Comparative Perspective
Lars Osberg Timothy Smeeding Dalhousie University Syracuse University
Are American attitudes toward economic inequality different from those in other countries? One tradition in sociology suggests American “exceptionalism,” while another argues for convergence across nations in social norms, such as attitudes toward inequality. This article uses International Social Survey Program (ISSP) microdata to compare attitudes in different countries toward what individuals in specific occupations “do earn” and what they “should earn,” and to distinguish value preferences for more egalitarian outcomes from other confounding attitudes and perceptions. The authors suggest a method for summarizing individual preferences for the leveling of earnings and use kernel density estimates to describe and compare the distribution of individual preferences over time and cross-nationally. They find that subjective estimates of inequality in pay diverge substantially from actual data, and that although Americans do not, on the average, have different preferences for aggregate (in)equality, there is evidence for: 1. Less awareness concerning the extent of inequality at the top of the income distribution in America 2. More polarization in attitudes among Americans 3. Similar preferences for “leveling down” at the top of the earnings distribution in the United States, but also 4. Less concern for reducing differentials at the bottom of the distribution.
re American attitudes toward economic alized nations, and that federal and state gov- Ainequality different from those found else- ernments in the U.S. do less to reduce the where, and if so, in what ways? It is widely rec- inequality of economic outcomes than do the ognized that economic inequality in the United governments of other countries.1 One hypoth- States is greater than in other affluent industri- esis is that this is what Americans want – that Americans have different attitudes toward inequality and redistribution than do the citizens Direct correspondence to Lars Osberg, Department of other countries, and that government of Economics, Dalhousie University, 6214 University (in)action therefore reflects the preferences of Avenue, Halifax, Nova Scotia, Canada (Lars. 2 [email protected]). Supported by the Russell Sage the electorate. However, Kelley and Evans Foundation. The authors thank the ASR editor and (1993), Kerr (1983), Kluegel, Mason, and anonymous referees for their comments. Participants at the Russell Sage Workshops on Inequality March 21, 2003 and May 21, 2004; the seminars at Yale, 1 For a detailed discussion see Osberg, Harvard, and Syracuse Universities; and Christopher Smeeding, and Schwabish (2004), Smeeding Jencks, Stephen Jenkins, Leslie McCall, John Myles, (2005), and the references therein. Förster and Shelley Phipps, Jeff Racine, and John Roemer have d’Ercole (2005) provide recent international notably improved the article. The authors also thank comparisons of inequality. Kim Desmond, Lihui Zhang, Zhouran Zhou, Laura 2 In the economics literature, Alesina and Turner, Kim Tran, Andrea Johnson, Nan Geng, Mary Angeletos (2005), Alesina, Di Tella, and MacCulloch Santy, Kati Foley, and Lynn Lethbridge for their (2001), Alesina and La Ferrara (2001), Bénabou and excellent work with the data and manuscript. Tirole (forthcoming), Glaeser (2005), and Piketty
AMERICAN SOCIOLOGICAL REVIEW, 2006, VOL. 71 (June:450–473) #2855-ASR 71:3 filename:71305-Osberg
ATTITUDES TOWARD ECONOMIC INEQUALITY—–451
Wegener (1995a), and Wilensky (2002) are lic opinion, we find little evidence for American among those who have argued the alternative exceptionalism in average attitudes. hypothesis: that Americans are not particularly However, “inequality” can be interpreted in different from the citizens of other affluent terms of income ratios or income shares. industrialized nations in social preferences for Individuals’ value-based attitudes toward economic equity and the reduction of econom- inequality (i.e., how much inequality respon- ic inequality. If so, then the explanation for dif- dents think would be “fair”) also are condi- ferences in economic, social, and policy tioned on their personal cognitive estimates of outcomes may perhaps be found in American the extent of inequality (i.e., how much inequal- attitudes toward government as an agent of dis- ity individuals believe actually exists). This arti- tributional change or in differences in the insti- cle begins, therefore, by discussing the tutional structure of American politics. But the conceptualization of “inequality.” It argues that prior question is whether, or how, American the battery of ISSP questions on what individ- attitudes toward economic inequality differ from uals in specific occupations “do earn” and what attitudes elsewhere. they “should earn” offer a particularly focused An international comparison of American way of distinguishing between individual value attitudes toward economic inequality faces, preferences for more egalitarian outcomes and however, three important challenges: other confounding attitudes and perceptions. Average attitudes toward aggregate inequality, 1. Distinguishing attitudes toward inequality of eco- nomic outcomes from beliefs about process equi- as summarized by the Gini index of “should ty or inequality of opportunity earn” inequality from the ISSP data, indicate 2. Clarifying what respondents may understand the that the United States is not particularly differ- meaning of “economic inequality” to be ent from other nations. To find differences 3. Summarizing the distribution of attitudes toward between the United States and other nations in economic inequality in the population. attitudes toward inequality of pay one must Historically, discussion of “American excep- therefore probe deeper and examine both atti- tionalism” (e.g. Lipset, 1996) often has empha- tudes toward inequality in different parts of the sized a presumed American belief in the income distribution and the range of individu- als’ attitudes toward inequality. ideology of mobility and opportunity, a refrain Because a seemingly simple summary term that recently has been reiterated by a number of such as “inequality” melds together perceptions authors in economics (e.g., Bénabou and Tirole, of income differences between the top and the forthcoming). This article starts by reviewing middle of the income distribution, attitudes briefly some of the sociology literature on these toward the gap between the middle classes and topics and by examining simple summary sta- the poor, and preferences for a general leveling tistics on American attitudes toward inequality of pay, this article disaggregates inequality of outcomes and the evidence for a presumed across the distribution. It examines average greater American belief in the prevalence of national perceptions of the maximum and min- equality of opportunity. Using the International imum that people “should earn” and “do earn” Social Survey Program (ISSP)3 surveys of pub- and finds some evidence that American respon- (1995) have discussed possible differences in attitudes dents are, on average, particularly likely to toward inequality in the United States, often in the underestimate the extent of top-end inequality. context of presumed differences in attitudes toward Furthermore, people disagree—sometimes economic mobility. This literature typically makes no quite vehemently—about inequality. The ongo- reference to the International Social Justice Project ing political debates on inequality within coun- or other sociological research that directly examines tries provide direct evidence of heterogeneity in attitudes. For example, Kelley and Evans (1993) and attitudes toward inequality. However, these Kluegel et al. (1995b) cannot be found in the bibli- internal disagreements are obscured when inter- ography of any of the aforementioned papers. 3 Since 1983 the International Social Survey national comparisons rely on average or medi- Program (ISSP) has coordinated the design of cross- an scores to summarize cross-national national surveys covering a variety of social science differences. This article therefore uses kernel topics. Full details are available at http://www.gesis. density methods to describe graphically the dis- org/en/data_service/issp/. tribution of individual preferences for equality #2855-ASR 71:3 filename:71305-Osberg
452—–AMERICAN SOCIOLOGICAL REVIEW in different countries, and shows that an impor- has produced significant socialist or social dem- tant difference between the United States and ocratic parties that have had the reduction of other countries is the bimodal distribution of socioeconomic inequalities as their major objec- American preferences for leveling. tive. Why has the United States been different? Although it is hard to find support for the Authors such as Lipset (1996) and, earlier, hypothesis of systematically different prefer- Lipset and Bendix (1959) have argued that the ences, on the average, for aggregate (in)equal- difference lies in distinctively American beliefs ity in the United States, there is evidence for about, and the reality of, greater socioeconom- 1. Greater underestimation of the size of top-end ic mobility. Belief in the promise of future suc- income differences in the United States cess, either for oneself or one’s children, is said 2. More polarization in attitudes among Americans to dominate any discontent with present inequal- (which is consistent with recent United States’vot- ities, to a uniquely American degree. Many ing behavior and opinion polling) political scientists concur (e.g., Iversen and 3. Similar preferences for “leveling down” at the Soskice, 2005), and Esping-Andersen (1990) top of the earnings distribution (as in other coun- has documented the enduring differences in the tries), but welfare state regimes of advanced capitalist 4. Less concern for “leveling up” at the bottom of nations. the distribution than in other nations. However, the United States is not alone in These findings are consistent with American thinking of itself as “a special case.” trends in political and social polarization, and Comparative historians have noted that nation- may have significant practical implications. al myths, in essentially every country, are almost Glaeser, Ponzetto, and Shapiro (2004) are rep- always based on some presumption of “unique- resentative of a recent political economy liter- ness,” and they also have noted that presumed ature which argues that “strategic extremism” national virtues may bear little relation to sta- by political actors (who must compete both in tistical evidence.4 A functionalist perspective effective mobilization of their own base of sup- (e.g. Parsons, 1960) would argue that there are port and in attracting support from their oppo- strong reasons to expect that affluent capitalist nents) may produce polarization in policy societies will have fundamentally similar atti- positions and attitudes. Although the same medi- tudes toward authority, inherited privilege, and an or average attitudinal score could be pro- economic inequality, given the common struc- duced in a society with a tightly compacted tural imperatives of a market economy and a unimodal distribution of attitudes, or one with democratic polity, together with common pres- a polarized or bimodal distribution of attitudes, sures from technological change, increasing political dynamics are likely to be quite differ- trade, and the globalization of economic and cul- ent in these two situations. Majority rule in a tural life. Inkeles (1998), Kerr (1983), and bimodal society means that the polity will be Wilensky (2002) have argued from the sociol- governed by whatever extreme can (perhaps ogy side that there is a convergence of welfare temporarily) tempt the median voter to its side states, including attitudes and values. (Iversen and Soskice, 2005). The article there- Furthermore, there is general agreement that fore closes with a discussion of the implications the United States is not, in fact, a particularly of a change in distribution of attitudes for eco- mobile society. Sociologists have a long histo- nomic inequality in the United States. ry of comparative studies investigating social class and occupational mobility (e.g., Breen EXCEPTIONALISM OR CONVERGENCE and Jonsson 2005; Erikson and Goldthorpe IN ATTITUDES TOWARD INEQUALITY? 1985, 1992, 2002; Grusky and Hauser, 1984) which conclude that whether income or occu- The intellectual background for this article is the long-standing debate about the “exceptional- ism” of the United States as compared with 4 See, for example, the review essays on “American other affluent, capitalist countries. Popular and Exceptionalism” in which Nelles (1997), Koschmann scholarly writers have, for at least 125 years, (1997) and Nolan (1997) compare Lipset’s claims for wondered why the political process in every American exceptionalism with Canadian, Japanese, affluent capitalist nation except the United States and German assertions of cultural uniqueness. #2855-ASR 71:3 filename:71305-Osberg
ATTITUDES TOWARD ECONOMIC INEQUALITY—–453 pation are used as an index of social status, the tries either “agree” or “strongly agree” with United States is not an exceptionally fluid soci- this statement (there was particularly strong ety, as compared with other nations (see agreement in the transition economies of the Björklund and Jäntti 2000, for both economic former Soviet Bloc). Although the United States and sociological perspectives). As Jantti et al. had a higher percentage that “strongly dis- (2005:2) have recently concluded, “the socio- agreed” with the statement than in most other logical approaches, such as that based on class nations, this represented only 3.2 percent of the mobility, suggest that the United States is fair- respondents. Indeed, in all countries, there are ly unexceptional (Erikson and Goldthorpe 1992, extremely few people who “strongly disagree” 2002). The economics literature, based on cor- with this statement. One message of Table 1 is, relation or regression coefficients, suggests that therefore, the ubiquity of a generalized prefer- the United States may, indeed, be exceptional, ence for “greater equality.” Although respon- not in having more mobility, but in having less dents in some countries are notably more (Solon 2002), a finding that our results with emphatic in saying they “strongly agree” that respect to intergenerational earnings mobility income differences are too large (e.g., France support.” Miles Corak (2004:9) similarly con- with 60.3 percent), several countries had less cluded that “the United States and Britain appear emphatic preferences for equality than the to stand out as the least mobile societies among United States (25 percent ), for example, those rich countries under study. The Nordic Australia (17.8 percent) and Germany (20.5 ).5 countries and Canada seem to be the most Do the data support a distinction between an mobile societies. Germany resembles the United “old Europe” (which may emphasize greater States and the United Kingdom more closely equalization of outcomes because of a greater than it does the other countries.” Finally, Entorf belief that there is inequality of opportunity) and and Minoiu (2004), Erikson et al. (2005), and a “new America” (which may believe that equal- Woessmann (2004) have examined education- ity of opportunity exists, so equalization of out- al opportunities for children from different fam- comes is less imperative)? When respondents in ily backgrounds in western European countries, different countries were asked what character- the United Kingdom, and the United States. istics were necessary to “get ahead in life,” their Woessman (2004:22) concluded that “the results perceptions of “equality of opportunity” can of this paper are generally in line with the broad perhaps be gauged partly by whether they pattern of the existing cross-country evidence thought “knowing the right people” was impor- on intergenerational earnings mobility, which tant. Coded responses ranged from 1 (essential) found that the United States and the United to 5 (not important at all). On this item, the Kingdom appear to be relatively immobile soci- United States’ 1999 score (2.58) was at the eties.” “fairly necessary” end of the spectrum. All this evidence on actual comparisons of “Knowing the right people” was seen in the intergenerational socioeconomic mobility does United States as slightly less essential than in not preclude the possibility that beliefs in future Canada (2.55), similar to the view in the mobility might preempt discontent with cur- Philippines (2.58), but considered to be slight- rent inequality, although it might seem to make ly more essential than in France (2.62) or the it less likely. But the crucial prior question is United Kingdom (2.65). American attitudes whether or not Americans actually differ from averaged 2.65 in 1992 and 2.61 in 1987. That other nationalities in their attitudes toward is, “knowing the right people” became seen as inequality. even more “essential” over this period. A seemingly straightforward way to find out Interestingly, in their subjective perception of whether people in different countries have dif- greater barriers to mobility than in Western ferent attitudes toward economic inequality is Europe, American respondents were in agree- to ask them directly. Table 1 reports the respons- ment with recent literature on intergenerational es in 27 countries to the ISSP 1999 survey mod- ule on Social Inequality when individuals were asked the seemingly simple question: “In (your 5 The 1992 and 1987 ISSP surveys cover fewer country), are income differences too large?” It countries, but with the same conclusion. See Osberg is noteworthy that clear majorities in all coun- and Smeeding (2006). #2855-ASR 71:3 filename:71305-Osberg
454—–AMERICAN SOCIOLOGICAL REVIEW
Table 1. Attitudes to Inequality: Are Income Differences Too Large? (1999)
Strongly Neither Agree Strongly Country Agree Agree Nor Disagree Disagree Disagree Total Bulgaria 84.0 12.8 1.4 .8 1.0 100 Portugal 82.3 13.8 1.7 1.4 .9 100 Russia 79.5 16.0 2.2 1.2 1.1 100 Slovakia 73.9 19.8 4.6 1.2 .6 100 Hungary 67.2 25.9 3.5 3.1 .4 100 Czech Republic 60.3 27.5 6.0 4.2 2.1 100 France 60.3 27.2 7.4 4.5 .7 100 Latvia 57.2 39.5 1.8 1.3 .2 100 Israel 53.9 36.0 3.9 5.5 .8 100 Slovenia 49.7 41.3 4.8 3.6 .6 100 Poland 46.8 42.3 6.2 3.9 .8 100 Germany East 45.0 48.6 4.4 2.0 .— 100 Chile 42.8 49.4 3.4 4.4 .1 100 Austria 40.4 45.8 9.1 4.7 .— 100 Japan 38.6 30.5 18.3 7.5 5.0 100 Spain 35.9 53.4 7.4 3.1 .3 100 Great Britain 31.7 50.6 11.6 5.4 .6 100 New Zealand 29.4 43.8 13.5 11.8 1.6 100 Sweden 29.2 41.9 18.1 8.4 2.4 100 Canada 28.1 42.5 15.7 11.2 2.6 100 United States 25.0 41.2 21.5 9.2 3.2 100 Norway 22.4 50.1 13.8 12.0 1.8 100 Philippines 22.3 43.1 16.9 14.6 3.3 100 Germany West 20.5 55.2 14.3 9.1 .9 100 Australia 17.8 53.1 17.1 11.6 .4 100 North Ireland 17.4 52.1 21.4 8.4 .7 100 Cyprus 12.2 53.4 21.8 12.5 .1 100 Source: ISSP 1999 income mobility (see references in Section 1 ferences in income are necessary for [R coun- earlier).6 try’s] prosperity.” A cell value such as 2.5 on the Table 2 also probes rationalizations for “benefits the rich” question can be read as say- inequality. Columns 2 and 3 report the popula- ing that, on the average, a country’s population tion average responses on a scale ranging from is about evenly split between “agree” and “nei- 1 (strongly agree) to 5 (strongly disagree) for ther agree nor disagree.” This particular ques- respondents’ evaluation of statements such as tion is a fairly strongly worded item that may tap “inequality continues to exist because it bene- into latent class antagonisms, particularly the fits the rich and the powerful” and “large dif- perception of capitalism as a rigged game and “unfairness” as the underlying explanation for inequality. Apparently, many people buy this 6 In responses to an item in the 1992 and 1987 ISSP idea, at least somewhat, in all the countries sur- asking whether “having well-educated parents” is veyed. For 1999, the average responses of important for getting ahead in life, the average score Americans (2.64) are bracketed by those of in the United States (2.72, 2.76) and Italy (2.78, 2.8) Hungarians (2.58) and Filipinos (2.67).7 were similarly situated in the range between 2 (very important) and 3 (fairly important), ascribing some- what more importance to well-educated parents than in Germany (2.99, 2.8) or Austria (2.95, 2.69). In 7 The 1999 U.S. survey is an outlier, taken near the 1992, Canadians averaged 2.97 on this item, where- peak of the stock market and information technolo- as Swedes averaged 3.16 and Norwegians averaged gy bubbles and at a time when unemployment was 3.48 (i.e., significantly closer to “not very important” at its lowest level for a generation. The comparable [4]). This item was not asked in 1999. 1992 value for the United States “benefits the rich” #2855-ASR 71:3 filename:71305-Osberg
ATTITUDES TOWARD ECONOMIC INEQUALITY—–455
Table 2. Opinions about Inequality (1999)
Knowing the right people— Inequality continues to exist Large income differences how important is that for because it benefits the rich are necessary for a getting ahead in life?a and powerful.b a country’s prosperity.b Cyprus 1.90 2.56 3.87 Slovakia 2.01 2.20 4.18 Poland 2.06 2.09 3.35 Austria 2.09 2.21 3.76 Bulgaria 2.16 2.01 4.12 Israel 2.18 2.40 3.34 Germany East 2.19 1.98 3.49 Russia 2.22 1.93 4.05 Spain 2.27 2.09 3.33 Slovenia 2.32 2.13 3.61 Latvia 2.34 2.03 3.76 Chile 2.41 2.12 2.91 Germany West 2.41 2.23 3.22 Portugal 2.41 1.83 3.59 Sweden 2.45 2.42 3.41 Czech Republic 2.46 2.36 3.70 Canada 2.55 2.38 3.65 Philippines 2.58 2.67 2.62 United States 2.58 2.64 3.19 France 2.62 1.91 3.74 Great Britain 2.65 2.42 3.48 Hungary 2.67 2.58 3.93 Australia 2.73 2.35 3.33 New Zealand 2.77 2.45 3.54 North Ireland 2.80 2.50 3.45 Norway 2.83 2.29 3.50 Japan 3.21 2.08 3.30 Source: The International Social Survey Programme. a Coded as: 1 (essential) to 5 (not important at all). b Coded as 1 (strongly agree) to 5 (strongly disagree).
Objectively, as Burtless and Jencks (2003) try’s] prosperity.” An average response such as and Osberg, Smeeding, and Schwabish (2004) 3.19 can be read as equivalent to about one noted, there is no good evidence that more fifth of Americans being on the “disagree” end inequality produces more of any good thing, of the range between “neither agree nor dis- especially economic and social prosperity. agree” (3) and “disagree” (4). It is notable that However, political trends depend on the sub- jective assessments by citizens of the rationale in 1999 the differences between the United for inequality. Presumably, even if greater States (3.19) and West Germany (3.22) were inequality is undesirable in itself, one might minimal.8 (In both the United States and the accept it as a “necessary evil,” a price that must United Kingdom there was a noticeable trend be paid if society as a whole desires prosperi- over time toward greater percentages of the ty. Do the citizens of modern capitalist nations, population disagreeing with this “instrumen- on the average, buy into this rationale for tal” rationale for inequality.) inequality? Column 3 of Table 2 reports aver- age responses to the item proposing that “large differences in income are necessary for [R coun- 8 With relatively large sample sizes, country dif- ferences in means generally pass a test of statistical significance, even if the empirical difference is not item was 2.51. It remains to be seen whether 1999 is large (i.e., one often can be statistically sure of a a blip or a true structural break. socially insignificant difference). #2855-ASR 71:3 filename:71305-Osberg
456—–AMERICAN SOCIOLOGICAL REVIEW
As Osberg and Smeeding (2006) have doc- Some people may have an idea of minimum umented in greater detail, the ISSP asks about adequacy in a just society, that is, a lower bound attitudes toward social inequality in a number (Y*min) on incomes, or what Smith of overlapping ways. The key point is that the (1776[1961]:339) referred to as “those things United States is not a clear outlier when mean which the established rules of decency have responses are compared across nations (see also rendered necessary to the lowest rank of people.” Kelley and Evans 1993; Kluegel et al. 1995a; Equation 2 expresses this idea. Suhrcke 2001:8; and Svallfors 1997). When A Yi > Y*min (2) Americans and Europeans are asked whether a good education, ambition, natural ability, or Furthermore, some individuals may have the hard work enable an individual to “get ahead in idea that it would be socially excessive if any life,” evidence of an attitudinal difference individual’s actual income exceeded some upper between the average respondent in the United bound (Y*max), as expressed in Equation 3. A States and those in other nations is hard to find. Yi < Y*max (3) If it were true that Americans tolerate more inequality of outcomes because they believe A just society could therefore be summa- there is more equality of opportunity in the rized as one that satisfies Equations 1 to 3, and that can therefore be described in graphic terms United States, then one would expect to find a as having a distribution of earnings resembling tendency for Americans to ascribe more impor- the 45-degree line in Figure 1. Up to this point, tance to personal characteristics for “getting the vocabulary does not exclude any of the pos- ahead” than is the case elsewhere. But, on the sible belief sets about an ethically acceptable dis- average, other countries are sometimes higher tribution of earnings. The beliefs of a complete and sometimes lower than the United States in egalitarian can, for example, be summarized as the importance their citizens ascribe, on the constraining Equations 2 and 3 such that average, to individual personal characteristics. Y*max = Y*min, in which case the line col- lapses to a single point, and there is a single CONCEPTUAL AMBIGUITIES IN THE answer to the twin questions “What should I MEANING OF “INEQUALITY” receive?” and “What should other people get?” Alternatively, some people might believe that However, although there may not be much dif- there should be no upper limit on ethically ference in average responses to summative ques- acceptable incomes. If so, Equation 3 loses any tions, what do survey respondents mean to say empirical content because Y*max is infinitely when they answer general questions about large. Alternatively, if one thinks there should “inequality” or the fairness of “income differ- be no lower limit to earnings, that amounts to ences”? specifying, in the terms of Equation 2, that One way to fix ideas about attitudes toward Y*min = 0. inequality of outcomes is to suppose, by con- In the ISSP data, very few people say they trast, that an individual believed he or she lived believe in completely equal earnings.10 Aside 9 in a just society. In this case, such a person from such complete egalitarianism, all belief A would believe that the actual earnings (Yi ) of systems about ethically acceptable earnings all persons (both him- or herself personally and inequality share the property that if a person all other individuals) are equal to what they believes he or she lives in a just society, and if should earn (Yi*). Equation 1 summarizes the that person is asked to estimate the relation- A idea that people should earn what they do earn. ship between what other people “do earn” (Yi ) A Yi* = Yi (1) 10 The ratios of egalitarians to respondents in the 1987, 1992, and 1999 Social Inequality waves of the 9 A huge and fascinating literature on procedural ISSP in the United States were, respectively, 7/1165, justice (e.g., Molm, Takahashi, and Peterson, 2003) 6/1132, and 2/988. Among the 35,656 respondents in invariably finds that “process matters” for fairness all surveys in all countries, only 212 (0.59 percent) judgments. But in this article, we focus on the per- replied that all individuals should have the same ceived equity of outcomes. wage. #2855-ASR 71:3 filename:71305-Osberg
ATTITUDES TOWARD ECONOMIC INEQUALITY—–457
Figure 1. “Fair Pay” and Actual Earnings
and what they “should earn” (Yi*), a regression is systematic inequity in earnings can be thought of the form of Equation 4 would yield the result of as the belief that some people get “too much” A A that b0 = 0 and b1 = 1. (Yi > Yi*) while others get “too little” (Yi < Y * = b + b Y A (4) Yi*). In graphic terms, such a perception of i 0 1 i inequity can be represented as the line de in As it happens (see later), some people appear Figure 1, whose slope (b1 < 1) can be taken as to believe, at least approximately, that the earn- indicative of an individual’s desire for “leveling” ings distribution is fair (i.e., there is a fraction of the earnings distribution, within their view of of the population whose personal estimates the acceptable range of incomes.12 11 imply b0 = 0 and b1 = 1) , and in Figure 1, the In the remainder of this article, we adopt the 45-degree line expresses this general idea that convention of referring to b1 as an estimate of A “should earn” equals “do earn” (Yi = Yi*). However, in all countries, many people do not share this belief. An individual’s belief that there 12 Note that the line segment de is drawn with a positive intercept b0. If some people get “too much” while others get “too little,” one must expect b0 > 0 11 This could be because individuals rationalize the and b1 < 1. Jasso (1978, 1980) expressed the “justice current reality of their society (“what is ought to evaluation” (JE) of an outcome as JE = ln (actual be”) or because reality fits their prior social justice earnings/just earnings) which, in terms of the current values (“what ought to be is”). For current purpos- discussion, implicitly assumes that b0 = 0 and b1 < 1 es, we do not need to distinguish between reasons why and implies that “just” incomes are always less than b1 = 1. Note that this article focuses on the individ- actual incomes. The “Jasso ratio” is equal to the uals’ evaluation of the fairness of the distribution of antilog of b1 under the assumption that b0 = 0. See economic rewards among others and does not address Alwin 1987; Shepelak and Alwin (1986); Alwin the determinants of any personal dissatisfaction that (1987); Wegener and Steinmann, ( 1995:156); Younts individuals may have with their own rewards. and Mueller (2001. #2855-ASR 71:3 filename:71305-Osberg
458—–AMERICAN SOCIOLOGICAL REVIEW individual “preferences for leveling,” which can effect to a given belief system, or whether that be estimated, for any given person, across that set of taxes and transfers is feasible.13 person’s responses identifying “should earn” When survey respondents use the term A pay (Yi*) and “do earn” pay (Yi ) in a set of “inequality,” they might mean to describe the occupations. However, Equations 2 to 4 also income ratios of individuals, or they might mean can be read as indicating that three numbers the income shares of groups in a population. So are needed to express the degree of a person’s far, this section of the article has been examin- egalitarian preferences: ing “economic inequality” in the sense of “dif- ferences between individuals in economic 1. The ethical floor to minimum earnings (i.e., Y*min) 2. The ethical ceiling to maximum earnings (i.e., outcomes.” Y*max) However, if individuals are to evaluate 3. The desired degree of leveling relative to the cur- inequality in the “distribution of income shares rent income distribution, among “acceptable” within a population” sense, they must estimate incomes (i.e., b1). both income ratios and how many people have A person with a belief system summarized particular levels of income (i.e., they must esti- graphically by line segment de would perceive mate f[y], the relative frequency of different levels of income). A good deal of evidence that someone at income Y1 “should earn” more A exists to show that survey respondents do not than he or she “does earn” (i.e., Y1* > Y1 , which implies a gap between actual and fair accurately estimate the proportion of the pop- incomes for people at the bottom of the distri- ulation with particular incomes. For example, A Kluegel et al. (1995a:201) have reported that bution with an actual income such as Y1 ). Graphically, because “should earn” is more than subjective estimates of the perceived frequen- “does earn,” the line de is plotted as lying above cy of “middle class” incomes depend heavily on the 45-degree line at that point. the respondent’s own socioeconomic position. Evans and Kelley (2004) also noted that there In Figure 1, one can call income level Yj the “just desserts” income because “should earn” is a tendency for survey respondents to place A themselves “in the middle” of the income dis- equals “do earn” income (Yj* = Yj ). Graphically, the line de intersects the 45-degree tribution. The problem for empirical work is line (which expresses the general idea that that asking people about their attitudes toward “should earn” equals “do earn”) at income Y . income shares implicitly requires respondents j to estimate both income ratios and the relative If the relationship between “should earn” (Yi*) A size of population groups, whereas asking only and “do earn” (Yi ) is linear, as in Equation 4, the point of intersection, or the “just desserts” about their attitudes toward income ratios makes much smaller informational demands.14 income, can be calculated as equal to b0/(1 – b1). On the other hand, in Figure 1, an individual In this regard, a fascinating series of questions A in the ISSP rounds of 1999, 1992, and 1987 dis- making more than Yj (i.e., at an earnings level A such as Y2 ) is someone who, according to belief system de, earns “too much” income (Y2* A < Y2 ). Graphically, because “should earn” is 13 Note that the political and ethical attitudes of A less than “do earn” at income level Y2 , the individuals are only in a very vague sense constrained line de lies below the 45-degree line. In practi- by actual budgetary feasibility, and that a different A cal terms, income level Y2 also could be seen belief system, as represented by different values of as a social problem of excess that might possi- b0 and b1, may identify differing income levels as bly be solved by taxation. defining “deprivation” or excessive rewards (see However, the question of how much society Section 1 of the Online Supplement on the ASR Web should tax or spend, in aggregate, cannot be site: http://www2.asanet.org/journals/asr/2006/ addressed by Figure 1, because it contains no toc051.html.) 14 In general, if y is a person’s income, and if the information about the percentage of the popu- i person’s characteristics are described by a vector Xi, lation that is at each level of actual income. and the returns to those characteristics are summa- Without information as to the population den- rized in the vector with the unexplained component A sity of Yi , one cannot know what the income ui, where E(ui ) = 0, then one can write individual shares of rich and poor are, or what aggregate income as yi = Xi + ui . The frequency distribution volume of taxes and transfers is required to give f(y) and any inequality statistics calculated from it #2855-ASR 71:3 filename:71305-Osberg
ATTITUDES TOWARD ECONOMIC INEQUALITY—–459
tinguished between subjective empirical esti- been in all three waves of the ISSP (notably the mates of inequality and the ethical evaluations United States, the United Kingdom, Germany, that people may have of those perceptions. and Australia), but others have been more Respondents were first asked to estimate what episodic. salaries people in various jobs do actually earn, General questions about inequality can min- then what persons in each occupation should gle empirical beliefs regarding the magnitude of earn. In contrast to the large literature that has income ratios, the frequency density of incomes, analyzed the statistical data to measure objec- and the processes that determine income levels, tive trends in income inequality, these data as well as ethical evaluations of both process and enable examination of the issues that actually are outcomes. In a general discussion of inequali- more relevant to individual behavior, namely, the ty, participants make implicit empirical esti- subjective estimates that individuals have of mates of the importance of capital income for income inequality and the subjective evaluation “the rich” and the processes that generated mar- of this perceived degree of inequality relative to ket income (e.g., discrimination or the extent of an individual’s own norms of “fair” income dif- inherited wealth). They implicitly guess the size ferentials. and frequency of transfer payments, and they In the 1999 ISSP questions about what spe- mingle these estimates with their attitudes cific jobs do pay and what they should pay, the toward inequality of outcome and opportunity. jobs considered included those of skilled factory Survey respondents’ subjective awareness of worker, doctor in general practice, chairman of the size and distribution of income sources is a large national company, lawyer, shop assistant, subject to great empirical errors, and there is owner/manager of a large factory, judge in the much controversy in the ethical evaluation of country’s highest court, unskilled worker, and income-generating processes. federal cabinet minister.15 These classifications A key advantage of using the “do earn/should are similar to those contained in the sociologi- earn” question format is that many of these cal “class” literature on occupations and socioe- confounding issues are held constant at the conomic status, most recently from Erikson and respondent level. In the ISSP data, attitudes Goldthorpe (2002) and Erikson et al. (2005), but toward what specific occupations “should earn” taken earlier from Erikson and Goldethorpe can be conditioned on what the individual (1985) and Hauser and Warren (1997), and later believes they “do earn” so that individual errors from Rose and Pevalin (2003). The occupations in estimating actual earnings can be directly considered in 1992 also included owner of a controlled for. Moreover, the “do earn/should small shop and farm worker, whereas the 1987 earn” ISSP questions are clearly restricted to dif- questionnaire also inquired about city bus driv- ferences in labor market earnings of specific er, secretary, brick layer, and bank clerk (but not occupations, thereby avoiding the complex set shop assistant or lawyer). Several countries have of issues surrounding the importance and eval- uation of different income sources. Respondents are not asked to consider any vignettes detail- ing complexities of household size, multiple (e.g., the coefficient of variation or the Gini or Theil earners, or other factors affecting household indices) depend on f(Xi ), , and ui, but inequality in composition or “need” for income. The ISSP the “average income ratio between types of persons” questions are phrased in terms of occupational sense is only about . earnings—the foundation of sociological “class” 15 Respondents also were asked about the income measurement as seen in Erikson et al. (2005) from their own occupations, but in this article, we and Rose and Pevalin (2003)—and there is lit- exclude these data because our focus is on attitudes tle reason for respondents to systematically toward inequality in society, not perceived personal impute a different age, race, disability status, injustice. We experimented with using or not using number of household members, or aggregate the data on what judges and cabinet ministers “do earn” and “should earn” because we worried that income of other household members to any of these responses may mingle individual attitudes the occupations listed. Hence, the “do toward government with preferences for leveling in earn/should earn” questions are not confound- occupational rewards, but in practice, it makes no ed by concern with the adequacy or excess of detectable difference. household consumption possibilities driven by #2855-ASR 71:3 filename:71305-Osberg
460—–AMERICAN SOCIOLOGICAL REVIEW number of household members, disability sta- among a range of occupations diverges from his tus, age, race, or the like. The implied context or her own estimate of “fair” inequality within for each occupation is full-time earnings, which this range of comparable occupations. abstracts from the differences in income pro- Table 3 presents the results for twenty seven duced by variations in labor supply, unemploy- nations, and Figure 2 plots the average values ment, or the number of earners in a family. The of GiniE and GiniA by country. The first col- ISSP data thus enable us to strip away many con- umn of Table 3 shows that the average percep- founding variables to see whether we find evi- tion of earnings inequality in the United States dence for “American exceptionalism” in was not very different from that in Australia, attitudes toward inequality, or evidence for a New Zealand, Canada, or Germany, despite broadly similar value base in affluent industri- very substantial real differences in earnings alized market economies. inequality in these nations (Gottschalk and One approach to the “do earn” and “should Smeeding, 1997, 2000). Indeed, the average earn” data is to use the Gini index to summa- subjective perception of earnings inequality in rize each ISSP respondent’s attitudes toward the United States was below the average of all inequality in pay. Specifically, in this article countries. we calculate both (1) the respondent’s estimate In Column 2, countries are compared in terms of the actual degree of pay inequality among the of the average subjective perception of inequal- listed occupations (as summarized here by ity in what people “should earn.” In all countries, GiniA, the Gini index of inequality16 of the some level of earnings inequality is accepted as respondent’s estimates of “do earn” income) ethically justifiable, but there is a substantial and (2) the respondent’s perception of “fair” range from the most egalitarian attitudes inequality in earnings (as summarized by GiniE, (Slovakia at 0.19 and Norway at 0.21) to the the Gini index of inequality across what occu- least egalitarian attitudes (Chile at 0.47 and pations “should earn”). Because the occupa- Philippines at 0.46). The United States is right tions enumerated in the ISSP questions are a in the middle, with an average level of “should subset of all occupations, because we have no earn” inequality at about 0.35, very close to information on the respondent’s estimate of the the European and all nations average of 0.34. frequency of each occupation in the popula- Column 3 of the table is the one that arguably tion, and because inequality of earnings with- has the most direct implications for the politi- in each occupation is not addressed, GiniA is not cal process because it presents the average dis- an estimate of actual inequality in the labor crepancy between perceived actual and market as a whole. Rather, GiniA is a summa- perceived fair outcomes, that is, the average ry estimate of perceived pay inequality among (across persons) of the ratio between each per- a broad range of internationally comparable son’s estimates of “should earn” inequality occupations. Hence, the ratio between GiniE (GiniE) and “do earn” inequality (GiniA). In and GiniA is, for each respondent, an indication every country, in every year, the average respon- of how much the respondent’s own personal dent thinks there should be less inequality than estimate of the actual degree of inequality in pay he or she thinks actually exists. The average “should earn” to “do earn” inequality ratio is always substantially less than 1. As Column 3 16 In doing this calculation, the implicit assump- indicates, in 1999 the average “tension” between tion is an equal number of people in each occupation, perceived fair earnings inequality (i.e., “should which clearly is not what any respondent actually earn” inequality) and perceived actual “do earn” believes is empirically true, but does standardize rel- inequality was about 0.75. For the average ative population weights for occupations across all American respondent, “should earn” inequali- respondents. Other summary indices (e.g., coeffi- ty was a bit closer to “do earn” inequality than cient of variation, Theil Index) of both “should earn” in most other nations (at 0.82), because “do and “do earn” inequality also have been calculat- ed—with very much the same implications—but to earn” inequality was estimated to be lower than conserve space, are not reported here. Szirmai (1991) elsewhere. used Dutch data to calculate the percentage difference Figure 2 depicts the data in another way, by in the Theil Index of “should earn” and “do earn” plotting the relationship, across countries, inequality as an index of “tendency to equalize.” between average perceptions of “fair inequali- #2855-ASR 71:3 filename:71305-Osberg
ATTITUDES TOWARD ECONOMIC INEQUALITY—–461
Table 3. Actual and Ethical Inequality - Gini 1999
Average Gini Index Average Gini Index of Salaries People of Salaries People Average Ratio of “Do Earn” (GiniA) “Should Earn” (GiniE) GiniE/GiniA Russia .66 .39 .61 Chile .60 .47 .79 Poland .58 .44 .77 Latvia .58 .41 .70 Hungary .56 .37 .67 Czech Republic .53 .39 .76 France* .52 .38 .74 Philippines .49 .46 .97 Great Britain .49 .36 .73 Slovenia .47 .34 .74 Japan .46 .37 .81 Israel .45 .36 .80 Canada .45 .33 .76 Portugal .45 .33 .73 United States .43 .35 .82 New Zealand .43 .32 .76 Germany East .43 .32 .74 North Ireland .42 .32 .76 Australia .42 .31 .74 Bulgaria .42 .28 .68 Germany West .41 .34 .82 Austria .41 .32 .78 Cyprus .40 .33 .82 Sweden .35 .22 .65 Spain* .34 .22 .65 Norway .30 .21 .73 Slovakia .25 .19 .82 Average – All Nations .46 .34 .75 Average of Europe .47 .34 .74 Source: International Social Survey Programme. Note: Respondents were asked what salaries people in various jobs do actually make and what they should make. (Spain and France reported “net income” but other nations asked for “Before Tax” salary) Jobs considered included skilled factory worker, doctor in general practice, chairman of a large national company, lawyer, shop assistant, owner/manager of a large factory, judge in the country’s highest court, unskilled worker and federal cabinet minister. Gini Indices were calculated for each respondent if they answered more than seven jobs in both the ‘do earn’ and ‘should earn’ categories, and if the jobs answered in the ‘do earn’ and the ‘should earn’ cate- gories were the same. ty” in what occupations “should earn” and aver- 2 does clearly indicate that the United States is age perceptions of “actual inequality” in what not an outlier, at least in average responses. occupations “do earn.” As the regression line There is, therefore, little basis in the ISSP data indicates, there is a strong correlation (R2 = for an argument that Americans are, on the aver- 0.78). At the margin, when average perceived age, more or less tolerant of earnings inequal- 17 actual inequality is higher, average “fair” ity than the citizens of other countries. inequality is higher by about two thirds (0.674) as much. Because a cross-sectional correlation 17 cannot show causation, Figure 2 cannot show This similarity in attitudes toward earnings whether habituation to higher actual inequality inequality occurs in the context of substantially differing levels of social transfers and public expen- produces higher norms of inequality, or whether ditures (see Osberg, Smeeding, and Schwabish 2004; less ethical aversion to inequality produces Schwabish et al. forthcoming). If the issue in evalu- greater actual inequality. Nevertheless, Figure ating inequality is “inequality in consumption #2855-ASR 71:3 filename:71305-Osberg
462—–AMERICAN SOCIOLOGICAL REVIEW
Figure 2. “Actual” and “Ethical” Inequality
However, Figure 2 presents a highly aggregat- rence across countries on which occupations ed picture of attitudes, in two senses: (1) the atti- “should earn” the most and which the least,19 tudes about the inequality of all individuals and the list of occupations contains an example within each country are averaged and (2) from both the very top (chairman of a large “inequality” is summarized by a single num- national company) and the very bottom ber— the Gini index. (unskilled worker) of the earnings distribution. But is it inequalities at the top or at the bottom PROBING DEEPER: IS IT INEQUALITY that people care about the most? In this article, AT THE TOP OR AT THE BOTTOM Equation 2 expresses the idea that individuals THAT MATTERS MOST? may believe in a minimum “should earn” Calculation of a single summary measure of inequality (such as the Gini index) does not show directly whether individuals are, on the income inequality (such as the Gini ratio, the Theil average, more accepting of inequalities at the top Index, or the coefficient of variation) often produce or at the bottom of the distribution.18 In the ambiguous international rankings when frequency ISSP data, there is a broad measure of concur- distributions of income differ such that the Lorenz curves of the cumulative distribution cross. Therefore, it is essential to specify which part of the distribution is of primary concern. It is “inequality” in this sense possibilities,” then a higher common “social wage” of the unequal shares of income in a population implies relatively less importance for market income which is the focus of much of the economics litera- as a source of effective consumption, an argument that ture, particularly that cited in Footnote 2. would have predicted less emphasis on inequality of 19 We have compared across countries the “should earnings in the Scandinavian countries. earn” and “do earn” occupational rankings, which are 18 Atkinson (1970) noted that comparisons of essentially the same in the countries examined (see inequality in income shares using different indices of also Kelley and Evans 1993). #2855-ASR 71:3 filename:71305-Osberg
ATTITUDES TOWARD ECONOMIC INEQUALITY—–463
income (Y*min), whereas Equation 3 describes tribution) show that both the 1992 and 1999 the maximum “should earn” income (Y*max) data put the United States almost exactly in the estimated by each respondent. middle of the nations surveyed. However, eth- To examine the full range of “fair inequali- ical values are conditioned on what individuals ty” in pay, the first four columns of Table 4 believe to be the actual inequality of earnings. present data on the maximum/minimum “should Even if the average American is not exception- earn” ratio in 1999 ISSP data for affluent, con- al in what the maximum/mean ratio should be, tinuously capitalist countries. Columns five to he or she differs from individuals in other eight present the maximum/mean ratio as an nations in the degree to which he or she under- estimate of aversion to excess at the top. That estimates top-end earnings. is, for each respondent, it compares the respon- Because the ISSP data identify specific occu- dent’s estimate of maximum “should earn” pations, respondents’ subjective estimates of income (Y*max) expressed as a ratio of the mean what occupations “do earn” can be compared “do earn” income that he or she estimates. The with objective data on actual earnings. Although last four columns attempt to get at dislike of dep- the objective data show a much larger, and rivation at the bottom by presenting the widening, gap between average earnings and mean/minimum ratio (i.e., the ratio of each executive compensation in the United States respondent’s average estimate of “do earn” than is characteristic in other countries, sub- income to their estimate of minimum “should jective (mis)perceptions of “do earn” inequal- earn” income [Y*min]). As indicators of the cen- ity are greater in the United States, a fact likely tral tendency of the distribution of attitudes to mute pressure for distributional change. toward each issue, Table 4 presents both the Table 5 indicates that the actual earnings mean and the median attitude, calculated across ratio between production workers and chief all respondents in each country.20 executive officers (CEOs) varies between In the 1999 data, there are big differences approximately 20:1 and 45:1, a ratio far greater between countries in the overall range of accept- than the subjective “do earn” estimates. In all able outcomes (e.g., the median French response countries, the average “do earn” estimate for for the maximum/minimum ratio was about manufacturing workers is remarkably close to three times the median Norwegian maxi- actual data.21 However, the subjective estimates mum/minimum ratio). However, these differ- of CEO compensation are well below objec- ences are driven largely by differing attitudes tive data. The degree of CEO compensation toward inequality at the bottom. Indeed, it is misestimate varies widely across countries, with remarkable how small the cross-national dif- the average American respondent particularly ferences are in ethically acceptable income likely to underestimate CEO pay. ratios at the top (in 1999, the median Spaniard’s How much do respondents think income dif- maximum/mean ratio was lowest, at 1.556, ferences should be compressed? The ISSP data whereas the median French ratio was largest, at show a general consensus of opinion, both with- 2.166). A look at median attitudes shows that in and across nations, on the rank hierarchy of cross-national differences are most apparent at occupations in both “do earn” and “should earn” the bottom of the distribution, where the range income. However, although individuals gener- is from 3.487 in France to 1.667 in Norway. Again, in these data on attitudes toward the range of inequality, there is little support for the hypothesis of “American exceptionalism” in 21 Some discrepancy might be expected because values. The median and mean maximum/mean the Bureau of Labor Statistics (BLS) data are for “should earn” ratios (i.e., the “average person’s” “production” workers in manufacturing, whereas the tolerance of inequality at the top end of the dis- ISSP asks about “skilled” workers in manufacturing. Both correspond to the “working class” occupations found in Erikson et al. (2005) and Rose and Pevalin (2003). For distribution of subjective estimates of the 20 Table S1 presents the comparable 1992 results objective CEO/worker “do earn” pay ratio in the and Table S2 presents the 1987 data in the Online United States, the United Kingdom, Canada, and Supplement, Section 2 (ASR Web site: http://www2. Germany, see Figure S2 (Section 2, Online asanet.org/journals/asr/2006/toc051.html). Supplement, ASR Web site). #2855-ASR 71:3 filename:71305-Osberg
464—–AMERICAN SOCIOLOGICAL REVIEW Country Rank Country Rank Country Rank Max/Min Ratio Max/Mean Ratio Mean/Min Ratio 9.7 6.7 5 2 2.1 2.0 6 7 4.0 3.2 5 4 Mean Median Mean Median Mean Median Mean Median Mean Median Mean Median a Mean and Median “Should Earn” Ratios Across Countries: 1999 Means, Medians and Rankings (All Individuals) Mean and Median “Should Earn” Ratios Max = maximum; Min minimum. Table 4. Table JapanFranceUnited KingdomCanadaUSA 1999 10.9 12.3 11.6 6.7 10.2 6.5 7.5 3 6.7 1 2 2United States 1987 4 3Note: 1 2 2.3 11.1 2.3 2.3 6.7 2.1 2.2 2.0 2.2 2 2.0 3 1 2 5 4 1 6 4.1 3.0 4.5 4.4 3.3 2.7 4.1 3.1 3.5 4 3.2 1 2 2 3 5 1 3 3.1 2.4 North IrelandAustria ZealandNew IsraelPortugalGermanyAustralia 8.1SwedenNorway 8.0Spain 5.6 8.1United States 1992 5.6 5.3 7.8 7.7 6 7.6 6.1 6.0 5.3 8 12.6 6.0 4.0 7 5 3.2 5.0 8.0 2.9 10 6 3.1 9 11 2.6 7 12 2.8 2.0 13 7 4 4 14 2.0 9 1.9 15 2.0 10 12 1.9 2.0 1.9 2.2 11 2.1 10 1.9 1.9 1.7 2.1 9 2.0 1.6 11 8 2.4 1.8 1.6 1.6 11 10 1.6 4 2.2 7 9 12 1.6 3.6 13 8 14 3 5 12 3.5 3.0 15 13 3.5 14 2.9 3.5 15 3.3 6 2.9 3.3 3.0 2.1 2.8 8 1.9 2.9 2.9 6 2.7 7 1.8 1.8 4.6 10 1.7 9 11 10 12 1.9 9 3.6 13 11 14 8 12 7 14 15 15 13 #2855-ASR 71:3 filename:71305-Osberg
ATTITUDES TOWARD ECONOMIC INEQUALITY—–465 CEO/Worker Pay Ratio Pay CEO/Worker f es.html); BBC New firms’ share-price return. three-year /workhours.html). 47FA126-D194-42F1-BDD4- in Manufacturing e Pay RatioPay Compensation d Production WorkerProduction Actual CEO/Worker Estimated CEOWorker Estimated Skilled Estimated Workers in Manufacturing, 2001 (US$)Workers 1999 (US$) from ISSP, Actual CEO Compensation and Pay of Production Actual CEO Compensation and Pay “Do Earn” Estimates Average Subjective (ftp://ftp.bls.gov/pub/special.requests/ForeignLabor/supptab.txt) (Table 5); Annual Hours worked per person Annual Hours worked 5); (Table (ftp://ftp.bls.gov/pub/special.requests/ForeignLabor/supptab.txt) 649,137542,622711,403442,188481,651461,738 19,582 16,699 22,654 21,192 23,436 26,465 33 32 31 21 21 17 141,987 259,313 292,715 116,439 250,422 158,165 20,556 15,307 24,383 24,202 27,695 24,408 7.3 17.5 12.2 4.8 9.4 6.6 1,305,012 29,391 44 218,601 30,161 8.3 Source: Source: a The Actual and Estimated Earnings of Chief Executive OfficersWorkers and Production Actual and Estimated Earnings of Chief Executive The a b b c b a CEO compensation data for Australia, France & Sweden from BBC; The Galt Review (www.galtglobalreview.com/world/world_ceo_salari The Galt Review from BBC; & Sweden Australia, France CEO compensation data for Manufacturing Pay: In ISSP, earnings of “the chairman national corporation”. of a large In ISSP, The National Post Business Magazine’s annual CEO Scorecard: Average CEO compensation of Canada’s 150 biggest companies by their 150 biggest companies by CEO compensation of Canada’s Average annual CEO Scorecard: Business Magazine’s The National Post Average of Total CEO Compensation from The Galt Global Review (1999) and from BBC News (2001). (1999) and from BBC News The Galt Global Review CEO Compensation from Total of Average In ISSP, earnings of a “skilled worker in a factory”. earnings of a “skilled worker In ISSP, Australia France UK Sweden Canada (www.dol.gov/ILAB/media/reports/oiea/chartbook/chart19.htm); Annual Hours worked per person in Canada (www.pbs.org/now/politics Annual Hours worked (www.dol.gov/ILAB/media/reports/oiea/chartbook/chart19.htm); Table 5. Table CountryUnited States CEO Compensation in Manufacturing Germany a b 247D44F89560). c d e f (http://news.bbc.co.uk/1/hi/business/1456723.stm); (www.nationalpost.com/nationalpostbusiness/archives/20021105/story.html?id=C (http://news.bbc.co.uk/1/hi/business/1456723.stm); #2855-ASR 71:3 filename:71305-Osberg
466—–AMERICAN SOCIOLOGICAL REVIEW ally agree that, for example, a doctor does make Norwegian”). Still, in every country (includ- more money than a skilled worker, and should ing these two), there is an ongoing political make more money, there is a lot of disagreement debate about income distribution and poverty. about how much more. The differences between These debates are fairly direct evidence that individuals in their assessments of the desir- people do not all agree, within nations, about able degree of “leveling” can be estimated from inequality, and that the “median/average nation- the ISSP microdata. Because each individual al attitude” can be a somewhat misleading con- respondent reported his or her personal esti- cept, one that is particularly misleading if A mate of “should earn” (Yi*) and “do earn” (Yi ) attitudes toward inequality are highly polarized. A income for a number of occupations, these data If a regression of the form Yi* = b0 + b1 Yi can be used to estimate, for each respondent, a is estimated on each respondent’s data, those simple linear regression following the specifi- people who think the existing distribution of A cation of Equation 4 in Section 1 (i.e., we esti- earnings is fair will report Yi* = Yi (i.e., mate a regression of the form Yi* = b0 + b1 “should earn” = “do earn”), implying that for A Yi ). The ratio between “should earn” (Yi*) them b1 = 1. To the extent that respondents sup- A and “do earn” (Yi ) income for occupation is, port the status quo, there will thus tend to be an at the margin, captured by the b1 coefficient, accumulation at b1 = 1 of the b1 estimates of which is taken here as an individual’s prefer- these “status quo” respondents. However, peo- ences for the leveling of pay. For most people, ple who disagree with the fairness of current pay b1 is less than 1, because most respondents inequalities, and who think that income differ- think that some leveling is desirable. However, ences are “too large” will report “should earn” attitudes toward inequality are bounded (i.e., pay rates which imply b1 < 1. The more strong- when b1 = 1) by the attitude that no leveling at ly a respondent disagrees with the fairness of all is desirable, because some respondents report current income differences, the more leveling that “should earn” equals “do earn.” the respondent will prefer, and the lower will be If one thought there was less egalitarianism that respondent’s implied value of b1. But all (in the sense of a desire for a leveling of earn- these disagreements among people are hidden ings) in average American values than in other if only the average or median attitude is con- countries, then one might expect to observe a sidered. systematically higher average b1 coefficient in To assess how the distribution of disagree- the United States than elsewhere, but that is ment about leveling (i.e., b1) varies across coun- not the implication of comparing the mean and tries, a picture may be worth a thousand words. median “leveling” coefficient (b1) estimated Figure 3 presents a graph showing the distri- from the ISSP data. In the 1987 and 1999 data bution of preferences for leveling in the United for the countries listed in Table 4, the median States in 1987, 1992, and 1999. It portrays the and mean b1 coefficient in the United States was percentage of the population at each value of the above the mean for all country years except b1 coefficient, as drawn using kernel density 1992, when it was below the mean. The average methods, which offer a way to smooth the his- rank of the United States (over all three surveys) togram frequency of the population at each value of the b coefficient.22 Its value lies in pre- was 16th for the median b1 coefficient and 13th 1 senting a picture of attitudes that conveys much for the mean b1 coefficient, which are very close to the middle of our set of 33 national sur- more information than summary statistics. veys. In particular, Figure 3 indicates that a notable feature of American attitudes is their bimodal- ity. In all 3 years there is a clear spike at b = 1, THE DISTRIBUTION OF 1 as well as a substantial number of respondents DISAGREEMENTS ABOUT EQUALITY clustering around approximately b1 = 0.5 (i.e., Up to this point, national preferences and atti- the United States is a society with both a sig- tudes toward inequality have been summarized nificantly large group in favor of substantially in terms of a measure indicating the central ten- dency of the distribution of attitudes within each nation (e.g., in terms of the attitudes held 22 See Greene (2002) or http://genstat.co.uk/ by the “average American” or the “median doc/8doc/html/stats/KernelDensityEstimation.htm. #2855-ASR 71:3 filename:71305-Osberg
ATTITUDES TOWARD ECONOMIC INEQUALITY—–467
more equality than now exists and a large group Figure 4 puts the United States, the United that agrees with the status quo). Over time, Kingdom, France, Norway, and Canada on the there appears to have been a migration of atti- same graph. It is limited to a five-country com- tudes among Americans, with a somewhat parison because additional countries are hard to increased tendency to respond that “what is distinguish visually, but its basic story also can should be” (i.e., b1 = 1) in the distribution of be told with the data of other nations. If a pic- earnings. ture of “social cohesion” in attitudes toward
Figure 3. The Distribution of Preferences for Leveling in the United States, 1987–1999
Figure 4. United States, United Kingdom, Canada, Norway, and France 1999: Distribution of Should Earn/Do Earn Slope Coefficient (Betas): Both Sexes #2855-ASR 71:3 filename:71305-Osberg
468—–AMERICAN SOCIOLOGICAL REVIEW inequality could be painted, it probably would In Figure 5, the modal value of the U.S. max- resemble Figure 4 for Norway.23 Whereas the imum/mean ratio is lower and significantly United States kernel density estimates paint a more concentrated than in similar kernel den- picture of polarized attitudes, the Norwegian sity graphs for 1992 and 1987, but in whatever picture is one of broad consensus. As other data year one chooses to analyze, the modal value of also have indicated, Norwegians, on the average, ethical attitudes toward fair “top-end” inequal- are in favor of reducing still further their already ity is at a level that is vastly different from the relatively small income gaps. Figure 4 indicates actual pay ratios reported in Table 5. In contrast, that there is a very strong convergence in atti- American attitudes toward inequality at the bot- tudes around a value of about b1 = 0.66. tom end have become more diffuse over time. The United States (with strong polarization) However, in both figures, the relative unanim- and Norway (with consensus) are poles of a ity of Norwegian opinion comes through very continuum, with Canada (not entirely American strongly. The modal values of the maxi- in attitudes), the United Kingdom (not entirely mum/mean ratio and the mean/minimum ratio European in values), and France (not as unan- of “should earn” incomes both are small, and the imously egalitarian as Scandinavia) being inter- distribution is tightly compacted. As Figure 5 mediate cases. All five countries have a large indicates, Canadian and American attitudes number of people who are “levelers.” In all five toward inequality at the top end are very simi- countries there are some who believe “do earn” lar, and there is a concentration of opinion that equals “should earn” (i. e. b1 = 1). The “excep- the maximum/mean ratio should be a little less tional” aspect of the United States is that it con- than 2:1. United Kingdom and French respon- tains a much larger group clearly satisfied with dents have a somewhat greater dispersion in the status quo than many other countries. The their acceptance of top-end inequality. In North slight tendency for a bump at b1 = 1 observed America, however, there still is a noticeable in France, the United Kingdom, Canada, and social consensus on the maximum level of Norway is replaced by a clearly bimodal distri- income someone “should earn.” bution in the United States. This bimodality On the other hand, Figure 6 indicates that among Americans is apparent among both men there was no consensus in the United Kingdom, and women, with some convergence around an Canada, France, or the United States on relative acceptance of the status quo, indicating little or minimum earnings in 1999.25 By contrast, the no leveling desired (0.9 < b1 < 1), and with consensus on a social minimum in Norway is another convergence of attitudes around sub- quite striking, which presumably makes it sig- stantial desired leveling (b1 = approximately nificantly easier to mobilize political support for 0.5). antipoverty policies. Although the b1 coefficient may capture an At this point, it is natural to ask why coun- overall preference for leveling, within the range tries differ in the distribution of their attitudes. of “acceptable” incomes, it does not directly This article has argued that people may have dis- address the issue of the ethically permissible tinct attitudes toward an ethically acceptable range of earnings, or whether there is more con- social minimum, toward a “fair” income max- cern with capping excessive rewards at the top imum, and toward the leveling of differentials of the distribution or limiting deprivation at the within that range, and that each set of attitudes bottom. Figures 4a and 4b therefore present the may have a distinct explanation. In considering distribution of attitudes in the United States each specific issue (e.g., attitudes toward income and elsewhere toward the maximum/mean24 and mean/minimum “should earn” ratios. compare the distribution of U.S. attitudes toward the mean/minimum and maximum/mean ratio over time 23 Norwegians stand out for social consensus and in 1987, 1992, and 1999, see Figures S3 and S4 trust in the social capital literature (see Helliwell (Section 4, Online Supplement, Section 2, ASR Web 2003:25) and for egalitarian and pro-welfare state atti- site). tudes (Svallfors 1997:295). 25 Kelley and Evans (1993) concluded, using 1987 24 One gets the same result if attitudes toward ISSP data, that cross-national differences in attitudes wage differentials are examined between named were primarily about appropriate income differences occupations, such as CEO and skilled worker. To at the top end, but opinions clearly have changed. #2855-ASR 71:3 filename:71305-Osberg
ATTITUDES TOWARD ECONOMIC INEQUALITY—–469
Figure 5. United States, United Kingdom, Canada, Norway, and France 1999: Distribution of “Should Earn” Max/”Do Earn” Mean (Max/Mean) Ratio: Both Sexes
Figure 6. United States, United Kingdom, Canada, Norway, and France 1999: Distribution of “Do Earn” Mean/”Should Earn” Min (Mean/Min) Ratio: Both Sexes
leveling, as portrayed in Figure 4), many econ- the distribution of individuals’ attitudes differs omists and sociologists have been trained to across countries is that there are cross-country think “in a regression perspective” (i.e., to exam- differences in the influence of personal char- ine the partial correlation of “explanatory” vari- acteristics (e.g., gender, age, or education) on ables with the dependent variable of interest, attitudes. An alternative explanation for cross- controlling for the influence of other variables). country attitudinal differences may be differ- In this perspective, one possible reason why ences across countries in the frequency of #2855-ASR 71:3 filename:71305-Osberg
470—–AMERICAN SOCIOLOGICAL REVIEW attributes (e.g., fundamentalist Protestantism26). ferent” about American values, as compared Moreover, a household characteristic such as with European attitudes, and that less redistri- income may vary across countries in both dis- bution is, essentially, what Americans want. tribution and impact on attitudes, and the struc- We question the assertion that Americans in tural form of estimating equations may differ general are uniquely satisfied with economic across countries. As well, it is plausible to won- inequality. On the average, Americans do not der whether some individual characteristics stand out as particularly different from other (e.g., political party or subjective class identi- countries in general attitudes toward inequali- fication) should be used as controls or consid- ty or in the “should earn/do earn” comparisons, ered to be endogenously determined. but comparisons of medians or means hide an Explaining international differences in the important part of the story. The United States distribution of attitudes is therefore a fascinat- appears to be a country with much more polar- ing and complex area for further research. Initial ization of attitudes toward income leveling than analyses suggest, if a common estimating equa- is common elsewhere, and increasingly so over tion linking personal characteristics and atti- time. The bimodality of American attitudes tudes toward leveling is presumed, that the toward income leveling is particularly striking. influence of standard variables such as age, Kluegel et al. (1995a:206) have argued that education, income, or gender on leveling pref- it is common for individuals to have a “split con- erences (b1) differs across countries. Although sciousness” about economic inequality, because American women are significantly more egal- the same person often will report support for itarian than American men, gender differences egalitarian principles (such as distribution in the determinants of b1 in other countries tend according to need) and inegalitarian attitudes to be statistically insignificant. Similarly, in (such as the moral depravity of the poor). They most countries (including the United States), note that this “presents a fertile ground for fram- both age and education are statistically insignif- ing effects as political actors compete to make icant as determinants of b1, but not in all coun- salient either the social explanations of pover- tries. Although higher-income households in ty and wealth in support of redistribution or the several countries (e.g., Canada, France, Norway) individual explanations to motivate opposition prefer less leveling, family income is not sta- to the welfare state.” tistically significant as a determinant of b1 in This article has argued that there is a trend either the United States or the United Kingdom. over time for American attitudes toward inequal- However, much more needs to be done to ity at the top end of the income distribution to explain individuals’ attitudes toward inequali- become less tolerant of inequality, even as at the ty. bottom end they have become more accepting of inequality. The United States is not very dif- CONCLUSION ferent from other countries in its aversion to wide differences in income between the middle The United States has more income inequality class and the very affluent. When it comes to dif- than other developed countries, but government ferences between the middle and the bottom of does less about it (Osberg et al. 2004; Smeeding, the income distribution, however, both France 2005). In partial response to the “missing redis- and the Anglo-American countries have a sim- tribution” of American public policy, an influ- ilarly diffuse set of attitudes, which contrasts ential literature has argued, at least since Lipset with the strong concern for a social minimum and Bendix (1959), that there is something “dif- in Scandinavia (see Jäntti and Danziger, 2000), and which may help to explain international differences in public policy and average pover- 26 The 1998 ISSP asked respondents whether they ty rates. agreed that “the Bible is the actual word of God and Our principal findings can be summarized as it is to be taken literally, word for word.” In their responses, 30.1% of Americans agreed, as compared follows: to 9.3% of West Germans, 6.1% of Australians, 9.6% 1. The empirical trend toward widening actual earn- of Canadians, 4.7% of Swedes and 5.1 % of UK res- ings and income differentials at the top of the idents. A further 49.2% of Americans believed the United States income distribution is not reflect- Bible to be “the inspired word of God.” ed in subjective estimates, which underestimate #2855-ASR 71:3 filename:71305-Osberg
ATTITUDES TOWARD ECONOMIC INEQUALITY—–471
top-end inequality more than is common in other “Fairness and Redistribution.” American Economic countries. Review 95:960–80. 2. There appears to be less concern over time for a Alesina, Alberto and Eliana La Ferrara. 2001. “social minimum” in the United States than in “Preferences for Redistribution in the Land of comparable nations. Opportunities.” Working Paper No. 8267, National 3. Public attitudes against excessive wage differen- Bureau of Economic Research, Cambridge, MA. tials at the top end may have hardened in the Alesina, Alberto, Rafael Di Tella, and Robert United States (at least up to 1999). MacCulloch. 2001. “Inequality and Happiness: 4. There is a strong, and increasing, polarization of Are Europeans and Americans Different?” attitudes toward income leveling in the United Working Paper No. 8198, National Bureau of States. Economic Research, Cambridge, MA. Alwin, Duane F. 1987. “Distributive Justice and Although it is hard to specify exactly the Satisfaction with Material Well-Being.” American long-term implications for political economy of Sociological Review 52:83–95. a polarization of attitudes and a widening dis- Atkinson, Anthony B. 1970. “On the Measurement crepancy between public perceptions of actual of Inequality.” Journal of Economic Theory and “fair” top-end inequality, this does not 2:244–63. sound like a likely recipe for social or political Bénabou, Roland and Jean Tirole. Forthcoming. stability. Under majority rule in a two-party “Belief in a Just World and Redistributive Politics.” system, one mode of the distribution may con- Quarterly Journal of Economics. Björklund, Anders and Markus Jäntti. 2000. trol the levers of redistributive policy for a time, “Intergenerational Mobility of Socioeconomic but the underlying polarization of attitudes Status in Comparative Perspective.” Nordic Journal implies that there is a substantial gulf in desired of Political Economy 26:3–32. public policies, and that a relatively small migra- Breen, Richard and Jan O. Jonsson. 2005. “Inequality tion of voters may suffice to tip the balance of Opportunity in Comparative Perspective” between two very different conceptions of “fair” Annual Review of Sociology 31:223–43. inequality. Burtless, Gary and Christopher Jencks. 2003. “American Inequality and Its Consequences.” Pp. Lars Osberg currently is the McCulloch Professor 61–108 in Agenda for the Nation, edited by Henry of Economics, and University Research Professor, at Aaron, James M. Lindsay, and Pietro Nivola. Dalhousie University, Canada. His undergraduate Washington, DC: Brookings Institute. education was at Queen’s University, Kingston, and Corak, Miles, ed. 2004. Generational Income the London School of Economics and Political Mobility in North America and Europe. Science. After serving as a CUSO volunteer in Cambridge: UK Cambridge University Press. Tanzania, he earned his Ph.D. from Yale University. Entorf, Horst and Nicoleta Minoiu. 2004. “PISA His major fields of research interest have been the Results: What a Difference Immigration Law extent, causes, and implications of poverty and eco- Makes.” IZA Discussion Paper No. 1021, Institut nomic inequality. His most recent book is The zur Zukunst der Arbeit (IZA), Bonn, Germany. Economic Implications of Social Cohesion (edited, Erikson, Robert and John H. Goldthorpe. 1985. “Are 2003). He was President of the Canadian Economics American Rates of Social Mobility Exceptionally Association in 1999–2000 and currently is Review High? New Evidence on an Old Issue.” European Editor for the Review of Income and Wealth. Sociological Review 1:1–22. ———. 1992. The Constant Flux: A Study of Class Timothy M. Smeeding is the Maxwell Professor of Mobility in Industrial Societies. Oxford, England: Public Policy, and Professor of Economics and Public Clarendon Press. Administration, at the Maxwell School of Syracuse ———. 2002. “Intergenerational Inequality: A University. He is director of the Maxwell Center for Sociological Perspective.” Journal of Economic Policy Research and of the Luxembourg Income Perspectives 16:31–44. Study Project, which he co-founded in 1983. His Erikson, Robert, John H. Goldthorpe, Michelle research focuses on national and cross-national com- Jackson, Meir Yaish, and D. R. Cox. 2005. “On parisons of income and wealth inequality and pover- Class Differentials in Educational Attainment.” ty among vulnerable groups, including low-wage Proceedings of the National Academy of Sciences workers, children, the aged, and the disabled. 102:9730–733. Esping-Andersen, Gøsta. 1990. The Three Worlds of REFERENCES Welfare Capitalism. Cambridge, UK: Polity Press. Evans, Mariah D. R. and Jonathan Kelley. 2004. Alesina, Alberto and George-Marios Angeletos. 2005. “Subjective Social Location: Data from 21 #2855-ASR 71:3 filename:71305-Osberg
472—–AMERICAN SOCIOLOGICAL REVIEW
Nations.” International Journal of Public Opinion Kingdom, and the United States.” Memorandum Research 16:3–38. No 34/2005, Department of Economics, University Förster, Michael and Marco Mira d’Ercole. 2005. of Oslo, Oslo, Norway. “Income Distribution and Poverty in OECD Jäntti, Markus and Sheldon H. Danziger. 2000. Countries in the Second Half of the 1990s.” “Income Poverty in Advanced Countries.” Pp. Working Paper No. 22, OECD (Organisation for 309–78 in Handbook of Income Distribution, edit- Economic Co-operation and Development) Social, ed by A. B. Atikinson and F. Bourguignon. New Employment and Migration, Paris, France. York: Elsevier Science. Glaeser, Edward L. 2005. “Inequality.” Working Jasso, Guillermina. 1978. “On the Justice of Earnings: Paper No. 11511, National Bureau of Economic A New Specification of the Justice Evaluation Research, Cambridge, MA. Function.” The American Journal of Sociology Glaeser, Edward L., Giacomo A. M. Ponzetto, and 83:1398–419. Jesse M. Shapiro. 2004. “Strategic Extremism: ———. 1980. “A New Theory of Distributive Why Republicans and Democrats Divide on Justice.” American Sociological Review 45:3–32. Religious Values.” Working Paper No. 10835, Kelley, Jonathan and Mariah D. R. Evans. 1993. National Bureau of Economic Research, “The Legitimation of Inequality: Occupational Cambridge, MA. Earnings in Nine Nations.” The American Journal Gottschalk, Peter and Timothy M. Smeeding. 1997. of Sociology 99:75–125. “Cross-National Comparisons of Earnings and Kerr, V. Clark. 1983. The Future of Industrial Income Inequality.” Journal of Economic Societies: Convergence or Continuing Diversity? Literature 35:633–87. Cambridge, MA: Harvard University Press. ———. 2000. “Empirical Evidence on Income Kluegel, James R., Gyorgy Csepeli, Tamas Kolosi, Inequality in Industrialized Countries.” Pp. Antal Orkeny, and Maria Nemenyi. 1995a. 261–308 in Handbook of Income Distribution, “Accounting for the Rich and the Poor: Existential Vol. 1, edited by Anthony B. Atkinson and François Justice in Comparative Perspective.” Pp. 179–207 Bourguignon. Amsterdam, Holland: North- in Social Justice and Political Change, Public Holland. Opinion in Capitalist and Post-Communist States, Greene, William. 2002. Econometric Analysis. Fifth edited by James R. Kluegel, David S. Mason, and edition. Upper Saddle River, NJ: Prentice-Hall, Bernd Wegener. Berlin, Germany, and New York: Inc. De Gruyter. Grusky, David B. and Robert M. Hauser. 1984. Kluegel, James R., David S. Mason, and Bernd “Comparative Social Mobility Revisited: Models Wegener, eds. 1995b. Social Justice and Political of Convergence and Divergence in 16 Countries.” Change: Public Opinion in Capitalist and Post- American Sociological Review 49:19–38. Communist States. Berlin, Germany, and New Hauser, Robert M. and John R. Warren. 1997. York: De Gruyter. “Socioeconomic Indexes for Occupations: A Koschmann, J. Victor. 1997. “Review Essay: The Review, Update and Critique.” Sociological Nationalism of Cultural Uniqueness.” The Methodology 27:177–298. American Historical Review 102:758–68. Helliwell, John F. 2003. “Do Borders Matter for Lipset, S. Martin. 1996. American Exceptionalism: Social Capital? Economic Growth and Civic A Double-Edged Sword. New York: W.W. Norton. Culture in U.S. States and Canadian Provinces.” Pp. Lipset, S. Martin and Reinhard Bendix. 1959. Social 19–42 in The Economic Implications of Social Mobility in Industrial Society. Berkeley, CA: Cohesion, edited by Lars Osberg. Toronto, Canada: University of California Press. University of Toronto Press. Molm, Linda D., Nobuyuki Takahashi, and Gretchen Inkeles, Alex. 1998. One World Emerging? Peterson. 2003. “In the Eye of the Beholder: Convergence and Divergence in Industrial Procedural Justice in Social Exchange.” American Societies. Boulder, CO: Westview Press. Sociological Review 68:128–52. Iversen, Torben and David Soskice. 2005. “Electoral Nelles, H. V. “American Exceptionalism: A Double- Institutions, Parties, and the Politics of Class: Why Edged Sword (Review Essay)” (review of Some Democracies Redistribute More Than American Exceptionalism: A Double-Edged Others.” Working Paper 2005–50, Institute of Sword, by Seymour Martin Lipset). The American Governmental Studies, University of California, Historical Review, Vol. 102, No. 3. (Jun., 1997), Berkeley, CA. 749–757. Jäntti, Markus, Bernt Bratsberg, Knut Røed, Nolan, Mary (1997) “Against Exceptionalisms” Oddbjørn Raaum, Robin Naylor, Eva Österbacka, (review of American Exceptionalism: A Double- Anders Björklund, and Tor Eriksson. 2005. Edged Sword, by Seymour Martin Lipset). The “American Exceptionalism in a New Light: A American Historical Review, Vol. 102, No. 3. (Jun., Comparison of Intergenerational Earnings 1997), pp. 769–774. Mobility in the Nordic Countries, the United Osberg, Lars and Timothy M. Smeeding. 2006. #2855-ASR 71:3 filename:71305-Osberg
ATTITUDES TOWARD ECONOMIC INEQUALITY—–473
“Social Values for Equality and Preferences for edition, edited by Edwin Cannan. Reprint, London, State Intervention in the USA and Europe” UK: Methuen and Company. Forthcoming in Welfare State Reform in the United Solon, G. 2002. “Cross-Country Differences in States and the European Union—Policy Choices Intergenerational Earnings Mobility” Journal of and the Constitution of the New Welfare Society, Economic Perspectives 16:59 – 66. edited by Christian Toft and Joseph Cordes. Suhrcke, Marc. 2001. “Preferences for Inequality: Cambridge, UK: Cambridge University Press. East vs West.” Innocenti Working Paper No. 89, Osberg, Lars, Timothy M. Smeeding, and Jonthan UNICEF Innocenti Research Centre, Florence, Schwabish. 2004. “Income Distribution and Public Italy. Social Expenditure: Theories, Effects and Svallfors, Stefan. 1997. “Worlds of Welfare and Evidence.” Pages 823–862 in Social Inequality, Attitudes to Redistribution: A Comparison of Eight edited by Kathryn Neckerman. New York: Russell Western Nations.” European Sociological Review Sage Foundation. 13:283–304. Parsons, Talcott. 1960. Structure and Process in Szirmai, Adam. 1991. “Explaining Variation in Modern Societies. Glencoe, IL: The Free Press. Attitudes toward Income Inequality.” Pp. 229–68 Piketty, Thomas. 1995. “Social Mobility and in Social Justice in Human Relations, edited by Redistributive Politics.” Quarterly Journal of Herman Steensma and Riël Vermunt. New York and London, UK: Plenum Press. Economics 110:551–84. Wegener, Bernd and Susanne Steinmann. 1995. Rose, David and David J. Pevalin. 2003. A “Justice Psychophysics in the Real World: Researcher’s Guide to the National Statistics Comparing Income Justice and Income Socio-Economic Classification. London, UK: Sage Satisfaction in East and West Germany.” Pp. Publications, Ltd. 151–75 in Social Justice and Political Change, Schwabish, Jonathan, Lars Osberg, Timothy M. Public Opinion in Capitalist and Post-Communist Smeeding, Michael Eriksen, and Joseph Marchand. States, edited by James R. Kluegel, David S. Forthcoming. “Income Distribution and Social Mason, and Bernd Wegener. Berlin, Germany, and Expenditures: A Cross-National Perspective.” In New York: De Gruyter. The Distributional Effects of Government Spending Wilensky, Harold L. 2002. Rich Democracies: and Taxation, edited by Dimitri B. Papadimitriou. Political Economy, Public Policy, and New York: Palgrave Macmillan. Performance. Berkeley, CA: University of Shepelak, Norma J. and Duane F. Alwin. 1986. California Press. “Beliefs about Inequality and Perceptions of Woessmann, Ludger. 2004. “How Equal Are Distributive Justice.” American Sociological Educational Opportunities? Family Background Review 51:30–46. and Student Achievement in Europe and the United Smeeding, Timothy M. 2005. “Public Policy and States.” Discussion Paper No. 1284, Institut zur Economic Inequality: The United States in Zukunst der Arbeit (IZA) Bonn, Germany Comparative Perspective.” Social Science Younts, C. Wesley and Charles W. Mueller. 2001. Quarterly 86:955–83. “Justice Processes: Specifying the Mediating Role Smith, Adam. [1776] 1961. An Inequiry into the of Perceptions of Distributive Justice.” American Nature and Causes of The Wealth of Nations, Fifth Sociological Review 66:125–45.