Jump to ContentJump to Main Navigation
Jobs with Equality$

Lane Kenworthy

Print publication date: 2008

Print ISBN-13: 9780199550593

Published to Oxford Scholarship Online: September 2008

DOI: 10.1093/acprof:oso/9780199550593.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2017. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber: null; date: 25 February 2017

Why Should We Care About Inequality?

Why Should We Care About Inequality?

(p.13) 2 Why Should We Care About Inequality?
Jobs with Equality

Lane Kenworthy (Contributor Webpage)

Oxford University Press

Abstract and Keywords

This chapter addresses the following questions: why should citizens and policy makers should want low inequality? Why should we object to inequality? Should we focus on equality of opportunity or equality of outcomes? Why not focus on poverty instead? Why emphasize incomes rather than wealth or material wellbeing? It argues that there is good reason to desire low inequality in and of itself — not instead of, but in addition to, high absolute incomes for those at the low end of income distribution.

Keywords:   inequality, poverty, income, wealth

For many, the aim of limiting inequality needs some justification. There are several issues here. Why should we object to inequality? Should we focus on equality of opportunity or equality of outcomes? Why not focus on poverty instead? Why emphasize incomes rather than wealth or material well‐being? Do people in fact care about inequality? I address each of these in turn.

Let me make clear at the outset that although I will frequently refer to “equality” in this book, I use this as a synonym for “low inequality.” Few if any egalitarians favor perfect equality of outcomes. Complete equality would substantially reduce work incentives, thereby violating the principle of reciprocity (all who are able to contribute do so; see Bowles and Gintis 1998; Galston 2001; Rawls 2001; White 2004a) and undermining the other aim I advocate: high employment. Imagine a society in which the social product is divided into an equal consumption allowance for each citizen. If the population were 10 million, the effective marginal tax rate on additional income would be 99.99999%, and an average earner who abandoned paid employment would reduce the value of her own consumption share by a mere 0.00001%. Plainly, the incentive to opt out of employment would be quite strong.


Objections to inequality tend to focus on considerations of fairness and on ways in which inequality may adversely affect other desirable outcomes.


One reason to favor societal institutions and policies that reduce inequality has to do with the importance of luck in determining earnings and incomes (Rawls 1971; Jencks et al. 1972; Dworkin 1985; Roemer 1998; Miller 2003: ch. 4; Miller 2005). Much of what determines people's earnings and income—intelligence, creativity, physical and social skills, motivation, persistence, confidence, (p.14) connections, inherited wealth, discrimination—is a product of genetics, parents' assets and traits, and the quality of one's childhood neighborhood and schools. With few exceptions, these things are not chosen; they are a matter of luck. A nontrivial portion of earnings inequality and income inequality is therefore undeserved.

Consequences of Inequality

Other arguments for low inequality focus on its consequences. A higher level of income inequality may result in slower economic growth, higher crime rates, poorer health, less educational attainment, heavier debt loads for the middle class, disproportionate political power wielded by the wealthy, and greater political polarization, among other societal ills. The extent to which it does so is an empirical question. It is a large question, and I will not attempt to answer it with empirical evidence here. Instead I will summarize recent arguments and findings.

Education. Higher levels of inequality may widen disparities in educational attainment. The main effect is likely to be on differences in college attendance. In the United States, even with substantial funds available for financial aid, many students from lower‐income households are forced to pay a relatively large amount to attend college. A study by the U.S. Census Bureau (2002: table 6a) found that among students from families with incomes below $25,000, the average yearly cost of attending college as of the mid‐1990s was $6,000. The average amount covered by financial aid was $3,000, leaving the remaining $3,000 to be paid by the student or her∕his parents. Thomas Kane (2004) reports that in 1980, 55% of children from families in the top income quartile attended a four‐year college, compared to 29% of those from families in the bottom quartile. By 1992, as income inequality increased, the difference had widened to 66% versus 28%. These quartile differences are smaller but still sizable when parents' education and student test scores and high school rank are controlled for (Kane 2004).

The effect of income inequality on inequality in college attendance is shaped by public policy. In the Nordic countries, access to college has been substantially enlarged in the past several decades by policies that fund such participation as a universal right of citizenship (OECD 2004e). Not surprisingly, the odds of a child with low‐educated (and thus likely low‐income) parents attending university are three to six times greater in Sweden, Denmark, and Norway than in the United States (Esping‐Andersen 2007).

Health. Higher income inequality contributes to greater disparity of health outcomes (Mullahy, Robert, and Wolfe 2004). As with education, the size of (p.15) the effect will be heavily influenced by government policy. But even in a system with universal health coverage, co‐payments coupled with lesser knowledge will tend to reduce use of health services among the poor and higher income will enable the rich to purchase better treatment.

A number of studies have linked income inequality with lower average levels of health (Wilkinson 1996; Lynch and Kaplan 1997; Kawachi, Wilkinson, and Kennedy 1999). The mechanism commonly posited to account for this link is relative deprivation: having an income below one's reference group may increase stress and thereby worsen health, regardless of the person's absolute income. However, a variety of empirical studies have cast doubt on this link both across countries and within countries over time (Burtless and Jencks 2003; Deaton 2003; Beckfield 2004; Mullahy, Robert, and Wolfe 2004).

Crime. Inequality might increase crime via material incentives or via frustration. To the extent inequality is associated with limited job and income opportunities for young persons (especially males), it may encourage greater pursuit of illicit income‐generating activities (Freeman 1996). High levels of inequality might also foster crime by breeding frustration. Perceived relative deprivation and blocked opportunity may contribute to greater criminal activity (Merton 1968).

Evidence on a link between inequality and crime is mixed (Burtless and Jencks 2003; Western, Kleykamp, and Rosenfeld 2004). At the individual level, the propensity to commit crime is higher among those with lower incomes. Cross‐sectional studies across US states or localities have sometimes yielded findings of a positive association. The United States has higher income inequality than other affluent countries and a much higher rate of violent crime, but other types of crime that seem likely to be encouraged by inequality, such as theft, do not correlate with inequality across countries. And over time, the association between inequality and crime in the United States is not particularly strong.

Consumption and debt. Robert Frank (1999, 2005) argues persuasively that housing is a “positional” good. That is, people's happiness with their home depends more heavily on relative comparison with other nearby homes than is true for many other goods (such as toothpaste or vacation time). Consider a hypothetical scenario, Frank says, in which you must choose “between world A, in which you will live in a 4,000‐square‐foot house and others will live in 6,000‐square‐foot houses; and world B, in which you will live in a 3,000‐square‐foot house and others will live in 2,000‐square‐foot houses. Once you choose, your position on the local housing scale will persist. If only absolute consumption mattered, A would clearly be better. Yet most people say they would pick B, where their absolute house size is smaller but their relative house size is larger” (Frank 2005: 137). Frank suggests that rising income at (p.16) the high end of the distribution in the United States (see below) has allowed the well‐to‐do to purchase increasingly large and elaborately equipped homes. Because housing satisfaction depends on relative comparison, middle‐class homeowners have felt compelled to follow suit, leading to dramatic increases in home prices and housing expenditures. The result is reduced spending in other consumption areas, greater debt, and increased middle‐class bankruptcy (Warren and Tyagi 2003; Mishel, Bernstein, and Allegretto 2007).

Economic growth. Although for a long time the conventional view held that income inequality is good for economic growth, in recent decades a contrary argument has emerged, suggesting that inequality may well impede growth. There are various mechanisms through which this effect might obtain (Lazear 1989; Akerlof and Yellen 1990; Levine 1991; Barr 1992; Alesina and Rodrik 1994; Persson and Tabellini 1994; Birdsall, Ross, and Sabot 1995; Kenworthy 1995: ch. 3; Bénabou 1996; Agell 1999; Gomez and Meltz 2001). One is reduced educational attainment, as discussed above. A second is that high levels of inequality may be viewed by those at the middle and bottom of the income distribution as excessively unfair, thereby reducing worker motivation and workplace cooperation. Third, the financial constraints and frustration generated by high levels of inequality may reduce trust, cooperation, civic engagement, and other growth‐enhancing forms of social capital. Fourth, social insurance programs that reduce inequality may encourage human capital investment and employment by providing—more efficiently than private insurance markets—a cushion against risk of job loss, illness, and on‐the‐job injury. Fifth, income polarization may foster extralegal demands for economic and∕or political reform. Rebellions, revolutions, and other forms of violent collective action diminish political stability, which may adversely affect growth. Sixth, higher levels of market inequality may generate popular demand for increased government spending, particularly on transfers, which might reduce growth. These latter two have limited applicability to affluent nations. In such countries, political stability has not been disrupted by violent collective action in recent decades, and higher levels of market inequality are not associated with higher levels of government transfers (see Chapters 3 and 7).

The empirical evidence on income inequality's effect on economic growth is mixed. A number of studies have found a negative cross‐country and over‐time association between inequality and growth in less‐developed nations (Birdsall, Ross, and Sabot 1995; Clarke 1995; Perotti 1996). This is not surprising, as effects of inequality on education and political stability are likely to be sizeable in this type of context. In affluent countries there is much less consensus among scholars (Barro 2000; Forbes 2000; Burtless and Jencks 2003; Voitchovsky 2003). In my own recent analysis of rich countries, of the US (p.17) states, and of over‐time trends in the United States, I found no indication that income inequality either increases or reduces economic growth (Kenworthy 2004a: ch. 4).

Economic mobility. The notion that “a mobile society is better than an equal one” (The Economist 2007) probably strikes many as sensible. But what if high levels of inequality impede mobility? There are several mechanisms through which greater income inequality might reduce intergenerational mobility (Corcoran 1995, 2001; Meyers et al. 2004; Esping‐Andersen 2007). The rich can pass their income on to their children directly via gifts and inheritance. Low parental income may inhibit opportunities for children via risk‐averseness, greater stress, lesser access to quality childcare and preschool, inferior schools, limited health care, and more dangerous neighborhoods. The United States, which has the highest level of income inequality among rich countries, also has less intergenerational income mobility than most others (Corak 2004; Jäntti et al. 2006). The degree of mobility in the United States appears to have increased in the 1960s, when income inequality was declining, but then remained flat since the 1970s as inequality has risen (Harding et al. 2005).

Civic engagement. Involvement in community affairs and organizations is widely thought to be beneficial both for its own sake and for its effects on other desirable societal outcomes (Putnam 1993, 2000). In the United States at least, such engagement tends to be strongly stratified by income (Putnam 2000; Brady et al. 2003; Skocpol 2004). There is disagreement, however, about whether rising income inequality since the 1970s has widened the dispersion in civic engagement (Brady et al. 2000; Putnam 2000; Wuthnow 2002; Costa and Kahn 2003a, 2003b).

Political participation. Political participation includes activities such as voting, joining or contributing money or effort to a political organization or a campaign, and participating in community activities or protests. Differences in resources due to unequal incomes are most likely to produce inequality in monetary contribution. There is considerable evidence that they are associated with inequality in other kinds of political participation as well (Verba, Schlozman, and Brady 2004). On the other hand, it is not clear whether the degree of inequality in participation in the United States has risen in concert with income inequality over the past several decades (Brady 2004; Freeman 2004; Verba, Schlozman, and Brady 2004: 644–5).

Political influence. To the extent that private money affects elections and political decision‐making, there is good reason to suspect greater income inequality will be associated with greater inequality in political influence. A task force commissioned by the American Political Science Association concluded recently that “Disparities of wealth, income, and access to opportunity (p.18) are growing more sharply than in other nations. …Progress toward realizing American ideals of democracy may have stalled, and in some arenas reversed. …The privileged participate more than others and are increasingly well organized to press their demands on government. Public officials, in turn, are more responsive to the privileged than to average citizens and the least affluent. Citizens with low or moderate incomes speak with a whisper that is lost on the ears of inattentive government, while the advantaged roar with a clarity and consistency that policy makers readily heed. …Our review of research on inequality and political participation as well as other components of American political life demonstrates an extraordinary association between economic and political inequality” (Jacobs et al. 2004: 651, 655; see also Phillips 2002; Verba, Schlozman, and Brady 2004; Krugman 2007).

Political polarization. Less obvious than its link with political inequality is the fact that greater income inequality may contribute to political polarization. In analyses of National Election Studies (NES) data, Nolan McCarty, Keith Poole, and Howard Rosenthal (2006) find that party preference and voting in the United States have become more closely correlated with income over time. The positions of legislators from the two parties have moved further apart since the 1970s, and over the course of the twentieth century the degree of polarization tracks closely with the level of income inequality.

The study of income inequality's consequences is in some respects in its infancy. Many of the hypotheses I have discussed here need further confirmation. Perhaps the most important task for researchers is to better gauge the magnitude of these effects. Are the adverse impacts of greater inequality sizable? Or are they relatively minor compared to other determinants? We have little in the way of answers to this question.


For most Americans, and very likely for many people in other affluent countries, inequality of opportunity is much more objectionable than inequality of outcomes. I share this view, but pursuing equal opportunity is not enough, for two reasons. First, because parents' income affects children's capabilities, true equality of opportunity would require something close to equalization of assets and incomes, at least among households with children (Jencks et al. 1972: 4; Duncan and Brooks‐Gunn 1997; Bowles, Gintis, and Osborne 2004). Second, even if asset and income equalization were accomplished, truly equal opportunities would remain unrealizable. Individuals' opportunities (p.19) are influenced by genetic endowments, parents and other adults, peers, and a variety of chance occurrences throughout childhood and adolescence. No liberal society—that is, one in which families and other institutions retain a sizable degree of autonomy—can ensure that its members reach adulthood with equal capacities for success. Institutions and policies that compensate for unequal opportunities by reducing inequality of outcomes are thus justifiable even for those who prioritize equality of opportunity.


Inequality should be distinguished from poverty. Inequality refers to the degree of dispersion in the distribution. Poverty refers to the incomes of those at the bottom of the distribution.

There are two broad approaches to thinking about poverty. The dominant one in Europe, and among researchers who study cross‐country differences in poverty, conceptualizes poverty as relative. The poverty line, below which a household or person is defined as being poor, is set relative to the median income in the country. The most common choice is to set the poverty line at 40%, 50%, or 60% of the median. The poverty line therefore differs across countries and shifts over time, depending on the median. The principal rationale for thinking about poverty as relative is that people tend to assess their living standard by comparing it to that of others in their society. As Robert Goodin and his colleagues (1999: 28) put it: “People feel themselves to be poor, and think others to be poor, in ways that matter both sociologically and ultimately morally, if they have substantially less than what is commonplace among others in their society” (see also Sen 1983; Brady 2003; Iceland 2003; Rainwater and Smeeding 2003). Poverty, in this view, is relative deprivation.

If we conceptualize poverty in relative terms, there is little need to distinguish it from inequality. A relative measure of poverty is a measure of inequality. It differs from other measures in that it focuses only on the bottom half of the income distribution, but relative poverty in essence captures the degree of inequality between the median and those at the bottom of the distribution. For the twelve countries on which I focus in this book, the correlation between posttax‐posttransfer household income inequality measured using the P50∕P10 ratio and relative poverty with the poverty line set at 50% of the median is .98.

But what if we conceptualize poverty in an absolute sense—that is, focusing on the absolute income levels of those at the low end of the distribution? Societies with more inequality may also have more absolute poverty. But that is not (p.20) inherently the case. A society could be highly unequal and yet have moderately high absolute incomes for those at the low end of the distribution. Conversely, an egalitarian society could have low absolute incomes at the bottom of the distribution. If these scenarios prevail, the distinction between inequality and poverty becomes potentially important for normative and policy debate.

Martin Feldstein (1999: 33) puts the issue in the following way:

According to official statistics, the distribution of income [in the United States] has become increasingly unequal during the past two decades. A common reaction in the popular press, in political debate, and in academic discussions is to regard the increased inequality as a problem that demands new redistributive policies. I disagree. I believe that inequality as such is not a problem and that it would be wrong to design policies to reduce it. What policy should address is not inequality but poverty.

The difference is not just semantics. It is about how we should think about the rise in incomes at the upper end of the income distribution. Imagine the following: Later today, a small magic bird appears and gives each Public Interest [the journal in which Feldstein's article was published] subscriber $1,000. We would all think that this is a good thing. And yet, since Public Interest subscribers undoubtedly have above‐average incomes, that would also increase inequality in the nation. I think it would be wrong to consider those $1,000 windfalls morally suspect.

The argument is that as long as the poor are not made worse off, we should not object to a rise in inequality that is produced by growing incomes for those at the top of the distribution. The only reason for doing so, according to Feldstein (and others), is envy or spite.

One response is to focus on the consequences of inequality. If the incomes of the rich pull too far away from the rest of society, growing frustration may lead to rising crime, withdrawal from civic engagement, loss of social cohesion, and increasing political influence by the rich. As noted earlier, these are empirical questions, and not ones I will address here.

Let's return to normative considerations. The Feldstein‐type challenge is consistent with a variety of other views about distributive justice, including that of John Rawls (1971). Rawls argued that the most reasonable way to decide upon a fair distributive principle is to imagine that you must make this decision knowing you will be born into the world but not knowing anything about what your assets and characteristics—intelligence, personality traits, parents, neighborhood, gender, skin color, etc.—will be. Rawls referred to this hypothetical scenario as the “original position.” He suggested that in such a situation a rational person would choose a set of basic liberties coupled with a distributive principle requiring that any increase in inequality raise the income of those at the bottom. In Feldstein's example, according to the Rawlsian criterion the $1,000 windfall given to the well‐to‐do would only be (p.21) justifiable if it was accompanied by some increase for those at the low end. In practical terms, one way to accomplish this would be to tax part of the $1,000 gain and redistribute it to the poor.

Rawls's distributive principle is a “maximin” one: whatever distribution maximizes the income of the poorest (and provides basic liberties) is to be preferred. As I suggested earlier, a system of perfectly equal (posttax‐posttransfer) incomes would substantially reduce work incentives and thus reduce the average income (the “size of the pie”). It is therefore unlikely to maximize the income of the least well‐off. It is an empirical question how unequal incomes would need to be in order to most effectively balance incentives for income‐maximization with redistribution, but the point is that inequality is not objectionable per se in the Rawlsian view.

Some experimental evidence suggests that Rawls may have been wrong in his assumption about what distributive principle people in the original position would choose. In experiments in which five or so participants are placed in a situation approximating Rawls's original position, most do not tend to choose his distributive principle. Instead, they tend to choose a principle in which the average income is maximized with a floor under the incomes of those at the bottom (Frohlich, Oppenheimer, and Eavey 1987). In this view, as long as the poor have “adequate” incomes, an increase in the incomes of the rich need not benefit the poor to be considered just. The results of such experiments are consistent with Rawls, however, in suggesting that (absolute) poverty should be of greater concern than inequality.

Rawls reaches his conclusion about the preferability of a maximin principle by a priori ruling out envy as a legitimate motivation in thinking about distributive justice (1971: 143–4, 530–41). This resonates with the intuition many of us have when considering distributive issues. But does it make sense? Branko Milanovic (2003: 6) argues that what we think of as “envy” in this context actually is simply an expression of the view that the higher incomes of the best‐off are undeserved: “What some people call envy is …not (the bad) envy but (the good) sense of justice.” It is indeed envy, but it need not have the negative connotation we usually attach to it. If some people have higher incomes than others and that difference is undeserved, because it is due largely or entirely to luck, then we need not necessarily approve of it morally.

I am not arguing that low inequality is more desirable than low (absolute) poverty, but simply that low inequality does matter in and of itself. In my view, both should be taken into account in assessing distributive outcomes.

Consider the income distributions in Table 2.1. The table shows income levels and inequality in four hypothetical societies: A, B, C, and D. Each has three households: POOR, MIDDLE, and RICH. These societies can be thought of as representing different countries at the same point in time or the same (p.22)

Table 2.1. Income distributions in four hypothetical societies.

Household income

Average income

Gini coefficient

Poor household

Middle household

Rich household

Society A






Society B






Society C






Society D






country at different points in time. Household incomes are expressed in a common currency (dollars or euros, perhaps) and are adjusted for cost‐of‐living differences (over time or across countries). The incomes are posttax‐posttransfer. The households are of equal size. Assume that incomes correlate perfectly with living standards, that is, there is no difference between the societies in household wealth or in publicly provided services. Assume also that these are recurring incomes; there is no upward or downward intragenerational mobility. The last column in the table shows a commonly used measure of inequality: the Gini coefficient. It ranges from zero to one, with larger numbers indicating more inequality.

First compare societies A and B. Which of these two should we prefer? The Rawlsian maximin criterion prefers A, because the income of POOR is higher in A than in B. The “maximize average income subject to a floor” criterion could prefer either A or B, depending on where the floor is set. If the floor is $18,000, A is preferred because in B the income of POOR is below the floor. If the floor is $15,000 or lower, society B is preferred, because POOR'S income meets the floor and average income is higher than in A. Inequality is greater in B than in A, but perhaps the difference is not large enough for concern about inequality to trump concern about the income level of the poor or about the average income level.

Now compare A and B to C. Both Rawls and “maximize average income subject to a floor” prefer C to either A or B: Rawls because the income of POOR is highest in C, “maximize subject to a floor” because C has the highest average income of the three and if the income of POOR meets the floor in A or B it necessarily does so in C. Inequality is considerably greater in C than in A. But is the difference large enough that a “sensible egalitarian” would object?

Finally, consider society D. Here POOR has a higher income than in the other three societies, MIDDLE has the same, while RICH has an immensely higher income. Rawls and “maximize subject to a floor” would both prefer D to any of the other societies. But should we agree? There is much more inequality (p.23)

Why Should We Care About Inequality?

Figure 2.1. Changes in posttax‐posttransfer household incomes at the 10th, 50th, and 90th percentiles, 1970–2000.

Note: Figures are in national currencies and thus permit only cross‐country comparison of income changes, not of levels for data definitions and sources, see the appendix.

in D than in the other three societies. Suppose D's RICH household has an income of 10,000,000 because Feldstein's magic bird brings the money to it, or for some other reason that has to do largely with luck. Should we prefer society D on normative grounds? Though I am not certain how many readers will agree, I do not think we should.

The point of this exercise is to suggest that there is reason to be concerned about both poverty and inequality, rather than only about poverty.

Figure 2.1 offers another opportunity to assess the relative importance of poverty and inequality, this time using real data for four countries. The figure shows trends in inflation‐adjusted posttax‐posttransfer incomes per equivalent person (I explain what “per equivalent person” means in Chapter 3) for working‐age households at the 10th, 50th (median), and 90th percentiles in (p.24) Sweden, Germany, the United Kingdom, and the United States. It includes all years for which Luxembourg Income Study data are available since the early 1970s. The incomes are shown in national currencies; the point is to compare income change over time, rather than income levels.

The incomes of the poor (proxied here by the 10th‐percentile household) remained constant in each of the four countries during this period of two and a half decades, the only exception being Sweden, where there was a bit of an increase. Incomes at the top (proxied by the 90th percentile) increased in three of the four countries. They grew most rapidly in the United Kingdom, but as a result the UK had the sharpest increase in inequality. The United States was next in terms of the degree of increase in 90th‐percentile income and in income inequality. Sweden followed. In Germany there was effectively no change in income at the bottom, middle, or top, and hence no rise in inequality.

Which of these countries had the most desirable performance, in terms of over‐time developments? According to Rawls's maximin criterion, Sweden did best, since income at the bottom increased. According to “maximize average income subject to a floor,” the United Kingdom performed best (if we assume that its minimum exceeded the floor). For someone who treats inequality as having the same weight in normative considerations as poverty, Germany might be judged to have had the best performance, since the trend for its poor was similar to that in other countries and it had the smallest increase in inequality.

For most readers, I suspect, there will be no simple or easy answer to the question. My inclination would not be to judge Germany's performance as best even though it had the least increase in inequality. I would probably conclude that Sweden's performance was the best among the four countries: incomes at the bottom increased somewhat, average incomes rose, and income inequality increased by less than in the United Kingdom or the United States.

Let me add one more illustration, again using real data. This time we will consider developments in the United States over a longer period—since the mid‐1940s. Figure 2.2 shows trends in inflation‐adjusted family incomes at the 20th, 40th, 60th, 80th, and 95th percentiles during this period (data for households, which is a more sensible unit to examine, are not available prior to the late 1960s). The data suggest that the principal difference between the “golden age” period from the mid‐1940s to the mid‐1970s and the ensuing period of rising inequality is the stagnation of incomes in the bottom half of the distribution (Kenworthy 2004b). Those at the high end have pulled away from the rest. But that is not because their income growth accelerated. Instead, it is because income growth slowed considerably for families in the bottom half.


Why Should We Care About Inequality?

Figure 2.2. Family income trends in the United States at various percentiles, 1947ff.

Note: Pretax‐posttransfer income.

In my view, we should worry about this development mainly because despite growth of average income, poverty has not decreased. This is indicated by the lack of increase in real incomes at the 20th percentile, which we can treat as a proxy for poverty. It also is reflected in the lack of decline in the official government poverty rate, which was 11% in 1973, 11% in 2000, and 12% in 2006. This judgment of course presumes a counterfactual in which average income could have increased just as rapidly with a larger share going to those at the bottom.

There is more to the story. The Current Population Survey (CPS), which is the source for the income data in Figure 2.2, does not have useful information on the incomes of the richest Americans, as they are top‐coded at $1 million ($300,000 prior to 1993). The best available time‐series data on incomes for those at the very top of the distribution are from Thomas Piketty and Emmanuel Saez (2007b), who analyzed tax records back to 1913. Figure 2.3 shows the share of total household income going to the top 1% of taxpayers according to the Piketty and Saez calculations. The chart suggests a massive rise since the late 1970s, with the top 1%'s share of household income more than doubling.

Though available only since 1979, data from the Congressional Budget Office on the income levels of the top 1% tell a similar story. These are shown in Figure 2.4. They too suggest a substantial rise in incomes for the very rich since the 1970s.


Why Should We Care About Inequality?

Figure 2.3. Income share of the top 1% in the United States, 1913ff.

Note: Taxpayer units. Pretax income. Includes capital gains, though the patterns are very similar if capital gains are excluded.

Source: Piketty and Saez (2007b: table 5A3).

Why Should We Care About Inequality?

Figure 2.4. Income trends for the bottom 20%, middle 20%, top 20%, and top 1% of households in the United States, 1979ff.

Note: Posttax‐posttransfer income.

Source: Unpublished calculations by Jared Bernstein from Congressional Budget Office data. See Mishel, Bernstein, and Allegretto (2007: 64–5).

(p.27) In the view of Feldstein (and others), income developments in the United States in recent decades should not be cause for concern. Real income for households at the tenth and twentieth percentiles has been stagnant, while income for the top 1% has soared. From this perspective, the only reason to object to these developments is envy. I suspect many observers would disagree, both because some (perhaps much) of the huge increase for those at the very top is a product of luck of one kind or another—genetic inheritance of cognitive ability or work ethic, good parenting, attending the right schools, being in the right place or industry at the right time, and so on—and because it seems likely that average incomes could have increased no less rapidly had they done so with more going to those at the bottom.

What do I conclude from these reflections on poverty and inequality? I very much agree with the emphasis Rawls, Feldstein, and others place on the absolute income levels for those at the low end of the income distribution. Yet I do not find compelling the argument that concern about inequality should be dismissed. Because some or much of the higher incomes of those in the top portion of the income distribution may be a product of luck, because in many instances it is plausible to believe that a less unequal distribution would have yielded no loss in average income, because many people care as much or more about their relative position as about their absolute well‐being, and because inequality may have undesirable effects on other aspects of society, I conclude that there is good reason to desire low inequality in and of itself—not instead of, but in addition to, high absolute incomes for those at the bottom.


Arguably, income is a less useful indicator of material well‐being than either wealth or access to basic material needs (Townsend 1979; Mayer and Jencks 1989; Wolff 2002). Unfortunately, comparative over‐time data on wealth and material hardship are very limited. For better or worse, I therefore focus on income.


Income inequality is measured with data from surveys that ask people about their income in a single year. It therefore captures the degree of dispersion at a particular point in time. What if the degree of income mobility—of upward and downward movement of households within the income (p.28) distribution—differs across countries? Suppose two countries have the same amount of income inequality using the standard single‐year measure, but country A has much more (intragenerational) mobility than country B. If households' income were measured over a period of multiple years, we would conclude that there is less inequality in A than in B.

To examine this possibility, we need data for the same households over multiple years. Although such data are scarce, the best available suggest that in fact there is relatively little cross‐country difference in income mobility. Thus, measures of income inequality based on income in a single year are fairly accurate in gauging the true degree of inequality (Gangl 2005; Gangl, Palme, and Kenworthy 2008).


Yes, they do. There is a great deal of evidence from experiments that suggests inequality matters to people, though it seldom is the dominant criterion (Konow 2003). For my purposes it is perhaps sufficient to point to some survey evidence. The best available comparative data come from the International Social Survey Programme (ISSP; for the United States see also Maxwell 2007; Pew Research Center for the People and the Press 2007). In 1999 the ISSP conducted a module specifically on social inequality.

Why Should We Care About Inequality?

Figure 2.5. Income distributions in five hypothetical societies.

Source: ISSP (1999, questionnaire, variable: v 93).

(p.29) One question asked whether respondents thought the difference in incomes between rich and poor in their country was currently “too large.” The share saying they agreed or strongly agreed that the difference in their country was too large was 71% in Australia, 71% in Canada, 88% in France, 82% in Germany, 73% in Norway, 71% in Sweden, 82% in the United Kingdom, and 66% in the United States (ISSP 1999).

Another question offered respondents pictorial illustrations of various income distributions and asked “What do you think the distribution in your country ought to be like—which do you prefer? ” The five choices were depicted as shown in Figure 2.5.

In each country some respondents did not answer and around 5% said they could not choose between the five types. Among those who did make a choice, a relatively small share said they preferred types A, B, or C: fewer than 2% chose A, fewer than 8% chose B, and fewer than 15% chose type C. The bulk of respondents selected either type D or type E.

Interestingly, in each of the countries included in the survey, a larger share of respondents chose type D than type E. D and E are identical in their population shares at the very bottom (the bottom two rows). The difference between the two is that D has a larger share in the middle, whereas E has a larger share at the top. E corresponds most closely to the “maximize average income subject to a floor” criterion, whereas the characteristic of D that differentiates it from the others is its lesser degree of inequality. Yet more respondents in the ISSP survey preferred D than preferred E: 55% versus 24% in Australia, 50% versus 35% in Canada, 46% versus 22% in France, 55% versus 20% in Germany, 57% versus 32% in Norway, 49% versus 34% in Sweden, and 47% versus 26% in the United States. This suggests strongly that people care about inequality in and of itself.


My aim in this chapter has been to suggest some reasons why citizens and policy makers in affluent countries should be interested in limiting income inequality. In the next chapter I consider the nature of the challenge they face.