Misperceiving Inequality

August 12, 2015 • Research Briefs in Economic Policy No. 32
By Vladimir Gimpelson and Daniel Treisman

Various theories associate countries’ levels of economic inequality with important political outcomes. These include the extent of redistribution (in democracies); the incidence of revolution and other political violence (in nondemocracies); and the emergence and stability of democracy and dictatorship.

One argument contends that democracies with greater market‐​generated inequality will redistribute more. In this view, the larger the gap between the median and mean incomes, the greater the fiscal transfer from rich to poor that majority‐​rule voting will produce. A second literature blames inequality for the outbreak of revolutions, coups, civil wars, and other forms of political violence. The greater the incomes and land‐​holdings of the rich, the more the poor stand to gain by forcibly expropriating them.

A third perspective sees inequality as driving the evolution of political regimes. Since elites in unequal autocracies anticipate high levels of redistribution under democracy, they fight hard to prevent it. This implies a negative relationship between inequality and democratization, unless the elites are somehow protected from expropriation: for instance, by the mobility of their assets. Further, the poor have weak incentives to seek democracy if inequality is very low (taxing the rich will benefit them little). Thus the odds of democratization should trace an inverted U: it is unlikely at either high or low inequality but more likely at intermediate levels.

Despite their intuitive appeal, all three theories have proved hard to substantiate empirically. According to recent reviews of the literature, inequality does not seem to matter for the politics of redistribution. Nor does the evidence suggest that economic inequality prompts political violence or change of regime type.

Scholars have suggested a number of reasons why the relations implied by theory might not hold in practice. A large gap between rich and poor might not lead to redistribution because attitudes toward inequality are mediated by beliefs about the fairness of the distribution,about social mobility, or about societal norms. Religious organizations might build coalitions spanning poor and rich, thus alleviating conflict between the two classes. Inequality might not translate into civil unrest if organizations do not exist to mobilize the poor or if assets of the rich are hard to expropriate.

While many of these factors may, indeed, break the link between inequality and political outcomes, we suggest a simpler explanation. All these theories assume that key actors have an accurate measure of income inequality in their society. Yet, given how difficult it is to estimate the distribution of income and property—for skilled professionals, let alone time‐​constrained, statistically unsophisticated citizens—this assumption is implausible. People may fail to respond to inequality in the ways posited because, quite simply, they do not know how high it is.

Our research shows that such uncertainty and misperception are ubiquitous. We present evidence from a number of large‐​scale, cross‐​national surveys that show in recent years that ordinary people have known little about the extent of income inequality in their societies, its rate and direction of change, and where they personally fit into the distribution. What they think they know is often wrong. This finding is robust to data sources, definitions, and measurement instruments. For instance, perceptions are no more accurate if we reinterpret them as being about wealth rather than income. We do find, however, that the perceived level of inequality—and not the actual level—correlates strongly with demand for redistribution and reported conflict between rich and poor.

The implications of these results for theories of redistribution, revolution, and democratization are far‐​reaching. If these are to be retained at all, they need to be reformulated as theories not about actual inequality but about the consequences of beliefs about it, with no assumption that the two coincide.

Showing that perceptions of inequality are a better predictor of political preferences than is the actual level, we do not argue that the latter is measured correctly. Indeed,there are many reasons why inequality is difficult to measure. But that does not invalidate our main point: it makes it in another way. If the experts cannot assess inequality accurately, it strains credulity to suppose the man in the street can gauge it intuitively. And the difficulty of measuring the actual income distribution does not affect our second point: that perceptions of inequality—whether or not they are accurate—do correlate with political preferences.

We defer to future work the analysis of what causes perceptions of inequality, but several hypotheses seem plausible. First, individuals may exhibit the “What You SeeIs All There Is” fallacy and generalize from their immediate reference group to the country at large. Those living in villages where incomes and property are relatively equal may underestimate their country’s gap between rich and poor. Those working in professions with large variation in earnings may believe inequality to be higher than those in occupations where wage differences are smaller.

A second likely cause of beliefs about inequality is the media in general, and television in particular. Globalization of media might cause people in poor countries to compare their consumption to that glimpsed in rich Western states,leading them to exaggerate the degree of inequality and relative deprivation within their country. The more sensationalistic and celebrity‐​focused is television, the greater may be popular awareness of the extremes of income and consumption. Meanwhile, greater travel—both domestic and international—could also influence the traveler’s beliefs about relative incomes.

A third possible determinant is ideology, which may predispose people to “see” the level of inequality that their beliefs and values convince them must exist. Citizens in post‐​socialist countries appear to be particularly sensitive to inequality, perhaps reflecting the ideological legacy of communism. Socialists in all countries may exaggerate the income gaps around them, while conservatives may underestimate them. A Marxist may assume that capital is becoming ever more concentrated. A believer in free markets may suppose that wealth is trickling down to all. Shocks experienced by individuals during their formative years can leave scars in values and perceptions that last for life.

Fourth, besides mistaking their reference group for a valid sample, individuals may succumb to other psychological effects. A desire to blend in may cause them to believe their own income is closer to average than it actually is. The “self‐​enhancement” bias leads people at times to see themselves as better off than they actually are. Individuals may also fail to distinguish clearly between reports of high inequality worldwide and high inequality in their own country. And they may confuse changes in the trend with information about the level: when inequality has been rising, they may assume it must be high; when it has been falling, they may believe it to be low. At the same time, they may suppose that a recession—which shifts the incomes of the poor downward—must increase inequality, when in fact recessions sometimes decrease the income gap by shrinking the capital income of the rich more than the wage income of the poor. On the other hand, reports of rapid growth that arouse unrealistic expectations may lead people to believe that income has been diverted to the rich.

Whatever the causes, the gap between perceptions and reality—or, at least, statisticians’ best estimates of reality—is clear. And misperceptions must have been even greater in less data‐​rich and scientifically sophisticated eras. Besides inequality, we suspect that misperceptions also affect how people respond to inflation, unemployment, and other economic phenomena.

Often, in the social sciences, it is useful to model social processes as if people had certain information or made certain calculations, even when we know this is unlikely to have been the case. An influential argument holds that a theory should be judged not on the realism of its assumptions but on the accuracy of its empirical predictions. However, several decades of work seeking empirical connections between actual inequality levels and political preferences or behavior have turned up little. At this point, it may be more fruitful to root theories in more psychologically plausible assumptions.

This research brief is based on “Misperceiving Inequality,” by Vladimir Gimpelson and Daniel Treisman,” National Bureau of Economic Research Working Paper no. 21174, May 2015, http://​www​.nber​.org/​p​a​p​e​r​s​/​w​21174.

Media Name: rb32-cover.jpg

Download the Research Brief in Economic Policy

About the Authors

Vladimir Gimpelson, Higher School of Economics; and Daniel Treisman, University of California–Los Angeles, and National Bureau of Economic Research.