Inequality has been a central concern, if not the defining concern, of the political Left since at least the French Revolution. There has been, and continues to be, enormous variation in emphasis, of course:

Inequality of what? Rights of birth and legal status? Control over the means of production? Income, consumption, wealth? Unpaid care work? Recognition?

Inequality between whom? Aristocrats and bourgeois? Capitalists and workers? Workers in different positions? Women and men? Blacks and whites? Migrants and natives? Countries of the North and the South?

What to do about inequality? Behead the king? Abolish servitude, slavery, wage labor? Socialize the means of production? Introduce representative democracies and redistributive welfare states? Lead struggles of national liberation? Change discriminatory social norms? Dismantle the system of nation-states and citizenship?

These concerns and struggles for equality have been reflected at times in philosophical discussions and in the social sciences developing over the same time period, particularly in economics. Distributional questions occupied economists from the outset of the discipline, when Ricardo famously stated that, “To determine the laws which regulate this distribution [between rent, profit, and wages], is the principal problem in Political Economy,” on which Marx followed up. In the twentieth century, Nobel memorial prizes in economics were given to economists working on inequality.1 Questions of (re)distribution feature prominently in the fields of labor economics and public finance.

There has furthermore been a resurgence of interest in inequality within economics recently. This resurgence might be partly owed to the massive increase in inequality within the Anglo-Saxon countries in a time of reduced economic growth since the neoliberal turn of the 1980s, and partly due to the media attention that the work of Piketty and others have received.

It remains to be seen how much of this resurgence amounts to an actual change of focus within the discipline. Despite the long tradition of research in inequality and despite the resurgence of interest, questions of inequality still remain marginal within the core economics curriculum. Introductory classes tend to either consider models where everybody is the same along key dimensions (so that inequality doesn’t even show up), or put normative emphasis on some notion of efficiency or some aggregates, both of which neglect questions of inequality. This neglect is somewhat surprising in the sense that a focus on inequality would be very natural given economists’ basic normative assumptions.

I will take some of these assumptions as a point of departure, and will take them to their logical conclusions, which are rarely spelled out.

One of the fundamental tenets of welfare economics is “normative individualism”: social welfare, which measures how good a society is, or how good the outcomes of some policy are, is a function of individuals’ welfare. A second tenet of welfare economics is that individual welfare, measured as “utility,” is not objectively comparable across individuals.

“Who benefits or loses from such changes, and by how much?”

The first of these assumptions, normative individualism, implies that any normative statement about policy alternatives or historical developments has to start with a description of who benefits or loses from such changes, and by how much—effectively, from a description of inequality and its changes.

The second assumption, non-comparability of utility across individuals, implies that any normative position on policies that generate winners and losers—which means pretty much all policies—boils down to taking sides in a distributional struggle. There is no way to make policy recommendations without being party to one or the other side in the distributional struggles that have characterized the Left since its inception. If a researcher wants to avoid taking sides in such a manner, she has to stop short at describing inequality and distributional impacts of changes, leaving normative evaluations to her reader. The rest of this essay elaborates.

Why do research on economic inequality?

Why should we care about inequality, and why should we do research on it? Several kinds of reasoning can be distinguished.

First, one might be worried about the consequences of economic inequality, which one might consider important in their own right. Possible consequences are:

Political consequences: An increasing concentration of income and wealth, and the rising influence of campaign donations and lobbying, might undermine democratic institutions predicated on the principle of “one person, one vote.”

Social consequences: Increasing inequality might further social segregation (residential, educational, etc.), thus reducing knowledge of how others live and undermining social cohesion and solidarity.

Economic consequences: Rising inequality might destabilize the economy. The increase in mortgage lending as a substitute for income growth of the bottom half of the distribution, for example, was at the origin of the financial crisis starting in 2008.

Whether rising inequality has these and other effects is a difficult empirical question.

Second, by studying economic inequality we learn how much it has changed over time and across countries, and in particular how much it is affected by policy decisions and other social factors. Recognizing these facts puts into question explanations of inequality that are ahistorical and asocial. These include biological explanations (such as biological racism, sexism, or justifications of inequality based on genetic differences in IQ) or explanations that reduce inequality to a matter of individual responsibility.

“One needs to study inequality as soon as one accepts the perspective of normative individualism.”

Third, as noted, one needs to study inequality as soon as one accepts the perspective of normative individualism, as do most modern theories of justice. This perspective implies that any normative evaluation of policy changes has to start by evaluating who wins and who loses, and by how much, and has to trade off the welfare of different individuals.

Finally, a long-standing line of ethical reasoning suggests that we should put the bulk of normative weight on those who are worst off and aim for an equalization of welfare. John Rawls for instance argued in A Theory of Justice2 that we should evaluate societies by imagining that we are behind a “veil of ignorance,” which prevents us from knowing who we actually are. In such a setting of uncertainty, we should try to make the worst-off person as well off as possible, ensuring a minimum standard of living for ourselves should we happen to be among the worst off. The point to be made here, though, is that we need to understand inequality for any normative judgment based on individual welfare, whether or not we follow Rawls’s reasoning.

Normative individualism

The framework of normative individualism implies that we have to start by considering a set of individuals.

This already raises the first set of difficult questions: Who is to be included in this set of individuals? Many discussions implicitly assume we are considering a given “society.” But does that just mean everybody of a certain citizenship, or everybody living in a certain territory? Why not all living human beings; should humans of another country count for nothing? And what about future generations? What about animals?

Given the set of individuals, we next need to decide how to measure their welfare. The goal is to assign a number to each individual that captures some notion of what it means for this individual to do well. How to measure individual welfare again raises a whole set of difficult questions.

A minimalist notion would only consider the formal legal rights enjoyed by individuals. A broader notion might also take into account various resources that allow individuals to achieve their objectives, such as education, income, and health. A comprehensive notion of opportunities might aim to take into account all factors that influence individuals’ options, and evaluate the options effectively available to them. And we might finally consider the outcomes actually achieved by individuals, evaluated either by some common criteria or by their individual preferences.

Utilitarianism (also called welfarism), the most common perspective in economics, evaluates individual welfare by the outcomes actually achieved as evaluated by individual preferences. This yields an ordinal, non-comparable measure of individual welfare.

“A ‘social welfare function’ maps the measures of individual welfare into a measure of social welfare.”

Given the set of individuals, and given evaluations of their welfare, we finally ask how well society as a whole is doing. Formally, we consider a “social welfare function,” which maps the measures of individual welfare into a measure of social welfare. This function determines how much we care about different individuals; usually it is increasing in the welfare of each individual. It tells us how much weight we assign to the welfare of Trevon relative to the welfare of Emily, the welfare of Sophia relative to the welfare of José.

The function might treat different people similarly, and not care about names. In that case it would still tell us how much we care about an additional dollar for a poor person versus a rich person, for a sick person versus a healthy person, since these would affect the levels of individual welfare. And the social welfare function might not depend on names, but on some characteristics of individuals. The function might for instance incorporate the belief that race, gender, or parental income should not determine individual welfare.

This individualist framework, where we evaluate social welfare as a function of individuals’ welfare, is compatible with radical or conservative viewpoints. It does impose some restrictions, however.

This framework is not compatible with fascist approaches (“you are nothing, your people is everything”), or perfectionist approaches (“greatness of philosophy and art justified slavery”). The framework is also not compatible with libertarian approaches which consider outcomes to be just as long as they are the consequence of private property, contracts, and voluntary exchange on markets. This framework, finally, does not explicitly take into account environmental concerns, to the extent that we care about biodiversity, say, for its own sake rather than because of its impact on future generations or because of the welfare of individual animals.

It is easily possible, of course, to consider objective functions that depend both on the welfare of individuals and on other objectives, such as environmental concerns.

Welfare weights

Whichever way individual welfare is measured, in order to make normative statements we must aggregate individual welfare to social welfare, using some function.

In order to understand the properties of such a function, it is useful to consider small changes of policy, which result in corresponding small changes of welfare for each individual. It is then useful to think about aggregation in terms of welfare weights. The welfare weight for a given person measures how much social welfare changes when we change individual welfare for that person by a small amount.

In a utilitarian context, individual welfare changes are often expressed in monetary terms. In this case, the ratio of welfare weights for two individuals measures how much we care about an additional US dollar for one individual relative to an additional US dollar for the other one.

An egalitarian position assigns a large welfare weight to those who are worst off in a society. More generally, we can think of the ratio between welfare weights for poor versus rich people as measuring how egalitarian some aggregation is.

We can also think about welfare weights “in reverse:” Given the policy choices actually made in a society, there is a corresponding set of welfare weights that justifies these policy choices. Such weights can be thought of as measuring effective social power—whose interests are represented by the powers that be. If, for instance, policies are implemented that mainly benefit the rich, these policies can only be rationalized by large welfare weights for the rich, etc.

The concept of welfare weights is also useful to understand what various popular statistics and measures of inequality are doing. How do these statistics change when we change the income (or some other input into individual well-being) of any given individual by a little?

GDP is the simplest case; it basically sums up the incomes of all residents. So an additional dollar for any given resident—whether homeless or billionaire—receives the same weight. This is of course a rather anti-egalitarian aggregation.

A more interesting case is the poverty rate (the share of individuals with incomes below some predefined poverty line). When we change the income of somebody richer than the poverty line, the poverty rate does not change, so they receive a weight of zero. Similarly, and maybe counterintuitively, the poverty rate does not change either if we change the incomes of people who are significantly below the poverty line, so they receive a weight of zero, too! The poverty rate does change, however, if we change the income of somebody just above or below the poverty line, moving them across the line.

A popular measure of inequality is the Gini coefficient. The Gini coefficient corresponds to welfare weights which are a linear function of the rank of a person in the income distribution.

The variance of incomes behaves similarly, in that it corresponds to welfare weights which are a linear function of the level of income of a person.

Let’s finally consider the share of the top 1 percent in the income distribution—a statistic receiving prominence in Occupy Wall Street’s slogan, “We are the 99 percent.” The top income share is affected equally by giving a US dollar to anybody in the bottom 99 percent, so they all receive the same weight; the same holds for everybody in the top 1 percent with opposite signs.


Let us conclude by summarizing what follows from a normatively individualist perspective for research practice.

First, researchers should always report disaggregated results: How does the distribution of relevant outcomes or resources look, how does it differ across demographic groups, how does it change as a function of policy changes or historical developments, who wins and who loses?

One good tool for such reporting are quantile functions (also known as Penn’s parade), which we used for instance on to visualize the wealth distribution, readers’ position in it, and how it would be affected by redistributive wealth taxes.

Second, be conscious that most changes and policies—in particular politically contentious ones—generate winners and losers. Making a policy recommendation in such a context always means taking sides; there is no escaping being party to distributional struggles.

Third, on a point I had less space to elaborate on, the variables available in common datasets do in general not measure individual welfare, and should not be taken as ultimate normative objectives. Wages, earnings, income, consumption, wealth, education, health, political rights, etc. are all important in their own right, but they are not identical to any well-justified measure of individual welfare.

This piece builds on ideas discussed in the open online textbook on empirical research on economic inequality available at For further technical detail and elaboration of the thoughts presented here, please consult the textbook.


Posted on October 4, 2016