INTRODUCTION

A number of areas of research in economics might sometimes be thought not to pass the “So what?” test: Do we really care about this issue? This would not seem to be the case for inequality, which looks like it passes the test with flying colors.

Given this sense of unanimity, it might seem to be churlish in the extreme to want to write a chapter about individuals’ attitudes to inequality; surely they are negative aren’t they? We believe that the situation is not quite as simple as might be imagined. First, we have to ask the rather fundamental question of what we mean when we talk about income inequality, and then why would we expect any measure of such inequality to be correlated with individual well-being. Following on from this setting-out of the scene, there are a number of open questions. Is inequality equally bad for everyone? And on an extremely practical level, how can we tell? Last, the term inequality is used perhaps rather loosely in the empirical literature. It is ofinterest to ask which measures ofthe distribution of income are the most important (to individuals) in this context: Is it (as is commonly assumed) the Gini coefficient, or rather something else? As we will discuss later, recent work using experimental and survey methods has allowed considerable progress to be made in answering some of these questions.

To set the stage, we first ask under which circumstances others’ incomes should affect our own well-being.^[852] We use the term income inequality to refer to any disparities in incomes between individuals (i.e., there is income inequality when some individuals have different incomes than others). As opposed to many of the other variables that have been related to individual well-being, the distribution of income does not exist at the individual level: income inequality is rather measured only at an aggregate, often societal, level.

We believe that people do indeed have preferences over inequality. It is helpful to consider two broad types of individual attitudes to the distribution of income in a society. The first can be thought of as the individual’s disinterested evaluation of income inequal­ity: IfI see two distributions of income in some society, which do I believe is better? We will call this the normative evaluation of inequality.

In addition to this disinterested reaction to income inequality, the individuals who we analyze when we carry out experimental or survey analysis do actually live in the society in question: Their own income then forms part of the income distribution in which we are interested. This second inequality effect is at the individual level: income inequality will directly impact both the absolute income that individuals receive, and how much richer and poorer they are compared to others. The attitude to inequality here is not dis­interested but rather self-interested, with the additional assumption that individuals care not only about how much income they receive but also about how much they receive compared to others. We will call this the comparative evaluation of inequality.

The effect of the distribution of income on individual well-being will likely run through both of these channels. Even though income inequality as such in a society is not an individual-level concept, any distribution of income will have individual-level effects due to the way in which it changes the individual’s own income and their standing with respect to those who are richer and poorer, as will be discussed later.

In the context of relative standing or comparisons, individual attitudes to inequality will depend critically on the reference group that the individual has in mind.

The first of these, the comparative reference group, is in the spirit of the original inter­pretation given by Hyman, whereby the reference group acts as the standard of compar­ison for self-appraisal. The normative reference group is the source of norms, attitudes, and values of the individuals concerned. Both groups can be further distinguished according to whether the individual in question is or is not a member of the reference group. Reinterpreting Shibutani’s (1955) proposed conception of the terms, a comparative reference group is the point of comparison allowing the individual’s own status to be calculated when the individual is part of the group (as in Hyman). However, the indi­vidual need not (yet) be part of the reference group. When the individual is not part of the group, but aspires to be, the reference group acts as a relative aspiration, that is, as the group of which the individual desires to be a member. A normative reference group is that whose perspectives constitute the frame of reference for the individual, and again a distinction between membership and nonmembership can be effected. In the latter case, individuals may adopt the behavior of the group as a result of anticipatory socialization (see Merton and Kitt, 1950).

Regarding the subject matter of this chapter, the reaction of an individual to income inequality will depend on both the role assumed by the reference group and member­ship status in the group.

Comparative reference groups may also matter even if the individual is not currently a member of the group. If the individual aspires to be part of the group in question, then comparisons with respect to richer individuals in the group may give rise to positive feel­ings, as the individual anticipates being as rich as the group members once they join the group. This idea of a comparative reference group to which the individual aspires is akin to that of the tunnel effect in Hirschman (1973), which will be referred to in Section 13.2.1.

The rationale behind this comparative view of reference groups is that one’s own position relative to others matters. We do not imagine that this is the only way in which others’ outcomes may be viewed by the individual. It is very likely indeed that some groups will not be considered comparatively, but instead viewed with some kind of extended sympathy. The individuals to whom one compares and those for whom one feels sympathy are probably not going to be the same. As such, we may well see individuals whose position relative to their neighbors or work colleagues is para­mount, but who at the same time vote for social programs for those in need or give money to international charities. Here individuals have a preference for making some others better off. We will explore this idea of empathy or altruism a little more in Section 13.4.4.

As opposed to the comparative view of reference groups, inequality in the normative view of reference groups is evaluated by the individual irrespective of where she appears in the distribution, or even irrespective of whether she appears in it at all. Concretely, a given distribution of income will be evaluated in the same way by an individual regardless of whether she is in the top or bottom quartile of the distribution, so that there is no role for comparisons to the richer and poorer in the normative reference group. Equally, we can all now have a normative opinion about the distribution of income in our own coun­tries in the nineteenth century, even though we do not appear in that distribution our­selves. The normative evaluation of an income distribution can also be thought of as a mirror of preferences over inequality under the veil of ignorance (where the individual does not know where she will eventually be situated in the distribution).^[853]

Both the normative and comparative views of income inequality will likely depend on how the distribution of income came about. We expect individuals to be more tol­erant of the income gaps that result from effort than those that come about by luck. We will consider some of the work on the fairness of the income distribution further in Section 13.4.3.

The remainder ofthis chapter is organized as follows. Section 13.2 considers empirical evidence for an impact of income inequality in the context of a comparative reference group. We appeal to two different ways via which we can evaluate whether income inequality does indeed reduce the well-being the individuals who are exposed to it. The first approach relies on various measures of subjective well-being as proxies for indi­vidual utility: these are used to establish whether income disparities are indeed signifi­cantly associated with measures of individual well-being (such as happiness or life satisfaction).^[854] The second is to see whether individuals behave as if they wish to avoid income inequality. This is tantamount to a revealed-preference argument. As it is any­thing but obvious to obtain clean measures of behavior and match these to income inequality in the field, we turn to experimental techniques in the laboratory to make pro­gress here. Section 13.3 then follows the same structure, but this time with respect to the normative evaluation of income distributions. We propose a number of extensions, out­standing issues, and suggestions for future research in Section 13.4. Last, Section 13.5 concludes the discussion.

13.2.