INTRODUCTION

The economic literature on inequality has traditionally focused on income inequality. One reason to be interested in income inequality is that it may be linked to potential economic growth, to aggregate consumption, and to the occurrence and size of cyclical movements (see Chapter 14).

One normative reason to be concerned with income distribution is that we are ulti­mately interested in the distribution of well-being and that we consider income as a proxy for well-being. A related argument emphasizes the right of everyone to have access to a minimum level of resources, income then being an indicator of these resources. These two approaches are closely related if we define well-being directly in terms of resources, but they may differ if one adopts alternative definitions of well-being, for instance, in terms of functionings or capabilities or in terms of subjective satisfaction with life. A third nor­mative reason why people are interested in income distribution has to do with the fairness of the process through which income is acquired. There are strong convictions in society that individuals should be paid in a fair way and that effort should be somehow rewarded. There is more discussion about the ethical desirability of remunerating productivity differ­ences stemming from differences in innate talent or in socioeconomic background. These considerations are linked to the debate about the content of “desert” and “merit” and their relevance for evaluating the income distribution. At first sight, such a focus on the process of income formation is very different from a concern for the final distribution of well-being.

It is not obvious that income is indeed an adequate proxy for well-being. It is well accepted that the same monetary amount may yield a different level of well-being for individuals with different needs. Moreover, individuals do not care only about their income. A consensus seems to be emerging that information on other dimensions of life (such as health, job quality, the natural and social environment in which people are living) should be integrated into a richer view of well-being (see Stiglitz et al., 2009, and the references therein). This broadening of the perspective on well-being has led to a grow­ing aversion against the use of (even a “corrected” or “extended”) income metric to mea­sure well-being on the ground that this would reflect a kind of “resource fetishism.” Yet, from an applied viewpoint, monetary measures have the obvious advantage that they yield an operational and cardinal measure.

This brings us to the main question for this chapter: is it possible to formulate an eth­ically attractive notion of individual well-being that is richer than monetary income and that is still sufficiently operational to be used in applied welfare analysis?^[52]

The concept of individual “well-being” can be approached from many different per­spectives. One could, e.g., take a psychological perspective and investigate what is the best measure of well-being for describing and explaining the emotion of “feeling well” (Kahneman and Krueger, 2006). However, according to the welfare economic perspec­tive taken in this chapter, the choice of an adequate measure of individual well-being is not a psychological, but rather a normative question. An adequate measure of well-being makes interpersonal comparisons such that redistribution from a better-off individual to a worse-off individual yields a better state of affairs as seen from the social welfare point of view.^[53] In other words, an adequate measure of well-being serves as equalizandum for an egalitarian policy.

The choice of a particular metric of well-being is inevitably a matter of value judg­ments. Selecting income as the measure of well-being, or deciding to go beyond the income dimension is a normative choice. Indeed, the argument that we should include other dimensions than income because people care about these other dimensions, for instance, is only valid if we accept the normative position that society should care about what people care about.

Because defining an adequate concept of well-being is a normative choice, it is no surprise that opinions differ about what is the best measure of well-being. As soon as one moves beyond the single income dimension to describe well-being, at least two sets of issues come to the fore. First, what additional dimensions should be included? Or, more fundamentally, what justification can or should be given for this choice of dimen­sions? Second, should these different dimensions be seen as incommensurable, or is it possible to aggregate them into one measure of individual well-being? If one takes the former position and sticks to a vector representation of well-being, how should one han­dle interpersonal comparisons involving a trade-off between the different variables? If one takes the latter position, what should be the normative logic underlying the aggre­gation across dimensions?

In this chapter, we will describe the answers proposed by different approaches to these two sets of questions, and we will discuss their normative implications. In par­ticular, we will look at the different approaches from two specific perspectives. First, we focus on the extent to which the proposed measure of individual well-being respects individual preferences. The principle of individual sovereignty has always been one of the main tenets of economics and remains a hotly debated issue (see, among others, Hausman and McPherson, 2009). One of the difficulties in the debate is that different interpretations have been given to the concept of preferences.

Second, we will focus on the implications of selecting a specific measure of well-being for the delineation of the domain of personal responsibility. Ifwe start from the idea that a redistribution from someone with a higher level of well-being to someone with a lower level of well-being is an improvement from the social point of view, this implies (even when this is not made explicit) that the person at the lower level of well-being is not held responsible for this lower level of well-being. This responsibility perspective helps to interpret some of the normative differences between the various approaches. As responsibility and freedom are closely related concepts, a focus on responsibility will also allow us to comment on the different interpretations of freedom embedded in the dif­ferent approaches.

The chapter is structured as follows. Section 2.2 gives a brief historical sketch of the development of the literature. Section 2.3 is the core of the chapter. It contains a critical discussion of the three prominent proposals for a measure of well-being cap­turing nonincome dimensions: the capabilities approach, the subjective well-being approach, and the equivalent income approach. In Section 2.4 we consider the liter­ature on multidimensional inequality measurement and on multidimensional stochastic dominance, which has taken up a direct concern for the distribution in multiple dimensions without introducing explicitly a measure of individual well-being.

Before starting, we make two remarks. First, this chapter is about evaluating states of affairs. For such a broad evaluation, the income distribution is not sufficient; it is necessary to work with a broader concept of well-being. This does not mean that it would not be relevant to argue in favor of a redistribution ofincome. In fact, as soon as we define well­being to include personalized and nontransferable characteristics (such as health), a direct redistribution of well-being is not feasible. If income has a positive effect on well-being, a redistribution of income can be an effective instrument to realize a more equal distribu­tion of well-being. Yet, this will not necessarily be a redistribution from the (income-) rich to the (income-)poor because the income-rich can be at a lower level of well-being than the income-poor. This is precisely where different approaches to well-being will make a difference.

Second, we will focus on inequality rather than on poverty. The two concepts are complementary, but may still involve different ethical intuitions, especially if one accepts that poverty has an absolute component. It is probably not a coincidence that the concept of multidimensional deprivation has traditionally played a more important role in poverty than in inequality research. Poverty researchers are generally more sympathetic toward rights-based approaches. They are also less inclined to accept the idea that different life dimensions can be traded off. We will discuss these issues where needed.

2.2.