4.10: Can the capability approach change welfare economics?

Of all the (sub)disciplines where the capability approach is relevant, welfare economics may well be the one where it is most difficult to describe its impact. The reason is that the capability approach could be seen in two very different lights, depending on one’s own position towards the current state of economics: either as an improved modification of mainstream welfare economics, or else as a path that could lead us to a very different type of welfare economics, which would radically break with some mainstream assumptions and practices. One could say that the welfare economists interested in the capability approach have two very different agendas: the first group only wants some changes in the normative focus, and possibly in some of the ontological and behavioural assumptions in the theory development, but no methodological or meta-theoretical changes, whereas the second group wants a paradigm change or a scientific revolution, in which there would be meta-theoretical and methodological pluralism (module B7) and much richer or thicker accounts of human agency (module B3) and structural constraints (module B5). In addition, it makes a difference whether we analyse the possibilities for a capabilitarian theoretical welfare economics or for a capabilitarian empirical welfare economics.

4.10.1 Welfare economics and the economics discipline

Before analysing the reach and limit of the capability approach in these various endeavours, a few general comments are in order about economics in general, and welfare economics in particular. Let us first ask: what is welfare economics? As Sen (1996, 50) writes, “Welfare economics deals with the basis of normative judgements, the foundations of evaluative measurement, and the conceptual underpinnings of policy-making in economics”. While, in practice, much of the economics discipline is concerned with policy advice, welfare economics is nevertheless a small subfield of economics, and is by some prominent welfare economists seen as unduly neglected or marginalised by mainstream economics (Atkinson 2001). One important reason is that welfare economics makes explicit the inevitable normative dimensions of economic policy analysis and evaluations, and most economists have been socialised to believe that ‘modern economics’ is value-free, and that anything to do with normativity can be outsourced to ethics or to a democratic vote. In reality, however, the imaginary science-value split that mainstream economists would wish for is, for many economic questions, impossible (Reiss 2013; Hausman, McPherson and Satz 2016). It would therefore be much better to face this inevitability upfront, and understand economics as a moral science (Boulding 1969; Atkinson 2009; Shiller and Shiller 2011) rather than as applied mathematics or a form of value-free modelling. But in economics, as in any other discipline, there are complicated sociological processes conveying views about authority and status, as well as unexamined beliefs about what ‘good science’ is and which type of objectivity is most desirable: one is not born an economist, but becomes one through one’s training, which is in part also a socialisation process (Colander and Klamer 1987; McCloskey 1998; Nelson 2002). Unfortunately, there is empirical evidence that many economists are unwilling to engage with these fundamental questions and hold on to the belief that economics is superior to other social sciences and has little to learn from other disciplines (Fourcade, Ollion and Algan 2015). Many economists who are interested in economic questions but are not endorsing the myth of value-free social science, or who crave more methodological and meta-theoretical freedom, have left for another discipline that offers them those liberties.¹⁵ After all, economists do not have a monopoly on economic topics, and there are many questions about such topics that are analysed by economic sociologists, economic historians, political economists, economic geographers, and economic philosophers. In my view, one cannot analyse the reach and limits of the capability approach in welfare economics if one does not acknowledge the high levels of discontent and methodological conservatism within economics, which cannot be found in any other discipline that engages with the capability approach.

With this background in mind, we can now proceed to ask whether the capability approach can make a difference to welfare economics. First, in section 4.10.2, we will look at the main theoretical contribution of the capability approach to welfare economics: its contribution to the development of non-welfarist welfare economics. In section 4.10.3 we will analyse what kind of empirical analyses a capabilitarian welfare economics could make, and what its challenges and possibilities are. Finally, in section 4.10.3, we analyse what challenges the development of a heterodox capabilitarian welfare economics would face.

4.10.2 Non-welfarism

The main theoretical contribution of the capability approach is that it contributes to the development of post-welfarism or non-welfarism in welfare economics. Welfarism is the position that social welfare depends exclusively on individual utilities, which are either understood in a hedonic or in a desire-satisfaction sense, and this has been the dominant position in economics for a long time. Post-welfarism broadens the informational basis of interpersonal comparison with non-utility information, such as deontic rights, or objective information such as people’s functionings and capabilities. In a series of publications, Sen has offered strong theoretical arguments to move from welfarism to non-welfarism (sometimes also called ‘post-welfarism’) and has inspired other welfare economists to work on a post-welfarist welfare economics (e.g. Gaertner and Xu 2006, 2008; Gotoh and Yoshihara 2003; Gotoh, Suzumura and Yoshihara 2005; Pattanaik 2006; Pattanaik and Xu 1990; Suzumura 2016; Xu 2002). Some reasons for this move are the same as the arguments against desire-satisfaction theories or the happiness approach that we reviewed earlier in sections 3.7 and 3.8. Another argument is that relevant information is left out of the informational basis. If two social states have exactly the same utility levels, but social state A has also a set of legal and social norms that discriminate against one group of people, whereas in social state B, the principles of moral equality and non-discrimination are protected, then surely, we should prefer social state B over social state A. But welfarism, because of its exclusive focus on utilities, is unable to take any type of non-utility information into account, whether it is the violation of deontic principles, information on rights, liberties and justice, or information on inefficient or unsustainable use of common resources. Many of the welfare economists who have embarked on the development of a post-welfarist welfare economics have focussed on the importance of freedom as an important part of the widening of the informational basis.

Non-welfarist welfare economics requires some changes to our approach to welfare economics. As Sen (1996, 58) noted in his discussion of the contribution of the capability approach to non-welfarist welfare economics, if we move to an informational basis with multiple dimensions of different types (as in the capability approach) then this requires explicit evaluations of the different weights to be given to the contributions of the different functionings and capabilities to overall (aggregate) social welfare. For Sen, the way to proceed is by public reasoning about those weights. This should probably not be seen as the only and exclusive way to determine them, since not all work in welfare economics is suited for public discussion — for example, it often entails desk-studies of inequalities or the analysis of the welfare effects of certain policy measures, and it is practically impossible to organise an exercise of public reasoning for every desk study that welfare economists make. Luckily, as the survey by Decancq and Lugo (2013) shows, there are various weighting systems possible that can give us the weights that are needed if one wants to aggregate the changes in different functionings and capabilities. For example, Erik Schokkaert (2007) has suggested that we derive the weights of the functionings from the contribution they make to the life-satisfaction of people, after those weights are cleaned of ethically suspicious information.

However, as we saw in section 4.10.1, many (possibly most) economists are unwilling to engage in explicit evaluations, since they believe in the science/value split and believe that economics can be value-free. This makes it harder for welfare economics to engage in such normative work, since they run the risk that their peers will no longer accept their approach as ‘economics research’. But it is inconsistent to reject all explicit evaluative exercises. Economists are happy working with GNP per capita and real income metrics as proxies of welfare, which uses market prices as the weights. But this is equally normative: it is assumed that the welfare-value of a certain good for a person is reflected by the price that the good commands on the market. This is problematic, for reasons that have been explained repeatedly in the literature. For one thing, market prices reflect demand and supply (and thus relative scarcity of a good) — diamonds are expensive and water (in non-drought-affected places) is cheap — but this doesn’t say anything about their importance for our wellbeing. Moreover, market prices do not take into account negative or positive welfare effects on third parties, the so-called externalities, despite their omnipresence (Hausman 1992).

Of course, it may be that, upon reflection of the various weights available, some capability theorists will conclude that the set of market prices, possibly combined with shadow prices for non-market goods, is the best way to proceed. That is quite possible, and would not be inconsistent with the general claim in the capability approach that weights need to be chosen. The point is rather that the choice of weights needs to be done in a reflective way, rather than simply using the weighing scheme that is dominant or customary. I take it that this is the point Sen is trying to make when he argues that “Welfare economics is a major branch of ‘practical reason’” (Sen 1996, 61).¹⁶

4.10.3 Empirical possibilities and challenges

When Amartya Sen introduced the capability approach in economics, there was some scepticism about its potential for empirical research. For example, Robert Sugden (1993, 1953) famously wrote:

Given the rich array of functionings that Sen takes to be relevant, given the extent of disagreement among reasonable people about the nature of the good life, and given the unresolved problem of how to value sets, it is natural to ask how far Sen’s framework is operational. Is it a realistic alternative to the methods on which economists typically rely — measurement of real income, and the kind of practical cost-benefit analysis which is grounded in Marshallian consumer theory?

What Sugden and other early welfare economic critics of the capability approach, such as John Roemer (1996, 191–93) were looking for, is a theory that is fully formalised and provides a neat algorithm to address questions of evaluation and/or (re-)distribution, resulting in a complete ranking of options. That requires two things: first, to be able to put the capability approach in a fully formalized model which can be econometrically estimated. This requires us to move beyond the welfare economic models as we know them, and may also require the collection of new data (Kuklys 2005). In addition, it requires us to accept that the different dimensions (functionings and/or capabilities) are commensurable, that is, have a common currency that allows us to express the value of one unit of one dimension in relation to the value of one unit of another dimension. One-dimensional or aggregated evaluative spaces are, ultimately, a necessary condition for conducting empirical work in contemporary mainstream welfare economics. Yet there may well be a trade-off between the number of dimensions and the informational richness of the evaluative space on the one hand, and the degree to which the theory can be formalised and can provide complete orderings of interpersonal comparisons on the other hand. Some welfare economists are working on the question of how to aggregate the many dimensions such that one has, in the end, one composite dimension to work with, but it should be obvious that this is not the only way to develop capabilitarian welfare economics. The alternative is to stick to the view that wellbeing is inherently multidimensional, which requires other methods and techniques that allow for fuzziness, vagueness and complexity (Chiappero-Martinetti 2008, 2000, 1994, 2006; Clark and Qizilbash 2005; Qizilbash and Clark 2005). One could also advance work on dominance rankings or incomplete rankings, which Sen has been defending in his social choice work for several decades (e.g. Sen 2017). As a consequence, there are several ways to develop a capabilitarian welfare economics, and to make the capability approach “operational” (Atkinson 1999, 185).

Sugden’s objection can be best answered by looking at the applications that have already been developed, and which have been listed in several overviews of the empirical literature (e.g. Kuklys and Robeyns 2005; Robeyns 2006b; Chiappero-Martinetti and Roche 2009; Lessmann 2012).¹⁷ However, whether the applications discussed in those surveys satisfy the critics depends on what one expects from empirical work in capabilitarian welfare economics. As was already shown in chapter 1, empirical applications of the capability approach do make a substantive difference to research using other normative frameworks, such as income-based metrics. But from that it doesn’t follow that capabilitarian welfare economics will be able to deliver alternatives for each and every existing welfarist study in economics. It may well be that sometimes the informational riches of the capability approach clash with requirements regarding measurability that certain empirical applications put upon the scholar.

4.10.4 Towards a heterodox capabilitarian welfare economics?

In the last two sections, we discussed how the capability approach can make a difference to contemporary welfare economics both theoretically and empirically. Those debates by and large stay within the mainstream of contemporary welfare economics, even though as Amartya Sen (1996, 61) notes, they require us “to go more and more in these pluralist and heterodox directions, taking note of a variety of information in making the wide-ranging judgements that have to be made”. The reference to ‘heterodoxy’ that Sen makes here is limited to the informational basis of evaluations, yet in other work he has challenged some of the behavioural assumptions underlying mainstream welfare economics (e.g. Sen 1977a, 1985b). However, other economists believe that we need a much more radical heterodox and pluralist turn in economics, which would also affect meta-theoretical views, the range of methods that can be used (e.g. including qualitative methods), giving up on the belief that economics can be value-free, and engaging much more — and much more respectfully — with the other social sciences, and indeed also with the humanities. What can these heterodox economists expect from the capability approach?

The answer to that question flows from the description of the state of economics that was given in section 4.10.1. The unwillingness of mainstream welfare economics to genuinely engage with other disciplines (Fourcade, Ollion and Algan 2015; Nussbaum 2016) clashes with the deeply interdisciplinary nature of the capability approach. The modular view of the capability approach that was presented in chapter 2 makes it possible to see that a heterodox capabilitarian welfare economics is certainly possible. It could not only, as all non-welfarist welfare economics does, include functionings and/or capabilities as ends in the evaluations (A1) and possibly include other aspects of ultimate value too (A6), but it could also include a rich account of human diversity (B3), a richly informed account of agency and structural constraints (B4 and B5) and it could widen its meta-theoretical commitments (B7) to become a discipline that is broader and more open to genuine interdisciplinary learning. Yet the modular account also makes it very clear that mainstream welfare economists can make a range of choices in those modules that are more in line with the status quo in current welfare economics, which would result in a very different type of capabilitarian welfare economics.

In short, while both types of capabilitarian welfare economics will depart in some sense from welfarist welfare economics, we are seeing the emergence of both mainstream capabilitarian welfare economics and heterodox capabilitarian welfare economics. The problem for the first is that it will have few means of communication with other capability theories, since it does not adopt many of the interdisciplinary choices that most other capability theories make. The problem for the latter is that it will not be taken seriously by mainstream economics, since it does not meet the narrow requirements of what counts as economics according to the vast majority of mainstream economists. Taking everything together, a capabilitarian welfare economics is possible, but (a) it will be harder to develop the capability approach in welfare economics than in some other disciplines because of the methodological and meta-theoretical clashes and restrictions, and (b) the difficult position that welfare economics occupies within mainstream economics will become even more challenging, since moving in the direction of the capability approach conflicts with the criteria that the gatekeepers in mainstream economists impose on anyone who wants to do something considered ‘economics’. This may also explain why there is much less work done in welfare economics on the capability approach, compared to some other disciplines or fields in which the capability approach has made a much bigger impact.

---

15 There are plenty of academic economists who have moved to history, development studies or philosophy in order to enjoy the greater methodological and paradigmatic freedoms in those disciplines.

16 Note also that for welfare economists, an important concern in examining and developing a capabilitarian welfare economics will be the question of how it can be formalized. On formalizations of the capability approach, see Sen (1985a); Kuklys and Robeyns (2005); Basu and López-Calva (2011); Bleichrodt and Quiggin (2013).

17 In addition to the first, rather rough empirical application in the Appendix of Sen’s (1985a) Commodities and Capabilities, the literature on empirical applications in welfare economics that used individual-level data started off with the paper by Schokkaert and Van Ootegem (1990), in which they showed that unemployment benefits may restore an unemployed person’s income level, but do not restore all of her functionings to the level they were at before she become unemployed.