Keywords

JEL Classifications

Justice and Utilitarianism

The concept of justice is often invoked in economic discussions. Its relevance to economic evaluation is obvious enough. However, it is fair to say that in traditional welfare economics, when the notion of justice has been invoked, it has typically been seen only as a part of a bigger exercise, viz., that of social welfare maximization, rather than taking justice as an idea that commands attention on its own. For example, in utilitarian welfare economics (e.g. Pigou 1952; Harsanyi 1955) the problem of justice is not separated out from that of maximization of aggregate utility. This situation has been changing in recent years, partly as a result of developments in moral philosophy dealing explicitly with the notion of justice as a concept of independent importance (see especially Rawls 1971, 1980).

In the utilitarian formulation the maximand in all choice exercises is taken to be the sum-total of individual utilities. The approach can be seen as an amalgam of three distinct principles: (1) welfarism, (2) sum-ranking, and (3) consequentialism. Welfarism asserts that the goodness of a state of affairs is to be judged entirely by the utility information related to that state, i.e., by information about individual utilities. All other information is either irrelevant, or only indirectly relevant as a causal influence on utilities (or as a surrogate for utility measures when such measurement cannot be directly done). The second principle is sum-ranking, which asserts that the goodness of a collection of utilities (or welfare indicators) of different individuals, taken together, is simply the sum of these utilities (or indicators). This eliminates the possibility of being concerned with inequalities in the distribution of utilities, and the overall goodness or ‘social welfare’ is seen simply as the aggregate of individual utilities. The third principle is consequentialism, which requires that all choice variables, such as actions, rules, institutions, etc., must be judged in terms of the goodness of their respective consequences. The overall effect of combining these three principles is to judge all choice variables by the sum-total of utilities generated by one alternative rather than another.

Sum-Ranking and Equality

A theory of justice can take issue with each of the principles underlying utilitarianism, and in fact in the literature that has developed in recent decades, each of these principles has been seriously challenged (see the papers included in Sen and Williams 1982). Some critiques have been particularly concerned with assessing and questioning the axiom of sum-ranking, and have considered the claims of equality in the distribution of well-being (see, for example, Phelps 1973; Sen 1973, 1977, 1982; Kern 1978).

The summation formula can be defended either directly (e.g., in terms of attaching equal importance to everyone’s ‘interest’: see Hare 1981, 1982), or indirectly through invoking some model of ‘impersonality’ or ‘fairness’ (e.g., involving a hypothetical choice in a situation of primordial uncertainty, in which each person has to assume that he or she has an as if equal probability of becoming anybody else: see Vickrey 1945; Harsanyi 1955). Other routes to deriving sum-ranking involve independence or separability requirements of various kinds (see d’Aspremont and Gevers 1977; Deschamps and Gevers 1978; Maskin 1978; Gevers 1979; Roberts 1980; Myerson 1981; Blackorby et al. 1984; d’Aspremont 1985).

Whether the defences obtainable from these approaches are convincing enough has been a matter of some dispute. There have also been some interpretative discussions as to whether giving equal importance to everyone’s ‘interest’ does, as alleged, in fact yield the formula of summing individual utilities irrespective of distribution, and also whether the additive formula that is obtained on the basis of hypothetical primordial choice is, in fact, a justification for adding individual utilities as they might be substantively interpreted in welfare economic exercises (see Pattanaik 1971; Smart and Williams 1973; Sen 1982, 1985a; Blackorby et al. 1984; Williams 1985). It is not obvious that this debate has been in any way definitively concluded one way or the other.

The Difference Principle and Leximin

Meanwhile, much attention has been paid to developing welfare-economic rules based on taking explicit note of inequalities in the distribution of utilities. A definitive departure on this came from Suppes (1966). Another major approach was developed in Rawls’s (1971) Theory of Justice, even though Rawls himself was concerned not so much with the distribution of utilities but with that of the indices of primary goods (on which more later). The concern with the utility level of the worst-off individual has been formalized and reflected in various formulae suggested or derived in the rapidly growing welfare-economic literature on this theme. In particular, James Meade (1976) has provided an extensive treatment of this type of distributional issues, and it has also been penetratingly analysed by Kolm (1969), Phelps (1973, 1977), Atkinson (1975, 1983), Blackorby and Donaldson (1977), and others.

In fact, the Rawlsian ‘Difference Principle’, which judges states of affairs by the advantage of the least well-off person or group, has often been axiomatized in welfare economics and in the social-choice literature by equating advantage with utility. In this form, the ‘lexicographic maximin’ rule (proposed in Sen 1970) has been axiomatically derived in different ways. The rule judges states of affairs by the well-being of the worst-off individual. In case of ties of the worst-off individuals’ utilities, the states are ranked according to the utility levels of the second worst-off individuals respectively. In case of ties of the second worst-off positions as well, the third worst-off individuals’ utilities are examined. And so on.

There is no necessity to interpret these axioms in terms of utilities only,s and in fact the analytical results derived in this part of the social-choice literature can be easily applied without the ‘welfarist’ structure of identifying individual advantage with the respective utilities. Various axiomatic derivations of lexicographic maximin – ‘leximin’ for short – can be found in Hammond (1976), Strasnick (1976), Arrow (1977), d’Aspremont and Gevers (1977), Sen (1977), Deschamps and Gevers (1978), Suzumura (1983), Blackorby et al. (1984), d’Aspremont (1985), among others. These can be seen as exercises that incorporate concern for reducing inequality, related to recognizing the claims of justice.

While the Rawlsian approach rejects the aggregation procedure of utilitarianism(i.e., ranking by sums), a major aspect of the Rawlsian theory involves the rejection of utility as the basis of social judgements (i.e., welfarism). Rawls (1971) argues for the priority of the ‘principle of liberty’, demanding that ‘each person is to have an equal right to the most extensive basic liberty compatible with similar liberty for others’. Then, going beyond the principle of liberty, claims of efficiency as well as equity are both supported by Rawls’s ‘second principle’ which inter alia incorporates his ‘Difference Principle’ in which priority is given to furthering the powers of the worst-off group. These powers are judged by indices of ‘primary social goods’ which each person wants (Rawls 1971, pp. 60–65).

Primary goods are ‘things that every rational man is presumed to want’, including ‘rights, liberties and opportunities, income and wealth, and the social bases of self-respect’. The Difference Principle takes the form, in fact, of maximin, or lexicographic maximin, based on interpersonal comparisons of indices of primary goods. This rule can be axiomatized in much the same way as the other ‘lexicographic maximin’ rule based on utilities, and all that is needed is a reinterpretation of the content of the axioms (with the objects of value being indices of primary goods rather than utilities).

The Rawlsian approach to justice, therefore, involves rejection both of welfarism and of sum-ranking. Furthermore, consequentialism is disputed too, since the priority of liberty might possibly go against judging all choice variables by consequences only. At least, in the more standard forms, consequentialism does involve such a conflict, even though it is arguable that the problem can be, to a great extent, resolved by a broader understanding of consequences, which takes into account the fulfilment and violation of liberties and rights, and also the agent’s special role in the actions performed (Sen 1985a).

Utilities, Primary Goods and Capabilities

The claim of primary goods to represent the demands of justice better than utilities is based on the idea that utilities do not reflect a person’s advantage (in terms of well-being or powers) adequately. It is arguable that in making interpersonal comparisons of advantage, the metric of utilities (either in the form of happiness, or of desire fulfilment) may be biased against those who happen to be hopelessly deprived since the demands of unharrassed survival force people to take pleasure in small mercies and to cut their desires to shape in the light of feasibilities (see Sen 1985a, b). The status of ‘preference’ may be disputed in view of the need for critical assessment (see Broome 1978; McPherson 1982; Goodin 1985, among others). Also, what types of pleasures should ‘count’ can itself be a matter for an important moral judgement. As Rawls (1971) points out, in the utilitarian formulation, we have the unplausible requirement that

if men take a certain pleasure in discriminating against one another, in subjecting others to a lesser liberty as a means of enhancing their self-respect, then the satisfaction of these desires must be weighed in our deliberations according to their intensity, or whatever, along with other desires (pp. 30–31).

These and other types of difficulties have been dealt with by some utilitarians through moving to less straight-forward versions of utilitarianism, for example Harsanyi’s (1982) exclusion of ‘all antisocial preferences, such as sadism, envy, resentment, and malice’ (p. 56); see also the refinements proposed by Hare (1981, 1982), Hammond (1982) and Mirrlees (1982).

Recently, it has been argued that primary goods themselves may be rather deceptive in judging people’s advantages, since the ability to convert primary goods into useful capabilities may vary from person to person. For example, while the same level of income (included among ‘primary goods’) may give each person the same command over calories and other nutrients, the nourishment of a person depends also on other parameters such as body size, metabolic rates, sex (and if female, whether pregnant or lactating), climatic conditions, etc. This indicates that a more plausible notion of justice may demand that attention be directly paid to the distribution of basic capabilities of people (see Sen 1982, 1985b). The approach goes back to Smith’s (1776) and Marx’s (1875) focus on fulfilling needs.

The achievement of capabilities will, of course, be causally related to this command over primary goods, and the capabilities, in their turn, will also influence the extent to which utilities are achieved, so that the various alternative measures will not be independent of each other. However, the basic issue is the variable that should be chosen to serve as the proper metric for judging advantages of people – the equity and the distribution of which could form the foundations of a theory of justice. On this central issue several alternative views continue to flourish in the literature.

Fairness and Envy

A view of justice that is not altogether dissimilar from Rawls’s concerned with primary goods is captured by the literature on ‘fairness’, inspired by a pioneering contribution of Foley (1967). In this approach a person’s relative advantage is judged by the criterion as to whether he or she would have preferred to have had the commodity bundle enjoyed by another person. This has been seen as a criterion of ‘non-envy’. If no one ‘envies’ the bundle of anyone else, the state of affairs is described as being ‘equitable’. If a state is both equitable and Pareto efficient, it is described as being ‘fair’ (even though the term fairness is also sometimes used interchangeably with only ‘equitability’).

There has been an extensive literature on existence problems, in particular whether equitability can be combined with efficiency in all circumstances. (The answer seems to be no, especially when production is involved: Pazner and Schmeidler 1974.) There has also been considerable exploration of the effects of varying the criterion of equitability and fairness to reflect better the common intuitions regarding the requirements of justice. Various results on these problems and related ones have been presented, among many others, by Foley (1967), Schmeidler and Vind (1972), Feldman and Kirman (1974), Varian (1974, 1975), Svensson (1980), and Suzumura (1983).

It should be remarked that the fairness criterion does not provide a complete ranking of alternative states. It identifies some requirements of justice, which makes the states fair. Varian (1974) has argued, with some force, that ‘social decision theory asks for too much out of the process in that it asks for an entire ordering of the various social states (allocations in this case)’, whereas ‘the original question asked only for a “good” allocation; there was no requirement to rank all allocations’ (pp. 64–5). While it is true that ‘the fairness criterion in fact limits itself to answering the original question’, the absence of further rankings may be particularly problematic if no feasible ‘fair’ allocation exists incorporating efficiency (as seems to be the case in many situations). Furthermore, while a ‘pass-fail’ criterion of justice may have attractive simplicity, it does not follow that two states, both passing this criterion, must be seen as being ‘equally just’. Various ‘finer’ aspects of justice have indeed been discussed in the literature (see particularly. Suppes 1966; Kolm 1969; Rawls 1971; Meade 1976; Atkinson 1983).

It should also be noted that the ‘fairness’ literature deals with commodity allocations, or incomes, or some other part of the set of things that figures in Rawls’s characterization of ‘primary goods’. The list is, in fact, much less extensive than that of primary goods as defined by Rawls (1971), and as such it leaves out many considerations that are regarded as important in the Rawlsian framework (e.g., the social bases of self-respect). On the other hand the criticisms – discussed earlier – of the Rawlsian focus on primary goods (based on recognizing inter-individual variations in the ability to convert primary goods into capabilities) would apply a fortiori to the fairness approach as well.

Liberty and Entitlements

A different type of consideration altogether is raised by the place of liberty in a theory of justice. As was mentioned before, Rawls gives it priority. This priority has been questioned by pointing to the possibility that other things (e.g., having enough food) may sometimes be no less important than enjoying liberty without restriction by others. Rawls does, of course, attach importance to these other considerations, but in view of the priority of liberty, they may end up having too little impact on judgements regarding justice in many circumstances, and this might not be acceptable (on this see Hart 1973).

On the other hand, in some other theories of justice, the priority of liberty has been given even greater importance than in the Rawlsian structure. For example, in Nozick’s (1974) theory of ‘entitlements’, rights are given complete priority, and since these rights are characterized quite extensively, it is not clear whether or not much remains to be supported over and above the recognition of rights. Nozick argues against any ‘patterning’ of outcomes, indicating that any outcome that is arrived at on the basis of people’s legitimate exercise of their rights must be acceptable because of the moral force of rights as such. These rights, in Nozick’s analysis, include not only personal liberty, but also ownership rights over property, including the freedom to use its fruits, to use it freely for exchange, and to donate or bequeath it to others (thereby asserting the legitimacy of inherited property).

This type of approach has been criticized partly on grounds of what has been seen as its ‘extremism’, since the constraints imposed by rights can override other important considerations, for example reducing misery and promoting the well-being of the deprived members of the society. In fact, it has been argued that a system of entitlements of the kind specified by Nozick might well co-exist with the emergence and sustaining of widespread starvation and famines, which are often the result of legally sanctioned exercises of property rights rather than of natural calamities (on this see Sen 1981). Although Nozick does refer to the possibility that in case of ‘catastrophic moral horrors’ rights may be compromised, it is not at all clear how his theory would accommodate such waiving of rights, in the absence of formulation of other, competing bases of moral judgements. On the other hand, there cannot be any doubt that Nozick’s theory does capture some notions of justice that can be found in a less clear form in the literature. Nozick’s analysis gives a well-formulated and illuminating account of an entitlement-based approach to justice.

Sources of Difference

To conclude, theories of justice explicitly or implicitly invoked in the literature show a variety of ways in which the demands of justice can be interpreted. There are at least three different bases of variation. One source of variation concerns the metric in terms of which a person’s advantage is to be judged in the context of assessing equity and justice. Various metrics have been considered in this context, including utility (as under utilitarianism and other welfarist theories of justice), primary goods index (as in the Rawlsian theory of justice), capabilities index (as in theories emphasizing what people can actually do or be, e.g., Sen 1985b), incomes or commodity bundles (as in the literature on ‘fairness’, and on statistical measures of poverty, e.g., Foster 1984), various notions of command over commodity bundles and resources (as in some notions of ‘equality’ developed in the literature, e.g., Archibald and Donaldson 1979; Dworkin 1981), and so on.

A second source of difference relates to the aggregating of diverse information regarding the advantages of different individuals. One approach, best represented by utilitarianism, sees nothing being needed to be ascertained other than the sum-total of the overall utilities of different people. Insofar as distributional considerations come into this exercise, they enter in the conversion of goods to be distributed into the appropriate metric of individual utilities. For example, inequality in the distribution of incomes may be disvalued in the approach of utilitarian justice because it may lead to a reduction in the sum-total of individual utilities, through (interpersonally comparable) ‘diminishing marginal utilities’. Other approaches are more concerned with distributional properties related to the different individuals’ relative positions (vis-à-vis each other). The Rawlsian lexicographic maximin is one example of such a distributional concern, and there are others than can be considered, such as adding concave transformations of the individual utility indices (e.g., the additive formula used by Mirrlees 1971, for his taxation assessment), and using various ‘equity’ axioms (e.g., Kolm 1969, 1972; Sen 1973, 1982; Atkinson 1975, 1983; Hammond 1976, 1979; d’Aspremont and Gevers 1977; Roberts 1980).

The third issue concerns the claimed priority of some particular aspect of a person’s advantage (e.g., Rawls’s insistence on the priority of liberty), or nonconsequentialist priority of some processes over results (e.g., Nozick’s 1974, view of rights serving as unrelaxable constraints; or ideas of exploration based on counterfactual exercises of shared rights to social resources, e.g., Roemer 1982).

Given the diversity of moral intuitions related to the complex notion of justice, which has been extensively used over centuries to arrive at normative assessment, it is not surprising that various theories of justice have been proposed in the economic and philosophical literature. The exercise of clearly understanding what the differences between distinct theories of justice consist of (and arise from) is, in some ways, the first task. This essay has been concerned with that task.

See Also