Envy is a deadly sin, but then so is avarice or greed, and greed seems not to trouble economists. Envy does, however, perhaps because it is an externality. Different economists have also used the term in different senses. Veblen (1899) avoids the word ‘envy’, but one feels that some of the pleasure of conspicuous consumption may come from the malicious belief that it induces envy in others. Brennan (1973) uses the term ‘malice’ to indicate negative altruism – a distaste for the income of others – and ‘envy’ to indicate that the marginal disutility of another’s income increases as their income increases. For other concepts of envy, see Nozick (1974) and Chaudhuri (1985).

Most economists now use the word ‘envy’ in a narrow technical sense due to Foley (1967), who was much more interested, however, in finding an adequate concept of ‘equity’. First, however, one should turn to Rawls (1971), whose Theory of Justice has 12 pages of very pertinent discussion.

Rawls on Envy and Justice

Rawls (1971) defines envy in a way which he attributes to Kant, and he is careful to distinguish ‘envy’ from ‘jealousy’, which can be thought of as a protective response to envy:

we may think of envy as the propensity to view with hostility the greater good of others even though their being more fortunate than we are does not detract from our advantages. We envy persons whose situation is superior to ours… and we are willing to deprive them of their greater benefits even if it is necessary to give up something ourselves. When others are aware of our envy, they may become jealous of their better circumstances and anxious to take precautions against the hostile acts to which our envy makes us prone. So understood envy is collectively disadvantageous: the individual who envies another is prepared to do things that make them both worse off, if only the discrepancy between them is sufficiently reduced (p. 532).

This is in section 80, on ‘the problem of envy’ in which Rawls asks whether his theory of justice is likely to prove impractical because ‘just institutions… are likely to arouse and encourage these propensities [such as envy] to such an extent that the social system becomes unworkable and incompatible with human good’ (p. 531). This is a positive question; a normative one arises when one recognizes the possibility of ‘excusable envy’ because ‘sometimes the circumstances evoking envy are so compelling that, given human beings as they are, no one can reasonably be asked to overcome his rancorous feelings’ (p. 534). In the following section 81, on ‘envy and equality’, Rawls argues carefully that his ‘principles of justice are not likely to arouse excusable… envy… to a troublesome extent’ (p. 537). Thereafter, he discusses the conservative contention ‘that the tendency to equality in modern social movements is the expression of envy’ (p. 538), and Freud’s lamentable suggestion that an egalitarian sense of justice is but an adult manifestation of childish feelings of envy and jealousy. Recently, indeed, a particular progressive tax in West Germany has been labelled an ‘envy tax’ (Neidsteuer), as noted by Bös and Tillmann (1985). Anyway, Rawls is careful to distinguish ‘rancorous’ envy from the justifiable feelings of resentment at being treated unjustly. While envy may often form the basis of an appeal to justice, the claims that all appeals to justice rely on envy often fail to distinguish envy from resentment. I shall return to this later.

Equity as Absence of ‘Envy’

More recently, however, ‘envy’ has acquired a precise technical sense in economic theory, following the (apparently independent) lead taken by Feldman and Kirman (1974) and by Varian (1974) in analysing a concept of ‘equity’ due to Foley (1967, p. 75). Apparently Foley was the first to use the term ‘envy’ in this sense, though only informally: Feldman, Kirman and Varian include it in their titles.

Consider any allocation (xig) (g = 1 to n; i = 1 to m) of n goods to each of m individuals. Suppose these individuals have preferences represented by ordinal utility functions Ui(xi) (i = 1 to m) of each individual i’s own (net) consumption vector xi. Then individual i is said to envy j if Ui(xj) > Ui(xi), so that i prefers j’s allocation to his own. Notice that this is a purely technical definition; it tells us nothing about i’s emotional or psychological state, whether i is unhappy because he prefers what j has, or whether i’s ‘envy’ makes him want to harm j. There is no sin in this unemotional economists’ concept of envy, but no particular ethical appeal either. Indeed, it might be better to say that ‘i finds j’s position to be enviable’, to minimize the suggestions of emotion.

Nevertheless, Foley was concerned to introduce a concept of equity of welfare which overcomes the deficiencies of equality of after-tax income – deficiencies which are obvious when there are different public goods in different areas, different preferences for leisure as against consumption, and different needs as well. Thus Foley proposes the absence of ‘envy’ as a test of whether an allocation is equitable. Formally, (xig) is equitable if Ui(xi) ≥ Ui(xj) for all pairs of individuals i and j.

Foley (1967) was careful to qualify this test. First, lifetime consumption plans must be considered so that the prodigal do not envy the higher later consumption enjoyed by the thrifty. Second, as he says:

if a gas station attendant has the desire to be a painter but not the ability, it may be necessary to make the painter’s life very unattractive in other ways before the gas station attendant will prefer his own; so unattractive, perhaps, that the painter will envy the attendant while the attendant is still envying him. These cases must be interpreted flexibly; either equivalents to the talents must be postulated which the gas station attendant does possess, or reasonable alternatives framed that abstract from the glamour and prestige of certain activities (p. 75).

Foley (1967, p. 76) concludes his discussion of ‘equity’ as follows:

If tastes differ greatly, there is very little gained by the analysis since a very wide range of allocations will meet the equity criterion. The definition is offered only as a tentative contribution to a difficult and murky subject and concludes the sketchy discussion this paper will make of welfare economics.

Fairness and Other Extensions of Equity

In a one-good problem of dividing a cake, procedures for achieving ‘fair’ allocations, without envy and with no cake wasted, were discussed in works such as Steinhaus (1948, 1949), and Dubins and Spanier (1961) before Foley. Fairness with many goods was considered by Schmeidler and Vind (1972), by adding Pareto efficiency to the requirement that nobody should envy anybody else’s net trade vector (as opposed to the consumption vector, which includes the endowment). Pazner and Schmeidler (1974) and Varian (1974) then came up with examples of economies with production in which there are no fair allocations, because Pareto efficiency requires skilled workers to supply more hours of labour than unskilled workers, and tastes are such that no allocation of consumption then avoids envy. Feldman and Kirman (1974) considered reducing the degree of envy, whereas Pazner and Schmeidler (1978) weakened the notion of equity to ‘egalitarian equivalence’ − finding an allocation (xig) in which each individual i is indifferent between xi and a consumption in an ‘egalitarian’ allocation (\( {{\overline{x}}^i}_g \)) with \( \overline{x} \)i ndependent of i. Of course, in this egalitarian allocation there is no envy. These later developments all seem like attempts to rescue a dubious notion of equality without giving up first-best Pareto efficiency, even though that is surely unattainable anyway in economies with private information.

Envy and Resentment Reconsidered

In the definition of envy, each individual i compares the consumption vector xj of another with his own, xi, using i’s own utility function Ui. But Ui(xj) > Ui(xi) is insufficient for i’s envy to be excusable, in Rawls’s sense. Indeed, if xj is preferable for i because j has some special needs met, i’s envy is quite unjustifiable. As Sen (1970, ch. 9) points out in his discussion of Suppes’s (1966) grading principles of justice, comparisons of xj and xi must allow for differences in tastes, needs, and so on. Thus the appropriate comparison in determining what is inequitable is rather whether Uj(xj) > Ui(xi). If this is true, we might say that i resents j. Absence of resentment then requires all individuals to have equal utility levels; of course, this requires interpersonal comparisons of utility levels, of the kind used to make decisions ‘in an original position’, behind the ‘veil of ignorance’, before each individual knows his tastes. The technical sense of envy defined earlier differs from this technical notion of resentment precisely because it ignores the original position; not surprisingly, then, envy has no moral force, whereas resentment may well have.

Complete absence of resentment in this sense is probably too strong; but one should look for there to be no resentment in the weaker sense that no individual can legitimately feel treated unjustly by the institutions that determine his welfare. That, of course, reverts to Rawls (1971), though not necessarily to his particular concept of justice.

See Also