Extending the Heuristic Argument: The Importance of the Most Advantaged Class to Implementation of the Difference Principle

One of Rawls's principal justifications for having liberals identify and afford special concern to the least favored class is that the economic expectations of this class serve an important heuristic role in a society's effort to realize the difference principle. Rawls came to recognize that, without further specification, the criterion that everyone benefit from a system of inequality is indeterminate. In Theory, Rawls admits that the notion of “everyone's advantage” is an “ambiguous phrase” (T, 61, 65/53, 57 rev; also see Reference Rawls1967, 134–40; Reference Rawls1968, 154)—and that defining injustice as “simply inequalities that are not to the benefit of all. . .is extremely vague and requires interpretation” (T, 62/54 rev). Rawls's later articulations of the difference principle correct this problem, now specifying that just inequalities, in benefiting everyone, should specifically advantage—and ideally maximally advantage—the expectations of the least favored (Reference Rawls1968, 163; also see Reference Rawls1967, 134–40; Reference Rawls1975, 263; T, passim; PL, 6, 261, 271; R, 42–43, 59, 65, 99–100).Footnote ¹² A related heuristic benefit of specifying the difference principle in terms of ideally maximizing the expectations of the least favored is greater simplicity in application. If the initial conception of the difference principle seemed to require measuring the benefits accruing to each representative person in a society, the expanded definition simply “selects one representative for a special role” (T, 97–98/83 rev).

My claim is that even if this specification of the difference principle provides implementers of social justice with a clearer criterion by which to evaluate and enact just socioeconomic policies, it is not without its own heuristic challenges—challenges that would be productively confronted and reduced by having ideal legislators explicitly integrate the economic expectations of the most favored class into their calculations. To understand both the endurance of heuristic challenges regarding implementation of the difference principle and the usefulness of the category of the most advantaged in helping address them, consider Figure 1, one of Rawls's depictions of the meaning of the difference principle (simplified in this case because it models only two representative positions).

FIGURE 1.

“In this figure the distances along the two axes are measured in terms of an index of primary goods, with the x-axis the more advantaged group (MAG), the y-axis the less advantaged (LAG). The line JJ parallel to the x-axis is the highest equal-justice line touched by the OP curve at its maximum at D. Note that D is the [Pareto] efficient point nearest to equality, represented by the 45-degree line. N is the Nash point, where the product of utilities is maximized (if we assume utilities to be linear in indexes of primary goods) and B is the Bentham point, where the sum of individual utilities is maximized (again with the same assumption). The set of [Pareto] efficient points goes from D to the feudal point F, at which the OP curve becomes vertical. . .Society aims, other things equal, to reach the highest equal-justice line measured by the distance from O along the 45-degree line. To do this it moves as far northeast as possible along the OP curve and stops when this curve bends to the southeast” (R, 62, emphasis added).

In attempting to legislate a just system of inequalities, ideal legislators are under two broad directives. On the one hand, they should seek a system of inequalities that approximates the point at which the least (or lessFootnote ¹³ ) favored maximally benefit (point D). How they do this depends on which part of the OP curve their society occupies. If it occupies a position of less than ideal levels of inequality (anywhere on the OP curve to the left of point D), they can favor a system of greater inequality so long as it raises the expectations of the least advantaged (and assumedly all other representative individuals). But if society occupies a position to the right of point D—where ongoing inequalities have been excessive insofar as they have not benefited, but in fact harmed, the expectations of the least advantaged (a space that Rawls calls the “conflict segment”; R, 124–27)—ideal legislators must reform the basic structure in a way that reduces inequality and lowers the expectations of the more favored. On the other hand, Rawls (T, 78–79/68 rev) is clear that equidistant misses of the ideal point D are not morally equal: a distribution to the left of D is “just throughout, but not the best just arrangement,” whereas one within the conflict segment is “unjust.”Footnote ¹⁴ This means that in attempting to approximate an ideal distribution D, legislators have an obligation to avoid the conflict segment and should prefer a point to the left of D over one to the right.

What Rawls did not perceive—but which I do not believe is in any way hostile to Rawlsian justice—is that the proper functioning of the difference principle, in particular the proper policing of the conflict segment, would be aided if implementers of social justice made appeal to as finely calibrated a notion of the most advantaged (e.g., the superrich) as Rawls suggests they rely on for the least advantaged (e.g., “the unskilled worker” or “the lowest income class”; see T, 94, 307/84, 270).Footnote ¹⁵ Specifically, these implementers should supplement the ideal that the least favored maximally benefit from an ongoing system of inequalities with the ideal that the most favored properly contribute (i.e., the advantages accruing to the most favored class within a system of inequalities should not be excessive in the sense of harming the prospects of the least favored). From the perspective of the most abstracted conception of Rawlsian justice, where it is assumed point D will be achieved, the latter ideal appears superfluous: it is already contained in the former. But from the perspective of ideal legislators, who must work to approximate point D in their specific societies, an explicit concern for the economic expectations of the most favored would serve at least three heuristic goals, all pertaining to the issue of how to hold on to a lively and rigorous conception of the conflict segment over and against various problems threatening to undermine its functionality.

First, insofar as determinations of whether and to what degree a society occupies the conflict segment depend on how finely the representative classes are calibrated, the formal integration of the most favored into the implementation of the difference principle promises to lend precision and greater effectiveness to that principle. To the extent that Rawls presents the difference principle almost always without appeal to the most advantaged—and usually includes only the “more advantaged” that he opposes either to the “less advantaged” (in which case the “more advantaged” is 50% of the population) or to the “least advantaged” (in which case the “more advantaged” is a majority of the population, and a significant one at that)Footnote ¹⁶ —the difference principle would seem to lack the precision and sensitivity to combat any but the grossest instances of maldistribution. That is to say, without a notion of the “most advantaged,” the conflict segment is limited to calculations that a great number of citizens (50%, if not a supermajority) prosper from a system of inequality to the detriment of the least or less advantaged, thereby potentially failing to include situations, which are perhaps more numerous, when a system of inequality enables a small class of citizens (the most advantaged) to prosper with negative social impact.

Second, when the prospects of the most advantaged are explicitly integrated into efforts to implement the difference principle, the importance of the conflict segment is underlined, rather than marginalized. This might seem like a relatively small issue, but in fact the idea of the conflict segment is difficult to keep in view without a representative class—the most advantaged—whose expectations can serve as a point of reference in evaluating justice. Rawls himself, who operates effectively without a notion of the most advantaged, also evinces a tendency to relegate or ignore the conflict segment, even as he also puts forward this crucial notion. Repeatedly, and indeed for the most part, Rawls presents the difference principle in terms of the forward-looking rightward movement from O to D, rather than the retrospective, corrective, leftward movement from P to D. Even if technically speaking the difference principle calls for both movements, Rawls's rhetoric and argumentation often emphasize the former to the exclusion of the latter.Footnote ¹⁷ Given that the difference principle is itself to a certain extent marginalized within Rawls's theory,Footnote ¹⁸ the further marginalization of the conflict segment threatens to effectively neutralize that segment's critical function. Insisting that the most favored always have their expectations analyzed with respect to the evaluation of ongoing or proposed aspects of a society's underlying economic structure serves to maintain a vital, if overlooked and threatened, Rawlsian notion.

Finally, the most advantaged class has a role to play—as the bearer of the “burden of proof” that a system of inequalities is not too excessive compared to the ideal distribution D (i.e., that a society does not occupy the conflict segment)—within the context of uncertainty likely to surround any implementation of the difference principle. Rawls is clear that “there is often a wide variety of reasonable opinion as to whether the principle is satisfied” (Reference Rawls1969, 184; also see 185; R, 48). Specifically, implementers face at least four sources of uncertainty: disagreement about the current diagnosis (who is relatively favored and unfavored at present), how a proposed policy will affect the relevant parties in the future, whether the policy will not only benefit the least favored but also maximize their expectations compared to all other alternatives, and whether the selected policy has in fact achieved its aims. So that the egalitarian element of the difference principle not be negated by the ability of opponents of justice to claim that their own preferred policies best serve the interests of the least favored (see R, 125), it is essential that the most advantaged class—as the best situated representative of the advantaged—bear the burden of proof within a context of uncertainty. In saying that the most advantaged are “best situated,” I mean three things: that the most advantaged have the most resources to withstand the costs of a less-than-perfect application of social justice; that their expectations most directly affect absolute levels of inequality; and that, for heuristic reasons, under conditions of uncertainty, their expectations can serve as a proxy of the advantaged more generally.Footnote ¹⁹

That is to say, when in doubt, legislators seeking to implement the difference principle should prefer to impose too many rather than too few economic costs on the most advantaged class, so as to veer toward too little rather than too much inequality. It might seem that, in attending to uncertainty in this way, I am disagreeing with Rawls, who argues (T, 201/176 rev) that citizens need to accept that on many questions of social and economic policy it is simply not clear what the difference principle requires—and that in such cases of indeterminacy, “we must fall back upon a notion of quasi-pure procedural justice,” according to which the justice of a distributive scheme should be seen as inhering not in its substance (whether a system of inequalities is indeed mutually beneficial) but in its being enacted by a legitimate legislature. But even if such an approach requires that ordinary citizens relent from rigorism regarding the difference principle—because they must acknowledge that “the appeal to justice is not sufficiently clear and its resolution is best left to the political process” (Reference Rawls1969, 184)—it still demands that legislators, within the political process, have some heuristic to guide their decision making in such contexts. Importantly, Rawls provides an account of what this heuristic should be. Recall that legislators committed to Rawlsian justice are under two moral directives: they should seek a system of inequalities that ideally maximizes the prospects of the least favored, but additionally, in the event that their policies miss the ideal distribution (D)—something virtually guaranteed under conditions of uncertainty—they are to prefer less than ideal rather than more than ideal levels of inequality (T, 78–79/68 rev). Taken together these two directives would appear to necessitate as a matter of effective legislative practice placing the burden of proof on the more—and in particular, as I am arguing here, the most—advantaged class. Legislators should identify the most advantaged as the class whose expectations are subject to special scrutiny in ensuring that a system of inequalities is not within the conflict segment and that therefore bears the risks of uncertainties in this regard.

When one considers the strong likelihood that the underappreciated conflict segment is precisely that part of the OP curve occupied by contemporary liberal-democratic societies,Footnote ²⁰ any marginalization of that segment within a theory of justice is surely regrettable for liberals committed to a notion of distributive justice that is more rigorous than the old, Lockean standard that the poor merely benefit at all from a system of inequality. An advantage of a direct and explicit concern for the most advantaged is that it counteracts such marginalizing forces.

Extending the Protective Argument: The Threat to Liberty from the Most Advantaged

A second way in which those at the economic bottom of society enter into Rawls's theory of justice involves Rawls's argument that some minimum amount of income and wealth is a precondition for the effective maintenance of the basic rights and liberties guaranteed in his first principle of justice (rights such as the freedom of speech, religion, person, assembly, and so on). Before obtaining whatever additional gains the least favored receive through the difference principle, all citizens according to Rawls are to be guaranteed a “social minimum” as a condition of their basic liberties, on the assumption that there is a threshold level of resources below which other liberties lose their meaning (PL, 228; R, 48, 162).

My claim—one that Rawls shares, even if its importance for singling out the most advantaged is something he does not strongly emphasize—is that the norm of protection justifies not only a floor on income and wealth but also a ceiling. Like insufficient resources, too much inequality can also undermine the value of certain basic liberties. Although Rawls's difference principle is explicitly opposed to making judgments about particular levels of inequality being intrinsically just or unjust (because it locates the justice of a system of inequalities in distributions that are mutually beneficial and that ideally maximize the prospects of the least favored, regardless of how equal or unequal these distributions turn out to be), the difference principle is not the only distributive principle within Rawls's theory of justice. There are two other principles, and even if they do not look to allocate specific shares in a predetermined manner, they are very much concerned about relative distributions and the problem of excessive inequalities. In an important footnote to Justice as Fairness: A Restatement, Rawls clarifies this point:

As Rawls makes clear here, beyond the difference principle there are two other crucial distributive principles within his theory: the fair value of political liberties (which “ensures that citizens similarly gifted and motivated have roughly an equal chance of influencing the government's policy and of attaining positions of authority irrespective of their economic and social class”; R, 46; also see 149) and the fair equality of opportunity (i.e., “those who have the same level of talent and ability and the same willingness to use these gifts should have the same prospects of success regardless of their social class of origin, the class into which they are born and develop until the age of reason”; R, 44). What is highly significant about these two principles is that Rawls understands them as being threatened by too much inequality—or what he terms “excessive accumulations” and “excessive concentrations” of wealth and income (Reference Rawls1958, 52–53; 1961, 161; Reference Rawls1963b, 77, 89; Reference Rawls1967, 140, 141; T, 73/63 rev; R, 44, 161).Footnote ²¹

Rawls's recognition of the problem of excessive inequality places him within a long tradition of republican thought concerned about the capacity of the wealthy to corrupt the political process and abuse the less fortunate. On the assumption that the balance of property inevitably shapes the balance of power, numerous republican thinkers historically argued for curbing wealth inequality both to protect ordinary citizens from harm threatened by the very rich (Aristotle Reference Aristotle and Rackham1932, 1295b4–15; Jefferson Reference Jefferson and Petersen1984, 841; More Reference More, Surtz and Hexter1965, 67) and to make political power more likely to be placed in the hands of the meritorious rather than the merely wealthy (Harrington Reference Harrington1977, 202, 231, 460; Montesquieu Reference Montesquieu, Cohler, Miller and Stone1989, 44–45, 94–96; Plato Reference Plato and Hamilton1989, 584d, 737e). Although underemphasized by present-day liberals, market utopians, and even many contemporary republican thinkers (see McCormick Reference McCormick2006, 163), the danger posed to liberty by the superrich is something Rawls does acknowledge. Yet, whereas this older republican tradition worried primarily about the threat of actual usurpation of liberties by the wealthy, with Rawls the concern is more that the raw discrepancy in wealth can make certain liberties come to lose their value. Although Rawls does not extensively elaborate the nature of the “political domination” (R, 44) engendered by excessive inequalities, Freeman suggests that in the case of fair equality of opportunity, “there comes a point at which the degree of inequality is so great that, even if it satisfies the difference principle, . . .it nonetheless causes [the least favored] to feel diminished and less than civic equals of those who are more advantaged, leads them to see themselves as failures, and therefore neglect taking advantage of opportunities to educate and develop their capacities.” In the case of the fair value of political liberties, Freeman speculates that Rawls has in mind situations in which “because of the enormous inequalities of opportunities and unequal political influence that the capitalist welfare state allows, the less advantaged tend to withdraw from political and civic participation, seeing it as pointless, and suffer therewith a loss of their self-respect” (Freeman Reference Freeman2007b, 131, 132–33; see T, 277–78/245–46 rev; R, 51, 159).

Rawls does not define when accumulations of property become excessive, instead referring the question to political bodies, such as legislatures (Reference Rawls1967, 143).Footnote ²² But this is precisely why implementers of justice, who follow Rawls in understanding that excessive inequalities can undermine the capacity of free and equal citizens to enjoy civic equality with respect to education and politics, need to police the upper bounds of the distribution of income and wealth and, in so doing, operate with a lively sense of the most advantaged class.

Extending the Redressive Argument: The Shadow of Unfairness within Well-Ordered Societies and the Need for Redress

A third reason Rawls provides for singling out the least advantaged concerns redress: because the least favored on average have suffered the most from arbitrary disadvantages of the natural and social lotteries (Rawls Reference Rawls1975, 258–59), they have a claim to be singled out by the difference principle as the representative class whose income and wealth (along with other primary goods) ideally are to be maximized in the implementation of a just system of inequalities. In justifying why the least favored in particular deserve to play a special heuristic role in the evaluation and implementation of justice, Rawls appeals to a redressive logic of compensation. Note that this form of compensation is distinct from the kind that is supposed to occur among reciprocally engaged free and equal citizens, whose differences in wealth and income are to be limited by the constraint that the less favored benefit—or are compensated—by any system of inequalities (T, 102, 536/87, 470 rev). Rather, the compensation I mean to highlight, and the one that links Rawls's philosophy to a theory of redress, is the notion that, when determining whether all citizens benefit from a system of inequalities, it is the expectations of the least advantaged that are to be privileged in the analysis because it is they who are most unfavored by the morally arbitrary distributions of the natural and social lotteries and by the imperfect capacity of a liberal society to partially correct for them even in the sense of fully establishing fair equality of opportunity.Footnote ²³ It is this latter kind of compensation that Rawls has in mind when he acknowledges that “it may appear that the difference principle is arbitrarily biased toward the least favored.” Rawls's response is that “those favored by social and natural contingencies [should] regard themselves as already compensated, as it were, by advantages to which no one (including themselves) had a prior claim.” Those who are not favored, the least advantaged, are entitled to their own kind of compensation in the form of a difference principle that privileges their interests when administering and evaluating justice (Reference Rawls1975, 263–64). Similarly, in one key passage from Theory (73–74/63–64 rev; also see R, 124), Rawls argues that the least advantaged do not morally deserve their (on average) inferior placement in the natural lottery, nor is the principle of fair equality of opportunity fully realizable, and for these reasons they are entitled to be singled out by the difference principle.Footnote ²⁴

To be sure, in drawing attention to the redressive elements of Rawls's justification for singling out the least favored, I do not mean to attribute to him the luck-egalitarian position that all undeserved inequalities are to be compensated. Clearly Rawls is not a luck egalitarian. He differentiates his theory from a wholesale philosophy of redress.Footnote ²⁵ In addition, he provides at least three reasons for allowing inequalities, even morally undeserved ones, to persist so long as they are part of a mutually beneficial social structure: the rationality of choosing absolute gains for all over the mere reduction of undeserved inequality (and, relatedly, the need for incentives for unequal rewards to motivate productive gains and the efficient allocation of resources) (T, 151/131 rev; R, 55, 203); the practical difficulty of enforcing a luck-egalitarian scheme (T, 87–88/76 rev; R, 68); and the entitlement (as opposed to moral desert) of individuals to be remunerated for their productivity (including individuals whose productivity is a function of morally arbitrary advantages; Rawls Reference Rawls1958, 48; Reference Rawls1967, 152; Reference Rawls1968, 170).

Still, if Rawls is opposed to efforts to neutralize all luck in an economic distribution, as I have indicated he is not altogether against redress, because his very particularization of the least favored as the class whose prospects are to be maximized in the implementation of the difference principle rests in part on redressive grounds. The question to ask of Rawls, then, is not whether luck should be mitigated at all, but how much mitigation is consistent with his principles of justice. The claim I want to make is that there are Rawlsian grounds for extending redress to include the most advantaged, so that implementers of justice would aim to mitigate not only the arbitrary disadvantages of the least favored but also the arbitrary advantages of the most favored. Specifically, my argument is that the two distributive principles beyond the difference principle that I discussed in the previous section—the fair value of political liberties and fair equality of opportunity—are important not only because they require for their protection policies that counteract excessive accumulations of property but also because they are principles, as Rawls acknowledges (at least in the case of fair equality of opportunity), that cannot be fully fulfilled. This circumstance means that even a so-called well-ordered society of the type Rawls hoped to see emerge would never be free of a kind of residual unfairness—or what I term a shadow of unfairness. Such unfairness legitimates a claim of redress against at least the most favored class of society—the class prospering the most within an imperfectly fair system and thus, through costs placed on it, the best situated to acknowledge (and also remedy, however partially) the persistence of unfairness in an otherwise well-designed liberal regime. Rawls does not make the point, but I take it as a justifiable extension of his philosophy.Footnote ²⁶

Consider the principle of fair equality of opportunity, the lexically prior portion of Rawls's second principle of justice. It is here where Rawls comes closest to explicitly recognizing the shadow of unfairness. The principle that all similarly talented and motivated children should have the same prospect of success regardless of their social class of origin will never be fully realized, Rawls admits, so long as there are institutions like the family:

As Rawls makes clear, he is aware of the problem of implementing fair equality of opportunity and invokes this circumstance as a justification for the difference principle: the difference principle, which privileges the least advantaged, “recognizes this fact” and mitigates it to some degree. Yet, even if Rawls appeals to shortcomings in the implementation of fair equality of opportunity as a justification for the difference principle, there is no escaping the fact that these shortcomings also disturb the difference principle itself. Although Rawls's argument in the previous passage is to appeal to the imperfection of fair equality of opportunity to justify the difference principle, he also repeatedly claims that the difference principle's refusal to understand particular distributions as being intrinsically just or unjust only works insofar as there is fair equality of opportunity:

Rawls may be correct in his claim that the difference principle reduces some of the residual unfairness that accrues from a world without full fair equality of opportunity, but he does not pursue the further implications of this phenomenon: namely, that without a more complete fair equality of opportunity, the difference principle's indifference toward wealth concentrations (so long as they are mutually beneficial) is no longer a fully fair mechanism of distribution. A liberal society's willingness to tolerate vast inequalities—to let “distributive justice. . .take care of itself”—erodes insofar as similarly talented and motivated individuals cannot expect similar opportunities and rewards. The family is clearly one key cause of this disequilibrium of expectations; relatedly, much social scientific research demonstrates how the class structure into which a child is born is a strong predictive factor of the economic and social opportunities experienced as an adult (Bowles, Gintis, and Groves Reference Bowles, Gintis and Groves2005; Sacks Reference Sacks2007).

This does not mean that a liberal society cannot make great strides in correcting how much the social contingencies of one's birth determine life expectations, but it does mean that institutions such as the family create limits to just how much fair equality of opportunity can be achieved. This problem has been addressed in different ways. Some take issue with the family itself on liberal grounds; however, this proposal seems undesirable (insofar as the family ought to be considered a “concession to human nature” as Rawls might have put it) and, in any case, potentially invalidated by the prior freedom to associate. Libertarian critics such as Nozick have argued that the inability to fully realize fair equality of opportunity demonstrates its inappropriateness as a norm of justice, but such a critique conflates unenforceability with invalidity (Nozick Reference Nozick1974, 167). Finally, certain defenders of Rawls, such as Freeman (Reference Freeman2007b, 98), have argued that the meaning of fair equality of opportunity is in fact more limited than meritocracy: that “Rawls does not understand FEO as requiring equal chances for the equally endowed,” but only the more minimal norm that “people, whatever their natural abilities and social circumstances, be given the means to fully develop and effectively exercise the talents and abilities that they are endowed with, so that they may engage in public life as equal citizens, and have a fair opportunity to compete with others of similar abilities for positions within the range of their developed skills.” The problem with this apologist position, however, is that it overlooks instances where Rawls explicitly calls for equal opportunities for the equally endowed as a condition of a just society,Footnote ²⁷ it appeals to a problematic teleological notion of “full development,” and even the more minimal standard it upholds (those with similar developed skills have fair opportunity to compete with others of similar abilities) is still clearly undermined, at least at the margins, by social class and the influence of the family.

Notwithstanding the substantial educational resources a liberal society will devote to all of its citizens, and especially perhaps to the least advantaged, it is nonetheless clear that, given two equally talented and motivated children, the one born into a privileged background will have greater prospects not only of wealth but also of culture, education, and “success” relative to a child born into an underprivileged background. This circumstance occurs not just in the United States but even within the world's most egalitarian educational systems (Blom, Melin, and Tanskanen Reference Blom, Melin and Tanskanen2009; Edlund and Svallfors Reference Edlund and Svallfors2009; Martin and Wajcman Reference Martin, Wajcman, Devine and Waters2004). Even if the impact of socioeconomic status on children's life prospects operates on the margins, it is unfair and it helps provide justification for a liberal society committed to free and equal citizenship to single out and seek redress from the most advantaged, the group that has prospered the most under an imperfectly fair system and is therefore best able to publically acknowledge and partially remediate such unfairness.

A similar situation obtains for the fair value of political liberties—even if Rawls does not make an argument in this regard. Although a liberal society can always try to approximate a situation in which all citizens, regardless of economic and social background, have an equal chance to hold office and influence election outcomes, it is difficult to point to a polity in human history where the very rich did not have superior political opportunities. Rawls, it is true, does not confront this circumstance as a permanent feature of political life. He denies that the scarcity of political space (the existence of a few high offices in a polity of millions) is itself a mark against equal political liberty (Rawls Reference Rawls1958, 50; Reference Rawls1963b, 75; Reference Rawls1971b, 195); and he thinks that various liberal reforms, such as campaign finance legislation, can make it so that the equally endowed and motivated “have roughly an equal chance of influencing the government's policy and of attaining positions of authority irrespective of their economic class” (R, 46). Although this ideal is noble and its approximation certainly worth striving for, the shadow of unfairness and, with it, the need for redress arise from the inescapable confluence of economic and political power that has made the economically well-off, on average, more likely to exert political influence than the disadvantaged.Footnote ²⁸ In the United States, this confluence is suggested by an abundance of social scientific literature pointing to the likelihood that politics is significantly shaped by inequality: Those with great wealth have more access to government, policy outcomes have been shown to be skewed to the interest of the wealthy, those who win office are on average far wealthier than ordinary citizens, and those who leave office have far greater opportunities for enriching themselves after their term is finished (Bartels Reference Bartels2008; Gilens Reference Gilens2005; Jacobs and Skocpol Reference Jacobs and Skocpol2005; Winters and Page Reference Winters and Page2009). But even within relatively more egalitarian societies, such as European and especially Nordic democracies, there is clear evidence of the disproportionate influence of the wealthy within politics.Footnote ²⁹

Importantly, extending this logic of mitigation to also include the particularization of the most advantaged as a distinct class deserving special regulatory attention is fully compatible with the three constraints on redress Rawls puts forward. First, because redress is constrained by the notion that all are supposed to benefit, there is no question of pursuing redress to a point that obviously leads to lower expectations for the rest of society.Footnote ³⁰ Second, although it is true that the extension of redress to the most advantaged would require an additional set of applications (focusing on the most advantaged and not just the least advantaged), it would not result in the determination of deservedness and specific allocative shares for an entire population, something Rawls finds, rightly I believe, highly impractical. Finally, it is not correct to say that the extension of redress disrespects the entitlement of the most advantaged class to benefit from the productive use of their labor and capital, because any redress would only regulate the highest levels of income and wealth, ensuring that those to whom it applied would retain their status as the wealthiest individuals (and retaining, too, their entitlement to remuneration for productive gains).

Clearly, the failure of a society to fully realize fair equality of opportunity and the fair value of political liberties is not as severe an injustice as despotism or other blatant violations of basic liberties. Nonetheless these two principles are fundamental not only to Rawls but also to multiple strands of liberalism more generally and to the political attitudes of ordinary citizens.Footnote ³¹ By using the terminology shadow of unfairness, I mean to refer to a level of residual unfairness that is not likely enough to destabilize a regime through generating disobedience and unrest, but nonetheless is real and important and thus sufficient to require some compensation from the class that most benefits from the inability of a liberal society to adequately escape the feudal confines of blood and lineage.Footnote ³²

Confronting the Accusation of Envy: In Defense of Reasonable Envy toward the Superrich

I have argued on three different grounds—heuristic, protective, and redressive—not simply that it is consistent with a Rawlsian conception of justice to single out and potentially regulate the upper bounds of the economic expectations of the most advantaged, but that as part of this process there are certain cases when it is legitimate to impose economic burdens on the most favored that have either a neutral or negative material impact on the rest of society. The protective argument, which aims to prevent inequalities that are so excessive that they threaten civic liberty, justifies regulation of the most advantaged even when the rest of society would be worse off in economic terms. The redressive argument, as I have explained, legitimates the imposition of economic burdens on the most advantaged that would have a neutral material impact on other representative members of a political community. And although the heuristic argument does not sanction lowering the prospects of the most advantaged class when doing so clearly would have a negative or neutral impact on the rest of society, its support of having the most advantaged bear the burden of proof that a society's system of inequalities is outside the prohibited conflict segment means that it legitimates economic bias against the most advantaged under conditions of uncertainty.Footnote ³³

Critics will object that in its defense of potentially subjecting the most advantaged to costs without compensating economic benefit for the rest of society, my argument is not simply guilty of envy—the disconcertion at the greater wealth of others and the seemingly irrational desire for those with more to have less, even if it results in neutral or negative economic consequences for everyone else—but, in its enviousness, guilty of something widely opposed within the broader political culture and, in particular, by Rawls himself.Footnote ³⁴ Rawls, after all, explicitly invokes the anti-envying mentality of “mutual disinterest” and defends it, arguing that citizens must not be expected to have other-regarding sentiments—whether altruism or envy—when determining the content of social justice behind the veil of ignorance (T, 128, 148, 253–54/111, 128, 223–24 rev). Rawls says citizens, so engaged, who are dismayed by another's economic success—if such success emerges from an underlying political and economic system that maximizes their own expectations—are “shortsighted” (T, 151)Footnote ³⁵ because they are in effect preferring to have less simply so that others have less too. Rawls's apparent view about the irrationality of envy is by no means idiosyncratic, but reflective of a general opposition to envy within the broader political culture. Libertarians and other critics of redistributionist schemes, for example, have long suggested that such schemes are rooted in envy and for this reason are invalid (Cooper Reference Cooper1982; also see Freud Reference Freud and Strachey1949; Nietzsche Reference Nietzsche1998; Nozick Reference Nozick1974). Many liberal egalitarians resist this charge, but not the assumption that were liberal egalitarian principles rooted in envy, this circumstance would be a check against those principles (Norman Reference Norman2002; Young Reference Young1989).Footnote ³⁶ Major contributions to liberal social philosophy have taken aim against envy, castigating it as a destructive, antisocial psychology (Mill Reference Mill and Gray1998, 10, 87, 250; Tocqueville Reference Tocqueville2000, 189, 479–82). The philosophical objection to envy from prominent liberal philosophers is paralleled by long-standing Christian objection to envy as one of the cardinal sins (see Epstein Reference Epstein2003, xi, 87; Fernández de la Mora and De Nicolas Reference Fernández de la Mora and De Nicolas1987, 19–42) and by the attitudes of ordinary citizens, especially Americans, whose opposition to envy—and in particular to capping the upper expectations of the most well-off citizens—has been documented in repeated studies for decades (Hochschild Reference Hochschild1986; Lane Reference Lane1986; McCall and Kenworthy Reference McCall and Kenworthy2009; Page and Jacobs Reference Page and Jacobs2009). This opposition to envy is political and not simply ethical, insofar as it has contributed to recent efforts in numerous nations to remove or reduce the estate tax, employ tax brackets without a category for the superrich, and increasingly object to progressive taxation as such (Graetz and Shapiro Reference Graetz and Shapiro2006; also see McCall and Kenworthy Reference McCall and Kenworthy2009).

Notwithstanding these considerations, however, such a one-sided account of Rawls's theory of envy is incomplete, unbalanced, and thus mistaken. Rawls's endorsement of mutual disinterest and his parallel critique of envy are not absolute: he makes an important distinction, often unobserved in the secondary literature, between ordinary, irrational, inexcusable envy and another form that he describes as “not irrational”—an envy “where it would be unreasonable to expect someone to feel differently” (T, 534/468 rev; PL, 284; R, 88). The crucial claim I want to defend is both that this latter form of non-irrational envy has an important role to play in the implementation of Rawlsian justice and that the kinds of envy implicated in my defense of extending liberal regulatory attention to include the economic expectations of the most advantaged fall within this category of non-irrational envy.

Rawls's delineation of the concept of non-irrational envy has not received sufficient attention. Indeed, numerous recent commentators have overlooked the notion altogether (e.g., Lehning Reference Lehning2009; Pogge Reference Pogge2007; Taylor Reference Taylor2011; Voice Reference Voice2011). In part this lack of attention stems from the fact that Rawls himself provides only a brief sketch of the concept. But it also arises because the usual treatment of the notion among scholars who have examined it is to marginalize non-irrational envy as having to do only with Rawls's arguments about the stability of his principles of justice (their capacity to generate the sources of their own support, in part by not producing rancorous sentiments, such as envy, in the citizenry) and not also about the implementation of the principles themselves (Freeman Reference Freeman2007b, 269; Mandle Reference Mandle2009, 130–32; also see Maffettone Reference Maffettone2011, 111).Footnote ³⁷ In other words, the usual reading of non-irrational envy is that Rawls appeals to it only as an excusable rancorous sentiment that he thinks will be more curtailed in societies governed by his proposed principles of liberal justice, not that it has any positive role to play in the effective instantiation of those principles as the willingness of implementers of justice sometimes to impose costs on the advantaged without compensating economic benefit for the rest of society (which might be called reasonable envy).Footnote ³⁸

Although I find it incorrect, the reduction of non-irrational envy to excusable envy is understandable, if only because Rawls's most prolonged engagement with the concept of non-irrational envy confronts the notion in precisely this sense (T, 531–32, 534–41/466, 468–74 rev; R, 88, 180, 202). In these passages, Rawls argues that, although normally it is irrational for individuals to want those with more to have less when doing so will have a negative impact on their own material welfare, there are circumstances when inequalities are so great that they harm the self-respect of ordinary citizens and thereby undo the irrationality normally attached to envious sentiments:

Rawls thinks it is an advantage of his proposed principles of justice that they are less likely than alternative schemes, especially utilitarianism, to generate such excusable rancorous sentiments (T, 537/469–70 rev).

But this usual treatment of non-irrational envy is overly narrow on two grounds: It neglects the way in which envy has a role to play in the implementation of justice (and is therefore not just excusable but reasonable), and it forgets that beyond the problem of excessive inequalities there are at least two additional situations in which the normal irrationality of envy is abrogated. Both aspects of this richer, more expansive conception of envy are suggested in the following passage:

The first thing to note about this passage is that Rawls acknowledges there are three instances when envy is not irrational: not just the problem of excessive inequalities but also inequalities that stem from an unjust basic structure (in which either basic liberties or fair equality of opportunity has not been fairly secured) and inequalities that do not satisfy the difference principle (but involve an economic system in which income and wealth accrue to winners of the natural and social lotteries without less fortunate members of society also benefiting from the inequalities being generated; see T, 530/464 rev).

The second crucial thing to note about this passage is that it comes not from the final Part III of Theory of Justice where Rawls engages with the issue of stability, but from Part I, “Theory” (specifically, chapter 3, “The Original Position,” section 25: “The Rationality of the Parties”), where Rawls discusses the general conception of rationality presumed to govern the adoption and implementation of the two principles of justice (T, 143/124 rev).Footnote ³⁹ The suggestion is that the three forms of non-irrational envy Rawls mentions arise in two separate contexts: not just in relation to the matter of stability (where theories of justice are evaluated with regard to how much excusable envy they generateFootnote ⁴⁰ ) but also in relation to the Original Position (where the three forms of non-irrational envy become contexts in which it is reasonable for implementers of justice to impose economic burdens on the advantaged with negative or neutral benefit to the rest of society.)

To be sure, I do not mean to say that the initial adoption of the two principles of justice is guided by envy for Rawls. In this regard, Rawls is clear that envy plays no role.Footnote ⁴¹ But, as Rawls explains, the Original Position can be disaggregated so as to refer to an initial adoption stage, where the two principles are selected, and to three successive stages—the constitutional convention (where the first principle, with its idea of equal basic liberties, is constitutionalized), the legislative stage (where legislators select the social and economic policies required by justice), and the judicial stage (where particular cases are adjudicated)—all of which have to do not with the selection of the principles, but their implementation (T, 195–201/171–76 rev). As the veil of ignorance is gradually lifted at each successive stage, each stage still models an ideal form of rationality pertaining to the implementation of justice.Footnote ⁴² If envy does have a potential role to play in the implementation of justice, as Rawls suggests, it would be at the legislative stage, where legislators ideally seek the passage of laws and policies that realize the social and economic conditions required by a just society—and do so with information about the nature of the distributions within their society, with “the full range of general economic and social facts. . .brought to bear” (T, 199/175 rev).

My claim is that any envy relating to the protective, heuristic, or redressive justifications for imposing regulatory burdens on the most favored can be seen as falling, respectively, within the three categories of reasonable envy that Rawls mentions: excessive inequalities, inequalities not bound by the difference principle, and inequalities founded on injustice.Footnote ⁴³ The first of these assertions—that to the extent the protective grounds for capping the economic expectations of the most advantaged involve elements of envy because such capping might very well lower everyone's material welfare, such envy is reasonable if it is part of the legislative effort to combat excessive inequalities—is straightforward and something Rawls clearly affirms. It is not simply that Rawls (Reference Rawls1967, 143) expects legislators to enact policies to counteract excessive accumulations of property. What also is key is that Rawls says explicitly that legislators so engaged are acting on, or at least responding to, a non-irrational form of envy. In Political Liberalism, Rawls makes it clear that non-irrational envy is a constraint which legitimates laws and policies that limit the upper bounds of wealth:

By treating envy here as a constraint, Rawls specifies that it functions both in terms of stability and in terms of implementation—or that stability itself is more than a criterion for passively evaluating rival regimes, but is a basis for implementers of justice to intervene in the economic structure of a just society. This passage is also important because it links the need for legislators to limit inequalities in the name of protecting the fair value of political liberties and fair equality of opportunity with the need for them to act on (or respond to) non-irrational envy—further supporting the notion that any envy pertaining to the protective grounds for regulating the most advantaged falls within one of Rawls's three categories of non-irrational envy: envy that responds to inequalities that “exceed certain limits” (T, 143/124 rev).

Second, the envious elements of the heuristic grounds for regulating the most advantaged should be seen as falling within the exception to the general irrationality of envy that arises when it is believed that “existing inequalities. . .are the result of letting chance work itself out for no compensating social purpose” (T, 143/124 rev). Given the uncertainty surrounding the successful implementation of the difference principle, the concern that a system of inequalities is too extreme, such that it passes into the conflict segment where it no longer serves compensating social purpose, will be an ongoing problem for legislators, even (and especially) ideally conceived ones. The context of uncertainty means that legislators committed to justice cannot reliably achieve a precise point on the OP curve (see Figure 1), but only a range. Their insistence that the most advantaged bear the risks of this uncertainty, and with it their preference for less than ideal rather than more than ideal levels of inequality, means that over time they will produce policies that tend to the left of the ideal distribution D rather than to the right: it means that they will engineer a society where all representative classes tend to have less than the ideal as opposed to one where only some do. Compare this to an alternate scenario where legislators, under no compulsion to treat uncertainty in a biased way, simply could do their best to reach the ideal distribution: over time they would produce a range of systems of inequalities that straddled a space equidistant from both sides of the ideal distribution D. What justifies the former scenario over the latter one—and, with it, a modicum of envy (because the former scenario is in effect the decision for all to have less income and wealth over time than they would under the latter scenario)—is that the two sides of the curve are not morally equal. Because the space to the right of D is one of injustice, but the space to the left is merely less than perfect justice (T, 78–79/68 rev), a society has a moral reason to prefer this otherwise irrational outcome. Since legislators cannot escape the possibility that their polices will lead society into the conflict segment, it is reasonable for them to operate on the basis of a heuristic device that, in privileging too much over too little equality, over time will have the real, if probably limited, envious effect of making all members of society materially worse off.

Finally, to the extent there is an element of envy in the redressive grounds for regulating the most advantaged, which as I have discussed justifies placing economic burdens on the most advantaged even when doing so has a neutral material impact on the rest of society, such envy finds legitimation in Rawls's third exception to the irrationality of envy: the circumstance that, within a society, “existing inequalities are founded on injustice” (T, 143/124 rev). Even if, as I have indicated, the injustice in question is less one of gross and correctible injustice than the residual unfairness stemming from even a well-ordered society's inability to fully realize fair equality of opportunity with regard to education and politics (see note 27), such unfairness—especially when it appears within a theory of justice as fairness—is a basis for imposing redressive burdens on the most advantaged that have no economic benefit for the rest of society.

I do not claim that the protective, heuristic, and redressive bases for regulating the most advantaged exhaust the potential meaning of non-irrational envy for Rawls, but only that to the extent envy is involved in my three arguments for extending Rawlsian justice to include a concern for the most advantaged as lively and explicit as normally evinced for the least favored, such envy falls within the category of reasonable envy.Footnote ⁴⁴ Nevertheless, that there is even a category of reasonable envy—that the imposition of economic costs on the advantaged without economic benefit for society has a role to play in the implementation of social justice—ought to be a corrective to the widespread denigration of the notion within contemporary liberal philosophy and the broader political culture. It is worthy recalling that the root meaning of “envy” comes from the word to “see”—so that envy literally means a “hostile look” (Schoeck Reference Schoeck1969, 197)—which reminds us that prior to any regulation and imposition of burdens, envy most basically involves a suspicion toward a particular person or class and, so, an identification of them.Footnote ⁴⁵ My arguments for regulating the most advantaged class are at the most fundamental level arguments for identifying this class—a stipulation that might seem meager at first, but is actually quite substantial in light of the surprising absence of the concept within major discourses of contemporary liberal thought, not to mention tax codes that define the upper bracket so low that the superrich remain hidden from view. Given the current context where leading philosophies, civic attitudes, and political practices all align with a refusal to seek out the most favored members of society, such an identification would itself be a major, if initial, accomplishment of a liberal society's willingness to evolve its understanding of justice to more directly reflect, and indeed extend and further particularize, Rawls's (R, 125) insight: “In the well-ordered society of justice as fairness it seems that those most likely to be discontent are the more advantaged.”