How much redistribution should there be?

• Author: Markovits, Daniel

1. THE EGALITARIAN IDEAL

   Egalitarianism ties people's fortunes together. It takes the good and bad things in people's lives--their blessings and their afflictions--and shares them out, or redistributes them, among their fellows. Where egalitarianism operates, each person's fortunes and misfortunes cease to be just her own and become, to the extent egalitarianism recommends, a part of communal fortunes and misfortunes, shared in by all those who come under egalitarianism's purview. Egalitarian political theories are not the only ones to tie people's fortunes together; feudal theories, for example, do so as well. But egalitarianism differs from these and other political theories in the manner of its tying. Egalitarianism insists that all people's lives are equally important and, accordingly, that no person's fortune may be subordinated to anyone else's. (1)

   Egalitarian intuitions call for redistribution that takes from the better-off--the rich, the healthy, and the fulfilled--and gives to the worse-off-the poor, the sick, and the desolate. But these intuitions stand in need of elaboration, and an egalitarian theory answers this need by presenting an articulate account of nonsubordination among persons. (2)

   A fully developed conception of nonsubordination must address many practical or applied problems, including many problems that are familiar in the law, and egalitarian political theory has indeed been brought to bear on such problems in legal scholarship. Thus lawyers have, for example, considered egalitarianism's implications for the choice of the tax base, (3) the legal treatment of disabilities, (4) health-care policy, (5) and even the role that cost-benefit analysis should play in administrative practice. (6) But before these and other more specific issues can be approached with any measure of confidence or clarity, an egalitarian theory must address two more general, theoretical questions, and efforts to answer these questions must proceed in a more broadly philosophical vein.

   The first question was made famous by Amartya Sen in the title of his 1979 Tanner Lecture Equality of What? (7) This question asks, in G.A. Cohen's words, "What aspect(s) of a person's condition should count in a fundamental way for egalitarians...?" (8) It suggests that egalitarians search, as Cohen put it, for an equalisandum, a thing whose equal distribution secures nonsubordination. (9) Here it is critical that egalitarianism calls for the nonsubordination of persons, and the equalisandum must be chosen bearing this in mind. The proposal, for example, that everyone should be given an equally long name fails as a conception of egalitarianism because the length of someone's name does not capture what is important about her person, so that equality among name lengths is entirely consistent with subordination among persons. (10)

   An egalitarian theory must also answer a second question, the question posed in my title, namely "How Much Redistribution Should There Be?" (11) To do this, an egalitarian theory must develop a conception of nonsubordination that explains how the equalisandum, whatever it is, must be distributed in order for no person's fortune to be subordinated to any other's. Put slightly differently, an egalitarian theory must develop a view of what counts as an equal distribution of the equalisandum. Because the equal distribution will almost certainly not be the distribution produced in the ordinary course of economic and social activity, this means that an egalitarian theory must develop an account of redistribution.

   This second question (the question concerning how much redistribution there should be) has been less intensively investigated than the first question (concerning the proper equalisandum). However, the second question is not any less central to egalitarianism (or any less important generally) than the first, and, as my title indicates, it is the question I explore here. (12) I shall develop a view that is implicitly accepted by most egalitarians, namely that nonsubordination requires redistribution to follow moral responsibility, specifically by eliminating luck's differential effects on persons' fortunes while leaving persons fully to bear the consequences of their (morally responsible) choices. I shall consider certain central features of this responsibility-tracking egalitarianism more carefully and completely than others have done, and this will lead me to conclude that the view's dual ambitions are unattainable, so that much of the most prominent philosophical discussion of egalitarianism involves a fundamental mistake. Finally, I shall introduce a new account of egalitarianism based on a new conception of nonsubordination that avoids some of the errors of the responsibility-tracking view.

2. RESPONSIBILITY-TRACKING EGALITARIANISM

   The relative lack of attention received by the question "How much redistribution should there be?" is, perhaps, a product of the patterns of disagreement and agreement within the community of philosophers writing in the modern egalitarian tradition. Philosophers within this tradition disagree deeply about what equalisandum should count in a fundamental way for egalitarians. (13) But they appear to display a remarkable consensus around the responsibility-tracking account of when an equalisandum is distributed equally, that is, so as to secure the nonsubordination of persons. I shall aim, ultimately, to break this consensus; but I begin by elaborating the view that sustains it.

   The responsibility-tracking view begins from the observation that choice and luck jointly determine a person's fortunes. (14) Her fortunes improve or decline as she chooses well or badly and as her luck is good or bad. The crucial distinction between these two influences is that whereas a person is responsible for her choices, she is not responsible for her luck. (15) (It makes sense in this context to say, for example, that whether or not a person has a happy marriage is significantly a matter of her choice of husband, whereas whether or not she had a happy childhood was significantly a matter of her luck in her parents.) Put in another way, one may understand people as being at once agents, who act upon the world in ways for which they are responsible, and patients, on whom the world acts in ways they cannot control. (16) And, once people are understood in this way, it becomes natural to say that egalitarianism should treat them differently in their roles as agents and as patients. On the one hand, the distribution of advantage should be sensitive to differences in people's choices--there should be no redistribution to counter the differential effects of choice. On the other hand, the distribution of advantage should be insensitive to differences in people's luck--redistribution should eliminate the differential effects of luck. Taken together, these principles entail that egalitarianism should track responsibility--that is, they entail the responsibility-tracking view of how much redistribution there should be.

   The responsibility-tracking view of redistribution is in this way intuitively attractive. Moreover, it is both an old position and one that has attracted widespread support in the contemporary philosophical literature on egalitarianism, often from people who cannot agree on other matters, including the proper equalisandum. (17) John Stuart Mill, for example, believed that "[t]he proportioning of remuneration to work done is really just only in so far as the more or less of the work is a matter of choice: when it depends on natural difference of strength or capacity, this principle of remuneration is in itself an injustice." (18) More recently, Ronald Dworkin has suggested that egalitarianism should track the distinction between people's persons and their circumstances, (19) that it should be ambition-sensitive but may not be endowment-sensitive. (20) Similarly, Cohen has argued that egalitarianism's "purpose" is specifically "to eliminate involuntary disadvantage," which he stipulates to mean "disadvantage for which the sufferer cannot be held responsible" because it does not "appropriately reflect" his choices. (21) And John Roemer (commenting on this tradition of thought) has proposed that "society should indemnify people against poor outcomes that are the consequences of causes that are beyond their control, but not against outcomes that are the consequences of causes that are within their control, and therefore for which they are personally responsible." (22) Accordingly, Roemer concludes, "all those who exercised a comparable degree of responsibility are equal, regardless of their circumstances." (23) In spite of their differences concerning the choice of equalisandum, these egalitarians (and others as well (24)) converge on the responsibility-tracking view of how much redistribution there should be.

   Furthermore, the widespread adoption of the responsibility-tracking ideal by egalitarians of all stripes is more than merely fortuitous. Responsibility-tracking egalitarianism is connected to an intuitively powerful conception of the ideal I began by claiming lies at the heart of egalitarianism--the ideal of the nonsubordination of persons. On this account, nonsubordination requires that people are treated differently in their capacities as agents and as patients. Specifically, if people are to be treated equally as agents, then their choices must distinguish them, but if people are to be treated equally as patients, then their luck may not distinguish them. One might say, in this vein, that responsibility-tracking egalitarianism requires that people are allowed to be agents separately but made to be patients together.

   This is certainly a plausible position. Thus, when people are approached as agents, it seems that nonsubordination requires leaving each person equally to bear the consequences of her own choices, so that there may be no redistribution when people are separated...