Rawls on Interdependent Preferences

by Will Wilkinson on January 18, 2006

While reading Randall Holcombe and Russ Sobel’s excellent paper “Consumption Externalities and Economic Welfare” (thanks to Michael Dennis, and about which more later), I ran across a cite to Rawls on interdependent preferences. It turns out that I’ve read this very important passage a bunch of times, but not since becoming obsessed with positional races, interdependent preferences, and suchlike. Here is what Rawls said:

[According to utilitarianism], [w]e arrange institutions so as to obtain the greatest sum of satisfactions; we ask no questions about their source or quality but only how their satisfaction would affect the total of well-being. Social welfare depends directly and solely upon the levels of satisfaction or dissatisfaction of individuals. Thus 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 deliberations according to their intensity, or whatever, alnong with other desires. If society decides to deny them fulfillment, or to supress them, it is because they tend to be socially destructive and a greater welfare can be achieved in other ways.

In justice as fairness, on the other hand, persons accept in advance a principle of equal liberty and they do this without a knowledge of their more particular ends. They implicitly agree, therefore, to conform their conceptions of the good to what the principles of justice require, or at least not to press claims that directly violate them. An individual who finds that he enjoys seeing others in positions of lesser liberty understands that he has no claim whatever to this enjoyment. The pleasure he takes in others’ deprivations is wrong in itself: it is a satisfaction which requires the violation of a principles to which he would agree in the original position. The principles of right, and so of justice, put limits on which satisfactions have value; they impose restrictions on what are reasonable conceptions of one’s good. In drawing up plans and in deciding on aspirations men are to take these constraints into account. Hence in justice as fairness one does not take men’s propensities and inclinations as given, whatever they are, and then seek the best way to fulfill them . . . The priority of justice is accounted for, in part, by holding that the interests reqiring the violation of justice have no value. Having no merit in the first place, they cannot override its claims. [ToJ, 2nd ed., p. 28.]

Although I didn’t have this passage explicitly in mind, if you subscribe to Reason, you can see how deep in my system Rawls’s sensibility is in my review of Layard’s Happiness. There I said:

Layard’s account of economic success as “pollution” is a striking illustration of what philosopher John Rawls had in mind when he argued that utilitarianism fails to take seriously the separateness of persons. If it is legitimate to use the coercive arm of the state to discourage work simply because it makes other people feel bad, then our liberty to pursue our own ends, for our own reasons, is hostage to the way the brains of strangers happen to light up. The aims and beliefs that make us distinct persons are reduced to nothing, except as they count in the summing.

And my implied point is: if the happiness center of your envious brain happens to light up when I do less well, or the unhappiness center of your bigoted brain lights up when a minority family moves in next door, then, when deliberating about policy, we ought graciously to ignore your brain and its preferences in these matters.

There simply is no way to do policy analysis without making some normative ruling about permissible and impermissible preferences at the outset. The virtue of Rawls is that he tries to draw the distinction in a principled way, and at a high level of generality, without leaning too heavily on any one comprehensive conception of the good.

  • Will Wilkinson
    Blar, Your point doesn't make any sense to me. From the perspective of justice, Chet's motivation in buying a boat means nothing. If the boat, or the status, is a value to him, then it is a value. And, of course, Chet's buying the boat can be quite valuable overall, regardless of his motivation -- to the people who made the boat, and made the bits that made the boat, and shipped that bits that made the boat, etc. The broken windows analogy is bad because a broken window is a loss of wealth. A yacht is a form of wealth. (Blowing up your yacht just show that you can would be more like it.) Becker's argument that absent status races people would take too few social-welfare enhancing entrepreneurial gambles is precisely the kind of argument that would lead parties behind the veil to endorse status races.

    But, more generally, action in general, and consumption in particular, is permissible as long as it does not constitute a harm. Chet's buying a boat may not actually make him better off, but justice doesn't demand maximal prudence. It simply prohibits injustice. My Rawlsian point was that your sense of having your relative position harmed by Chet's yacht does not amount to real injustice on Chets part, so does not demand rectification or correction.
  • Jadagul, if people are spending their money on something that is of no value, then that's a problem, and an inefficiency in the economy. To say otherwise - that it helps the economy to have people like Chet buy worthless status-enhancers - is to commit the broken window fallacy. At least under the Rawls view, producing a status-enhancer is like breaking and fixing a window, in that it costs resources but adds no value. If clever policies made it so that society didn't need to produce so many worthless yachts, then that would be an improvement, just like getting society to stop needing to fix so many windows.
  • Blar: Even if we decide that the status-seeking desire is of no value, that doesn't mean it's of negative value. The argument that racism should be disincentivized has to rest on the fact that racism is harmful to others; the fact that racism is an invalid preference just means that we shouldn't weigh bigots' desire to discriminate against minorities' desire not to be discriminated against.

    In the status game, Will and you both argue that preferences for status should be discounted (or something similar. Take that as a simplification for purposes of argument). But if we discount Chet's desire to have a bigger yacht than his neighbor, we also have to discount his neighbor's desire to have a bigger yacht than Chet; thus for policy purposes, we would not want to encourage or discourage Chet from producing on those grounds alone. Chet is attempting to benefit himself at another's expense, apparently, but the premise of the post is that we shouldn't register the harm to the other on our utilitometer. So the only utility change we're left with is "Chet works hard, produces stuff, raises standard of living," which seems to me a good thing.
  • Will, those are four important points, all worth discussing, but I don't want these comments to be a debate of every argument you've made against Layard. I'd rather focus on this post, on Rawls and interdependent preferences. The main point of my comment was that the Rawlsian approach of deciding that some preferences don't count does not weaken Layard's general argument. If people are working to satisfy preferences that don't have any value, then that's just as much of a problem as people working to benefit themselves in a way that harms others. In this comment to one of your earlier posts, I identified both of these problems with status-seeking (people striving for something that isn't really of value and people striving for something that benefits them at others' expense), and now I can add that, depending on how we define value, we might transform one of these problems into the other, but that doesn't make the problem disappear. Your comment did not respond to this argument.

    I suppose that it's possible for the Rawlsian approach to add to your anti-Layard argument if you can come up with a principled way to have the benefits that people gain by getting ahead in the status game count, while the harms to others from falling behind in the status game do not count, but this seems hard to do. It is more likely that ruling out certain preferences in our normative judgments would weaken the argument made in your first quick response (that status games create new status rather than merely redistributing existing status), since there's a good chance that a principled application of the Rawlsian approach would lead us to decide that at least some of these status-benefits have no value.
  • Will Wilkinson
    Well, then you ought to have your head on straight!
  • john g
    Sobel teaches at my alma mater WVU. I competed in the Business Competition he ran. I also worked for a summer with Mr. Wilkinson at Cato.

    Creepy, huh?
  • Will,

    I don't see what's inconsistent about advocating individual strategies to reduce positional preferences, while saying that public policy still has to deal with the reality of a human nature in which those preferences are prominent. This is just an Augustinian stance: we have to simultaneously reside in a hoped-for City based on universal love, and the City based on empirical human nature, including greed, envy and malice.

    Take Amy Chua's writings about the relationship between market dominant minorities and politically dominant majorities. You can say that the Malays ought not to care about the material success of the Chinese, since they are still absolutely better off with Chinese entrepenurial skill than they would otherwise be. But if your trying to avoid race riots, you need to take human nature as it is, not as you would like it to be.
  • Will Wilkinson
    Blar, A few quick responses. First, I think status is fantastically multi-dimensional and contextual, and don't think status-seeking, in general, is a zero sum game at all. Second, it is by no means clear that, even if status-seeking was zero sum, that it would be inefficient. Becker and Mulligan argue that absent the highly valued prospect of a big upward move in status, people would take too few entrepeneurial gambles, slowing growth, hurting everyone, but mostly the poor. Third: there is actually no evidence that I can find that shows that a Frank/Layard tax-truce in the income status race will actually have a positve welfare effect, in hedonic terms. There is no evidence that people who work more are therefore less happy than people who work less. And there is no good reason to believe that status races don't just jump to a new dimension of comparison when comparison on the old dimension is de-incentivized. Fourth, Chet: Both Chet and his neighbor have the option to care less, or not at all, about the size of each other's yachts. It is not clear how their failure to care less gives their preferences normative salience. Closely related, Coase: it takes two to externality tango. The neighbor's yacht can't be an neg. externality for Chet but for Chet's preference. Since Chet can change his preferences, that might be the "least cost" stategy of harm mitigation. And it probably would be, since he'll gain other big benefits from becoming less fixated on social comparison. The entire back of Layard's own book is full of solid individual strategies for reducing one's disposition to social comparison. He shadily argues that policy must take preferences about relative position into account, because that's just human nature, and then tells us how to get rid of or constructively rechannel our preferences for relative position. WTF!?
  • Will, I think that a Layard-style argument against the rat race goes through just as well on a Rawls-style view of what's valuable as it does on a utilitarianism-style view of what's valuable. Layard's basic point is that a lot of what people do is directed at succeeding in a positional game which he calls status-seeking. People like it when they're doing well in this game relative to other people, and they don't like it when they aren't. This is a problem, since it's a waste of people's efforts (even if the game is not strictly zero-sum, and it does produce some things of value rather than simply redistributing positional goods, it is still inefficient). The utilitarianism-style objection to the positional game is that every positive movement in relative position creates an equivalent negative movement which cancels out its value. The Rawls-style objection is that the whole game is worthless, since people's preferences about their relative position have no value. Either way, it seems like good public policy to get people to direct their time and energy at something other than this positional game.

    To take a crude example, say that Chet and his neighbor each have a yacht. The neighbor's yacht is bigger, which makes Chet unhappy, so Chet works really hard and buys a new, bigger yacht. Now Chet is happy that he has a bigger yacht than his neighbor, and his neighbor is unhappy about being bested by Chet. Layard says that these two changes cancel out, Rawls that happiness and unhappiness based on who has a bigger yacht is irrelevant and lacking in value.

    In real life, the effects of relative position are probably not so crude. For the most part, I think that they are not based on emotions like envy or schadenfreude, but rather on indirect effects like how people evaluate how well their lives our going. This could make things harder for the Rawls-style argument, since we aren't just talking about bigoted preferences.
  • Austen
    Sure, Rawls's is one sort of way of thinking about such constraints. Seems to me it's not the only way, or necessarily the best way. Normative theorists from all quarters are compelled to articulate and justify, or at any rate, to accept, some "principles of right", as limits on the acceptability of preferences. If utilitarians have to "lean on" some very general conception of the good (like, a Mill-style hedonism, or a diluted eudaimonism) in order to arrive at maxims like "Don't respect bigoted preferences" -- is that any less feasible an approach than Rawls's own beset-with-problems, intuition-heavy one?

    Or, maybe your "Rawls sensibility" is irreversibly entrenched, and this post is little more than a pumping of it. In which case: hey, rock on.

    Or, maybe what really bothers Rawls and Kantians (and you?) about the value-based approach is the likely concession of impermissible preferences' (albeit, comparatively lesser) ~value~. The miscreant's thrill during the jacking of cars is not, a utilitarian could say, utterly devoid of value. The miscreant shouldn't do it, and we shouldn't tolerate it. Nor are we wrong to feel disgust at the thrill. But yeah, dispassionately considered, there's a tiny bit of value in the thrill, and thus in the preference for it. So what?
blog comments powered by Disqus

Previous post:

Next post: