Jessica recapitulates the part of John Rawls’ theory that economists care about (concerning any other parts, I generally must plead ignorance). She recently read Will Wilkinson’s thoughts on it (linking to this particular post), and though I don’t believe that Rawlsekianism has much in the way of legs, pushing the left towards freer markets isn’t such a bad thing. At any rate, I promised that I would critique Rawls, and that is what this is all about. In a few short pages, James Buchanan laid down a short but very strong critique of Rawls. Although I no longer have access to that paper*, I am going to use some of that argument (to the best of my recollection).
Rawls, Buchanan, and FA Hayek each came up with their own version of constitutionalism at roughly the same time, each fleshing out basic ideas in the 1950s but really putting things together during the 1970s, so it is interesting to see how three thinkers can go down the same theoretical road but end up using entirely separate methods and at different results. Draw your own comparisons to the first batch of social contract philosophers, noting that Buchanan relied heavily upon Hobbes in his theory.
Each of the three sees a subset of the entire set of possible sets of constitutional rules as normatively reasonable and others normatively unreasonable. None of the three, for example, wishes a Leviathan-style government in which one individual controls everything without any checks. The question, however, is how to describe the subset of normatively reasonable solutions and from there, how do we get to one of those? I think Jim Buchanan’s depiction of the set is the clearest and best, at least in certain spheres of rules: look at things as m n-person cooperation games or prisoner’s dilemmas, where m is the number of areas in which a rule can exist. Within each of the m areas, if the area manifests itself historically as a cooperation game, either no rule is required or a rule is required to have citizens choose the best cooperative option**. On the other hand, the games which look like prisoner’s dilemmas will always require rules, and the best rule is to eliminate the off-diagonals, leaving only options in which all individuals perform the same action and the temptation from defecting when others cooperate is gone. After doing that, individuals will all cooperate.
Given a set of game theoretic mechanics, how do we convince individuals to submit to rules? After all, individuals already exist and there is a status quo in place which we have to upset in order to introduce our set of normatively reasonable rules. These rules will likely have a negative effect on some individuals and a positive effect on others, so how will we convince those who may lose to play along? For Rawls, he first assumes risk-averse individuals and then strips them entirely of their history and knowledge of particular features, placing a veil of ignorance around them. Without any knowledge of who you are and what you could be, you enter into debate with all of the other ethereal spirits who will exist in your world. Because of your risk-aversion, you would like conditions to be as good as possible for the worst among you***, which is his difference principle. To see it in graphical form, look to your left. This is a simple version of a graph with two individuals, where the utility is equal to the minimum utility of the individuals in the society. In other words, we have indifference curves following a Leontief function occurring about a 45-degree line from the origin. A sufficiently risk-averse individual, with no clue whether he be person X or person Y, will submit to a set of rules which maximizes the function min(U(X), U(Y)). U(X) and U(Y) will not necessarily be equal, so there can still be relative winners and losers, but there is no set of rules which will assist the worst-off (individual Y, for example) without making some better-off individual (X) worse than the previous worst-off individual (so U(X, new) < U(Y, old) < U(Y, new)). This outcome is called the difference principle.
Now that we have a quick overview of the major points—the veil of ignorance and the difference principle—I can go into a critique, working in ascending order of seriousness. The first critique is that Rawls assumes that all individuals are risk-averse. This is a fairly common assumption in neoclassical economics, so I can’t fault him for it, but to get his outcome, you have to consider individuals to be seriously risk-averse, not just mildly. Suppose you have a thousand individuals in a society and two potential sets of rules. With rule set 1, all 1000 individuals will end up with a utility of 2. With rule set 2, 999 individuals will have a utility of 23 million, whereas one unlucky soul has a utility of 1 (where 1 is slightly above the bare minimum required to live, so that person will still be able to survive). If you take Rawls seriously here, everybody would be so supremely risk-averse that they would turn down an almost-guaranteed life of luxury for guaranteed decrepitude simply because otherwise, one person’s living in a van down by the river. This result just does not make sense, and though a Rawlsian would likely counter by arguing that this is an edge case and that the people with a utility of 23 million apiece could surely come together and give a few bucks to Matt Foley, a reasonable theory would have to deal with cases similar (but less extreme) than these.
A stronger critique, which Buchanan uses, is to note that we are never in an ex nihilo situation but instead deal in a world which already exists, rules which already exist, and outcomes which have already come about. Even if you have a veil of ignorance around the present and future, it is impossible to have it around the past. I know what kind of utility I had with the previous set of rules and I am going to want that at least that utility with a new set, especially if I am as risk-averse as Rawls makes me out to be.
Finally, the strongest critique—another which Buchanan brings up—is that our veil of ignorance is necessarily incomplete. You do have at least some knowledge about your characteristics, contra the Rawlsian assumption. I may not know what my exact characteristic makeup will be twenty or thirty years from now, what my precise beliefs and standing will be, etc., but I have a reasonable idea and a better idea of what my position will be two years from now. Because of that, when it comes time to come up with a constitution, I do have an idea of where my interests lie and as a self-interested individual, I am obviously going to use that knowledge to further them. Having me pretend that I don’t know my probabilistic makeup (e.g., I believe there is a 30% chance that I will end up in situation X, a 90% chance that I will adhere to belief z, or a 0.015% chance that I will be hit by a bus and become mentally incapable) simply does not work because you have a prisoner’s dilemma situation for Rawlsians. If everybody else (“all y’all,” in South Carolinian) were to adhere to the veil of ignorance and forget about their situations, I get an advantage by remembering my situation and jiggering the constitution to favor me. If everybody else were to remove the veil of ignorance, I would be at a deficit if I kept mine. Thus, defection and removing the veil of ignorance is the winning strategy.
Given this, how do we respond? Again, I can take Buchanan as the model here (even though I tend closer to the Hayekian model of constitutionalism). In Buchanan’s book, The Limits of Liberty, he draws up how a group of individuals could, starting with knowledge of their own situations and in a Hobbesian state of nature, create a value-increasing constitution. For Buchanan, extreme risk-aversion is not a necessary aspect and his form of constitutional economics will work even with risk-neutral individuals. In addition, Buchanan has a “veil of uncertainty,” which is like a drastically weakened veil of ignorance. With the veil of uncertainty, we all know our present and past characteristics and circumstances. Furthermore, we understand the history of our society (and possibly other societies as well) and even have an idea of what our future will hold given sets of rules. We will not necessarily be right—Knightian uncertainty can play a role here, as well as our future characteristics following probabilistic models—but we are not entirely in the dark. As a result, Buchanan’s constitutional prescriptions are much more circumscribed and modest than those of Rawls. Buchanan argues that we should limit constitutional rules—as much as possible—to abstract and long-term situations, increasing the percentage of rules which fall in a hazy area of the veil of uncertainty and thus increase the degrees of freedom available for negotiation. In fact, it is even possible to get oligarchs on board with a constitution which eventually leads to their own destruction, by timing rules so that the current generation will not be affected by certain rules or they receive up-front a one-time cash payout (or a pension) equivalent to their losses from the normatively superior set of constitutional rules.
This sums up my critique of Rawls and scratches the surface of why I would consider Buchanan a much better starting point for constitutional economics. You don’t need as restrictive of assumptions (really, all you need is the “hard core” of methodological and normative individualism and self-interest, as well as unanimity and the capability for wheeling and dealing) and Buchanan-style constitutionalism is actually possible in the realm of man.
* – Rawls on Justice as Fairness,” Public Choice 13 (fall 1972): 123–28. In Social Justice and Classical Liberal Goals, vol. 3, Social Choice Theory, ed. Charles K. Rowley (London: Edward Elgar, 1993), 75–80.
** – Viktor Vanberg, in Rules and Choice in Economics, chapter 5, argues that Hayek saw such cooperative games as evolving to a set of best practices, but it is possible to get stuck in a sub-optimal equilibrium if enough people start out in one. If driving on the left side of the road when everybody else drives on the left nets you a utility of 1 but driving on the right side when everybody else drives on the right nets you a utility of 2 (because you can travel to other countries more easily, for example), you would think that everybody would drive on the right side. But if all individuals start out driving on the left side of the road, they will stay on that side even if every individual sees the utility gain from _all_ individuals switching. When one individual switches, though, you have a utility loss: some moron driving on the wrong side of the road. What you can do about this situation, Vanberg argues, is to have a rule-making authority decree that from a particular day on, everybody shall drive on the right side of the road. He uses the example of Sweden quite often, as they did exactly this. On the other hand, somebody once wrote a paper arguing that with two societies with open immigration and a stochastic function representing which action an individual would take (drive right, drive left, move to the other island and drive right, move to the other island and drive left), eventually the higher-utility option would win out. I can’t find this paper anymore to credit the person, but I thought it was in the Review of Austrian Economics.
*** – Throw in your own Christian reference yourself; I don’t do that stuff for you.