Three Thought Experiments About Wealth Distribution

Rawlsian, Nozickian, and Mischelian Worlds

In A Theory of Justice John Rawls famously argued that the just society is necessarily one in which the society's architects do not know ahead of time what their roles will be in it. If we don't know what our role will be in a society that we're going to be part of, rationally self-interested decision-makers would choose to create and be part of a society with a very egalitarian distribution of power. It might have been fun to be a plantation owner in the antebellum American South or a patrician in Rome, but if you fell out of the sky randomly into a social role in either place, chances were much greater in both cases that you would be a slave. The implication is that society-designers with foreknowledge are suspect because they can bias the game in their favor; even Thomas Jefferson surely believed that a man like himself would prosper in the kind of nation he was designing.

Robert Nozick took issue with the idea that the distribution of power in a just society would be close to egalitarian. He argued that if you can get from that just, egalitarian distribution to a non-egalitarian distribution through steps that are all just and un-coerced on their own, the resulting society must be just as well, even if it is no longer egalitarian. Making the thought experiment concerete: imagine an egalitarian Rawlsian world of a million people where everyone has $100,000, and one of these million is a gifted pianist named Steve. Word of Steve's amazing talent spreads, and he gives a concert attended by all his fellow citizens, all of whom gladly pay a dollar for the experience of hearing him play. Afterward, Steve has $1,099,999, and everyone else has $99,999. Steve is now the richest person in the world by a factor of 11. Is this unjust? If so, exactly which of these voluntary steps was the unjust step in moving from the Rawlsian to the Nozickian world?

A dimension which Nozick did not consider but which might affect the moral equation is the degree to which rational decision-makers vary in their abilities to make those decisions. Imagine after the first performance, all of Steve's fellow citizens are satisfied that they got their money's worth, but most of them say "once is enough." Some fraction of his audience was so moved that they go to a second performance of the same piece. Steve gets richer, and the repeat customers get another dollar poorer. Eventually his audience is reduced to a group of hardcore loyalists who find his performances so powerful, so emotionally rich, that nothing else in their lives compares, and they cease caring about anything else; they go to all his performances, they buy concert T-shirts, etc. and their wealth is transferred a dollar at a time (or more, if the pianist raises his prices) to the pianist, until they are broke. Whether or not Steve is aware of his role in their destitution is an interesting but separate question; for the time being let's say the lights in the concert hall are such that Steve can't see that he's reduced his fans to rags and bones, and he doesn't mingle after the shows. We can call this world the Mischelian world.

The dimension that Rawls and Nozick neglected in their thought experiments is variation of self-interested decision-making ability among the agents in the experiments (which may be transmissible across generations). If we add to the agents of the Nozickian world a distribution of rationally self-interested decision-making ability, we create the the Mischelian world described above. In the Mischelian world, some agents will be consistently better able to act in their rational self-interest because of superior working memory, rationality, critical thinking, or ability to delay gratification, and this will have an impact on the preservation of their material resources and the subsequent distribution thereof, all without any coercion. For those of us in liberal democracies, this is the world in which we are now living. Whether these traits are dictated by genetics or upbringing is irrelevant - what's relevant morally is that where agents fall along the distribution of these characteristics is not under the control of the agents, or for that matter of the Jeffersons designing new societies (not yet anyway.)


What Happens to Theories of Justice When the Agents Differ in Important Ways?

There are interesting implications for policy that fall out of these experiments, practical as much as moral, and they offer unpleasant suggestions to both side of the political spectrum. Perhaps most poignantly to redistributivists, it was a socialist (George Orwell) who pointed out in fiction the stubborn re-emergence of class structures in human societies even after violent attempts to cleanse it had been executed. (The "Inner Party"; hence members of the Chinese Communist Party explaining, with no hint of irony, why the uneducated rubes working in the factories and fields cannot be trusted with free elections and free speech.) This certainly seems to be bad news for attempts at economically egalitarian societies. This tendency of class structures to resist disturbance seems to be a socioeconomic parallel to Le Châtelier's principle: class re-emerges even where strong efforts have been made to obliterate it, even if the new structure rests on slightly different characteristics - blood relationships, ability to amass wealth, maneuver in bureaucratic hierarchies, or to parrot dogma as a loyalty signal have all been criteria for these structures at various points in history, including now. There are not many people who believe that individual differences do not matter to economic success, and that individual differences appear entirely randomly across populations rather than in consistent, inter-generationally robust association - the disagreement tends to be over the mechanism that made the agents different (i.e. whether due to upbringing, opportunities, genetics, etc.) Even Marx had to recognize that individuals could not be expected to produce incommensurate with their ability. Therefore, it's hard to see how strongly wealth-redistributive policies could matter unless they were carried out continuously. In that case, it is also very hard to see why this would not necessarily result in overall slowing of economic growth and therefore, less happiness for everyone.

Some implications of the Mischelian world are not entirely pleasant to the more libertarian among us either, as they justify some degree of paternalism. All but the most absolutist libertarian concedes that there is a cut-off in the distribution of rational decision-making ability, beyond which individuals cannot be responsible only and entirely for him or herself. If you went to Steve the pianist's concerts and found that there were children or retarded or senile or insane people spending their last dimes on him, would you be comfortable with that? What if they were "normal", but instead of piano concertos Steve were selling lots in gambling games, or opium, or sexual acts? All of us have more trouble behaving with rational self-interest with certain goods and services. Keep in mind, the question is whether a just society allows Steve to sell such things, not whether he is moral in doing so - if indeed there is a difference between those two questions. If legal distinctions limiting commerce and other types of decision-making responsibility are seen as necessary for justice in a Mischelian world, this could ironically be seen simultaneously as a justification for elitism and paternalism as well as a reward for less responsible, productive agents.

Perhaps the solution in the Mischelian world is to define a cut-off for humans at the margins of rational decision-making ability, and then keep them being responsible for their decisions. In fact in our own Mischelian world, that's exactly what we do, for children and the mentally disabled, though especially with children we don't have the means or resources to determine exactly where each individual's rationality falls at every age. So, in most countries we arbitrarily choose an age of majority. (Straight-ahead as it seems, this practice is still attacked, as you can see with one hot-under-the-collar commenter at my outdoor-activities blog.)

The problem with cut-offs (of products - "alcohol isn't addictive and harmful enough to ban, but heroin is") or agents ("my daughter is only 14 but she's mature enough to be allowed to drive") is that they're coarse-grained approximations (often using proxy indicators like age), and in fact even if we're not below the lower cut-off on the rational decision-making spectrum, there are no doubt lots of people with overall better rational decision-making capabilities than us. These cut-offs are also based on rational decision-making over the long-term, which is often obvious after brief interactions (with the retarded or demented) but not always. Should a just society require you to assess the rationality of everyone with whom you engage in commerce, or of the particular decisions involving this particular transaction? When I was in college, a mom-and-pop sandwich shop opened up that had 25-cent burger specials on Friday nights. Like every other bottomless-stomached college student, I showed up every Friday and got in line and ordered 6 of them. Eventually they put up a sign asking us to please consider ordering something else because they lost money on every burger. Of course, we all had a good laugh at this, and ordered more burgers, and soon after they closed. Clearly their belief that our good hearts would prevent us from taking advantage of their kindness was false, and we knew it. Should they have been protected from making this decision, or from us mean students from taking advantage of them? Having met the couple and talked to them briefly, I can attest that they seemed non-retarded and non-demented, but I still might be smarter than them. Does that make a difference? Should we be protected, from heroin or from Steve's heart-rending performances (or should the more rational people be handicapped?) Is the cut-off for children and the disabled only a result of our limited methods and resources? That is to say, in a future of finer-grained social justice, is it desirable for technology to allow us a gradient of protection for individuals who are even a little bit unable to control themselves in certain circumstances, for example your blogger who has confessed to his chocoholism?

The issue with the Mischelian world is that there are differences between the agents that make up these societies that result from nature and are not eliminable by the application of justice (not yet), even though they definitely affect whether a just state obtains. It also seems that the coarse-grained nature of the way we protect the irrational from injustice results only from limitations in technology and resources, rather than from a positive decision that our protections should be only so detailed, and not go further. I believe the discussion will proceed in this century, and if it goes far enough in my own lifetime, Safeway will start refusing to sell me Kit Kats.


Conclusion: Wealth is Only a Contributor to Utility, But So Is Genetics

It seems obvious that the reason anyone cares about economic egalitarianism is because wealth relates to utility; that is, wealth is a proxy indicator of happiness, which is what we care about and why you're reading this. It is not clear that there can be such a thing as a just society that cares about economic equality but not utility equality. Gross National Happiness is the best known direct quantitative indicator of utility. As with wealth, utility is in part dependent on agent-specific traits which differ along a spectrum, and there is some indication that there is such a thing as a happiness set point. Needless to say, like many factors influencing rationality, happiness set point is a given, not subject to the decisions of a just society but absolutely impacting them. In a just society, do the naturally happier individuals owe cheering-up efforts to the naturally sadder? To refer to another of Nozick's thought experiments, in a sense the individuals at the low end of the happiness set-point curve, in a just society that works to raise their utility, are anti-utility monsters: in an effort to bring the unhappiest up to a certain cut-off point, the utility of the happier is consumed, and overall utility decreases.

Finally, I have alluded twice to our sense of justice not yet being able to eliminate these differences and therefore having to enact justice in the way society deals with the hands that its agents are dealt. It is my sense at least in the U.S. that those who are most in favor of wealth-redistributive policies are likely to be the most strongly opposed to even investigating whether there is a genetic basis for the rationality differences between agents, much less whether our sense of justice dictates that we do something to improve them. Such differences could certainly not be the only reasons for the persistence of inequality, but if they exist they would certainly be a root cause. In the coming decades we will understand much more about the genetics of cognition, and going forward, policy discussions about wealth inequality must make reference to these findings.