In the 1990s and in his 1999 book “Equality of opportunity”, John Roemer set the stage for what has since been a flourishing area of inequality studies—inequality of opportunity. Roemer’s key insight was to divide the factors that influence one’s income into three parts: circumstances or factors that are exogenous to the individual, that is over which he has no control (gender, race, parental income and education etc.), those that are the product of his effort, and finally those that are the result of what Roemer called “episodic luck” (I got a good job because I just happened to be available at the time when the job was advertised).
Roemer’s approach has led him also to propose a very radical way to reward (pay) people. Consider two groups of individuals, defined by an exogeneous marker like gender. One group (men) tends to be physically stronger and produces on average 10 widgets per day. Another group, women, is physically weaker and produces only 5. Should everybody be paid according to the number of widgets he/she produces (which a simple-minded “meritocratic” approach would suggest)? No, Roemer says, the reward should be proportional to our contribution compared to the average of our group. So, if I produce 12 widgets, why is 20% above men’s average, I should be paid the same as the woman who produces 6 widgets, which is also 20% above the average of her group. The reason is that we are both paid according to our (differential) effort—and the attempt is made to control (abstract from) our innate characteristics which may privilege or punish some of us.
Consider the radicalness of this proposal as applied in another context. Students’ grades should also follow the same rule. If say rich parents’ kids do better than poor parents’ kids on average, then a rich kid who has scored 12 points on a test should get the same grade as the poorer kid who has scored only 6 points. And so on.
But recently as I was rereading Phelps Brown’s book “The Inequality of Pay”, published in 1977, I ran across a different scheme of rewards applied in Beijing in the 1960s, around the time of the Cultural Revolution. All men were paid according to the average number of widgets produced by men, and all women were paid according to the average number of widgets produced by women. Here is the quote from Phelps Brown:
This story brings out what to Western observers may seem a contradiction in Chinese pay structure: if it is right and proper to pay a man more than a woman because the man being stronger produces more, why should not a man who exerts himself and produces more than another man likewise be paid more? To the Chinese the answer is simply that the latter differential appeals to self-interest whereas the former cannot. Strangely but intelligibly, the Chinese treat payment in proportion to the amount of work done as a self-evident principle of natural justice while differences in that amount are not within the worker’s own control, but as mischievous when they are. (p. 53; emphasis mine)
The reader, probably having thought how radical and left-wing is Roemer’s proposal, is now suddenly thrown into this most radical left-wing experiment ever where—the very opposite principles rule! It seems that there is no continuity: a more left-wing approach is not just slightly more to the left than the less leftist approach—it is the very opposite of it!
To see that, recall that in Roemer’s case we do not want to pay somebody for his or her circumstances, but only for his or her effort. In the Chinese case, it is the reverse: we pay somebody only for his or her circumstances, but not for his or her effort. Why is that? The philosophy is entirely different. Circumstances are viewed as “natural” and one should be paid according to them. But payment according to effort is viewed as corrosive of moral norms since it means that people respond to economic incentives. People should work either because they want to contribute to the community (without expecting anything in return) or because they like to work. “Incentivizing” --appealing to self-interest--in such a setting is considered as bad, as in a different setting paying somebody for an exogenous advantage that he or she does not deserve.
The ultimate outcome of the Chinese system is an equal pay for everybody, both men and women, and regardless of individual productivity. It would be at the polar extreme of the “meritocratic” pay where everybody is paid simply according to the number of widgets he or she produces.
What is the best way? Meritocratic pay responds to out feeling of justice that everybody should be paid according to their contribution. Presumably it would lead to the highest output. Roemer redefines justice so as to extract only differential effort according to which people should be paid. They will be paid the same amount for different number of widgets produced. Empirically, it will be always very difficult to determine what are the factors that should fall under the heading of circumstances and hence should not affect the reward. The Chinese system has a moralistic element in it: it is bad to be incentivized by pay. The downside is that it is likely to lead to very low effort of most participants.
When we design systems of rewards, we are obviously always led by some principles of justice or ethics. The problem is that these principles do not all come up with the same solution. In many cases, as we have seen here, depending on what our guiding principle is, the reward structure will be very different. On top of that we need, in principle, to take into account the effects on the overall output—unless of course our philosophical principle Is such that the quantity of that output is immaterial.
John Roemer who kindly commented on the text asked me to make clear that he never advocated direct application of the principles explained here (nor thought that this would be possible to do in a market economy), but argued that policies like affirmative action should be designed with the objective of reducing or eliminating the impact of circumstances on one's income.