I don't know how many of you out there have read John E. Roemer's (1998) book Equality of Opportunity, or the article of the same name in Arrow, Bowles, Durlauf (eds, 2000) Meritocracy and Economic Inequality. Anyway, I have been trying to think about his arguments for a while (as I am currently researching this topic) and I have a little hypothetical that I want to test to see if anyone thinks it's valid. First I will discuss Roemer's proposal, then on to my hypothetical.

He says that much has been written on the former, but not enough (rigorously) on the latter. Roemer argues for an algorithm, of the sort which takes individuals circumstances into account and then partitions them into typesets, where 'type' is meant to be indicative of level of advantage. He then goes on to argue that the individual's level of effort within their type set is what should determine our policy for receiving benefits. Everyone in the top centile of each type should receive similar outcomes. He argues therefore that his algorithm (which he sets up mathematically I should add) would therefore take types into account and provide for each individual contingent on the effort that they choose to expend and reward them accordingly - a just desert if you will. Now, this is a summary, and a poor one at that, but I think that it is quite an interesting proposition. One of the consequences of which is that disadvantaged individuals will require dramatically more compensation in order to get them to an equivalent level of say education as those who are historically advantaged. Note though that he argues that his EOp mechanism should apply in terms of admissions to schools, universities, etc, but should not apply for professions where a specific level of excellence is required (such as say, surgery, dentistry, high level physics and mathematics). At this point the nondiscrimination principle should be implemented.

[Aside: he comments how this makes him enemies on the right and left either because he goes too far, or not far enough]

Roemer's rule for allocating them to university, employment, 'future earnings capacity' would not allocate them (or at least allocate them to the lowest positions) because they expend (virtually) no effort (or epsilon's worth of effort).

Moreover, as the individual is unique, i.e. they cannot be classified in terms of the given type set, how do we classify them in terms of a decile of effort? Do they necessarily fall in the last decile of some imagined group of super-individuals? Surely that would be inaccurate because, if this group existed, none of them would be able to spend more than (epsilon or) near zero effort.