I quote Charlie Munger, who is always a fascinating read:
"It is not always recognized that, to function best, morality should sometimes appear unfair, like most worldly outcomes. The craving for perfect fairness causes a lot of terrible problems in system function. Some systems should be made deliberately unfair to individuals because they’ll be fairer on average for all of us. I frequently cite the example of having your career over, in the Navy, if your ship goes aground, even if it wasn’t your fault. I say the lack of justice for the one guy that wasn’t at fault is way more than made up by a greater justice for everybody when every captain of a ship always sweats blood to make sure the ship doesn’t go aground. Tolerating a little unfairness to some to get a greater fairness for all is a model I recommend to all of you. But again, I wouldn’t put it in your assigned college work if you want to be graded well, particularly in a modern law school wherein there is usually an over-love of fairness-seeking process."
In other words, fairness of outcome (as opposed to fairness of opportunity) strips participants of incentives to try and better their own outcome, because at some level, you face forgiveness for nonperformance and thus don't put in quite as much effort. A completely unforgiving world seems obviously negative - failure needs to be possible without destroying you, or nobody would take risks, which is the basis of common law bankruptcy procedures. But a comfortable level also should not be guaranteed, or else people will not strive for improvement. Fundamentally, this is the basis of many of the "social safety net" arguments. As society gets richer and it's easier for everyone to become complacent, it's especially important not to foster that complacency. If you want to avoid people failing miserably, then you need to facilitate big, big rewards for success - big enough that not getting those rewards feels like failure. The problem with high marginal tax rates (over 100% for the lower half of the distribution) is that it gives no incentive to keep working.
In addition to economic gain, I actually wonder how much this applies to Affirmative Action. I've outlined three other significant objections to affirmative action in my "Government Priorities" list, but this one seems to apply as well: do very high-performing minorities put in less work because they know they're held to lower standards? Wouldn't that cause the racial gap to persist longer?