On Decentralized Affirmative Action Policies and Their Duration

Abstract

Successive decentralized policy makers must decide whether to implement an affirmative action policy aimed at improving the performance distribution of future generations of a targeted group in their district. Workers receive wages corresponding to their expected performance, suffer a feeling of injustice when getting less than their actual performance, and employers do not observe district by district whether workers benefited from affirmative action. We find that welfare-maximizing policy makers choose to implement affirmative action perpetually, despite the resulting feeling of injustice that eventually dominates the anticipated beneficial effect on the targeted group's performance. This is in contrast with the first-best that requires affirmative action to be temporary.

Keywords: Affirmative Action, General Equilibrium, Loss Aversion, Prospect Theory, Moral Hazard, Game Theory

JEL Classification: D40, I28, I30, J15