Charles Manski from the Department of Economic, Northwestern University Seminar

Economists studying choice with partial knowledge typically assume that the decision maker places a subjective distribution on unknown quantities and maximizes expected utility. Someone lackinga credible subjective distribution faces a problem of choice under ambiguity. This article reviews recent research on policy choice under ambiguity, when the task is to choose treatments for apopulation. Ambiguity arises when a planner has partial knowledge of treatment response and,hence, cannot determine the optimal policy. I first discuss dominance and alternative criteria for choice among undominated policies. I then illustrate with choice of a vaccination policy by a planner who has partial knowledge of the effect of vaccination on illness. I next study a class of problems where a planner may want to cope with ambiguity by diversification, assigning observationally identical persons to different treatments. Lastly, I consider a setting where a planner should not diversify treatment.