Diversity, Disagreement, and Information Aggregation
-Micro Theory Seminar
Joint with Xienan Cheng
Abstract: "Imagine two equally well, but imperfectly informed experts are hired to advise a decision maker. It seems obvious that the decision maker will not want the experts' private signals to be perfectly correlated. But, if not perfect correlation, which joint distribution of the experts' private signals are most helpful to the decision maker? To address this question we analyze Blackwell comparisons of joint distributions that are compatible with given marginal distributions of the two experts' private signals. We identify conditions under which an increase in the probability that the experts disagree improves, in Blackwell's sense, the informativeness of the advice that the decision maker receives."