Title: Cost Efficiency in Incomplete Markets

Abstract: We establish a cost efficiency principle that allows to characterize the optimizer of general portfolio optimization problems in distributional terms: The price of a cost-minimizing strategy is identified as the minimal superhedging price of random variables dominated in convex order. This result relies on a minimax theorem on the cone of nonnegative random variables and lends itself to a wide array of applications, such as the distribution builder approach to portfolio optimization, the evaluation of the performance of hedge fund managers and the re-interpretation of results of behavioral finance in classical terms. This is joint work with Carole Bernard.