Nonparametric Tests of Conditional Mean-Variance Efficiency of a Benchmark Portfolio

In this paper we propose three nonparametric methods for testing conditional mean-variance efficiency of a benchmark portfolio. These approaches avoid functional form misspecification and share a pleasant feature that the test statistics are based on estimators that converge at the fast parametric rate. We derive limiting distributions of the test statistics and compare the three methods in a simulation study. We also present an application of the methods to testing the conditional CAPM