Hedge Funds and Optimal Asset Allocation: Bayesian Expectations, Time-Varying Investment Opportunities and Mean-Variance Spanning

Abstract

In this paper we analyze the contribution of hedge funds in optimal asset allocations for different investor clienteles. The preferences of specific institutional investors are captured by implementing a Bayesian asset allocation framework that incorporates heterogeneous expectations regarding alpha. Mean-variance spanning tests are used to draw inferences on the ability of hedge funds to enhance the efficient frontier. A novel variance decomposition procedure is employed for analyzing the co-movement of hedge fund returns with the benchmark assets. The empirical findings strongly indicate that portfolio benefits of hedge funds are time-varying and crucially depend on investor optimism regarding hedge fund alpha. Allocations to hedge funds improve the global minimum variance portfolio even after controlling for short-selling restrictions and minimum diversification constraints. However, the factor structure of hedge fund returns has become more similar to the benchmark assets due to dynamics underlying the composition of the aggregate hedge fund universe.