Statistical inferences for realized portfolio weights

Statistical inferences for weights of the global minimum variance portfolio (GMVP) are of both theoretical and practical relevance for mean-variance portfolio selection. Daily realized GMVP weights depend only on realized covariance matrix computed from intraday high-frequency returns. Both finite sample and asymptotic distributional properties of the realized GMVP weights are deduced. Then, statistical tests for the GMVP proportions are developed in order to provide sequential monitoring with on-line decisions whether a given portfolio composition deviates from the current GMVP significantly. The theoretical results are illustrated both in Monte Carlo simulations and in an empirical application.