Large-scale portfolios using realized covariance matrix: evidence from the Japanese stock market

The objective of this paper is to examine effects of realized covariance matrix estimators based on intraday returns on large-scale minimum-variance equity portfolio optimization. We empirically assess out-of-sample performance of portfolios with different covariance matrix estimators: the realized covariance matrix estimators and Bayesian shrinkage estimators based on the past monthly and daily returns. The main results are: (1) the realized covariance matrix estimators using the past intraday returns yield a lower standard deviation of the large-scale portfolio returns than the Bayesian shrinkage estimators based on the monthly and daily historical returns; (2) gains to switching to strategies using the realized covariance matrix estimators are higher for an investor with higher relative risk aversion; and (3) the better portfolio performance of the realized covariance approach implied by ex-post returns in excess of the risk-free rate, the standard deviations of the excess returns, the return per unit of risk (Sharpe ratio) and the switching fees seems to be robust to the level of transaction costs.