The intertemporal stability of the covariance and correlation matrices of Hong Kong stock returns

Abstract

The intertemporal stability of the covariance matrix of stock returns is important in using ex-post factor structures on the APT and in portfolio optimization, while that of the correlation matrix is important in examining the ex-ante diversification benefits and stock return co-movements. By using Box's M test and an extension of the test, the intertemporal stability of the covariance and correlation matrices, respectively, of Hong Kong stock returns during the 1981–1992 time period are studied. Empirical results show that the covariance matrix of stock returns is less stable intertemporally than the corresponding correlation matrix and the results are robust across four different investment horizons. In general, the longer the investment horizon, the larger is the degree of intertemporal stability on both the covariance and correlation matrices of stock returns.