Correlation Structure of the Permanent and Temporary Components of International Stock Market Prices

Abstract

This paper studies short-and long-horizon correlations among stock returns of six major stock markets using canonical correlation analysis that decomposes stock prices into permanent and transitory components. We introduce an efficient approach to estimate the term structure of correlations (for different investment horizons) by incorporating the dynamics of different price components. We demonstrate by both simulation and bootstrap procedure that our estimation method yields substantial efficiency gain relative to a more traditional approach. We further apply our methodology to study intertemporal stability of the correlations among the permanent return components and find results consistent with increasing international capital market integration.

JEL Classification: F39