Cointegration and rolling window cointegration analysis of a selected group of stock market indices

Abstract

In order to explore the degree of integration of international stock markets we select a group formed by the most important indices considering their market capitalization and geographical distribution. After testing for unit roots using the Augmented Dickey Fuller (ADF) and the Kwiatkowski Phillips Schmidt Shin (KPSS) tests, a full sample cointegration analysis is done provided that all the indices are found to be I(1). The existence of cointegration relationships can be interpreted economically as the existence of markets integration. Further, a rolling window cointegration testing procedure is implemented in the programming language R, in order to characterize the dynamic of the degree of integration ofstock markets. This analysis provides valuable information, given that an increase in the degree of cointegration can be interpreted as a signal of the presence of contagion among markets.