Interdependence and contagion in international stock markets: A latent Markov model approach

Abstract

In the last decades, interdependence among financial markets of different countries has represented a preferred topic in both theoretical and empirical studies. The prime focus of this paper is on the detection and evaluation of financial contagion and herd behaviour, which are accounted as predominant features of international markets. To this purpose we suggest to exploit the framework provided by latent Markov modelling, which we find extremely useful for detecting and defining the different stock market phases, referred to as regimes. Furthermore, we investigate the transitions between market phases by means of the regime switching probabilities, still provided within latent Markov models. The comparison between the dynamics of the latent stochastic processes underlying the observed time series of the international stock market returns provides significant insights on the level of their interdependence and on the measurement of contagion effects during financial crises. Our work contributes to the existing literature on both financial time series analysis and clustering and introduces a powerful and innovative approach for evaluating the linkages of the stock markets in different countries.