Contagion or interdependence: An application to the stock markets using unconditional cross-market correlations

In this paper, we consider stock market contagion as a significant increase in cross-market linkages after a shock to one country or group of countries. Under this definition we study if contagion occurred from the U.S. financial crisis to the rest of the major stock markets in the world by using the adjusted correlation coefficient approach which consists of testing if average cross market correlations increase significantly during the relevant period of turmoil. We would not reject the null hypothesis of interdependence in favour of contagion if the increase in correlation only suggests a continuation of high linkages in all state of the world. The majority of our results based on conditional correlation suggest contagion, but these results are biased though. Conditional correlation not only suggests contagion for all the countries in the sample, but also indicates no contagion but interdependence at the same time; this is clearly a lack of robustness.