Dynamic cross-correlation and dynamic contagion of stock markets: a sliding windows approach with the DCCA correlation coefficient

摘要

How stock markets relate to each other is very important because this could have positive effects (such as enhancing economic growth) but also negative effects (possible contagion risks). Considering this issue, this study proposes continuous evaluation of the cross-correlations between markets, applying a sliding windows approach based on the detrended cross-correlation analysis correlation coefficient. Measuring the cross-correlations between the USA and other eight stock markets (the remainder of the G7 plus China and Russia), this allows dynamic analysis of the evolution of cross-correlations and also continuous analysis of the contagion effect. The results show that in the period before the crisis the correlation levels with the US stock market decreased, while post-crisis the results point to a contagion effect. As the proposed approach could be used for continuous monitoring of cross-correlations, this kind of information could be important for the different agents involved in stock markets.