Examination of country capital markets interaction by Granger Causality and Clustering Analysis

Globalization has led to an increase in international capital mobility. This has enabled the country's capital markets to become more integrated, and the benefits of international financial integration for the country markets, as well as its costs, have emerged. For this reason, it has been among the most studied subjects in finance science. This study examines the interaction between country capital markets through stock price indices. For the analysis, stock exchanges of 45 countries were taken and the period from August 2001 to December 2019 was examined by more than one method. With factor analysis, it was examined how the country stock markets formed a grouping in the international market, and then the causality relationship within and among factors was tested with Granger Causality Analysis. In addition, the causality relationship with Turkey and the remaining 44 countries was examined. As the third analysis, in order to observe the effects of the crisis in cluster analysis, the data set was divided into three periods and it was investigated how clusters were formed before, during and after the crisis. Finally, the hierarchical clustering results for the whole period were taken and compared with the factor analysis results. According to the results of the analysis, factors formed a grouping similar to developed and developing countries. However, it is seen that Asian countries seperate in a different factor. With Granger causality analysis, a significant integration relationship was found within and among factors. In total, 152 causality relationships were identified. It has been seen that the USA is a dominant stock market in the factor, while the USA and the UK are the dominant stock markets among the factors. In addition, it has been concluded that Turkey is integrated with especially developed countries. With the cluster analysis, it was observed that different clusters were formed before the crisis, after the crisis and during the crisis periods, and the number of clusters increased especially during the crisis. Although cluster analysis and factor analysis give similar results, cluster analysis is more geographically septated.