Clustering global equity markets with variance ratio tests

In this paper we employ variance ratio tests of the random walk hypothesis to investigate the interdependence of global equity markets in terms of the predictability of equity index returns and how the clustering pattern has evolved in recent years. The study is based on almost 15 years of daily returns of free float-adjusted market capitalization equity indices from 46 countries. First, we examine the validity of the random walk hypothesis in individual markets using conventional, rankand sign-based variance ratio tests. Second, we employ multidimensional scaling and clustering techniques to examine the interdependence of variance ratios among equity markets. The empirical findings suggest that multivariate analysis of variance ratios provide significant insights that single variance ratio tests fail to capture. In particular, the results indicate that developed and emerging markets have become considerably more integrated over time. JEL Classification: G14; G15 Key-words: Random walk hypothesis; Variance ratio tests; Cluster analysis; Multidimensional scaling; Emerging and developed equity markets.