Abstract
This article attempts to characterize the pattern of information flows between the stock markets by determining mean and variance causal relationships. A two-step procedure proposed by Cheung and Ng (Citation1996) is used. Stock market returns are specified as Autoregressive–Generalized Autoregressive Conditional Heteroskedasticity (AR–GARCH) models with Monday and Friday effects. Stock markets of our sample are chosen to analyse the main causes of information flows documented in the literature: linkage between economic fundamentals and the time lag between the stock markets' opening hours. Results provide evidence of nonlinear causality between stock markets, even when linear Granger causality is rejected. Causality seems to be attributed to both economic linkage and time lag between market openings.