Causal structure of international stock markets

This article examines the causal structure of international stock markets using causal discovery algorithms across a six-market system. Unlike methods that infer connections without an assumption of cause and effect, causal discovery methods strive to uncover genuine causal relationships directly from observational data. Our findings reveal significantly fewer causal links compared to previous studies. Notably, during recessions, information circulates so rapidly that its impact rarely extends beyond a single day. However, in other periods, information from the previous day continues to affect returns, positioning the U.S. stock market as a leading market. Leveraging the identified causal relationships, we backtest simple cross-border timing strategies that achieve significant improvements in both risk and return relative to buy-and-hold benchmarks. These findings point to a previously unexplored class of trading signals for cross-border market timing.