ABSTRACT Recent disruptive events have renewed interest in whether U.S. excess returns predict movements in global equity markets. This study investigates the asymmetric predictive effects of positive versus negative U.S. excess returns and assesses whether economically motivated constraints enhance forecasting performance. We find that, although the U.S. remains an important driver of international markets, its dominance has weakened. Predictability is markedly asymmetric: negative U.S. excess returns exhibit stronger and more persistent predictive power, and imposing economic constraints—specifically, the CT constraint that truncates negative return forecasts at zero and the PTV constraint that limits forecasts by bounding the implied Sharpe ratio within a plausible range—further improves forecast accuracy. Cumulative squared forecast error evidence shows that U.S. spillovers intensify during recessions, and the predictive advantage of negative U.S. returns remains pronounced at longer horizons, with the PTV constraint delivering additional gains in long‐horizon forecasts. Importantly, these statistical improvements translate into economic value: in a mean–variance asset‐allocation exercise, constrained forecasts generate economically meaningful utility gains relative to benchmark strategies. Finally, the predictive power of U.S. returns is substantially weaker in BRICS markets than in developed economies. Overall, our results indicate that global market interconnectedness is evolving, highlighting the importance of international diversification for policymakers concerned with cross‐border shock transmission.
Guo et al. (Mon,) studied this question.