Purpose This study investigates how global trade-policy shocks, including tariff revisions and export restrictions, generate cross-border financial contagion through informational rather than purely capital-based channels. Design/methodology/approach A Brown and Warner (1985) event-study is combined with a hierarchical Bayesian model using daily returns from fifteen economies (2023–2025). AI-derived sentiment and sentiment variance from policy announcements capture narrative tone and uncertainty, with governance and macro controls explaining heterogeneity. Findings Trade-policy announcements produce statistically significant cumulative abnormal returns within short event windows. Developed markets exhibit relatively rapid adjustment, whereas emerging markets experience deeper and more persistent negative reactions. After adjusting for global benchmarks, contagion networks become denser, suggesting that synchronization is driven by shared interpretation of policy signals. Stronger institutional governance moderates volatility, while weaker governance amplifies uncertainty. Negative AI-derived sentiment significantly depresses returns, and higher sentiment variance reflects elevated disagreement among market participants. Originality/value The study integrates hierarchical Bayesian inference with large-language-model sentiment, offering a probabilistic framework for information-driven contagion.
Rasyidi Faiz Akbar (Fri,) studied this question.