Algorithmic Intermediation and the International Transmission of U.S. Monetary Policy
创建于 更新于
摘要
This paper develops a two-region macro-financial framework to analyze how algorithmic and AI-driven fund management shapes the international transmission of U.S. monetary policy to emerging markets. It argues that instability arises not from intermediation per se but from the similarity of models across funds: when algorithms share signals and forecast errors, their trades reinforce one another and amplify capital-flow responses during stress; diverse models offset errors and stabilize flows. The empirical test using equity portfolio flows to nineteen EMs from 2000–2024 supports a state-dependent picture: algorithmic herding amplifies outflows after U.S. shocks only in high-volatility regimes, while faster adjustment alone does not produce a comparable effect. The amplification is robust across robustness checks but hinges on defining stress via the VIX and on keeping the 2008–09 crisis in the sample. Policy implication: emphasis should be on preserving model diversity rather than restricting non-bank intermediation. [page::1][page::5][page::9][page::11][page::15]
