We apply machine learning techniques to predict international stock returns using firm characteristics. Market-specific training is important, as neural network models (NNs) achieve stronger results when they are trained in each market separately than in a global model trained with U.S. data. NNs outperform linear models in predicting stock return rankings and forming profitable portfolios. In contrast, regression trees underperform linear models when the number of observations is low. We also show that adding variables constructed from U.S. firm characteristics, which may contain information beyond the characteristics of international stocks, further enhances the return predictability of market-specific NNs.

Mutual funds investing in illiquid corporate bonds actively manage Treasury positions to buffer redemption shocks. This liquidity management practice can transmit non-fundamental fund flow shocks onto Treasuries, generating excess return volatility. Consistent with this hypothesis, we find that Treasury excess return volatility is positively associated with bond fund ownership, and this pattern is more pronounced among funds conducting intensive liquidity management. Causal evidence is provided by exploiting the U.S. Securities and Exchange Commission’s 2017 Liquidity Risk Management Rule. Evidence also suggests that the COVID-19 Treasury market turmoil was attributed to intensified liquidity management, an unintended consequence of the 2017 Liquidity Risk Management Rule.

There is causal evidence that mortgage credit expansions increase house prices. Does an expansion of margin lending increase stock prices? Because unconstrained arbitrageurs are more important for pricing stocks than homes, the impact is not obvious. Tests are limited because sizable shocks to margin lending are rare. We examine a major Chinese margin-lending expansion between 2010 and 2015. Institutional holding, regression discontinuity, and event study evidence—exploiting the rollout of margin lending across stocks—shows that arbitrageurs anticipated and bought in advance of a significant causal effect of credit. We develop a model to rationalize our findings. Our estimates suggest that margin debt contributes to stock market fluctuations.