Predicting financial market stress with machine learning

Using newly constructed market condition indicators (MCIs) for three pivotal US markets (Treasury, foreign exchange, and money markets), we demonstrate that tree-based machine learning (ML) models significantly outperform traditional timeseries approaches in predicting the full distribution of future market stress. Through quantile regression, we show that random forests achieve up to 27% lower quantile loss than autoregressive benchmarks, particularly at longer horizons (3–12 months). Shapley value analysis reveals that funding liquidity, investor overextension and the global financial cycle are important predictors of future tail realizations of market conditions. The MCIs themselves play a prominent role as well, both in the same market (self-reinforcing dynamics within markets) and across markets (spillovers across markets). These results highlight the value of ML in forecasting tail risks and identifying systemic vulnerabilities in real time, bridging the gap between highfrequency data and macroeconomic stability frameworks.