Economic policy uncertainty as an indicator of abrupt movements in the US stock market

A two regime switching model is developed in an attempt to relate expected US stock market returns to deviations from fundamentals and to Economic Policy Uncertainty (EPU). The analysis is based on monthly data that cover the period from January 1900 to October 2022 and the EPU index is used as an explanatory variable. The findings suggest that the US stock market spends most of the time in a low-volatility regime, periodically switching to a high-volatility regime during times of financial instability. In an attempt to examine the forecasting ability of the model, out-of-sample probabilities of a crash and a boom are estimated recursively. The results provide evidence that our model is able to depict periods of abrupt movements in the US stock market. Finally, the estimated model and the associated probability of a crash are used to develop and evaluate a proposed trading strategy, in order to analyse the financial usefulness of the model. A simple simulation reveals that our trading rule produces statistically significant abnormal returns and manages to outperform the simple buy-and-hold strategy for the period before the Covid-19 crisis.