Economic Policy Uncertainty and Long-Run Stock Market Volatility and Correlation

Abstract

We study how uncertainty affects stock market volatility and correlation. We model the realized volatility and correlation for both stocks and bonds using alternative measures and sources of uncertainty and control for risk aversion and general macroeconomic conditions. Our results indicate a strong predictive ability of the risk-aversion variable and economic-uncertainty measures constructed from financial information, at both the daily and monthly frequencies. We find leverage effects at the daily but not the monthly frequency and that the effect of risk aversion on stock volatility increases when returns are negative. We find also that higher risk aversion and uncertainty increase the probability of flight to safety. The international analysis shows that US investors benefit from using information about risk aversion and uncertainty when forming their portfolios. However, including too many predictors may lead to overfitting and worse out-of-sample results.