Does the macroeconomy matter to market volatility? Evidence from US industries

Abstract

The paper employs a generalized autoregressive conditional heteroskedasticity-mixed data sampling (GARCH-MIDAS) model to examine the relationship between macroeconomic conditions and US stock market volatility at the industry level, with three main findings. First, some macroeconomic factors, such as the term spread, the housing starts, the National Activity Index, the change in unemployment rate, and the default rate, have noticeable effects on industry volatility. Second, out-of-sample test results show that the GARCH-MIDAS model with macroeconomic series helps improve forecasting performance mainly in the consumer staples sector, but its performance is comparable to that of the GJR-GARCH(1,1) model in other sectors. Finally, on average, the term spread contributes the most to fluctuations during expansions, whereas the default rate assumes the most significant role during recessions. As for quantitative easing, although it seems to reduce the scale of stock volatility and to cause the volatilities across sectors associated with macroeconomic series to reach the values observed in expansions, the volatility attributed to the default rate is approaching its average level in recession periods.