Breaking Down Stock Market Volatility

This paper analyzes the link between long-term stock market volatility and macroeconomic variables through the GARCH-MIDAS model. By using daily stock returns and monthly macroeconomic variables, the results show that the long-term volatility is strongly influenced by the realized variance, producer price index and industrial production. Specifically, an increase in inflation leads to an increase in the long-term volatility, whereas changes in industrial production have the opposite effect on this component. These results are confirmed and validated by the new GAS-MIDAS and GARCH-AMIDAS models in terms of an out-of-sample forecasting exercise. The GARCH-AMIDAS model outperforms the other component models in the short-run, while the GAS-MIDAS model works best in the long-run.