Abstract:
In this paper, we employ the generalized autoregressive conditional heteroscedasticity-mixed data sampling (GARCH-MIDAS) framework to forecast the daily volatility of state-level stock returns in the United States (US) based on structurally decomposed four monthly oil shocks associated with oil supply, global economic activity, oil consumption and oil inventory. We find that over the daily period of (February) 1994 to (December) 2022 and various forecast horizons, in 46 out of the 50 states, the GARCH-MIDAS model with at least one oil shock can outperform the benchmark, i.e., the GARCH-MIDAS-Realized Volatility (RV), with 24 states depicting the importance of all the four shocks. In general, oil market-specific shocks, whether supply or demand, tend to matter more than a global economic impact driving the oil market in forecasting volatility of regional stock returns across with better forecasting performances related to states with higher CO2 emissions based on underlying energy consumption data. Our findings have important implications for investors and policymakers, with the observations for the former group depicted by an analysis of economic significance, i.e., utility gains.