Revisiting the predictive prowess of economic policy uncertainty (EPU) in stock market volatility: GEPU or NEPU?

Abstract

We employ the Generalized Autoregressive Conditional Heteroskedasticity (GARCH) variant of the Mixed Data Sampling (MIDAS) [GARCH-MIDAS] model to revisit the forecasting prowess of economic policy uncertainty (EPU) in the predictability of stock market volatility. We show that both the global economic policy uncertainty (GEPU) and its domestic dimension, namely the Nigerian Economic Policy Uncertainty (NEPU), discourage future investment in the stock market. However, compared to NEPU, we show results that give credence to the relatively higher effect of GEPU on the volatility dynamics of the stock market. We also show the comparative out-of-sample forecasts of the NEPU and GEPU to be episodic in the predictability of volatility in the stock market. The fact that our results are robust across multiple forecast horizons is an indication that events that form the premise upon which a particular dimension of the EPU is measured matter in their forecasting prowess.

Keywords

• C53
• C58. D80