An Information-Based Index of Uncertainty and the predictability of Energy Prices

We develop an index of uncertainty, the COVID-19 induced uncertainty (CIU) index, and employ it to empirically examine the vulnerability of energy prices amidst the COVID-19 pandemic using a distributed lag model that jointly accounts for conditional heteroscedasticity, autocorrelation, persistence, and structural breaks, as well as day-of-the-week effect. The nexus between energy returns and uncertainty index is analyzed, using daily price returns of eight energy sources (Brent oil, diesel, gasoline, heating oil, kerosene, natural gas, propane, and WTI oil) and four news/information-based uncertainty proxies [CIU, EPU, Global Fear Index (GFI) and VIX]. The CIU and alternative indexes are used, respectively for the main estimation and sensitivity analysis. We show the outperformance of CIU over alternative news uncertainty proxies in the prediction of energy prices. News (aggregate) and bad news are found to negatively and significantly impact energy returns, while good news has a significantly positive impact. Imperatively, energy variables lack hedging potentials against the uncertainty occasioned by the COVID-19 pandemic, while we find no strong evidence of asymmetry. Our results are robust to the choice of news variables, forecast horizons employed, with likely sensitivity to energy prices.