Abstract:
In this study, we examine the predictive value of tail risks for oil returns using the longest possible data available for the modern oil industry, i.e., 1859–2020. The Conditional Autoregressive Value at Risk (CAViaR) of Engle and Manganelli (2004) is employed to generate the tail risks for both 1% and 5% VaRs across four variants (adaptive, symmetric absolute value, asymmetric slope and indirect GARCH) of the CAViaR with the best variant obtained using the Dynamic Quantile test (DQ) test and %Hits. Overall, our proposed predictive model for oil returns that jointly accommodates tail risks associated with the oil market and US financial market improves the out-of-sample forecast accuracy of oil returns in contrast with a benchmark (random walk) model as well as a one-predictor model with only its own tail risk. Our results have important implications for academicians, investors and policymakers.
