Abstract:
This paper attempts to investigate the information content of risk reversal in estimating VaR of oil futures. Using CAViaR PlugIn models, we incorporated risk reversal into CAViaR models. We tested the performance of our model with daily returns of WTI crude oil futures. Our simulation results shows that CAViaR risk reversal PlugIn model significantly improves the performance of standard CAViaR models in terms of out sample hit percentage. We also tested the properties of VaR violation series. Coverage tests results indicate that our newly proposed model not only beats benchmark Riskmetrics and CAViaR models in out sample VaR forecast performance but also produce VaR violation series with properties of true coverage and randomness.
