Estimating and Evaluating Value-at-Risk forecasts based on Realized Variance: Empirical Evidence from ICE Brent Crude Oil Futures

Abstract

In this article we examine the properties of estimates of realized volatility at various intra-daily sampling frequencies for Brent Crude oil futures traded at the IntercontinentalExchange (ICE). The estimates of realized volatility are subsequently modeled and forecasted to predict day-ahead Value-at-Risk. We suggest a new method for evaluating the whole distribution of the variance forecasts by examining a simple PP-plot. Our results show that the distribution of ICE Brent Crude oil futures returns standardized with predicted volatility for the next trading day is very close to Gaussian, which significantly simplifies the Value-at-Risk estimation. Finally, our results suggest that the ideal choice of sampling frequency is between one and ten minutes for this commodity.