Given the impact of changes in the price of crude oil on expected cash flows, rate of interest and inflation as well as investors’ increasingly utilization of crude oil as a financial asset, it is reasonable to suspect that we could substantially improve the accuracy of equity return volatility forecasts by conditioning on the price of crude oil. To evaluate this hypothesis, we carry out an out-of-sample forecasting study for monthly aggregate equity return realized volatility using an autoregressive benchmark and alternative specifications that employ the price of crude oil. Our approach also takes into account the possibility that relative predictive performance changes over the out-of-sample. Several interesting findings are unraveled: First, there is evidence of threshold nonlinearity in forecasting. Second, this statistical evidence also results in economic gains. Third, the evidence of predictability is strong at long forecast horizons while somewhat weak one-month ahead. Fourth, the predictive power associated with certain nonlinear crude oil price measures concentrates from the onset of the Great Recession and onwards.

鉴于原油价格变化对预期现金流量，利率和通货膨胀以及投资者越来越多地利用原油作为金融资产的影响，我们有理由怀疑我们可以大幅提高权益的准确性。通过以原油价格为条件来返回波动率预测。为了评估此假设，我们使用自回归基准和采用原油价格的替代规范，对每月总股本收益实现的波动率进行了样本外预测研究。我们的方法还考虑了相对预测性能在样本外变化的可能性。揭示了一些有趣的发现：首先，有证据表明在预测中存在阈值非线性。第二，这种统计证据也可以带来经济收益。第三，在很长的预测期内，可预测性的证据很强，而未来一个月的预测性却很弱。第四，与某些非线性原油价格测度相关的预测能力集中于大萧条的开始和以后。