RISK PREMIUM EFFECTS ON IMPLIED VOLATILITY REGRESSIONS
研究发现期权隐含波动率预测实际波动率的偏差源于波动率风险溢价，该溢价由风险中性密度的高阶累积量刻画，调整风险溢价后预测偏差可消除。
Abstract This article provides new insights into the sources of bias of option implied volatility to forecast its physical counterpart. We argue that this bias can be attributed to volatility risk premium effects. The latter are found to depend on high‐order cumulants of the risk‐neutral density. These cumulants capture the risk‐averse behavior of investors in the stock and option markets for bearing the investment risk that is reflected in the deviations of the implied risk‐neutral distribution from the normal distribution. We show that the bias of implied volatility to forecast its corresponding physical measure can be eliminated when the implied volatility regressions are adjusted for risk premium effects. The latter are captured mainly by the third‐order risk‐neutral cumulant. We also show that a substantial reduction of higher order risk‐neutral cumulants biases to predict their corresponding physical cumulants is supported when adjustments for risk premium effects are made.