In this paper, we predict realized volatility of stock return by utilizing time-varying risk aversion based on a simple linear autoregressive model. Our in-sample results suggest that time-varying risk aversion have significant impact for stock return volatility. In terms of out-of-sample forecasting performance, the empirical results indicate that the incorporation of time-varying risk aversion in the benchmark model can yield more accurate stock return volatility forecasts. Notably, the out-of-sample forecasting results confirm that our conclusions are robust when we apply alternative lag orders and alternative prediction evaluation periods. Finally, we study links between the prediction ability of time-varying risk aversion and the volatility of other stock indices and two kinds of crude oil, and find that the new predictor can effectively strengthen forecasting performance in most case. In view of the importance of volatility risk in the asset pricing process, our research is of great significance for financial asset participants.

在本文中，我们利用基于简单线性自回归模型的时变风险规避来预测股票收益的已实现波动率。我们的样本内结果表明，时变风险规避对股票收益波动有显着影响。在样本外预测性能方面，实证结果表明，将时变风险规避纳入基准模型可以产生更准确的股票收益波动率预测。值得注意的是，样本外预测结果证实，当我们应用替代滞后顺序和替代预测评估期时，我们的结论是稳健的。最后，我们研究了时变风险规避的预测能力与其他股票指数和两种原油的波动性之间的联系，并发现在大多数情况下，新的预测器可以有效地增强预测性能。鉴于波动风险在资产定价过程中的重要性，我们的研究对金融资产参与者具有重要意义。