混频模型下经济政策不确定性与股市波动研究

Study on Economic Policy Uncertainty and Stock Market Volatility under the Mixed-Frequency Model

摘要: 实证研究表明，经济政策不确定性(EPU)对股市波动具有重要影响，而股指收益率分布存在显著的时变高阶矩特征。传统的波动率模型难以有效捕捉这些特征，往往导致对股市波动的低估或高估。为此，本文以沪深300指数和中国经济政策不确定性指数(CEPU)为研究对象，基于Gram-Charlier展开式(GCE)构建包含时变高阶矩的GARCH-MIDAS-SK模型，并引入CEPU指数，构建了GARCH-MIDAS-SK-CEPU模型，旨在分析经济政策不确定性对中国股市波动的影响。实证结果表明，股指收益率分布呈现出明显的时变高阶矩特征，CEPU对股市波动率具有显著的正向作用。引入CEPU指数和时变高阶矩特征能够提升模型的预测性能。此外，DM检验结果显示，本文构建的模型在股市波动率预测中优于传统基准模型，为股市波动率预测提供了更有效的工具。

Abstract: Empirical research shows that economic policy uncertainty (EPU) has a significant impact on stock market volatility, while stock return distributions exhibit significant time-varying higher moment characteristics. Traditional volatility models often fail to effectively capture these features, leading to underestimation or overestimation of stock market volatility. To address this issue, this paper uses the CSI 300 Index and the Chinese Economic Policy Uncertainty Index (CEPU) as research objects, and constructs a GARCH-MIDAS-SK model incorporating time-varying higher moments based on the Gram-Charlier expansion (GCE). Furthermore, the CEPU index is introduced to construct a GARCH-MIDAS-SK-CEPU model, aiming to analyze the impact of economic policy uncertainty on stock market volatility in China. The empirical results show that stock return distributions significantly exhibit time-varying higher moment characteristics, and CEPU has a significant positive effect on stock market volatility. Incorporating the CEPU index and time-varying higher moments improves the model’s predictive performance. Moreover, the DM test results demonstrate that the proposed model outperforms traditional benchmark models in stock market volatility prediction, providing a more effective tool for stock market volatility forecasting.