Stock Market Volatility Predictability: A Transfer Entropy‐Determined Model‐Switching Strategy
提出一种基于转移熵的HAR-TEDMS模型，通过检测石油与股票隐含波动率指数间的信息传递来动态切换模型状态，显著提升股市波动率预测精度，对投资者有参考价值。
ABSTRACT This article proposes a Transfer Entropy‐Determined Model‐Switching (HAR‐TEDMS) strategy within the HAR‐RV framework to improve the predictive accuracy of stock market volatility. The core mechanism of the HAR‐TEDMS model is based on transfer entropy, which is used to identify whether the market is in a state of dependence or independence by examining the significance of information transmission between the oil implied volatility index (OVX) and the stock implied volatility index (VIX). This mechanism enables the model to dynamically switch between incorporating interactive or independent information from OVX and VIX, thereby effectively adapting to different market states. Empirical results reveal the superior forecasting performance of the HAR‐TEDMS model across different forecasting horizons. Furthermore, we confirm that the predictive ability of the HAR‐TEDMS model is primarily reflected in its capability to capture asymmetric information transmission and its adaptability to turbulent environments. This novel HAR‐TEDMS model enhances the understanding of information transmission mechanisms across financial markets, underlining its potential value in guiding investment strategies.