基于动态因子模型的上证主板A股协同波动研究

Research on the Coordinated Volatility of Shanghai Mainboard A-Shares Based on the Dynamic Factor Model

摘要: 本文以2018年3月至2026年2月上证主板92只代表性股票的月度对数收益率作为研究对象，用动态因子模型来分析A股协同波动和风险传导。实证结果表明，前三个主成分共解释了样本真实波动方差的36.76%。向量自回归和脉冲响应分析表明因子之间存在着明显的跨期溢出效应，在本样期内周期因子对于市场因子有着很强的负向领先作用，这一结果与A股市场板块轮动中的“存量博弈”特征较为吻合。另外模型还找到了一个与市场、周期无关的波动率因子，给非平稳资产定价和动态对冲提供量化的依据。

Abstract: In this article, we use the monthly logarithm return of 92 representative stocks listed in SSE Main Board for March 2018-February 2026 to build a Dynamic factor model to study co-movement and risk transmission inside China’s A-shares. According to the empirical results, the first three PCs explain 36.76% of the sample true volatility variance. Vector autoregression and impulse response analysis indicate significant intertemporal spillover effects among the factors. Within our sample period, the cyclical factor exhibits a strong negative leading effect on the market factor, a result that is broadly consistent with the “zero-sum game” feature often observed in A-share sector rotation. And there is a “volatility” factor that’s not connected to markets or cycles giving us some numbers behind non-stationary assets and moving hedges.