Idiosyncratic Volatility Puzzle and Its Estimation Model

• 1、School of Economics and Business Administration,Chongqing University,Chongqing 400030
• 2、Innovation Instiute of Corporate Finance and Accounting Governance,Chongqing University 400030

Abstract：Taking the CAPM, Fama-French three-factor, and Fama-French five-factor models as the mean equation, the unconditional standard deviation and the conditional variance of the GARCH and EGARCH model were used as benchmarks to calculate the idiosyncratic volatility of the Chinese stock market, and study the relationship between idiosyncratic volatility and stock returns. The results show that the pricing anomaly of low expected returns caused by high idiosyncratic volatility risks stems from the inconsistency of the idiosyncratic volatility estimation model. Specifically, when using the unconditional standard deviation to calculate idiosyncratic volatility, there is a pricing anomaly. When using conditional variance models such as GARCH and EAGRCH, there is no pricing anomaly. The result is still robust after changing the residual estimation equation and using a series of control variables such as size and liquidity. This research is useful to explain the "Idiosyncratic Volatility Puzzle" that has long plagued the field of asset pricing.

Keywords： Idiosyncratic volatility expected return asset pricing capital markets