摘要
本文在投资组合难以充分分散化的前提下,综合考虑系统偏度(协偏度)和特质偏度的影响,将风险资产的统计学偏度纳入无套利分析理论框架中,推导出风险资产的期望收益率受其方差的正向影响,受其偏度的负向影响。为了克服样本极端值对偏度测度的影响,本文通过混频分位数回归先估计个股收益率的条件分位数,再借助Cornish-Fisher展开式来间接估计偏度,这是一种稳健的估计方法。基于构造的个股月度稳健偏度指标建立月度偏度因子,并在学术界已有的两大经典因子模型——CAPM单因子模型和Fama-French三因子模型中,通过时间序列检验、横截面检验和抗干扰检验,发现偏度因子可以改进既有因子模型的定价效果,是资产定价的重要因子。
Under the premise that it is difficult to fully diversify the portfolio,this paper comprehensively considers the influence of systemic skewness(co-skewness)and idiosyncratic skewness,brings the statistical skewness of risk assets into the theoretical framework of no arbitrage analysis,and deduces that the expected return of risk assets is positively affected by its variance and negatively affected by its skewness.In order to overcome the influence of extreme value on skewness measure,this paper estimates the conditional quantile of individual stock return by mixing quantile regression,and then estimates the skewness indirectly by Cornish Fisher expansion,which is a robust estimation method.Based on the constructed monthly robust skewness index of individual stocks,the monthly skewness factor is established.In the existing two classical factor models——CAPM single factor model and Fama-French three factor model,through time series test,cross-sectional test and anti-interference test,it is found that the skewness factor can improve the pricing effect of the existing factor model and is an important factor in asset pricing.
作者
夏仕龙
Xia Shilong(School of Economics,Sichuan University,Chengdu 610065,China)
出处
《数量经济研究》
2022年第3期117-133,共17页
The Journal of Quantitative Economics
基金
2019年度教育部人文社会科学研究青年基金西部和边疆地区项目“不确定环境下我国宏观经济政策协调搭配研究”(19XJC790015)的资助
关键词
偏度
资产定价
因子模型
无套利原理
混频分位数回归
Skewness
Asset Pricing
Factor Model
No Arbitrage Principle
Mixing Quantile Regression
