收益率厚尾分布、特质性波动与股票价格行为  

Fat-tailed Distribution of Stock Returns,Idiosyncratic Volatility,and Stock Price Behavior

作　　者：王春峰[1,2] 姚守宇 房振明[2] WANG Chun-feng1,2 , YAO Shou-yu1,2 , FANG Zhen-ming2(1. College of Management and Economics,Tianjin University,Tianjin 300072,China; 2. Financial Engineering Research Center,Tianjin University,Tianjin 300072 ,Chin)

机构地区：[1]天津大学管理与经济学部,天津300072 [2]天津大学金融工程研究中心,天津300072

出　　处：《系统工程》2017年第12期1-14,共14页Systems Engineering

基　　金：国家自然科学基金资助项目(0401210010)

摘　　要：作为当前资产定价领域研究的热点问题之一,特质性波动与股票横截面收益间的正负向关系一直以来争议较大。不同于前人学者,在对特质性波动采用了更加科学的估计方法后,本文主要从收益率厚尾分布视角出发,在考虑到所有收益信息真实表达的前提下,全面考察了特质性波动率与股票横截面收益间的关系。研究表明,特质波动率与股票横截面收益率间的关系是存在分位数依赖的,"特质波动率之谜"仅存在于收益率较低的样本股票中。在收益率较高的股票中,特质波动率与股票收益呈正向关系,且随着股票收益率分位数的升高这种正向关系显著增强。这个结论与以往研究均不相同,本文也试图从投资者追捧所引致的媒体及舆论监督效应为这种复杂的关系提供了一种可能的合理统一解释。As one of the hot topics in the current asset pricing field,the positive and negative relationship between idiosyncratic volatility and cross-sectional stock returns has been controversial.Different from previous literatures,after adopting a more scientific estimate method of idiosyncratic volatility,this paper mainly focuses on the perspective of fat-tailed distribution of stock returns,aiming to take the true expression of all income information into account,making a comprehensive analysis for the relationship of idiosyncratic volatility and cross-sectional stock returns.The result demonstrates that the idiosyncratic volatility and stock return’s relation is quantile dependent.Idiosyncratic volatility puzzle only exists in sample stocks with lower-return.In higher-return stocks,idiosyncratic volatility and stock return show apositive relationship.Further more,with the stock return quantile increases,this positive relationship is significantly enhanced.This conclusion is not the same with the previous researches and we also tried to provide a reasonable and unified explanation for this complex relationship from the media supervision effect caused by investor attention.

关 键 词：特质波动率 股票横截面收益 厚尾分布 分位数回归 

分 类 号：F830[经济管理—金融学]