基于分位数回归的基金风格分析与业绩评价

摘要

基金风格分析旨在判定基金投资取向、识别基金投资风险、评估基金投资业绩,对于机构投资者与个人投资者都具有重要意义。基于分位数回归,对Fama-French三因子模型进行了拓展,并将其应用于基金风格分析,进一步讨论了基金风格漂移、基金经理业绩评价等主题。运用Fama-French三因子模型,从2010年12月前成立的704只基金中筛选出＂长盛债券＂和＂易方达策略＂两只基金,对比了均值回归与分位数回归的异同,认为两种方法互相补充,共同完善基金风格分析效果。%The purpose of fund style analysis is to judge the fund investment orientation,identify the risk of the fund investment and evaluate fund investment performance,which is very important to the institutional investors and the individual investors.Based on the quantile regression,we expanding the Fama-French "three factor" model,and apply the method to the fund style analysis.Then,we discuss the themes of fund style drift and the fund manager's performance evaluation.Using Fama-French "three factor" models,we select the two funds named "ChangSheng Zhaiquan" and "YiFangDa Celue" from all 704 funds which founded before December of 2010.Comparing the different results of the mean regression and quantile regression,we think that the two methods complement each other and make the result of fund style analyses better.