管理评论

›› 2018, Vol. 30 ›› Issue (4): 20-32.

• 经济与金融管理 • 上一篇    下一篇

基于CVaR的基金业绩测度研究

黄金波1, 李仲飞2, 丁杰1   

  1. 1. 广东财经大学金融学院, 广州 510320;
    2. 中山大学管理学院, 广州 510275
  • 收稿日期:2016-01-22 出版日期:2018-04-28 发布日期:2018-04-22
  • 通讯作者: 李仲飞(通讯作者),中山大学管理学院教授,博士生导师,博士;丁杰,广东财经大学金融学院讲师,博士。
  • 作者简介:黄金波,广东财经大学金融学院副教授,博士
  • 基金资助:

    国家自然科学基金创新研究群体项目（71721001）；国家自然科学基金重点项目（71231008）；国家自然科学基金青年项目（71603058）；教育部人文社会科学研究青年基金项目（16YJC790033；17YJC790023）；广东省自然科学基金项目（2014A030312003；2016A030313656；2014A030310428；2015A030313629）；广东省哲学社会科学规划项目（GD15YYJ06；GD15XYJ03）。

Research on Fund Performance Measure with CVaR

Huang Jinbo1, Li Zhongfei2, Ding Jie1   

  1. 1. School of Finance, Guangdong University of Finance & Economics, Guangzhou 510320;
    2. Sun Yat-Sen Business School, Sun Yat-Sen University, Guangzhou 510275
  • Received:2016-01-22 Online:2018-04-28 Published:2018-04-22

PDF (PC)

506

摘要：

基于条件在险价值（CVaR）建立新的基金业绩测度指标，该指标在理论上拓展了经典的夏普比率。在正态分布下，该指标是夏普比率的增函数，二者对于基金业绩排名是一致的；在非正态分布下，该指标克服夏普比率没考虑高阶矩、不满足随机占优单调性的缺陷，能给出更为合理的基金业绩排名。利用方差法、经验分布法和核估计法对新指标进行估计，蒙特卡洛模拟结果表明，方差法仅在正态分布下有效，在非正态分布下其估计结果存在系统性偏差；同属于非参数方法的经验分布法和核估计方法在任意分布下都具有大样本性质且估计精度相当。最后运用新指标对我国开放式基金的业绩进行测算和排名，结果显示：当各基金的偏度系数和峰度系数差异较小时，夏普比率和新指标给出的基金业绩排名基本一致；而当各基金的偏度系数和峰度系数差异较大时，二者给出的基金业绩排名差异较大，新指标因考虑了高阶矩信息给出的排名更为合理，这与理论预期是一致的。

关键词: 基金业绩测度, 随机占优单调性, 高阶矩, 夏普比率, 条件在险价值

Abstract:

This paper presents a theoretically sound fund performance measure that generalizes the Sharpe ratio. The new performance measure is monotone with respect to stochastic dominance and consistently accounts for mean, variance and higher moments of the return distribution. It is equivalent to the Sharpe ratio if returns are normally distributed, while the new measure can produce more reasonable performance rankings than Sharpe ratio under non-normal distribution. We suggest a parametric and two non-parametric estimators for the new performance measure and compare their estimation errors by Monte Carlo simulation. Finally we provide an empirical illustration using mutual funds data in China. The empirical results show that performance rankings induced by the Sharpe ratio and new measure are consistent when funds' higher moments are not different from each other, while performance rankings induced by Sharpe ratio and new measure are inconsistent when differences of funds' higher moments are significant. The new measure is more reasonable since it accounts for the higher moments, which confirms theoretical analysis.

Key words: fund performance measure, monotonic with respect to stochastic dominance, higher moments, Sharpe ratio, conditional value-at-risk