节点文献

证券投资基金业绩持续性研究

Study on Mutual Funds Performance Persistence

【作者】 李德辉

【导师】 方兆本;

【作者基本信息】 中国科学技术大学 , 管理科学与工程, 2006, 博士

【摘要】 证券投资基金的业绩持续性是指业绩优秀的基金在其后一段时间内更倾向于继续保持优秀,而业绩差的基金继续表现出差的业绩,也就是人们常说的“强者恒强,弱者恒弱”。从本质上看,基金业绩持续性是考察基金历史业绩对未来业绩是否存在一定程度的揭示作用,其思想与有效市场假说相抵触,因此被视为金融市场异常现象。基金业绩持续性研究主要包括两大问题:其一,基金业绩持续性现象是否存在?其二、基金业绩持续性来源的解释。 首先,本文对基金业绩持续性研究进行了细致的梳理,主要内容包括: 1.针对“基金业绩持续性现象是否存在”,本文从四个角度进行了综述,分别是业绩持续性的期限长短与稳定性、各类基金的业绩持续性、业绩差的基金较业绩好的基金持续性更强、能否利用业绩持续性盈利。 2.针对“基金业绩持续性来源的解释”,本文介绍了现有的四个假说,分别是数据质量假说、股票惯性假说、基金经理才能假说、费率税赋等假说。 3.针对基金业绩持续性的检验,本文首先详细介绍了两种主流的方法——列联表、横截面回归,并对相关性检验、Kolmogorov-Smimor检验等方法进行了简介,最后讨论了在对现有检测工具基础上可能的改进。 其后,本文提出新的检测方法进行实证研究,主要内容、贡献和创新包括: 1.引入新的检测方法——扫描统计量,证实部分基金存在持续性,并给出了持续性强度的上限。 本文引入扫描统计量后,从事件聚集性的角度出发,使得单只基金业绩持续性检测成为可能。其必要性有三:传统方法的共性是从基金行业层面整体检测持续性;根据以往文献,业绩持续性主要体现在业绩优劣两端;投资者一般不大可能买卖一篮子基金,对单只基金更为关心。在利用最长链模型、基于贝努利过程的其它扫描统计量模型,对中美基金业绩持续性进行实证分析后,证实部分基金存在持续性,并给出了持续性强度的上限。

【Abstract】 Persistence in mutual fund performance is well documented in financial literature. It refers to a fund manager’s ability to consistently deliver investment return above or below a benchmark return, and is also called hot hand and icy hand phenomenon sometimes. The question of whether there is persistence in mutual funds has long been important, both for academic research and for practitioner, since the efficient market hypothesis implies that past performance is no guide to future performance. Most researches focus on the existence of persistence and explanation for persistence.This article first provides a comprehensive review of fund performance persistence.1. As for the hypothesis of the existence of persistence, four issues are introduced, such as the length of persistence and its stability, the persistence of non-stock funds, underperformance, and investment strategy using fund persistence.2. As for the explanation for persistence, four hypotheses are introduced, such as data quality hypothesis, stock momentum hypothesis, fund manager ability hypothesis, fees and taxation hypothesis.3. For the issue of persistence detecting, the method of cross-sectional regression and contingency table are traditionally used. Rank correlation approach and Kolmogorov-Smirnor test are also introduced in this article. In addition, potential improvements of these four methods are discussed.In the second part of article, we provide a new method for persistence detecting. The main contribution and innovation are:1. Provides a new method for persistence detecting, and confirms the existence of persistence among some funds, the degree of persistence is also studiedTraditional methods are ineffective when each individual fund is concerned. In this article we find that the method of scan statistics can be used to deal with this task efficiently. Up to now, scan statistics is a brand new method in persistence research.

  • 【分类号】F830.91;F224
  • 【被引频次】16
  • 【下载频次】1720
  • 攻读期成果
节点文献中: 

本文链接的文献网络图示:

本文的引文网络