公募FOF中常用量化方法及相关模型的实证研究

发布时间：2018-03-15 10:17

本文选题：FOF　切入点：大类资产配置　出处：《山东大学》2017年硕士论文　论文类型：学位论文

【摘要】：中国证监会在经过大概三个月的征求意见之后,在2016年9月23日正式颁布并实施了《公开募集证券投资基金运作指引第二号——基金中基金指引》,从此公募基金正式迎来FOF新时代,FOF的出现极大地丰富了我国金融产品体系,为投资者提供了一个新的投资渠道,FOF不仅能替普通投资者精选优质基金,还能够对风险进行二次分散,同时为更多投资者实现全球资产配置,满足流动性宽裕的高净值用户的需求,总之,FOF的出现极大推动了我国公募基金行业的发展。随着《指引》的发布,FOF相关业务正在稳步推进,但相关理论研究还比较落后。本文正是在这样的背景下,对FOF组合构建中所涉及到的具体量化模型进行理论介绍与实证研究。FOF组合构建的核心大概可以分为两步,第一步是进行大类资产配置,第二步是对每一类资产的子基金进行筛选,即基金评价内容。对于第一步,在对FOF进行大类资产配置研究时,我们主要讨论的模型有以下三种,如Markowitz的均值-方差模型、B-L模型(BlackLitterrman)以及风险平价模型,在分析了构型的构建方法之后,基于最新数据从实证的角度对三种模型进行了比较分析。至于基金评价相关内容,我首先介绍了基金评价过程中常用的一些量化指标,如基于CAPM模型的夏普比率、詹森α、特雷诺指数、超额收益率以及信息比率等,利用这些指标可以粗略地对基金进行排序与筛选,并构造相应的基金组合。随着近年来研究的深入,研究者们逐渐发现,投资组合的业绩来源中包含一个十分重要的因素,那就是"风格",后来便衍生出一系列基于风格归因的基金业绩评价方法。在对风格进行归因分析时,根据模型选择数据的不同,主要有两种方法,一种是基于收益率时间序列回归法,一种是居于持仓明细回归法,因为公募基金每日公开披露净值,净值时间序列数据相对易获得,故本论文主要使用的数据是前者。在本文后半部分详细地介绍了基于风格归因的量化模型,如Fama-French三因子模型,Carhart四因子模型等等,并对Carhart四因子模型进行了实证分析,力求比较清楚地解释FOF组合构建中子基金的筛选过程,希望能给组合管理者以及基金研究员一定的参考作用。
[Abstract]:After about three months of soliciting opinions from the China Securities Regulatory Commission, In September 23rd 2016, the "guidelines for the Operation of Public offering Securities Investment funds" No. 2 was formally promulgated and implemented. From then on, the public offering funds formally ushered in the new era of FOF, which greatly enriched the financial products system of our country. It provides a new investment channel for investors. FOF can not only select quality funds for ordinary investors, but also redistribute risks and achieve global asset allocation for more investors. In short, the emergence of FOF has greatly promoted the development of the public offering fund industry in China. With the release of the guidelines, FOF related businesses are advancing steadily. However, the relevant theoretical research is still relatively backward. Under this background, the theoretical introduction and empirical study of the specific quantitative model involved in the construction of FOF combination can be divided into two steps. The first step is to allocate assets in large categories, and the second step is to screen the subfunds of each type of assets, that is, the content of fund evaluation. For the first step, we mainly discuss the following three models when we study the allocation of large categories of assets in FOF. For example, the mean-variance model of Markowitz, BlackLitterrman model, and the risk parity model, after analyzing the construction methods of the configuration, the three models are compared and analyzed based on the latest data from an empirical point of view. First of all, I introduce some commonly used quantitative indicators in the process of fund evaluation, such as Sharp ratio based on CAPM model, Jason 伪, Traineau index, excess return rate and information ratio, etc. With these indicators, funds can be roughly sorted and selected, and corresponding fund portfolios can be constructed. As recent studies have deepened, researchers have come to discover that the source of portfolio performance includes a very important factor. That is "style", and then a series of performance evaluation methods based on stylistic attribution have been developed. In the attribution analysis of style, there are two main ways to select different data according to the model. One is based on the return time series regression method, the other is based on the position breakdown regression method, because the public offering funds disclose their net worth publicly every day, and the time series data of net value are relatively easy to obtain. In the second half of this paper, the quantitative model based on style attribution, such as Fama-French three-factor model, Carhart four-factor model and so on, is introduced in detail, and the Carhart four-factor model is analyzed empirically. This paper tries to explain clearly the selection process of the FOF portfolio to construct the neutron fund, hoping to give some reference to the portfolio manager and the fund researcher.
【学位授予单位】：山东大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：F832.51

【相似文献】