我国开放式证券投资基金绩效评价研究

发布时间：2019-05-06 15:48

【摘要】：从我国第一只证券投资基金诞生至今,我国证券投资基金市场经历了近20年的发展,无论是基金数目还是基金规模都实现了飞速的发展,为资本市场的稳定发展提供了坚实基础。但是,相比成熟市场,我国的证券投资基金对于证券市场的控制力还有待提高。要实现证券投资基金的全面发展,核心之一就是要构建一个坚实有效的市场信用体系,而这又与基金的综合评价机制的建立是分不开的,因此本文探索研究一个全面、客观和科学的基金绩效评价体系。在理论研究部分,首先,对国内外有关基金绩效研究的文章进行分析研究,从风险收益的单因素模型、基金管理人能力模型、风险收益多因子模型、多元统计模型、数据包络分析(DEA)等角度出发,分类分析了国内外较为经典的文献资料;其次,对经典的基金绩效指标进行详细评析,本文对特雷诺(Treynor)指数、夏普(Sharpe)指数、詹森(Jensen)指数、索提诺(Sortino)比率和信息比率进行了理论分析。并对计算选股择时指标的理论基础T-M模型、H-M模型、C-L模型、Fama-French三因子模型和Carhart四因子模型进行了分析;最后,通过阐述分析VaR模型、因子分析法和超效率DEA模型,构建本文的实证综合模型“VaR+因子分析+超效率DEA”模型。在实证研究部分,首先,本文改进传统绩效评价方法中的风险测度方法,引入VaR风险测度模型来测度样本基金的风险,通过对不同时期的样本基金进行风险测度可知,样本基金在震荡期、强势期和弱势期的周度VaR均值分别为0.0527、0.1137和0.0724,其中最高的是强势期,其次是弱势期,震荡期的最低。此外,利用得到的VaR值改进Sharpe指数,通过改进的Sharpe指数对样本基金进行绩效排名;其次,采用因子分析法对本文所选取的指标体系进行处理,经过因子分析的处理得到风险测度指标、风险收益测度指标、选股择时测度指标、成本费用测度指标和基金结构测度指标共五大类投入指标,在因子分析法中通过计算可得到样本基金的因子得分,根据因子得分对样本基金的绩效进行分类评价;再次,将因子分析得到的投入产出数据导入超效率DEA模型,通过计算效率值对样本进行绩效评价,根据效率计算结果可知,导致基金无效的原因主要是技术无效率,从投影分析可知,无效基金主要需要从结构因素和选股择时因素方面入手来提高基金的投资效率。根据综合绩效排名分析表明,基金的类型和投资风格对综合绩效的影响不显著;最后,对样本基金不同时期的绩效排名进行持续性分析。实证结果显示,从震荡期到强势期,基金绩效表现微弱的持续性。从强势期到弱势期,基金绩效不存在绩效持续性,反而表现出较强的绩效反转性。
[Abstract]:Since the birth of China's first securities investment fund, China's securities investment fund market has experienced nearly 20 years of development, both in terms of the number of funds and the size of funds have achieved rapid development. It provides a solid foundation for the stable development of the capital market. However, compared with the mature market, China's securities investment funds have yet to improve their control over the securities market. In order to realize the all-round development of the securities investment fund, one of the core is to construct a solid and effective market credit system, which is inseparable from the establishment of the comprehensive evaluation mechanism of the fund. Objective and scientific fund performance evaluation system. In the part of theoretical research, firstly, the article about fund performance research at home and abroad is analyzed and studied, including the single factor model of risk and return, the ability model of fund manager, the multi-factor model of risk and return, the multivariate statistical model, and so on. From the point of view of data Envelopment Analysis (DEA), this paper classifies and analyzes the domestic and foreign classical literature materials. Secondly, the classical fund performance indicators are analyzed in detail. The Traineau (Treynor) index, Sharp (Sharpe) index, Jensen (Jensen) index, Sotino (Sortino) ratio and information ratio are analyzed theoretically in this paper. At the same time, the theoretical basis of calculating the stock selection time index is analyzed, such as T _ (m) model, H ~ (m) model, C ~ (?) L model, Fama-French 's three-factor model and Carhart's four-factor model. Finally, through the analysis of VaR model, factor analysis and super-efficiency DEA model, this paper constructs the empirical synthesis model "VaR factor analysis super-efficiency DEA" model. In the part of empirical research, firstly, this paper improves the risk measurement method of traditional performance evaluation method, introduces the VaR risk measure model to measure the risk of sample funds, and through the risk measurement of sample funds in different periods, we can know that the risk of the sample funds in different periods can be measured by using the risk measurement model. The mean weekly VaR of the sample funds in the periods of shock, strong and weak were 0.0527, 0.1137 and 0.0724, respectively. Among them, the highest was the strong period, the next was the weak period, and the lowest was the shaking period. In addition, using the obtained VaR value to improve the Sharpe index, through the improved Sharpe index to carry on the performance ranking to the sample fund; Secondly, the index system selected in this paper is processed by factor analysis. The risk measure index, the risk benefit measure index and the stock selection time measure index are obtained by the factor analysis. The cost and cost measure index and the fund structure measure index have five kinds of input index. The factor score of the sample fund can be obtained by calculating the factor analysis method, and the performance of the sample fund can be classified and evaluated according to the factor score. Thirdly, the input-output data obtained from factor analysis is introduced into the super-efficiency DEA model, and the performance evaluation of the sample is carried out by calculating the efficiency value. According to the result of efficiency calculation, the main reason for the invalidity of the fund is the inefficiency of technology. From the projection analysis, it can be seen that the ineffective fund needs to improve the investment efficiency of the fund mainly from the aspects of structure factor and stock selection time factor. According to the comprehensive performance ranking analysis, the types of funds and investment styles have no significant impact on the comprehensive performance. Finally, the performance ranking of sample funds in different periods is analyzed on a continuous basis. The empirical results show that from the period of shock to the period of strength, the performance of fund performance is weak persistence. From the strong period to the weak period, the fund performance does not exist performance continuity, on the contrary, it shows a strong performance reversal.
【学位授予单位】：江西财经大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：F832.51

【参考文献】