我国债券投资组合业绩归因方法研究

发布时间：2018-06-18 06:56

本文选题：业绩归因 + 投资组合　；参考：《南京大学》2012年硕士论文

【摘要】：业绩归因方法是科学度量投资经理投资决策效果并帮助其改善投资业绩最有效的方法之一,它是投资组合绩效评价体系中的重要环节。业绩归因方法本质上是将投资组合(Portfolio)的实际绩效与一个市场基准(Benchmark)的收益进行比较,同时,将两者之间的差额即超额收益分解成与投资经理决策过程对应的几种效应(Effect),以解释超额收益的来源。本文以债券定价公式为切入点,应用微积分方法分解债券收益率从数学角度证明了债券的收益主要来自三个方面,分别是持有收益、利率曲线收益和利差收益,由于债券组合是众多债券的集合,本文进一步用数学公式分解了债券投资组合收益率。在这一基础上本文从理论以及实务角度探讨了债券组合的投资经理为获取相对基准组合的超额收益可以采取的投资决策过程即实务中常用的“自上而下”投资决策过程,根据不同投资策略所要捕捉的不同超额收益将整个投资决策过程分解成五大投资决策过程因子并且与分解出的债券组合超额收益一一对应最终得到构成债券组合超额收益的五大归因效应,分别为持有效应、久期效应、期限结构配置效应、债券品种配置效应以及个券选择效应。最后本文为了验证模型得出的五大归因效应是否能够完整解释债券投资组合的超额收益还运用国内某投资公司的真实数据进行了实证研究,实证结果表明：本文所提出的模型在期间较短如一个交易日的情况下比较有效,上述五大归因效应能解释90%多的超额收益的来源,但是当考察期间较长如一个月时误差则较大,误差较大的原因来源于诸多方面,首先当考察期间较大时,本文通过求微分得到的近似公式不再有效,其次,由于跨度较大,外部现金流的干扰较大,导致Dietz近似收益率计算方式误差较大,这一结果进一步证明有必要采用跨期处理手段以更精确的得出较长期间下的归因结果。
[Abstract]:Performance attribution method is one of the most effective methods to measure the investment decision effect of investment managers and help them to improve their investment performance. It is an important link in the portfolio performance evaluation system. The performance attribution approach essentially compares the actual performance of portfolio with that of Benchmark, a market benchmark, while, The difference between the two is decomposed into several effects corresponding to the decision process of the investment manager to explain the origin of the excess return. In this paper, the bond pricing formula is used to decompose the bond yield. From the mathematical point of view, the paper proves that the bond yield mainly comes from three aspects, namely, holding income, interest rate curve income and interest difference income. Since bond portfolio is a collection of many bonds, this paper further decomposes the yield of bond portfolio by mathematical formula. On this basis, this paper discusses the investment decision-making process that the investment manager of bond portfolio can take in order to obtain the excess return of the relative benchmark portfolio from the perspective of theory and practice, that is, the "top-down" investment decision-making process commonly used in practice. According to the different excess returns to be captured by different investment strategies, the whole investment decision-making process is decomposed into five major investment decision process factors, and corresponding to the decomposed bond portfolio, the excess returns can eventually be obtained to form the bond portfolio. The five attributional effects of excess returns, They are the holding effect, the duration effect, the term structure allocation effect, the bond variety allocation effect and the security selection effect. Finally, in order to verify whether the five attributional effects obtained by the model can fully explain the excess returns of the bond portfolio, this paper also makes an empirical study using the real data of a domestic investment company. The empirical results show that the proposed model is more effective when the period is as short as one trading day. The above five attributional effects can explain the origin of more than 90% of the excess returns, but the errors are larger when the period of investigation is longer, such as one month. The reasons for the larger error come from many aspects. Firstly, when the period of investigation is large, the approximate formula obtained by differential is no longer valid. Secondly, because of the large span, the interference of external cash flow is greater. As a result, the error of Dietz approximate rate of return is large. This result further proves that it is necessary to use the method of intertemporal processing to obtain more accurate attribution results in the longer term.
【学位授予单位】：南京大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F832.51;F224

【参考文献】