多标度投资组合绩效度量非系统误差分析及校正

发布时间：2019-04-02 00:23

【摘要】：传统金融学研究在运用样本数据计算不同期限组合收益和方差时,通常会任意选取一个样本标度来对多标度投资组合绩效进行度量,这一简单的处理实际上隐含了时间标度与组合期望收益和方差相互独立的假设。而在现实投资实践中,由于投资环境的复杂性,导致投资者需要不断调整资产的投资期限,造成投资者在进行跨期组合绩效测度时经常忽视时间标度对组合绩效的影响效应,进而使投资决策产生由于基准时间标度任意选择而引起的非系统误差。因此,如何对多标度投资组合绩效进行合理测度是金融经济学领域中具有较高现实意义的研究问题之一。论文主要对多标度条件下的投资组合绩效测度非系统误差进行实证分析并对其校正进行探讨。首先从理论分歧与投资实践困惑两个角度出发,提炼论文研究的价值和意义,并提出相应的研究思路和方法。其次从跨期投资组合绩效度量、金融市场多标度研究框架以及投资期限与组合绩效关系三个方面出发,对金融学关注的投资期限、投资组合绩效以及绩效测度三大核心问题进行逻辑综述。在多标度投资组合夏普比率计算改进基础上,分时间标度大于基准标度(TkT0)和时间标度小于基准标度(TkT0)两种情况对理论与实际SR值之间的非系统误差进行对比,发掘非系统误差规律特征。最后对多标度投资组合绩效误差校正函数设计及有效前沿获取进行探讨,通过设计非系统误差关于时间标度T的校正函数,以沪深300指数为例对误差校正函数的实际校正效果进行比较,并基于校正误差函数获取投资组合绩效单标度和多标度有效前沿。实证结果分析表明：在TkT0条件下,GTSR RSR,且而当TkT0时,RSRGTSR,存在基于二项式拟合误差函数比基于泰勒展开获得的误差函数校正效果更为理想,多标度有效前沿随着时间标度扩大呈现出非线性特征。通过对多标度投资组合绩效度量误差研究,采用沪深300指数对组合绩效测度存在的非系统误差进行实证分析,并提出相应的误差校正思路和方法,旨在校正投资者在多标度投资组合绩效测度过程中产生的非系统误差,进而正确指导投资决策和风险管理。
[Abstract]:When using sample data to calculate income and variance of portfolio with different maturity, traditional finance studies usually choose a sample scale to measure the performance of multi-scale portfolio. This simple treatment actually implies the assumption that time scales are independent of portfolio expected returns and variances. In the real investment practice, because of the complexity of the investment environment, investors need to adjust the investment duration of the assets constantly, which results in the investors often ignore the effect of time scale on portfolio performance when they measure the cross-term portfolio performance. Furthermore, investment decision-making results in non-systematic error caused by arbitrary selection of reference time scale. Therefore, how to measure the performance of multi-scale portfolio reasonably is one of the research problems with high practical significance in the field of financial economics. This paper mainly analyzes the non-systematic error of portfolio performance measure under multi-scale conditions and discusses its correction. First of all, from the perspective of theoretical divergence and investment practice confusion, the value and significance of the study are refined, and the corresponding research ideas and methods are put forward. Secondly, from three aspects: cross-term portfolio performance measurement, financial market multi-scale research framework and the relationship between investment duration and portfolio performance, this paper focuses on the investment duration of finance. The three core issues of portfolio performance and performance measurement are summarized logically. On the basis of improved Sharp ratio calculation of multi-scale portfolio, the non-systematic errors between the theoretical and actual SR values are compared in the case that the time scale is larger than the benchmark scale (TkT0) and the time scale is less than the benchmark scale (TkT0). Explore the characteristics of non-systematic error law. Finally, the design of performance error correction function of multi-scale portfolio and the acquisition of effective frontier are discussed, and the correction function of non-systematic error about time scale T is designed. Taking Shanghai-Shenzhen 300 index as an example, the actual correction effect of error correction function is compared, and the effective frontier of portfolio performance card scale and multi-scale is obtained based on correction error function. The empirical results show that the error function based on binomial fitting in RSRGTSR, is better than the error function based on Taylor expansion in the case of, GTSR RSR, in TkT0 and when TkT0 is in existence, and the correction effect of error function based on Taylor expansion is better than that in the case of Taylor expansion. With the expansion of time scale, the effective frontier of multi-scale presents nonlinear characteristics. Through the research of multi-scale portfolio performance measurement error, this paper uses the Shanghai-Shenzhen 300 index to analyze the non-systematic error of portfolio performance measure, and puts forward the corresponding error correction ideas and methods. The purpose of this paper is to correct the non-systematic errors produced by investors in the process of multi-scale portfolio performance measurement so as to guide the investment decision-making and risk management correctly.
【学位授予单位】：湖南大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F830.59;F224

【参考文献】