基于Copula观点整合策略的大类资产组合管理——与Black-Litterman模型的比较研究

Portfolio Management based on Copula Opinion Pooling: Compared with Black-Litterman Model

Black-Litterman模型克服了均值—方差模型对输入参数过于敏感的弊端,但是其假设市场和主观观点均符合正态分布,却与资本市场常见的尖峰厚尾现象相悖。Copula观点整合策略(COP)在Black-Litterman模型的基础上引入Copula函数,可以有效解决市场和主观观点的非正态分布现象。根据股票、商品和债券等大类资产的历史收益率形成主观观点,采用等权重组合作为市场基准组合,比较研究了COP组合与BL组合及传统等权重组合间的绩效差异,研究结果表明:COP组合和BL组合通过融合投资者主观观点,在控制投资组合风险、降低净值曲线波动方面表现得远优于传统投资组合;由于考虑了资产的非正态分布,COP组合相对BL组合能够获得更高的动态配置绩效,研究结论在控制了时间窗口、风险规避系数及信心水平后仍然稳健。

Black-Litterman model overcomes the disadvantages of the mean-variance model being too sensitive to input parameters,but it assumes that the market and subjective views are consistent with the normal distribution,which is contrary to the peak and tail phenomenon common in the capital market.The Copula Opinion Pooling(COP)introduces the Copula function based on the Black-Litterman model,which can effectively solve the non-normal distribution phenomenon of the market and subjective views.This paper forms a subjective view based on the historical returns of major assets such as stocks,commodities and bonds,and uses the equal-weight portfolio as the market benchmark portfolio,and then compares the performance difference between COP portfolio and BL portfolio and traditional equal-weight portfolio.The results show that:COP portfolio and BL portfolio show far superior performance to traditional investment portfolio in controlling portfolio risk and reducing net value curve fluctuation by combining investor subjective views.Due to consideration of non-normal distribution of assets,COP portfolio can achieve higher dynamic configuration performance than BL portfolio,and the research conclusions are stable after controlling the time window,risk aversion coefficient and confidence level.