对基于价值无差异的资产定价模型的实证检验——兼解释股权高溢价之谜  

An Empirical Investigation for Value-Indifference-Based Asset Pricing Model—With an Explanation for the Equity Premium Puzzle

作　　者：孙力强[1] 陈小悦[1] 

机构地区：[1]清华大学会计研究所

出　　处：《中国会计评论》2008年第1期41-60,共20页China Accounting Review

摘　　要：陈小悦和孙力强(2007)在价值无差异的基础上建立了一套全新的定价模型,本文采用股票市场的数据对该模型进行了实证检验,模型检验的同时也是对股权溢价之谜进行解释。研究结果表明,本文的定价模型在美国、中国内地和中国香港三个市场的检验都取得了良好的效果,即市场风险溢价均值都向模型的理论值收敛,实际风险溢价与理论值差异很小且不显著,采用该模型可以准确地描述股票市场组合收益率与风险的关系,并对股权溢价之谜做出合理的解释。The classical theories of asset pricing are being suspected and leaving ＂puzzles＂ unsolved for many years, and utility function has been the obstacle of development of asset pricing theory. Assuming asset prices follow Ito processes Chen and Sun （2007） established a new asset pricing model based on value indifference platform, where the value indifference judgments become the precondition of the utility indifference judgments. And then the model is extended to the case where asset prices follow normal distributions. This model establishes the direct quantitative relationship between return and risk, and hints value indifference among stocks, risky bonds and riskless assets, and also the long-term equilibriums based on the value indifference. A new risk factor investment evaluation period T is imported to the unified equilibrium structure, which displays a wide research field in the future. Using stock market data this article tries to provide empirical evidences to supporting the quantitative relationship and investment evaluation period indicated by the model, and also explain the equity premium puzzle at the same time. The results show that this model can interpret the equity premiums in America, China mainland and Hong Kong stock markets respectively, even the returns of these markets have different standard deviations. For details the average risk premiums converge to the theoretical values calculated by the model, and the differences between actual risk premiums and their theoretical correspondences are very small （less than 2% in most cases） and not significant. By simulations of 100 000 times, we find that the curve of actual values crosses that of theoretical values at the point where the investment evaluation period is about one year, which is consistent with our prior analysis. The simulation results are perfect in America stock market while there are a little larger errors in China mainland and Hong Kong stock markets. The probable reason is that the theoretical values are sensitive near th

关 键 词：资产定价 实证检验 股权溢价之谜 

分 类 号：F830.91[经济管理—金融学] F832.5