[论文解读] Risk-return relationship: An empirical study of different statistical methods for estimating the Capital Asset Pricing Models (CAPM) and the Fama-French model for large cap stocks

Linh Nghiem|arXiv (Cornell University)|Nov 23, 2015

Financial Risk and Volatility Modeling参考文献 15被引用 4

一句话总结

这篇本科论文通过实证方法评估了使用大盘美国股票估计CAPM和Fama-French三因子模型的统计方法。研究发现，没有任何一种统计方法能可靠地恢复CAPM所预测的正风险-收益关系，且Fama-French模型在预测个股收益方面并未优于CAPM，进一步证实了这些模型在解释资产定价方面的局限性。

ABSTRACT

The Capital Asset Pricing Model (CAPM) is one of the original models in explaining risk-return relationship in the financial market. However, when applying the CAPM into reality, it demonstrates a lot of shortcomings. While improving the performance of the model, many studies, on one hand, have attempted to apply different statistical methods to estimate the model, on the other hand, have added more predictors to the model. First, the thesis focuses on reviewing the CAPM and comparing popular statistical methods used to estimate it, and then, the thesis compares predictive power of the CAPM and the Fama-French model, which is an important extension of the CAPM. Through an empirical study on the data set of large cap stocks, we have demonstrated that there is no statistical method that would recover the expected relationship between systematic risk (represented by beta) and return from the CAPM, and that the Fama-French model does not have a better predictive performance than the CAPM on individual stocks. Therefore, the thesis provides more evidence to support the incorrectness of the CAPM and the limitation of the Fama-French model in explaining risk-return relationship.

研究动机与目标

• 评估各种统计方法在估计CAPM和Fama-French模型中的表现。
• 评估Fama-French模型在预测大盘股票收益方面是否优于CAPM。
• 调查是否存在能够可靠恢复CAPM所预测的正风险-收益关系的统计方法。
• 为CAPM和Fama-French模型在解释资产收益方面的局限性提供实证证据。

提出的方法

• 本研究使用大盘美国股票的数据集进行实证分析。
• 应用多种统计估计方法来估计CAPM和Fama-French模型的参数。
• 通过样本外预测准确率比较CAPM和Fama-French模型的预测表现。
• 通过检验回归模型中贝塔值和因子载荷的符号与显著性来测试风险-收益关系。
• 所用统计方法包括普通最小二乘法（OLS）、稳健回归以及其他在资产定价中常用的估计器。
• 分析聚焦于个股收益，而非投资组合，以评估模型在个股层面的适用性。

实验结果

研究问题

• RQ1 不同的统计方法是否能可靠地恢复CAPM所预测的系统性风险（贝塔）与预期收益之间的正相关关系？
• RQ2 Fama-French三因子模型在预测个体大盘股票收益方面是否优于CAPM？
• RQ3 是否存在能够提高CAPM在现实数据中实证有效性的估计技术？
• RQ4 Fama-French因子在解释个股收益变动方面，其作用在多大程度上超越了市场风险？

主要发现

• 没有任何一种统计方法能一致地恢复CAPM所预测的正风险-收益关系。
• 当应用于个股时，Fama-French模型并未表现出优于CAPM的预测表现。
• 实证结果进一步为CAPM在解释资产收益方面的有效性提出质疑。
• 本研究证实了CAPM和Fama-French模型在捕捉个股层面真实风险-收益权衡方面存在持续的局限性。