【95周年校庆系列讲座】Evaluation of Asset Pricing Models: Optimal Risk Premia and Goodness-of-Fit Measures

时间:2020-07-22         阅读:

光华讲坛——社会名流与企业家论坛第 5761 期

(线上讲座)

主题Evaluation of Asset Pricing Models: Optimal Risk Premia and Goodness-of-Fit Measures

主讲人新加坡南洋理工大学  Robert L.Kimmel兼职教授

主持人金融研究院  Philip H. Dybvig教授

时间2020年7月24日(周五)9:00-10:30

直播平台及会议IDTencent会议

主办单位:金融研究院 科研处

主讲人概况:

Robert L. Kimmel received his PhD from the University of Chicago, and has held various positions at Princeton University, Ohio State University, EDHEC Business School, and the National University of Singapore. Since leaving NUS last year, he has been an adjunct lecturer at Nanyang Technological University in Singapore, and a freelance consultant. His research focuses on continuous-time modelling of stochastic volatility and interest rate processes, and methods for estimation and evaluation of asset pricing models.

Robert L. Kimmel在芝加哥大学获得博士学位,曾就职过普林斯顿大学、俄亥俄州立大学、EDHEC商学院和新加坡国立大学。自去年离开新加坡国立大学后,他一直是新加坡南洋理工大学的兼职教授和自由职业顾问。他的研究重点是随机波动率和利率过程的连续时间建模,以及资产定价模型的估计和评估方法。

内容提要:

We consider linear factor models, in which expected returns are linear functions of beta coefficients on explanatory factors. Such models have been in common use in finance for decades, and many methods for assignment of risk premia to the factors and evaluation of the model's fit have appeared in the literature. We show that there is essentially a unique method for assigning risk premia, and a unique method for assessing the fit of the model, based on only two assumptions: the fit of the model is judged solely by its predictions of the assets' expected returns, and the fit improves when the prediction error for an asset decreases, holding the prediction error of all uncorrelated assets fixed. The unique (to within monotonic transformation and additional “tie-breaking” criteria) goodness-of-fit is based on the maximum Sharpe ratio that can be achieved using the factor mimicking portfolios, and the unique risk premia assigned to factors that are themselves excess returns are simply the expected excess returns of the factors.

本文考虑预期收益是因子β系数的线性函数的因子模型,该类模型在金融学中已被广泛使用了几十年,文献中出现了许多将风险溢价分配给因子和评估模型拟合度的方法。大家发现,本质上存在一种分配风险溢价的独特方法,以及一种评估模型拟合度的唯一方法,仅基于两个假设:模型的拟合度仅由其对资产预期收益率的预测来判断,此外,当对资产收益的预测误差减小,保持对所有不相关资产收益的预测误差不变的情况下,拟合度会提高。拟合优度(在单调变换和附加的“加时决胜”标准内)是基于因子模拟投资组合可以实现的最大夏普比率,而分配给因子独特的风险溢价就是这些因子的预期超额收益,这些因子自身就是超额收益。

参与方式

诚邀所有财大师生、教职员工参加我院举办的讲座!如您想参与讲座,请通过您的学校邮箱给gongmx@swufe.edu.cn发邮件,获取Tencent会议ID。

XML 地图 | Sitemap 地图