Upcoming Events at CAFIN

UC Santa Cruz CAFIN Lecture:

On Comparing Asset Pricing Models

Siddhartha Chib (joint work with Xiaming Zeng and Lingxiao Zhao) 

Wednesday May 15, 2019
 3:30-5:00 PM
Engineering 2, room 499
1156 High Street, Santa Cruz, California 95064

 


    Abstract

  • A fundamental goal of theoretical and empirical finance is to explain and measure the risk premium (the difference between the expected return of the asset and the risk-free return) for the cross-section of traded assets in financial markets. According to the factor theory of asset pricing, financial assets earn a risk-premium because the returns on those assets are systematically related to the underlying pricing factors. Such factors are called risk factors and represent sources of systematic risk affecting all assets. Recently, Barillas and Shanken (2018, Journal of Finance) develop a Bayesian approach for finding such risk factors. Our first goal is to point out that their Bayesian marginal likelihood based model comparison method is unsound and cannot be used to locate the risk factors. We establish this point with theoretical arguments and empirical evidence. Our second goal is to show how this problem can be done correctly. Along with the method given in Chib and Zeng (2018, Journal of Business and Economic Statistics), and in this work (forthcoming in the Journal of Finance), we have a reliable way to select which factors are risk factors in asset pricing models.


  • Speaker

  • Siddhartha Chib

    Siddhartha Chib

    Harry C. Hartkopf Professor of Econometrics and Statistics
    Olin Business School, Washington University in St. Louis

    Professor Chib's research is in the area of Bayesian statistics and Markov chain Monte Carlo computational methods. In his work, he has developed new statistical approaches and methods for diverse problems involving binary, ordinal and censored data, Metropolis-Hastings algorithms, model choice, stochastic volatility, Markov mixture models, asset pricing, causal inference, and moment-based Bayesian inference. He is a Fellow of the American Statistical Association, Journal of Econometrics, and the International Society of Bayesian Analysis. Since 2003, he has directed the annual NBER-NSF sponsored Seminar in Bayesian Econometrics and Statistics, which features presentations by young and established researchers working on the theory and application of Bayesian methods.