Bayesian Parametric and Semiparametric Factor Models for Large Realized Covariance Matrices

发布时间:2018-04-10

ShanghaiTech SEM Working Paper No. 2018-005

  

Xin Jin

Shanghai University of Finance and Economics - School of Economics

John M. Maheu

McMaster University - DeGroote School of Business

Qiao Yang

ShanghaiTech University - School of Entrepreneurship and Management



Abstract

This paper introduces a new factor structure suitable for modeling large realized covariance matrices with full likelihood based estimation. Parametric and nonparametric versions are introduced. Due to the computational advantages of our approach we can model the factor nonparametrically as a Dirichlet process mixture or as an in nite hidden Markov mixture which leads to an in nite mixture of inverse-Wishart distributions. Applications to 10 assets and 60 assets show the models perform well. By exploiting parallel computing the models can be estimated in a matter of a few minutes.

  

Keywords: in nite hidden Markov model, Dirichlet process mixture, inverse-Wishart, predictive density, high-frequency data

JEL Classification: G17, C11, C14, C32, C58

  

Date Written: September 2017

  

Available at SSRN: http://ssrn.com/abstract=3159716

Download PDF Version here:

【No.2018-005】SSRN- 2 Bayesian Parametric and Semiparametric Factor Models for Large Realized Covariance Matrices.pdf