Generalized Least Squares Model Averaging

54 Pages Posted: 1 Jul 2013 Last revised: 15 Apr 2015

See all articles by Qingfeng Liu

Qingfeng Liu

Otaru University of Commerce - Department of Economics

Ryo Okui

Seoul National University

Arihiro Yoshimura

Kyoto University - Graduate School of Economics

Date Written: December 21, 2014

Abstract

In this paper, we propose a method of averaging generalized least squares estimators for linear regression models with heteroskedastic errors. The averaging weights are chosen to minimize Mallows’ Cp-like criterion. We show that the weight vector selected by our method is optimal. It is also shown that this optimality holds even when the variances of the error terms are estimated and the feasible generalized least squares estimators are averaged. The variances can be estimated parametrically or nonparametrically. Monte Carlo simulation results are encouraging. An empirical example illustrates that the proposed method is useful for predicting a measure of firms’ performance.

Keywords: model averaging, GLS, FGLS, asymptotic optimality, Mallows’ Cp

JEL Classification: C51, C52

Suggested Citation

Liu, Qingfeng and Okui, Ryo and Yoshimura, Arihiro, Generalized Least Squares Model Averaging (December 21, 2014). Available at SSRN: https://ssrn.com/abstract=2287602 or http://dx.doi.org/10.2139/ssrn.2287602

Qingfeng Liu

Otaru University of Commerce - Department of Economics ( email )

3-5-21 Midori
Otaru City, Hokkaido 047-8501
Japan

Ryo Okui (Contact Author)

Seoul National University ( email )

Seoul
Korea, Republic of (South Korea)

Arihiro Yoshimura

Kyoto University - Graduate School of Economics ( email )

Japan

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
173
Abstract Views
1,133
rank
210,136
PlumX Metrics