Research Highlights

Focused Information Criterion and Model Averaging for Large Panels with a Multifactor Error Structure (with Shou-Yung Yin and Chang-Ching Lin, published in JOURNAL OF BUSINESS & ECONOMIC STATISTICS)

2019-08-23

Author Shou-Yung Yin
Chu-An Liu
Chang-Ching Lin
Abstract This paper considers model selection and model averaging inpanel data modelswith a multifactor error structure. We investigate the limiting distribution of thecommon correlated effects estimator (Pesaran, 2006) in a local asymptotic frameworkand show that the trade-off between bias and variance remainsin asymptotic theory.We then propose a focused information criterion and a plug-in averaging estimatorfor large heterogeneous panels and examine their theoretical properties. The novelfeature of the proposed method is that it aims to minimize thesample analog of theasymptotic mean squared error and can be applied to cases irrespective of whetherthe rank condition holds or not. Monte Carlo simulations show that both proposedselection and averaging methods generally achieve lower mean squared error thanother methods. The proposed methods are applied to examine possible causes thatlead to the increasing wage inequality between high-skilled and low-skilled workers inthe U.S. manufacturing industries.
Link https://www.tandfonline.com/doi/full/10.1080/07350015.2019.1623044(Open New Window)