SMU SOE Practice Job Talk (Dec 1, 2017): Nonstationary Panel Model with Latent Group Structures and Cross-sectional Dependence

Please click here if you are unable to view this page.

TOPIC: NONSTATIONARY PANEL MODEL WITH LATENT GROUP STRUCTURES AND CROSS-SECTIONAL DEPENDENCE

ABSTRACT This article proposes a novel approach, based on Lasso, to handle unobserved heterogeneity in nonstationary panel model with cross-sectional dependence. We employ the penalized principal component (PPC, hereafter) estimation method to jointly estimate the group-specic long-run relations, unobserved common factors and identify individuals' membership. We obtain three types of estimators- C-Lasso, post-Lasso and Cup-Lasso estimators by iteratively performing the PPC-based method. In post-Lasso and Cup-Lasso estimators, we apply the fully modied procedure for bias-correction. Taken together, our estimators achieve the oracle property so that the group-specic coefficients can be estimated as well as if the individuals' membership were known. We establish the convergence rates and limiting distributions of the C-Lasso, post-Lasso and the Cup-Lasso estimators, which are normal and permit inference using standard test statistics. An empirical example is presented based on growth convergence puzzle through the channel of global technology diffusions. It empirically confirms the multiple steady states of growth convergence. Keywords: Nonstationary; Parameter heterogeneity; Latent group patterns; Penalized principal component; Cross-sectional dependence; Lasso; Growth convergence puzzle JEL Classification: C13; C33; C38; C51; F60; O32; O40 Click here to view the paper. Click here to view the CV.
		PRESENTER Wenxin Huang SMU
		RESEARCH FIELDS Econometric Theory Macroeconomics
		DATE: 1 December 2017 (Friday)
		TIME: 4pm - 5.30pm
		VENUE: Meeting Room 5.1, Level 5 School of Economics Singapore Management University 90 Stamford Road Singapore 178903

© Copyright 2024 by Singapore Management University. All Rights Reserved.