Estimation and Inference in Interactive Effects Panel Data Models with a Constrained Factor Structure

• Activity Number: 70
• Type: Contributed
• Date/Time: Sunday, August 9, 2015 : 4:00 PM to 5:50 PM
• Sponsor: Business and Economic Statistics Section
• Author(s): Mohitosh Kejriwal* and Evan Totty
• Companies: Purdue University and Purdue University
• Keywords: common factors ; interactive effects ; constrained estimator

Abstract

This paper considers estimation and inference in interactive effects panel data models when the factor structure is subject to constraints. The constraints are useful tools for incorporating prior information or economic theory in applied work. In particular, we investigate the effects of such constraints on the statistical properties of the estimator for the slope parameter and the associated test statistics. Our analysis is motivated by the well-known incidental parameter problem when the cross-section dimension is large. We describe an efficient procedure for estimating such a model as well as study its large sample properties. Monte Carlo simulations are presented to assess the efficacy of the proposed procedure in finite samples.