Download this Paper Open PDF in Browser

Panel Data Estimation in Finance: Preliminary Assumptions and Parameter Consistency

40 Pages Posted: 6 Apr 2016 Last revised: 1 Jul 2016

William David Grieser

Texas Christian University

Charles J. Hadlock

Michigan State University - The Eli Broad College of Business and The Eli Broad Graduate School of Management

Date Written: June 22, 2016

Abstract

We examine the strict exogeneity assumption, a necessary and often overlooked condition for consistency of many panel data estimators that are used in financial research. We test this assumption in several common settings and find that these tests very frequently reject. We offer guidance on detecting violations of the strict exogeneity assumption and provide evidence on the possible magnitude of the resulting estimation errors. Alternative estimation approaches when strict exogeneity is violated are discussed. We investigate the testable assumptions underlying these alternative approaches and characterize current practices. Implications of our findings for empirical financial research strategies are discussed.

Keywords: Corporate Finance, Panel Data, Fixed Effects, First Differences, Strict Exogeneity

JEL Classification: C23, G30

Suggested Citation

Grieser, William David and Hadlock, Charles J., Panel Data Estimation in Finance: Preliminary Assumptions and Parameter Consistency (June 22, 2016). Available at SSRN: https://ssrn.com/abstract=2759386

William David Grieser (Contact Author)

Texas Christian University ( email )

Fort Worth, TX 76129
United States

Charles J. Hadlock

Michigan State University - The Eli Broad College of Business and The Eli Broad Graduate School of Management ( email )

315 Eppley Center
East Lansing, MI 48824-1121
United States
517-353-9330 (Phone)
517-432-1080 (Fax)

Paper statistics

Downloads
162
Rank
154,316
Abstract Views
423