Panel Data Estimation in Finance: Preliminary Assumptions and Parameter Consistency
40 Pages Posted: 6 Apr 2016 Last revised: 1 Jul 2016
Date Written: June 22, 2016
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: Suggested Citation