A Note on the Application of Dynamic Panel Data (DPD) Models in Economic Models of Crime

Abstract

The dynamic panel data (DPD) model and application of Difference and System GMM estimators has great advantages to address endogeneity of law enforcement variables and measurement error in reporting of crimes in economic models of crime as well as controlling for fixed effects and avoiding dynamic panel bias. Meanwhile, neglecting instrument proliferation which is very common and uncommonly detected in these models can lead to misleading and unreliable results. Bolding this warning and presenting available techniques to control for this potential danger is the main mission of this note.