Abstract
Panel data modeling is being used increasingly as a versatile tool to study various economic relationships. In such studies, sometimes, the dependent variable depends not only on the pure exogenous variables, but also on its own lag values. This intervention leads us to use dynamic panel data models. While using least squares dummy variable estimator, we show the attractive performance of such models as compared to that of simple panel data models. The choice of dynamic panel data model brings salient results in terms of lower standard error of regression, improvement in R2 and right specification of the model