Common Correlated Effects Estimation of Hybrid Panel Data Models

Abstract

In this research, a model of panel data models was reviewed which is hybrid coefficients model which is characterized by a portion of the regression coefficients being fixed slopes while the other portion of the coefficients are random slopes meaning that they have a normal distribution with an unknown mean and variance. Several methods were used to estimate the parameters of this model in the case of unbalanced panel data. these estimation methods depend on the common correlated effects estimator which is composed of three estimators, common correlated effect mean group estimator(CCEMG), common correlated effect pooled (CCEP) and half jackknife panel (HJP)estimator to estimate the parameters of the hybrid coefficients model represented by the first fixed slope and the mean of the random slope coefficient. Monte Carlo experiments and different sample sizes (NT) are small, medium and large, with different variance levels to compare between estimation methods, the simulation results showed that the (CCEP) is the best estimation method because it has the less average mean absolute error (AMAE). The (CCMG) is the best method after (CCEP).