This paper proposes an ML estimator for short panel data models with interactive fixed effects and endogenous variables. We show that such a model can be analysed in terms of covariance structure analysis. Monte Carlo simulation results show that the proposed ML estimator outperforms the GMM estimator. |