Robust M Estimation for Poisson Panel Data Model with Fixed Effects: Method, Algorithm, Simulation, and Application

Abstract

The fixed effects Poisson (FEP) model is one of the most important for the count data when the data containperiods and cross-sectional units. The maximum likelihood (ML) estimation method for the FEP model provides good results in the absence of outliers, but it is affected by outliers. So, we introduce in this paper robust estimators for the FEP model. These estimators yield stable and good results in case of the presence of outliers. The Monte Carlo simulation study and empirical application were conducted to assess the performance of the non-robust fixed Poisson maximum likelihood (FPML) estimator and the robust estimators: fixed Poisson Huber (FPHR), fixed Poisson Hampel (FPHM) and fixed Poisson Tukey (FPTK). The results of simulation and application show that robust estimators are better than FPML estimator when the count panel data contains outliers. In addition, FPTK is more efficient than other robust estimators.