Robust Estimation for the Fixed Effect Panel Data Model

Nor Mazlina, Abu Bakar@Harun and Habshah, Midi (2017) Robust Estimation for the Fixed Effect Panel Data Model. In: 8th International Conference on Numerical Optimization and Operation Research (ICNOOR-VIII), 27 August 2017, Bukit Tinggi, Sumatera.

Text FH03-FESP-18-12638.pdf Restricted to Registered users only Download (485kB)

Abstract

Ordinary Least Square parameter estimation for the fixed effect regression suffers biasness in the presence of high leverage points. In this study, a robust alternative is proposed by incorporating new robust weights in Generalized M-estimator determined by a superior outlier detection method, Robust Diagnostic-F (RDF). The method helps to obtain new robust weights for Robust Within GM-estimator (RWGM). The performance of the newly proposed method, RWGM(RDF) is investigated using real and simulated data sets. The ratios of root mean square error are evaluated and compared with the existing RWGM under robust centering procedures. The newly proposed estimator is found to be more efficient and resilient towards high leverage points due to the success of the new robust weights. The results are confirmed through reanalyzing numerical examples.

Actions (login required)

View Item