2020/11/22 by Beste Hamiye Beyaztaş, Beyaztas, Beste Hamiye, Soutir Bandyopadhyay +1
Mathematics · Decision Sciences · #Advanced Statistical Methods and Models #Advanced Statistical Process Monitoring #Statistical Methods and Inference
paper · pdf · doi:10.48550/arxiv.2011.11123
The panel data regression models have gained increasing attention in\ndifferent areas of research including but not limited to econometrics,\nenvironmental sciences, epidemiology, behavioral and social sciences. However,\nthe presence of outlying observations in panel data may often lead to biased\nand inefficient estimates of the model parameters resulting in unreliable\ninferences when the least squares (LS) method is applied. We propose extensions\nof the M-estimation approach with a data-driven selection of tuning parameters\nto achieve desirable level of robustness against outliers without loss of\nestimation efficiency. The consistency and asymptotic normality of the proposed\nestimators have also been proved under some mild regularity conditions. The\nfinite sample properties of the existing and proposed robust estimators have\nbeen examined through an extensive simulation study and an application to\nmacroeconomic data. Our findings reveal that the proposed methods often\nexhibits improved estimation and prediction performances in the presence of\noutliers and are consistent with the traditional LS method when there is no\ncontamination.\n