COMPARING SOME PANEL DATA ESTIMATORS IN THE PRESENCE OF AUTOCORRELATION

Authors

  • K. Ajao Department of Statistics, Nasarawa State University, Keffi, Nigeria & NSUK-LISA Stat Lab, Nasarawa State University, Keffi,
  • M.O. Adenomon Department of Statistics, Nasarawa State University, Keffi, Nigeria & NSUK-LISA Stat Lab, Nasarawa State University, Keffi,
  • M.U. Adehi Department of Statistics, Nasarawa State University, Keffi, Nigeria & NSUK-LISA Stat Lab, Nasarawa State University, Keffi,

Abstract

In this work, panel data that were characterized by features of no first order autocorrelation was modelled using three estimation models: Pool Regression, Fixed Effect, Random Effect models. Panel data like other aspects of econometrics, exploits regression analysis as one of the statistical tools to formulate and illustrate models. The regression analysis requires some assumptions which, if violated, results to one problem or the other. In such case, the Pooling method of estimation remains linear, unbiased and normally distributed but might not be efficient as the estimates of the parameters might become indeterminate, the confidence intervals may be too wide and the standard errors might become large. Simulation studies were carried out at different panel structures and autocorrelation level. The experiment was repeated for 10,000 times and Root Mean Square Error (RMSE) was used to judge the performance of the models. The results from this work showed that for small sample panel structure N = 25, T = 5 and n = 5, irrespective of autocorrelation levels, fixed effect model is preferable at all level. But when moderate for N = 50, T = 10, n = 5, irrespective of autocorrelation level, random effect model is preferred, while for large panel structure for N = 450, T = 30, n = 15, irrespective of autocorrelation level, random effect model is preferred.

Downloads

Published

2023-10-07

Issue

Section

ARTICLES