Predictive Efficiency Of Random Effects Approach: A Real Model Simulation Study

Main Article Content

Samad AmirKhalkhali
Sal AmirKhalkhali

Keywords

Simulation, Panel Data Regression Model, Fixed Effects, Random Effects Methods

Abstract

This real model simulation study attempts to shed more light on the predictive performances of two of the most commonly used panel data regression methods - fixed effects and random effects. In particular, this paper attempts to address the question, How do these two alternative estimators perform in prediction when errors follow non-normal distributions? The simulation results support the random effects approach as the better choice.

Downloads

Download data is not yet available.
Abstract 97 | PDF Downloads 115