Panel Data Analysis: A Simplified Summary

Abstract

The purpose of this article is to provide a summary of linear panel data analysis in simple language without mathematical expression. The summary serves as a quick reference for many researchers who want to utilize panel data techniques in their research. The article starts by explaining what a panel dataset is and how it differs from time series and cross-section datasets. Next, it explains the main linear panel data models which are then summarized graphically. Panel data techniques solve two main problems: (1) omitted variable bias (individual-fixed effects or unobserved heterogeneity) and (2) endogeneity bias. As shown in Figure 1, dynamic (GMM) models and some instrumental variable (IV) models can solve these two problems.

Keywords: Panel Data, Fixed Effects, Random Effects, Endogeneity, GMM

JEL Classification: C00, C10, C23, C26