Stata tip 149: Weighted estimation of fixed-effects and first-differences models

Abstract

This tip clarifies estimation of weighted panel-data models in Stata in two ways. First, it extends the well-known deviation-from-means interpretation of fixed-effects models and the equivalence between fixed-effects and first-differences models with two time periods to the case of weighted estimation. Second, it highlights several ways to fit weighted fixed-effects (WFE) models in Stata. Of course, the tip also applies to models that are weighted for reasons other than heteroskedasticity arising from group averaging.