Traces of the past in income inequality

Abstract

The cumulative effect on income inequality of structural economic changes can only be observed in years. We discuss the implications of this point for the empirical approach to the inequality data through a simple linear regression setting. We show that determinants of inequality can only be found by (i) specifying a dynamic model, or (ii) using the average of changes in within-cohort inequalities as the dependent variable. The results obtained with the US and UK data show that these two methods deliver similar results, which significantly differ from those obtained through standard static regression analyses.