ESTIMATING THE INEQUALITY OF HOUSEHOLD INCOMES: A STATISTICAL APPROACH TO THE CREATION OF A DENSE AND CONSISTENT GLOBAL DATA SET

Abstract

The deficiencies of the Deininger and Squire data set on household income inequality are well known to include sparse coverage, problematic measurements, and the combination of diverse data types into a single data set. Yet many studies have relied on this data due to the lack of available alternatives. In this paper we show how the UTIP-UNIDO measures of manufacturing pay inequality can be used, with other information, to estimate measures of household income inequality. We take advantage of the systematic relationship between the UTIP-UNIDO estimates and those of Deininger and Squire. The residuals from this exercise provide a map to problematic observations in the Deininger and Squire data, and the estimated coefficients provide a way to construct a new panel data set of estimated household income inequality. This new data set provides comparable and consistent measurements across space and through time.