Estimating the Inequality of
As described in chapter 2, the University of Texas Inequality Project (UTIP) has produced a global inequality dataset, based on the Industrial Statistics database published annually by the United Nations Industrial Development Organization (UNIDO). This dataset has approximately 3,200 observations over thirty-six years (1963–1999), in the original version, with updates into the early 2000s so far.1 It is based on source data that are likely to be accurate and consistent, both through time and across countries. It offers a direct way to measure the change in inequality in manufacturing earnings over time, and we have argued that these measures are good instruments (or approximations) for the overall change in earnings inequality in most countries and a good basis for making international comparisons of inequality levels. The density of coverage permits detailed checking for consistency through time and between countries, and despite their quick-and-dirty nature these measures hold up under this scrutiny very well.2
However, the calculations do not measure household income inequality. UTIP-UNIDO is a set of measures of the dispersion of pay using the betweengroups component of a Theil index, measured across industrial categories in the manufacturing sector. Though there is evidence that the UTIP-UNIDO measures provide a sensitive index of changes in distribution generally, the exact nature of the correlation between an establishment-based measure of manufacturing pay inequality and a survey-based measure of household income inequality is not clear, particularly in comparisons across countries.
This chapter offers a way to combine the information in the DS data with the information in UTIP-UNIDO, along with a certain amount of additional information, in order to accomplish two objectives. The first is to separate the useful