On the Measurement of Horizontal Inequity

This paper makes a new attack on the old problem of measuring horizontal inequity in the income tax. Local measures of inequality of posttax income among pretax equals are proposed, which reflect alternative value judgments about the nature and magnitude of an inequity. These measures are aggregated into global indices. The welfare gain from eliminating horizontal inequity revenue-neutrally, and the revenue gain from eliminating it welfare-neutrally, in each case preserving the vertical performance of the tax, are captured by these indices. Difficulties of implementation arising from the "identification problem" are discussed. A variation in the methodology validates banding the income data to create "close equals" groups. Simulations show that the banding procedure works well. A range of potentially fruitful applications is discussed.