Measuring variation in tax liability among economic equals

Abstract

The coefficient of variation (CV) and coefficient of residual variation (CRV) have been used as measures of horizontal equity. Both, however, are noisy measures in that they overstate the amount of variation due to inequity in the tax system (Grasso & Frischmann 1992). The purpose of the current study, therefore, is to use the CRV of tax liability and provide an estimate of the portion of this CRV measure that is due to the tax system alone and that portion that is due to specification error. This precision is an improvement over the Grasso and Frischmann model. The first purpose is achieved by computing the difference between two CRV measures. The first CRV measure is derived from a regression of an expanded total income amount (ETI) on tax liability; the second CRV measure is from a regression of several explanatory variables on tax liability. To the extent that the more fully-specified regression captures the Internal Revenue Code provisions, the reduction in the CRV measure can be attributed to the tax system. A second purpose of this study is to estimate the extent to which the various types of tax code provisions cause this variation. The second purpose is achieved by an iterative process of omitting one explanatory variable from the full regression and determining the change in CRV due to this omitted variable.

Citation

Westort, P.J. (2001), "Measuring variation in tax liability among economic equals", Advances in Taxation (Advances in Taxation, Vol. 13), Emerald Group Publishing Limited, Leeds, pp. 169-203. https://doi.org/10.1016/S1058-7497(01)13011-5