Attempts have recently been made to assign inequality contributions to various components of income. This paper discusses the issues involved in such assignments and highlights the problems that follow from having a number of possible decomposition rules. U. S. data on the distribution of family incomes are used to examine the relative influence of these income components and to evaluate empirically the performance of different decomposition rules. A wide range of inequality contributions can be obtained, even when restricted to only “naturally” derived decomposition rules. Some of the results are plainly absurd and serve to warn against the indiscriminate use of decomposition formulae without first investigating their properties.

You do not currently have access to this article.