Economists are now familiar with “between” and “within” group inequality decompositions, for race, gender, spatial units, etc. But what exactly is the normative significance of the empirical results produced by these decompositions? This paper raises some basic questions about policy interpretations of decompositions that are found in the literature.