Assessing income redistribution: what are the key analytic choices?

Abstract

We present the extent of divergence in the literature on the stylised facts about income redistribution in rich countries. Analytical choices that underpin this divergence are then identified and investigated empirically using microdata for 30 European countries. In terms of direct redistribution via cash transfers and direct taxes, whether social insurance pensions are treated as redistribution - the conventional approach - or as market income - as in some recent studies - is seen to be critical. When the analysis is extended to include indirect taxes and non-cash benefits from state spending, they work in opposite directions and generally have only a limited net redistributive impact. Being able to attribute the benefits of such spending to households in more satisfactory ways is a priority. Whether household survey data are 'corrected' to include missing incomes at the top as well as imputed rent of owner-occupiers and undistributed profits of companies is also seen to have a substantial impact on the scale of measured redistribution. Finally, extending the scope of redistributive analysis to include all of national income, as in recent studies from a Distributional National Accounts perspective, is investigated. This underlines the implications of including state spending on collective goods such as security and infrastructure, without a clear rationale for how it is meaningfully allocated across households.