Horizontal inequity estimation : the issues of close equals identification

This paper focuses on the estimation of horizontal inequity, remaining within the framework of the close equals groups approach started by Aronson, Johnson and Lambert (1994), and systematised by Urban and Lambert (2008). Within the framework of the close equals groups the choice of bandwidth, which determines the intervals of close equal income units, is fundamental. Following the existing literature on the redistributive effect decomposition, we propose a new criterion for the identification of the optimal bandwidth: in this article the identification of close equals groups is mainly oriented to the estimation of horizontal inequity. Our criterion intends to be a contribution to empirical work that focuses on comparing the effects of different tax systems on a particular population of taxpayers. In order to test the robustness of the new criterion, different tax systems, characterised by different degrees of tax progressivity, are applied to Italian and Polish personal income tax data. Our results suggest the bandwidths, chosen according to our methodology, are more stable than those obtained by maximising potential vertical effects