Are conditional government transfers a politically acceptable form of redistribution?

Abstract

This paper examines whether characteristics of government transfers that do not affect actual levels of redistribution affect support for such transfers. We employ survey experiments in Brazil, Chile, Uruguay, and Turkey to determine whether, and in what contexts, making government transfers conditional on behavior of bene ciaries increases support conditional government transfers among non-bene ciaries. In these experiments, we prime some respondents to think of themselves as \different' from potential bene ciaries in regional and ethnic/racial dimensions, and manipulate the type of conditionality imposed on bene ciaries of government transfers. We seek to determine whether a \conditionality premium' exists among non-bene ciaries, and whether it is affected by how similar non-bene ciaries perceive themselves relative to bene ciaries of government transfers. Results show that conditional transfers are generally more popular than similar unconditional ones, and also support our `otherness hypothesis', whereby the \conditionality premium' is greater when non-bene ciares are primed to think of themselves as being different from bene ciaries in non-economic dimensions. This is only the case in the two more heterogenous countries that we examine (Brazil and Turkey). The fact that some results only seem to hold in heterogenous societies suggests that conditionalities might be more necessary in such societies