Fractionalization and the size of government

I study the effect of voters with a group-based social conscience. Voters then care more about the well-being of those belonging to their own group than the rest of the population. Within a model of political tax determination, both fractionalization and group antagonism reduce the support for redistribution. Whereas within group inequality increases support for redistribution, inequality between groups has the opposite effect. All these results hold even if a poor group is in majority. Using a panel data set for the US constructed from micro data, I find support for the hypothesis that within race inequality increases and between race inequality decreases redistribution.