In this paper, we assume that redistribution modifies the community structure of an economy and generates both intra and intergenerational mobility. In a world in which neighbourhood attributes and family backgrounds are important to determine the investment effort and the productivity of education, we abandon the stylised median voter in favour of more nuanced pressure-group reasoning and study (i) the conflicts of material interest, (ii) both the income distribution and growth dynamics. We find multiple politico-economic regimes where the pressure for redistribution is highly non-linear depending on social affinity. In some range of inequality, the ends may oppose high rates of redistribution against the middle and more unequal and segmented economies may redistribute less, not more. Our framework displays multiple steady states describing either a situation of equality and integration or segregation with persistent inequality depending on historical economic discrimination. Re-distribution increases the number of candidates for the integrated equilibrium, but an economy’s institutions will have first to overcome ranges of inequality in which voting cycles are likely to emerge.