I construct a successive generations model of appropriative competition between ethnicities where de facto power puts an upper-bound on the consumption of each clan. My aim is to derive the implications of the structure of the ethnic and political divisions in terms of economic prosperity. I expose a mechanism reducing negative externalities in consumption implying that long-run economic growth is larger when influence is concentrated rather than dispersed. I find that competition intensity sometimes is beneficial if it is above a thresh-old or if small groups initially have a large influence. I provide an argument in favour of the Ethno-Linguistic Fractionalization index over a polarization index when considering the impact of diversity on growth. This paper fills a gap in the literature by allowing for the possibility of more than two groups of different sizes in a dynamical setup and by being consistent with the evidence of proportionality between ethnic and political share sin sub-Saharan Africa.