This paper asks how variations in trade openness contribute to cross-country income differences. We approach this question using counterfactual experiments within a quantified general equilibrium model of trade. We find that trade costs gain their relevance only by amplifying the effects of existing differences in endowments, population sizes and technologies. If, for example, market entry costs were the same in all countries, inequality would be about 13% lower. Variable trade costs are found to have a similar effect. In contrast, if countries differed only by their degree of trade openness, the resulting variance of per capita income would be negligible.