Compensation hypothesis, which has established a link between trade openness of countries and levels of government spending, has been widely accepted in the literature on trade policy and international globalization. However, the nature of the distribution effects produced by trade is likely to determine the existence of more or less redistribution demands from the median voter, and therefore government growth. In this paper I hypothesize that the effects of trade openness on redistribution demands are not homogeneous between countries, and I argue that they depend both on the type-of-factor endowment of the economy and the size of the sectors more likely to be affected by trade. I test this hypothesis with ISSP data for 23 countries, both with a country level and an individual level analysis. The results show that redistribution demands issued from trade openness of the median voter of a country are largely conditional on GDP per capita and size of potential loser sectors such as manufacturing: while trade has a negative effect on pro-redistribution preferences in “poor” and/or in “low manufacturing” countries; it positively affects pro-redistribution preferences in “rich” and/or in “high manufacturing” countries. Additionally, I empirically observe that the size of the loser sector plays a more important mediating role than the type-of-factor endowment of the economy.