Advanced industrial democracies experience increasing inequalities or at least a new trade-off between equality and growth: liberal welfare states opted for growth and accepted rising inequality, while conservative welfare states tried to hold back inequality, thereby accepting lower growth, and only the social democratic welfare states were partly able to overcome that trade-off. The rise in inequality is widely interpreted with regard to globalization and technological change. This paper contrasts this interpretation with an alternative based on the argumentation of Kuznets' inverted U-turn which is individually reformulated as the diffusion process of some qualification (which remains unspecified throughout this paper). These two alternative mechanisms which are identical with regard to inequality measured using the Gini coefficient or the standard deviation of logged incomes, can be differentiated through different trend expectations with regard to the skewness of income distributions: In the globalization model, increasing inequality is accompanied first by a fall and later-on by a rise in skewness, while the qualification diffusion model shows the opposite sequence: rising to a maximum and falling back later on. Due to their different position in the inequality-growth trade-off, liberal and social democratic welfare states are assumed to be ahead in this evolution, while conservative welfare states lagging behind. Based on the Luxembourg Income Study, skewness estimations of logged monetary income distributions form an unbalanced panel with 69 observations from 16 OECD countries. A fixed effects regression for the skewness time trend in conservative welfare states and the trend difference for the two other welfare state groups shows strong support for the qualification diffusion model.