Multigenerational Mobility in Earnings and Education: Evidence from Administrative Data

This study measures the persistence in income and education over three generations in Switzerland. I use administrative data covering the universe of labor income since 1982 and family linkages over three generations. Most studies rely on mobility estimates from two generations to predict the long-term dynamics of inequality. However, recent studies show that extrapolating estimates from two generations to more than two generations underestimate the persistence in inequality. My results show that the traditional two-generation paradigm is well suited to predict persistence in earnings, but strongly underestimates persistence in education. The intergenerational income elasticity (IGE) is around 0.22 between two and 0.05 between three generations. The elasticity for education is 0.31 between two and 0.08 between three generations. I test two theories that could explain this excess persistence: the Clark hypothesis and the direct grandparental effects model. I can reject Clark’s hypothesis of a latent persistence as large as 0.75, but cannot reject the direct grandparental effects model.