We propose a twin family model linking twins with their spouses and children to quantify the relative importance of genetic and environmental factors in explaining the variance of socio-economic outcomes. Using data from the Danish Twins Registry and population registers, we test and relax the assumptions of the standard behavioral genetics model most frequently applied in economics using twins or adoptees. Exploiting an education reform differentially affecting parents, we find no evidence of gene-environment interactions. While we find some assortative mating based on genetic factors, differentially shared environments are key: they explain half of the variance in years of schooling, whereas genetic factors explain only nine percent. We find similar percentages for earnings, income, and wealth. Decomposing intergenerational elasticities reveals that shared environments explain 50% for earnings, 60% for income, 70% for wealth, and 80% for schooling. Family environments are more important than previously understood.