Family background and income during the rise of the welfare state: Brother correlations in income for Swedish cohorts born 1932-1968

Social scientists from many academic disciplines have for a long time been interested in the association between family background, in a wide sense, and socioeconomic status during adult life. In the last 10-15 years, economists have become more active in this research field and have, in particular, focused on family background and subsequent income. Empirical research on such issues has become feasible in some countries thanks to household panel data sets (such as the Panel Study of Income Dynamics, PSID, and the National Longitudinal Survey, NLS, in the United States, and the German household panels, the GSOEP) in which families have been followed from one generation to the next. In other countries (such as the Nordic ones and Canada) administrative register data have been used to connect family members to each other. The most frequent analytical approach has been to estimate intergenerational income elasticities (or correlations) between fathers and sons. Solon (1999, 2002) summarizes such estimates for several countries. A less frequent, but very powerful, approach is to estimate sibling correlations in income. The power of this approach comes from the fact that a sibling correlation tells us what fraction of the variation in the variable of interest (such as income) that can be attributed to factors that siblings share. Siblings, who have grown up together, share both family factors (e.g., parental income) that impact on their subsequent income and neighborhood influences of various sorts. Thus, a sibling correlation is a broader measure of the importance of childhood conditions than an intergenerational one. 1 1 Note also that as long as closely spaced siblings are used to estimate a sibling correlation, income is measured for more or less the same income distribution. Parents' and offsprings' incomes are, however, by necessity measured many years apart and thus intergenerational estimates are complicated by the fact that income distributions might be very different for the two generations. In such cases the intergenerational elasticity will deviate from the intergenerational correlation. 2 Empirical research following these two approaches has produced a variety of interesting results. One of the most striking ones is the clear cross-national pattern that family background is more important for labor market achievement in the United States than in most