Search Results

The present paper examines the extent of intergenerational earnings mobility in Denmark, Finland, Norway, and the United States. We examine income mobility among pairs of fathers and sons as well as fathers and daughters using both mobility matrices and regression and correlation coefficients. Our results suggest that while all countries exhibit substantial income persistence across generations, especially in the tails of the distribution, there is greater persistence of rich rather than poor incomes among men, there is less income persistence in the Nordic countries and daughters are more mobile than men.

We use two data sets for the U.S. The first sample is taken from the National Longitudinal Survey of Youth (NLSY79) and the second is the 2000 wave of the NLSY-Child which contains information from 3,425 women with 8,323 children. Combining information from the NLSY79 marital/cohabitating history of their mothers and questions about family structure in the NLSY-Child we create a family structure history. This allow us to distinguish between children living with married biological parents, cohabiting parents, single mothers, mothers married to stepfathers, and mothers cohabiting with unrelated males. We use a number of educational outcomes. For children from ages 5 to 15 we have assessment instruments including three Peabody Individual Achievement Tests (PIAT) for reading and math and the Behavior Problems index which measures a child's anti-social behavior. For children over the ages of 15 we observe school enrollment status and highest grade completed. Explanatory variables include demographic characteristics, parental education, number of siblings, and family income.

For Sweden we use a random sample of children born 1974-84 drawn from the population registers. The data sample roughly 20 percent of Swedish children born each year from 1974-84 and their siblings. The total sample size is over 300,000 child observations. This data is combined with family and individual information from the censuses from 1965, 1970, 1975, 1980, 1985 and 1990. Our outcome variables are grade point average at age 16, high school graduation at age 19, and years of schooling and earnings for older children in the sample. We create marital history--length of cohabitation and length of marriage-- for the parents using information from the bidecennial censuses from 1990 and before and tax records after 1990. Our explanatory variables include the sibling composition of the household (his children, her children, and their joint biological children), the educational attainment and earnings of the adults in the household, and whether the family lives in an urban area.

Identifying the causal effect of marriage on outcomes is complicated by the selection problem. We use IV-methods as well as fixed-effect models to deal with this problem.