Oct 27 2015

Research Seminar: Using Linked Census and Administrative Data to Analyze Multigenerational Mobility: 1850-2013 -Catherine Massey (U.S. Census Bureau)


An abundance of recent research examines intergenerational mobility from parent to child. There is strong evidence that mobility is the same for children entering the work force today as it was for those born in the 1970s. However, estimates over longer periods suggest that mobility has not been stable, including some evidence of decline or of large gains from 1940 to 1970 with potential declines afterwards. In addition, research suggests that the mobility of children observed from 1850 to 1880 was greater than the mobility of children observed in the 1970s. These analyses, however, cannot speak to changes in mobility over the entire course of the nineteenth and twentieth centuries or how accounting for more generations of a family may affect our understanding of intergenerational transmission of status.

Due to data limitations, analyses of mobility are often limited to two generations. Attempts to examine more than two generations have been limited to small samples, non-U.S. data, and comparisons of income averages by surname over time. To overcome these issues we link individuals across U.S. census data to examine changes in intergenerational and multigenerational mobility in the U.S. from 1850 to the present. Our analysis examines both occupational and income mobility, as well as mobility in education attainment.

We create six-generational linkages using the 1850-2000 decennial censuses and the 1973-1990 Current Population Surveys (CPS). Then we create a three-generational link using the 1850, 1880, and 1910 decennial censuses and another three-generational link between the 1940 Census, the 1973-1990 CPS, and Census 2000 to compare multigenerational mobility over time. Lastly, we observe and compare mobility across multiple two-generational links using linked 1850-80, 1880-1910, 1910-40 decennial census data, the 1940 Census linked to the 1973-1990 CPS, and the 1973-1990 CPS linked to Census 2000.


Catherine Massey, of CARRA, the Center for Administrative Records Research and Application, at the U.S. Census Bureau is working on the Core Longitudinal Infrastructure Population Project (CLIPP) that is developing methods to improve linkages across large demographic data sets. One CLIPP project examines intergenerational mobility using the 1940 census, CPS data from the 1970s, and the 2000 census and is an excellent illustration of the capabilities of CLIPP.