The slope of these lines represents an inverse measure of relative mobility, because steeper lines indicate greater relative advantage for children born to parents with higher education or income. Figures 5 and 6 display children’s final schooling by parental final schooling and parental income deciles by year, focusing on the white population. As a first application of the method, I examine long-term intergenerational mobility trends in the US. The method outlined here opens up a wide range of new possibilities for research on intergenerational mobility because large, cross-sectional data sets such as the US census are more widely available over time and space than panel data sets. Intergenerational mobility estimate in census data by intergenerational mobility estimate in tax data in 2000, by state This congruence is not surprising given that the assumptions underlying the census adjustment appear valid.įigure 4. The figure suggests a striking correlation between the adjusted intergenerational mobility statistics in census data with conceptually analogous intergenerational mobility statistics based on administrative tax data covering the full US population. Figure 4 compares these mobility estimates by state. The resulting mobility estimates can also be validated directly against recent estimates from tax data in the 2000s. Schooling by parental home value decile for dependent and independent children in matched 1930-40 census data Schooling by parental income decile for dependent and independent children in the PSID, pooling years 1968-2011įigure 3. Schooling by parental schooling for dependent and independent children in the PSID, pooling years 1968-2011įigure 2. The assumption makes it possible to adjust intergenerational mobility statistics in census data to account for the missing 65% of children who cannot be linked directly with their parents at ages 22-25 over the 1940-2000 period.įigure 1. I find that this assumption cannot be rejected in every panel data set in which it can be tested in the US historical record. The key takeaway is that children’s average final schooling is related to parental characteristics for dependent and independent children in similar ways. Figure 3 shows this again for parental home values in a panel linking children from the 19 censuses. Figures 1-3 show this insight for recent decades in a leading panel data set, the Panel Study of Income Dynamics, by plotting children’s final schooling at ages 22-25 by parental education and income, respectively. The main insight is that children who live independently from their parents in their mid-20s are much more similar to children who still live with their parents at these ages than researchers had previously assumed. I develop a simple adjustment for these ‘missing’ independent children that allows for estimation of the relation of children’s final schooling to their parents’ final schooling or their parents’ income. This is because the census only links parent and child outcomes while children still live with parents, and children rapidly become independent after age 17 but before any adult outcomes can be observed meaningfully (Cameron and Heckman 1993). Prior research on intergenerational mobility has largely ignored census data. In this column I discuss a new method to estimate intergenerational mobility statistics on US census data (Hilger 2015). Measuring intergenerational mobility in census data The lack of reliable, longer-term trends is unfortunate because the high school movement, early GI Bills (formally known as Servicemen’s Readjustment acts), Great Society programmes, several key Supreme Court decisions, and the Civil Rights movement all predate the availability of most panel data sets. While new administrative data sets are improving intergenerational mobility estimation in more recent periods, they do not shed light on long-term historical trends and they lack information on individual race and parental education (Chetty et al. Many panel data sets contain this information but they begin in the 1960s and are too small to examine mobility over time or subgroups with precision (e.g. No standard government data set has collected this information historically in the US. The main empirical problem is that measuring intergenerational mobility requires data on labour market outcomes for both parents and children. However, surprisingly little is known about intergenerational mobility variation over time, space and groups. Intergenerational mobility is an important social objective for many individuals and policymakers, and may affect public attitudes toward other social objectives such as equality and growth (Piketty 1995, Benabou and Ok 2001, Corak 2013).
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |