Working Paper: CEPR ID: DP17788
Authors: Gregory Clark; Neil Cummins; Matthew Curtis
Abstract: There is considerable interest in comparing intergenerational social status mobility across time and place. But such attempts are vitiated by unknown measurement errors in status indices, errors that also vary over time and place. Typically the more error, the more apparent social mobility. Using a new database of 1.7 million marriages in England 1837-2021, we show how improving the quality of an occupational status index generates lower implied social mobility. Without control of the errors embodied in all social status indices, attempts to compare social mobility across time and place are impossible. This paper develops a solution to this problem using not absolute correlations, but the relative correlations between different family members. The method deployed here suggests that occupational status persistence in England 1837-2021 was always much greater than conventionally measured, and was largely unchanging 1837-2021.
Keywords: No keywords provided
JEL Codes: No JEL codes provided
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Measurement errors in occupational status indices (J70) | Traditional estimates of intergenerational social mobility in England are significantly overstated (J62) |
| Using new method focusing on relative correlations (C10) | Find that occupational status persistence has been much greater than conventionally measured (J69) |
| Intergenerational correlation in occupational status (J62) | Remained largely constant across the period from 1837 to 2021 (N11) |
| Without controlling for measurement errors (C29) | Comparisons of social mobility across different time periods and locations are fundamentally flawed (J62) |