Estimating Social Mobility Rates from Surnames: Social Group or Dynastic Transmission versus Family Effects

Gregory Clark's webpage at UC Davis

Abstract: Grouping individuals by surnames, whether or not these surnames correspond to ethnic subgroups, results in high intergenerational correlations of status: in the range 0.7-0.8. The Son Also Rises (Clark et al., 2014) argues this corresponds to a pattern of individual family mobility, where whatever measured short term mobility is, there is strong underlying status correlation within families. Torche and Corvalan, 2016, argue that surname status persistence instead just captures group level persistence. Within social groups there will be the conventionally measured social fluidity. Adermon, Lindahl and Palme, 2016, claim instead that the observed persistence stems from social network effects within family dynasties, not again the individual families. In this paper I show that the individual persistence model of The Son Also Rises has different empirical implications to group or dynastic persistence models. I show also that data from an extensive English lineage born 1790-1929 supports the individual persistence model better than these alternatives.