“The Shadow of Peasant Past: Seven Generations of Inequality Persistence in Northern Sweden”, Martin Hällsten, Martin Kolk2023-05 (, )⁠:

[Twitter] The authors use administrative data linked to parish records from northern Sweden to study how persistent inequality is across multiple generations in education, occupation, and wealth, going from historical to contemporary time.

The data cover 7 generations and allow the authors to follow ancestors of individuals living in Sweden around the new millennium back more than 200 years, covering the mid-18th century to the 21st century. In a sample of around 75,000 traceable descendants, they analyze (1) up to fifth cousin correlations and (2) dynastic correlations over 7 generations based on aggregations of ancestors’ social class/status.

With both approaches, the authors find that past generations structure life chances many generations later, even though the results align with traditional stratification research in that mobility across multiple generations is high.

The results imply that today’s inequality regime may have been formed many generations back.

Figure 4: Cousin correlations in years of education with iteration based on sibling correlation. Men and women born 1940–87 in Sweden with a fifth-generation descendant in northern Sweden.

…We find that the results for occupation closely resemble those for education, whereas wealth, not surprisingly, is slightly different. This is due mainly to mobility patterns in proximate generations. However, when we focus on the role of more distant kinship, the similarity in persistence is very striking. It is interesting to note that, similar to several recent studies, the persistence rate across generations beyond grandparents is rather stable and incompatible with a first-order (parent-child) Markov process. However, even though we can fit up to a 6th-order Markov model with the dynastic correlations approach (ie. with generational lags under mutual control), the direct associations from more distant generations are small and do not contribute much more inequality than what already resides in parental and grandparental coefficients. Taken together, these findings suggest that (1) most transmissions in the earliest generations are sequential from generation to generation, but what is transferred remains mostly intact rather than decaying, and (2) what is transferred is only a partial explanation of life chances within that generation. This is close to the latent factor model of Clark & Cummins2014 and Braun & Stuhler2018, where some endowment is transferred across generations and then translated into inequality within that generation. In their model, high multigenerational correlations will occur without the existence of transfers that skip a generation. The important point here is that the model implies that the transfer of a trait from parents to children is strong and that the translation of the trait into outcomes is weak; this is what creates this persistent pattern.

See Also:

View PDF:

The Shadow of Peasant Past: Seven Generations of Inequality Persistence in Northern Sweden