Friday, December 20, 2013

Family wealth persistence over several centuries

Social mobility has been much studied to understand how the poor have a shot at becoming rich and how the rich manage to preserve their status. Such studies are usually limited to mobility during a lifetime for a single individual or for a family from one generation to the next. Going beyond this time frame is virtually impossible, because there is no panel dataset for wealth or income that spans over several generations. One can, however, discover some interesting proxies that allow to create such a dataset.

This is what Gregory Clark and Neil Cummins do in a pair of papers that exploit the fact that people with rare surnames are highly likely to be from the same family. Using national birth and death registries for England and Wales as well as probate registries that recorded wealth at death, they gather records for 21,618 people over about 150 years in the first paper. The second paper focuses on educational status instead of wealth over eight centuries and uses registries of students at Cambridge and Oxford universities as well as censuses for the rest of the population. In both cases, intergenerational correlations are estimated to be much higher than in studies with shorter samples. It can take 20 to 30 generations for an initial status to disappear. This may be indicative that social mobility has increased in recent generations in England and Wales (my interpretation, although Clark and Cummins argue that intergenerational persistence is stable over centuries despite stark changes in inheritance taxation) or that families have an underlying social status that changes much more slowly than characteristics that are easier to observe (the authors' interpretation).

PS: If you are looking at the papers, do not be surprised to see the same abstract on both. Very negligent LSE staff posted similar cover pages on both papers.


Quid pro quo said...

150 years is a long time -spread over two world wars and drastic change in economic power distribution amongst countries. If statuses are so sticky, then the question is what drivers would cause them to change?

David Stern said...

Do people not change their names to those of rich or well-known people to try to improve their social status?