A Non-PC View of Social Mobility

Bloomberg's Clive Crook has written a review of economist Gregory Clark's book The Son Also Rises. This is an explosive book; Clark finds that there isn't much social mobility today, and never has been. Nor is there much in places we believe it is common, like Sweden.

Clark studies rare surnames across the centuries, small families with distinctive names, and finds their social status is remarkably stable from generation to generation, indeed century to century. Crook writes of Clark's findings:

The surname evidence asks us to believe that cultural and environmental differences don't much matter, that the advent of universal education and the modern welfare state, political upheavals, revolutions and world wars, made little difference. Through it all, social status was inherited as strongly as though it were a biological trait, like height. The surname data don't prove that genetic transmission is the main driver, but it's hard to see what else could be.

Socioeconomic status is a bundle of characteristics: income, wealth, education, profession, type of residence and so on. These individual measures are noisy. (snip) Underlying social status, the whole bundle, is much less variable, and that's what gets passed along to the next generation. So the individual measures that formed the consensus have systematically understated the real persistence of social status.

Crook concludes by expressing feelings not unlike my own:

Impressed as I am by this fascinating book, I'd love Clark to be proved wrong. And he must be wrong, mustn't he? Mobility in modern welfare-state societies is no higher than in pre-modern times? Please. Yet that's what his data -- ingeniously gathered and carefully analyzed -- seem to show. Don't tell me he's a closet racist to say these things. Tell me where his analysis went wrong.

I predict Clark's book will attract as much hate, vituperation and denunciation as Charles Murray and Richard J. Herrnstein's The Bell Curve, and for similar reasons. Because it alleges as true things we wish were not true, we are inclined to blame the messenger.

The other DrC points out to me that Clark's hypothesis may be more valid for males than for females, who have mostly tended to assume the social status of their husbands. Perhaps that is why Clark referred to "son" in his title, Crook doesn't mention gender in his review.