How universal are rates of social mobility across time and societies?

Gary Solon, in a new survey paper, takes issue with the earlier results of Greg Clark, which had suggested social mobility was roughly constant across a wide spectrum of cases.  Solon writes:

…the results reported by Clark do not reflect a universal law of social mobility.  Quite to the contrary, other studies based on group-average data, even surnames data, frequently produce intergenerational coefficient estimates much smaller than Clark’s.

A second testable prediction of Clark’s hypothesis…is that instrumental variables (IV) estimation of the regression of son’s log earnings on father’s log earnings should yield a coefficient estimate in the 0.7-0.8 range if father’s long earnings are instrumented with grandfather’s log earnings.  When Lindahl et al, estimated that regression with their data from Malmo, Sweden, the IV coefficient estimate was 0.15, considerably higher than their ordinary least squares (OLS) estimate of 0.303.  They obtained a remarkably similar comparison of IV and OLS estimates when they used years of education instead of log earnings as the status measure.  The pattern of IV estimates exceeding OLS estimates is consistent with Clark’s general story about measurement error in particular indicators as proxies for social status.  It is equally consistent with all the alternative stories listed in section II for why grandparental status may not be “excludable” from a multigenerational regression.  What the results are not consistent with is a universal law of social mobility in which the intergenerational coefficient is always 0.7 or more…

A third testable prediction…is that using an omnibus index that combines multiple indicators of social status should make the intergenerational coefficient estimate “much closer to that of the underlying latent variable.”  [But]…The resulting estimate was not “much closer” to the 0.7-0.8 range.

In sum, when Clark’s hypothesis is subjected to empirical tests, it does not fare so well.

Here is an ungated version.


My question regarding surname analyses has been how much does name-changing toward higher class names fouls up Clark's findings:

For example, the Sailers used to be the Seilers until one social climber in Wil, Switzerland decided that was too bluecollar of a name.

The Wellesleys were before that the Wesleys, and before that the Cooleys. I am referring to the family of the Duke of Wellington.

And when go back far enough spelling was fluid and there was no agreement, even within families, how a name should be spelled. The late medieval social-climbing English clan we know as the Woodvilles never once spelled their surname name that way; it appears as Wydville, Wideville, Wydeville, Widville, and several other permutations.

Since grandfather's income and wealth doesn't matter, then no harm in having higher estate taxes.

"Since... doesn't matter...". The paper indicates the jury is still out. 'not... "excludable" ' means it does matter. I'd be interested in the Lindahl data set; whether it is pre- or post- the time point when Sweden abolished their inheritance tax.

Not excludable means not proven as well.

Fair enough. I should have written: "it may matter".

Or, it may matter, not, but since there is no proof...raise the estate taxes.

'First, do no harm ....' What exactly is your goal, Bill? To increase social mobility by reducing the incentive to leave your loved ones better off? Doesn't matter to me, personally, I'll pass it all on before I die anyway. Can't take it w me. See you in heaven.

Ryan, Read Picketty. Your loved ones will have to work for a living like everyone else, unless you do not believe that there efforts are too weak to be rewarded in the market.

Doesn't this assume that the average persistence of a group of families (names) equals that of an individual family? But if there are reinforcing tendencies to group support (networking, etc.) then focusing on individual families ignores the effects of strong outsiders marrying up (and into) the family while some of the less successful descendants marry out of the family.

The Sailer issue about people taking the names of the elite isn't a problem because Clark compares (for example) the relative prevalence of rare names in elite professions compared to average names using the distribution of names at the beginning and end points. If commoners take on elite names, this will only make it less likely the rare names to outperform unless the commoners who take elite names are at least as likely to do well as the socially and genetically connected elites. That is, if some names go from being 0.1% of the population but 2% of Oxford, to 1% of the population, the latter group will have to outperform approx tenfold relative to the average name in the population at the same rate as the former group (or rather 0.7 times the old performance since there is some regression to the mean).

I think SS was referring to the case where people adopt a posh name *after* they make it to oxford. i.e. a lot of the posh names observed in the data for oxford, are actually non-posh names in disguise. After people ascend in class they would feel pressure to change their name to try and conceal their lower class roots. This seems like a common enough theme. To adopt a posh name if you are obviously lower class, would only bring the contempt of everyone. Perhaps that passes for a mechanism.

Clark uses some data from countries where changing ones name isn't even possible (if I recall) or is really onerous (more then usual).

The thing about the study is not the thing you can digest without putting a lot of work in and going to the source materials. In some cases they claim the correlation is very low, in other a bit lower then clark but higher. Is it income or "status"? Grandfathers or fathers? Group or individual? This stuff was easier to follow in his book and would take more work to sort through in this paper.

You get stuff like this: "What the results are not consistent with is a universal law of social mobility in which the intergenerational coefficient is always 0.7 or more…"

I mean, doesn't that feel really gotcha? You come away from Clark's book thinking that the popular generational mobility story isn't quite as cracked up as people make it to me. I don't think I'm going to change my mind if I find out that under such and such assumptions a scenario comes out to less then 0.7, therefore throw out the guys whole book.

If they could write it in a more digestible form like Clark did I might take more of a look at it. Probably not happening during my busy season as it is.

I doubt that, until 20th century, there was, in any country, big difficulty in changing the name

I doubt that you are correct.

Right, the Sailer family story is that we were named Seiler (ropemaker) until a Seiler became mayor of a small town in Switzerland. Then the glory of it all went to his head and he felt he had to change the family name to something classier.

Clark's predecessor in surname analysis, Nathaniel Weyl, observed that people named Smythe were higher achieving than people named Smith even though Smythe is just a social climber's variant.

Anyway, Clark wrestles with this issue a few times in his book, partly reassuring my worries, but I'm not wholly convinced he's completely disposed of them.

Comments for this post are closed