Surnames and the laws of social mobility

by on October 19, 2012 at 5:21 am in Uncategorized | Permalink

Here is some new work by Gregory Clark (pdf):

What is the true rate of social mobility? Modern one-generation studies suggest considerable regression to the mean for all measures of status – wealth, income, occupation and education across a variety of societies. The β that links status across generations is in the order of 0.2-0.5. In that case inherited surnames will quickly lose any information about social status. Using surnames this paper looks at social mobility rates across many generations in England 1086-2011, Sweden, 1700-2011, the USA 1650-2011, India, 1870-2011, Japan, 1870-2011, and China and Taiwan 1700-2011. The underlying β for long-run social mobility is around 0.75, and is remarkably similar across societies and epochs. This implies that compete regression to the mean for elites takes 15 or more generations.

Here is NPR coverage:

“If I just know that you share a rare surname with someone who was wealthy in 1800, I can predict now that you’re nine times more likely to attend Oxford or Cambridge. You’re going to live two years longer than an average person in England. You’re going to have more wealth. You’re more likely to be a doctor. You’re more likely to be an attorney,” Clark says.

Dylan Matthews offers some charts.  For the pointer I thank Fred Rossoff.

1 Michael Heller October 19, 2012 at 6:07 am

Drat. Is mother’s maiden name worth a try?

2 prior_approval October 19, 2012 at 6:33 am

Seems interesting, until recognizing that surnames may not be precisely the best measure – after all, 50% of the sample loses theirs every single generation, though intermarriage is certainly part of the process. (I believe not a single country studied uses matrilineal naming conventions.)

Not to mention that other markers of being elite, such as profession, where prohibited or restricted in the countries studied for that same 50% of the sample.

And comparing multiple countries with monarchies will almost certainly pretty much prove what people who look at monarchy already know – family relations count. One could almost say family feudal, in the pre-TV game show era.

3 Millian October 19, 2012 at 10:39 am

50% of a very large sample.

However, there is an excellent point in prior_approval’s contribution that we are talking about social mobility through the male line only. People devoted more resources to male children.

4 Bob Jones October 19, 2012 at 10:53 am

Depends on if you’re a member of the Tribe.

5 prior_approval October 19, 2012 at 6:49 am

Having just finished my honey and jogurt (boring peasant food or elite foodie combo?), then cleaning up, it occurred to me that this fearless research conclusively proves that social classes that perpetuate themselves remain social classes over a longer time horizon, given enough stability. (Students of monarchy/aristocracy, stop yawning – and it is noticeable that neither Russia nore France were included in this study, both societies having made a fairly thorough sweep of their elites during extended and bloody revolutions.)

And equally tautalogically valid, those who enter the elites but don’t become part of the social class, will likely fall out of the elite class due to tiny size. For example, think about the billionaires Jobs, Ellison, and Buffet -compared to Gates, they are meaningless in terms of social class. Whereas Gates is part of a fairly long running family of what can be identified as the American elite social class (it also helped Gates to make his billions – nothing like mommy having her ear to ground at a major computer company at the dawn of the PC era).

In other words, current research demonstrates that Marx was correct in this area of social observation.

6 Cliff October 19, 2012 at 9:55 am

Marx was correct about what, exactly?

7 dearieme October 19, 2012 at 10:07 am

It doesn’t matter, Cliff. If Marx was right, he wasn’t original; if original he wasn’t right. Well known fact, that.

8 bonami October 23, 2012 at 12:01 am

For once, Cliff, try in good faith to answer your own question.

9 Wonks Anonymous October 19, 2012 at 10:32 am

China is included and it did have a bloody revolution.
“Figure 22 shows these rates of relative representation. Note the relatively smooth decline in the over represention of these surnames, both through the Imperial Period, but also through the Republican Era (1912-49) and the Communist Era (1949-).”
Clark notes that he does not have data to see whether the Cultural Revolution changed anything.

10 Brian Donohue October 19, 2012 at 1:01 pm

your awesome tautologies are awesome!

11 Steve Sailer October 19, 2012 at 7:30 pm

Warren Buffett’s father was a U.S. Congressmen, and Buffett spent much of his younger years in Washington D.C., which helps explain his very close personal relationship with the Washington Post dynasty, which helps explain his relentlessly good press.

12 Rory Sutherland October 21, 2012 at 4:32 pm

And Buffet is descended directly from French Hugenots, a pretty elite business caste in their way.

13 Andreas Moser October 19, 2012 at 7:29 am

There should be a surname lottery every generation. – But then rich folks would bestow pompous first names upon their offspring. – So maybe there should be a complete name lottery. Or we will all just have numbers.

14 IVV October 19, 2012 at 9:30 am

Pompous first names are already bestowed upon offspring, which are then picked up in turn by progressively lower classes (no one can stop a parent from bestowing a name) until the elite have moved on to new names and the old name is considered lower class–compare the trajectory of Courtney over the years. Nowadays, Henry and Cecilia are the new rich names.

I know the situation is different in Europe. My German wife had a wonderful time when she realized that we could give our child a “von” surname without challenge in the United States.

15 Piros October 19, 2012 at 11:20 am

That is terrible, do you have any idea how much toadying and bootlicking it took to get that von in the old country?

16 Rory Sutherland October 21, 2012 at 4:33 pm

Is that what Treyvon Martin’s parents were planning to do?

17 Unanimous October 19, 2012 at 4:10 pm

If we all get numbers, who gets to be 1? I’d like 007.

18 Tom West October 19, 2012 at 6:03 pm

Sorry, you are number 6.

19 Russ R. October 19, 2012 at 7:41 am

A narrow focus on the wealthy perpetuates envy, and ignores the real reason why we should care about social mobility, namely, the poor and the opportunities available to them.

Does sharing the same last name with someone who was poor in 1800 have any real predictive ability on one’s social status today?

20 Millian October 19, 2012 at 10:44 am

Probably not, but that’s because (a) there were so many poor people when inequality was so much larger, and (b) rich people were more diverse and cosmopolitan than poor people back then, so there are proportionally more data points when discussing names. You provide no evidence why we should care about social mobility for only one reason.

21 Cyrus October 19, 2012 at 7:48 am

Both results can be true if at any level of status, those who differ substantially from their parents are less likely to reproduce

22 Corporate Serf October 19, 2012 at 8:24 am

At least for tracking surnames and family histories in Bengal, a better source would have been temple records from the famous temples/ pandas. When people visited the temples for pilgrimage, they would write down their names etc. These records do go back generations and the family linkages are usually kept.

Language / knowledge of this particular cultural tidbit might have been an issue for these researchers.

23 Rahul October 19, 2012 at 8:32 am

Wasn’t there a similar paper using surnames to prove nepotism in Italian academia? I forget the authors but was posted on MR a while ago……

24 Roy October 19, 2012 at 10:00 am

How the heck are surnames of any use in Taiwan? A colonial country with vast social dislocations and a very limited number of surnames? Not everyone named Lin is related.

25 Matt W October 19, 2012 at 11:03 am

The paper focuses on rare surnames that have, at some point in the past, been indicative of social class.

26 Roy October 19, 2012 at 11:16 am

That is exactly my point. It is either useless or it fundamentally can not understnd the nature of surnames in frontier Taiwan.

I understand Taiwanese geneological records are aweome and they can be usedfor some fantastic research, but I hav no idea how this sort of work would be valid just using surname data.

The conclusion strikes me as very reasonable, and a surname technique sounds great in England, Sweden, and Japan. But it makes little sense in a Taiwnese context.

27 Linda Seebach October 19, 2012 at 11:28 am

Not much use in China, either, where the idiom for “Joe Sixpack” is “Old Hundred Names,” because that’s roughly how many surnames there are. “Rare surnames” doesn’t mean the same thing in China as in England. Also, the “same” surname may include people whose names in characters are different.

28 Matt W October 19, 2012 at 11:58 am

I have no idea if this is valid or not, but he lays out the methodology pretty specifically:

“[I]f we turn to the level of specific provinces, we can find relatively infrequent surnames which differ in average status from the mean. Here we look at the province of Zhejiang, lying on the coast just south of Shanghai. We establish the relative frequency of surnames in the Province in the general population
taking as a metric the surnames of 5,673 soldiers from this Province dying in the war against Japan, and in the Civil War, 1937-1949. We then establish which surnames were overrepresented among the Jinshi, the highest achievers under the imperial exam system before 1905, who were born 1670-1699. These surnames, less than 4.5% of the 1937-49 population, identify an early imperial elite, 19.5% of the Jinshi
from the Province born in this era. They are more than four times as likely among the Jinshi from Zhejiang as among the general population. We measure what happens to the relative representation of these surnames among Jinshi, and then among equivalent elites, for those born 1700-1969…”

29 dearieme October 19, 2012 at 10:09 am

How many Britons were surnamed Guelph in 1800?

30 whatsthat October 19, 2012 at 10:25 am

How old is the idea of the welfare state? In practice I mean. When did countries get rich enough to look after the poor?

31 david October 19, 2012 at 10:35 am

Germany under Bismarck, and Britain under Lloyd George. In both cases it was enacted by establishment figures concerned that socialist parties were gaining momentum (compare FDR and Huey Long).

32 Millian October 19, 2012 at 10:51 am

The British welfare state predates Lloyd George. Some education and public health services were provided by local governments. Lloyd George’s great contribution was to fund pensions through progressive taxation rather than regressive tariffs or death duties.

33 JonF October 20, 2012 at 5:19 pm

Welfare states are very, very old, allowing for different limits of the possible in the past. Genesis relates how Pharaoh on Joseph’s advice stored up grain during the seven fat years then doled it out during the seven lean years. The story may or may not be true, but it was regarded as believable at the time it was written. And we do know that the vast monuments of Egypt were built with otherwise unemployed labor– in effect they were early examples of public works.

The Roman grain dole is well known. Greek and Roman cities also paid public doctors to treat the poor (though citizens only). And in the Middle Ages the Church provided the functions of the welfare state.

34 Yancey Rothschild October 19, 2012 at 12:15 pm

I have taken appropriate measures.

35 Brian Donohue October 19, 2012 at 1:03 pm


36 Rich Berger October 19, 2012 at 4:03 pm

Oy vey ist mir! More fodder for the perennially aggrieved.

37 Steve Sailer October 19, 2012 at 5:49 pm

Popular culture in Britain is increasingly dominated by aristocrats and high bourgeoise types. Singer James Blunt is from the officer-class Blount family (he shortened the name so that proles could have an easier time pronouncing hit) that traces its military lineage to back before William the Conqueror. Actor Xander Keynes (Edmund in the Narnia trilogy) is a direct descendant of Charles Darwin and a collateral descendant of John Maynard Keynes.

38 Steve Sailer October 19, 2012 at 5:56 pm

Nathaniel Weyl wrote two fascinating books of American surname analysis about 50 years ago. Here’s a summary:

39 Steve Sailer October 19, 2012 at 6:01 pm

In American history, probably the most consistently accomplished surname is “Huntington,” which traces to a single Puritan family that arrived in the 1630s: Wikipedia lists a couple of dozen noteworthy American Huntingtons, such as the late Prof. Samuel Huntington of Harvard. Huntington Gardens, Huntington Beach, and Huntington Park can be found in Southern California alone.

40 Steve Sailer October 19, 2012 at 6:22 pm

The British liberal intellectual elite of the mid-19th Century — the Darwins, Galtons, Wedgewoods, Arnolds, Huxleys, Keyneses, etc. — tended to marry amongst themselves, and remained extremely over-represented in the intellectual elite for at least a century, perhaps longer.

Another example was the Waugh-Cockburn clan of writers, which is going strong today.

41 JonF October 20, 2012 at 3:16 pm

This study tells us nothing about the general case for social mobility, only about how hard it is to break into the elite ranks. However that’s a tiny slice of society, and one with the resources to create high barriers to entry, and also insure its current members against falling out of the favored space. A more important question is how easy is it for a lower income person to reach the middle class.

Comments on this entry are closed.

Previous post:

Next post: