Choose Your Ancestors Carefully

The Son Also Rises: Surnames and the History of Social Mobility
by Gregory Clark
Princeton University Press, 2014, 384 pages, US$24.95

The Son Also Rises could well have been subtitled “a critique of wasted breath”. After 1945, Western countries became obsessed with social mobility. The Left talked about access and equity. The Right talked about opportunity. Governments of all stripes spent trillions of dollars to increase social mobility. The lesson of Gregory Clark’s fascinating book is that it was all pointless.

Social mobility is no greater in the West today than it was in medieval England. Upward mobility in Sweden today is no better than in the eighteenth century. Nor is it better in contemporary Sweden than in the United States. The wealth of East Asia exploded in the past half-century. Modernity came to Asia aggressively. But this had relatively little impact on who climbed up or slid down the social ladder. Neither huge spending on higher education nor even mass-scale totalitarian murder of millions made a difference. All government schemes to increase social mobility end in failure.

This does not mean that social mobility does not exist. Individuals, families and social groups rise and fall over time. But the time scale is very long. Social mobility is glacial. It occurs at a snail’s pace over centuries. Yet states, encouraged by social science, invest massively in the expectation that they can effect major changes over short periods. This is an illusion. Across two generations, between parents and their children, churn occurs. Children do better or worse than their parents. But the net result of millions of movements up and down the social scale is only ever incremental change to the long-term underlying social pattern.

We can compare the wealth, income, education and occupation (WIEO) of parents and children. If the two are perfectly correlated then we can predict everything about a child’s social position from the parents’ social position. No mobility, up or down, occurs in this case. The correlation coefficient is 1. Conversely, 0 represents complete intergenerational mobility. In this circumstance, nothing can be predicted about a child’s outcome from their birth. Social science conventionally looks at mobility data across two or three generations. That data shows significant changes in WIEO. Looking at the historical short-run creates an illusion. Suppose there is a strong, 0.75, correlation between one generation and the next. Using regression analysis, 50 per cent of the child’s position can be predicted from the parents’ WIEO. Repeat that for a couple of generations, and family WIEO quickly regresses to the social mean, in three to five generations.

Nothing lasts, it seems. Except that it does. Clark’s remarkable insight is that the strongest correlation of generations is not that of parent and child but between grandparent and grandchildren, or great-grandparent and great-grandchild. Thus a child from a high-status group may slide down the social ladder but there is a high chance that the child’s child or the grandchild’s child will slide back up the social scale.

Novelists and ordinary people observe this all the time. Social science screens it out, in part because social science has a progressive bias. Social democrats and social liberals desire a correlation coefficient of 0. They deplore a correlation of 1. Yet social reality over the long run is actually much closer to 1 than 0. That is true across a thousand years of data. A hundred years of social democracy has made no difference to the historical pattern.

Clark uses a biological metaphor to explain what happens. He points to the studies of adopted children. The WIEO of the biological parents of these children is a massively better predictor of adult socio-economic status than the WIEO of the adopting parents. Thus family environment is a minor cause of mobility. The major cause is inherited traits. But Clark notes that these operate in two ways. The first is the phenotype of individuals. An individual’s inherited traits interact in contingent ways with one another and with the social environment. Distinct personalities emerge, and vary across generations. The second way of looking at traits however is the genotype of families. Some families have very powerful traits, which play out in individuals across generations in different ways. Traits will recede and dominate, clash and combine, negate or underscore each other and the surrounding social environment. Sometimes the results are brilliant, sometime disastrous, but more often than not above average.

Clark has found an ingenious way to map long-term social mobility. He identifies rare high-status surnames. He calculates the incidence of these names in the general population, then compares that with the incidence of the surnames in key professions (medicine and law), in property and income tax registers, and the registers of major universities. He and his co-investigators draw on a large body of data from England, Sweden, the United States, Japan, China, Taiwan and Korea. They find rare surnames that are uncommon among the general population but that frequently occur in tax, property, university and professional registries. They track these names over centuries. They find that the incidence of the rare surnames in the registries declines over time but does so slowly. For successful rare surname cohorts, the correlation between generations is typically 0.75 over the long run. Sometimes it is as high as 0.9. This happens under almost all imaginable circumstances, modern and medieval. That means that the sum effect of all of the participation, equity, access and opportunity schemes of social liberalism and social democracy is zero.

The Maoist dictatorship in China killed seventy million people in the name of revolutionary equality. Yet the long-term correlation coefficient for Qing era elite surnames (the Shens) from 1912 to 2008 is 0.8—highest for professors (0.9), 0.8 for company board chairs, 0.74 for central government officials. Neither mass murder nor the era of mega-growth after Mao had any profound effect on the historic pattern. Rare samurai surnames in Japan likewise remain socially influential. High-status surname groups begin at a point in time with WIEO vastly above average. Eventually this will regress to the social mean. All elevated positions erode over time, but it takes centuries, and even in some cases, millennia. Even snails move faster. Take the case of the high-status surname group of the post-Norman-era English indigenous elite: the Berkeleys, Hiltons and Pakenhams. From 1170 to 1350 the incidence of these names at Oxford and Cambridge is four times their incidence in the general population. It then takes to 1980, six centuries, for that rate to regress to one.

Another rare surname group derives from landowner names appearing in the Inquisitions Post Mortem (IPM) for the years 1236–99. This was a mix of Norman and English names: the Barclays, Stanleys, Keynes, Mainwarings and Mortimers. Clark tracks that cohort from 1440 to 1858 through the records of the Prerogative Court of Canterbury (PCC), the old probate court of upper-class England. The incidence of IPM rare surnames in the PCC registry initially is eight times more common than its occurrence in the general population. By the mid-nineteenth century, four hundred years later, that has dropped to around twice as common. The long-run correlation coefficient between generations of IPM names is 0.84. The Keyneses do not quickly fall and the Smiths do not rapidly rise. The Keynesian welfare state made no difference to this social law.

Social structure is durable. It erodes only very slowly over time. Its characteristics resist, ignore and defeat progressive social policies. States might as well take the multiple billions they spend on social mobility and burn them in a bonfire for all the effect they have.

The evidence that Clark uncovers is intriguing. Even if the reader is not interested in the social science, what Clark reveals about the history of surnames is eye-opening. Take the case of the very rare surname Pepys, made famous by the diarist Samuel Pepys. The name group emerged from obscurity when one of them enrolled at Cambridge in 1496. Since then at least fifty-eight Pepyses have enrolled at Cambridge or Oxford, most recently in 1995. The occurrence of the name in the general population would have predicted two or three enrolments. The nine Pepyses who died between 2000 and 2012 left estates with an average value of £416,000. This was more than five times the average estate value in England in the period. This surname group has maintained a high status over seventeen generations.

The roll-call of persistent surnames is impressive. Tim Berners-Lee is a descendant of a prominent early nineteenth-century family. The great-great-grandfather of the left-wing Guardian newspaper editor Alan Rusbridger was the well-to-do land steward to the Duke of Richmond. Sweden is still the acme of social democracy, yet rare aristocratic surnames (Gyllenstjerna) and old Latinate clerical, academic and merchant family names (Linnaeus, Celsius) from the eighteenth century remain socially prominent. They have higher taxable incomes than the average and a significantly higher rate of participation in the learned professions and the universities. Sweden’s free university tuition and high taxation have made no difference to the long-term dawdling pace at which surname groups return to the social average.

In the United States intergenerational social mobility rates are equally slow. Neither free-market opportunity nor left-liberal regulation makes an iota of difference. Surname groups that begin well maintain their status over the long term. Low-status name groups rise only very gradually. This confirms what everyone already tacitly knows. Descendants of Ashkenazi Jews (the Goldmans), Sephardic Jews (the Baruchs), the colonial Dutch rich (the Vanderbilts), the Puritan New Englanders (the Clevelands), colonial Ivy League names (the Rutgers) and Japanese immigrants (the Ishidas), having established a high status, retain it. You can put them in internment camps, like the Japanese-Americans during the Second World War, and it makes no difference over time.

No amount of entrance quotas and diversity programs in universities and no tsunami of “tax the rich” and equality campaigns have any discernable effect on the net long-term social result. Social patterns, as in the nature of patterns, persist. The position of surname groups with historic low socio-economic status changes at a listless pace. This is true of pre-Civil War African-Americans (the Merriweathers) and French-Canadian immigrants (the Gagnons). The participation of these name groups in the legal and medical professions is very low. The socio-economic standing of African-Americans has declined during the term of the first African-American President. Tellingly, Barack Obama’s father was a Kenyan. Contemporary black Africans have a better representation among US doctors than do white Americans.

Race, that marker of nineteenth-century pseudo-science, is no explanation of either social mobility or enduring socio-economic status patterns. Family, on the other hand, is a powerful conduit. Parents can’t ensure that their children do as well as they do. But the fortunes of high-achieving families often rally in later generations. The traits of the family genotype persist across generations, appearing, disappearing and reappearing. Winston Churchill (1874–1965), was the descendant of the great general John Churchill, the 1st Earl of Marlborough (1650–1722). Winston’s mother was a Jerome, a New York French Huguenot family. Winston’s son Randolph was (shall we say) a disappointment. He had something of his father’s drive but not his discipline.

The tantalising mystery left unexplained by Clark is how and why prominent surname groups come into being. This is a question not of inheritance but of origins. It is hard enough to understand how social aptitude is passed on; it is even more difficult to understand how it begins in the first place. Looking at Clark’s evidence, it is possible to start to make some guesses about the origins of successful surname groups by considering the shared characteristics of these groups across time and space. Literacy is a common theme. The descendants of highly literate (which also means tiny) religious groups achieve and maintain high socio-economic status. That is true of the Egyptian Copts, the Ashkenazi Jews, the New England Puritans, the French Protestant Huguenots, and the Dutch Republic’s mercantile Calvinists. High levels of literacy point to social exclusiveness through achievement. Exclusiveness is also often the converse of exclusion. In fact, exclusion may be more closely associated with achievement than inclusion.

This is true of high-achieving non-assimilated endogamous religious minorities in the Muslim world, such as Assyrian and Armenian Christians and Zoroastrians. The same applies to the caste-exclusiveness of various Christian minorities in India or to the Banias, the Indian occupational caste, whose off-shoots are influential in the United Kingdom, the United States and Canada. You can find online sites advertising “Bania IT Software Engineers Matrimonial Grooms and Brides”. Egyptian Christian Copts today are the most sociologically over-represented cohort among US doctors. Their ancestors were a dhimmi group in Muslim Ottoman Egypt. They were punitively-taxed second-class citizens who still ended up possessing half of Egypt’s wealth by the end of the Ottoman era. They then lost much of that property when Nasser nationalised it.

But, as Clark observes, material fortune in one historical moment tells us almost nothing about long-term social fortune. Winning a lottery, for example, makes little difference to one’s children’s WIEO. Grants, give-aways and hand-outs have no lasting effect on who has what. It is clear that neither the direct inheritance nor grants of WIEO are significant. Nor is the key difference religion in the broad sense. High-status surname groups often overlap with religious groups. But religious membership is just as likely to correlate with low social status.

What religion and high status do appear to share is literacy. The same may also apply to secular high-achieving surname groups. It would be interesting to see a study of persons like Robert Fitzharding (c. 1095–1170), the effective founder of the Berkeley family dynasty, one of England’s most successful. This merchant financier of Henry Plantagenet was an Anglo-Saxon nobleman who thrived in Norman England. He founded St Augustine’s Abbey in Bristol and a family line that even today still owns much of its eleventh-century lands. Fitz is Norman French for “son of”. Harding is Old English for hardy, brave or strong. If hardy-brave-strong character traits can be inherited, that would be very useful in the ever-shifting sands of high politics especially if combined with an inherited capacity for literate life. The founding of an abbey suggests a literate disposition. The result, a mix of hardiness and literacy, is a powerful thing.

The illusion of social liberalism and social democracy is that public policy can over-write this complex story of inheritance, trait and personality. Modern liberals in England thought they could tax inheritance out of society. They couldn’t, any more than free university tuition or the attempt to liquidate social classes in China could. Society is not progressive. It is conservative. It does change but it changes slowly. Through change, continuities assert themselves. From agrarian to industrial to post-industrial society, technology, art, occupations, tastes and judgments changed. Yet through even the most momentous transformations, some things remained quietly impassive.