Editor’s Note: If you think you’re too short, you should have chosen taller parents. Thankfully, inherited physical traits, like height or eye color, aren’t often considered predictors of future socioeconomic success. In America, we’re taught that your social status, unlike your stature, can rise; it’s not pre-determined.
But what if social status is inherited, just like stature? In that world, surnames would be a pretty good indicator at birth of future social mobility. That is our world, says economic historian Gregory Clark. He makes that case in one of the most provocative economics books of the year, “The Son Also Rises,” using surname-based intergenerational inequality correlations.
Over the past week, Making Sen$e has featured Paul Solman’s never-before-published conversation with Clark about his 2007 book, “A Farewell to Alms,” broken into five parts. Clark traces economic history, beginning with hunter-gatherers, and argues that in England, where the Industrial Revolution was born, the takeoff was due, above all, to “the survival of the richest.”
That argument, like the one we present today, strikes at the heart of the belief in equal opportunity for economic growth and social mobility. “We can predict the majority of status variation among people at birth just from their lineage,” Clark writes. In other words, our society’s divergence of fortunes — which as Clark points out, isn’t just about income, but also social status – is relatively fixed. That’s something no one ever wants to talk about.
Clark is here to talk about it, and to explain the implications for public policy of a world where, he thinks, social mobility is, and is likely to remain, slow.
— Simone Pathe, Making Sen$e Editor
Everything we thought we knew about social mobility is wrong. And the evidence is hiding in plain sight. Just look at our surnames. Surname evidence suggests social mobility rates are slow, and impervious to social policy. Practically speaking, this means the poor will be much better served by policies to reduce the inequality of social outcomes than by those designed to speed up social mobility.
What we thought we knew was that high rates of social mobility were possible, but only achieved by some societies; that nurture mattered more than nature in social outcomes; that the right social policies could generate near equality of opportunity; that the U.S. was a laggard when it came to such policies; and that the right policies would eliminate in the U.S. the legacies of past disadvantage within two to three generations.
A study of the changing social status of surnames across many societies and epochs, however, dashes all such fond hopes. There is, everywhere, social mobility. All elites eventually get replaced. All underclasses eventually rise. Unfortunately, the process happens over a very long timeframe: many generations, in fact. The study of surnames shows that in the near term, social mobility rates are everywhere low and utterly unaffected by social policy; the pattern of mobility is consistent with nature rather than nurture being the main determinant of social status.
The hard truth is that underlying social status is inherited from parents as strongly and mechanically as height. This iron law of status means there is no way that a merit or incentive system can be used to justify a winner-take-all society.
Social mobility rates can be summarized by one number: the degree to which the social outcomes of children – earnings, education, wealth and health – correlate with that of their parents. The closer that correlation is to zero, the less family, lineage, race and ethnicity matter to the next generation. All is possible at birth. The closer that number is to one (in essence, 100 percent) the more status is predictable from family circumstances alone. Birth is fate.
The figure below, adapted from egalitarian economist Miles Corak, summarizes the conventional wisdom, as seen in estimates of the correlation between income inequality within families and in the economy as a whole across various countries. The correlation varies greatly. But where the income distribution is more unequal, the correlation between the economic status of children and their parents is higher. The correlation is low in prosperous, egalitarian Nordic economies, but high in in-egalitarian and poor Latin America. At the extreme, outcomes in countries like Sweden seem largely unpredictable at birth. There seems to be near equality of opportunity. (The U.S. lies in the middle both in terms of social mobility and in inequality.)
The varying strength of conventional status inheritance across countries suggests that institutions control social mobility rates, and that many societies have suboptimal social mobility rates. By this argument, nurture, not nature, determines social success and social policy is the answer. In particular in the U.S., many children of talent are denied their potential by the circumstances of their families, according to this story. The U.S. is failing children from disadvantaged families. So, in general, is any society with high inequality. But as it turns out, that’s far from the whole story.
Conventional social mobility estimates look at only the inheritance of individual aspects of social status, such as income, across single generations.
How your income correlates with your parents’ income turns out to be a very imperfect measure of how your underlying social status correlates with theirs. The most famous philosopher of the 20th century, Ludwig Wittgenstein, maintained high social status despite giving away all the money left to him by his father, the steel magnate Karl. The crucial point is that a lot of what is conventionally estimated as social mobility is just the froth around much more slowly changing underlying social status.
The key to why the correlations are misleading is not that they are based on income, but that the current measures mainly capture random transitory fluctuations in income and wealth at the individual level. But for elite groups as a whole, the systematic downwards movement of income or wealth or education is much lower.
A different approach is to use the average status of surnames distinctive enough to track families across generations. Using surnames, we can measure the inheritance of that underlying status between single generations, and also over multiple generations.
Take supposedly mobile Sweden, for example. Conventional estimates imply that social mobility among Swedes should wipe out all past advantages and disadvantages within three generations, reminiscent of the 19th century English adage, “Shirtsleeves to shirtsleeves in three generations.” But consider the name Leijonhufvud, for example. It is a Swedish name created for a family ennobled more than 300 years ago, 10 generations in the past. According to the conventional notion of Swedish social mobility, Leijonhufvuds today should be no more distinguished than Swedes bearing surnames of low-class origin such as Andersson.
But Leijonhufvuds remain significantly wealthier and more educated. They live in fancier suburbs. Their social status has survived for three centuries. Measured through surname status, contemporary Sweden has rates of underlying social mobility that are extremely slow, and no higher than in pre-industrial Sweden or medieval England.
The result of my research is that this same strong conservation across generations of surname status is seen everywhere: in Thatcher’s England, in Tony Blair’s England, in medieval England, in Industrial Revolution England, in the huddled masses of the U.S., in Communist China, in capitalist Taiwan, in socially homogenous Japan, in the Chile of Pinochet and the Chile of Allende. Indeed there seems to be a universal constant rate of social mobility, with the few exceptions being even more immobile societies such as India. This rate is so slow that it implies that we can predict the majority of status variation among people at birth just from their lineage. The figure below shows this pattern.
In the U.S., for example, high status surnames of Ashkenazi Jewish origin such as Katz are converging so slowly on low status surnames mainly held by black Americans such as Washington that if current trends hold, it will be hundreds of years before they achieve equality.
The fact that Katz and Washington are ethnic surnames might suggest this immobility is a peculiar function of ethnic differences. But exactly the same slow mobility in the U.S. is observed for overwhelmingly white surnames of unusually high or low status. Thus the surnames of the New France settlers of North America – Hebert, Cote, Gagnon — are all now of low social status, 300 years after most of these settlers arrived in the New World. And they are converging on mean status no faster than black surnames.
The reason conventional estimates show much more social mobility in egalitarian societies such as Sweden is that partial aspects of status, such as income and wealth in that country, are not as strong indicators of underlying status as they are in more unequal societies. Since architects in Sweden earn little more than bus drivers, for example, you cannot infer much about underlying social status from income.
The correlations in the second figure imply that the descendants of the current high- and low-status families will all eventually have an expected average status. But this will take many, many generations. When we look at social mobility at the group level – where the group is defined by race, ethnicity or by a shared surname – we observe for such groups the slow underlying rate of social mobility. This is why the American Jewish population has seemed surprisingly impervious to expected downwards social mobility, and the black population surprisingly bereft of expected upwards mobility. These groups are not exceptions to a general rule of fast social mobility, created by a unique cultural inheritance or by racial prejudice. They are illustrations of the general rule that social mobility is everywhere slow.
Since social democratic Sweden, with publicly provided education from kindergarten to college, and with universal health care pre-natal to grave, has underlying social mobility rates not perceptibly greater than those of the more laissez-faire U.S. or the English feudal era, social status seems to be inherited within families functionally in the same way as genetic traits such as height. In practical terms, nature dominates nurture. The only way to attain better social outcomes is to have chosen different parents.
However, if we think that social status is as much outside the control of the individual as height, then we have to re-evaluate our acceptance of high degrees of inequality in modern America. Unlike social mobility, the degree of inequality is something that societies do have significant ability to change through the tax and expenditure systems. Suppose we had a social system that allocated enormous rewards to tall people and little reward to short people? We would think such a system arbitrary and absurd. But if social status is inherited as strongly within families as height, then shouldn’t we insure those who get a bad draw in the parental lottery against their misfortune?
Sweden, for example, transfers resources from those of high to those of low social status to minimize inequalities of income, wealth and health. Yet in Sweden, those at the upper ends of the status distribution are as faithfully reproducing their place in the social system as in the U.S. Redistribution has not taxed away enthusiasm for achievement among Swedish elites.
The U.S., in particular, will face a social landscape of persistent inequalities over many coming generations. There are well known differences in social status among such long-established populations as Native Americans, African-Americans, Jews and Franco-Americans. But restrictive immigration since World War II has led to a highly heterogeneous immigrant influx. Immigrants from countries distant from the U.S. tend to be strongly positively selected from the top of the status and ability distribution. Thus Indian, Iranian, black African and Philippine immigrants to the U.S. are heavily drawn from the elites of these societies. While the average Indian in 2000 had less than five years of education, those in the U.S. had more than 15 years, more even than the native U.S. population.
In contrast, immigrants from Mexico and other Central American countries are strongly selected from the poor of these societies. Much of this immigration is illegal, and this path is unattractive to those with education.
The accidents of U.S. history and of recent American immigration policy have thus created a society of unusual divisions of ability and status. The evidence from surname social mobility is that this will magnify the existing ethnic class divisions of the U.S. We need to consider how to mitigate the consequences of these forces. Measures to promote social mobility can help only on the margins. So the main work has to be to create a social system that can accommodate such persistent divergence of fortunes.
One thing we should be clear on, however, is that the measured low rate of social mobility is in itself not a social tragedy. It crucially depends on what is driving the high correlation of status between parents and children. If that is driven by environmental deprivation, exclusion, nepotism and connections, then it would indeed be a tragedy. But there is considerable evidence that the biological inheritance of talent and drive is what underlies most of the correlation between the social status of parents and children. Whatever the social system — Communist China or Republican Texas — families of greater social competence will navigate themselves to the better social positions. In that case, the strong hereditability of social status revealed by surnames is no more a social tragedy than the strong hereditability of height. What would be tragic would be to not recognize, in light of these social mobility facts, the imperative to limit the rewards from social status.