What's in a Name?
The Son Also Rises: Surnames and the History of Social Mobility
By Gregory Clark (Princeton University Press 364pp £19.95)
Why is Mr Neville richer than Mr Smith? Will your children go to Oxford? Who should you marry if you want to win at the game of life? Gregory Clark, a Scottish economist at the University of California, Davis, offers some answers in his fascinating new book, The Son Also Rises. In short, he argues that the contribution of distant relatives matters as much as our parents in implementing our fate.
Whereas conventional work on social mobility tracks outcomes over a single generation in one dimension, such as income or education, Clark steps back to take in outcomes in all aspects of a human life over many generations. What counts, he argues, is 'social competence', a genetically driven underlying set of skills related to both income and education. If you inherit competence, you can draw on it to become rich, even if your parents are not. In any generation, a person with high competence can select a fulfilling but low-paid existence as a monk or artist. A low-competence individual may prosper through luck or grit. Yet over generations, underlying patterns assert themselves. How can Clark be sure? By tracking the success of elite and low-status surnames across generations and measuring the rate at which they converge.
In an egalitarian society such as Sweden's, parental education or income is only 40 per cent correlated with a child's educational performance or future income. This suggests they are largely architects of their own fate. Against this conventional wisdom, Clark discovers that, in the long term, a Swede's ancestry accounts for 70 to 80 per cent of where she will wind up in life. Swedes with noble or latinised surnames, such as the botanist Carl Linnaeus, originally constituted an elite. Names such as these remain over-represented in top universities and among physicians. At the other end, those with surnames ending in '-sson', originally lower status, are under-represented. The same is true in England, where the surnames of Norman conquerors and noble families are over-represented in elite professions. They are followed by locative surnames, such as Hilton, which describe places. Names that describe occupations, such as Baker and Smith, are under-represented in elite circles, with common Traveller surnames such as Loveridge at rock bottom.
The spread between a Norman name such as Baskerville and the plebeian Smith on the Oxbridge entrance rolls or in the size of legacies bequeathed in wills is nowhere near as large as it was in the 1300s because, in ten or fifteen generations, the gap closes significantly. In this sense, no elite can defy gravity. The flipside is the glacial pace at which equality is established: parent-child correlations for surname status average 70 to 80 per cent in equal societies, rather than the more rapid 40 or 50 per cent currently assumed by social scientists. From communist China to authoritarian Chile, multicultural America to homogeneous Japan, Clark's 'iron law' of inheritance holds up. Surname achievement correlates at 70 to 80 per cent between generations.
The implications of this are profound. A level playing field will produce as much inequality as a tilted one, so the Left's policy interventions are futile. Talent will rise to the top whether market incentives exist or not, so the Right's insistence on free markets is misguided. It's a bold thesis based on copious research and one that is refreshing in its willingness to tread on sensitive ground, such as ethnicity and IQ. But is it waterproof?
Yes and no. The iron law is remarkably consistent: inequality is more enduring than many previously thought, though its grip is weaker when examined over the course of a lifetime than over generations. Yet Clark's insinuation that the iron law is underpinned by genetics is flimsy. Our relatives and ancestors beyond our parents are not just a source of genes, but of social networks, culture and resources. Even if our parents flop, we draw on our relatives' multiple advantages. Clark argues that his case is strengthened by the persistence of his iron law in times of democracy, feudalism, communism and mass democracy. But if the persistence of inequality is based on genetic competence, one would expect the iron law to emerge only in a meritocracy. Social orders based on might or nepotism should yield different intergenerational correlations to those based on unalloyed IQ. The fact that the same relationship is found in extractive and knowledge-based societies suggests that factors other than intelligence - networks, power, resources, culture - are more important. Bravery, strength and confidence characterised the founders of most royal lines. Aristocrats initially spurned reading. The fact the iron law has held, despite the shift from a brawn- to a brain-based order, actually refutes an IQ-based argument for inequality.
The argument's simplicity - in focusing on the correlation from one generation to the next - is a strength but also a weakness. Proper statistical analysis needs to control for confounding factors. Take the United States. Clark finds that people with French-Canadian names do worse than those with English ones. But the French (having, for the most part, come by way of Canada) disproportionately settled backwoods areas such as the bayous of Louisiana and the forests of northern Maine. Other groups, such as the Irish, English or Germans, settled in cities as well as the countryside. The fair method would be to control for urban residence by comparing, say, French-Canadian and Irish migrants who arrived in Boston in the 1880s.
There are other unwarranted leaps in the story. Clark informs us that in Ireland in 1911, Catholics with Scottish and Irish surnames had high levels of illiteracy compared to Protestants with Scottish or Irish names. Clark asserts that the top end of the native Irish married into the Protestant population while the bottom end of the Scots planter population joined the Catholics. There is little historical evidence for this. Instead, we know that ethnic and religious boundaries were fluid at all social levels in Ireland. For example, holders of high office in the Protestant Orange Order were known to occasionally marry Catholics in the early 19th century. Later, they would be expelled for doing so. In Newfoundland, the pattern of Protestant and Catholic was initially set by the spread of English or Irish settlement, then altered by the availability of priests from different denominations in a remote region. The Irish case would support Clark's argument if Catholics with Scottish surnames, such as John Hume, achieved at a higher level than Protestants with Irish names, such as Terence O'Neill. Instead, ethnicity and religion trump surname completely, wiping out the effects of social competence.
The same mistake is made in the case of India. Muslim underperformance is not a legacy of low social competence or the absorption of bad genes through the conversion of Hindu rejects. Muslims were the elite in Mughal India, had high status under the British, but lost out after independence. Focusing merely on their representation among doctors is too narrow a measure of status. Likewise, Brahmin achievement cannot be ascribed to superior genetics but rather to a social structure, network and culture that buoys up even those with unsuccessful parents. Too often the book skates around anomalies such as the worsening performance of Muslims after Indian independence or the post-Independence rise of scheduled castes, attempting to bracket these as political anomalies. What this misses is that politics is always in play alongside social influence, culture, resources and, yes, genes. Nothing in the book convinces me that genes underlie the pattern of inequality Clark commendably lifts from the detritus of history. Most surname-based inequalities remaining in Western countries result from ethnicity. Clark is too pessimistic about social policy: it can help to overcome ethnic inequality, though he is right in thinking it is a slow process. The book is also over-optimistic in its suggestion that African-Americans will automatically rise to the mean. Ethnic inequality has little to do with genes, but everything to do with culture and social networks. Regardless of intermarriage, inequalities often persist. For some groups, the son never rises.
Exclusive from the Literary Review print edition. Subscribe now!
Eric Kaufmann is Professor of Politics at Birkbeck College, University of London. His latest book is Political Demography (OUP, 2012).