Tuesday, March 04, 2014

"The Son Also Rises" - Gregory Clark

In The Times today, Hugo Rifkind writes:
"Stay with me. For I recognise that look upon your face. The other day, you see, I figured out how to replace every halogen light bulb in our bathroom with an LED equivalent. Man, I was pleased with myself. “The wattage will be cut by 98 per cent!” I shouted at my wife, maybe for the eighth time. And she looked back at me wearily and explained that although she agreed that this was terribly important, she’d prefer not to have to talk about it ever again."
By coincidence, we have just done exactly the same thing in our house. Two data points surely create a trend line.

Here is how The Economist reviewed Gregory Clark's "The Son Also Rises: Surnames and the History of Social Mobility."
"Periods of great inequality are good for social theorising. Last year Charles Murray, a libertarian columnist and pundit, warned in “Coming Apart: The State of White America, 1960-2010” that mating among people of similar means is increasing the divide between a motivated elite and a listless underclass. In “Twilight of the Elites: America after Meritocracy”, another commentator, Christopher Hayes, showed how the elite are using their resources to circumvent the meritocracy and establish a permanent upper class. Now the fray is joined by Gregory Clark, a Scottish-born economic historian at the University of California, Davis. In “The Son Also Rises” he argues that social mobility is low everywhere and always will be, and there is nothing society can do about it.

"Mr Clark has waded into these waters before. His 2009 book, “A Farewell to Alms”, reckoned that Britain’s Industrial Revolution was attributable to faster breeding among the rich, which raised the population’s underlying competence. Critics pounced, crying that the book was thinly sourced. Perhaps for this reason, he has armed his new book with reams of data. Most of the text is given over to methodical presentation of research, with the uniting theme that optimistic assessments of mobility are badly wrong.

"Conventional research has it that society is highly fluid: the effect of inheritance is almost nil in some Nordic societies, and family explains no more than 25% of an adult child’s status in America, which has always been less mobile. But Mr Clark points out that these studies track change over just two or three generations and are therefore biased by quirks of fate: the working-class lottery-winner or the scion who chooses social work over high finance. Longer projects average out this randomness and paint a darker picture.

Mr Clark draws upon research that uses surnames to track status over centuries. The academics he follows have mined sources as varied as the Domesday Book, the Royal Society’s records, even membership of the American Medical Association, in order to find surnames that are over-represented in elite positions. Researchers then track how long it takes those monied surnames to lose their wealth-predicting power.

"With surprising consistency across countries and eras, mobility is found to be painfully slow. Birth has predicted more than 50% of one’s income or education status, Mr Clark reckons. Erasing the legacy of past prosperity takes 10-15 generations rather than the three or four implied by sunnier estimates. So the shadow of 18th-century wealth still darkens income distributions today.

"That is the most unexpected finding. Efforts to democratise education and eliminate discrimination over the past century appear to have had no discernible effect on mobility, leading Mr Clark to conclude that mobility is strongly linked to underlying social competence—an “inescapable inherited” trait. Only the intermarriage of people who are more prosperous and educated with those less fortunate will dilute the genetic resources of well-off families, slowly pushing them back towards the average and preventing the rise of a permanent overclass.

"Oddly, Mr Clark judges the world to be “a much fairer place than we intuit.” He explains this by stating that the rich acquire their wealth because they are clever and work hard, and not because the system is rigged. The world is less corrupt and nepotistic than people might think.

"This conclusion gives the book a cheery tone, but there are also plenty of nasty conclusions to be drawn. One inescapable judgment is, as Mr Clark says, that “a completely meritocratic society would most likely also be one with limited social mobility.” He does not say that American blacks are poor because they are black. His work implies, however, that poor blacks remain so because they are descended from people with low social competence; discrimination is irrelevant, except to the extent that it limits intermarriage with other groups. “The Son Also Rises” may not be a racist book, but it certainly traffics in genetic determinism.

"That is a weakness. Mr Clark is too quick to write off the promise of recent social changes. The oldest Americans born after the passage of the Civil Rights Act are barely 50. Impressive work on the effect of good teaching or well-targeted poverty assistance suggest such programmes make a difference. Yet Mr Clark follows his logic to an unexpectedly egalitarian end. Redistribution is sensible, he argues, not in order to boost mobility but because mobility is intractably low. The cream will rise regardless, and so paying extraordinary salaries to capable workers is unnecessary. If high rates of mobility are used to excuse or justify inequality, he suggests, then the reality of low mobility implies something quite different: that great inequality serves little purpose and redistributing income from the rich to the poor might raise overall welfare at little economic cost. This makes for uncomfortable reading for those of all ideological persuasions."
Clark is largely content to examine the data he and his team researched in the context of his mathematical model of social mobility (a first-order Markov social-mobility diffusion process) without imposing any personal agenda. His public-policy thoughts are restricted to a chapter at the end where, as The Economist notes, he expresses an ethical preference for Nordic-type societies.

It's useful at this point to note the correlation at the state level between individual willingness to countenance large transfer payments and the degree of ethnic cohesion within that state - Clark doesn't and as a consequence fails to note that this pleasant outcome is unlikely to come to pass in the United States.

A small aside: reviewers who wish to exhibit their impeccably PC credentials include words such as "nasty", "genetic determinism" and "racist" in their reviews. These terms, devoid of scientific meaning, invariably demonstrate high-minded distaste replacing possibly-dangerous rational thought.

Clark's book is an easy read if you know some statistics* and are interested in the history of diverse populations from Chile to China via the US, UK, Scandinavia, the Middle-East, India and Japan. His probabilistic social-mobility model is simple yet makes powerful counter-intuitive predictions. These are backed up by the voluminous data-sets he's managed to analyse. You also get advice on who to marry.

The world still divides between those social scientists who think that genetics (i.e. inherited differences between people and populations) are irrelevant as everyone is born identical for social-outcome purposes, and those who are tracking the latest research in population genetics and DNA analysis as they shed new light on sociological, economic and political issues. The latter camp is slowly gaining respectability as more hard data becomes available and Clark's book is a breath of fresh air in this developing paradigm.
* Correlation coefficient, regression to the mean, normal distribution, standard deviation, variance, time series.