Thursday, April 10, 2014

Book Review: "The Son Also Rises"--the anti-Piketty.

On occasion, I'll put up a book review. There are books I'm reading that pique my interest and then drive me either to huzzahs or harrumphes. I go so far as to order these books and plow through them as some nerd's form of "pleasure." But sometimes they raise such cackles inside of me that I need to scream them out into the blogosphere. I came across one of them lately.

A few Sundays ago an article appeared in that (slightly trashy) Sunday Review section of the New York Times by Gregory Clark, professor of economics at UC-Davis. In it, Clark described some amazing data. I was so intrigued by the data, and love economic history so much, that I just wanted to know more about his data and explanation. A few days later, The Son Also Rises, arrived from Amazon with that new book smell that you get when you take a whiff of its pages. As I plowed through all the correlations and figures, I began to detest its smell.

His central argument is that a more equal society, as a product of better social mobility, in our or future lifetimes from now is a pipe dream. Going back into history, to the Norman invasion in England, for example, and across many countries (US, Japan, China, Chile, England, India, Korea and Sweden), we see that the same families and kin groups, indicated by "surname," have occupied the top rungs of the social status ladder for centuries and have not moved, contrary to our statistical expectations. We would expect that all families would regress (or progress) to the mean in terms of social status and income. They don't. The richest families with the highest status remain at that position for centuries, with some rare exceptions. Despite all the efforts by governments to engineer more social mobility (not sure governments are always doing that when reducing social inequality, but anyway), their projects have failed, even in extreme cases like Communist China.

I was really impressed by the data Clark gathered. Based on an analysis of surnames from many ancient and far-flung sources that must have been difficult to collect and scour, Clark found that prominent families disproportionately attended Cambridge and Oxford (again, contrary to statistical expectations, not common sense). The descendants of the Swedish nobility disproportionately occupy slots at national universities. The descendants of samurai and Kazoku are disproportionately represented in many elite professions in Japan. In the US, the same family names from some of the usual suspected ethnic groups are overrepresented among the "elite profession" of doctors--Jews, South Asian Indians from the Brahmin class, Japanese with special surnames, and blacks from Africa. Not surprising, Native Americans and African American blacks are underrepresented in that profession.

The fancy new data held my attention for a while, especially how global in scope it was. But I was dying in suspense to get to the explanation. Boy, was I disappointed when I found out it was BIOLOGY...or genetics. Uh oh, Professor Clark! Sure you want to go down that road???

He does. After trying to impress you with all of his data, he puts forward a "law of social mobility." Now, just like when you see a utopian philosopher walking down the street wanting to chat, you better run from a social scientist who calls any hypothesis of his a "law." This law is so oversimplified it boggles the mind how this got through peer review at Princeton UP. Basically, a family's income in one generation is determined by the "family's underlying social competence" and some random component. What is "social competence?" Well, it's a latent variable that "cannot be directly observed." This social competence is reflected in a family's earnings, wealth, occupation, education, residence, health, and longevity. But there's something more "underlying" all of this. None of these factors have any independent effect of their own. That's just their phenotype; not genotype. [Body starts quivering.]

That's when Clark's argument turns scary. Because all this underlying social status doesn't change across time, doesn't regress to the mean as expected, a family's status must be inherited! By inheritance, Clark might be saying it was all accident of birth. But the genetic metaphor slips into earnestness a little too easily. He does argue that there's some genotype that is reproducing the qualities necessary to stay at the top of the social ladder. Clark finds that almost all of a child's status is determined by the status of the parent. The correlation between a child's status and its parents is close to 1, especially among people at the top 1%. Clark is careful to constantly say "inherited" rather than genetic. In fact he writes, "This does not imply that the social genotype is actually derived from genetics, just that it behaves in a way that mimics genetic transmission of characteristics (p. 282, fn. 9)." Is it genetics then? Again, Clark is very cautious in his language, but you get an overwhelming sense that he thinks it is. After citing Galton's study on the relationship between parents' heights and their children's, how a Danish child's income is strongly correlated with its parents' wealth, and the issue of assortive mating (whereby people of a high status seek to marry others like them), there is a strong impression that Clark thinks high social status is more than just inherited, it's also genetic. And let's just forget about the approving nods given to the "work" of "sociologist" Charles Murray, which only adds evidence to the indictment of Clark as a social darwinist.

To his credit, maybe by inheritance Clark just means accident of birth, and there's text to support that interpretation. Therefore, at best, he's guilt of taking a metaphor too far, which can mislead the reader. At worse, he really does believe that a family's social competence, whatever that is, is in our genes. But Clark quickly, harshly, and over broadly dismisses any of the typical structural variables most political scientists and sociologists would use to explain the persistence of a society's social structure.

He sweeps away, often with one big arrogant swoosh, alternative explanations for no good reason and without citations. When trying to explain why children in Jewish families persist in having high social status while blacks do not, all we need to do is invoke his model of inherited social competence, "there is no need to invoke racial discrimination, social networks, or ethnic capital to explain such effects." Earlier, he writes, when referring to the possibility that ethnic capital, the ability of ethnic social networks to help strengthen and maintain a particular group's social status, Clark writes, "There is little to no evidence for these hypothesized community benefits to individuals. The descendants of Japanese immigrants to the United States, as noted, are a group of high average status. Yet, this is a small, dispersed community that has integrated strongly into white American society through intermarriage. With the simple model of a slowly changing underlying social competence of families, we can explain all the above results without having to invoke racial barriers, ethnic capital, or the importance of social connections (p. 124)." In effect, Clark is saying "My simplistic theoretical model explains it all very simply; why bother with messy things like racism, generations of discrimination and social exclusion, persistent grinding poverty, geographic limitations and economic dislocation over decades."

This is an absurd thesis. First, just because your theoretical model predicts social status better than others does not make it an accurate or truthful model. All models are only representations of reality, not reality itself. So, it could be just a  magician's/social scientist's trick that his model fits the data. Anyone can construct a model that explains the data without providing a valid theoretical explanation. It's also called confusing correlation for causation.

Second, it would be nice to see just what this "latent variable" of "social competence" is and how it works "in the flesh" sort of speak. Clark's dismissal of social institutions is completely without merit. Social status, BY DEFINITION, is socially constructed. High incomes, wealth, education, professional success can only happen because society is set up for those things to be rewarded. In this case its capitalism. If high status is genetic, I would love to see how these families perform in the wild (across generations) or on Survivor. Just because people who inherited wealth and high social status stayed that way doesn't mean that social and political institutions had NOTHING to do with it. Many political scientists and historians know for example that in order to produce democratic institutions and extend mass suffrage in most Western European countries at the turn of the 20th century, great bargains had to be struck between the working class and bourgeoisie. These bargains kept most of the status elite groups had in place--for a century and a half. It wouldn't be surprising to me that elites have always found ways throughout history to maintain their high social status, EVEN in Communist China. A wealth of social and political research should not just be dismissed out of hand (especially when the author presents no review of the existing literature explaining social mobility or the absence thereof).

Finally, Clark's policy prescriptions are pretty patronizing. He does conclude that if the social status of some groups will remain high or low (forget about how they got there) for a long time, then the best we can do is make life a little better for them, so long as it doesn't impede the "free market" too much. This can come in the form of poverty relief and public education. But, he would argue, let's not fool ourselves. The poor will stay poor (whichever groups they are in a particular country), and the rich will remain rich because, well, their parents were! Whether you invest in Baby Mozart or Head Start programs, it doesn't really matter. You can't marginally improve your chances of rising or prevent yourself from falling down the social ladder. You and your kin are going to stay at the place on the economic ladder for generations.

What I find problematic about this is that we shouldn't make the giant leap from the supposed lack of social mobility to policy prescriptions and conclude that if the latter doesn't help the former we're all screwed to live in a terribly unequal society. Instead, let's focus on social and economic inequality, its causes, and how it changed across time, groups, and generations as well as countries. These public programs that were enacted aren't there to address social mobility per se. Actually, very few social programs advertise that as their goal. Medicare and Medicaid in the US, Social Security, pensions in Europe, wage bargaining institutions, etc. exist in the advanced capitalist democracies to address the problems of social and economic inequalities that, perhaps!, these "underlying social competencies" produce. That's where the real story lies.

And that's why I'm so eager to get to Piketty's book, "Capital in the Twenty-First Century." It's also rich with data, like Clark's. It also has a simple theory, about how income inequality grows or contracts across groups and time. It's also about the power of inherited wealth. But the model is steeped in empirical evidence. It's not some abstract picture mounted on top of data and with the messy factors of history, politics and economics swept away from underneath. And correlations are not confused for causation. After this sour (and sorrowful) piece of scholarship, I can't wait to get to the sweetness of Piketty.






No comments:

Post a Comment

There was an error in this gadget