It’s the social mobility of their respective elite classes. What does this mean? Well, two professors, Gregory Clark and Neil Cummins, have recently published a very interesting paper called “Surnames and Social Mobility in England, 1170-2012.” Bear with me while I explain their research, before I turn to the surprising correlation and explanation of why China has had a similar outcome as England.
Clark and Cummins take a long-term view of staying power of the elites, estimating the correlation between the status of families through multiple generations—that is, how much of the current elite status of a person can be explained by the status of his parent, and the parent of his parent, and so on. As a proxy for the families, they take distinguished surnames that are not common and are overrepresented among the elites. One name is overrepresented when it appears more frequently in the elites than what you could expect given the number of people who use it. They define elite in various ways: in England, they use the attendance to Oxford and Cambridge since 1170.
Using educational status in England throughout this period they show that “Oxbridge attendance suggests a generalized intergenerational correlation in status in the range of 0.70-0.90. Social status is more strongly inherited even than height. This correlation is unchanged over centuries. Social mobility in England in 2012 was little greater than in preindustrial times. Thus there are indications of an underlying social physics surprisingly immune to government intervention.”
It is highly tempting to blame this on capitalism, and some people have insinuated this already.
Blaming this persistence of the same families in the elite on capitalism would be jumping to conclusions too rashly. For that, we should have a comparator country in a period in which it has not been capitalist. It happens to exist, and the two authors of the recently published paper (along with several other co-authors), have documented it. In fact, the paper mentioned above summarizes two chapters of a book they published earlier this year, The Son Also Rises: Surnames and the History of Social Mobility, Princeton University Press, 2014, available in Kindle. They studied the elite’s surnames’ behavior for eight countries and several centuries and found that in all of them the persistence of the elites was similar to modern England (which was similar to that of medieval England.)
In addition to England, the studied countries were Sweden (supposedly a social democratic country with high social mobility), the United States (land of opportunity), India (admittedly rigid but ruled by socialists for several decades), Taiwan, Japan, Korea and Chile. To crown the list, they studied China under communism and compared it with previous stages of the almost interminable history of the country.
And, surprise, surprise, the social mobility of communist China, including the Mao years, and the previous nationalist years, is very similar to that of England (and to all the other countries in the list.) Maybe you are thinking that, yes, of course, there was an elite in communist China, there is one everywhere—but surely they would be different names and people from those that formed the pre-communist elite. After all, as Clark and Cummins note, a million mainland Chinese fled to Taiwan when the communists defeated the opposing Nationalists—most of them members of the elite. Under the communist agrarian reform in the late 1940s and early 1950s the land owned by the landlord class was seized and redistributed—amounting to 43% of all the land in China; in the process, 800,000 landlords were executed. Then, in the 1960s and 1970s, about 10 million of the relatives of former landlords, businessmen, and apparent bourgeois were killed during the Cultural Revolution. All in all, the communists killed about 60 million people on the excuse that they were bourgeois. This included teachers, intellectuals, professionals and anybody that sniffed at being a member of the previous elites. Large numbers of students in the urban regions were sent to the countryside and denied education—to facilitate the equalization of society.
The authors ask whether Mao created a period of unusually rapid social mobility through the elimination, repression, and dispossession of the upper and middle classes?
Surprisingly, he did not. The authors identified 13 surnames that appear with unusual frequency in the Qin examination system—the Chinese test to identify who will become a member of the highest elite in the country, the state bureaucracy. They selected them from more than 50,000 successful candidates (the most successful) in the Yuan, Ming, and Qin dynasties, starting in 221 BCE (we are talking China). These surnames are overrepresented in the modern imperial era and in modern Chinese elites—the high officials in the Nationalist government from 1912 to the triumph of the communists in 1949; professors at the ten most prestigious universities in the country in 2012; chairs of the boards of companies listed in 2006 as having assets of $1.5 million and above; and members of the (still communist) central government administration in 2010.
The intergenerational correlation of status between the Nationalist period (just before the Communists escalated power) and 2006 (that is, covering almost thirty years of Maoism and then the current variety of communism) was 0.9 for professors in 2006, 0.8 for company board chairs and 0.74 for central government officials. This means that if you predict that the surname of a member of the elite in 1912-1949 would still be a member of the central government elite in 2006, you would be right in 74% of the cases in each generation. Other names come and go, but these ones stay. That is staying power. And it happened under a communist regime that killed scores of millions of people suspected of being members of the elite.
Clark and Cummins made other estimations using different data sets. They came to the conclusion that we can be 95% confident that the true intergenerational correlation of status for Communist China lies in the range of 0.71-0.92. If you were a member of the elite in 1949, your immediate descendants had a probability of being members of the elite themselves of 71 to 92%. This is very close to capitalist England—and to all the other countries in the sample.
These results contradict many studies that estimate much lower intergenerational correlations. These, however, are based on the comparison between two generations only. Clark and Cummins’ results are more reliable because they cover many generations, and everybody knows that the higher the number of observations, the better the estimate because it tends to eliminate chance events.
The authors say that their results do not imply that some individual families do not decay or that others fail to escalate the heights of society. They estimate that any elite family in any of these countries will go up and then down, following a process called regression to the mean. The process, however, could take from 600 to 800 years, even in communist countries. Take that, Marx!