The Son Also Rises Read online

  Another source of evidence of the persistence of marital endogamy is websites advertising for potential wedding partners in Bengal. We recorded the characteristics of two hundred women identified as Kulin Brahmin on one of these sites. 83 percent specified that they were seeking a Brahmin husband, 2 percent specified Brahmin or other high-caste, and only 15 percent stated that caste status was no barrier to a potential union.19 However, that 15 percent includes 8 percent who listed their preference in a form such as “Brahmin-Kulin, caste no bar.” Thus a full 93 percent of respondents indicated a preference for a Brahmin spouse. And listings on websites presumably are biased toward the least traditional families of the Brahmin community.

  If the marital endogamy of castes and religions in India explains low average social mobility for surname groups, we should find higher rates of social mobility for individual families within these groups. Families sharing the surname Banerjee, for example, will have the same rates of mobility as in any other society. It is just that the average status of the Banerjees will not converge toward that of the Shaws. We should also find that over time, all the major Kulin Brahmin surnames have the same average social status. This hypothesis is borne out by the incidence of these surnames among physicians.

  This consideration implies that in India, the estimates of social mobility for surname groups correctly estimate persistence at the group level but overestimate persistence at the family level. While the Brahmins may have stayed at the top of the social scale for thousands of years, individual family lineages within the Brahmin caste should show the normal slow turnover of elite and underclass families every three to five hundred years.

  The Beneficiaries of the Reservation System

  We show above that the reservation system has been of enormous benefit to some families in admission to university and government positions. Some surnames that are associated with scheduled castes, such as Mandal/Mondal, now appear, according to many measures, to belong to the social elite of Bengal. If the reservation system were to end now, what would be the long-run status of these surnames? Would they decline toward their 1947 social status? If not, could we conclude that affirmative-action policies can permanently change the outcomes for social groups? And if affirmative-action policies bring about such effects, does this prove that biology is not the major factor determining the social status of the next generation?

  The information does not exist to answer these questions. There is, however, one piece of information that suggests that the reservation system in India may have less effect on the overall social status of its beneficiaries than the data on physicians and university admissions in Bengal would suggest. The horizontal axis of figure 8.8 shows the percentage of physicians in Bengal registered between 1950 and 2011 from each of the six surname groups discussed above, plus those with Christian surnames. Large numbers of physicians who complete medical school in Bengal end up working in the United States. The American Medical Association directory for 2012 shows 1,168. The vertical axis of the figure shows what percentage of these Bengali physicians in the United States come from each surname group.

  For some surname groups, such as the Kulin Brahmins, the percentages of physicians employed in Bengal and the United States are the same. Other groups, such as Christians, have a much greater presence in the United States. Notably, the group for which it was estimated that as many as 70 percent were admitted to reserved places in medical schools is heavily underrepresented among Bengali physicians working in the United States. Their representation in the United States is only a quarter of that in Bengal. The mixed Hindu surname group, of whom an estimated 45 percent were admitted to reserved places, show up in the United States at about 40 percent less than the rate expected from their representation in Bengal.

  FIGURE 8.8. Surname group shares among Bengali physicians in West Bengal versus in the United States.

  There could be many explanations of these findings. But one possibility is that physicians admitted to Indian medical schools through the reservation system find it harder to meet the onerous requirements to practice medicine in the United States: passing the U.S. Medical Licensing Examination and completing the required residencies. Though the reservation changes the measured social status of these individuals, it does much less to change the underlying social competence of these families. If the reservation system were ended, the measured social status of these groups would soon also decline substantially. Another possibility, however, is that groups that benefit from the reservation system have fewer of the family resources needed to relocate to the United States and practice medicine there. But the observed patterns contain nothing that clearly contradicts the biological explanation.

  Conclusions

  Long-run social mobility rates in India, as measured by the frequency of surname types in high-status occupations such as physicians or judges, are even lower than in England, the United States, and Sweden. The reservation system may actually be restricting mobility because it helps only a small portion of the lower classes, and even there it tends to benefit families who were not of particularly low social status in the first place. But without the reservations system, the estimated status persistence rate would be, on average, a very high 0.91.

  This unusually low rate of social mobility is consistent with the argument in chapter 7 that group marital endogamy leads to persistent classes of the advantaged and disadvantaged. In Kolkata, for example, there is little or no intermarriage between the very low-status Muslim community and the prosperous communities of high-caste Indians and Christians. Matrimonial listings on the Internet still show strong preferences, even within the Hindu community, for partners of the same caste.

  1 There is contention about this claim, but a paper in Nature concludes that “allele frequency differences between groups in India are larger than in Europe, reflecting strong founder effects whose signatures have been maintained for thousands of years owing to endogamy” (Reich et al. 2009). Chapter 13, by contrast, shows that subcastes in England and Ireland, such as the Traveller population, are not genetically distinct from the general population.

  2 The list of scheduled castes was initially promulgated by the colonial government in 1936. The British classification was largely adopted by the government of India in 1953 in establishing the current system of reservations (Jadhav 2008).

  3 The Indian Supreme Court has ruled that no more than half of available positions can be reserved for disadvantaged groups, but some states have higher quotas. Tamil Nadu, for example, has reservations for 69 percent of positions. The legality of these quotas is still being litigated.

  4 The large agricultural population makes it difficult to classify occupational status. Mobility studies based on occupation are thus difficult to interpret and to compare with those from more developed economies. See, for example, Nijhawan 1969; Kumar, Heath, and Heath 2004; Hnatkovska, Lahiri, and Paul 2013.

  5 Corak 2013; Hertz et al. 2007.

  6 Hnatkovska, Lahiri, and Paul 2013. The researchers actually estimate the income elasticity, which equals the correlation only if the variance of log incomes is constant.

  7 Government of Bengal, Political Department 1930.

  8 The association of these surnames with the Kulin Brahmin subcaste can be confirmed by examining the surnames of those listing themselves as Kulin Brahmin on matrimonial websites. All these surnames are also found, however, among other subcastes of Brahmins.

  9 In 2013 a law was passed reserving 17 percent of places in state-run universities for other backward classes.

  10 The press now publishes articles on the plight of lower-class Brahmins, who are disadvantaged in competition for university places and government positions.

  11 Candidates from scheduled castes and tribes who score high enough on the entrance requirements are allotted unreserved places. So the reservation system reduces the number of places available to members of the higher castes by the share of reserved places.

  12 Bankura Medical College, entry yea
r 2012, and Kar Medical College, entry years 2010 and 2011, had admissions lists showing candidates’ reservation-system status. These give the status of 395 admitted students in total.

  13 This assumes that absent the reservation system, admissions rates for this surname group in the unreserved category would remain as they are currently.

  14 Susan Bayly notes that the British caste designations of 1931 were often so broad as to include many groups not suffering any social disabilities (Bayly 1999, 277).

  15 The relative representation of the surnames in the period 1920–47 was calculated assuming (counterfactually) that the Muslim population share was the same as in 1980–2011. This was done to exclude from this exercise the effects of a shifting Muslim population share on measured social mobility rates.

  16 Corwin 1977.

  17 Dalmia and Lawrence 2001.

  18 Leonard and Weller 1980, tables 1–3.

  19 Bengali Matrimony, n.d.

  NINE

  China and Taiwan

  Mobility after Mao

  IN ALL THE SOCIETIES WE HAVE LOOKED AT so far, mobility rates have been slow—or, in the case of India, almost nonexistent—at the group level. But these are all societies whose institutions and social structures have been stable and continuous over many centuries.

  England famously had only one lasting Political Revolution, the Glorious Revolution of 1688–89.1 And even that revolution was made by the upper classes of English society for their own benefit and did not involve fighting on English soil. A new king and queen were invited in, and the tiresome old one, bereft of support, conveniently fled the scene. There were precious few imprisonments, dispossessions, or exiles as a result. The United States was born in a long and bloody revolution, but it was a revolution of the colonial upper classes against an external tyrant.

  In Sweden, there is debate about whether there ever was any political revolution. Thus, according to one author, “Conventional wisdom holds that Sweden’s transition to democracy was exceptionally gradual, stepwise and non-revolutionary.”2 In India, the British were expelled, and a new country was born in the violence of Partition and its mass population movements in 1947. Even so, in most parts of India the transition was smooth. The new rulers of India in the Congress Party were the same Brahmin elite who had governed as part of the Indian Civil Service under the British. In fifty of the sixty-six years of Indian independence, the prime minister has been a Brahmin.

  China, by contrast, experienced in the twentieth century a revolution unparalleled in its ferocity, bloodlust, class hatred, and mass dispossession. The sclerotic Qing imperial regime collapsed by 1912. Although the Nationalist Party won the election of 1912, warlords remained in control of the central government in Beijing. The Nationalists by 1925 established a rival capital in the south. It took more than a decade of fighting for Chiang Kai-shek, leader of the Nationalists, to defeat his northern rivals and reunify China under Nationalist rule in 1928. But this unity was fragile, beset by continuing armed conflict with remnant warlords, the Communist Party, and in 1937 open conflict with the Japanese, who had previously seized Manchuria. The Communists came to power in 1949 after decades of social turmoil.

  The final Communist victory produced a generation of unprecedented social dislocation. Perhaps a million mainland Chinese fled to Taiwan with the Nationalists, including many Nationalist functionaries and sympathizers from the middle and upper classes. Many Shanghai businesses transferred their operations to Hong Kong, which also received hundreds of thousands of refugees.

  Under the agrarian reform of 1946–53, as much as 43 percent of the farmland of China was seized from the landlord class and redistributed.3 In the process, some 800,000 landlords were executed.4 This was because land reform was mainly a political movement to eliminate opponents and potential opponents of the revolution. These executions represented only perhaps one in a thousand of the rural population, but this figure represented a significant fraction of the old elite.5

  The Cultural Revolution of 1966–76 saw another round of mistreatment and purges of “class enemies”—relatives of former landlords, businessmen, and anyone suspected of having a bourgeois background. Teachers and intellectuals, particularly those in positions of authority, were frequent targets. Thousands died at the hands of the Red Guards, the militant youth wing of the Communist Party. As many as half a million in the cities were shipped to the countryside for reeducation through labor. The institutions of society—universities, hospitals, government ministries—were paralyzed, placed under the control of erratic teenage revolutionary committees. But by 1967 the Cultural Revolution had dissolved into a bloody mess, into which the army entered both on the side of the Red Guards and in defiance of them. All across the countryside, alleged class enemies were being confronted, abused, and killed (figure 9.1). The total death toll from the chaos of the Cultural Revolution is unknown but has been estimated to be as high as ten million.

  FIGURE 9.1. Administrator of the Heilongjiang Daily denounced by the newspaper staff for following the capitalist line, 1966.

  Institutions of higher education effectively closed between 1966 and the early 1970s. Even after this, large numbers of urban students were sent out to labor in the countryside, denied the opportunity for higher education. It is estimated that as many as eighteen million people were forced to relocate to the countryside in these years. Admission to universities in the years 1970–76 was largely determined by political qualification and connections: only students without the taint of bourgeois background were admitted. In 1977 the National College Entrance Examination was restored, allowing “bourgeois” students to compete for university places. But not until 1980 were all those exiled to the countryside during the Cultural Revolution allowed to return to the cities. Thus the universities were effectively closed to those with bourgeois family backgrounds for more than ten years.

  Mao Zedong may not have had the lasting impact on Chinese society that he sought through such movements as the Cultural Revolution. But did he create a period of unusually rapid social mobility though the elimination, repression, and dispossession of the upper and middle classes?6

  Some recent studies suggest that postreform China has achieved high rates of social mobility. The urban intergenerational correlation of income in this period, for example, is only 0.3–0.6.7 But these estimates are based on one dimension of status and on only two generations, and we see above from examination of such examples as Sweden that such estimates can be completely misleading about underlying social mobility rates. This chapter again applies surname-group analysis to attempt to arrive at a more realistic estimate of social mobility.

  Chinese Surnames

  The problem with measuring social mobility in China using surname distributions is that the Chinese have few surnames, and these surnames have been employed for millennia. There are estimated to be only about four thousand surnames in use among Han Chinese. The hundred most common Chinese surnames are held by nearly 85 percent of the population, with the three most common Chinese surnames, Wang (王), Li (李), and Zhang (張), held by more than 270 million people (21 percent of the population).8 These “big three” surnames are used below as a benchmark for the average representation of surnames among elite groups. In England and Wales, by contrast, in 2002 there were 270,000 surnames shared by five or more people. Because almost all Chinese surnames are so common, they typically carry little information on the social status of their holders.

  FIGURE 9.2. Imperial examination hall with 7,500 cells, Guangdong, 1873.

  However, it is possible to identify thirteen relatively rare surnames that appear with unusually high frequency among those who attained the highest qualification under the Qing exam system, that of jinshi (进士). A complete list has been published of all the Qing jinshi. Indeed, in the front courtyard of the Confucius Temple in Beijing are hundreds of large stone tablets bearing the names of the more than fifty thousand jinshi of the Yuan, Ming, and Qing dynasties. These thirteen elite sur
names were identified as those that showed a relative representation among Qing jinshi between 1820 and the end of the imperial exam era in 1905 that was at least four times that of the big three surnames.9 Though relatively rare by Chinese standards, the thirteen elite Qing surnames are held by nearly eight hundred thousand people today.10 So there is plenty of evidence by which to judge the current status of these surnames.

  The imperial examinations were the meritocratic path to high-level positions in the bureaucracy. The gains from various levels of achievement on these exams were so great that to limit cheating and nepotism, they were held under strict supervision, with the candidates isolated in individual cells for the two to three days of the exam (see figure 9.2). Individual examinations were numbered, and candidates’ answers were copied out by clerks so that the graders could not recognize candidates by their calligraphy. Candidates spent years studying to attempt the various levels of the exams. The gains were not just the emoluments of imperial service but also the protection and business advantages the degrees afforded to the families of successful candidates. Extended families thus had interests in furthering the success of the most academically able of their relatives.

  Geography still matters to social status in China, a fact that slows mobility at the national level. The populations bearing the thirteen Qing elite surnames are all concentrated in the lower Yangzi River valley. These surnames are overrepresented among both exam passers in the imperial era and in modern Chinese elites. Here, to exclude geographic causes of persistence in status, social mobility rates are measured relative to the common surnames of the lower Yangzi. The mobility rates estimated for modern China would be even lower were the geographical elements in immobility not excluded.