The Son Also Rises Page 18
The relative representation of these thirteen Qing elite surnames among modern elites was calculated by comparing them to three equally regionally favored surnames, Gu (顾), Shen (沈), and Qian (钱), the “regional three,” that have only average status in the lower Yangzi. These three surnames, held by more than ten million people now, offer a large and stable comparison group.
Figure 9.3 shows the relative frequency of the Qing elite surnames in various modern Chinese elites with respect to the “big three” and “regional three” surnames. The modern elites employed are high officials in the Nationalist government in China from 1912 to 1949; professors at the ten most prestigious Chinese universities in 2012; chairs of the boards of companies listed in 2006 as having assets of US$1.5 million and above; and members of the central government administration in 2010. Compared to the big three surnames, the thirteen elite Qing surnames are nearly as prominent among Chinese elites in recent years as are Brahmin surnames among elites in India. But, as noted above, geography is responsible for much of this effect.
Relative to the regional three surnames (Gu, Shen, and Qian), the thirteen Qing elite surnames are less overrepresented but still distinctive in the Qing and modern eras. Their relative representation was 2.28 among high Nationalist officials, 1.88 among professors at elite universities in 2006, 1.62 among chairs of company boards, and 1.46 among central government officials. Had these surnames declined to average status by now, all the measures of relative representation would be one.
The stock of professors, company board chairs, and government officials in 2006–12 consists almost entirely of people born after 1940 and therefore educated under communism. (Only those age 60 or older would have completed their education before the Cultural Revolution and the disruption of higher-education institutions between 1966 and 76.) These three groups are observed roughly seventy-five years, or 2.5 generations, after the high Nationalist officials of 1912–49. Assuming that the Nationalist officials were a similarly exclusive elite, the intergenerational correlation of status for these surname groups between the Republican era and 2006 is 0.9 for professors, 0.8 for company board chairs, and 0.74 for central government officials.
FIGURE 9.3. Relative representation of Qing elite surnames among modern Chinese elites.
Assuming that the jinshi elite of 1820–1905 and the Nationalist officials of 1912–49 were also comparably elite groups, it is also possible to calculate the mobility rate between these two eras.11 The decline of the relative representation of the thirteen Qing elite surnames among the Nationalist officials implies an intergenerational correlation of status of 0.8.
These results are supported by data on student admissions to Tsinghua University from 2003 to 2011. As one of the top ten universities in China, Tsinghua again represents an elite at the pinnacle of the education distribution. The relative representation of the thirteen Qing elite surnames, even relative to the three regional surnames, was still 1.62. Since these students, entering university at age 18, are a full three generations removed from the Nationalist officials of 1912–49, this figure again implies an intergenerational correlation of 0.83. Despite the disruptions of civil war, land reform, and the Cultural Revolution, the data show a very slow decline in status of the Qing elite within Communist China.
These results are all subject to random errors, because the sample sizes of some of the elites are modest. The samples include 9,363 Qing jinshi, 26,738 Nationalist officials, and 26,429 elite university professors. Of these samples, only 39, 62, and 51, respectively, hold one of the thirteen Qing elite surnames. Just by chance, the number of Qing elite surnames could deviate significantly from the underlying trend in any of these samples. So none of the individual estimates of intergenerational correlations from these samples can be relied on. The true value of each could deviate substantially from the estimate.
Nevertheless, it is clear from these samples that these thirteen Qing surnames remain elite. And there is another sample so large as to give absolute confirmation of their continuing enhanced status: the household registration details of all 1.3 billion Chinese in 2008. This database records whether individuals have completed a university degree. There are 29,604 people with one of the Qing elite surnames who have a university degree. Now the sampling error is minuscule.
The household registrations reveal that in 2012, 2.6 percent of all Chinese held university degrees; for holders of the regional three surnames, the figure was 3.5 percent. Degree attainment is 58 percent higher for the Qing elite surnames than for the big three surnames and 30 percent higher than for the regional surnames. The relative representation of the Qing surnames is lower among university graduates than among professors, but this is exactly what we would expect given that university graduates in general are a much less elite population.
Degree attainment has expanded rapidly in the last ten years, and therefore the average date for the awarding of these degrees is 1998. Thus this population is three generations separated from the Nationalist elite of 1912–49. Allowing for this, and for the lower exclusivity of the pool of modern university graduates, the implied intergenerational correlation over these three generations is 0.78, very similar to what was implied by the smaller more elite samples.
There is still random error associated with the smaller sample of Nationalist officials. This could lead to under- or overestimation of the rate of social mobility between 1912–49 and 2008. Taking this into account, we can be 95 percent confident that the true intergenerational correlation of status for Communist China lies in the range 0.71–0.92. Even at the lower bound of this range of estimates, this is a remarkable degree of status persistence by the elite in a society that experienced the degree of turbulence and anti-elitist actions in the early years of Communist rule.
Surname–Place of Origin Identifiers
Other evidence for rates of social mobility in Communist China is supplied by identifiers that combine surnames and places of origin. Given how common most surnames are in China, families and lineages distinguished themselves in the imperial and Republican eras by associating the name with the family’s place of origin. Such combinations include the Fan family of Ningbo (宁波范氏), the Zha family of Haining (海宁查氏), and the Weng family of Changshu (常熟翁氏). The surnames themselves—Fan, Zha, and Weng—are average in social status at the national level. But these particular combinations of surnames with places of origin represent an imperial elite whose later regression to the mean can be measured.
To measure social mobility in this way, the focus here is on two regions in the lower Yangzi, South Jiangsu and North Zhejiang, which lie, respectively, north and south of Shanghai (see figure 9.4). This has long been one of the most economically advanced areas of China. Required for this local mobility estimate are the frequency in these regions of each surname–place of origin combination and the frequency of these combinations among various local elite groups over time.
For want of a better source, surname–place of origin population shares are estimated from the records of twenty thousand soldiers from the lower Yangzi who died in the civil wars of 1933–36 and 1945–49 and the external wars of 1937–45 and 1950–53.12 (Modern population censuses might seem better sources from which to estimate these population frequencies, but no modern population count gives name frequencies by surname and place of origin.) It is assumed that the share of surnames among soldiers who died between 1933 and 1953 represent the population shares in these areas as a whole and that the population share of surnames has been constant over time.13
FIGURE 9.4. The areas of the lower Yangzi used in the surname–place of origin mobility estimates.
Again, unusual Qing-era success on the imperial exams is used to identify the early elite surname–place of origin combinations in these two regions. The records used in this case are the far more numerous ones of the juren, those who succeeded at the provincial level of the imperial examinations. Surname-location combinations that occurred at five times or more of the e
xpected frequency among juren in each location from 1871 to 1905 were classified as constituting local surname elites.
To trace the status of these groups into the present, we need to know the family place of origin for members of various later elites. This information is available for the Republican era because it was included in university rosters and alumni records, which provide data for eleven thousand students.
Under communism, such forms of identification became less common. The Communists, after all, wanted to remake society, not perpetuate the ancient lineage of successful families. But fortunately, two sources from the Communist era still reported family places of origin at the county level. For modern Zhejiang, the Zhejiang Provincial Archive lists notable Zhejiang people born between 1930 and 1990 and includes the family place of origin. This list covers a spectrum of occupations, such as entrepreneurs, artists, professors, and officials; of these, only those with a university degree were counted as elite. The notable people list also includes, conveniently, people who migrated from the region.
Data on later elites is obtained for South Jiangsu from the Nanjing University archive, which records students by name, place of origin, and year of graduation for the years 1952–2010. This institution ranks third among Chinese universities, and two-fifths of its students are from Jiangsu. To enroll, students must rank among the top 3 to 5 percent of those taking the National College Entrance Examination (gaokao [高考]). Since less than half of each cohort typically took this exam in the Communist era, Nanjing students represent the top 0.6 to 2.5 percent of the South Jiangsu distribution of scholastic achievement.
Comparing the relative representation of the 1871–1905 exam elite in two subsequent Republican elites—students and officials—in the years 1912–49 gives an estimate of persistence rates for the Republican era of 0.78–0.81, consistent with estimates in many other societies (see table 9.1).
The implied social mobility in North Zhejiang for the Communist era is not significantly higher: the persistence rate is still 0.74. This mobility estimate is partly based, however, on the achievements of people born in Zhejiang who emigrated to Taiwan, Hong Kong, or the United States but were still included on the list of notable people.
In South Jiangsu, based on the Nanjing student records for 1952–2010, social mobility rates for the Communist era were higher than before. But there was still a strong connection between surname–place of origin combinations overrepresented in 1952 and those overrepresented in 2010. (Again, given the modest sample sizes, there may be considerable sampling error underlying these estimates.)
Overall, although these local estimates do suggest faster social mobility in the Communist era, they also imply a remarkable persistence of status in a society convulsed at times by revolutionary fervor and a determination to root out “class enemies.”
TABLE 9.1. Persistence rates for the lower Yangzi, 1871–2010 Period Surname groups
Mobility among the Nationalists: Taiwan, 1949–2012
More than a million supporters of the defeated Republican government of China retreated to the island of Taiwan in 1949. These people, most of whom had been part of the mainland elite, immediately formed a new political and social elite on an island populated by only six million people before their arrival. With the arrival of this immigrant elite, what were social mobility rates in Taiwan after 1949?
As a matter of policy, the Taiwanese government does not collect or publish information about the relative status of the descendants of natives and mainlanders in modern Taiwan. But as Yu Hao shows in his recently completed PhD dissertation, it is possible to identify groups from each community using surnames.14 Using information from 1956, when the population census reported surname totals by ethnic status, Hao identifies three hundred surnames whose holders were at least 95 percent mainlanders, and thirty surnames whose holders were at least 98 percent native Taiwanese. These surnames had average status in their respective communities.
Using records of university students, doctors, and business leaders, Hao then estimates social mobility in Taiwan for the period 1949–2012. The isolation of Taiwan from the mainland before 1990, and the ban even today on students from mainland China pursuing graduate degrees in Taiwan, means that we can be relatively certain that the holders of the three hundred mainland surnames in Taiwan now are the descendants of those who held the name in the previous generation.
Table 9.2 summarizes Hao’s persistence rate estimates across the last two generations for three different elite groups: National Taiwanese University students, medical doctors, and business leaders. As a background to these estimates, note that individual-level estimates show high rates of intergenerational income mobility in Taiwan. The intergenerational correlation is estimated to be on the order of 0.17–0.24, as low as in Nordic countries.15
As always, the estimates in table 9.2 show much greater persistence than those for individual families. However, these estimates, in the range 0.55–0.64, are low compared to those seen in the other surname studies. They certainly suggest, ironically, that social mobility after 1949 was faster in class-ridden Taiwan than in the class-free society Mao built.
TABLE 9.2. Persistence rates for Taiwan, 1949–2012
Persistence rate
Group
Unadjusted
Adjusted for emigration of mainlander elite
National Taiwan University students
0.55–0.66
0.61–0.72
Medical doctors
0.53–0.65
0.59–0.71
Business leaders
0.58–0.62
0.64–0.68
However, after the 1950s there was significant migration from Taiwan to the United States and Canada, mainly by the more-educated mainlander population. In the United States, these immigrants and their descendants have among the highest university graduation rates of any community. Such migrants in the United States and Canada now number 0.7 million, equivalent to 3 percent of the population of Taiwan. The subtraction of this mainlander elite from the population of Taiwan by emigration increased estimated social mobility rates within Taiwan. There are consequently fewer mainlanders at National Taiwan University or working as medical doctors and business leaders. There are correspondingly more native Taiwanese in these positions who are descended from families that were lower class in 1949.
The information does not exist to allow Hao to estimate what the true persistence rates would have been absent this migration. He cannot observe exactly how selective the migration was or how concentrated among mainlanders. We can be confident, however, that the mobility rates reported in table 9.2 overstate true rates of social mobility.
Yu Hao estimates counterfactually that this effect could have reduced measured persistence rates for educational status by as much as 0.06.16 Table 9.2 shows the estimated persistence rates with this upward adjustment. The average measured persistence rate is 0.66, which is now not far below the estimates found for other countries. And, as noted, the effects of emigration are hard to estimate.
Conclusion
The startling observation in this data is that social mobility rates in China under communism, even with the execution and exile of significant numbers of the middle and upper classes, have been just as low as in countries that have not experienced such social turmoil. The Qing elite are still overrepresented at the top of Chinese society. In particular, despite Mao’s best efforts, even in the machinery of central government, the thirteen Qing elite surnames are more prevalent than expected. Despite Mao’s best efforts, “class enemies” are strongly entrenched within the current Communist government of China.
These results are in line with some more-conventional studies of social mobility in modern China. Studies that examine social mobility by classifying people into groups according to the social background of their families should, according to the interpretation of social mobility in chapter 6, show much less mobility than would be expected from conventional estimates of mobility rates i
n recent years. One such study is that of Robert Walder and Songhua Hu, who classify people as descendants of the old elites (地富反坏右), landlords and businessmen before 1949, or descendants of the old poor (贫下中农), those whose forebears were tenants and workers before 1949. The descendants of the old elites had to pursue professional careers instead of politics after 1949 because they were discriminated against in the Communist Party. Nonetheless, they maintained advantages in educational attainment. Once discrimination inside the party was officially ended after 1976, they performed much better in all occupations relative to the descendants of the old poor.17
This slow rate of mobility in the modern era echoes similarly slow social mobility rates in the imperial era (see chapter 12). Despite the classic study of Ping-ti Ho, suggesting rapid social mobility and an open meritocracy in China in the exam era, surname distributions imply strong persistence of elite surnames among exam-degree attainers.18
Some scholars, such as Cameron Campbell and James Lee, have interpreted this slow mobility and long persistence of elites in China as revealing the importance of kin networks and extended-family strategies.19 In the imperial era, given the low pass rates on the provincial and national exams (the pass rate for the provincial exams was typically less than 1 in 200), it was unlikely that a nuclear family could produce juren or jinshi over consecutive generations. For any individual child, luck was as important as talent in determining success in the exams.20 The best strategy, then, was for the heads of kin groups to pool resources and try to secure an exam pass for the best candidates from each generation. The successful candidates then had to reciprocate by providing aid to the kin group that had secured their appointment.