Cadre parents and their entrepreneur children? The dual-track intergenerational reproduction of elites in China: 1978–2010

This article documents and conceptualizes a mode of reproduction of elites in a society in transition from state domination to market orientation. By focusing on China’ marketization, we explore how parents’ advantageous backgrounds have influenced the chance of their children’s attainment of certain elite positions (administrative, technocratic, or market) and whether these patterns have varied across three periods (1978–1992, 1993–2002, and 2003–2010). Using data obtained from the 2011 China Social Survey, we find that although parents’ advantageous status has a persistent effect on children’s status attainment, the reproduction of the state elite and market elite still follows two separate tracks: the children of cadres do not show significant advantage in the process of becoming entrepreneurs and managerial elites, and the children of entrepreneurial and managerial elites are less likely to join cadres. We also find that the effects of the reproduction model are still enhanced and shaped by state power in different periods. These findings demonstrate the important interplay between family background and contextual inequality and give a deeper understanding of the different trajectories of elites in contemporary China.


Introduction
In a society where economic capital outweighs political power, the positions of the economic elites and political elites seem interchangeable. This is reflected not only in the so-called revolving door between these two groups (Cohen 1986;Vidal et al. 2012) but also the fact that the offspring of the economic elites are able to take advantage of their parents' success when pursuing a political career, and vice versa (Corcoran 1995;Schubert et al. 2013;Schwartz 1987;Zeitlin 1974). After all, despite ongoing changes in the composition of elites (Savage and Williams 2010), the political elites and economic elites still belong to the upper echelon of the social hierarchy (Khan 2012;Mills 1956). They are seen as members of the same inner circle (Useem 1984), the upper class (Scott 1991), or, to put it more eloquently, the "state nobility" (Bourdieu 1998).
In a society where political power outweighs capital, however, there is no cross-sector reproduction of elites in intergenerational mobility. In state socialism "without capitalists," (state) power is dominant. Political elites are selected from the politically privileged groups (Djilas 1957;Whyte 1975). Children of the former economic elites (capitalists) from the pre-revolutionary society have no way to become market elites, as the market per se has been eliminated (Davis 1992;Goodman 2000). 1 Yet what happens in a transitional society that stands between these two ideal types? Politicized capitalism is one of these transitional forms (Nee and Opper 2007;Szelenyi 2013). It is characterized as "capitalism" rather than, for example, state socialism because even those who hold state power rely on the market for profit. However, it differs from classical capitalism in that (political) power is more useful than (economic) capital in accumulating personal wealth. Thus, in such an economy, are the children of cadres more likely to become businessmen? Or are the children of businessperson more likely to become cadre members to amass political capital for their families?
Though many scholars labeled China as a nation of politicized capitalism-or similar terms (Szelenyi 2008;Walder 2009;Zhao and Zhou 2017), it is not capitalism. Yet it provides a natural social laboratory where we can seek answers to questions about the patterns of elite mobility in a hybrid economy (Burawoy 2001;Eyal et al. 2003). For some Chinese, the combination of a cadre parent and an entrepreneurial child (usually male) is envisioned as one of the most advantageous ways to amass wealth-a situation confirmed by many corruption scandals. At the same time, the combination of a parent in business and a child in politics also conforms to the ideology of "upward mobility" for some Chinese families, which valorizes the pursuit of an official position after success in business and the idea that a child should bring honor to his or her ancestors and family. Yet, despite the common impression derived from news reports on corruption cases, there have been no quantitative empirical studies that systematically tackle this phenomenon.
Using data from a national survey, we contribute to the literature in two ways. First, we adopt a historical perspective-that is, we examine whether the status attainment model of the two elites has changed during the 30 years since China embarked on its market transition in 1978. Second, we focus on the relationship between power and capital. Although many quantitative studies, by analyzing the mobility rate of each social stratum, have examined the general intergenerational mobility of all classes in China (Li 2002;Liu 2003;Wu and Treiman 2007), we focus on the frequency and trajectory of status changes among the two most powerful groups, state elites and business elites.
Our historical focus requires us to pay attention to the temporal variations in the process of status attainment. Many researchers have pointed out that the dominant modes of intra-and intergenerational mobility are likely to change over time (Nee and 1 When we talk about classical capitalism or classical socialism, we speak of them as ideal types. At different stages of "really existing socialism," the market always exists in one way or another. A very small number of the children of the "old bourgeoisie elites" still participate in some forms of self-employment or underground commercial activities (Osborn and Slomczynski 1997;Szelenyi 2002). In China during the planned economy, some children of the old bourgeoisie still received political titles, though they may have lost control over their parents' material wealth because of the economic nationalization. Their political statuses were not strictly equal to those of the socialist political elites, but rather were a form of recognition by the Party of their "red capitalist" identity. Walder 2002;Wu 2006;Wu 2010). Change may be caused by many factors, including loss of political status after market transition (Walder and Hu 2009;Zhou and Hou 1999), changes in dominant economic rules (Szelenyi and Szelenyi 1995), and shifts in industrial policy (Walder et al. 2013) or personnel systems (Sun 2011).

Cao
Our focus on the power-capital relationship requires us to consider our topic relative to two schools of literature on the status attainment of state elites and market elites. These two schools have represented the frontier of discussion of China's social stratification since the beginning of the debate on market transition (Bian 2002;Guthrie 2000;Wu 2019). However, as Wu (2008) points out, most researchers have focused on intragenerational mobility and career advancement-in particular, how the political elites of state socialism become the economic elites of the market economy (Liu 2006;Nee 1989;Peng 2004;Walder 2002;Wu 2006). Some have even argued that virtually all of the literature on market transition, including the transitions of Central and Eastern Europe, shares the problem of paying more attention to intra-generational mobility than intergenerational mobility (Gerber and Hout 2004).

Intergenerational mobility of elites in market transition
The literature on the intergenerational mobility of China's political elites since market transition displays a consensus on the issue of "reproduction or circulation": since 1978, the positive effect of advantageous family background, whether deriving from old elites of the era before the founding of People's Republic of China (PRC) or new elites made up of state officials, has re-emerged. In particular, the offspring of officials enjoy a greater competitive edge than the offspring of other social groups (Walder et al. 2000;Zhou and Hou 1999).
In contrast to the consensus on the intergenerational mobility of political elites, there has been no empirically supported conclusion about the intergenerational mobility of market elites. Unlike its political elites, China's market elites re-emerged just after the market transition in the late 1970s. In other words, China's market economy is built from a society without capitalists (Eyal et al. 1998). Therefore, many researchers see these new businessmen as the first generation of private entrepreneurs.
There are a few stories of entrepreneurial parents producing entrepreneurial children. 2 Some researchers have studied whether the children of cadres become entrepreneurs; others discuss how family background prior to the revolution affects the third generation or beyond. Many of them find that the effect skips a generation, similar to what Szelényi (1988) refers to as "interrupted embourgeoisement" in his study of Hungarian peasant entrepreneurs in the 1980s. For example, based on studies of local cases, Goodman (2000) and Chen (2012) find that many private entrepreneurs have parents or grandparents who were from so-called (petty) bourgeois families before the liberation in 1949, and their families' mercantile tradition has had a significant impact on the new entrepreneurs.
This argument, however, has not been fully empirically tested in China. Some studies have shown that in China's countryside, the children of both the old elites (i.e., political or economic elites before the liberation) and the new elites (i.e., political or economic elites in state socialism) have been more likely to enter non-agricultural professions since the market transition (Wu 2008). This research, in effect, demonstrates the significance of advantageous family background, although the data are restricted to China's countryside before 1996. A study by Lu (2017), in contrast, shows a limited impact of advantageous family background in the case of the wealthiest Chinese entrepreneurs. Over half of those on the list of the richest people in China come from peasant or working-class families, and another 30% are the children of professionals (Lu 2017). His research, nevertheless, focuses on a limited subsection of the rich.
Studies that apply a single model to analyze the intergenerational mobility of both state and market elites are rare. Among them, two stand out as particularly relevant. In one article, Jia and Lan (2014) find that in provinces where fiscal expenditure accounts for a larger percentage of the gross domestic product (GDP) (i.e., those with big governments), people whose parents are cadres are more likely to become entrepreneurs. In general, however, they argue that the mobility pattern of cadre parents with entrepreneurial children does not exist. Their research touches on how different institutional conditions (e.g., the size of government) affect the mode of intergenerational mobility. Yet the size of government may not be an effective or direct explanatory mechanism. Moreover, the authors fail to test the possibility of entrepreneurial parents with cadre children, let alone investigate possible changes in modes of status attainment under different historical periods.
The second article is a study by Zheng and Li (2009), which uses sampling data from three cities. This study discusses only intergenerational mobility among state elites (i.e., administrative, technocratic, and professional elites, to use the authors' terminology). Some economic elites (including the managers of state-owned enterprises and entrepreneurs) are categorized in the analysis as cadre elite. 3 Nevertheless, their article echoes the question raised in our paper: do China's elites originate in different segments of society or come from the same group? We agree with them on one point-observation of intragenerational mobility alone cannot solve the puzzle. Only an analysis of intergenerational mobility can answer this question.
Zheng and Li provide very interesting empirical findings. They argue that the whole elite stratum, more precisely the state elites, is reproducing its advantageous position. This reproduction is achieved through two mechanisms. The first is exclusion, which means the elite group excludes others from joining them. The children of the elites are more likely to be among the elite. The second is intergenerational transfer. There is a free flow of personnel among the administrative, technocratic, and professional elites. These groups are mutually permeable, thus forming an elite stratum that is united and cooperative.
Some findings of this paper are in line with the conclusions of these two articles. However, our research seeks to further advance knowledge in this area. First, the two articles only examine how paternal family background affects the children's current profession. We consider both the children's first job and current job. The inclusion of the first job helps us study the extent of advantage resulting from path dependency in career mobility during intergenerational transmission. Second, by involving both state elites and market elites, we have considerably expanded the subject compared to the previous literature on the intergenerational mobility of Chinese elites, which mainly deals with either political or economic elites. Third, we not only provide a full model to analyze the trend from 1978 to 2010 (the latest year for which data are available) but also divide the data into three historical periods (1978-1992, 1993-2002, and 2003-2010) to further investigate the different patterns of intergenerational mobility caused by changes in the macroeconomic and political environment.
We conceptualize our findings as a dual-track intergenerational reproduction of elites; we explain this concept later. This concept sounds somewhat like Walder's famous conceptualization of the "dual track" of the Chinese political elite's career mobility (Walder 1995) but has little relationship with it. 4 We wish to further explore the mechanisms of this model of reproduction. Studies that concentrate primarily on the intragenerational mobility of state or market elites ironically provide important clues to this puzzle: political connection and educational level are the two most common factors.
Many arguments about the status attainment of Chinese political elites center on the "reds" versus "experts" debate. Most researchers perceive the Communist Party of China (CPC) membership and college-level education as important mediating mechanisms for upward mobility but disagree on which one has precedence. Party membership indicates political loyalty (being "red"), whereas a university diploma signifies one's educational level (being "expert") (Li and Walder 2001). The division between red and expert is also the starting point for the theory of elite dualism (i.e., bureaucrats and technocrats) (Zang 2001). Despite the differences, most researchers agree that as the meritocratic reform of the personnel system deepens and Chinese society advances, the importance of educational level, in comparison with Party membership, has increased (Tsai and Dean 2013;Zang 2004). Based on the above existing literature, we propose: Hypothesis 1: The children of China's market elites and political elites follow two different paths in the reproduction of intergenerational status.
Research on the intragenerational mobility of market elites also examines the effect of political connection on status attainment. The difference, however, is that the previous career experience of market elites (particularly careers in the Chinese state ) is a more widely used indicator than Party membership. Some argue that people with a cadre background are more likely to become entrepreneurs (Song 1998). Others, however, believe that cadres have no advantage in becoming entrepreneurs in firms created by rural elites at the township level (Nee 1991). Walder (2002) and Wu (2006) argue that the chance of cadre members becoming entrepreneurs has been decreasing over time, whereas others contend that political elites are more likely to turn into corporate entrepreneurs (Rona-Tas 1994). Moreover, the significance of higher education for the status attainment of entrepreneurs has increased (Chen 2015;Lu 2017). Other researchers, however, hold that the inherence of pro-business family culture is the key to understanding the reproduction of market elites (Wu 2008). Based on the above existing literature, we propose: Hypothesis 2: The mode of intergenerational reproduction of Chinese elites will vary in different historical periods.
In the following sections, we introduce our research design and methods after a brief review of the data source and sample characteristics. We then present the model results, which mainly address the impacts of first jobs, current jobs, and elite status attainment. In the "Main findings" section, we discuss the mechanisms underlying our findings and reveal changes in status attainment during different historical periods since 1978. The conclusion discusses how the findings of this paper can inform scholarly understandings of Chinese society now and in the future. We also suggest directions for future research.

Data
We use 2011 Chinese Social Survey (CSS2011) data gathered by the Institute of Sociology at the Chinese Academy of Social Science. CSS is a widely used national biennial longitudinal project launched in 2005, covering all provinces in the Chinese mainland, including 151 counties and city districts. CSS ensures rigorous survey procedures on several levels. In the sampling stage, this survey uses statistics from the Fifth and Sixth National Censuses of China to create the sampling frame of counties and city districts; it adopts map-and address-based sampling methods in the survey destinations to cover migrant populations. It is now widely regarded as one of the most authoritative national social surveys in China. After data cleaning, 4125 cases are retained for statistical analysis. 5 We include variables that measure social status, workplace, family background, and socio-demographic information. Summary statistics, including the percentage/mean, standard deviation, and range for all variables, are presented in Table 1.

Measurement of independent variables
The core independent variable is family background. We measure it mainly by the current/last occupation of the father or mother at the time of the investigation, depending on who is in the highest social class. 6 For those who are retired, we measure it by the last job before retirement. For the very small number of respondents whose father had passed away, we substitute the profession of their mother. With regard to data processing, we categorize the occupation codes in CSS into four main groups depending on the nature of the sector of working place (danwei) and occupation: state elite, market elite, technocratic elite, and non-elite.
The state elite is measured as administrators and officials of the Party, government, military, state-owned enterprises (SOEs), and public institutions with political rankings above the department-level (chuji). 7 Market elites include holders of middle-to senior- We put managers of SOEs into the category of administrative elites for the following two reasons: (1) Despite the ongoing de-politicization reforms, the regulation of management in the SOEs is in general the same as that of the Party cadres (Shambaugh 2008). The reform has not been truly enforced (Chan 2004). Managers of SOEs are not real entrepreneurs (Brødsgaard 2012).
(2) Managers of SOEs receive the same or similar political treatment as state cadres and have similar opportunities for political promotion. In recent years, state officials have often become managers of SOEs and vice versa. Thus, these managers can be perceived as "cadres-to-be" (Yang et al. 2013).
level managerial positions in private or foreign companies, owners of private enterprises, and self-employed entrepreneurs (getihu). 8 Technocratic elites are professionals and technocrats with middle-and upper-level professional titles in SOEs and public institutions, including senior engineers, associate professors and above, and director-level physicians. 9 The rest are coded into the "non-elite" category, including ordinary employees within the state system (e.g., workers at state-owned enterprises) and people in the market sector in a broad sense (e.g., workers at foreign or private-owned companies, freelancers who are flexibly employed). Appendix Table 4 reports the distribution of occupations.
In addition to family background, the time at which the respondent obtained his or her first job is also a key independent variable. One aim of this research is to test whether the path to attaining elite status in urban and rural China from 1978 to 2010 was different in different historical intervals. After careful consideration of the generational political succession and benchmark events of the market transition, we divided the period into three stages : 1978-1992, 1993-2002, and 2002-2010. The first stage started with the reform and opening-up launched by Deng Xiaoping in 1978, during which market-directed economic reform spread from rural cooperatives to urban factories. During this period, state-owned enterprises, which occupied the dominant position in the national economy, were not seriously affected by the reform. The second stage started in 1992 and ended in 2002. After a short hiatus in reform policies, the state initiated a new and more radical marketization with the objective of "building a socialist market economy with Chinese characteristics." Massive restructuring and privatization of state-owned enterprises provided abundant opportunities for cadres, former managers of SOEs, and a new generation of entrepreneurs but was subjected to heavy criticism from those whose interests were encroached upon during the process (Huang 2008;Lin 2008).
The third period ran from 2003 to 2012, but this study is cut off in 2010, which marks the limit of the most recent data. In 2003, the state experienced a peaceful political succession, and then in 2012, the 18th Party Congress took place. In this period, although the private market was still growing quickly in terms of the percentage of GDP, the rise of large state-owned companies generated controversy over whether the state 8 The distinction between proprietors and managers adheres to a classic Erikson-Goldthorpe-Portocarero framework. In fact, since the management revolution in the 1940s, many senior managers worldwide are, in fact, major stakeholders of their enterprises. This is also the case in today's China, and few of these major stakeholders would describe their professions as merely managerial. Both self-employed entrepreneurs and owners of private enterprises are proprietors in the EGP framework. 9 A critical concern, not only for this paper, is how far the "elite" can meaningfully be derived from a national sampling survey: "elite" might not be "elite" enough (Savage 2015; Savage and Williams 2010). The problem is that even scholars in "elite studies" cannot achieve a consensus on the categorization (Khan 2012) and normally terms as "super elite," "one percent," and "double one percent" (Keister 2014;Keister and Hang 2017) are used to describe those very top positions. Our argument is that the definition of "elite" must be contextual: it depends on who one is comparing oneself to, and it is very likely that the specific nature of the elites varies considerably across national contexts. In the everyday Chinese context, even a graduate from a prestigious university can sometimes be called an elite. Our definition might not be perfect, but it makes sense.
was advancing as the private sector retreated. The criticism of the offspring of cadres benefiting from doing business, and widening social inequality, had become more acute (Lu 2012;Walder et al. 2013).
CPC membership and education 10 are two key independent variables for our mechanism-based explanation of status attainment. CPC membership is a dichotomous variable understood as "Yes" or "No." Education is a three-level categorical variable: junior high school or less, senior high school, and college or above.

Measurement of dependent variables
The first dependent variable is the type of sector of the respondent's working place. We make a distinction between the state and non-state sectors (see Fig. 1). The state sector includes the CPC, the government, the military, SOEs, and public institutions. Nonstate consists of private and foreign-invested firms and all other economic entities that cannot easily be identified, such as self-identified freelancers, collective enterprises, hybrid working units, and farmers. This division concerns only the sectors the respondents work in, not the level of their positions.
Respondent's occupation is another dependent variable in this study. As a Goldthorpe-inspired work based on a representative national survey, we observe the respondents' first job position and their current position. Then, we divide their occupation into four categories: state elite, market elite, technocratic elite, and non-elite, the same coding as in the independent variable. The distribution of the occupations can be found in Fig. 2.

Measurement of control variables
Control variables include the interviewees' gender, age, registered permanent residence (hukou in birth), number of siblings, and region of residence (when interviewed). Because the marketization level is relatively high in the coastal area of south-eastern China and its municipalities (Fan et al. 2011), we group Beijing, Tianjin, Shanghai, Shandong province, Jiangsu province, Zhejiang province, Fujian province, and Guangdong province into one category, "developed coastal area," 11 measured by a dummy variable.

Methodology
To examine changes in the attainment of elite status across sectors in China during different periods between 1978 and 2010, we adopt logit models. Regarding heteroskedasticity in cross-sectional data, we use robust standard error in all models.
This study uses the conditional multinominal logistic regression (CMLR) (Hendrickx 2000) to test influential factors for status attainment in three different historical periods because the children's social class, which would be the dependent variable in a standard multinominal program, is now entered as an independent variable. As discussed in the 10 We note that using Party membership as an indicator for political loyalty ("redness") has been increasingly criticized. For example, Zang Xiaowei argues that the seniority of Party membership is a better indicator (Zang 2001). We are also aware that the educational level of Party members and entrepreneurs may be the result of on-the-job education, rather than full-time study (Lu 2017;Tsai and Dean 2013). Unfortunately, because the CSS2011 did not ask when the interviewees received their Party membership and diploma, we are not able to treat these as time-varying variables. 11 It is usually called "Eastern" China in many other literature. literature review section, the status order may be less stable due to China's ongoing changing social transformations. We, therefore, decide not to make any assumption about their order by using ordinal logistic regression. Finally, in all model estimates, data are weighted to represent the general population.

Main findings
Obtaining the first job: segregation of sectors Table 2 displays the determinants of the first job sector. Compared with children of non-elite families, those from families of state elite have the advantages in obtaining first jobs in the state sector significantly. Specifically, individuals from families of state elites are more likely to find first jobs in state sectors, public institutions, and SOEs as a starting point for their careers. Moreover, the children of state elites are more likely to work in state sectors than the children of market elites (model 1), which demonstrates some differentiation within the state elites. In model 2, when the control variables are employed, the advantage of state elites' children is still significant. Meanwhile, regardless of both education level and CPC membership which increases the probability of the state sector (model 3), children from the family of state elite have a higher probability of entering the state sector. It reminds us that when analyzing the acquisition of professional status, we need to consider the role of the type of initial working place in the process.

Elites' status attainment: impact of family background
We employ CMLR to further examine the effect of family background on status attainment by controlling other relevant variables (see Table 3). In the scaling metrics of Fig. 1 Distribution of first job sector by family background model 1, the differences between scale values show how the logit of one occupation versus another is affected by education, CPC membership, first job sector, control variables, and scaled father's occupation. The market elite rather than state elite as father increases the likelihood of the children getting a better occupation (.503 vs. .348). The Row and Columns model 2 (Goodman 1979) estimate indicates that there is a strong association between father's social class and children's class (p<0.05). The impact of having a father who is a state elite rather than a market elite on the logit of becoming a state elite rather than a market elite is .09 ((.348 − .503) × − .524 × (.632 + .428)).
Model 1 also indicates that the inheritance effects of the state elite fail to confer a significant advantage, but it has a significant positive impact on the market elite (noting that the parameter has a large standard error). It means that the background of the market elite has much more advantage in elite reproduction than the other elite group. Meanwhile, there are the stereotyped ordered regression (SOR) effects of education, CPC membership, and the first job sector. The impact of one rank of education on the Logit of becoming a state elite versus a market elite is 2.03 (1.912 × (.632+.428)), the impact of being state sector rather than non-state sector is 1.11, and the impact of being CPC membership rather than others is .74.
As a result, we note that though children of state elites and market elites have an advantage in reproducing their elite position, the effect of crossing sectors is not apparent. That is, the children of state elites are more likely to join the state elites, whereas the children of market elites have more advantage in becoming market elites. Our statistic results show that education, CPC membership, and experience of the state sector have interaction effects in elite status attainment. The average number of siblings is 2.34. We control this variable in all the models. The number of siblings has a significant negative effect on entry into the state system (P < 0.01) but has no significant effect on the acquisition of elite status.

Elites' status attainment: differential impact of family background in different periods
We further test the impact of family background on state and market elites using CMLR models again. As shown in Table 3, models 2, 3, and 4 deal with subsamples in three historical periods: 1978-1992, 1993-2002, and 2003-2010. We find that the inheritance effect had a positive impact in 2003-2010, but no significant effects in the other two periods. Regarding scaling metrics, the advantage of state elite in status reproduction is stable (noting the difference between state elite and market elite in 1993-2002 is rather small). Therefore, the segregation of elite attainment is tested again.
The results of SOR analyses in the sub-sample indicate that both education and first job sector are significantly influencing children's elite position. In regard to the first job sector, the experience of working in the state sector has a positive effect on joining the elite position later in life, whereas the effect is decreasing from 1978 to 2010. However, CPC membership does not show significance in the 1978-1992 period but has a significant impact in the other two periods, increasing significance after 2003.

Conclusion and discussion
This study focuses on a rarely discussed aspect of the intergenerational mobility of elites in China by studying the trajectories of the state elite and market elite from a  historical perspective. In this section, we highlight several major contributions of this study to the field. We conceptualize the main findings of this research as a dual-track intergenerational reproduction of elites. This concept has two aspects. First, in the reproduction of elite status, both state elites and market elites are likely to pass down their advantageous status to their children. Second, although there are some cases of two trajectories within one family, 12 in general, the national statistics show that the barriers between the two tracks still exist. There is no apparent crossover effect in the reproduction of these two elite statuses. In other words, the children of market elites in China do not have the same advantage as the children of state elites when trying to enter politics; there are still two trajectories of intergenerational career mobility for market elites and state elites. We note that in the analyses of contemporary capitalism, the relationship between money and power is also changing (Khan 2012;Savage 2015) and various forms of social closure exist corresponding with different forms of capital (Flemmen et al. 2017). However, the distinction between the two trajectories of mobility in China is still sharp and dominated by fundamentally different logics.
Some researchers may argue that the current situation is only temporary: as generations succeed each other, cross-sector intergenerational reproduction will finally become the dominant mode of elite reproduction. Without longer-term data, we are not able to predict historical trends in elite reproduction. Nevertheless, we cannot rule out the possibility of cross-sector reproduction based on two of our findings. (1) A diversified trend of status attainment for both market and state elites is emerging, which may encourage greater openness in upward mobility. Our analysis of the intermediary factors for elite status attainment shows that the importance of education and Party membership for becoming state elites increases; at the same time, the importance of first jobs has been decreasing since 1992. If this diversification of intra-generational mobility continues, it may exercise an effect on intergenerational mobility.
(2) Our research shows that since 2003, fathers' status as members of the state elite has begun to show a small effect on their children becoming business managers in the market sector. This effect may have been underestimated. As one recent study shows, people who have at least one parent who is an administrative cadre are more likely to secure a position in finance, bureaucracy, public institutions, or international organizations (Li et al. 2012). Our observation tells us that even in the case of cadre parents with businessperson children, instead of becoming entrepreneurs who have their own businesses, the children of cadres prefer jobs as well-paid managers in domestic or overseas companies in the areas of finance, real estate, energy, or brokers between companies and government departments. Because these people usually want to remain low profile or invisible, social research using common methods cannot fully account for them.
For the children of entrepreneurs or managers of private enterprises, the possibility of going into politics is less likely to be discounted. From what has been reported in the news and a number of qualitative studies, cases of businessperson parents with cadre children seem to be increasing. This phenomenon is not a point of shame for such families but rather an achievement worth boasting of, in many cases. Unlike businessperson sons with cadre fathers, who may be accused of taking advantage of political capital in exchange for economic benefits, 13 the combination of businessperson parents and cadre children creates less temptation with regard to economic and political interests. Under the current political framework in China, market elites can use their channels in the state agency (e.g., People's Congress, People's Political Consultative Conference, the Association of Industry and Commerce, and even the Party Congress) to secure political status for themselves or their families (Chen 2015;Dickson 2008;Lu 2014). At the local level, some entrepreneurs are even given a (ceremonial) deputy position in the local government to reward their work (Zhang 2016).
Thus, if the intergenerational mobility of the elites is increasingly cross-sector, the faster track is probably cadre father/businessperson son rather than businessperson father/cadre son, as a result of the unequal status of state elites and market elites-in Bourdieu's words, the market elites are the "dominated dominant class" (Bourdieu 1998).
The study of elites is an examination of their resources and a study of the conversion of those resources. Some scholars have observed that the study of elites is, in fact, the study of resource control and distribution. We have accordingly used a top-down approach to examine power and inequality (Khan 2012). China falls in the middle of the spectrum between classic state socialism and classic capitalism and is closer to the end where state power is dominant. This condition is likely to remain stable in the near future. In the long run, whether state elites and market elites will remain two separate groups that reproduce themselves on dual tracks across generations, or whether they will incorporate, depends on the interaction between these two elite groups and the rest of society. This process will have a significant impact on the unity of the elite group and the legitimacy of state power.
This study also has some limitations: our analysis is based on cross-sectional survey data, and we need to be cautious in inferring results. In the future, we can use panel data, if available, to explore the causal mechanism. At the same time, use non-survey data, such as wealth list, to conduct more advanced research on the top elite groups. This is what we are doing in other research projects.