Social networks in China: A thematic review of guanxi scholarship in the past decade

Abstract

This paper is a thematic review of the theoretical analysis and empirical research carried out by sociologists about guanxi networks in the past decade. Among other topics of sociological interest, the authors have chosen to focus on five aspects of guanxi networks in China: diverse theoretical approaches, changing trends, causal inference, positive and negative effects, and relationally altered social stratification. In each of these aspects, the authors synthesize the main points and conclusions of both Chinese- and English-language publications, while providing a concise discussion on the directions of future research of sociological significance and highlighting potential avenues for further exploration in the field.

Guanxi networks are a vantage point for understanding the social logic of how China works (Bian, 2019). As a form of social connection among individuals, guanxi networks have a profound influence on both individual behaviors and social integration. Chinese scholars have dedicated themselves to exploring the conceptual connotations, cultural origins, and theoretical mechanisms of guanxi networks over the past half-century. Their research has examined the role of guanxi networks in various aspects of economic actions, resource allocation, social support, and social integration (Bian, 1997, 2019; Chen and Volker, 2016; Huang, 2008; Li and Guo, 2022). In parallel, the international academic community has utilized guanxi networks as a vantage point from which to gain a deeper understanding of Chinese behavioral patterns, cultural traditions, and economic operating rules (Barbalet, 2021; Burt et al., 2018). However, debates among scholars have emerged around the theoretical orientations, dynamic trends, and causal effects of guanxi networks (Bian, 2018; Chen and Volker, 2016; Guthrie, 1998; Lin and Ao, 2008; Yang, 2002; Zhai, 2023). In response to these debates, substantial progress has been made in empirical research on guanxi networks in the past decade, facilitated by the availability of survey data and the application of quantitative methods. Nevertheless, there remains a lack of systematic reviews that comprehensively assess the progress made across different research topics in the field of guanxi research.

This paper is a thematic review of the theoretical analysis and empirical research carried out by sociologists about guanxi networks in the past decade or so. It is important to clarify that this study focuses on individual-level guanxi networks and does not encompass collective-, organization-, or macro-level social networks. Thematic reviews of guanxi scholarship in earlier decades were based mostly on English-language publications (Bian, 2018; Gold et al., 2002), while the present review covers a selection of the most influential publications in both Chinese and English since 2010. The selection of Chinese-language publications targeted top sociology journals in China, such as Sociological Studies (社会学研究), Chinese Journal of Sociology (Chinese version)/Society (社会), Sociological Review of China (社会学评论), and Open Times (开放时代). A similar criterion was used in the selection of guanxi-related publications from top sociology journals and China studies journals in English, such as the American Journal of Sociology, American Sociological Review, China Quarterly, and Social Networks. It is worth noting that the selected literature approximately consists of 40% Chinese-language publications and 60% English-language publications.

Among other topics of sociological interest, we have chosen to focus on five aspects of guanxi networks: diverse theoretical approaches, changing trends, causal inference, positive and negative effects, and relationally altered social stratification. These five aspects encompass not only some of the most extensively debated topics involving guanxi networks but also areas where empirical research has achieved notable advancements. The selection of these five themes would satisfy a theory-informed empirical sociologist who wants to move guanxi scholarship forward from the current status quo. With this intention, the sociologist would ask a set of interrelated questions: (1) What are the theoretical debates about the essence of guanxi networks? (2) What have been the long-term trends in research on guanxi networks that would give us empirical clues on which theories may be falsified while other theories confirmed? (3) No matter what theories are falsified or confirmed, do guanxi networks really matter? That is, are they causally consequential? If so, (4) What are the positive and negative effects of guanxi networks on dimensions of social life? And finally, (5) What do these guanxi effects mean for the dynamics of China's social stratification—the center of attention of most sociologists, inside and outside the country, for several decades? In addition to a synthesis of the main points and conclusions of the covered publications, for each of the five aspects we will provide a brief discussion on directions of future research of sociological significance.

Theoretical approaches to guanxi networks

How can one understand the essence of guanxi networks? For a long time, scholars have developed diverse theoretical approaches to exploring the multi-faceted meanings of guanxi networks (Barbalet, 2021), which continued to be reflected in the post-2010 publications on guanxi scholarship. In this section we pay attention to three key aspects of guanxi networks: (1) cultural connotations, (2) conceptual definitions, and (3) measurement methods. These studies serve as a valuable theoretical and empirical foundation for comprehending the essence, trends, and effects of guanxi networks in China.

Cultural connotations

What is guanxi? “The Chinese word guanxi is not a term which can adequately be expressed by an English-language equivalent of one word, the concept is too culture-specific” (Parnell, 2005: 35). The so-called cultural specificity of guanxi is, in fact, multi-dimensional, and it is deeply rooted in the Confucian codes of behavior which have lasted for more than two thousand years (King, 1991). Among them are the five cardinal relations (wu lun 五伦), emphasizing, in order of importance, (1) the benevolent and loyal monarch–subject relationship in the court, (2) the benevolent and filial father–son relationship in the family, (3) the benevolent and obedient husband–wife relationship in marriage, (4) the brotherly relationship of love and respect among siblings, and (5) trustful and responsible bonds in non-kin relationships (Bian, 2019).

Post-2010 guanxi scholarship shows that the lasting legacies of the five cardinal relations have been translated into three normative expectations for the social interactions of Chinese individuals even today. These refer to the expectations of ganqing (感情), renqing (人情), and mianzi (面子), which can respectively be understood as (1) the expectation of affectionate interaction between interacting parties valuing mutual sentiment for one another, (2) the expectation of reciprocal obligations between the parties for future favor exchanges, and (3) the expectation of respect to be obtained from the parties when mutual affection is felt and reciprocal obligations are fulfilled (Shen, 2022). Many studies by Chinese researchers demonstrate that despite 45 years of marketization and economic takeoff, China today is still suffused with these cultural connotations of guanxi, normalizing social interactions in all life domains (Zhai, 2016). These cultural connotations form the primary contents of an emerging school of social thought (Yang and Xiang, 2022), providing the theoretical origins and foundational reasoning behind a call for creating a Chinese sociology (Zhou, 2023).

Conceptual definitions

Cultural connotations cannot substitute for conceptual definitions. The former are culturally specific and largely descriptive, and the latter are cross-cultural and theoretically constructive. Based on his review of pre-2000 guanxi scholarship, Yanjie Bian (2017) summarizes diverse understandings of guanxi into three conceptual definitions. The first conception defines guanxi as the social extension of familial ethics into society at large. That is, close non-kin ties are culturally expected to be like kin ties (Fei, 1947; Liang, 1949), thus engendering the term “pseudo-kin relations” (Lin, 2001a). The second conception defines guanxi as “instrumental-particular ties” (Walder, 1986), a term that emphasizes the instrumental functions of guanxi ties and guanxi networks. The third conception views guanxi as social networks of asymmetrical exchange relations. For example, Lin (2001a) argues that, unlike economic exchanges, social exchanges are the exchanges of expressive and instrumental favors whose circulation is based on shared sentiments of interactive parties, thus having the nature of asymmetric exchange. According to Lin, social exchange networks are universal, and guanxi networks as characterized certainly exist in all cultures and societies beyond China.

In the English-language literature, post-2010 researchers have added three new conceptual definitions of guanxi to the social science literature. One addition defines guanxi as connections of high trust independent of third parties (Burt and Burzynska, 2017). This definition emphasizes the importance of mutual trust in guanxi networks and highlights that guanxi involves strong personal connections. The authors show that this definition of guanxi is universally applicable, although guanxi practices in the business world are more frequent and prevalent in China than in America and Europe (see also Burt et al., 2018, 2021; Burt and Opper, 2024). Another addition defines guanxi as ties of instrumental particularism (Barbalet, 2021), which excludes kinship ties from the confines of guanxi research but broadens its applicability to non-kin relations in all cultures and societies around the world. Still another addition defines guanxi as a categorical variable of five kinds of ties that are mutually exclusive from one another in terms of how relational sentiments and obligations are configurated within them (Bian, 2018, 2019): (1) ties of connectivity recognize the shared community of interacting parties and flows of information between them; (2) sentimental ties value human affection, including sympathy, care, love, and a sense of altruistic, non-reciprocal help for significant others; (3) sentiment-derived instrumental ties entail the circulation of substantial resources that are unintended from the formation of the tie; (4) instrumental-particular ties circulate valuable resources according to the norms of reciprocity, “face”, and favor, and such norms are well understood at the formation of the tie; (5) obligational ties are a well-understood informal contract that ignores relational sentiments and only values the principle of reciprocity, thus having the propensity to facilitate corrupt or illicit exchanges.

In the Chinese-language literature, Zhai (2023) proposes a theory relational dimensionality to contextualize different social relationships in a time–space interactive scheme. In a two-by-two scheme, social interactions can be categorized based on the duration of interactions (shorter or longer) and the rate of spatial mobility (higher or lower), thus creating four ideal types of relationships: (1) fixed relationships emerge under the conditions of longer duration and lower spatial mobility, and family and kinship relations are characteristic of such relationships that are centered around the ethics of piety; (2) friendship relationships emerge under the conditions of longer duration and higher spatial mobility, and they are centered around the ethics of righteousness and obligation; (3) contractual relationships emerge under the conditions of shorter duration and lower spatial mobility, and superior–subordinate relationships are characteristic of this ideal type that is centered around the ethics of loyalty; (4) loose relationships emerge under the conditions of shorter duration but higher spatial mobility, and relationships among strangers are characteristic of this ideal type that is centered around the ethics of politeness and mutual respect. Zhai argues that these four ideal types of relationship exist in all societies, although relational cultures vary among them depending on in which of the four modes a nation-state has been contextualized in its history and most recent status quo (Zhai, 2023).

Measurement methods

Neither cultural connotations nor conceptual definitions are deterministic. Their analytic usefulness is to be verified in empirical studies, which depend largely on measurement methods. Since the 1970s, social network analysts in Western countries have explored many methods to measure personal networks, and the most influential are those of name generators (Burt, 1984; Marsden, 1987), position generators (Lin and Dumin, 1986), and resource generators (van der Gaag and Snijders, 2005; van der Poel, 1993). Most recently, the International Social Survey Programme has developed the Social Resources and Social Capital module of multiple survey items, which have encouraged fruitful empirical analyses of changing patterns of social networks in the twenty-first century around the world (Sapin et al., 2020).

In the past decade or so, China-based researchers made a collective effort to experiment with different methods to measure guanxi networks. This collective effort has been led by researchers at the Institute for Empirical Social Science Research (IESSR) at Xi’an Jiaotong University, under the directorship of Yanjie Bian. It has generated the following measurements: Chinese New Year greeting networks, job-search networks, social eating networks, and inter-ethnicity networks. A Chinese-language review describes each of these measurements (Bian and Yang, 2019), and these measurement methods are widely used in the Chinese General Social Survey, Job-Search Networks Survey (JSNET), and China Labor-Force Dynamics Survey. Here we briefly discuss their analytic efficacies.

Chinese New Year greeting networks

This method is extended from the position generators first developed by Nan Lin and Mary Dumin (1986). In a selection of socially significant occupations, Lin and Dumin's original work asked American survey respondents to identify those positions in which their social contacts of any kind had jobs. The Chinese modification situates this method in the context of the annual Chinese New Year celebration, a week-long period of socialization. In this way, the Chinese modification emphasizes the sentimental foundation of guanxi networks by linking the position generators to an event of cultural significance among the Chinese. The modified method has proved to be a measurement innovation in the study of how Chinese guanxi networks are related to class stratification (Bian, 2008a), wage attainment (Zhang and Cheng, 2012), and the dynamics of social capital (Li and Guo, 2022).

Job-search networks

The theory of the strength of weak ties has sparked a longstanding research tradition about the role of social contacts networks in labor markets (Granovetter, 1973). At issue is not merely the relative efficacies of strong ties and weak ties for job searches (Mouw, 2003), but also about how network-mobilized information and influence matter for job attainment (Bian, 1997) and wage returns (Bian et al., 2015a). This interest has led to a series of periodic surveys by the JSNET project, conducted in multiple Chinese cities from 1999 to 2021. A battery of 15 items has emerged from the JSNET surveys that permit analyzing the causal effects of guanxi networks on subjective and objective measures of job quality in transitional China (Bian, 2022).

Social eating networks

Eating with significant others is both an ancient and contemporary mechanism of maintaining social bonds (Dunbar, 2017; Simmel, 1997). Research on guanxi treats social eating as both a network builder and a resource mobilizer for favor exchanges. Bian and He's (2022) study reports that in the first aspect, 89% of the participants involved in social eating gained new friends; in the second aspect, approximately 50% of the social eating participants were asked to provide favors to other participants within the social eating context. This method of analyzing social eating networks has proven to be analytically valuable in research on social capital mobilization in China (Li and Li, 2016b) and East Asia (Bian and Guo, 2015), and has also been incorporated into the International Social Survey Programme (see Sapin et al., 2020).

Inter-ethnicity networks

Measuring social networks in multi-ethnic countries also poses challenges due to cultural differences, language barriers, and group diversity or heterogeneity. In China, for example, ethnic minorities have developed heterogeneous cultural circles due to different historical origins and natural environmental conditions, giving rise to at least two challenges. The first is how to measure and compare guanxi networks between different ethnic groups in the same social environment. Scholars in China have developed the Important Festivals networks to measure the mutual greeting networks of different ethnic groups during their respective important festivals, which has proven to be a valid measurement method in recent studies (Guo and Li, 2022; Li, 2020, 2022). The second is how to capture guanxi networks that cross ethnic or cultural boundaries. The ongoing survey project by the IESSR team is working through an innovative application of position generators and resource generators to measure social networks across ethnic groups. These efforts aim to promote a more comprehensive and nuanced understanding of guanxi networks in multi-ethnic countries.

Looking ahead

Measuring different types of guanxi networks is an area in which future research needs to break new ground. Bian's (2018, 2019) classification provides a valuable framework for understanding various manifestations of guanxi and lays the foundation for future studies in this field. However, measuring these dynamic and complex guanxi networks poses challenges. Researchers need to carefully consider the emotional, obligatory, and fundamental dynamics of these different types of guanxi networks and may employ quantitative methods to assess the presence and strength of each tie within specific guanxi networks. Additionally, observing and comparing behaviors, emotions, and interactions in different social contexts can offer valuable insights into understanding the nature and functions of guanxi networks. It is also important for future research to explore other forms of measurement and examine the types of guanxi networks that are prevalent in Chinese society but have received limited attention in the literature. These include alumni networks, associations of people from the same county/city, professional associations, business associations, and more. Understanding the characteristics, functions, and effects of these networks can provide valuable insights into their role in social dynamics and individual outcomes.

The incorporation of knowledge on China's localized concepts, particularly the concept of guanxi, within international systems of codification is a necessary pathway for future guanxi studies. Faced with the boundaries arising from language differences, promoting communication in the social sciences becomes crucial in enhancing understanding and consensus within a globalized context. It is essential for scholars to incorporate China's local concept of guanxi into international knowledge-codification systems and establish connections between guanxi-related knowledge and sociological theories. As previous research has revealed (Bian, 2019; Zhai, 2023), guanxi networks are not exclusive to China but can vary across cultural contexts in different countries. Therefore, measurement methods of guanxi networks need to consider cultural differences and capture the generalized essence of guanxi networks for cross-national research.

Changing trends of guanxi networks

Guanxi networks, like all social networks, are not static but dynamic. Two changing trends of guanxi networks have been studied in the post-2010 scholarship: (1) the changing significance of guanxi networks in economic life and (2) the changing stock of network social capital. These works offer insights into guanxi networks from diverse theoretical perspectives, leading to starkly divergent directional trends in the research on guanxi networks.

The changing significance of guanxi networks

Research conducted in China during the Mao Zedong period (1949–1976) provided a standard understanding about the significance of guanxi networks under state socialism. During this period, state planning created a series of bureaucratic controls over valued resources and opportunities for access in the resulting shortage economy, and consequently guanxi networks of personal connections were mobilized by cadres and ordinary citizens to break free from bureaucratic controls in order to receive favorable treatment in resource allocation under the systems of party clientelism (Walder, 1986), organized dependence (Walder, 1992), and guanxi favoritism (Bian, 1997). When the post-1978 market reforms gradually relaxed some bureaucratic controls and eventually led to an economy of abundance and growing inequalities, did guanxi networks fade away from economic life? Sociologists have offered three different propositions, each based on distinct theoretical logics and empirical evidence. The main debates lie the interplay among guanxi networks, market rationality, and formal institutions.

One proposition is widely known as the “declining significance of guanxi” hypothesis (Guthrie, 1998, 2002). The basic theoretical assumption of this proposition is that there exists a trade-off between irrational guanxi favoritism and market rationality. Guthrie argues that China's market-oriented reforms meant an increase in economic rationality, which is measured by the rise of open competition, price mechanisms, hard budgetary constraints, and meritocracy; when these rational criteria play a decisive role in resource allocation in the process of marketization, the role of irrational guanxi favoritism should decline accordingly. Empirical studies have shown that in the late 1990s, whether in emerging market sectors or reforming state sectors, employers prioritized measures of economic rationality over guanxi favoritism in personnel, financial, and material decision-making (Guthrie, 1998). As for hiring, one qualitative study found that guanxi networks did not play a role in the screening of first-time job applicants in the emerging private sector (Hanser, 2002), and another qualitative study found that the influence of guanxi contacts was “highly limited” and giving way to a criteria of meritocracy in the hiring of college graduates into professional and technical positions (Huang, 2008). A China-based researcher found that the dependence of private entrepreneurs on social networks for hiring gradually decreased during market transformation (Yang and Huang, 2005).

Different from the declining significance of guanxi hypothesis, the “guanxi persistence” proposition stresses the theoretical logic of network embeddedness, which posits that all economic activities and institutions are embedded in the networks of ongoing social relationships (Granovetter, 1985). The logic of network embeddedness holds that social networks supply economic transactions with such intangible resources as trust and cooperation, which promote the circulation and sharing of information, suppress opportunistic behavior, reduce transaction costs and future uncertainty, and thereby improve market efficiency. Likewise, proponents of the guanxi persistence proposition (Zhang and Zhang, 2012) argue that the operation of China's emerging market economy does not merely rely on price mechanisms, among other measures of market rationality, but also on legal systems, social trust, and moral values, each of which is embedded in guanxi networks of trust and cooperation. Empirically, it has been widely found that Chinese legal practice is strongly shaped by elite and non-elite networks of political power and favor exchanges (Liu and Halliday, 2011), that business founding and development are predominately reliant on guanxi networks of high trust and mutual support (Burt and Burzynska, 2017), and that contemporary moral values in socioeconomic life are aligned with Confucian ethnics (Warner, 2014) and guanxi favoritism (Zhou, 2023). In fact, the resilience of guanxi norms is so strong (Yang, 2002) that even Western bankers have had to adjust their business strategies in running their China operations (Nolan, 2011).

The third and final proposition is known as the “increasing significance of guanxi” proposition. Among others, Chang (2011) defines guanxi as a set of network strategies for accessing scarce resources, bridging social relationships of power, and embedding trust and reciprocity. To her, the prevalence of guanxi varies with the degree of institutional uncertainty; the greater the uncertainties in resource allocation, the greater the significance and instrumental utilities of guanxi. Bian and Zhang (2014) take this one step further to argue that China's market transition is characteristic of rising institutional uncertainty and market competition. Their historical analysis reveals that institutional uncertainty has been generated by the three strategies of gradual reform: the strategy of reform without design caused the problem of institutional nontransparency to rise, the strategy of trial-and-error tactics caused the problem of institutional ambiguity to rise, and the strategy of market–non-market institutional coexistence caused the problem of institutional incompatibility to rise. The good news, however, is that these institutional uncertainties paved the road to a market economy and increased market competition from local to national and international scales. Empirically, the joint forces of institutional uncertainty and market competition enlarged the social space of guanxi networks that filled “institutional holes” (Bian, 2002), as evidenced in the growing proportion of jobseekers that are using guanxi contacts to facilitate job searches—rising from 40% in 1978 to 90% by around 2014 (Bian, 2019). In the meantime, other large-scale surveys also show that the importance of guanxi ties has shown sustained growth in job hunting (Tian and Liu, 2018), the impact of guanxi networks is higher in the private sector than in the state sector (Knight and Yueh, 2008), and job-search contacts have lasting effects on the wage incomes of jobseekers in both entry-level positions and in later career development (Bian et al., 2015a), with this effect being stronger in the more marketized eastern region than in the less marketized central and western regions (Lu and Li, 2008).

The changing stock of network social capital

Like all egocentric networks, guanxi networks generate social capital for the individuals within them. Lin (2001b) defines network social capital as the resources that are embedded in and can be mobilized from the structures of egocentric networks. Cross-national research indicates that the higher the degrees of extensity, heterogeneity, and upper reachability of the egocentric networks that people develop, the greater the amount of network social capital generated for them (Lin and Erickson, 2008): the same pattern is observed for China (Bian, 2008b).

However, unlike the declining trends in network social capital observed in many Western countries such as the US, UK, and Australia since the Second World War (Putnam, 2004), the JSNET survey series shows that China has enjoyed an upward trend in network social capital since the start of the post-1978 market reforms (Li and Guo, 2022). While kinship-centered social capital has been stable, social capital generated by work-related social connections has risen significantly over the four decades of market reforms (Bian et al., 2020a). What have been the driving forces behind this trend? Chinese researchers have revealed three structural sources of change: (1) economic marketization, (2) age-cohort variation, and (3) the rise of internet networks.

China's post-1978 market reforms have significantly transformed its economic system from state redistribution to a market–non-market mixed system. This hybrid economy has two all-encompassing structural features that have immediate implications for the rising importance of network social capital in economic life (Bian and Zhang, 2014). The increasing institutional uncertainty, as revealed above, is manifested by the growing problems of information asymmetry and social distrust in economic life, which call for the expanding role of network social capital. The increasing scale and intensity of market competition, on the other hand, requires economic actors to be empowered with comparative advantages in human, financial, and social capitals. Consequently, the joint forces of increasing institutional uncertainty and market competition enlarged the social space of work-related network social capital (Bian et al., 2020a).

Beyond the impact of economic marketization, the changing stock of network social capital varies across age groups and birth cohorts (Li and Guo, 2022). An age-period-cohort model analysis has corrected an earlier estimation bias of the age effect on social capital in previous studies (Li and Li, 2016a), and it shows that the stock of network social capital was relatively high for the pre-1948 cohort, lower for the 1949–1977 cohort, and higher again for the post-1978 cohort (Li and Guo, 2022). This deeply reflects the impact of the macro political-economic environment and social conditions experienced by different age groups and birth cohorts during their lifetimes, especially revealing the lasting imprint of macro political-economic changes on patterns of individual social interactions.

Finally, the changing stock of network social capital has been closely related to the rise of online social networks. Undoubtedly, the advent of the digital age and the formation of virtual spaces are reshaping human communication patterns and socioeconomic life (Lin et al., 2018). In this changing context, researchers are particularly concerned about whether social networks in virtual spaces can be transformed into interpersonal friendships in real spaces. Recent research has shown that not only can increased online interactions and friendship ties be transformed into offline network ties, but some offline interpersonal relationships and network structures (in terms of network heterogeneity and network ceiling) can also be maintained and expanded in online virtual spaces (Bian and Lei, 2017; Bian and Miao, 2019). Another study has shown that while charitable donations mainly originate from strong-tie networks, online networks and emerging virtual spaces can promote charitable donations through weak-tie networks (Lin and Huang, 2017) and generate medical crowdfunding opportunities through both strong-tie and weak-tie networks (Yang et al., 2023).

Looking ahead

The extensive research discussed above strongly supports the growing recognition of the importance of guanxi networks and their impact. However, there are still several unresolved questions and areas for further exploration regarding network dynamics and trends. On the increasing significance of guanxi networks, future research needs to further clarify the structural, institutional, and cultural causes of guanxi dynamics in the long run. Theoretically, a critical question is whether or not there is Weberian elective affinity between guanxi favoritism and market rationality (Bian and Shuai, 2020). Empirically, this requires systematic data monitoring of the entire reform era before an evidence-based answer can be obtained. A related issue is evaluating the changing trends in returns of guanxi networks to outcome variables, such as employment opportunities, wage returns, and job quality measures. Have the comparative advantages generated by guanxi networks stabilized, decreased, or increased? In each scenario, why and how has this happened? These research gaps present opportunities for future investigations into the changing effects of guanxi networks.

On the changing stock of network social capital, there are several issues for further research. The first concerns the validity and reliability of measurement methods. While positional generators have been the main method to measure network social capital, it is an open question whether they capture important dimensions of the concept, such as relationship types, tie strengths, and pathways of network influence. A related issue concerns different life domains that generate different kinds of network social capital, which in turn also function differently. Beyond family and workplace domains, how is network social capital generated through social interactions in social dining, sports, and other domains, and does how each of these generating sources contribute to generalized network social capital on the individual and collective levels? A third issue concerns the contribution of online networking to network social capital. This requires a more rigorous empirical design to sequentially identify, evaluate, and compare guanxi networks between virtual and real spaces.

The causal status of guanxi networks

Is the effect of guanxi networks on any given outcome variable causal? Or is it a statistical correlation that cannot survive a rigorous test of causality? In this section, we will first review the causality dispute in Western network research and then summarize research achievements that demonstrate the causal status of guanxi networks in China studies. These research works explore the theoretical sources of homogeneity in guanxi networks, introduce innovative measurement techniques, and utilize counterfactual frameworks and other methodologies to improve causal inference.

The causality dispute in Western network research

Ever since the inception of social network analysis in the 1960s (Freeman, 2004), sociologists and other social scientists have always believed that social networks causally influence social behavior and attitudes (White, 1981). For example, Granovetter's (1973) “strength of weak ties” theory argues that weak ties of infrequent interaction and low intimacy link people of different social circles and thus are more frequently utilized than strong ties of high trust and closeness for circulating non-redundant information about valued opportunities across group boundaries. Lin (1982) advances this reasoning by arguing that human society is hierarchical and therefore weak ties linking socioeconomic ranks are better able to mobilize information and other social resources (i.e. power, prestige, wealth) than strong ties of equal social status for purposive actions. Through a job-search research program, both Granovetter (1995) and Lin (1999) demonstrate that weak ties matter for job-search outcomes because of the influence of network-mobilized information or social resources. Moreover, social networks among producers not only give rise to markets through mutual cooperation and competition (White, 1981), but in certain contexts networks can even replace markets and hierarchies as a means of resource allocation and circulation (Powell, 1990)

However, Mouw (2003) challenged the causal role social networks are believed to play in the job-search process, arguing that social contacts are theoretically homogeneous in personal attributes and therefore the non-random acquisition of friendships and other social ties ensures “a positive correlation of friends’ income and occupational status even in the absence of a causal effect” (Mouw, 2003: 869). In other words, the observed effect of social contacts on job status and wage income may be largely endogenous to human capital because job seekers with higher education tend to have greater access to high-status social contacts. Mouw's analysis of several US survey datasets showed that the widely claimed network effects on job status and wage income are, after all, spurious. This triggered a round of scholarly debates and empirical testing of causal inference in network research about job-search processes and outcomes.

Lin and Ao (2008) were among the very first to respond to Mouw's critiques. While admitting the validity of Mouw's theoretical analysis about endogeneity, they find his statistical analysis flawed due to the mistaken encoding of key variables of occupational homogeneity. The ideal measurement should be traced back to the respondent's previous work and contact characteristics. Based on this approach, Lin and his associates replicated Mouw's statistical models with correctly coded variables, which recovers the causal role of social contacts in job searches (Lin et al., 2013). Fernandez and Galperin (2014) tackled the problem of endogeneity from the demand side of the labor market. Among nearly 30,000 job applications to a retail bank, they found that for both first-time and repeat applicants, those with referrals have a significantly higher (10 times higher) probability of getting interviews and, especially, job offers than those without referrals. The referral effect was confirmed in counterfactual tests when repeat applicants were reanalyzed through a referral–non-referral comparison. From the supply side of the labor market, McDonald (2015) analyzed the National Longitudinal Surveys and found significant wage returns for network-based job searching over formal job searching, and individuals who were informally recruited into their jobs as non-jobseekers enjoyed much higher wage returns. Among jobseekers, those who simultaneously used both formal and informal channels (i.e. joint channel users) were found to make more job-search attempts and receive higher income offers than either formal or informal channel users, demonstrating the causal role that the social capital of network contacts plays in job-search success above and beyond the role of human capital (Shen and Bian, 2018).

Endogeneity and causal effects of guanxi networks

Since 2010, Chinese scholars have approached the problem of endogeneity of guanxi networks with three sets of research achievements. The first explores the sources of homogeneity in guanxi networks. Unlike Mouw (2003, 2006), who attributes network homogeneity to actor preferences, Liang (2010) found that Chinese guanxi networkers initially construct a large number of heterogeneous, rather than homogeneous, alters for favor exchanges at social events of personal importance. However, because of the relatively high costs of maintaining heterogeneous relationships, heterogenous alters tend to have a shorter survival cycle in egocentric networks, leading to the pattern of network homogeneity observed in survey data. Yet, heterogeneous relationships established in past events can be “woken up” or remobilized when needed (Zhang and Liu, 2008). This insight has strong implications for design improvements in the study of guanxi networks.

The second is a methodological guide to systematically dealing with endogeneity problems in guanxi networks. Chen and Fan (2011) summarize four biases, each leading to a different kind of endogeneity problem in statistical analysis of guanxi networks: (1) unobserved variable bias, which affects model estimations; (2) self-selection bias, which leads to network homogeneity; (3) sample selection bias, which confines dependent variables to a non-random sample; and (4) simultaneity bias, which leads to two-way causality. To accurately address these endogenous biases, Chen and Fan provide a useful discussion of various statistical methods, including fixed effects models, propensity score matching, Heckman selection models, simultaneous equation models, instrumental variables, and natural experimental methods. Chen and his associates have used some of these statistical techniques to successfully demonstrate the causal role guanxi networks play in labor markets (Chen et al., 2013, 2014; Chen and Fan, 2011; Chen and Volker, 2016).

The third set of achievements comes from the JSNET project, whose researchers use measurement innovations and counterfactual models to uncover underlying causal mechanisms whereby guanxi networks affect, directly and indirectly, labor market processes and outcomes. These achievements can be summarized into four theories. First, network resource theory distinguishes between information mechanisms and influence mechanisms, and innovative measures of these mechanisms allow for rigorous testing of the causal role that guanxi networks play in job allocation and wage attainment (Bian et al., 2015a; Guo, 2017). Next, network selection theory maintains that guanxi users are highly heterogeneous among themselves (Wu, 2011), and it is found that job seekers with the propensity to use fewer high-profile guanxi contacts have higher income returns (Guo, 2019). Third, network activation theory argues that a rational jobseeker follows game-theoretic logic and that choosing the right timing to activate network contacts and resources is consequential for job and wage offers (Bian, 2022; Yang et al., 2022). Finally, network mobilization theory points to specific strategies of network mobilization, and it is found that the strategies of mobilizing multiple rather than single contacts, mobilizing strong or weak ties, and mobilizing network high- and low-profit contacts matter for heterogenous returns on network mobilization efforts (Li and Li, 2016b, 2019).

Looking ahead

Post-2010 Chinese research achievements demonstrate that guanxi networks are both homogeneous (in terms of alter attributes) and heterogeneous (in terms of mobilized resources). Theoretically, future research needs to specify the configuration of guanxi networks to explain the extent to which heterogeneous resources can be mobilized from the networks of seemingly similar alters. Whereas the theory of “structural holes” (Burt, 1995) directs research attention to structural differences and behavioral implications between dense and sparse networks, the theory of guanxi habitus (Bian and Miao, 2020) points to the direction of studying how Confucian ethics of sentiment and obligation interplay with each other in social interactions under the resilient Chinese guanxi culture.

Empirically, the analysis of causal inference in guanxi networks must be strengthened. To tackle the causal status of guanxi networks, one research task is to continue to explore innovative measurement methods to allow for rigorous statistical analysis through counterfactual models. In light of emerging methods of data science, such as data crawlers, machine learning, and social experiments, future researchers also need to utilize some of these new techniques in order to meet the goal of causal analysis of guanxi networks. Importantly, none of these techniques will matter much should the research into guanxi networks be decoupled from substantive issues. Social change is rapid in China, as well as around the world, and causal analysis of social networks, including guanxi networks, must help uncover, understand, and resolve the emerging real problems facing humanity.

Positive and negative consequences of guanxi networks

Informal networks are positively and negatively consequential, and guanxi networks are no exception. In the past decade, research has explored the bright sides of guanxi networks in areas such as educational attainment, jobs and careers, entrepreneurship, collective action, and subjective well-being. It has also delved into the dark sides of guanxi networks, including economic costs and risks, political corruption and injustice, social contagion and exclusion, as well as associated social costs. Here we summarize the main findings of empirical analysis about the positive and negative implications of guanxi networks.

Positive consequences of guanxi networks

The positive consequences of guanxi networks refer to the use of guanxi contacts to generate goal-attainment results through mobilizations of tangible and intangible resources (Li et al., 2023). China-based scholars have uncovered several important sets of positive consequences of guanxi networks in studies of different topics.

Educational attainment

Influenced by Bourdieu (1986) and Coleman (1988), this topic focuses on educational effects of two types of guanxi networks: peer networks and parental networks. Peer networks, such as those between university classmates and dormmates, are found to increase academic performance through mechanisms of peer influence in educational values, study methods, and problem-solving tactics (Cheng, 2017) and to contribute to maintaining healthy and forward-looking mentalities (Zhang, 2023). Parental networks, on the other hand, promote children's academic achievement through sharing information and advice about children's school engagement (Li and Zhang, 2022) and extracurricular opportunities (Zhao and Hong, 2012). While these studies shed light on the network mechanisms at play, they raise the question of how these mechanisms work for students at different educational levels and across various school contexts.

Jobs and careers

Since the 1990s, job acquisition has been the most studied topic in the research on Chinese guanxi networks (Bian, 1994, 1997; Bian and Huang, 2009; Lin and Bian, 1991), and three new advancements have emerged in post-2010 studies. One is a measurement innovation in distinguishing network-mobilized resources between information and favoritism, which allows for a rigorous empirical analysis of the relative efficacies of strong and weak ties in China's transition economy (Bian et al., 2012; Bian et al., 2015a; Zhang and Guo, 2011). Another is the discovery of several causal mechanisms that explain how guanxi networks matter for job-search success: guanxi networks are found to reduce recruitment and job-search costs (Zhang and Hao, 2013), improve risk resilience (Sun and Bian, 2017), and increase returns on guanxi investments (Bian and Sun, 2019). A third achievement is a macro–micro interactive analysis, which shows that guanxi networks are most relevant and effective when local and national economies on the whole increase market competition and institutional uncertainties (Bian and Zhang, 2013; Cheng and Bian, 2014; Liang, 2012). On the topic of career promotion, one's bureaucratic connections in the workplace increase one's probability of obtaining promotion for administrative roles, and one's market connections built up around one's career increase one's probability of obtaining promotion in professional/technical roles (Wang and Bian, 2022).

Entrepreneurship

It is well known that guanxi networks are the cornerstone of entrepreneurial success in China's transition economy (Bian, 2006; Bian and Qiu, 2020). This is why self-employed entrepreneurs are significantly more active than salaried employees in making guanxi investments, as evidenced in Spring Festival visitation networks and social eating/drinking networks (Zou and Ao, 2011). In a random sample of 700 entrepreneurs in the Yangtze River Delta, Burt and associates found that a great majority of the guanxi contacts obtained at the time of a business's founding continue to be helpful resources throughout that business's development (Burt et al., 2018), and this is true even after a family business has developed into a sizable corporation (Burt et al., 2021). Some successful corporate managers decide to leave career jobs when personal networks are conducive to starting their own businesses (Ma, 2015a). Whether small or large, private businesses perform better when they have developed institution-crossing networks than if their networks are concentrated in either state or non-state sectors (Wang et al., 2021; Wang and Zhao, 2012). Nonetheless, while structural-holes networks are found to improve corporate performance in family businesses, they have no significant effect in “modernized” enterprises where kinship networks are not dominant (Nee et al., 2017).

Collective action

Collective action is voluntary in democratic societies, in which informal networks of participants are an organizing mechanism of collective action (McAdam and Paulsen, 1993). In China, a party-state under the auspices of the Communist Party of China, guanxi networks are also active in mobilizing and coordinating different forms of collective action. For example, clan networks are found to mobilize peasants to participate in village head elections (Kennedy et al., 2018). Villagers as well as urbanites with more local network connections have higher likelihoods of participating in grassroots elections than their counterparts with fewer local network connections, and local networks also pave the way for them to participate in formal associations, which in turn increase their election participation (Hao and Ke, 2024). When clan networks are constrained by their informal status, officially sponsored associations of the elderly are found to effectively promote the success of collective resistance (Lu and Tao, 2017). Finally, networks of migrant workers can effectively spread information, knowledge, and outlooks, thereby stimulating the cognitive and social foundations of migrant communities and shaping the scale, form, and success rate of collective action among migrants (Lu, 2019).

Subjective well-being

Guanxi networks not only have instrumental effects, but also emotional or expressive effects on the subjective well-being of individuals. Subjective well-being research has been an important growth point in China's social network analysis in the past decade, and research topics have bloomed. Scholars focus on self-evaluations of health, happiness perception, fairness, social trust, and life satisfaction, as well as a series of empirical studies on issues such as emotional support (Bian et al., 2015b; Guo and Wang, 2019; Li and Li, 2019; Song, 2015). Despite theoretical debates about the underlying mechanisms through which guanxi networks promote subjective well-being (Song, 2015), the results of empirical studies on youths and migrant workers in China have validated the social support theory (Guo et al., 2020; Guo and Wang, 2019; Yue and Wang, 2022).

Emerging issues of sociological significance

The positive effects of guanxi networks have been analyzed in research programs on emerging issues. One emerging issue is inter-ethnic networks. Researchers have analyzed rates and patterns of inter-ethnic communication, exchange, and integration among selected ethnic populations, and they have found that inter-ethnic social capital significantly shortens the psychological distance between different ethnic groups and enhances social integration, generalized trust, and national identity (Guo and Li, 2022; Li, 2020, 2022). Another emerging issue is China's strategic goal of building a strong sporting country, and sport-related social capital has been found to be an important mechanism to help realize this objective (Bian, 2020b; Bian and Lu, 2022), by means of enhancing individuals’ physical and mental health (Liang and Jia, 2022; Liang and Zhang, 2020). A third emerging issue is epidemic-specific social capital. In the context of the COVID-19 pandemic, Chinese sociologists have found two forms of social capital that helped people cope with the pandemic: family-centered and neighborhood-based core networks that provided care, comfort, assistance, and companionship during community lockdowns, and online networks of confidants, friends, colleagues, and acquaintances that spread useful information for epidemic prevention (Bian, 2020a; Bian et al., 2020b). These forms of social capital were also found to significantly contribute to maintaining a positive mindset during the pandemic (Bian et al., 2021).

Negative consequences of guanxi networks

Social networks are negatively consequential when they hinder individuals’ goal attainment, erode organizational operations, and undermine social justice (Horak et al., 2020; Li et al., 2023; Song et al., 2021; Wang and Uzzi, 2022). In the past decade, researchers explored the instrumental and expressive negative consequences of guanxi networks.

Economic costs and risks

Guanxi networks may lead to unexpected consequences and negative effects in the internal management of organizations and external market operations. For example, guanxi networks may increase implicit contract costs between organizations due to maintaining relationships and rewarding favors. The internal guanxi circle of an organization may lead to tension in relationships within and outside the circle, and the internal allocation of an organization may also replace performance-based principles with relativism (Horak et al., 2020). These modes of action undoubtedly increase enterprise costs and constrain organizational efficiency. In the internal management of enterprises, relationship circles have embedded roles in processes such as recruitment, promotion, and dismissal. When the relationship circle is separated from the formal hierarchical structure of the organization, it increases the risk of collective resignation (Shuai, 2019). In the operation of external financial markets, it has been found that although high social capital helps to form financing networks, it unintentionally exacerbates the risk contagion between enterprises (Wu et al., 2011).

Political corruption and injustice

Guanxi networks may also result in unintended social or political consequences such as factions, nepotism, bribery, collusion, and the breeding of corruption. For instance, guanxi networks may give rise to guanxi protection, interest alliances, implicit oligarchies, and so on (Horak et al., 2020), resulting in an unfair distribution of social resources, increased corruption and bribery, and erosion of social justice. The negative effects of guanxi networks include rent-seeking behavior, power abuse, nepotism, favoritism, and corruption and bribery at the individual level, as well as social inequality, network capitalism, oligarchic decision-making, and persistent inequality at the societal level (Barbalet, 2018; Horak et al., 2020; Karhunen et al., 2018). Empirical research from China has summarized three forms of networks where corruption thrives: nepotism based on open networks, bribery based on closed networks, and extortion based on negative networks (Barbalet, 2018). These forms of corruption are rooted in different network structures and have been found to have deeply shaped the economic growth models in Russia and elsewhere (Karhunen et al., 2018).

Social contagion and exclusion

Guanxi networks also exert negative emotional consequences. Recent research has identified three theoretical mechanisms governing the negative effects of guanxi networks: social contagion, social exclusion, and social anomie (Li et al., 2023; Rook, 2015; Song et al., 2021). The social contagion mechanism is one in which guanxi networks cause psychological pressure by spreading harmful information, illicit attitudes, and immoral behaviors (Christakis and Fowler, 2012; Zhang and Centola, 2019). The social exclusion mechanism emphasizes that a prominent feature of guanxi networks is the exclusion of outsiders, which may lead to empathy failure and social differentiation from members outside the network circle (Allan and Phillipson, 2017; Li et al., 2023). Finally, the social anomie mechanism suggests that in real life, some guanxi networks may deviate from their intended purposes, leading to the formation of illicit networks such as those involved in bribery, drug abuse, and other criminal activities (Horak et al., 2020). Recent empirical evidence shows that the effects of guanxi networks on the perception of social isolation follows an inverted U-shaped curve, that is, moderate social interactions reduce the perception of isolation, while excessive social interactions increase individuals’ perception of social isolation (Li, 2023). This is because the increased heterogeneity in network structure and the increase in distant social interactions lead to increased tension and conflicts, resulting in higher social costs and perceived social isolation.

Social costs and consequences

To deal with the negative consequences of guanxi networks reviewed above, a newly developed social cost model suggests several cost categories of guanxi networks, such as hidden social interaction conflicts, asymmetric social exchanges, unfavorable social control, and potential social comparisons (Song et al., 2021). The social cost model utilizes a tripartite typology (structural forms, structural composition, and contents) to categorize relationship and network properties, offering an alternative perspective to the social resource model. Empirical evidence about subjective well-being and political trust supports the social cost model of guanxi networks. Despite the fact that guanxi networks enhance subjective well-being and trust, they can also have a profoundly detrimental impact on subjective well-being through spreading rumors and conveying negative emotions (Li et al., 2023). As a consequence, guanxi network pose a substantial threat to individuals’ mental and emotional well-being, while simultaneously eroding social trust and undermining the cohesion of communities. Chen and Bian (2015) also demonstrated that social eating networks have a negative effect on political trust, and that the process of accumulating, maintaining, and mobilizing guanxi capital through social eating is often accompanied by an erosive effect on political trust. It is worth noting that informal guanxi networks erode political trust by providing channels for the transmission of political stimuli to the general public, while formal networks promote political trust through assisting organizations in advancing their agendas and interests, leading to the assimilation of political trust (Chen et al., 2020).

Looking ahead

While we have obtained a good deal of empirical knowledge about the positive consequences of guanxi networks, future research needs to further explore the underlying mechanisms through which guanxi networks generate these positive effects. Existing research has identified three mechanisms for positive effects: social support, social companionship, and social control (Li et al., 2023; Rook, 2015). Are there other mechanisms? We suggest looking into three potential areas: social identity mechanisms (how the shared identity of network members increases their willingness to transfer resources), social learning mechanisms (how network members learn from one another), and social domination mechanisms (how network centrality diffuses the influence of power). We further suggest that the exploration of these additional causal mechanisms must be conducted by paying attention to group heterogeneities (such as gender, ethnicity, and class) as well as contextual variations (along cultural, institutional, organizational, and cross-national boundaries). For example, the positive effects of guanxi networks may diverge between Confucian and Christian cultures: whereas Confucian culture encourages primordial particularism, personalized connections, and deliberate network building (Fei, 1947), Christian culture legitimizes generalized particularism, group-oriented favoritism, and the unintended consequences of personal networks (Arrow, 1998). Accordingly, what are considered to be “positive” consequences of social networks vary across cultural contexts, and a theoretical model for an America–China comparison can be seen elsewhere (DiTomaso and Bian, 2018).

As compared to the research achievements on the positive consequences of guanxi networks, post-2010 research development has only begun to explore the negative effects of guanxi networks. Indeed, we have not seen a systematic effort looking into various manifestations of the dark side of guanxi networks. There is also the need to identify the underlying mechanisms that make sense of how guanxi networks generate negative consequences, beyond what we already know about the mechanisms of social contagion, social exclusion, and social anomie (Song et al., 2021). If network embeddedness varies from “shallow embedding” to “adequate embedding” to “deep embedding” (Uzzi, 1999), then future researchers should study the unintended consequences of different degrees of network embeddedness in Chinese guanxi culture. Unlike positive effects, the negative effects of guanxi networks are hidden, elusive, tacit, illicit, or illegal, thus requiring skillful research designs to uncover them. In theory, we can capture negative effects from at least three vantage points of how they are generated, such as source, process, and manifestation, and each of these requires adequate measurement methods to capture them. Finally, future research needs to pay special attention to avoiding negative network effects through social intervention, the empirical results of which have the potential to contribute to policy evaluation and social progress.

Guanxi networks and social stratification

Guanxi networks provide an important theoretical perspective for understanding the origins of social inequality. From this perspective, the process by which network social capital shapes inequality arises in two stages: deficits of capital and deficits of capital return (Li, 2022; Lin, 2001b). Deficits of capital emphasizes a lower stock of network social capital possessed by group A than by group B, and this inequality can be attributed to differences in the investment or accumulation process of network social capital by different groups (Lin, 2001b). Deficits of return, on the other hand, emphasizes a lower market return to a unit of network social capital for group A than for group B, and this inequality is caused by both group differences in social capital mobilization and market discrimination in returns to different groups (Lin, 2001b). In the past decade, China-based scholars have examined both forms of social inequality caused by network social capital in the research areas of gender, ethnicity, and rural–urban inequalities.

Gender inequality

Gender inequality is one of the fundamental dimensions of inequality in human society. Recent research has examined gender gaps in the stock of network social capital and gender gaps in market returns to network social capital.

Gender gaps in the stock of network social capital

The JSNET data reveal that working women often face a deficit in social capital due to gender homogeneity in their social networks, and women who rely on strong ties to secure employment often find themselves entering female-dominated occupations, reflecting the combined effects of gendered occupational segregation and homogeneous networks (Tong, 2012). Moreover, while social capital can lead to significant increases in earnings, females tend to have lower levels of or lower quality social capital compared to males, generating wage disadvantages for women (Cheng et al., 2015). Finally, compared to men, women have fewer social connections with domestic and international academics, which explains women's lower academic performance compared to men (Zhu and He, 2016).

Gender gaps in returns to network social capital

Job-search analysis reveals that the lower capital returns for women partially reflect demand-side gender bias or discrimination in the hiring process, particularly in the private sector rather than the public sector (Tian and Liu, 2018). In terms of wage returns, women experience significantly lower returns of social capital compared to men, thus contributing to the wage disadvantage faced by women (Cheng et al., 2015). Moreover, in both market-dominated and government-dominated sectors men have higher returns from their network ties compared to women, and the gender gap in network returns widens with increased marketization (Chen and Zhang, 2015). In the field of science research, when male and female scientists have similar network ties, women have significantly lower research outputs (Zhu and He, 2016). Similarly, even when male and female entrepreneurs have the same network structures, their network returns differ significantly in men's favor (Burt et al., 2021). Finally, the role of guanxi networks on gender inequality is profoundly shaped by the institutional environment and cultural beliefs. Studies have found that gender gaps in job-search success based on networks are greater in market-dominated sectors than in government-dominated sectors, as gender-bias norms have greater legitimacy in the market sector (Tian and Liu, 2018).

Ethnic inequality

Previous studies have examined the sources of ethnic stratification in education, employment, and segregation (Gustafsson et al., 2020; Ma, 2015b; Ma, 2016; Wu and Song, 2014). Post-2010 studies have investigated ethnic stratification from social network perspectives.

Ethnic gaps in the stock of network social capital

Empirical evidence shows that the stock of network social capital among ethnic minorities is significantly lower than that of the Han Chinese (Li, 2022; Li and Li, 2016a; Wang, 2023). Existing research suggests that ethnic minorities primarily face language barriers that prevent them from developing inter-ethnic connections (Guo and Li, 2022). Once these barriers are removed in cross-ethnic interactions, it can significantly optimize the network structure and improve social capital for ethnic minority groups (Li, 2020). Taken together, guanxi networks that cross ethnic boundaries not only enhance social trust and ethnic tolerance (Guo and Li, 2022) but also serve as an important way to diminish wage disparities by ethnicity (Li, 2022).

Ethnic gaps in returns to network social capital

There is no significant ethnic difference in labor market returns to social capital (Li, 2022). However, the effects of social capital on ethnic stratification are conditional upon employment sectors and geographic concentration: while ethnic concentration does not impact wage returns of social capital in the state sector, it increases wage returns of social capital in the market sector but diminishes wage return of social capital in the agricultural sector (Li, 2022). These findings underscore the complex interplay of government power, market forces, and social networks in shaping ethnic stratification in China. It appears that government power limits the role of guanxi networks through protective policies for ethnic minorities in the state sector, but market forces facilitate the role of guanxi networks in ethnic inequality in non-state sectors.

Urban–rural inequality

The urban–rural inequality resulting from China's household registration (hukou) system is an important topic of concern for sociologists. The hukou system has led to severe disparities in educational attainment, employment opportunities, wage levels, welfare benefits, and health conditions between urban and rural populations (Li, 2016; Li and Li, 2007; Wang, 2017; Wu, 2013). Post-2010 researchers have identified hukou gaps in both the stock and returns of network social capital.

Hukou gaps in the stock of network social capital

Compared to urban workers, rural migrant workers exhibit stronger homogeneity in guanxi networks and lower levels of social capital, making it difficult for them to access higher-wage occupational positions (Cheng and Bian, 2014). Additionally, rural migrant workers have a much lower likelihood of establishing social connections with local urban residents, and the average occupational prestige of network members is significantly lower than that of urban residents (Lu et al., 2013). This can be attributed to the longstanding institutional exclusion, social exclusion, and discrimination that rural migrant workers have faced due to the hukou system, which severely limits their accumulation and mobilization of social capital. In addition to rural migrant workers, those who transition from rural to urban hukou also suffer challenges in social connections. For example, research on marriages between rural and urban hukou holders reveals that while these unions can enhance the social connections of individuals transitioning from rural to urban hukou status with urban managers or professionals, they do not significantly facilitate guanxi networks with urban relatives and friends (Lu and Bian, 2018). This highlights the persistent social segregation resulting from the hukou system, which creates enduring barriers that are challenging to overcome in terms of social connections and integration.

Hukou gaps in returns to network social capital

Capital return analysis indicates that both general social capital and localized guanxi network resources generate lower wage returns for rural migrant workers compared to urban workers (Lu et al., 2013). One important reason is that rural migrant workers have limited bargaining power within the labor market, leading them to be more likely to enter lower-wage occupations. Even when they are in the same occupation, rural migrants tend to work in employment sectors and industries with lower wages (Cheng and Bian, 2014). Additionally, case studies on wage arrears suggest that the strong relationships between labor contractors and construction workers, to some extent, suppress collective action (Wei and Chan, 2022). Overall, the deficits in returns to social capital exacerbate economic inequality between rural migrant workers and urban residents in China, serving as micro mechanisms that perpetuate and solidify rural–urban inequality.

Looking ahead

The aforementioned studies provide valuable insights into the sources of social inequality by gender, ethnicity, and rural–urban division in China from a social network perspective. However, there remains a need for further comprehensive and nuanced research within each of these research lines. In terms of gender inequality, future studies could explore how the role of social networks in perpetuating gender inequality is shaped by supply-side and demand-side factors such as tertiary education expansion, occupational gender segregation, changing gender norms, and economic institutional transitions. In examining ethnic inequality, it is crucial to further differentiate the structural characteristics and social resources embedded within cross-ethnic networks. Developing innovative measurement methods is necessary to capture the distinctive characteristics and effects of inter-ethnic guanxi networks across different ethnic culture circles in China. Regarding rural–urban inequalities, although the distinction between rural and urban hukou has been officially abolished in Chinese society, the impacts of the hukou system still persist and may not disappear immediately. We suggest that future research should examine the life chances of second-generation migrants from rural to urban areas from a social network perspective. The second-generation migrants may face institutional barriers in accessing educational and occupational opportunities in urban areas. Thus, it would be worthwhile investigating whether social networks can help overcome these institutional barriers and provide better educational opportunities and occupational positions for second-generation migrants.

Beyond gender, ethnicity, and rural–urban inequality, future research on guanxi networks and social stratification could be expanded in several areas. First, there is a need to delve into the intermediate mechanisms through which social networks influence social stratification. While existing research has examined the relationship between social networks and stratification in terms of capital deficits and return deficits, less attention has been given to the structural characteristics of social networks. For example, exploring differences in network density, heterogeneity, and reachability across different social groups and identifying which structural indicators play the most significant role will contribute to a deeper understanding of social stratification from a social network perspective. Second, we suggest that future research investigates the contributions of social networks to the dynamics (or trends) of social inequality. The literature reviewed in this discussion has primarily focused on cross-sectional data, examining the static effects of social networks on social stratification. However, both social stratification and social networks are dynamic phenomena. Understanding how changes in social networks influence changes in social inequality/stratification over time requires a clear understanding of the mechanisms through which social networks affect trends. Lastly, we suggest that future research explores practical pathways to reduce social stratification from a social network perspective. Despite extensive research on social inequality, persistent levels of inequality remain. It is crucial to draw lessons from the study of social inequality and identify effective strategies to reduce stratification. To this end, it will be necessary for social science research to explore practical approaches to reducing social inequality from a social network perspective.

A concluding remark

Guanxi networks continue to provide an excellent vantage point from which to understand Chinese society. This article has reviewed post-2010 guanxi-related scholarship in terms of conceptual controversies, changing trends, causal inferences, positive and negative effects, and patterns of relationally altered social stratification. Taken together, these studies shed light on a broad conclusion: the guanxi ethics that are deeply rooted in Chinese culture have a persistent influence in socioeconomic and political life in the process of Chinese-style modernization (Bian and Ma, 2023). To be sure, an adequate understanding of guanxi networks not only contributes to our knowledge of the enduring features of Chinese civilization but also serves as an important foundation for social order (Yang and Xiang, 2022) and social thought (Zhou, 2023) in China today. Thus, scholars of contemporary Chinese society should take guanxi networks seriously and theoretical innovations and empirical developments in the sociology of guanxi networks in the future should be expected.