Networks in general and Chinese guanxi networks in particular are widely understood to consist of dyadic relations founded on trust. The justifications typically offered for these assessments are subjected to critical scrutiny in this article and alternative conceptualizations are provided in the following account.

The centrality of trust to consideration of networks in general is indicated in the first part of the discussion below through a brief treatment of the migration of “trust” from sociological characterizations of markets to institutional characterizations of networks. The second part of the article goes from networks in general to guanxi networks in particular by providing a critical examination of the operation of the latter in terms of trust relations. This is developed in the third part in which the Chinese sense of “trustworthiness” is outlined. Against the grain of the consensus in the prevailing literature, it is shown that guanxi networks cannot be based on trust as the term is normally understood. The following section proposes an alternative basis of network functioning, through assurance rather than trust. This development leads to a reformulation of the structure of guanxi networks, from dyadic to triadic, as public visibility of guanxi exchanges introduces a means of third-party sanction of transgression against guanxi norms, ensuring network maintenance in the face of defection. Finally, changes in the form of guanxi, from network relations based on strong-ties to weaker-ties subsequent to developments in China’s transitional economy, raises questions concerning changes in the means of sanction through application of different media of communication in affecting the reputation, and therefore the face (mianzi), of wrongdoers in provision of assurance to network members.

The mechanism of loss of face in preserving guanxi networks, from gossip to disclosures in the mass media, is seldom discussed in accounts of guanxi. But this becomes of central importance when the structure of guanxi relations is understood to be not fully accounted for in its atomic dyadic form but rather in the molecular triadic form in which third-party assurance is crucial and also when the current changes in the form of guanxi, led by widespread labor migration and market transition, are appreciated.

The location of trust: from markets to networks

The received wisdom in sociology is that markets presuppose trust. Durkheim’s statement concerning pre-contractual trust (Durkheim 1964, pp. 200–229) and Simmel’s claim regarding the necessity for social integration of inter-personal trust (Simmel 1978, pp. 178–179) provide a vocabulary for analysis concerning market behavior and institutions, and their foundations. Granovetter’s account of the relationship between embeddedness and trust (Granovetter 1985, pp. 487–493), while not directly consonant with these earlier statements, supplements them in being concerned with the production of trust in social networks (Granovetter 1985, p. 491). From the perspective of economic analysis, on the other hand, the role of trust in market relations is seen to be less decisive than it is in sociology. Market transactions, in Arrow’s terms, may be “lubricated” with trust (Arrow 1974, p. 23). Trust facilitates exchange in light of the fact that markets are ambiguous arenas of activity in which independent actors operating not only through self-interest but possibly with “opportunism” and “guile” necessarily exercise distrust (“buyer beware”) in relation with others, and in which there is no necessary expectation of future commitment.

In a detailed account, in which distinctions within the generic category of trust are drawn, Williamson (1993) argues that “calculativeness” rather than trust more accurately indicates the animus and character of the facilitating disposition underlying market exchanges. While sociological disquiet with Williamson’s arguments abound, a slight of hand in institutional analysis has resulted in the formation of a perspective in which trust becomes of secondary relevance in understanding the basis of market behavior although primary in understanding networks. Rather than trust standing behind contract, as with Durkheim, property rights are typically held to provide institutional support to market transactions, and third party (legal) enforcement of these rights effectively render trust redundant as an underpinning element of them. In an analysis of the historical evolution of institutions in medieval trade Greif (2006) contrasts developments among Genoese traders and those of the middle-eastern Maghribi. He shows that the bi-lateral trading patterns of the Genoese were generative of legally tempered property rights prefiguring modern western market institutions, whereas the practices of the Maghribi traders evolved into an alternative pattern in which “agency relations were characterized by trust … [and] not based on legal contracts” (Greif 2006, p. 63; see also pp. 86–87).

The different historical trajectories indicated by Greif points to a widely accepted institutional distinction between markets and networks, respectively grounded in legally-enforced property rights and trust relations. A representative statement succinctly summarizes the convention:

In market transactions the benefits to be exchanged are clearly specified, no trust is required, and agreements are bolstered by the power of legal sanction. Network forms of exchange, however, entail indefinite, sequential transactions within the context of a general pattern of interaction. Sanctions are typically normative rather than legal (Powell 1990, p. 301).

The normative sanctions indicated here are predominantly regarded as those of trust. Indeed, it is widely accepted that there is a likelihood of network patterns emerging when exchanges are based on long-term and continuous reciprocity in which reputational profiles are evident and commonality of background of participants (ethnic, geographic, ideological, or professional) is established (Powell 1990, p. 326). This last point seems to reinforce the other two Powell mentions in the quotation above because the “more homogenous the group, the greater the trust, hence the easier it is to sustain network like arrangements” (Powell 1990, p. 326; see also Molm, Takahashi, and Peterson 2000; Wellman and Wortley 1990).

The displacement of trust from markets and its relocation in networks is reflected in the conventional conceptualization of social capital in terms of trust and the correlative assumption that networks are expressions of or a mechanism in the formation of social capital (Coleman 1990; Fukuyama 1995; Putnam 1993). The assumption that trust is the basis of networks and social capital is central in the discussion of guanxi networks as forms of social capital. Indeed, the contrast of guanxi as trust-based, in contradistinction to relations based on legal contract, and the transaction cost benefit (including the containment of fraud) of the former over the latter are frequently mentioned in the literature (Boisot and Child 1996, pp. 612–613, 617, 619, 625; Lee and Dawes 2005, pp. 48–52; Lovett, Simmons, and Kali 1999, p. 243; Tong and Yong 1998, p. 84; Tsang 1998, p. 66; Whyte 1996, p. 12; Yeung and Tung 1996, p. 63). Not only is there an almost universal assumption that guanxi is based on trust but also an acceptance that trust is generated in guanxi exchanges (Smart 1993, 400, 403; Lee and Dawes 2005). We shall see that guanxi networks are not regarded as unique in these regards, for they constitute “traits of social exchanges in general … [in which] a sentimental basis [can be located] and obligations are very much in line with the theory of trust” (Lin 2001a, p. 159). Indeed, it is an ironic feature of the discussion of guanxi that while it is frequently regarded as an indigenous Chinese cultural form it is typically explicated in terms, such as trust, that are themselves arguably indigenous European concepts associated with a political tradition (Silver 1985) that has never had meaningful institutional expression in China, as indicated below.

Guanxi and trust

Guanxi literally means “relationship” or “personal connection” although it cannot be understood in those terms alone. China is often described as having a collectivist culture, but this is a mistake; rather, Chinese culture is best described as relational and the effective relations are between roles based on familial ties, with non-family ties taking a familial form. The pervasive relationships constitutive of Chinese society have been summarized by Chang (2010) as a number of different types of reciprocities (wanglai) founded on or distinguished by different particular criteria (lishang). The term guanxi, in this account, is reserved to indicate only a narrow band of gift-exchanges, described as instrumental wanglai. This is a useful caution against seeing guanxi in all personal connections and exchanges. “Instrumental” in this context implies not only orientation to utilities but also social bonding, which comes from mutual assistance or dependence through which a sense of moral obligation is experienced. Guanxi relations, which are affective and obligatory, always involve provision of a favor for which the benefits are compound, including immediate and also potential or future elements for both the favor seeker and provider that are substantive as well as reputational. Guanxi relations are cultivated and, as they involve reciprocity, obligation and affectively charged moral assessments and commitments, endure as a network of structured engagements. The costs of membership in guanxi networks are not negligible but typically regarded as acceptable in terms of the available advantages: “The Chinese are willing to sacrifice short-term interests for long-term favor exchanges, since they know that the benefit of group effort will be much greater than that of individual endeavor” (Luo and Yeh 2012, p. 65).

Guanxi, then, can be understood as a form of asymmetrical exchange of favors between persons on the basis of enduring sentimental ties (renqing) in which enhancement of public reputation or face (mianzi) is the aspirational outcome. Guanxi networks (guanxiwang) are widely recognized as a fundamental element of traditional Chinese social structure that continues to be pervasive in China and Chinese cultural areas. The discussion of guanxi in the sociological literature has been focused on the role of guanxi in the transitional economy of mainland China. Two alternate positions can be identified; one holds that guanxi is integral to Chinese culture, including its Confucian heritage (Fei 1992; Hwang 1987; King 1991), another argues that guanxi is best understood institutionally as a component of social organization characterized by distributional disarticulations and an underdeveloped rights-and-law framework (Guthrie 1998, 1999; Wank 2002).

Given the significance of informal networks in all modern economies, it is unlikely that economic and legal development in China would lead to a total demise of guanxi (Chen and Easterby-Smith 2008; Hammond and Glenn 2004), although as the broader institutional landscape changes so the form and application of guanxi will change, moving from strong-tie to weaker-tie relations, as the range of available contacts expands, and from expressive to more instrumental purposes, about which more will be said in what follows. In recognizing a dynamic element in the incidence of guanxi, a more recent discussion has identified variation in the form of guanxi commensurate with changes in the societal circumstances in which it is manifest, including difference among an “accessing form” of guanxi, a “bridging form,” and an “embedding form” (Chang 2011), as well as favor-seeking and rent-seeking guanxi (Qi 2013). Discussion of variation through differential application of the form of guanxi leaves unexplored, however, another aspect of the structure of guanxi that is referred to here as the difference between guanxi’s atomic and molecular structures. This latter distinction brings into question the role of trust in guanxi, as we shall see. Practically all discussions of guanxi mention trust as both a mechanism and an outcome of guanxi relations. While there is a persistent use of the term “trust” in these sources, there is seldom found a clear appreciation of the complexity of trust as a property of social relationships nor is there a clear theoretical appraisal of the issues involved in assuming its role in guanxi. Before examining the concept of trust itself, it is first necessary to consider how the notion of trust is integrated into accounts of guanxi.

As most discussions of guanxi assume underlying strong ties there is a corresponding inference that trust plays a formative role in the incidence and development of guanxi relations. The assumption, mentioned above, that the “more homogenous the group, the greater the trust” (Powell 1990, p. 326) is frequently given expression in discussions of guanxi: “The closer a relationship is, the stronger the sense of trust and the feelings of affection are between members” (Chan 2009, p. 718). Family and kinship bonds in particular are assumed to entail affective closeness and high levels of trust: “The family- and kin-centered character of Chinese social relations … gives ascribed social circles and fictive kin friends the privilege of sharing an extraordinary level of trust and affection” (Chan 2009, p. 718). In fact, significant ruptures in relations between kin are not uncommon in China. At the same time the pervasive power of family obligation and the structural priority of family in social relations means that friendship and similar non-kin relations are experienced as kin-analogues, as fictive kinship, and subsumed in kin terminology (Baker 1979, pp. 162–167; Fei 1962, pp. 83–94). Indeed, the closest affection may be between friends rather than kin (Yan 1996, pp. 99–100). The assumption of affection and especially trust among family members, while frequently made, requires substantiation and when it is subjected to close examination other forms of solidarity instead of trust tend to emerge.

In his discussion of “Chinese Entrepreneurs and Business Trust” Wong (1996, pp. 19–20) treats personal trust exclusively in terms of trust between family members. The obvious question, of whether personal trust among family members may create distrust outside the family, is answered by claiming that “trust and distrust are not mutually exclusive [but relate] in graduation not dichotomy” (Wong 1996, p. 20). This optimistic appraisal is contradicted by the evidence (Ermisch and Gambetta 2010), however, and Wong might have heeded more the implications for his argument concerning trust between family members of the quotation he provides from Maurice Freedman (Wong 1996, p. 20):

… competition and conflict are inherent at all levels of the [traditional Chinese] social system; brother contends with brother, segment with segment, lineage with lineage, the lineage with the state. But there is also harmony because each contender must be united against its opponent.

The harmony that comes from uniting against an opponent in these circumstances is not based on affection and trust. Indeed, given the interdependence of egoistic selves constrained by familial obligation that is characteristic of Chinese networks (Fei 1992), a more appropriate summary term than “trust” for these circumstances might be “calibrations of expedient alliance”; this latter produces solidarity but not trust.

The confident reliance on another that typically comes with trust assumes an independent and unconstrained orientation that is absent in the relations outlined in Freedman’s quotation above and the “social system” to which it refers. Indeed, Peng (2004, p. 1052), in attempting to characterize the mutual support between family members offers an oxymoronic notion of “enforceable trust” to refer to “a norm of mutual commitment between group members that promotes trustworthy and cooperative behaviour”. In a moment of clarity, Peng (2004, p. 1052, note 5) admits that “solidarity” may be a more suitable term than “enforceable trust” in light of the pattern of family relationships he goes on to describe, which resemble those indicated by Freedman in the preceding paragraph. In more general discussions of solidarity, including familial solidarity, there is an absence of references to trust (Hechter 1987; Komter 2005) and the classic account of the emergence of solidarity in a group through conflict with another is also without reference to trust (Simmel 1969) even though its author elsewhere provides an account of trust, indicated above, that scholars continue to draw upon.

Rather than characterize Chinese family relations in terms of affection and trust, it is more appropriate to understand them in terms of the functions families must perform, especially the social and economic functions that respectively entail responsibilities of caring for elderly parents as well as marrying in order to produce heirs and contributing to the family’s collective economic fortunes (Baker 1979). In discharging these responsibilities, which obtain still in present-day China, obligations are prior to affections and where the two most strongly align traditionally is in the relationship between father and son, which alone has the quality of “love” as described in the five human relationships (wu lun) set out by Mencius (2004, p. 60) but which is typically experienced as conditional and possibly distant. In noting the “unavoidable obligations,” the “involuntary relations,” and “unbreakable relationships” of Chinese family ties maintained by “strong normative pressures,” Luo and Yeh (2012, pp. 56, 62) refer to what they describe as “behavioural trust,” which must be distinguished from “real trust” because while “benevolence and loyalty” are to be found in the relations between family members, such relations are “enforced” and “monitored.” It is clear that while the language of trust persists in the accounts mentioned above and similar ones there is an alternative vocabulary through which the solidarity of groups and the reliability of the support their members provide to each other can be characterized. What is required for clarity of expression is a theoretical framework explicating the structure of guanxi, as indicated below, in which the concept of trust is replaced by the concept of assurance. But before that stage of the argument is reached, it can be noted that even though the term “trust” is widely used in discussion of guanxi, another concept is actually operative.

Guanxi and trustworthiness

When trust is identified in studies of guanxi, it almost always refers to “trustworthiness,” in the sense of signals of estimableness or sincerity, rather than “trust,” which implies a sense of unconditional acceptance of dependence on another. Tong and Yong (1998) provide an extended discussion of guanxi and xinyong. They note that xinyong is a notion of Chinese personalism that refers in general terms to integrity, credibility, trustworthiness, and reputation, and in the more limited context of business dealings may refer specifically to credit worthiness (Tong and Yong 1998, p. 85). Guanxi and xinyong are “causally linked,” according to Tong and Yong (1998, p. 85), “with the improvement in one increasing the quality of the other.” The discussion of xinyong in their paper is drawn from interviews in the early 1990s with businessmen in Malaysia and Singapore. Tong and Yong provide a commonsense appreciation of xinyong that follows the usage of their informants. Although Tong and Yong discuss sociological distinctions within the general term “trust,” especially in distinguishing between system trust and personal trust (Tong and Yong 1998, pp. 88–90), their purpose is to restate the vernacular usage of their informants and their informants’ judgment of the suitability of business partners on the basis of their trustworthiness or interpersonal reputation. Indeed, reputation as the currency of network exchanges in general is the symbolic resource principally acquired through them, in which utilitarian resources may incidentally be obtained (Lin 2001a, p. 156). In practical terms, then, there is a virtual identity of trust and reputation in discussion of network relationships (Powell 1990, p. 326).

It is important to note that the term xinyong never means more than trustworthiness and particularly reputational trustworthiness. This is acknowledged by Tong and Yong (1998, p. 85) when they write that xinyong “literally means the use or usefulness of trust.” A similar slippage is found in a study by Chen and Chen (2004) of guanxi development, in which the authors state that “guanxi is … characterized by the mutual trust and feeling developed between the two parties through numerous interactions following self-disclosure, dynamic reciprocity, and long-term equity” (Chen and Chen 2004, p. 306). Yet only trustworthiness (xinyong or kexing) and sincerity (cheng xin cheng yi) are indicated in the discussion (Chen and Chen 2004, pp. 313–314). The Chinese word for trust in the sense of giving trust or being a truster is xinren. Neither Tong and Yong nor their informants use this latter term, xinren. Chen and Chen (2004, p. 313) misleadingly and unhelpfully translate trust as xing, a Chinese character that has to be used with another character if it is to indicate a trust-related term. Trust is a dispositional quality of persons relating to expectations about outcomes generated through reliance on others (Barber 1983, p. 9; Barbalet 2009). To give trust, then, is to accept dependence on another (Luhmann 1979, pp. 15, 22, 81; Rousseau et al. 1998, p. 395), “to make oneself vulnerable to the other person despite uncertainty regarding motives, intentions and prospective actions” (Chua, Morris, and Ingram 2009, 491). Trustworthiness, on the other hand, concerns a socially available perception of reliability typically constructed through a social production of reputation. The important distinction between trust or trustfulness, on the one hand, and trustworthiness, on the other (see Hardin 1993, pp. 512–553), is not drawn by Tong and Yong (1998) or Chen and Chen (2004). Hardin’s comment can directly apply to both sets of authors, when he writes that “many discussions of trust run trust and trustworthiness together, with claims about trust that might well apply to trustworthiness but that seem off the mark for trust” (Hardin 1993, p. 512).

Trust is always based on inferences the trust giver makes about the reliability of another in a cooperative engagement. Such inferences may draw on a perception of another’s trustworthiness but trust cannot be reduced to such a perception as “trust is located in the choice of the trust giver to depend on another in spite of an absence of information concerning the outcome of that dependency” (Barbalet 2009, p. 373). In networks in general and guanxi networks in particular, trustworthiness and reputation do not function in relation to trust-giving but in terms of a very different facility, as we see below, including monitoring of network participants (Luo and Yeh 2012, p. 56; Hardin 1996, pp. 41–42). Monitoring is irrelevant for trust because trust generates expectations about another’s behavior before it can be monitored (Gambetta 1990, p. 217). The point of the discussion here concerning the nature of trust and the distinctions within the notion of trust is not to settle the endemic conceptual and theoretical disagreements in the enormous and growing literature concerning trust (Shapiro 2012, p. 99) but to draw attention to the negative relationship between personal trust and guanxi on the one hand and the particular nature of trustworthiness in guanxi relations on the other.

In Chinese society social relations, not only family and kinship relations but also guanxi relations are founded on personal obligation. The obligations underlying social relations derive from and express the role requirements and the interactions of roles that constitute the particular qualities of Chinese sociality (Hamilton 1990): “The self in Chinese society is embedded in social relationships and is emotionally tied to personal obligations as defined by those relationships” (Hamilton and Zheng 1992, p. 25). This is in contrast with the notion of self that is commonplace in European societies as activated not by obligation and therefore dependency but through voluntary or autonomous “exercise[s] of will” (Hamilton and Zheng 1992, p. 26). It is of signal interest that the concept of personal trust is compatible only with the Western conception of self as described above but not the Chinese. This is because personal trust presupposes the freedom of participants from prior role obligation or commitment to another. The giving of trust is necessarily a voluntary action in which persons are not constrained by any requirement to trust; correlatively, it is “not possible to demand the trust of others” (Luhmann 1979, p. 43). In his classic account of personal trust Luhmann (1979, p. 41) emphasizes that the “first basic prerequisite for building up personal trust is that human actions are perceived in general as personally determined.” The emphasis on personal determination here reinforces the idea that trust is provided by an individual free of the determination of other persons or institutional norms in which a “selective representation of self” provides a basis for an “anticipated continuity” of engagement through trust in which “new ways of behaving become possible” (Luhmann 1979, p. 40). The conception of the person in which trust operates is therefore of an autonomous self, determining its own identity through attachments and actions voluntarily willed.

The trustworthiness held to be pervasive in guanxi relationships does not correspond with the self-willed freedom to choose that is associated with the concept of trust as it is understood in the specialist literature and in vernacular usage in European societies. It is associated, rather, with the habituated behavior expressive of role obligation in which signals of sincerity or estimableness are conveyed through the reassurance of familiarity generated through repeated and close contact. The conflation of trust and trustworthiness that pervades the discussion of guanxi is clear in Kao’s (1996) discussion; at the same time, however, the nature of trustworthiness indicated in the present article is also expressed in his depiction of the basic characteristics of “personal trust” in Taiwan:

This type of ‘trust’ is certainly particularistic, but it is not based upon ascribed relationships alone. … It depends upon [a person’s] achievements, upon demonstrating that they can be trusted. Therefore, in Chinese society “trust” is inseparable from “personal intimacy” (Kao 1996, p. 63).

If a person is “able to demonstrate his loyalty, then he becomes ‘trustworthy’” (Kao 1996, p. 62); trustworthiness derives from being “familiar” with “somebody you already know” (Kao 1996, p. 67) and the relations of dependence and obligation are therefore the structural prerequisites of trustworthiness in guanxi relations.

In addition to the above considerations there are further grounds to doubt the veracity of trust in networks, especially in China. As noted above, a principal mechanism of network relationships is held to be trust borne of group homogeneity (Powell 1990, p. 326), a principle reiterated in the treatment of guanxi discussed in the preceding section. But group homogeneity is likely to be a counter-indicator for trust. Family enterprise in China and Chinese cultural areas is often seen as demonstrating the value of family solidarity as a means of successful entrepreneurial engagement through the provision of business-related resources by way of kinship links, including access to credit and reliable workers (Peng 2004; Whyte 1996). Operational costs can be reduced in such enterprises through employment of family members, thereby increasing an enterprise’s competitiveness (Wong 1988, pp. 143–144). The employment of family members in top positions, which does not necessarily mean neglect of ability or training, provides “more [trustworthy and] highly motivated … employees” (Wong 1988, pp. 144–145). The downside, however, is that family loyalty operates as an effective basis of distrust of non-family members, which makes cooperation with non-relatives outside the enterprise difficult and possibly expensive (Whyte 1996, pp. 3–4). The association of strong bonds between network members on the one hand and distrust of non-network participants on the other is widely acknowledged (Edelman et al. 2004; Kern 1998). The causal relationship between strong family ties and low levels of trust of non-family members (Ermisch and Gambetta 2010) is arguably limited by the cultural generation in China of fictive kinship ties between business associates (Bian 2001, pp. 275–278), but it is not clear that the efficacious product of fictive kinship is trust.

Communal and customary relationships are treated as the basis of trust by social capital theorists (Fukuyama 1995; Putnam 1993). In addition to the problem of relations with non-communal members analogous to the generation of low trust of outsiders among close family members, there is empirical evidence that community undermines rather than promotes trust within a community itself. Strong community networks tend to generate envy and envious gossip and are therefore likely sources of not trust but distrust, which is an indifference to trust relations, and also mistrust, which undermines trust when it exists (Bergmann 1993; Boissevain 1974; Eisenstadt and Roniger 1984). The literature concerning trust in China provides contrasting if not apparently contradictory findings, ranging from reports of low trust to high trust, depending on the focus and scope of the study in question (Child and Möllering 2003; Gabrenya and Hwang 1996; Huff and Kelley 2003; Newton 2001). Given the invasiveness of official powers and the prevalence of corruption in China (Wedeman 2012) the basis for social trust is low. Business practices suggest also low personal trust in business communities. It is reported that entrepreneurs in mainland China keep three sets of accounts; one for personal use, in which there is an accurate report of business transactions; one for taxation purposes, in which the accounts are false or inaccurate; and one to show to business partners, which also presents a false or inaccurate report of balances (Wank 1999, p. 73, note 4). Similar practices are reported for other Chinese cultural areas (Kao 1996, p. 66). Assuming that these practices are common—and anecdotal evidence suggests that they are—then Chinese business persons will know not to trust their business partners and associates.

Assurance and the triadic structure of guanxi networks

Given the prevalence of guanxi in China and, in light of the above remarks, the effective absence of interpersonal trust, it is necessary to consider what is called here the molecular structure of guanxi as based on something other than trust relations. This structure of guanxi networks is partly indicated in the distinction between trust and assurance. Yamagishi and Yamagishi (1994, p. 132) write that:

Trust is … an expectation of goodwill and benign intent. Assurance, on the other hand, is … an expectation of benign behavior for reasons other than goodwill of the partner. Trust is based on an inference of the interaction partner’s personal traits and intentions, whereas assurance is based on the knowledge of the incentive structure surrounding the relationship.

Assurance, then, in contrast to trust, relates to an incentive structure external to the individual participants in a network exchange and their personal expectations of others in network relations. Reference to assurance takes us only part of the way, however, in understanding the molecular structure of guanxi. Assurance as an incentive structure requires institutional grounding. There are at least two possible ways of grounding assurance as an incentive structure: one is in legal institutions and another is in networks. Law and networks as assurance structures share a common quality that is acknowledged for law but neglected in appreciating how guanxi networks operate. Institutionalized assurance functions in terms of third-party enforcement. Recognition of this fact leads to clarity in conceptualization of the molecular structure of guanxi networks.

Networks are typically seen as series of dyadic relations. Summarizing the conventional understanding of them, Aspers (2011, p. 20) defines networks as “several interwoven, dyadic (binary) ongoing interaction relations between actors (nodes) characterized by reciprocity” (see Granovetter 1973). Guanxi is similarly defined as a “dyadic, particular and sentimental tie that has the potential of facilitating favour exchanges between the parties connected by the tie” (Bian 2006, p. 312) and as an “informal, particularistic personal connection between two individuals who are bounded by an implicit psychological contract to follow the social norm of guanxi such as maintaining a long-term relationship, mutual commitment, loyalty, and obligation” (Chen and Chen 2004, p. 306). To understand the operation of assurance in networks, and especially guanxi networks, and the institutionalization of assurance within them it is necessary to appreciate that while binary relations constitute the most elemental aspect of networks, their atomic structure, they are not characteristic of the operational form of networks that have instead the molecular structure of triads. This is because the assurance of compliance within network exchanges is achieved through the third-party enforcement of reputation, which can therefore occur only within triadic relations.

Through a conceptual analysis of guanxi, it is possible to distinguish two distinct phases, one private and the other public. The initial negotiation of a favor involves only the favor seeker and the one from whom the favor is sought; it is therefore dyadic in its structure. An “operating principle” of this first phase of guanxi relations, to use Chen and Chen’s term, which they say is “essential for establishing guanxi bases,” is personal- or self-disclosure (Chen and Chen 2004, pp. 310, 316). A characteristic and unique feature of dyadic relations is indeed the possibility of secrecy between the participants (Simmel 1950, p. 123). A key element of the first phase of a (potential) guanxi exchange is the need for both parties to maintain confidentiality (Lin 2001a, p. 158). Because there are social costs in failing to acquire a favor and also in refusing to provide favor, the interests of both parties are safeguarded by avoiding public knowledge of their negotiations. Any suggestion that trust is necessary in this confidential exchange (Chen and Chen 2004, pp. 313–314) may be dispelled, for the negotiations will typically be conducted on the basis of a sense of obligation, which for family members is likely to be sufficient, and with an orientation guided by calculability concerning creditworthiness, the accessibility and valorisation of means, and so on (Yan 2009, pp. 192–202).

On the successful conclusion of the dyadic and confidential first phase, guanxi enters a second phase in which public propagation and recognition of the exchange is essential (Lin 2001a, p. 159). Indeed, it is in the public second phase of guanxi where the payoffs for both parties engaged in the first phase negotiations are located:

For the giver, being recognized in the social networks for one’s ability to render favours increases one’s standing or reputation in the community … [which is] the payoff sought by favour givers in guanxi …

… For the favour seeker, obtaining a successful favour indicates his or her capability in vertical and upward access to valued resources in the society … [so] word-of-mouth diffusion of a successful guanxi enhances the reputation or social standing of the favour seeker as well (Lin 2001a, p. 157).

In each instance, for both favor seeker and favor provider, the success of their transactions is registered in the public visibility of their exchange. This public visibility is frequently noted. In his discussion of relations within the business community in the Chinese city of Lukang DeGlopper (1995, pp. 205–206) indicates that the “firm’s most valuable asset is hsin-yung [xinyong] … a reputation for meeting one’s obligations.” The achievement of this reputational trustworthiness, he goes on to say, is in the fact that “All transactions take place before an audience or chorus of nei-hang-ren [fellow businessmen], who continually observe and comment on each other’s doings” (DeGlopper 1995, p. 206). But this public visibility is not always properly understood, for when “everyone supervises the actions of others” it cannot be said, as these authors maintain, that the “unit of control is the dyadic relationship” (Hamilton and Zheng 1992, p. 27). That guanxi exchanges necessarily involve third-party observation through which the currency of reputation is maintained and sanctions against possible defection from agreements are executed indicates a triadic not a dyadic form. Indeed, a principal mechanism in the operation of guanzi is face (mainzi), comprising a projected self-image (Cheng 1986, pp. 337–344) which is validated through public perception and approval (Hwang 1987, p. 960–962). The characteristic feature of triads, in contradistinction to dyads, is that in triads the third member plays a mediating and sanctioning role (Simmel 1950, pp. 148–149).

The molecular structure of guanxi networks in their triadic form entails an assurance of enforcement against defection through the mechanism of reputation subject not only to public recognition but more importantly to public scrutiny and sanction. Loss of reputation through defection leads to exclusion from future network exchanges; enhancement of reputation through successful guanxi leads to continuing access to future exchanges and the possibility of accessing increasingly beneficial favors. Without the public visibility internal to triadic forms such an assurance regime would not be available and reputation would be without meaning or purchase. It is in this context that the importance of trustworthiness is derived, and the characteristic form of trustworthiness in interactive familiarity based on interdependent relations of normative obligation associated with role expectations, discussed above, is established.

Reference to triadic form here does not imply that guanxi exchanges necessarily incorporate triangular relationships, although some do. Ambrose King (1991, p. 74) mentions that third-party brokerage may facilitate guanxi building when two parties wishing to establish guanxi with each other have had no previous interaction. It is commonly assumed that an intermediary (zhongjian ren—literally, middle person) known to the two parties of an eventual guanxi exchange may provide an element of assurance, insofar as each party to the exchange owes obligatory feelings (renqing) not only to the other party but also to the intermediary (zhongjian ren), who by virtue of his or her role in bringing the participants of a guanxi exchange together has an interest in its success. In fact the function of zhongjian ren is to establish a shared social identity through a common link with a third person (Hamilton and Zheng 1992, p. 23; Yang 1989, pp. 40–41). Triangular relations incorporating zhongjian ren therefore primarily relate to the conditions of possibility of entering a guanxi relation in the first phase described above and not necessarily to the conditions of realization in the second phase, as some authors have assumed (Lui 2001, p. 391; Qi 2013, pp. 316–317), in which triadic relations provide the basis of assurance. In any event, not all guanxi relations require intermediaries. Chen and Chen (2004, pp. 311–312) indicate that the bases of guanxi relations are of three types, only one of which involves a triangular relationship (a common third party); the other two are binary relationships founded respectively on, first, common social identity (of birthplace, educational institution, and workplace) and, secondly, what they call an “anticipatory” base of future intention to form a guanxi relation, what more generally might be called a common trajectory.

With increased marketization in China, it is likely that guanzi will be based less on the strong ties of common social identity and increasingly involve more weakly connected parties participating in multilateral interactions that will unavoidably depend upon third parties, either as intermediaries consciously selected or implicitly in the latent and structural interdependency that arises in multilateral interactions (Lin 2001b, pp. 55–77; Lin 2002, pp. 58–59; Chan 2009, pp. 726–733). This change in the form of guanxi, from strong-tie to weaker-tie form, is associated with a difference in its application, usually described in terms of a distinction between expressive and instrumental relationships, although the distinction is problematic in light of the self-interested nature of many strong-tie expressive relations (Wilson 2002, pp. 174–176). Whatever the base from which a successful guanxi exchange arises, strong-tie or weaker-tie, binary or triangular, the public disclosure of the exchange involved, either by dissemination, gift exchange, or some other social declaration, generates public reputational visibility which introduces a third element into the relationship providing all guanxi exchanges with a triadic form.

Assurance and face

As we have seen, most of the research on guanxi is focussed on the first phase of guanxi exchanges and tends to treat the second phase in a taken-for-granted manner. The literature typically discusses the pattern of the dyadic ties that constitute guanxi. In this vein particular themes emerge. One is the move from strong-tie to weaker-tie guanxi, a development that occurs as opportunities for social relations and contacts become more expansive as the constraints of household registration (hukou) and work unit (danwei), especially, have become relaxed during the period of economic transition. Another concern in much of the literature has been to characterize the use or purpose of guanxi in terms of either its expressivity or instrumentality. Thirdly, but by no means final, there is extensive discussion of the role and future of guanxi in the developing institutional mix of the transition period and especially whether formal and legal forms will crowd out and displace informal guanxi. In all of these discussions the mechanism of guanxi has been understood in terms of trust assumed to be implicit in the relationships involved.

When an appreciation of the structure of guanxi comes to focus not on its atomic or dyadic form, but rather on its molecular and triadic form, then assurance rather than trust is seen to function as the efficacious mechanism of the relationship. Under these circumstances, in considering guanxi through the operations of its second or public phase, then the research question that comes into focus concerns the ways in which public visibility operates to maintain guanxi networks and therefore the sanctioning of guanxi exchanges. The role of face (mianzi) in this process is central. The importance of face to guanxi is frequently acknowledged, but in a taken-for-granted manner that therefore leaves it largely unexplored. Yet the operations of face are central to an understanding of the triadic processes of guanxi and especially the sanctions within them.

Being able to provide a favor to another has the consequence of enhancing a person’s reputation within the social network in which the exchange occurs, not only because it demonstrates a material capacity but more importantly the application of that capacity in the affective response to another in a way that satisfies the norms of sociality. That is, the provision of a favor demonstrates a person’s having “human feelings” (renqing). The recipient of the favor also enjoys enhancement of reputation by demonstrating their worthiness in being extended a favor and the expectation on the part of the favor-giver of the recipient’s ability at some future time to reciprocate (bao), therefore demonstrating his or her own renqing. The reputations in question, which amount to claims on public judgments on the adequacy of the performance of the expectations of the positions respectively held by these persons in the relationships of the network, are experienced as enhancements of face (mianzi). The involvement of public judgments of reputational claims, through face, is the basis of any sanction required for the maintenance of the network and punishment of any wrongdoing within it.

In these considerations, one must agree with Ho (1976, p. 872) that “the social dynamics involved in losing face are more deserving of our attention than those involved in gaining face … everyone who care for maintaining a minimum level of effective social functioning must see to it that his face is protected from being lost.” Indeed, a person’s concern for preservation of their “face” cannot be conceived as merely an individual or personal matter: “the concern for face exerts a mutually restrictive, even coercive, power upon each member of the social network … the individual’s actions, far from being directed by [their] own wishes, are in effect dictated by the necessity of meeting the expectation of others” (Ho 1976, p. 873). In any failure to meet these expectations there will therefore be a consequential loss of face that imposes limits on the person’s ability to engage in future guanxi exchanges and therefore affect the structure and constitution of the network itself. In this sense, then, face is a social resource that exists through a complex interaction of performance, self-evaluation, public visibility, and the approbation of others. Withdrawal of the last of these nullifies the value of the first two and intensifies the third, which now has negative consequences for the one who has “lost face.” Fear of losing face generates a practice of the maintenance of ledgers or accounts in which a record is kept of gifts and favors, given and received, and their value (Wilson 2002, p. 170; Yan 2009, pp. 184–191, 202–204).

Under conditions of strong-tie guanxi, which typically operates in more-or-less closed communities, including villages, urban neighbourhoods, and family businesses, the principal means of assurance through which face may be lost as a consequence of transgressing guanxi expectations is gossip. Farrer (2002, p. 199) holds that guanxi relies on gossip in a number of ways; first, gossip is a means of exchanging information about network members; second, gossip is not simply a by-product of networks but definitive of membership in a network and therefore indicates the boundaries of guanxi networks; finally, gossip may be a source of moral authority for its purveyors. Of particular interest here is the way in which gossip functions as a kind of moral discourse, directed to the transmission of information concerning departures from the expected norms of behavior, commitment, and decorum. In this sense, then, gossip is not only a means of censure but a basis of assurance for all members of the network that transgressions of the expectations of appropriate behavior shall be exposed, with a consequent loss of reputation and face for the transgressor.

Gossip may operate also in the general processes of assurance in guanxi networks based on weaker- and not only strong-ties, but its effectiveness will be limited by the scope or range of communicative interactions among increasingly distant participants. If gossip is considered principally as a medium of communication it will be clear that it cannot hold a monopoly position in China today, given both the changing nature of networks and the availability and role of informal electronic communication (Herold 2008) and the mass media in exposing various sorts of transgressions of social expectations. The role of these latter and other communication forms in providing assurance over guanxi exchanges cannot therefore be ignored. Farrer (2002, p. 199) insists, though, that gossip remains the most effective means of communication because it is less limited in the topics and perspectives it offers, in comparison with the Chinese mass media, which are subject to political control. While this last point is true, it overlooks the fact that “the Chinese media enjoy significantly more autonomy in reporting on financial misconduct than they do reporting on most other areas of Chinese law and society” (Liebman and Milhaupt 2008, p. 981). Liebman and Milhaupt (2008, p. 981) go on to make the important and counter-intuitive claim that the “[mass] media are perhaps the most effective regulator of corporate wrongdoing in China today.” While not directly related to the maintenance of assurance in guanxi networks, Liebman and Milhaupt’s study of “Reputational Sanctions in China’s Securities Market” indicates the capacity of mass media communication to function as a mechanism of assurance through direct exposure of stock-market transgressions leading to loss of face.

China’s stock exchanges, in Shanghai and Shenzhen, since their establishment in the late 1990s, have rapidly become among the largest in the world in terms of total market capitalization and other standard measures (Liebman and Milhaupt 2008, p. 937). Legal regulation of the exchanges is slow and inefficient, with 85 % of eligible target companies escaping prosecution (Liebman and Milhaupt 2008, pp. 943, 955). A much more effective but under-researched sanction is a reputational mechanism engaged by the exchanges themselves, namely public criticism or censure (gongkai qianze). Unlike legal enforcement gongkai qianze is applied soon after discovery of transgression and at a low-cost (Liebman and Milhaupt 2008, pp. 955, 977); stock-exchange gongkai qianze is “virtually always reported in the Chinese media, which ensures broad public exposure of the fact that a company or individual has received scrutiny and criticism by a stock exchange” (Liebman and Milhaupt 2008, p. 973). Because gongkai qianze is public disclosure of wrongdoing their major consequences are reputational (Liebman and Milhaupt 2008, pp. 973, 977) and are therefore directed to the perpetrator’s face (mianzi).

The assurance mechanisms of guanxi networks rely on public exposure of transgressions of social expectations; such public exposures affect the reputations of offending individuals and lead to their loss of face. The operation of such mechanisms therefore provide re-assurance to network members in general that as sanctions apply to transgressors of network norms then the continuing operation of the network and the practices that constitute its proper operations are secure. As guanxi networks change in concert with changes in social relationships and economic opportunities, so too shall the medium of communication of information and censure concerning the behavior of network members also change. Gossip, the activities of netizens (wangmin), the mass media and other communication means all play a role, the significance of each varying with the nature and breadth of the network involved. The research reported by Liebman and Milhaupt (2008) indicates the continuing importance of public exposure and also the need to recognize the role of practices that from a non-Chinese perspective may be unlikely sources of assurance efficacy, but that in the context of Chinese institutions are highly salient in understanding the continuing importance of guanxi networks that not only persist but develop in unexpected ways, and the means of assurance that operate within them.

Conclusion

A number of counter-intuitive and revisionist conclusions can be drawn from the preceding discussion. First, in spite of the exponential growth of treatments of networks, including guanxi networks, in terms of trust relations there is limited intellectual value and in fact a number of conceptual and factual problems that ensue in supposing the effectiveness of trust in this context. A second conclusion, then, is that guanxi and possibly other forms of networks operate not in terms of trust but expectations concerning third-party capacities to provide assurance to members of the network that transgressions against its norms shall, through public exposure, suffer loss of reputation or face and thereby effective expulsion from the network. When social interaction is structured triadically then the behavior of any pair of participants is subject to the influence and sanction of third parties. The third party (each is a third party to the others) interprets the interaction from the point of view of their own interests. The interest of guanxi participants is to maintain the network. The triadic structure of guanxi networks, also concluded from the preceding discussion, is evident in the relations between publicly displayed feelings of obligation (renqing) and face (mianzi), which play a role in maintaining guanxi through their visibility to others who can affect the reputational standing of those participating in the mutual obligations entailed in guanxi exchanges.

The operation of guanxi in terms of its payoffs and reproduction depends, then, not on personal trust between participants but on the assurance participants have of affecting the public reputation of those engaged in guanxi exchanges, both positively and negatively. There is an assumption in much of the literature that guanxi is necessarily based on strong-ties and therefore that the loss of face subsequent to norm transgression is effected only by gossip. Two other conclusions can be drawn from the present article. One of these is that the social changes and economic opportunities generated through China’s present transitional stage have meant that relationships of favor-giving constitutive of guanxi now occur outside of the confines of village, neighborhood, and family and therefore that weaker-tie bases of guanxi are not anomalous but now routine. Another conclusion, then, is that in these circumstances it is necessary to appreciate that face-to-face gossip can no longer remain the dominant form of communication through which the assurance of face-loss resulting from norm transgression is achieved and that other media have a role to play as marketization in China leads to an increased salience of weaker-tie guanxi exchanges, relations, and networks.