Guanxi as Social Exchange: Emotions,Power and Corruption

Introduction

Social life entails interdependence between individuals, frequently manifest in exchange through which something given by one person to another precipitates a return provision. The idea, that exchange ‘is the purest and most developed kind of interaction, which shapes human life when it seeks to acquire substance and content’, is reinforced by the claim that it is ‘often overlooked how much what appears at first a one-sided activity is actually based upon reciprocity’ (Simmel, 1978: 82). Simmel goes on to indicate how orators, teachers and journalists feel ‘the decisive and determining reaction of the apparently passive mass’, implying that this ‘reaction’ is an enabling provision in return for the orator’s, teacher’s and journalist’s activity, an exchange between those who act and those who are acted on.

Simmel’s account of exchange arguably answers his question: how is society possible? ‘Society consists of […] beings which, on the one hand, feel themselves to be complete social entities, and, on the other hand – and without thereby changing their content at all – complete personal entities’ (Simmel, 1971: 18). Social exchange is a significant mechanism in this transition. A quarter-century after Simmel, answering a different question, Mauss (2002) established another approach to social exchange. Through a perspective paralleling the idea associated with Hirschman (1977), that capitalism emerged through the displacement of one set of passions (bloodlust) by another (greed), Mauss (2002: 104–105) holds that: ‘[t]wo groups of men who meet can only either draw apart [and possibly] fight – or they can negotiate […] Nothing better interprets this unstable state between festivals [of collective exchange] and war’.

Implicitly critiquing Durkheim’s idea that shared beliefs generate mechanical solidarity Mauss claims instead that it is ‘exchange in archaic societies’, to quote the subtitle of his essay, that produces social solidarity. In archaic societies,

it is not individuals but collectivities that impose obligations of exchange […] upon each other […] [C]lans, tribes, and families confront and oppose one another either in groups who meet face to face in one spot, or through their chiefs. (Mauss, 2002: 6)

This collectivist strand of exchange theory not only exemplifies anthropological rather than sociological approaches (Ekeh, 1974) it requires acceptance of principles that become entirely contingent when the focus shifts from group to individual exchange, including the idea that exchange obligation is wholly constraining, precluding a right to refuse a gift (Mauss, 2002: 52). Refusal of a potlatch or kula gift may indeed precipitate war; gift refusal between individuals, though, simply disrupts the exchange in question.

Certainly Mauss’ (2002: 3) opening claim, that gift exchanges appear to be ‘voluntary, [but] in reality they are given and reciprocated obligatorily’, is axiomatic in sociological exchange theory, as is his acknowledgement of the significance of exchange for honour and prestige (Mauss, 2002: 47–49). In addition to these widely shared notions is another, the idea that interaction is sustained by exchange rather than the other way around: in Simmel’s words, ‘exchange is not conducted for the sake of the object that the other person possesses, but to gratify one’s personal feelings which he does not possess’ (Simmel, 1978: 82). The reference here to the necessary role of emotion, ‘one’s personal feelings’, in understanding the operations of exchange is not, however, widely shared although it deserves close attention.

Simmel’s insight, concerning the importance of exchange for social interdependence, independently developed by Mauss, is later expanded on the premise of operant conditioning: ‘interaction between persons is an exchange of goods, material and non-material […] [in which] each is emitting behavior reinforced to some degree by the behavior of the other’ (Homans, 1958: 597, 598). The other’s ‘behaviour’ either reinforces the original action and therefore encourages its repetition or through negative sanction dissuades its recurrence. Subsequent elaboration of social exchange theory in this tradition (Emerson, 1972a, 1972b) gave rise to an important experimental approach in social analysis (Cook and Emerson, 1978). An alternative to Homans’ operant-conditioning approach, developed by Blau (1964), privileges rational expectation rather than reinforcement. None of these accounts provides a role for emotions in exchange.

More recently social exchange theory has undergone an ‘emotions turn’ (Lawler, 2001; Lawler et al., 2009). This approach holds that the consequences of emotions formed through exchange generate individual commitments to groups and institutions. Lawler draws on the operant conditioning tradition insofar as social exchange is held to produce ‘global feelings’ that are either positive or negative and which are internally rewarding or punishing (Lawler et al., 2009: 35, 79–80). As the social units in which the exchange takes place are perceived by those involved to be the source of these global feelings, so different types of exchange generate different levels of commitment (Lawler et al., 2009: 80–85).

The present article examines a particular type of social exchange in order to demonstrate how emotions are formative in rather than simply consequences of social exchange. It will also be shown that by addressing the determining role of emotions in exchange, social exchange can be distinguished from other forms of interaction, including bribery. The type of exchange to be discussed below is Chinese guanxi, a form of exchange frequently assumed to be either inherently corrupt or a facilitator of corruption, thus making it an ideal case for the purposes of the present argument.

Guanxi as Social Exchange

Guanxi can be understood as a form of asymmetrical exchange of favours between persons on the basis of enduring sentimental ties in which enhancement of public reputation is the aspirational outcome. The benefits of guanxi exchanges are thus normative and reputational, although material or instrumental advantage is always possible. For it to remain a guanxi exchange, though, the realization of an instrumental outcome must be embedded in reputational enhancement, as shown below. Networks of guanxi exchanges are a fundamental element of traditional Chinese social structure that continues to be pervasive in mainland China and Chinese cultural areas. Research on guanxi has over the past four decades become an expanding concentration in anthropology and business studies as well as in sociology (Chen et al., 2013; Liu and Mei, 2015; Qi, 2012). Guanxi literally means ‘relationship’ or ‘personal connection’ but it is important to acknowledge the essential mediating role of gift exchange that, first, signifies the existence of a relationship and, second, consolidates the connections between participants. The term for gift in mandarin Chinese is liwu, with li meaning rite; in guanxi exchange has ritual force. The ‘ritual’ dimension of guanxi is performed with respect to a number of practices associated with exchange decorum often summarized in terms of renqing (literally, human feelings), explained below. The role of gift exchange in guanxi is acknowledged by all who discuss it, but its significance is diminished if the explanatory framework adopted to treat guanxi fails to foreground its social exchange dimensions.

The rise of guanxi research incidentally parallels a declining interest in social exchange theory and a proportionate rise in sociology of social network analysis. The relations between these latter two approaches is complex, given the role of social exchange in the formation, consolidation and development of social networks as well as significant explanatory differences between them (Cook and Whitmeyer, 1992). The purpose of mentioning these cross-cutting trends in the fortune of academic theories is to note that sociological discussion of guanxi typically takes place in terms of its network rather than its social exchange properties (Chang, 2011; Huang, 2008; King, 1991; Tong and Yong, 1998), although it is possible to question the extent to which guanxi networks satisfy the conditions of standard network analysis (Barbalet, 2015). There is, however, an exception in an account of guanxi ‘on the basis of social exchange theory’ (Hwang, 1987: 945).

Hwang’s social psychological discussion of guanxi is informed by the experimental and operant conditioning tradition mentioned above. For Hwang (1987: 947) ‘face’ and ‘favour’ in guanxi means that ‘the process of using power to influence other people is basically a social exchange process’. For reasons indicated below, the idea that exchange and guanxi in particular operate through power is not accepted here. But the definition of power used by Hwang (1987: 947), ‘socio-moral suasion, or peer-group pressure, that one may use to change the attitude, motivation, or behavior of another to conform to one’s will during the process of social interaction’, is so broad as to include a range of relations from which power is usually distinguished. Indeed, a source of the approach Hwang draws on frequently slides, in discussion of power as dependence, from ‘power’ to ‘control’ and then to ‘influence’ (Emerson, 1962: 32–33) and in a later statement claims that ‘power is redundant and unnecessary […] given our conception of dependence’ (Emerson, 1972b: 64). While he uses the language of power it is clear that for Hwang (1987: 958, 960–962) social standing, prestige or status is the source of influence between persons in guanxi exchanges.

The possible significance of power in exchange is not denied here. Indeed, the importance of power in exchanges constituting bribery will be explored below. In treatments of Chinese business and politics guanxi is frequently held to be indistinguishable from corruption (Fan, 2002; Luo, 2008; Smart and Hsu, 2007; Wang, 2014). When guanxi and corruption are distinguished the basis on which they are separated is not always clear:

Because guanxi provides particular instead of general access to resources and operates through personal relations rather than formal structures, there is a tendency to associate it with corruption, bribery and malpractice […] [but the] impropriety associated with guanxi is not inherent in it […] While guanxi is not itself a cause of corruption, if corruption occurs guanxi is likely to be one of its mechanisms. The emphasis in guanxi, however, is on relationships, whereas in corruption it is on unwarranted material gain achieved through inappropriate means […] Guanxi frequently appears to be associated with bribery and corruption because of the manner in which it is established and maintained, namely through social occasions, such as lunches, dinners, and gift-giving. (Qi, 2013: 311–312)

It is shown below, rather, that while gift-giving and exchange may be features of both guanxi and corrupt practices, they cannot be distinguished in these terms. But it is first necessary to say more about the nature and basis of social exchange in guanxi.

The sense of obligation internal to gift exchange is treated in various ways. The perspective developed here emphasizes the importance of emotions that directly inform the obligatory relations maintaining social exchanges, including guanxi. The importance of personal feelings or emotions over objects of possession in exchange, mentioned by Simmel (1978: 82) above, will be elaborated in order to ground guanxi as a relation of social exchange. In doing so it will be shown how guanxi can be clearly distinguished from bribery. Bribery is based on coercive power differentials that do not operate in guanxi, founded instead on the esteem of participants, characteristic of social exchange in general. This distinction is reflected in the emotional complexion of each, as indicated below.

Use of the term guanxi here does not necessarily correspond with vernacular usage. In south-west China the dialect term goudui is used, meaning relationship cultivation (Osburg, 2013). In north-east China people are more likely to use the term renmai, made up of two characters ren (person) and mai (pulse, vein or mountain range), which together literally mean being part of a stream or range, part of a larger unit, so that renmai refers to sustaining contacts with others. Another term often used in rural China, li shang wanglai, means ‘reciprocally personalizing relationships’ (Chang, 2009: 478). Both renmai and li shang wanglai have a positive connotation and avoid the possibly negative connotation of the term guanxi as it is used in present-day China, which has become associated with only instrumental exchange based on self-interest, if not corruption. Indeed, terms such as self-interest (zishen liyi) and even ambition (ye xin) carry negative connotation even though the behaviour referred to by them is rampant in today’s marketized China.

These last remarks raise the question of historical context. Like all enduring social institutions quanxi undergoes modification as the circumstances in which it operates change. It was indicated above that guanxi functions in terms of normative and especially reputational outcome and also that guanxi may be mobilized to provide material benefits. These may be summarized as expressive and instrumental aspects of guanxi respectively (Gold, 1985: 659; Hwang, 1987: 949–953; Yan, 1996: 226–229). In rural communities and in imperial China guanxi is a means of social bonding and therefore its expressive form predominates. During the Mao era guanxi was used to circumvent rationing and its instrumental form was therefore dominant. In China’s current marketized economy guanxi provides businesspersons access to administrative favour and its instrumental aspect predominates. Other accounts refer to the changing forms of guanxi in different terms (Chang, 2011; Qi, 2013). Behind these and similar distinctions there remains a stable core of relations facilitating exchange in which the participants are committed to the propriety and decorum of their association through emotional ties and the obligations entailed in them. The following section will consider the nature of obligation in social exchange and the next will explore the particular emotions that animate guanxi.

Obligation in Exchange

Like social exchanges in general guanxi exchanges operate through a sense of obligation: a ‘person who supplies [goods and] services […] to others obliges them to reciprocate’ (Blau, 1964: 28; see also 29, 89, 92, 98–99, 133–136, 141–142). Obligations are ‘existential conditions’ of exchange and not the result of a norm of reciprocity, according to Blau, even though such a norm may arise to reinforce and stabilize ‘tendencies inherent in the character of social exchange itself’ (1964: 92). For Blau (1964: 91) the obligatory nature of exchange is contingent upon its voluntary nature, insofar as the provision of a resource from one person to another in exchange is based on an expectation of a return provision that cannot derive from negative sanction or coercion. The enforcing mechanism of obligation thus requires considered explication.

Blau’s account of the mechanisms of exchange obligation remains the most developed in sociology. He holds that the unspecified and prospective nature of social exchange is at the root of its obligatory form as the giver has to take the receiver’s future return on trust: ‘[s]ince there is no way to assure an appropriate return for a favor, social exchange requires trusting others to discharge their obligations’ (Blau, 1964: 94). While social exchange ‘may originate in pure self-interest’ it nevertheless paradoxically generates ‘trust in social relations through their recurrent and gradually expanding character’ (Blau, 1964: 94). One’s trust of another derives from demonstration of their trustworthiness. Blau (1964: 94, 98) says that in reciprocating a favour one ‘proves [one]self trustworthy of continued and extended favors’, and, as the ‘establishment of exchange relations involves making investments that constitute commitments to the other party’, then ‘the initial problem is to prove oneself trustworthy’.

The question is thus: can trust be trusted? Even if the provision of a favour gives rise to an expectation of a return that expectation may be disappointed, and it is always possible that it will be, in which case exchange will not occur or, with ongoing exchange, it will cease. As Blau (1964: 6) says, social exchange ‘is limited to actions that are contingent on rewarding reactions from others and that cease when these expected reactions are not forthcoming’ (see also Hwang, 1987: 958). For Blau (1964: 94), then, trust is both a requirement of social exchange and its consequence: he acknowledges that ‘there is no way to assure an appropriate return for a favour’ and that social exchange ‘requires trusting others to discharge their obligations’. This requirement is seen to be satisfied by the operation of social exchange itself. Blau assumes, then, that trust emerges from the need to trust; but, as he acknowledges, a need to trust cannot itself generate or explain trust formation.

One problem with Blau’s and similar accounts is that they assume that any exchange relation is necessarily an exclusive relation between two individuals, parallel to trust relations as dyadic or two-party interactions between truster and trusted (Wilson and Eckel, 2011). An implicit assumption here is that objects of exchange are discrete possessions: something in the possession of one person is given to another person in exchange. But this notion neglects a signal social dimension of exchange. The provision of a gift or favour leads to enhancement of participants’ social standing, status or esteem. In this way social exchange is indeed a social relation; the private provision of a gift from one person to another draws into the relation between them the appraisal of others. Blau is not unaware of this prospect. He observes that ‘[f]avors make us grateful […] [and] if we express our appreciation […] publically […] [we] help establish a person’s reputation’ (Blau, 1964: 16). The point is not assimilated into Blau’s argument concerning the basis of obligation, however, in which exchange obligation remains a dyadic relation of trust-giving on grounds of the other’s perceived trustworthiness.

The role of third parties, suggested by Blau above, is not confined to exchange. One objective of competition between two persons is to attract the attention of a third (Simmel, 1978: 290–291, 2008). In exchange, though, the flow of attention is in the opposite direction. Third-party observation of exchange participants is directed to appraisal of adherence to or departure from exchange expectation, thus conferring or denying social status or esteem. This is loosely the argument of moral sentiments (Smith, 1979). In explaining obligation or moral commitment Smith prioritizes neither norms of reciprocity nor trust but emotional feelings formed in relation to how self is regarded by others. In particular, central to Smith’s account are the emotions of sympathy for others and pride or shame provoked by whether responsibilities are discharged and expectations realized (Smith, 1979: 110–113, 137). This approach and these emotions, or their variants, can be inserted into an account of how obligation arises and operates in social exchanges, including guanxi.

While social science accounts of guanxi typically operate in terms of the primacy of trust (Chen and Chen, 2004; Lee and Dawes, 2005; Luo and Yeh, 2012; Tong and Yong, 1998; but see Barbalet, 2014), paralleling Blau’s explanation of the mechanism of obligation in social exchange, there is also acknowledgement in the literature of the importance of third-party observation as a mechanism of assurance in guanxi relations, as when participants in guanxi relations ‘anticipate that some people in their respective networks may know what is going on between them and may evaluate their interaction in accordance with their social standards’ (Hwang, 1987: 953; see Barbalet, 2014: 62–63). The point here is that to fall short of those standards means losing the esteem of others, losing face and therefore the capacity to participate in future guanxi exchanges (Hwang, 1987: 960–961; see Barbalet, 2014: 63–65). In order to keep face, then, to retain one’s social standing, it is necessary to adhere to prevailing standards and show that one is doing so. In the Chinese cultural context this requires a display of renqing, which is typically translated as ‘human feelings’ even though the term refers to much more than that.

The Emotions of Guanxi

The emotions implicit in renqing are fundamental to guanxi, as are those in ganqing. These emotions complexes (they are not single emotions such as fear or guilt although more concrete than positive or negative ‘global’ feelings) not only situate guanxi as social exchange they also encourage recognition of different types of guanxi in terms of the relative strength and contribution of each.

Discussions of renqing typically acknowledge that the term has a number of meanings (Chang, 2009: 461–462; Yan, 1996: 145–146). The differences do not arise through alternative definitions but indicate different connotations or implications deriving from variation in context and application. The oldest meaning of renqing refers to the emotions inherent in ‘natural’ human relations, especially those of family life (Hwang, 1987: 953; Yang, 1994: 67). Family life in China is governed by hierarchically organized roles and ritual relations between roles. In this context renqing means feelings appropriate to human relationships in which harmony and order in descent and inter-generational relations are maintained, based on principles of seniority and gender. Renqing thus refers to feelings connected with how to act appropriately; it includes those emotions associated with a sense of proportionate relational expectation and the moral or normative patterns of social life and a person’s sensibility to such norms. Renqing therefore also includes the actions and engagements that contribute to concurrent and conforming relations, such as providing sympathy where it is due as well as respect for others, acknowledging a favour, repaying a debt and so on. These latter can be described as giving renqing. Indeed, the actual gifts given in sympathy and out of respect can also be called renqing (Hwang, 1987: 953–954; Yang, 1994: 67–70).

Renqing thus captures a broad spectrum of feelings concerning what is culturally appropriate for a given situation involving persons occupying particular roles. It includes those emotions expressive of custom, propriety and social etiquette. Failure to express the appropriate emotions and to behave accordingly is likely to lead to withdrawal of approval and therefore to loss of face (Hwang, 1987: 960–961; Yang, 1994: 69). In this sense renqing functions to maintain guanxi; it provides both a sense of the codes of conduct appropriate to the social exchanges that constitute guanxi and also the emotional forms that are the means of personal commitment to its practices. Renqing therefore includes what Gouldner (1960) calls the norm of reciprocity and also what Hochschild (1983: 6–7) calls emotional labour, the labour required ‘to induce or suppress feelings in order to sustain the outward countenance that produces the proper state of mind in others’ or more broadly the emotional management that aligns feelings and their expression in accord with the social expectations encapsulated in ‘feeling rules’ (Hochschild, 1983: 113–114, 250–251). Guanxi is achieved through renqing and the associated fear of loss of face in failing to act with renqing.

Renqing is thus a source of emotional involvement in ways of doing things, including the decorum and norms of exchange that constitute expectations and responsibilities in guanxi. In addition to the obligations inflected through renqing, guanxi relations call upon the commitment or, more properly, attachment of one person to another. The term ganqing literally means sentiment or affect and refers to the emotional feelings of affection or attachment to a person or a place (Chang, 2009: 462). It is possible to have bad or negative ganqing as well as positive. Thus, whereas renqing refers to a type of relationship, one in which the discharge of obligation is primary, ganqing indicates more the quality or criterion of a relationship (Chang, 2009: 463; Gold, 1985: 660, n. 1; Jacobs, 1979: 259). In the context of guanxi, if there is ganqing between two persons their relationship will be enhanced by emotional commitment to an enduring intimate bond (Yang, 1994: 121). Without ganqing guanxi is likely to be distant and possibly unreliable (Luo, 2007: 6).

Ganqing is more situational and therefore less norm-dependent than renqing, less about correct decorum and more about personal attachment. Ganqing forms through direct experience with another, in work through cooperative engagement; it is ‘utilized’ in such mutual activities in the sense that ganqing removes disagreement between persons, thus making it possible for things to get done (Jacobs, 1979: 263). This gives depth to the notion that ganqing refers to the quality of relationships. Ganqing does not arise only in work relations, but derives from shared activities in general (Luo, 2007: 15–16). These may include social drinking. In Chinese communities toasting, drinking games and competitive drinking are means of not only relaxing inhibitions but explicitly encouraging ganqing (Kipnis, 1997: 53–57). Since post-Mao marketization shared consumption of commercial sex as well as banqueting and drinking serve the purpose of generating ganqing among businesspersons and between businesspersons and officials (Barbalet, 2015: 1047–1048; Osburg, 2013: 45–65). Such activities ‘would often end with declarations of friendship and affection, plans for future forms of entertainment, and offers of mutual aid’; these activities are explicitly constructed to generate ganqing (Osburg, 2013: 57, 44). The importance of ganqing to business guanxi, in which instrumentalism is high, indicates that emotional attachment is not confined to expressive guanxi, as sometimes assumed (Yang, 1994: 68). Ganqing is a means of creating solidarity, of emphasizing an emotional connection between two persons rather than necessarily indicating a forceful emotional state within any single person. In this regard it is unlike friendship, based on ‘mutual affection and sympathy’; ganqing, on the other hand, ‘presumes a much more specific common interest’ (Fried, 1974: 226). It is the bond of attachment that is generated through the situational manipulation of ganqing that is primary in going beyond the normative structure of renqing and thus more firmly personalizes guanxi relations.

Some conclusions can be drawn from this account. First, it is possible to distinguish different types of guanxi in terms of the relative contributions of renqing and ganqing. Authors tend to emphasize either renqing or ganqing in their accounts of guanxi, but not both. Rather than survey the field it is sufficient to identify some cases. Hwang (1987) is exclusively concerned with renqing in his discussion of guanxi, focusing on guanxi compliance through fear of loss of face. Kipnis (1997), on the other hand, discusses guanxi almost exclusively in terms of ganqing because of his focus on the personal construction of obligation. As with different accounts of guanxi so experience of guanxi and the form it socially acquires varies in terms of whether renqing or ganqing is dominant. What will determine which is dominant must be left to empirical investigation. The differential depth of the guanxi relationship, with ganqing reflecting close and renqing more distant connections, is suggested by Yang (1994: 121–122). In a similar vein Yan (1996: 226–229) distinguishes ‘primary’ and ‘extended’ guanxi, one based on moral obligation and emotional attachment and the other on rational calculation. Irrespective of the particular mix of these two complex emotional forms it is unlikely that guanxi could be performed without involvement of both, one relating to the normative structure of obligation and the other to the bond or connection in which the participants are implicated.

A second question that arises from this discussion is the extent to which the underlying emotional mechanism of guanxi, as manifest in renqing and ganqing, encourage its distinction from other types of relations, such as bribery or corruption, with which it is frequently associated. Yang (1994: 62–63) reports the difficulty even educated Chinese people have in clearly articulating the difference between guanxi and bribery, although she suggests that a discerning element of distinction is the emotional dimension of the relationships: ‘bribery relations are the weakest in terms of emotional affect [whereas] guanxi, renqing and ganqing [are] the most filled with emotional content’ (Yang, 1994: 123). The relation between the briber and bribed, however, is unlikely to be emotionally weak. The briber will hope that the bribe is appropriate and be anxious about whether it will be accepted, the bribed may be fearful of being exposed and so on. A greater continuity than the dissimilarity assumed by Yang between guanxi and bribery has been suggested by Smart (1993). While it is not difficult to distinguish a gift from a bribe in terms of their legal standing or intention Smart (1993: 405) insists that they are ‘not real entities that can be scrutinized, fitted to a formal definition, and labeled’. Rather, they are performances, he says, that must be properly executed and appropriately understood, or possibly misunderstood, in which case the performance can be said to have failed.

Guanxi, Bribery and Corruption: Continuity or Disjuncture?

The idea, proposed by Smart (1993), that guanxi cannot meaningfully be distinguished from bribery as a form of social exchange, is too open-ended. Indeed, earlier in his discussion Smart (1993: 403) indicates another approach:

In guanxi, immediate instrumental purposes are subordinated to the greater aim of developing relationships that may serve as resources for solving problems over long periods of time […] Where the concern of the exchange is not to create such relationships, but simply to achieve some immediate objective for which the relationship would be a useful means, then although the form of the gift may be outwardly followed, its content is different – a deal or a bribe rather than a gift exchange.

Smart’s reference here to enduring relationships in guanxi, and their absence in a bribe, suggests a preliminary framework in which the particular emotional commitments to the decorum of the exchange relation and to the bond between participants (through renqing and ganqing respectively) may distinguish a guanxi exchange from a bribe in which more furtive emotions are likely to operate. Behind these different sets of emotions, as Smart in effect implies, are different structures of relationships. In simple terms both a gift (pertaining to guanxi) and a bribe involve favour exchange. Indeed, the Chinese term for bribe, huilu, literally means wrongfully giving money or gifts. The idea that bribes are one-off and guanxi enduring will not be correct, however, when bribery is continuous or when it melds into what is euphemistically called ‘protection’.

The claim to be developed here, rather, is that a difference between guanxi and bribery, which accounts for the emotional patterns of each, is whether there is coercion in the exchange. Bribery is not only primarily self-interested or instrumental, rather than secondarily or incidentally so as with guanxi as gift-giving, but it typically entails threats and generates risk or vulnerability qualitatively unlike those of a non-returned gift, a point frequently obscured in the literature (Steidlmeier, 1999). The covert nature of bribery and its improper inducements means that it has a coercive form, emergent if not original, and as such relates not to the esteem of the participants, as with guanxi, but to the power differences in their relationship. The accompanying emotions, unlike those of guanxi, are hubris-type and fear-type, depending on location in the relationship, which reflect the structure of power in bribery, its inverse form blackmail (Coase, 1988: 674–675) and corruption in general.

The role of power and the likelihood of coercion – and the emotions associated with them –distinguishes corruption from guanxi, with which it is frequently confused. This can be demonstrated by critically considering Jean Oi’s (1985) classic discussion of rural politics in China during the late 1970s and early 1980s in which she argues that guanxi operates in clientelist power relations. Oi’s account raises questions of whether the relations she discusses can meaningfully be characterized as clientelism and also whether the unequal power relations she describes are indeed guanxi? Oi (1985: 252) writes:

The Chinese terms for the cultivation of personal relationships with those in positions to control the allocation of desired […] resources is using guanxi or ganqing. Both of these expressions may be roughly translated as having ‘connections’, or a personal relationship with someone […] guanxi and ganqing are not equivalents; and neither necessarily indicates a patron–client relationship. Having guanxi or ganqing with someone can represent either a horizontal or a vertical dyadic alliance. It can be a special system of exchange between persons of equal or unequal status, and it can be clientelistic or non-clientelisitic […] Ganqing comes closer to describing a clientelist type of relationship, but it can also refer to a strong horizontal dyadic alliance.

Apart from a certain hesitancy in this statement, which comes from the fact that the ‘terms guanxi and ganqing are applied indiscriminately by Chinese to all forms of personal ties’ (Oi, 1985: 252), there is absence of reference to renqing. The ties that are held to operate in Chinese rural clientelism are primarily instrumentally opportunistic connections between persons in which consideration of norms and decorum is secondary. For these reasons it is possible to say that a form of ganqing may operate within vertical dyadic alliances or power relations. Indeed, according to one report, mutual favourable regard in ganqing may seemingly reduce class differences between landlord and tenant so that these unequal roles may be more smoothly performed (Fried, 1974: 102–109, 227). It is in this sense that Eisenstadt and Roniger (1984: 139–142) argue that ganqing underlies a range of patron–client relations in pre-revolutionary China. Do such power relations, though, characterize guanxi?

Clientelism includes a number of elements which gravitate around institutionally based power asymmetries (as in landlordism) in which the patron provides protection or support to the client (or tenant) and the client, in return, provides service and loyalty to the patron. In general terms ‘patron–client relations are based on a very strong element of inequality and of differences in power between patrons and clients [and that] the most crucial element of this inequality is the monopolization, by the patrons, of certain positions which are of crucial importance for the clients’ (Eisenstadt and Roniger, 1984: 49; see also Oi, 1985: 240, 253). While the clientelist relationship is based on power asymmetry it is also symbiotic; there is interdependence through which the ‘interests’ of each party are ‘satisfied’ in a context in which there are no feasible alternatives, particularly for the client. The ‘clientelism’ Oi (1985: 253) refers to, on the other hand, is distinguished by the fact that ‘Chinese peasants pursue clientelist politics not so much out of fear for their subsistence, but in order to live above the subsistence level’, and ‘[l]ocal patrons derive their power from their position to allocate essential goods and services rather than from private ownership of the bases of patronage’. These qualifications are arguably sufficient to indicate that rather than clientelism in Eisenstadt and Roniger’s sense, quoted above, some other form of relationship is indicated here by Oi.

In the patron–client relationship of landlordism each party deploys inherent resources, resources to which they have undisputed right; in political corruption, on the other hand, a resource is available to the patron through their management of it; it is not an inherent resource as it belongs to another, typically the state, or, in the case of cadres, the collective. Oi (1985: 246, 251, 253) relates instances of cadres playing patron with resources they manage on behalf of the collective and at its expense, which might therefore be better described as corruption rather than clientelism. A cadre’s relation with peasants is predicated on the former’s relations with the party-state. It is true, as Oi (1985: 263) indicates, that cadres and peasants typically share the same socio-economic background, but the cadres’ responsibilities to the party-state means that they represent the party to the peasant rather than the peasant to the party. Cadres cannot be patrons in the clientelist sense because they are never principals but always agents and agents of the party rather than the peasant collective.

In the period subsequent to Maoist collectivization market relations provide a setting for patron–client ties between private entrepreneurs and state officials (Sun et al., 2014). These relations are with much more justification held to operate through guanxi exchanges (Osburg, 2013: 76–112; Wank, 1996, 2002). This relationship, which can be characterized in terms of an exchange of resources between public officials and market actors, should not be confused with the clientelism described by Oi. This is because the relationship of favour between entrepreneurs and officials does not operate through chronically asymmetric structures of power but through Bourdieuian conversions of different forms of capital between distinct institutional spheres, between state and economy, officials and business operatives. This type of political patronage assumes not significant power inequality, then, but qualitatively different forms of power, of political and economic capital respectively. Significant power disparity would in these circumstances prevent guanxi being applied in this type of patronage, in which official preferment is exchanged for financial consideration.

Oi (1985: 246) in fact shows how cadres engage in corrupt favouritism rather than clientelism. Indeed, she uses the terms ‘favouritism’ when discussing the details of cadre activity (Oi, 1985: 251; see also Yan, 1996: 165–166). Cadres exploit opportunities for self-aggrandizement provided by the capacities of their administrative and political office; this is corruption not clientelism. The currency of guanxi, on the other hand, involves enhancement of reputation or face through favour exchange that is based not simply on ganqing but the normative relations of renqing including its moral precepts and code of social etiquette. These are practised in both horizontal as well as vertical dyadic exchanges. But vertical relations in guanxi are not coerced; they are based on hierarchy pertaining to social role expectation rather than political power differentials. This is not to deny that power inequality and the patronage, favouritism and corruption that are subsequent to it do not operate in China; on the contrary. Indeed, the provision of generous ‘gifts’ to cadres, who in turn reward the givers with desirable job assignments or similar exercises of favouritism, is often called by the people involved la guanxi (instrumentally cultivated guanxi) ‘and the term guanxi in this context has a more negative meaning than in its common usage’ (Yan, 1996: 165). Yan (1996: 168) holds that this ‘form of transaction is no longer gift exchange in its original sense’. But this is not because gifts must be like-for-like, or because there is an instrumental outcome (Yan, 1996: 168). Rather it is because the relations in question are based on significant power differences and (covert) coercion.

There is nothing to be gained analytically in subsuming the relations described above under the singular term guanxi, which is more properly reserved for understanding qualitatively different types of connections between persons who exchange favour in advancement of social standing or esteem and therefore which necessarily operates in terms of non-coercive and voluntary relationships, as revealed by the respective emotions involved. This is not to deny the instrumental possibilities of guanxi exchanges, which are implicit; but such instrumentalism is a function of esteem relations in guanxi, not a replacement of them (Wilson, 2002). This distinction thus empirically distinguishes guanxi from corruption in the nature of their sustaining emotions, reflecting esteem or coercion, rather than in their legality or instrumentality (Smith-Crowe and Warren, 2014).

Conclusion

The present article makes a number of contributions to both social exchange theory and to the sociology of guanxi. Indeed, this article departs from the predominant mode of treating guanxi as a variant form of social network by instead foregrounding its social exchange attributes. The treatment of social exchange here also departs from conventional approaches by providing emotions with a core explanatory role. This approach makes it possible to distinguish guanxi from bribery or corruption, a distinction blurred in practically all previous treatments of guanxi.

In its classic forms social exchange theory neglects emotions, focusing instead on operant conditioning or rational expectation. The recent ‘emotions turn’ in exchange theory regards emotion only as an intermediary variable between meso-level exchange and micro-level commitment. By developing the insight, that ‘exchange is not conducted for the sake of the object that the other person possesses, but to gratify one’s personal feelings which he does not possess’ (Simmel, 1978: 82), the determinative role of emotions in the structure and process of exchange is elaborated here in a manner not found elsewhere in the literature. It is shown above that rather than dyadic relations of trust, the staple of network analysis as well as exchange theory, emotions inherent in third-party assurance animate and shape social exchange relations. A detailed account of two emotions fundamental to guanxi, namely renqing and ganqing, which draws on an extensive empirical literature, shows that the differential contribution of each produces distinctions within the general category of guanxi.

Bribery is a form of exchange typically characterized in terms of its ethical, legal or dramaturgical properties. Guanxi is widely regarded as indistinguishable from bribery. The approach to social exchange developed in the present article, on the other hand, provides a consistent and distinctive means of distinguishing guanxi from corruption in terms of characteristic sets of emotions pertaining to each, based on esteem and coercion respectively. This is demonstrated through a detailed re-examination of a classic case study. Gift exchange and corruption are not confined to Chinese societies, of course, and the case of guanxi permits demonstration of the general relevance of arguments in Simmel, Mauss and Blau, among others, concerning the consequences of gift-giving for the formation of social esteem and a sense of obligation. The theory of social exchange and of guanxi as a form of exchange is recast and advanced in the present article, in which the elements and relations of exchange are treated in terms of particular and signifying emotions.