2.1 Risk and Commitment

Due to power-dependence inherent to exchange relations, actors face structural risk of being exploited by unknown partners. This can make actors in an exchange network more likely to interact with past partners who did not behave opportunistically rather than turning to new partners whose trustworthiness is uncertain, even if potential rewards could be higher. When this occurs, commitment between two exchange partners can develop (Cook & Emerson, 1978; Molm & Cook, 1995). Therefore, exchange relations can solidify and become cohesive through repeated interaction.

An experimental test of this hypothesis was provided by Kollock (1994), who found that intensifying information asymmetry in simulated market transactions increased subjects’ preference for long-term exchange relations as a strategy to cope with uncertainty. This mechanism then cumulatively led to more commitment and increasing mutual dependence between partners, which in turn generated stronger relational cohesion. Further tests were performed across U.S. and Japan by Yamagishi, Cook, & Watabe (1998), who similarly found that committed relations were more likely to develop between subjects who displayed a lower level of generalized trust. This confirmed that structural risk and uncertainty reduction are linked to commitment and cohesion. Following these experiments, Back & Flache (2006) showed that strategies based on commitment can be more efficient than others in coping with the risk of being exploited by free-riders in simulated game-theoretic exchanges (see also Bravo, Squazzoni, & Boero, 2012; for a review of these studies, see Bianchi & Squazzoni, 2015).

2.2 Theory of Relational Cohesion

Unlike these studies, other scholars have suggested that commitment in exchange relations could be driven by positive emotions generated by a successful exchange, rather than on strategic behaviour. A Theory of Relational Cohesion (TRC) was developed upon experimental research testing this hypothesis (Lawler & Yoon, 1993, 1996, 1998; Lawler & Thye, 1999; Lawler, Thye, & Yoon, 2000; see Thye, Yoon, & Lawler, 2002 for a review). Unlike previous conceptualizations, in TRC cohesion is defined as the perception by partners that their exchange relation is a “unifying force” in the social situation in which it occurs (Lawler & Yoon, 1996, p. 91). Similarly, commitment is then defined as the perceived strength of the tie existing between a person and a social unit, such as relations, groups, networks, organizations, or communities (Thye, Yoon, & Lawler, 2002).

The first block of TRC is a causal model developed by Lawler and Yoon (1993), which elaborated on behavioural aspects of commitment. In the model, the generation of positive emotions acted as an intervening mechanism between the structure of power-dependence in an exchange relation and behaviourally-observed commitment. As a condition, the model posits a process of negotiation of an agreement between partners with equal relative power. When this happens repeatedly, reaching agreements generates actors’ gratitude. Then, satisfaction for a successful exchange brings actors to eventually develop commitment. First experimental tests showed that subjects negotiating exchange terms face-to-face within an equal power-dependence relation were more likely to display committing behaviour. For example, not only did they tend to commit to a partner despite more profitable alternatives, they also were more prone to concessions during negotiation.

Furthermore, other experiments confirmed the importance of positive emotions, which were generated by the joint activity of negotiating a successful agreement. Moving from the premise that humans tend to attribute responsibility for their emotions to identifiable sources, these experiments showed that successful exchanges allowed subjects to attribute their satisfaction to their partners’ behaviour. This eventually generated relational cohesion, displayed through increased commitment and willingness to engage in collective-oriented behaviour (Lawler & Yoon, 1996). Later, new experiments compared the effect of different network structures on exchange. Unlike research on typical situations of balanced power-dependence, results showed that greater exchange frequency in unequal power relations was not sufficient to produce commitment, because it could not elicit positive emotions. In turn, shared group identity also displayed a positive effect on commitment (Lawler & Yoon, 1998). Finally, emotional enforcement of commitment in exchange relations was experimentally disentangled from strategic behaviour of uncertainty reduction (Cook & Emerson, 1978; Kollock, 1994). This suggested that the two mechanisms, rather than concurrent, might rather be complementary depending on different social situations (Lawler, Thye, & Yoon, 2000).

In conclusion, studies in TRC provide evidence of an emotion-based mechanism generating cohesion in an exchange network (Blau, 1964; Thye, Yoon, & Lawler, 2002). Furthermore, this mechanism is likely to trigger stable group cohesion because it generates affective and integrative bonds within an exchange network, differently from commitment driven by strategies of uncertainty reduction (Cook & Emerson, 1978; Kollock, 1994).

However, generalizability of TRC is considerably limited. First, the postulated mechanism was tested in cases of repeated interaction between individuals who jointly negotiated the exchange terms. Therefore, TRC does not apply to most types of social exchanges, where actors rarely negotiate exchange terms. Moreover, framing an exchange within the terms of an agreement leaves out most common social exchanges involving unilateral transfer of benefits from one individual to another (e.g., voluntary provision of help). Furthermore, TRC assumes equal-power exchange networks, which minimize actors’ temptations of opportunism (Lawler & Yoon, 1998), thereby allowing actors to perceive the social situation in a more cooperative frame.

3.1 Affect Theory of Social Exchange

Further studies building on TRC looked at the effect of different forms of exchange on cohesion and solidarity (Lawler, 2001; Lawler, Thye, & Yoon, 2008). By relying on previous experimental evidence, Lawler (2001) proposed an Affect Theory of Social Exchange (ATSE). The theory moves from the assumption that actors engaging in joint social activities attribute their perceived emotions to the social units in which the activity occurs (social relationships, networks, or groups). This would account for variations in actors’ affective attachment to higher-level social formations.

This theory aims to study the mechanisms which cause the emergence of a micro social order, defined as “a recurrent pattern of interaction among a set of actors, from which they come to perceive themselves as a unit (i.e., a group) and to develop feelings about that unit” (Lawler, 2002, pp. 4–5). It is interesting to note that this concept comprises what could be conceived as interpersonal bonds of solidarity. Moreover, the definition of a “micro social order” appears as a pre-condition as to how Lawler (2001) defined “solidarity,” namely as “the strength and durability of person-to-group and person-to-person relations” (p. 329). In this way, solidarity becomes something more than just integrative interpersonal bonds. Following Parsons (1951), solidarity is conceptualized here in two dimensions: a) affective ties between group members, b) a relation between each group member and the collectivity as a social object.

The strength of group attachment implied by a micro social order can determine solidary behaviour, such as providing unilateral benefits without expecting reciprocation, increasing one’s availability to collaboration despite risk of exploitation, forgiving isolated free-riders, and commitment (Lawler, 2001). Another basic premise of the model is that each form of exchange requires different types of joint activity, which vary along a structural and a cognitive dimension: a) the degree of separability of actors’ tasks and contributions; b) the perception of shared responsibility on exchange outcomes. Hence, increasing non-separability should generate stronger perception of shared responsibility, which in turn should make actors attribute positive emotions to the exchange relation as a social unit. Therefore, ATSE predicts that different forms of social exchange generate micro social orders of different strength, as they entail different degrees of task separability and perceived shared responsibility. More specifically, stronger micro social orders would be produced by direct rather than indirect exchange and by negotiated rather than reciprocal exchange.

These predictions were tested by comparing different forms of three-actor exchange networks in laboratory experiments (Lawler, Thye, & Yoon, 2008). The strength of generated micro social orders was measured by four indicators: 1) frequency of exchanges; 2) positive feelings about exchanges; 3) perceptions of network cohesion; 4) affective attachment about the social unit. However, results only partially supported the theory. As predicted by ATSE, subjects in direct exchange reported higher levels than those in indirect exchange across all indicators. Interestingly, perceptions of network cohesion decreased over time among indirect exchange partners. However, no significant differences between negotiated and reciprocal exchange were observed in the experiment.

3.2 Criticism to Affect Theory of Social Exchange

As seen above, ATSE sought to expand the scope of TRC by hypothesizing the occurrence of an emotion-based mechanism in the context of non-economic exchanges. However, experimental results were not entirely supportive. Furthermore, other scholars challenged general implications of ATSE by shifting the analytical focus back to structural risk and further developing its relationship with trust and commitment (Molm, Peterson, & Takahashi, 1999, 2001; Molm, Takahashi, & Peterson, 2000; Molm, 2003; Molm, Schaefer, & Collett, 2009).

In order to emphasize risks in exchange, Molm, Takahashi, & Peterson (2000) concentrated the analysis on direct exchanges in unequal-power relations. Their hypothesis was grounded on the classic SET proposition that a necessary condition for an actor to initiate a social exchange interaction with no binding assurance structures (e.g., the provision of help or advice) is the actor’s trust of being reciprocated (Blau, 1964). Defined later as the “structural or situational potential for incurring a net loss” (Molm, Schaefer, & Collett, 2009, p. 5), risk is in turn a necessary condition for trust to develop between exchange partners, as it provides actors the opportunity to prove their trustworthiness through reciprocation (Dasgupta, 1988; Gambetta, 1988; Hardin, 2002; see also Kollock, 1994; Yamagishi, Cook, & Watabe, 1998).

Consequently, higher risk of non-reciprocation in reciprocal exchange should generate more trust than negotiated exchange, as in the latter partners can rely on a previously negotiated agreement that ensures reciprocation. Risk connected to negotiated exchange is then only related to the possibility of failure in establishing an agreement, which is less threatening than the one implied by the lack of assurance of reciprocation in reciprocal exchange (Molm, 2003). Therefore, if actors succeed in establishing an agreement, trust between actors can be rather unnecessary to yield a positive outcome in a negotiated exchange, as this is ensured by the assurance structure provided by the agreement (Yamagishi & Yamagishi, 1994).

In an experimental test of these predictions, Molm, Takahashi, & Peterson (2000) found stronger self-reported levels of trust and affective commitment towards partners in reciprocal exchanges. They argued that while negotiated exchanges provided actors the opportunity to display only behavioural signs of commitment (Kollock, 1994), the risk of non-reciprocity in reciprocal exchanges allowed subjects to learn about their partners’ trustworthiness once reciprocation occurred. Reciprocation was then attributed to partners’ intrinsic motivations rather than to strategic compliance to exogenous enforcement structures. This generated affective commitment together with trust. Interestingly, higher risk in reciprocal exchange could even bring actors to strategically commit to partners more than in negotiated exchanges. In an earlier study contrasting power and exchange forms, Molm, Peterson, & Takahashi (1999) showed that subjects with higher structural power in exchange networks (i.e., more opportunities to shift to other partners) disproportionately opted for reciprocity and were less likely to explore other partnerships in reciprocal rather than in negotiated exchanges, in order to cope with higher risk. Finally, further experiments reported the importance of the expressive value of reciprocation in itself, as it was perceived by actors as a sign of affective regard beyond trustworthiness (Molm, 2003; Molm, Collett, & Schaefer, 2007; Molm, Schaefer, & Collett, 2007).

In conclusion, while not directly questioning the occurrence of emotionally-driven commitment between economic exchange partners, these studies cast serious doubts on the possibility of economic exchanges to generate solidarity as claimed by ATSE, at least if partners had different power in the network. Molm (2003) argued that the same properties of economic exchange which lubricate cooperation — i.e., joint decision-making, bilateral flow of benefits, the relationship created by successful negotiation — might as well exacerbate conflict-related aspects linked to partners’ diverging interests. While Lawler (2001) suggested that perception of shared responsibility reduces competition bias in negotiated exchanges, the mere fact of interacting in the frame of an agreement prevents actors from inferring trustworthiness and developing affective commitment for exchange partners (Molm, 2003).

In sum, TRC and ATSE provide important insights to explain the formation of solidarity from economic exchanges. However, its scope seems to be limited to exchange networks where power differences between actors are not relevant, i.e. where competition between partners is not salient. In this context, successful negotiated exchanges can trigger an emotional mechanism that enhances solidarity between partners besides trustworthiness (see also Barrera, 2007). However, it is worth noting that most market-based economic exchanges in real social settings occur in situations of resource asymmetry and competition.

3.3 Reciprocity Theory of Solidarity

Building on criticism to ATSE, Molm, Collett, & Schaefer (2007) elaborated a more comprehensive view on the relationship between exchange forms and solidarity, which is called Reciprocity Theory of Solidarity (RTS). By extending previous results on risk and trust (Molm, Peterson, & Takahashi, 1999; Molm, Takahashi, & Peterson, 2000; Molm, Collett, & Schaefer, 2006; Molm, Schaefer, & Collett, 2007), RTS predicts varying effects of exchange forms not only on actors’ dyadic commitment, but on “the integrative bonds that develop between persons and between persons and the social units to which they belong” (Molm, Collett, & Schaefer, 2007, p. 207), as they defined solidarity. In this way, the aim of RTS is very close to how ATSE claims to explain the emergence of “micro social orders.” However, unlike ATSE, RTS argues that: a) reciprocal exchange generates higher solidarity than negotiated exchange; and b) generalized exchange generates higher solidarity than direct exchange.

The model assumes that each exchange form entails a different “structure of reciprocity,” defined by two dimensions: a) whether benefits are reciprocated directly or indirectly; and b) whether benefits flow unilaterally or bilaterally (Emerson, 1981; Molm & Cook, 1995; Molm, 2003). These variations are responsible in the model for the generation of three causal mechanisms: 1) the “risk of non-reciprocity;” 2) the “expressive value of reciprocity;” 3) the “salience of conflict” (Molm, Collett, & Schaefer, 2007, p. 211). As a first consequence, the risk of non-reciprocity is higher in reciprocal than in negotiated exchange (see also Molm, Takahashi, & Peterson, 2000; Molm, 2003). Moreover, generalized exchange is more risky than direct reciprocal forms, because actors are dependent on the actions of multiple others, such that risk increases proportionally to the length of the exchange chain. Therefore, the production of trust should also be higher in these contexts.

Secondly, reciprocation allows actors to infer the expressive value of the exchange because it can more easily be attributed to the partner’s intrinsic motivations rather than compliance to a binding agreement. This is magnified by the unilateral character of benefits’ flow, as an indefinite time gap occurs between initial transfer and reciprocation (see also Molm, 2003; Molm, Schaefer, & Collett, 2007). This is even stronger in generalized rather than direct reciprocity, given that the former does not imply even a tacit obligation.

Finally, the conflict between exchange partners is less salient in reciprocal than in negotiated exchanges, as in the latter the competitive nature of splitting resources is underscored by negotiation (see also Molm, Collett, & Schaefer, 2006). This conflict can be even less salient in generalized exchanges, where no direct relation between incoming and outgoing resources is established.

An experimental test of RTS was run by comparing all three forms of exchange (negotiated, direct reciprocal and chain-generalized) in both three- and four-actor exchange networks. This was to disentangle potential confounding effects of network size and power, according to the typical setting of SET laboratory experiments (see also Molm & Cook, 1995). Molm, Collett, & Schaefer assessed affective components of solidarity by measuring self-reported feelings of: a) trust, i.e., “the belief that the exchange partner will not exploit the focal actor;” b) affective regard, i.e., “positive feelings and evaluation towards the partner;” c) “the perception of the exchange relation as a social unity, with actors united in interests and purpose;” d) “feelings of commitment to the partner and the relationship” (2007, p. 207). Experimental results confirmed theoretical predictions, as subjects reported: a) higher feelings of solidarity in reciprocal than in negotiated exchanges; b) higher feelings of solidarity in generalized than in direct reciprocal exchanges.

To summarize, Molm, Collett, & Schaefer (2007) and Lawler, Thye, & Yoon (2008) provided conflicting experimental results on the effect of forms of exchange on solidarity. However, a closer look at the differences between their experimental designs reveals that the validity of findings depended on certain structural characteristics of the network where exchanges occurred.

Results reported by Molm, Collett, & Schaefer (2007) do not only support RTS, but also directly challenge the validity of ATSE. This is because subjects engaged in exchanges within equal-power networks, i.e., no subject was more advantaged than others in terms of availability of alternative partners. Although previous experiments on risk and trust in direct exchange were performed with varying levels of power inequality (Molm, Takahashi, & Peterson, 2000; Molm, Peterson, & Takahashi, 2001), experiments by Molm, Collett, & Schaefer (2007) were designed so that they could be compared with predictions by Lawler (2001), which were conditional to equal-power networks. Furthermore, equal-power networks provided a conservative test for RTS, because the hypothesized mechanisms should have even stronger effects in unequal-power networks.

Moreover, experimental designs by Molm, Collett, & Schaefer (2007) and Lawler, Thye, & Yoon (2008) differed in one important feature, which anyway weakens the scope of ATSE. While in the former study direct exchanges occurred in negatively-connected networks — i.e., subjects could choose only one exchange partner at a time (Molm & Cook, 1995) —, the latter experiment was set in null-connected networks, where actors could simultaneously engage in direct exchange with multiple partners. As noted by Lawler, Thye, & Yoon (2008), conflict is less salient in null-connected networks, because they do not necessarily imply exclusion. Unlike direct exchanges, generalized exchanges always imply the transfer of resources to one actor at a time. This discrepancy could have made potential conflicts in direct exchange substantially less salient in the experiment by Lawler, Thye, & Yoon (2008), thereby producing stronger solidarity in direct than generalized exchange forms.

3.4 More Recent Developments

In conclusion, experimental research in Social Exchange Theory provides relatively discouraging results about the capacity of economic exchange to generate solidarity. Reported evidence suggests that interpersonal transactions framed by previously negotiated agreements can generate feelings of solidarity only under strict conditions, i.e. if partners are similar in terms of bargaining power and cannot simultaneously engage in similar exchanges with others. Otherwise, structural conflict between the partners’ interests would prevent the development of feelings of solidarity. Considering that most market-based transactions occur in contexts of power imbalances and competition, we might conclude that economic exchanges in market-based economies structurally hinder the development of interpersonal bonds of solidarity. At the same time, the above presented research claims that feelings of solidarity need trust to establish between interaction partners, which in turn is possible as long as the partners can experience each other’s trustworthiness through successful risky exchanges (Molm, Collett, & Schaefer, 2007).

Further studies seem to confirm this picture, although in a more nuanced way. Barrera (2007) found that repeated successful negotiated exchanges could generate interpersonal trust in a laboratory experiment where subjects played an Investment Game (Berg, Dickhaut, & McCabe, 1995) with previous exchange partners. However, displayed trust seemed to be related to subjects’ individual characteristics than to the partners’ perceived trustworthiness. Moreover, experiments run by Kuwabara (2011) showed that negotiated exchanges could generate relational cohesion as subjects reported feelings of belonging to a relationship with their partners only if negotiation was framed as a cooperative task by the experimental design. On the contrary, subjects were more inclined to frame the exchange as an opportunity for conflict when they had to negotiate within a zero-sum game or in a clear trustor-trustee relationship. Anyway, cooperative forms of negotiated exchanges could not generate trust along cohesion. Despite providing more context-dependent outcomes, these studies have substantially confirmed the difficulty of economic exchanges to generate interpersonal bonds of solidarity.

However, while it surely points at limitations of Social Exchange Theory, recognizing “context-dependence” of highly abstract concepts of economic exchange is not enough to advance our understanding of their consequences. If economic exchanges inhibit the generation of solidarity because risk is substantially suppressed, then it is useful to disentangle the link between economic exchange and the assurance structures which manage risk. In this vein, Molm, Schaefer, & Collett (2009) have studied trust generated by negotiated exchanges with non-binding agreements and by reciprocal exchanges where partners were allowed to communicate. Although these experiments showed that reciprocal exchange produced a more resilient and affect-based form of trust, this form of negotiated exchange could generate interpersonal trust, albeit cognition-based and “fragile.”

Most importantly, these studies show that relaxing the boundaries between different forms of exchange allows to integrate experimental research with observational studies in order to look at economic exchanges in actual social contexts. Following this approach, Bianchi, Casnici, & Squazzoni (2018) have analyzed patterns of solidarity within a network of independent professionals who occasionally partnered on business projects by outsourcing parts of their work within the frame of informal or loose agreements. Results showed that successfully partnering in business was not sufficient for subjects to expect solidary behaviour, such as support in need situations, from their partners. On the other hand, support was likely expected from those professionals whom were considered as trustworthy partners in risky business, which in turn was predicted by previous successful collaborations. In sum, expectations of solidarity were not directly generated by economic exchange relations, although they were mainly framed by risky informal agreements. Instead, solidarity was generated as a by-product of the trust elicited by: a) direct experience in collaborations; b) indirect information on trustworthiness circulating within the network. To summarize, these results first suggest that interpersonal economic exchanges can indeed generate trust if they are framed by loose non-binding agreements (as also in Molm, Schaefer, & Collett, 2009). More importantly, this study showed that partners in successful economic exchanges who get to trust each other for risky business can develop bonds of solidarity.