“Both enormous trust and enormous malfeasance, then, may follow from personal relations.”

(Granovetter, 1985, p.492)

Boundary spanner corruption as a dark side complication of marketing relationships threatens the well-being of organizations, their representatives, and business-to-business (B2B) relationships. We define boundary spanner corruption as voluntary collaborative behavior of individuals representing different organizations that violates their own organizations’ norms. This can occur in any business relationship, but for illustrative purposes we focus our discussion primarily on channel relationships between customer and seller organizations. Boundary spanner corruption can occur between any counterparts from firms in marketing relationships—frontline employees, such as salespeople and buyers; mid-level supervisors, such as sales and procurement managers; or executives. It ranges from deviance that violates only organization norms (e.g., personal favors, excessive social entertaining and gift-giving, or unreasonable discounts) to more severe malfeasance that also contravenes norms, laws or regulations of the broader society (e.g., bribery, bid-rigging, fraud, and embezzlement implemented through boundary spanner collusion). A 2021 survey of 1336 senior executives found nearly half felt vulnerable to corruption in business relationships (Kroll, 2021). The World Bank estimates that bribery alone amounts to approximately 2% of annual global GDP (Kaufmann, 2005). Evidence like this suggests that it is the rare firm whose representatives have never been solicited to engage in corruption; it is highly likely that some boundary spanners in nearly every firm have been tempted to—or engaged in—corrupt activities.

Organizations whose boundary spanners engage in joint corruption can pay a very high price when violations of societal norms are exposed publicly. Fines, criminal charges, and tort litigation regularly follow public exposure, imposing often substantial costs for the firms. Even in the absence of such consequences, scandals about nefarious collusive activities often have deleterious effects on reputation and/or sales, which negatively impact stock market valuations (Wei et al., 2017). When it was revealed that managers of GlaxoSmithKline’s Chinese subsidiary bribed doctors to increase sales (The Wall Street Journal, 2014), the company suffered an 18% decline in sales in China the next year (GSK, 2016). Although a few major scandals make headlines, smaller, garden-variety corruption occurs much more frequently and is more prevalent in marketing relationships. Public knowledge about any form of boundary spanner corruption damages the firms. For example, personal favors or excessive gifts for one individual may encourage colleagues or representatives of other business partners to seek the same arrangement. Alternatively, publicity may result in the termination of marketing relationships with other partner firms that seek to deter boundary spanner corruption or the appearance of impropriety. Costs also accrue before public exposure and, indeed, even if the deviance is never publicly revealed. Boundary spanner corruption inherently involves the misallocation of resources, a sub-optimal decision, and opportunity costs for at least one—and often both—of the involved organizations; it also generates enhanced latent financial risk for both firms (Palmatier, Scheer, & Steenkamp, 2007b).

Despite the economic and reputational damage that boundary spanner corruption can wreak, academic research about this phenomenon in the marketing and sales literature is limited, arguably due to the topic being notoriously difficult to study (Aven, 2015). Although the potential dark side of close business relationships has long been acknowledged (Grayson & Ambler, 1999; Moorman et al., 1992), boundary spanner corruption in interorganizational relationships has been overlooked. Similarly, the sales and services literature has examined various types of individual-level deviance (e.g., Brady et al., 2012; Ramaswami, 1996), but surreptitious collusion between boundary spanners representing different firms has not been specifically studied.

We draw insights from marketing research on both B2B relationships and deviance as well as from corruption research in developing our conceptual framework and propositions. Building on Fang et al.’s (2008) multi-level trust framework, we develop a model of boundary spanner corruption that integrates the triumvirate of interpersonal trust, intraorganizational agency trust, and interorganizational trust with the concept of the social cocoon. A social cocoon is “a micro culture created within a group where the norms may be very different from those valued by society or even the wider organization” (Anand et al., 2004, p.46). We contend that interpersonal trust between individuals representing different firms in a B2B relationship can enable the development of a social cocoon. We theorize that a boundary-spanning social cocoon is particularly susceptible to developing corrupt behavior when it is accompanied by simultaneous intraorganizational agency trust in the representatives and interorganizational trust, as these factors weaken control and monitoring of trusted representatives. A social cocoon increases the likelihood that any motivation to engage in deviance arising from individual-level or intraorganizational deviance-inducing factors will be manifest in boundary spanner corruption.

Our research makes several contributions. We break new ground by conceptualizing the phenomenon of boundary spanner corruption, a hitherto neglected phenomenon in the sales, marketing, and B2B literatures. We advance marketing theory by introducing the concept of the boundary-spanning social cocoon, which illustrates the mechanisms through which many of the same elements that generate benefits in B2B relationships can foment corrupt activities. We extend Fang et al.’s (2008) framework by integrating deviance-inducing factors that demonstrate how trust creates vulnerability to boundary spanner corruption. Our conceptual framework thus addresses calls for multi-level modeling of the dark side of business relationships (Oliveira & Lumineau, 2019). We illustrate how even a firm that successfully manages internal deviance-enabling factors and curbs intrafirm deviance is still at risk of its boundary spanners forming social cocoons with counterparts from firms less effective in squelching potential deviance.

We begin by elaborating on the nature of boundary spanner corruption, contrasting the phenomenon with the existing literature on deviance in sales and service organizations and the dark side of close relationships, with particular emphasis on the vulnerability generated by trust. Drawing on these literatures, we present a multi-level conceptual model of boundary spanner corruption that integrates concepts of trust and the social cocoon. In conjunction with insights from a qualitative study of 25 marketing and purchasing executives, we develop theoretical propositions. We conclude by providing directions for future research on boundary spanner corruption and offering important factors for managerial consideration.

Explicating and exploring boundary spanner corruption

Identifying characteristics of boundary spanner corruption

Boundary spanner corruption has several essential identifying characteristics, some of which are found in previous literatures on the dark side of business relationships and deviance in sales and service organizations, and some of which distinguish boundary spanner corruption from those extant literatures. Our conceptualization of boundary spanner corruption exports and adapts individual-level deviance (i.e., contrary to organizations’ norms; voluntary; threatens the well-being of organizations) to the interorganizational context (i.e., collusive deviance involving representatives of two or more organizations who surreptitiously collaborate across organizational boundaries). We draw examples from news coverage to support the identifying characteristics of boundary spanner corruption.

First, boundary spanner corruption is undertaken by representatives of two or more organizations who collaborate across organizational boundaries. By definition, boundary spanner corruption excludes individual deviant activities. It can involve frontline representatives, managers, executives, or other any individuals representing any type of organization in any type of vertical or horizontal interorganizational relationship. For example, representatives of Turner Construction paid Bloomberg executives $15 million in bribes and kickbacks for systematically rigging bidding processes in Turner’s favor and purchasing $240 million of more expensive, higher-commission products and services (Associated Press News, 2018). In the LIBOR scandal, traders from Bank of America, Barclays, Citigroup, JP Morgan Chase, Royal Bank of Scotland, and UBS colluded in fraudulent manipulation of the London Interbank Offer Rate (The Telegraph, 2015); an investigation revealed that requests for rate manipulation were reciprocated with small personal favors and gifts (The Economist, 2012).

For simplicity and clarity in exposition, we focus our discussion primarily on the seller-customer B2B relationship, the type of interorganizational relationship most frequently examined in marketing. Similarly, we focus on the salesperson-buyer interpersonal relationship within the B2B relationship, as the nature of these ongoing, longer-term interactions usually offer greater opportunity for corruption compared to other boundary-spanning relationships.

Second, the boundary spanners’ collaborative behavior is contrary to their organizations’ norms, contravenes formal and informal organizational policies and procedures, and violates their agency obligations to their organization-principals (Robinson & Bennett, 1995). If a firm’s representatives are encouraged, promoted, or required to participate in corruption by their organization (Pinto et al., 2008), those sanctioned collusive activities are consistent with deviant organization norms and thus constitute organizational corruption but not boundary spanner corruption. The top management of neither Bloomberg nor Turner Construction were involved in the elaborate pay-to-play scheme engineered collusively by their representatives. When the corruption was publicly exposed, Turner stated: “Turner rejects the conduct alleged against two former employees as an absolute betrayal of Turner’s core values of integrity, which are followed by the 9,000 Turner employees who work hard, honestly and well every day” (Associated Press News, 2018). The collaborative manipulation of LIBOR rates explicitly violated norms of the traders’ banks. Citibank declared that the revelation was “an embarrassment to our firm, and stands in stark contrast to Citi’s values” (ABC News, 2015).

Third, the corrupt behavior is voluntary. One representative proposes an action that violates their agency obligations; the boundary spanner counterpart agrees voluntarily to engage in the activity, which also violates the partner organization’s norms. If an individual is coerced or forced to commit deviant acts (Elangovan & Shapiro, 1998; Robinson & Bennett, 1995), it is not boundary spanner corruption, which by definition excludes blackmail and all other involuntary deviant activities. In boundary spanner corruption, although lured or tempted, either individual can decline to participate without experiencing personal negative consequences. No one forced Bloomberg employees to accept Turner Construction’s bribes. There may be professional consequences, however, such as when refusal to participate results in the loss of a customer, with a concomitant negative impact on the salesperson’s compensation. For this reason, some firms limit personal negative consequences if a solicited bribe is declined and reported. Often, however, the voluntary nature of collusive corrupt behavior is embedded in the personal relationship between boundary spanners; traders in the LIBOR scandal had close, jovial relationships in which rate manipulations were based on personal reciprocity and the provision of favors (The Economist, 2012).

Fourth, the representatives’ behavior threatens the well-being of their organizations, even if one or both organizations benefit in some manner. For example, when a buyer signs a major contract due to a bribe or personal favor from the salesperson, the customer purchases goods that it otherwise would not have acquired. This imposes on the customer greater costs, poorer quality, or other opportunity costs. The seller may potentially benefit, if the illicitly acquired sales revenues could not have been obtained cleanly from another customer. Despite any short-term benefits, however, both involved firms incur significant latent financial risk (Palmatier, Scheer, & Steenkamp, 2007b) associated with the potential future discovery and exposure of the corruption. The Turner Construction scandal undermined customer satisfaction and severely endangered future business industry-wide. “[It is] a shame that a few rogue employees may have acted in a non-compliant and criminal manner and tarnished the image of the companies they worked for and the industry as a whole” (The Real Deal, 2018). Banks involved in the LIBOR scandal were fined a total of $6 billion, leading a top executive to remark: “The lesson here is that the conduct of a small group of employees, or of even a single employee, can reflect badly on all of us, and have significant ramifications for the entire firm” (ABC News, 2015).

These identifying characteristics highlight the multi-level nature of boundary spanner corruption—an interpersonal relationship between representatives of two firms in an interorganizational relationship leads to activities that violate the intraorganizational norms of each firm. A full understanding of boundary spanner corruption therefore necessitates consideration of factors at multiple levels within the relationship.

Insights from research on deviance in sales and service organizations

Deviant, self-serving intraorganizational behaviors are similar to boundary spanner corruption in that they, too, involve norm violations vis-à-vis one’s own firm. Research on sales and service has identified several types of deviance: organization-targeted deviance such as manipulating control systems through shirking or expense-padding (Ramaswami, 1996), colleague-targeted interpersonal deviance such as sabotaging co-workers (Jelinek & Ahearne, 2010), and customer-related deviance such as service sweethearting (Brady et al., 2012) or service sabotage (Harris & Ogbonna, 2006). This literature suggests that dysfunctional salesperson activities are motivated and enabled by both intraorganizational factors and individual factors (Brady et al., 2012; Robertson & Anderson, 1993).

Intraorganizational deviance-inducing factors include triggers that motivate deviant intentions and opportunities that arise from failure to block or suppress those intentions from becoming manifest in behavior (Fox et al., 2001; Marcus & Schuler, 2004; Robinson & Bennett, 1995). Triggers include factors such as perceived inequity of sales bonuses or unfair allocation of sales leads; opportunities can result from boundary spanner autonomy or organizational latitude regarding norm compliance. Furthermore, some representatives will have greater individual-level deviance-inducing propensities, such as competitiveness and risk-seeking, and lack of internal controls to suppress deviance, such as low self-monitoring or lack of moral obligation (Marcus & Schuler, 2004). These deviance-inducing intraorganizational and individual factors provide insight into the conditions that induce a boundary spanner to consider or desire to pursue deviant activities, which is the first step toward boundary spanner corruption. (See Table 1 for an overview and Web Appendix A for details).

Table 1 Key literature on deviance in sales and service organizations
Full size table

However, the deviance examined in the sales and service literature focuses almost entirely on individually motivated and executed acts. This research offers limited insights regarding how deviant intentions become manifest in cross-firm collusive deviance. The research on service sweethearting is an exception, as it examines a frontline employee’s collusion with a consumer to provide unauthorized discounted goods in return for a reciprocal personal or financial benefit (Brady et al., 2012). In service sweethearting, the frontline employees violate their organization’s norms, but the involved consumers act on their own behalf. In boundary spanner corruption within a B2B relationship, both involved individuals are agents of their respective firms and, by definition, violate the norms of both firms.

In conclusion, the sales and service research on deviant behaviors provides helpful insights regarding deviance-inducing intraorganizational and individual factors and how the discontent can manifest in individually-executed deviant acts. However, it sheds little light on the interorganizational factors or interpersonal considerations that lead representatives of different organizations to engage in more complex, collaborative deviance. Seeking insight on these factors, we turn to the literature on dark side behaviors in interorganizational relationships.

Insights from research on the dark side of interorganizational relationships

Research on the dark side of close business relationships distinguishes between intentional behaviors such as opportunism (Rindfleisch & Heide, 1997) and unintentional dysfunctional behaviors such as those arising from cognitive biases (Villena et al., 2011). We limit our review to intentional behaviors. Like other intentional dark side behaviors (Wuyts & Geyskens, 2005), boundary spanner corruption is voluntary and detrimental to organizational well-being. Among dark side behaviors, we focus primarily on opportunism, as it shares key characteristics with boundary spanner corruption. Those superficial similarities are misleading, however.

Research on opportunism in interorganizational relationships typically examines surreptitious actions by one firm or representative to take advantage of a partner firm or individual (Rindfleisch & Heide, 1997). Opportunism is similar to boundary spanner corruption as it occurs within a B2B relationship, entails surreptitious behavior, and typically involves frontline representatives. However, the differences are highly significant. Opportunism research in marketing has focused on activities that benefit the focal firm at the expense of a partner firm; representatives that engage in opportunism typically do so with the tacit or explicit approval of management. In addition, representatives that engage in opportunism do not collude with counterparts at other firms, but often mislead, misdirect, obfuscate, or lie to counterparts. Thus, the literature on intentional dark side behaviors in B2B relationships primarily addresses opportunism against a partner firm. In contrast, boundary spanner corruption involves collusive activity between individuals representing different firms in which opportunistic actions such as secrecy and deception are directed against their own firms. Such boundary spanner corruption has been overlooked in the marketing literature. (See Table 2 for an overview and Web Appendix B for details).

Table 2 Key literature on the dark side of business relationships
Full size table

In research on intentional dark side behaviors, trust has been recognized as a pivotal factor (Villena et al., 2019; Wang et al., 2013). Trust, “the belief that one’s partner can be relied on to fulfill its future obligations and to behave in a manner that will serve the firm’s needs and long-term interests” (Scheer & Stern, 1992, p.134), has been examined extensively in the relationship marketing literature. Trustors exhibit decreased desire, will, and ability to monitor highly trusted relationship partners, diverting scarce resources to monitor those less trusted. However, decreased monitoring of highly trusted partners makes the trustor inherently more vulnerable to intentional violations of trust (Wang et al., 2013), malfeasance (Dyer & Singh, 1998; Gulati & Nickerson, 2008) and opportunism (Anderson & Jap, 2005; Villena & Craighead, 2017). Trust thus provides the essential discretionary space that enables surreptitious activities to go undetected. Intentional dark side behaviors are often explicitly associated with a violation of trust, such as salesperson opportunism violating the buyer’s interpersonal trust. Boundary spanner corruption is similarly a violation of trust but differs from the extant literature with regard to whose trust is violated. Boundary spanner corruption requires interpersonal trust between salesperson and buyer; the salesperson violates the seller firm’s agency trust and simultaneously the buyer violates the customer firm’s agency trust.

Although the literature has acknowledged trust as a key factor in enabling dark side behaviors, discussions often meld trust at different levels or simply fail to distinguish the specific levels of trust. Even when the focal trustor-trustee are identified, there is little investigation of the simultaneous or interactive effects of trust at multiple organizational levels, with intraorganizational trust being particularly neglected. This vague conceptualization of the dark side of trust in the extant literature may help explain why boundary spanner corruption, a multi-level phenomenon, has not yet been investigated.

Integrating insights from literature

To summarize, research on deviance in sales and service organizations has focused on intraorganizational and individual considerations that motivate individually-executed deviant actions (Ramaswami & Singh, 2003). Investigations of the dark side of business relationships have considered interorganizational relationships and acknowledged the role of trust, but have largely neglected differences between interorganizational and interpersonal levels (Oliveira & Lumineau, 2019). Nor does the dark side literature capture intraorganizational norm violations. Neither literature stream has directly examined boundary spanner corruption. Although each stream provides insights on some factors that contribute to boundary spanner corruption, neither paints a complete picture regarding explanatory mechanisms. We contend that to understand the drivers and enablers of boundary spanner corruption, a multi-level approach is required. We offer a multi-level framework that draws insights from these literatures to elaborate on the phenomenon of boundary spanner corruption.

A multi-level framework of boundary spanner corruption

The role of trust at multiple levels

We adapt Fang et al.’s (2008) multi-level trust framework to explicate how boundary spanners’ behavior is impacted by interpersonal relationships between representatives of different firms, the intraorganizational relationships, controls, and practices that exist internally within each firm, and the interorganizational relationship between the firms. (See Fig. 1.) We theorize that simultaneous trust at all three levels plays a role in the development of boundary spanner corruption.

Fig. 1
figure 1

A multi-level model of trust and boundary spanner corruption

Full size image

Boundary spanner corruption requires mutual interpersonal trust between representatives of involved firms, but this alone is not sufficient. Interpersonal trust provides opportunity for the development of a social cocoon, which may under certain conditions generate boundary spanner corruption (Anand et al., 2004). A boundary-spanning social cocoon is a microculture within the representatives’ relationship that provides the discretionary space for the development of intra-cocoon norms that may conflict with cocoon members’ organizational norms (Anand et al., 2004; Ashforth & Anand, 2003; Aven, 2015). While we focus on single representatives on each side of the B2B relationship, a social cocoon may involve multiple representatives in each firm.

Close salesperson-buyer relationships and accompanying interactions allow a social cocoon to develop through encapsulation. The social cocoon can create and sustain a new set of norms that deviate from those of the seller and customer firms. As cocoon-specific norms become more dominant, a slippery slope towards boundary spanner corruption evolves. If deviance-inducing factors in the intraorganizational task environment or at the individual level generate deviant intentions in either the salesperson or buyer, the social cocoon provides the opportunity for collaboration with the counterpart. Although boundary spanners appear to be faithful representatives of their firms, in reality, they collaborate to engage in surreptitious behavior that violates their firms’ explicit or implicit norms (Zaheer et al., 1998).

The formation and strengthening of a boundary-spanning social cocoon is facilitated as greater intraorganizational agency trust is placed in each representative by their respective managers. Agency trust makes the firm more vulnerable to exploitation, as it weakens the imperative to develop and use intraorganizational controls to detect potential corrupt behaviors (Gargiulo & Ertug, 2006) and enables risk-taking by the boundary spanner (Mayer et al., 1995; Ring & Van de Ven, 1994; Rousseau et al., 1998). Mutual interorganizational trust similarly encourages the weakening of controls in relationships with trusted partner firms (Gulati & Nickerson, 2008; Wuyts & Geyskens, 2005).

Given the dearth of directly relevant empirical research on boundary spanner corruption, we sought to evaluate and enrich our theoretically derived multi-level framework with qualitative field interviews. These interviews were catalysts for refinement and exemplification (Grewal et al., 2007; Workman Jr et al., 1998) and provided additional insight as we developed propositions.

Methodology of qualitative field interviews

Informants were selected from a database of professional contacts of a leading international procurement association, which facilitated initial contact. To explore if boundary spanner corruption occurs in diverse settings, we used a purposeful sampling strategy to recruit informants from firms of different sizes, geographies and industries (Homburg et al., 2014). We sought key informants who have both current oversight over boundary spanners and previous experience in a boundary-spanning role. We recruited and interviewed 25 senior-level executives who have knowledge of relevant interpersonal, intraorganizational, and interorganizational factors. Executives from both seller and customer firms were included to obtain perspectives from both sides of the B2B relationship: nine seller-side, thirteen customer-side, and three with experience in both seller and customer roles. Several informants were interviewed multiple times to clarify and explore emergent topics at greater depth, which afforded opportunities to enhance the multi-level framework and propositions. A total of 34 interviews were conducted, which is in line with recommended sample sizes for similar studies (McCracken, 1988). Web Appendix C summarizes interview informants.

Informants were invited to participate in an interview via email by a member of the research team, who subsequently conducted all interviews. All who were invited agreed to be interviewed. In-person interviews were conducted in English in various venues ranging from formal locales, such as the informant’s workplace, to informal cafes and restaurants. Nine of the interviews were conducted via video calls. Informants were not compensated. To encourage honesty and accuracy, we assured complete anonymity, de-identified informants and companies, and asked about practices observed in the industry in general. The interviewer used unobtrusive and nondirective questions, avoiding active listening (McCracken, 1988). No questions addressed personal experiences of informants or specific practices of their firms; nevertheless, in the course of discussion, some informants spontaneously offered examples from their own or their firm’s experience. A semi-structured interview guide probing the various aspects of the multi-level framework was developed and continuously refined throughout the interviews. Web Appendix D presents a summary of the interview guide.

Although we structured the interviews around the different levels of our multi-level framework (individual, interpersonal, intraorganizational, and interorganizational) and their roles in boundary spanner corruption, informants had the opportunity to discuss additional aspects they considered important. Each interview began with general questions that covered various aspects of close business relationships and deviant behavior of boundary spanners. Follow-up questions elicited further details, both in terms of facts as well as the informant’s perceptions of the phenomenon. We ended the interview process when no further insights emerged from the last field interviews. Interviews lasted between 40 and 60 min. All were audio-recorded and transcribed with permission of informants, generating 463 double-spaced pages of field interview transcripts in Times New Roman font size 12 (136,095 words).

Analysis of qualitative data from field interviews

The analysis of the interview transcripts was carried out through several coding steps. Initially, we tagged words or phrases of interest through a line-by-line analysis of each transcript. This process was guided by our conceptual understanding of boundary spanner corruption, but also allowed for new aspects to emerge. Next, for each transcript, we grouped similar tagged words and phrases into zero-order categories, which capture their general essence of meaning (Yin, 1994). As individuals often use different terminology to refer to the same underlying concepts, the 667 zero-order categories were examined and aggregated into first-order categories representing similar concepts despite divergent terminology (Homburg et al., 2000). This process resulted in 98 descriptive first-order categories. Next, we grouped related first-order categories into higher-level second-order categories. While the first-order categories are descriptions formed from informants’ considerations, the second-order categories have been synthesized using the literatures on deviance in service and sales organizations and the dark side of marketing relationships (Homburg et al., 2017). This process resulted in 17 second-order categories (see coding tree in Web Appendix E and example category aggregation in Web Appendix F). These second-order categories were then related to our multi-level framework of boundary spanner corruption (Grewal et al., 2007).

The initial analysis process was done by the researcher who carried out the interviews. We ensured the trustworthiness of our coding process in several ways: we triangulated data by relating emerging categories with the relevant literature streams, we checked refutability across key informants with a diverse industry and geographic background as well as different firm sizes, we ensured triangulation by engaging a second researcher unfamiliar with the field to independently code the data, we resolved any incongruent coding through discussion between the two coders, and we validated our initial results by returning to selected informants to verify our first- and second-order categories (Homburg et al., 2017). Furthermore, we recontacted three informants, asking them to relate specific examples from their initial interviews to our final framework (Homburg et al., 2014). There was strong agreement that our framework is aligned with actual practices.

Utilization of insights from field interviews

The findings from our qualitative analysis were used in conjunction with our multi-level framework and an extensive review of the associated literature to develop detailed propositions about boundary spanner corruption. Combining our conceptual work with our qualitative findings was an iterative process (Grewal et al., 2007), with the field interviews providing both breadth and depth for proposition development. The qualitative data enabled us to gauge the criticality of, and linkages between, themes relating to each proposition. As we elaborate our multi-level model of boundary spanner corruption in the following, we enrich our theoretical discussion with evocative quotes capturing illustrative managerial experiences.

A multi-level trust model of boundary spanner corruption

Boundary spanner interpersonal trust & boundary-spanning social cocoon development

That trust in business relationships can pay substantial dividends has long been acknowledged in both academic research and the popular press. Nearly 25 years ago, a meta-analysis revealed that “the effect of trust on satisfaction and long-term orientation is even substantially larger than the direct effect of economic outcomes” (Geyskens et al., 1998, p.242). Trust is highly prized because, as noted in the January 31, 2020, Harvard Business Review: “Trust is the social glue that holds business relationships together. Business partners who trust each other spend less time and energy protecting themselves from being exploited, and both sides achieve better economic outcomes in negotiations” (Brett & Mitchell, 2020, p.2).

Relationship-building practices that are inherent to the boundary spanner role require the establishment of personal relationships and frequently include the exchange of personal benefits such as meals and gifts (Palmatier et al., 2009). Some of these interactions evolve into friendships (Haytko, 2004; Heide & Wathne, 2006). The autonomy afforded to the firm’s representatives (Perrone et al., 2003; Zaheer et al., 1998) and their frequent, intensive interactions (Doney & Cannon, 1997; Palmatier, 2008; Rousseau et al., 1998) engender an environment in which strong, mutual interpersonal trust can develop across organizational boundaries (Richter et al., 2006). Mutual interpersonal trust drives the development of bilateral relational norms, such as solidarity, mutuality, and flexibility (Palmatier, Dant, &; Grewal, 2007a), which allow boundary spanners to negotiate more effectively and interact more efficiently, improving relationship performance and organizational outcomes (Palmatier et al., 2006; Weitz & Jap, 1995). Nearly all informants espoused similar perspectives, such as:

There needs to be personal trust there, that enables the clearing of all the red tape, all of the process and hurdles in order to ideate and come up with the new processes and new innovations that would generate value for both companies.Footnote 1

However, mutual interpersonal trust also can enable the development of a boundary-spanning social cocoon. Social cocoons facilitate the evolution of new intra-cocoon norms valid only within the cocoon (Anand et al., 2004; Gambetta, 1993; Husted, 1994). Through encapsulation processes, cocoon members become increasingly close to each other and more isolated from their own firms. The resulting social cocoon can create and sustain a new set of norms that differ from or conflict with those espoused by their firms. To understand the link between mutual interpersonal trust and boundary-spanning social cocoons, we first discuss the encapsulation mechanisms that contribute to the initial formation of a social cocoon and subsequently relate mutual interpersonal trust to such encapsulation.

Boundary-spanning cocoons are cultivated through social, physical, and ideological encapsulation (Greil & Rudy, 1984). Boundary spanners are particularly vulnerable to social encapsulation, adopting behaviors that manifest micro-culture solidarity with their counterparts such as dress code, jargon, non-business activities, or interactions on social media. Indeed, individuals are often selected for boundary spanner positions based on their sensitivity to social cues and ability to adapt to other organizational cultures (Caldwell & O’Reilly, 1982). Training encourages mirroring behaviors such as reflecting a counterpart’s body language, tone, or mood to build rapport (Gremler & Gwinner, 2000). In the words of one informant:

[W]e have ... a negotiation course, which is around building rapport and starting the relationship. Finding what you have in common and finding the commonalities, especially at a personal level, both within the business and outside of the business, allows you to develop a common ground and then start to build a friendship on it. ... [Y]ou do have a level of trust, and I think salespeople generally are trained to do that.

Such social activity between boundary spanners fosters social closeness and deep personal relationships:

So lots of time is spent forming friendships and developing intimacy through social activities that are non-business, like tickets to tennis and things like that. ... little things that you do for each other that form that intimacy. And then later, we can help each other in different ways. I think that’s just how it all goes around. ... They’re really invested in me as a person and they do it for me because they like me, because I’m their friend. The ones where I have a high personal relationship, it’s like a joint love for each other.

Another informant identified seeking social closeness as a nearly inevitable aspect of human nature:

I don’t think you’ll ever replace having a cup of coffee, it’s a social lubricant. That’s how you develop allegiances. And you do it because it’s human … to interact; we are social beings and we’ve always worked socially … through history. It’s unusual for humans to be solitary.

Social closeness is intensified by increased use of social media in the workplace, which allows representatives to be in more frequent social contact. Organizations have very little influence and oversight over social media interactions and thus are increasingly isolated from this aspect of their representatives’ work life, as one of our informants describes:

[S]ocial applications and the increasing interconnectedness between people are blurring the boundaries between work and private lives. ... I was very clear with my team to say, “If you’re talking to a supplier, it does not happen through WhatsApp. You go through email and telephone calls and the company-approved channels”. All of that’s gone by the wayside. ... [T]hat is really quite a challenge because from a compliance standpoint, in most cases it’s extremely hard to get access to that data.

Many boundary spanners experience physical encapsulation, separation from colleagues in their firms who otherwise would reinforce organization norms. Their duties are frequently performed out of the office and require ongoing interactions with counterparts in partner organizations. In extreme cases, the boundary spanner is physically located in or near a counterpart’s facilities. Many salespeople have no office in their firm’s facilities, but rather work out of their homes and spend more time with customers than with their fellow employees (Robertson & Anderson, 1993), a trend exacerbated by the recent pandemic. Even when located in a facility with other colleagues, often more time is spent on video calls or other interactions with boundary spanner counterparts than with intrafirm colleagues. Some boundary spanners realize the usefulness of physically separating counterparts from their firm’s environment:

Like the old adage “you need to be in the car park with the customer”. You see someone in the office, they’ll have their game face on. ... But when we say “get in the car park” it is like take them out of their space, bring them somewhere else and then you really find out what is going on. They might start opening up because they don’t feel that they are shackled by meeting rooms and people around them and then they can start to tell you what their real problems are or what is really going on in their organization.

Such temporal physical isolation together with aspects of social isolation can progressively cause representatives to detach from their firms, as described by one of our informants:

One of the sales guys from our side and one from the purchasing side had a very close personal relationship. The guys would stay at each other’s houses when they traveled and see each other as old mates. Definitely, our guy would let them get away with all sorts of things because he had a closer connection with the individual than he had any sort of connection with our company. So, the personal relationship was definitely everything.

A boundary-spanning cocoon can develop ideological encapsulation, a state in which the worldview of the social cocoon becomes paramount. In the early stages of development, intra-cocoon norms are often influenced by and co-exist with norms of the representatives’ firms; behaviors are generally aligned with accepted organizational practices. In later stages, intra-cocoon norms may begin to diverge from partner firms’ norms, leading to questioning of firms’ decisions, greater allegiance to intra-cocoon norms than firm norms, and potentially deviance. Based on our field study, this appears to happen through developing unique ways of working, shared business objectives, and shared identities and interests:

We both have similar business drivers and goals, right. So, the fact that I understand how my counterpart is being rewarded, what his job is, what success looks like to him, and that he understands what success is like for me, and that we are able to line up those two things with each other. Then that creates an opportunity for us to really create a win-win situation. ... [T]here’s certain unwritten understandings that develop as well. Especially through repetition. In that sense ways of working develop that are outside of strict corporate governance and bureaucracy.

The potency of intra-cocoon norms was noted by one informant:

You can end up in a situation where you might have processes in place that are circumvented because of the personal relationships, for the best of reasons. The circumvention could take place because of a norm: “In this situation we are circumventing and this is why”. But there could be risks, which are realized, which are understated. Assessment of the risk may be distorted by the personal relationship.

Through physical, social and ideological encapsulation processes, boundary spanners assume vulnerabilities. For example, when meeting for non-business activities, bringing their families together (physical encapsulation), or exchanging stories about families or personal histories (social encapsulation), boundary spanners assume social vulnerabilities. When adopting ideas and business practices that deviate from their firms’ positions (ideological encapsulation), boundary spanners also assume professional vulnerabilities. Interpersonal trust can instigate, lubricate, or even legitimize these encapsulation processes by reducing the risk of the associated vulnerabilities being exploited by the trusted counterpart. As interpersonal trust deepens over time and boundary spanners increasingly accept more mutual vulnerability, they become more susceptible to the influence of their counterpart, gradually strengthening encapsulation (Lofland, 1969; Pillemer & Rothbard, 2018; Zahra et al., 2006). Thus, for encapsulation to thrive, interpersonal trust is essential (Haytko, 2004):

So, if you ask me to detail elements of my clients’ personal lives, I can probably tell you quite an astonishing amount of detail. About who they are, their marriage, their girlfriend’s or boyfriend’s name. What they do for a living. How long they’ve been together. What they do at the weekend. You know all of these things. Do they live together? How long do they live together? Do they own any property together in other countries? ... [P]eople tend to share personal information with people that they like or that they feel comfortable sharing that sort of information with, and when people start to open up about their personal lives, it makes the business more personal as well. ... If you can’t develop that relationship where they go “I trust you”, getting those sorts of information and deals done is infinitely more difficult.

Each encapsulation process in turn strengthens the cocoon members’ interpersonal relationship. Physical encapsulation increases the frequency and intimacy of interactions, social encapsulation promotes solidarity, flexibility, and mutuality, and ideological encapsulation involves a perception of uniqueness and an “us against the world” mentality. Increasing familiarity and trustworthiness in the interpersonal relationship confirms that any vulnerabilities are not being exploited, which strengthens both trust and further encapsulation. This self-reinforcing process pivoting around mutual interpersonal trust provides fertile ground in which a boundary-spanning cocoon can develop and strengthen. We thus propose:

Proposition 1

Mutual interpersonal trust between representatives of different firms in a B2B relationship leads to encapsulation of the representatives, which increases the likelihood of the emergence and strengthening of a boundary-spanning social cocoon.

Boundary-spanning social cocoon & boundary spanner corruption

Although social cocoons can be beneficial (Huang et al., 2016; Pratt et al., 2006), they also can become fertile ground for the development of corrupt behavior (Ashforth et al., 2007; Ashforth & Anand, 2003). A social cocoon can create an environment in which organization norms have diminished or no relevance (Ogino, 2007) and where intra-cocoon norms are prioritized over those outside the cocoon (Kong et al., 2014; Pillemer & Rothbard, 2018; Zahra et al., 2006). When a social cocoon forms, the members “actively seek to compartmentalize themselves from external influences” (Anand et al., 2004, p.16). Representatives’ obligations to their respective firms are gradually superseded by allegiance to the counterpart. This sentiment clearly emerged from our field interviews:

People think that because they can trust each other that they can go their own way and forget about the process.

Quite regularly we hear [boundary spanners] having super strong opinions. They say, “This is the only one I’m gonna work with.” And then you look at the quotation and it’s very high. There are other companies who could do it at a better rate. It could be bundled for example, or it could be other solutions, and they don’t even want to discuss it. The personal relationship [of the boundary spanner] detracts from what is technically the right thing to do.

Intra-cocoon norms do not necessarily conflict with or violate partner firms’ norms, but they have the potential to do so. This is particularly likely to occur when corrupt behavior is inculcated through three normalization processes that make “the extraordinary seem ordinary” (Ashforth & Kreiner, 2002, p.215)—socialization, rationalization, and institutionalization (Anand et al., 2004; Ashforth & Anand, 2003).

Socialization represents a slippery slope, that is, a progression from initial cooperation, via incrementalism of ever greater participation in corrupt behaviors, to finally fully compromising and committing to corruption (Bandura, 1999; Gino & Bazerman, 2009). As part of this slippery slope, intra-cocoon norms progressively diverge more seriously from organizational norms. Disregarding their firms’ norms in favor of adherence to the boundary-spanning cocoon’s norms is de facto evidence that the representatives are “in”, thereby validating them as insiders (Aven, 2015; Misangyi et al., 2008). Cooperation is motivated via social rewards for activities consistent with divergent intra-cocoon norms, as illustrated by:

One of my clients recently said to me “You scratch my back, I’ll scratch yours.” … [S]he’s definitely told me some stuff internally that’s being discussed ... I took her out for a nice dinner and we’re due another one soon. ... It’s a combination, it’s saying thanks for your support, but also it’s another opportunity to get to know them a bit better to have a good time with them and show gratitude as well. It’s not bribery because they’ve already done the deal. I will take them out because then I’ve got a business case, right?

Incrementalism begins with minor and isolated violations of extant firms’ norms, gradually escalating to more serious deviance. Thus, the road to corruption begins with an initial voluntary decision to propose or assent to what are often only marginally questionable behaviors. Psychological closeness and personal loyalty motivate cocoon members to seek favor with each other, even if they must suppress their own moral compasses or disconnect from their firms’ norms (Gino & Galinsky, 2012; Hildreth et al., 2016). Consider the following quote that illustrates this incrementalism:

There is a certain part of give and take that is part of the grease that makes the personal relationship. You may ask for a favor of a supplier, through a back-channel communication, for reducing a price. It’s not a contract variation request, “can you process the paper-work faster than the 30 days, can you rush through this?” That is the soft side. The supplier may come back and say “remember when I did you a favor, can you give me this favor back? Can you make an exception on the payment terms?” That is where the slippery slope starts, which starts in a completely reasonable realm, but could lead to more deviant behavior. It’s OK, until it’s not.

Compromise frames a questionable activity as preferred vis-à-vis a greater evil. One informant slipped into the respondent role and provided an evocative personal example where a buyer brought into a compromising situation by salespeople is taking a personal risk to avert the greater evil of jeopardizing the relationship:

Suddenly you’ll find yourself in a slippery place. I’m just having a nice dinner with them, now they’ve just brought in some girls here and I don’t want to do anything. I don’t want to push them away, as that will look bad in the culture and you don’t want to look bad. … But at the same time, they make you sing with them, to kiss them (just on the cheek) and you think “if someone is taking a picture now, I may be in a bad shape.” You are just like “ok, I just want to play my role and have fun,” but they are putting me in a difficult position.

Rationalization reclassifies corrupt behavior as compatible with the representatives’ moral compasses. These self-deceptions can also develop into a shared social resource by which cocoon members jointly justify corrupt activities through moral disengagement and ethical fading mechanisms (Ashforth & Anand, 2003; Bandura, 1999; Tenbrunsel & Messick, 2004). In some cases, informants who act in accord within intra-cocoon norms suppress contemplation of their firms’ norms, become oblivious to violations of those norms, and do not cognitively acknowledge their behavior as deviant, as exemplified by:

When I am doing something kind for someone, or when I facilitate something, or when I help something happen, does this mean this is a kind of passive, informal bribery? Probably not. But, when I help someone, somehow, somewhere, I have in mind that the guy owes me something. … One day, he may help you to do business. It is not direct bribery, but “being kind” to one guy, should we say that he’ll help me? This must happen. It’s an asset, it’s not bribery, but it is informal and it’s almost deviant. It’s impossible for me to say what is acceptable and what is not. It’s personal relationships. You cannot put it on a scale.

Institutionalization embeds deviant actions into routines and processes within the social cocoon. Requests move from ad hoc questionable decisions to actions more systematically contravening the firms’ norms and hold influence over a larger body of cocoon members’ decisions (Bandura, 1999; Gino & Bazerman, 2009). Such institutionalization is further supported by growing intra-cocoon memory regarding previous collaboration in deviant activities. Deviant intra-cocoon norms become self-reinforcing, as members seek to operate in accord with the dominant ideas guiding intra-cocoon behavior (Earle et al., 2010). One informant describes how routinized norm violations can escalate over several episodes:

You know, you’re not going to go and give 100 grand from day one or ask for money from somebody you just got to know today, right? They all started getting coffee, getting a drink, and then just progressively getting something else. ... You can get $100, you accept it. Next time it’s becoming $2000, you accept. Next time is going to be $10,000. That’s how it works, right?

Ultimately, through systemic momentum (Ashforth & Anand, 2003), the corruption becomes routine and, in some cases, can become apparent to those outside the boundary-spanning cocoon:

The most severe one I have heard is a large sum of money in a red packet from a supplier during a special celebratory dinner given to a customer—a senior individual who was a decision-maker. It sounds legitimate, as it’s a lucky draw, but I have been told that this particular senior always won the lucky draw.

In summary, the normalization of corruption in boundary-spanning social cocoons is often given legitimacy by relationship-building practices. In fact, the link between the boundary spanners’ interpersonal relationship and corruption was top of mind for most informants. Thus, we propose:

Proposition 2a

A boundary-spanning social cocoon comprising representatives of different firms in a B2B relationship allows for the normalization of corrupt behavior, which increases the likelihood of boundary spanner corruption.

The potential for corrupt behavior is enhanced in the presence of individual or intraorganizational factors that induce deviant intentions (Marcus & Schuler, 2004). Boundary spanners may lack internal controls or have individual propensities that make them particularly susceptible to corrupt behavior. Lack of internal controls such as weak ethical values or low self-monitoring can make deviant behavior appear less problematic (Brady et al., 2012; Jelinek & Ahearne, 2006a). Albeit many companies provide norm-enforcing ethics and compliance training, one informant outlines how this proves ineffective for boundary spanners with weak ethical values:

He just thought there’s no conflict of interest here. He didn’t see it. He had training from Ethics and Compliance, he had signed the declaration that he had done the training. This was over several years, but he still couldn’t see the conflict of interest. It was never proven that he was receiving a payment under the table for awarding contracts. But presumably he was getting some sort of benefit because we were talking about in total around about five and a half million of invoices that were completely unsubstantiated.

In addition, many of the propensities associated with boundary spanner success—competitiveness, drive, risk-taking, extroversion—can result in greater willingness to engage in deviant behavior (Harris & Ogbonna, 2002, 2006; Jelinek & Ahearne, 2010). One informant outlines how boundary spanner competitiveness motivates corrupt behavior:

So, the local business invited my procurement manager. It was a very important supplier and my procurement manager, without asking me for permission, went ahead and attended the session. It was supposed to be a seminar, but the seminar is just the front. It was actually a drinking session with a cash gift. ... It was his ego because he wanted to outshine you. He wanted to show to his peers across the region that he can socialize with a very important supplier and being seen. So, this one clouded his judgment.

Although boundary spanners may not be more personally disposed to deviance than other employees, their competitive work environment and organizational role inherently involve intraorganizational triggers and opportunities that make them particularly susceptible to deviant behavior (Jelinek & Ahearne, 2006a, 2006b, 2010; Robertson & Anderson, 1993). In sales roles, triggers are prevalent; quantitatively-based individual performance targets, public performance rankings and awards, quotas that are tied to significant leaps in compensation, and other individual performance pressures create a competitive environment beyond that experienced by most other employees. Our informants highlight intraorganizational pressures as a prevalent trigger for potentially corrupt behaviors within a boundary-spanning social cocoon:

Huge pressure from the business to grow sales, to grow the platform, to grow the presence in customers. The question is how. The how is all about interpersonal relationships. It’s all about having my friends in the distribution networks who can play God and decide whether I stock your products, whether I don’t stock your products, and so on.

When individual accountability is coupled with perceived unfairness (e.g., in the allocation of leads, territories, or bonuses), boundary spanner trust in the firm and job satisfaction erodes. The accompanying resentment, frustration, exhaustion, and loss of direction and shared purpose can cause moral disengagement that may trigger deviant intentions, which may then be acted out through deviant behaviors, often justified as actions taken to restore equity (Jelinek & Ahearne, 2006a, 2006b; Ramaswami & Singh, 2003). Given the intraorganizational complacency that grants boundary spanners wide latitude and autonomy in pursuit of organizational objectives, a motivated boundary spanner has many opportunities to reduce perceived inequity by taking actions that enhance their personal outcomes at the firm’s expense (Agarwal & Ramaswami, 1993; Ramaswami, 1996). One informant describes a specific example of how generally lax procurement controls create an intraorganizational opportunity enabling deviant activities:

[Firm] has a fairly high dollar threshold before they start worrying about this type of thing. It’s a lot of money to an individual, but at [firm] level … $10 or 20 million here or there doesn’t really register much. So our … procurement control framework, I would say it’s lax, it has a higher tolerance than other companies. … [Y]ou don’t even need a contract for anything that’s going to cost $2 million or below. ... It really boils down to [firm’s]‘s risk appetite ... and there are lots of opportunities to engage in a fraudulent manner without detection.

If no boundary-spanning cocoon exists, deviant intentions are likely to be pursued via individually-executed deviant behaviors. Although individual and intraorganizational factors can drive individually-executed deviance, they are not sufficient to generate boundary spanner corruption. A boundary-spanning social cocoon provides an excellent opportunity for a deviance-minded individual—a viable, highly attractive, and convenient venue in which collusive deviant behaviors can be normalized and pursued, with the added benefit that the cross-firm collusion is difficult for either firm to detect. Information asymmetry inherent in boundary spanning roles creates opportunity for representatives to extract economic rents to the detriment of their firms (Jaworski, 1988; Jaworski & MacInnis, 1989). Actions, decisions, orders, and payments are not internal to one firm, but cross firm boundaries. Boundary spanners can capitalize on this complexity and use their technical and bureaucratic expertise to conceal deviant actions from their firms (Jávor & Jancsics, 2016). One informant describes how the unique complexities of the B2B context provide the opportunity for boundary spanners to collude on KPIs in response to a specific intraorganizational trigger:

It happens because the pressure that everyone is under. The KPIs, it hurts you when it hits the pocket. There is no official transaction, there is no transfer between banks, there is no evidence, I communicate with you offline, no email, nothing. It boils down to the individual: “You scratch my back”. It happens when the relationship goes to a very good level. You have built a relationship where I can trust you.

Therefore, although individual and intraorganizational factors can drive individually-executed deviance, they are not sufficient to cause boundary spanner corruption. We propose:

Proposition 2b

A boundary-spanning social cocoon comprising representatives of different firms in a B2B relationship allows for the normalization of corrupt behavior, which results in greater likelihood of boundary spanner corruption as individual or intraorganizational deviance-inducing factors increase.

Intraorganizational agency trust & boundary spanner corruption

Intraorganizational trust can develop between many different trustor-trustee dyads within the firm (Fang et al., 2008). With regard to boundary spanner corruption, trust between a representative and their supervisors as direct line managers is most relevant (e.g., a salesperson and their sales manager or a buyer and their procurement manager). However, interorganizational trust may also relate to other internal relationships of the representative (e.g., a salesperson with a product manager or a buyer with an operations manager).

In this context, agency trust is the belief of relevant managers that their representative will perform well, fulfill complex responsibilities, pursue the organization’s best interests, and adhere to organization norms (Fang et al., 2008; Perrone et al., 2003). In a business relationship, the partner firms each bestow fiduciary responsibilities upon their respective representatives to enact their firm’s policies and practices. A significant amount of agency trust is placed in boundary spanners (Villena & Craighead, 2017; Wang et al., 2013), both in the customer and the seller firm, particularly when representatives perform their duties independently and in isolation from others within their firm (Rindfleisch & Heide, 1997). Although agency trust can serve as an effective governance structure in some circumstances (Brown et al., 2000; Heide & John, 1992; Zaheer & Venkatraman, 1995), relational safeguards are often ineffective in preventing opportunism (Jap et al., 2013; Jap & Ganesan, 2000; Wang et al., 2013; Wathne & Heide, 2000). Strong agency trust can allow a boundary spanner to enact deviant intentions:

Because they could get away with it. Because they were already trusted. Human relationships and trust are a double-edged sword. In that case the trust that was given to those people was abused. And it killed the company, a hundred years of legacy gone. So personal relationships are extremely important, but trust requires to be worked upon, it’s not just set and forget. And if you do think you’re going to set and forget it, then temptation creeping in is going to be that much more alluring.

A serious complication of agency trust is that it motivates managerial laxity and exceptions in the development and use of behavior-based controls such as close monitoring, direction, and oversight and intervention in the trusted boundary spanner’s activities (Anderson & Oliver, 1987). Agency trust encourages increased reliance on outcome-based controls, which allow boundary spanners to be “left alone to achieve results in their own way using their own strategy” (Anderson & Oliver, 1987, p.76). Outcome-based controls, however, can inadvertently increase performance pressures, perceived unfairness, and resentment, providing motivation for deviant actions (Ramaswami, 1996; Robertson & Anderson, 1993). Regardless of the specific types of control used, scarce managerial resources are likely to be devoted to monitoring less-trusted boundary spanners, with highly trusted representatives given much less oversight as the following quote by one of our informants exemplifies:

He’s the number one person in that function, and owns the budget for [region]. He became very bold, I think, because he’s a main decision maker. He only needs to explain to his boss and he signs off. So because he’s trusted by his boss, he’s trusted by the audit. ... But he’s not playing to the market standard.

Behavior-based controls are implemented through intensive interactions between managers and boundary spanners in which the representatives discuss calls made or correspondence with counterparts as well as the content of those interactions, such as product demonstrations or social entertaining. These controls not only allow managers to support the development of boundary spanners’ skills, they also provide a venue in which managers can reinforce the firm’s norms. Behavior-based controls such as these signal the importance the firm places on responsible behavior and the firm’s intent to enforce and evaluate representatives’ adherence to organization norms (Jaworski, 1988; Robertson & Anderson, 1993). One informant observes regarding the importance of having behavioral oversight:

It boils down to people and process. You need to have strong compliance and periodic audit so that people are aware that their actions are being monitored or being subject to scrutiny.

Behavior-based controls impede the development of boundary-spanning social cocoons by inhibiting the encapsulation process. Frequent and in-depth interactions between representatives and managers reduce the representatives’ physical encapsulation with counterparts. Unfortunately, agency trust can undermine the utilization of behavior-based controls that would otherwise discourage or safeguard against encapsulation (Gargiulo & Ertug, 2006; Marcus & Schuler, 2004). If managers meet less frequently with highly trusted representatives, those individuals are at greater risk of physical encapsulation. Regular meetings enable the deepening of representative-manager social ties, thereby also undermining social encapsulation. But if a trusted boundary spanner’s expenses are less carefully scrutinized, excessive social entertaining is more likely to occur, along with concomitant social encapsulation. Prominent controls also mitigate against ideological encapsulation by continually reinforcing the firm’s expectations, practices, and norms (Robertson & Anderson, 1993). When a trusted representative is given more autonomy, however, the firm’s norms are typically discussed and reinforced infrequently.

Weak behavior-based controls also contribute to the normalization of deviant norms or behaviors within a boundary-spanning cocoon by failing to inhibit the socialization of undesirable behaviors by boundary spanners that can act as a precursor to more serious corruption. High agency trust and the resulting weak behavior-based controls can bolster a representative’s confidence in their own judgment and self-control, consequently undermining the self-reflection and introspection that could lead to awareness of, more careful evaluation of, and ultimately resistance to initially minor deviations from organization norms. Furthermore, it encourages morally malleable reasoning (Jap et al., 2013; Tenbrunsel et al., 2003) and thus the rationalization of corrupt behavior by boundary spanners. The absence of clearly prohibited behaviors, or failure to sanction behavior that contravenes the firm’s norms, signals that those behaviors are acceptable or even desirable (Ashforth & Anand, 2003; Robertson & Anderson, 1993). Weak behavior-based controls can signal that organization norms are discretionary, particularly when violated for a “greater good” such as acquiring business. Non-intervention allows corrupt behavior to proliferate and become endemic and thus institutionalized within the social cocoon.

One informant singled out the importance of controlling the encapsulation and normalization processes early on:

At [Firm], we did a lot of staff rotation. ... Two years. It's very common. They rotate between different accounts. ... that would prevent the emergence of a kind of social relationship. No one is going to give you 100 grand on the first day they meet you. They always start from, “OK, let's get a drink and I give a dinner or something”. The relationships develop progressively from the bottom to all the way. You can control the start and stop that going too far. ... knowing them too well. ... Most companies have a code of conduct. So there’s like a limit, say $200. ... Everything works step by step. So, if you manage the things on a very low value, you slow things down before they get out of control. ... Don’t try to stop $100,000 cash. ... You control at the beginning when it is small.

Although agency trust is generally beneficial, it can become dysfunctional when it undermines sufficient monitoring of boundary spanners. Agency trust does not cause boundary spanner corruption, but it helps enable encapsulation of representatives within boundary-spanning cocoons and provides greater opportunity for an individual with deviant intentions to normalize corrupt behavior. Boundary spanner corruption is most likely to occur when agency trust exists in both firms simultaneously. A representative employed by a firm with strong behavior-based controls is much less likely to encapsulate into a boundary-spanning cocoon, adopt deviant intra-cocoon norms, and normalize corrupt activities. Therefore, if one of the partner organizations has effective behavior-based controls, boundary spanner corruption is less likely to occur. We propose:

Proposition 3

Simultaneous agency trust placed in representatives of different firms in a B2B relationship weakens behavior-based controls, which increases the likelihood that a boundary-spanning social cocoon will develop as well as result in boundary spanner corruption.

Interorganizational trust & boundary spanner corruption

Research indicates that partner firms can form a trusting interorganizational relationship (Fang et al., 2008; Morgan & Hunt, 1994). In this case, trust is not specific to a particular relationship between representatives of two firms, but cuts across most, if not all relevant touchpoints between these firms. This may involve different functions and hierarchy levels in either firm (Doney & Cannon, 1997). Such organization-level trust manifests as increased confidence and positive expectations within the wider organization about the partner firm, including adherence to relational norms that have formed within the interfirm relationship. Mutual interorganizational trust has been found to amplify the dark side effects of strong interpersonal trust (Fang et al., 2008; Jeffries & Reed, 2000). Interorganizational trust can create a false sense of security for both firms and their respective managers regarding the risk of representatives’ encapsulation and normalization in boundary-spanning cocoons.

As with agency trust, interorganizational trust provides supervisors or managers of representatives with a justification for devoting or reallocating scarce resources to monitoring and policing the actions of representatives in relationships with less-trusted business partners, while relaxing such controls in relationships with trusted partners. Dealing with extra-relational threats to the firm’s mutual interests may be given more attention by managers than safeguarding against internal threats. Development and use of boundary spanner controls in a strong, ongoing B2B relationship is unlikely to be a high priority. Vulnerability increases as the firms’ managers have little knowledge of, or influence over, the norms and practices developing within a boundary-spanning cocoon. In addition, mutual interorganizational trust can encourage representatives in a social cocoon to view their interpersonal trust as consistent with their firms’ positions (Fang et al., 2008), thus aiding in normalizing deviant behavior and rationalizing that behavior within that cocoon. The following quote based on the strong relationship between two CEOs signaling mutual interorganizational trust to managers exemplifies the possible detrimental effects on oversight over boundary spanners:

A CEO had a personal relationship with another CEO, where they were excited about the opportunities and were both putting their energies into making a partnership work. Twenty years later ... the relationship is just running on autopilot, junior people are managing one another and there isn’t any oversight, so who knows what is happening? There’s not any framework to measure against.

Furthermore, when there is interorganizational trust, managers have a default tendency to presume that the partner’s representatives are also trustworthy (Doney & Cannon, 1997). Consequently, within a successful, trusting interorganizational relationship, developing and deploying behavior-based controls can be viewed as less important (Gulati & Nickerson, 2008; Villena et al., 2011; Wuyts & Geyskens, 2005). Lack of clarity in responsibilities for monitoring and lax enforcement of norms fosters both encapsulation and normalization within a cocoon (Ashforth & Anand, 2003; Lofland, 1969; Luo, 2004). The detrimental effect of interorganizational trust on relationship-specific controls was highlighted by multiple informants:

Due diligence increases, to a point, but when trust increases, it starts to slow down. If one doesn’t hear anything bad from the company, one reduces due diligence. But this only happens over time. For it to reach that stage, one needs multiple engagements to build trust.

If one has higher trust when you’re working with a supplier, then you may put less metrics in place to manage that supplier, and you may be more relaxed in managing that supplier until something goes wrong. ... [T]hings could just carry on for quite a while.

Similar to agency trust, mutual interorganizational trust does not cause boundary spanner corruption, but it cultivates an environment that provides opportunities for boundary spanners to more readily engage in corruption. Therefore, we propose:

Proposition 4

Mutual interorganizational trust between firms in a B2B relationship weakens behavior-based controls, which increases the likelihood that a boundary-spanning social cocoon will develop and result in boundary spanner corruption.

Consequences of boundary spanner corruption

Boundary spanner corruption generates immediately realized costs, opportunity costs, and latent risks that harm the representatives’ organizations (Luo, 2004). With regard to realized costs, boundary spanner corruption inherently inflicts immediate harm to at least one firm, as its representative acts contrary to the role as an agent of the firm. For example, a salesperson may sell to a complicit buyer at a lower price than the customer would have been willing to pay or a buyer may misuse company funds to acquire a desired supply relationship. One informant describes how excess inventory can be procured below market price through boundary spanner corruption:

It’s called a “broker fee” which will probably be paid in cash.

Boundary spanner corruption also may cause immediate harm to both firms as it results in misallocated resources, leads to sub-optimal decisions, and harms the productivity of the firms’ business relationship (della Porta & Vannucci, 1999). One informant describes how boundary spanner corruption can lead to doing business with sellers for which no business case exists:

The team in procurement was channeling money to certain suppliers who could then arrange structures in the distribution network to enhance or influence sales. The mix of preferred suppliers was changing dramatically, and there is nothing wrong with that, if there is a good business case to do it, but there wasn’t. There were examples of non-preferred suppliers, who hadn’t, as it transpired, gone through an onboarding process who were basically being charmed business.

Such sub-optimal decisions generate opportunity costs associated with the foregone alternative course of action displaced by the corrupt activity; the marginal benefits of the decision that would have been made on the merits, in the absence of corruption, are forfeited. It can undermine employee morale and productivity or damage development of organizational capabilities (Luo, 2004). Boundary spanner corruption has the potential to dissuade future relationship-specific investments and interfirm cooperation, consequently undermining both organizations’ ability to obtain full value from the relationship. In compound or multiplex relationships (Ross Jr & Robertson, 2007; Tuli et al., 2010), this may affect and create risks for other ventures between the partner firms, such as R&D or marketing alliances, joint ventures, or licensing agreements. Due to the resulting loss of interorganizational trust, such other ventures may not come to fruition at all, may be discontinued or faded out, or their governance may change to rely on contracts and controls that create additional transaction costs as well as stifling the value-creation potential of these ventures.

Even if one firm receives short-term financial benefits from a representative’s corrupt behavior, both firms inherently incur increased latent financial risk of potential exposure (Palmatier, Scheer, & Steenkamp, 2007b). “Even when the actors ostensibly direct the corrupt behavior on behalf of rather than against the organization, the behavior may ultimately impede the organization’s ability to accomplish its legitimate purpose and may threaten its very survival” (Lange, 2008, p.710). Firms face increased potential for public relations damage, marring of brand image, loss of sales, and extensive direct costs associated with audits, fines, and penalties—all of which are realized if the corruption is publicly exposed (Johnson et al., 2018), as exemplified by:

It is not an individual career risk anymore. It is the organization you represent. The reputation risk is so high, you don’t want to be investigated and appear on the front page of international newspapers saying “this organization has been involved in forging transactions”.

In the GlaxoSmithKline scandal, for example, GSK not only suffered a severe loss of sales in China. Its brand reputation was damaged worldwide. U.S. prosecutors initiated actions under the Foreign Corrupt Practices Act. Beyond reputational damage, legal authorities in many jurisdictions mandate that firms police their representatives. Failure to prevent corruption can result in large fines or serious legal ramifications (Luo, 2004), such as the $6 billion in fines for banks involved in the LIBOR scandal. Under the newly implemented UK Bribery Act, organizations can be prosecuted for their representatives’ corruption, even if the organizations were unaware of the activity (Nicholls et al., 2011). Given the realized costs, opportunity costs, and latent risks association with boundary spanner corruption, we propose:

Proposition 5

Boundary spanner corruption between representatives of different firms in a B2B relationship increases organizational and relational costs, which negatively impacts each firm’s performance.

Figure 2 provides an overview of the propositions and depicts the resulting multi-level model of trust and boundary spanner corruption in business relationships.

Fig. 2
figure 2

Overview of propositions

Full size image

Discussion

Theoretical implications

Our research makes several key contributions. First, we identify and conceptualize boundary spanner corruption, a distinct and important dysfunctional complication of marketing relationships that has been overlooked in prior research. This deleterious behavior is unique in that it involves cross-firm collaboration of representatives in a B2B relationship, making it particularly difficult to police or detect. It is also unique in that it negatively affects both firms. Corruption is particularly likely to arise in boundary-spanning relationships, where employees often face individual-level numerical targets and organizational pressures to deliver results (Ramaswami & Singh, 2003). Boundary spanner corruption is especially insidious when it offers short-term financial benefits for the boundary spanner’s firm, for this initiates a deleterious reinforcement process—questionable or deviant action delivers positive firm outcomes, which leads to accolades and financial rewards for the boundary spanner and bolsters trust in that individual, which in turn motivates less vigilant control over that person, who subsequently has both greater motivation and opportunity to engage in future corrupt behavior both within the same B2B relationship and in other relationships managed by that representative. Unethically obtained bonuses or other compensation can skew the targets and quotas set within the firm, setting goals for boundary spanners that are difficult to achieve by honest means. In addition, other boundary spanners within the firm who are aware of a colleague’s questionable behavior and observe the laurels that a deviant colleague receives may intuit that the organization implicitly approves of such behavior, making it easier to justify that boundary spanner’s own questionable or corrupt practices. Boundary spanner corruption has not been sufficiently addressed in the extant literature. Seemingly related concepts, such as the dark side of business relationships (Moorman et al., 1992) or deviance in sales and service organizations (Ramaswami, 1996), omit consideration of boundary spanner corruption, the circumstances that enable and promote that corruption, and the threats it poses for marketing relationships. Our conceptual model addresses this gap and offers theoretical propositions to be examined in future research.

Second, we introduce the concept of the boundary-spanning social cocoon, a key factor that enables a deviance-minded boundary spanner to engage in cross-firm collusion rather than taking independent action (Robertson & Anderson, 1993). Boundary spanner corruption is a form of deviance that is particularly difficult for a firm to detect. Actions taken to hinder the formation of boundary-spanning cocoons and the normalization of corrupt behaviors will reduce the likelihood of boundary spanner corruption. Understanding and detecting factors that enable or encourage the development of a boundary-spanning cocoon and the normalization of corrupt behaviors are particularly important, for such knowledge can empower firms to devise better controls to prevent the manifestation of deviant behavior (Ramaswami, 1996). While most existing compliance controls are aimed at uncovering corrupt behaviors post factum, such as through regular audits, our conceptualization suggests that boundary spanner corruption can be more effectively addressed through ex ante monitoring of boundary spanners and prevention mechanisms. Such mechanisms hold promise for firms to reap the benefits of close cross-firm interpersonal relationships while simultaneously minimizing the negative consequences that can stem from those relationships (Jap & Anderson, 2003). Although a firm can have full knowledge of its own agency trust and behavior-based controls and some understanding of interorganizational and interpersonal trust (Wang et al., 2013), it knows much less about the partner’s agency trust and internal controls. Relationship marketing at the interfirm level requires both firms to collaborate to enact and enforce joint controls, which we anticipate, based on insights from our framework, to be much more effective than a firm independently trying to impede or police boundary spanner corruption.

Third, we demonstrate the importance of understanding how factors at multiple organizational and interorganizational levels interact to impact employee behavior. Building on Fang et al.’s (2008) findings, we theorize that boundary spanner corruption is impacted by interactions among trust at different levels. Our framework not only highlights that potential dark side influences equally reside on those levels and can be exacerbated as an unintentional result of trust-building efforts and relationship marketing, but that the dark side potential can be greatly underestimated unless deviance-inducing factors at the intraorganizational and individual levels are taken into account (Brady et al., 2012). Our conceptual framework cautions that such deviance-inducing factors play an important role in turning value-generating marketing relationships into value-destroying boundary spanner corruption. Distinguishing between factors of motivation—intraorganizational triggers and individual propensities—and factors of control—intraorganizational opportunity and individual self-control—provides a useful framework for assessing which factors are most pertinent in a specific B2B relationship. The failure to consider, recognize, and examine the multi-level nature of antecedents of marketing and relational behaviors limits the explanatory power of extant marketing research. We advocate the development of a holistic approach to research on marketing relationships, an approach that takes into account both bright and dark side effects operating at multiple levels in B2B relationships. It is critically important to understand the circumstances in which the same interpersonal, intraorganizational, and interorganizational factors that generate highly favorable outcomes (Doney & Cannon, 1997) can simultaneously—and unintentionally—lay the foundation for dysfunctional consequences such as boundary spanner corruption.

Future research directions

Empirical research on surreptitious activities such as boundary spanner corruption is challenging. Any attempt to obtain self-reports of boundary spanner corruption is essentially impossible, as an individual willing to engage in corruption is certainly willing to lie about it. Although quantitative methods using secondary data regarding the impact of publicly-exposed corruption can be obtained, available data sources rarely capture crucial aspects in our framework such as trust, behavior-based controls, encapsulation and normalization mechanisms and social cocoons in interfirm dyads. Thus, qualitative historical methods drawing on archival or primary data (e.g., case studies, interviews) may be a good first step to gain a nuanced understanding of boundary spanner corruption. In addition, research designed to gather quantitative primary data such as critical incident techniques and experiments also offers potential insights. However, non-projective or direct approaches are likely to be overwhelmed by lying or social desirability bias. Thus, projective techniques are needed that focus on behaviors of others to glean some information about conditions that enable, promote, and impede boundary spanner corruption. Focusing the research on “a colleague” allows informants to draw on the observed deviance of others or project from situations in which they were tempted to engage in deviant behavior.

Another promising avenue would be to pursue an indirect approach that focuses on detecting a boundary-spanning cocoon, rather than striving to uncover the corruption directly. A cocoon-likelihood index could be developed to detect specific encapsulation behaviors and attitudes that, individually, are innocuous. However, the totality of the set of cocoon-enabling or cocoon-promoting factors indicate that a boundary-spanning cocoon exists or is likely to form. As each warning-sign behavior or attitude in isolation is accepted or even desirable, a boundary spanner is likely to honestly and accurately report the various elements. Intermixing these indicators in a battery with other innocuous behaviors and attitudes would further reduce their apparent importance and thus also reduce the likelihood that an individual who is engaging in corruption would be motivated to lie about the seemingly innocent behaviors.

The development of a cocoon-likelihood index could be based on a process such as the following example. First, the salesperson’s level of engagement in various activities and attitudes vis-à-vis a specific buyer is assessed. Second, all activities and attitudes are presented again, but focusing on the salesperson’s typical or average behavior vis-à-vis all accounts. Third, focusing on the set of warning-sign elements, the salesperson’s activity with the specific buyer is compared to the salesperson’s typical or average behavior, with the comparative values then compiled into a formative index of cocoon likelihood. Higher involvement in warning signs with a specific buyer compared to the salesperson’s typical behaviors or attitudes would not indicate that corruption is occurring, but rather that the pre-requisite conditions for boundary-spanning cocoon formation exist and warrant further investigation.

Relevant warning-sign behaviors and attitudes would need to be developed for the specific research context. Consideration of the three encapsulation processes may provide insight regarding relevant warning-sign behaviors and attitudes. Social capital theory and its distinctions between structural, relational, and cognitive aspects (Nahapiet & Ghoshal, 1998) may provide a useful theoretical grounding to develop items to capture physical, social, and ideological encapsulation in the boundary spanner dyad. Potential examples in a sales context may include: time spent on official sales interactions, time spent on informal interactions outside the typical workplace, number of casual network-building interactions (meals, entertainment, etc.), degree of confidence that the buyer “has your back,” confidence that “you can share proprietary information without concern it will be divulged to others within the customer firm,” interaction with the buyer before they were employed by the customer firm, acquaintance with the buyer before the salesperson joined the seller firm, etc. As a distractor task, such items can be interspersed with others such as “amount of newly-introduced products the buyer has purchased,” “use of adaptive selling with this buyer,” “uniqueness of the customer firm’s needs,” the salesperson’s estimate of growth potential with the customer, etc.

Given the purpose of our theory-focused conceptual framework, we offer no specific hypotheses. Rather, our conceptual framework and our propositions serve as theoretical way markers regarding the trust mechanisms that operate at different levels, the interactions among those mechanisms, and their effects on performance, providing a foundation for empirical research. We implicitly focused on high levels of trust at all three levels for explanatory purposes. However, there is a need to explicate and examine the precise nature of the interactions among interpersonal, interorganizational, and agency trust and their effects on the potential for boundary-spanning cocoon formation, boundary spanner corruption, and overall relationship performance. Although boundary spanner corruption is undoubtedly a problem, it is important to avoid adopting a cure that eradicates trust and all of its well-documented benefits. Trying to suppress trust across different levels in a business relationship is not practical, for the negative consequences would be certain, while the benefits would be in doubt.

A better approach would be for researchers to seek to determine the level at which agency trust and interorganizational trust become counterproductive and how different trustor-trustee configurations weaken behavioral controls. Is agency trust in boundary spanners more or less deleterious when it emanates from different hierarchy levels, for example, from immediate supervisors or division managers? Is interorganizational trust emanating from higher hierarchy levels, such as from the C-suites, more problematic in creating a false sense of security for buyer-salesperson interpersonal relationships? How much trust is too much, such that the negative effects of excessive trust overwhelm its beneficial effects? We speculate that agency trust constitutes the key moderator and that it has an inverted-U shaped effect, that is, a threshold point after which the negative consequences of interpersonal trust overwhelm its positive effects. We posit that greater interpersonal trust pays increasing benefits until agency trust exceeds the inflection point and monitoring is relaxed too much. It is the combination of high interpersonal trust and high agency trust that enables the boundary-spanning social cocoon to form and enable boundary spanner corruption. Similarly, interorganizational trust may interact with interpersonal trust to encourage formation of a boundary-spanning social cocoon, but this only becomes a breeding ground for corruption if agency trust exceeds that inflection point. This suggests a positive interaction of interpersonal trust and interorganizational trust, which has negative effects on boundary spanner corruption only under the boundary condition of excessive agency trust (i.e., a three-way interaction).

We have focused on basic processes within the boundary-spanning relationship and within the firms involved in the B2B relationship. However, the exact way these processes, activities, and controls are manifest could differ for the seller firm and customer firm. Future research is needed to determine potential nuances in how the trust interactions and cocoon formation factors play out in the seller firm versus the customer firm.

Managerial considerations

The conceptual and qualitative nature of our research provides important theoretical insights, based on which we offer tentative suggestions for business practice. Boundary spanners do not always act in the best interests of their firms. Seemingly well-functioning, close business relationships have the potential to foment and disguise corrupt behaviors by trusted boundary spanners; that corruption exposes their firms to potential immediate financial consequences, long-term negative performance implications, and massive latent financial risks. Although boundary spanners must be trusted due to the nature of their roles, trust need not necessarily extend across the whole bandwidth of the firm-boundary spanner relationship. Thus, trust-but-verify is the way to go. For example, boundary spanners may be trusted more fully and be given greater autonomy in some aspects, such as sales-support asset utilization or adoption of new sales technologies, yet simultaneously not be given free rein over other activities. In particular, relationship marketing activities, exception pricing involving customer-specific discounts, visit scheduling, bidding processes, use of discretionary marketing funds and many similar activities should entail oversight and a more closely managed approval process. We also suggest that decisions regarding boundary spanner monitoring not be delegated to their immediate supervisors, as interpersonal relationships between supervisor and boundary spanner can impact those decisions. The default systems for monitoring boundary spanners should be decided and mandated at a higher level within the firm.

All elements of the firm’s formal and informal controls should be reviewed and redesigned with consideration of how they encourage, enable, or deter boundary spanner corruption. This is particularly important because the more trust that is placed in boundary spanners, the higher the tendency to employ controls that focus on results rather than how the results are achieved. Formal output-based controls create pressures that can provoke deviant intentions or increase willingness to comply with questionable or deviant actions instigated by a counterpart. When using output-based controls with boundary spanners, firms should strive to set realistic performance targets and judiciously select key performance indicators. Further, a boundary spanner that declines and reports a solicited bribe or other corrupt proposal should not be penalized for missing targets or falling short on key performance indicators, but instead be compensated fully. Firms should provide assurances that they will reward representatives who act with integrity rather than penalize representatives who follow firm norms.

We caution against exclusively relying on performance-based controls even with trusted boundary spanners. Behavior-based and input-based controls should be reviewed with an eye to impeding or detecting encapsulation-related behaviors that enable boundary spanner corruption. This could include specifying unacceptable practices, particularly actions that could be cognitively reframed by boundary spanners as serving the firm’s interests. Training should present a message consistent with this perspective. Carefully delineating the scope of boundary spanner autonomy and concurrently establishing management approval systems to validate decisions beyond that scope reduce gray areas regarding what are and are not acceptable practices and signal to boundary spanners that their actions are monitored. However, it is also important to realize that excessive behavior-based controls may severely stifle boundary spanners in carrying out their roles, can undermine boundary spanner trust in supervisors and top management, and thereby promote undesirable reactance. Thus, it is important that firms find the right balance between trusting and controlling the behavior of boundary spanners.

Informal social and cultural controls offer a possible practical way to reduce behavior-based controls and to leverage trust as a governance mechanism. For example, internal conferences and social events can help establish, promote, and provide opportunities to impede or undermine boundary-spanning cocoons by developing and reinforcing shared firm norms. Developing regular, frequent interactions between boundary spanners and firm colleagues through “office Fridays” or jour-fixe meetings to provide sounding boards and feedback mechanisms can promote decision-making consistent with firm norms. Such interactions help facilitate boundary spanner identification with, and trust in, the firm and colleagues. Firm trustworthiness can also be enhanced by providing a comprehensive, effective, and fulfilling task environment, showing boundary spanners that their interests are at the heart of organizational considerations. For example, this can be done by supporting the development of transferable skills through training or executive education that provides more immediate benefits to the individual than to the firm. Furthermore, developing sales processes and tools, demonstrating their effectiveness, offering useful training, designing an adequate and fair compensation system, can all enhance confidence in the firm and reduce boundary spanners’ likelihood of resorting to corrupt behaviors to fulfill their role and achieve objectives.

Our framework highlights the unique difficulty of detecting and dealing with boundary spanner corruption—the initial impetus for collusive deviance can emerge externally from the firm. Boundary spanner corruption is a dyadic phenomenon and requires dyadic solutions. Even if one firm in a B2B relationship has strong systems to detect internal, individual deviance, they are not designed to detect cross-firm, collaborative deviant behavior. Controls implemented unilaterally by that firm may not be effective in mitigating the risk of corruption. The most comprehensive approach for addressing potential boundary-spanning corruption requires partner firms to enact similar controls to impede, prevent, and detect development of a boundary-spanning cocoon and associated negative behaviors. For example, detection of corruption may require the firm and partner to design and implement joint policies for monitoring boundary spanner behavior. Relationship marketing at the interfirm level can build a system in which both firms collaborate to enact and enforce joint controls, which we expect to be more effective than an independent effort.

Counterintuitive to prevailing relationship marketing thought, the more trust at the various levels that exists within a B2B relationship, the more important it becomes for firms to coordinate their controls. Trust and controls together can safeguard close marketing relationships from the detrimental influence of boundary spanner corruption. We suggest that relationship marketing programs encourage open dialogue about current controls within both firms, how well those controls operate, how vigilantly boundary spanners are monitored, and what activities and decisions should be within the scope of boundary spanner autonomy versus requiring managerial approval. Ultimately, criteria for boundary-spanning control system audits could be developed based on existing CRM systems and data captured. Periodic surveys that include cocoon-likelihood indicators and boundary spanner risk profiles can be used to highlight situations needing further investigation or the need to calibrate controls across boundary spanners. Ideally, interventions based on this data could be integrated into pre-existing developmental or performance appraisal meetings. Although these managerial suggestions are tentative and require further investigation, our framework suggests that a coordinated interfirm approach is likely to be most effective in reducing boundary spanner corruption.

Conclusion

Our conceptual framework and exploratory research highlight the importance of considering how concurrent trust at the interpersonal, intraorganizational, and interorganizational levels can enable the development of boundary spanner corruption. We offer a first step on the road to a greater understanding of this dark side complication of relationship marketing and close business relationships. The prevalence of boundary spanner corruption and its huge financial implications indicate that this phenomenon is worthy of future empirical research and managerial attention.