When Do Greedy Entrepreneurs Exhibit Unethical Pro-Organizational Behavior? The Role of New Venture Team Trust

Abstract

Greed is an excessive form of self-interest that leads individuals to pursue material wealth and immaterial desires with little or no concern for the well-being of others. While the literature suggests that greed results in managers acting unethically, a situational strength perspective suggests that this may not always be the case. Therefore, this paper aims at understanding how the new venture context, particularly the role of a team member's affective and cognitive trust in the new venture team, shapes the manifestation of greed in entrepreneurs’ unethical pro-organizational behavior—behavior that allows new venture team members to advance their ventures despite the violation of social norms. Consistent with our theorizing, we find that new venture team members’ affective and cognitive trust in their teams shape the relationship between greed and unethical pro-organizational behavior in opposing ways. Particularly, higher levels of greed are more likely to be connected to unethical pro-organizational behavior when a member's affective trust in the team is high and cognitive trust in the team is low. Our study offers implications for the entrepreneurship and management literatures, alongside implications for practice.

Greed is “the tendency to always want more and never being satisfied with what one currently has” (Seuntjens, Zeelenberg, van de Ven, & Breugelmans, 2015: 917). It is an excessive form of self-interest (Haynes, Campbell, & Hitt, 2017; Wang & Murnighan, 2011) that leads individuals to pursue material wealth, such as money and luxury goods, and immaterial desires, such as status or fame (Seuntjens, Zeelenberg, Breugelmans, & Ven, 2015), with little or no concern for the well-being of others (Haynes, Hitt, & Campbell, 2015; Mussel, Rodrigues, Krumm, & Hewig, 2018; Wang & Murnighan, 2011). Prior research has shown that managers in established organizations who are high in greed often act unethically (Seuntjens, Zeelenberg, van de Ven, & Breugelmans, 2019) and may harm their organizations to pursue their personal desires (Bruhn & Lowrey, 2012; Haynes et al., 2017; Haynes, Josefy, & Hitt, 2015). Indeed, managerial greed has been attributed a major role in organizational scandals (e.g., Levine, 2005) and in the financial crisis in the late 2000s (Seuntjens, Zeelenberg, van de Ven et al., 2015; Wang & Murnighan, 2011).

However, although existing work has indicated that greed translates into managerial behavior, theoretical arguments suggest that this may not always be the case. Specifically, situational strength theory (Cooper & Withey, 2009; Meyer, Dalal, & Hermida, 2010) proposes that situations shape the extent to which an individual's personality manifests itself. Particularly, situational strength is expected to “result in psychological pressure on the individual to engage in and/or refrain from particular courses of action” (Meyer et al., 2010: 122). Thus, context can constrain personality-consistent behavior or provide the freedom to display it (Judge & Zapata, 2015). This notion also holds for greed: Following discussions on whether greed is a trait or a state, a consensus has recently emerged that greed is best conceptualized as a trait that manifests itself to a different extent across situations (Lambie & Haugen, 2019; Mussel, Reiter, Osinsky, & Hewig, 2015).

For a manager of a new venture, the social context of the new venture team (NVT)—defined as “the group of individuals that is chiefly responsible for the strategic decision making and ongoing operations of a new venture” (Klotz, Hmieleski, Bradley, & Busenitz, 2014: 227)—may shape the extent to which he or she pursues greedy desires. NVTs are usually small (Friedman, Carmeli, & Tishler, 2016), and their members “cooperate intensively, are interdependent, and see one another more frequently than do members of TMTs [top management teams] in large, established organizations” (De Jong, Song, & Song, 2013: 1845). While NVTs often face limited or no formal control from boards (Hendry & Kiel, 2004), nor auditing (Adams, 1994; Krishnan, 2005) or risk governance mechanisms (Sheedy, Garcia, & Jepsen, 2021), nor control from the financial market (Byrd, Parrino, & Pritsch, 1998), NVT members are able to closely monitor their teammates’ behavior (Breugst, Patzelt, & Rathgeber, 2015). Thus, the context of an NVT is likely to shape the strength of the situation for its members, particularly in terms of an NVT member's perceived freedom to act in line with his or her personality within the team (Judge & Zapata, 2015).

As a particularly important attitude toward an NVT, trust is likely to affect an NVT member's perceived latitude in how he or she can behave in the new venture. Trust reflects the extent to which a member believes that he or she can rely on teammates and be vulnerable to their actions (Mayer, Davis, & Schoorman, 1995; Tsai, Chi, Grandey, & Fung, 2012). Consequently, the extent to which an NVT member acts consistently with his or her level of greed is likely contingent on his or her trust in the NVT. Studying the interplay between NVT members’ greed and the trust in their teams may yield novel theoretical insights into the origins of managers’ unethical behavior and into how greed and social environments impact managers’ personality-based behavior. Thus, this paper pursues the following research question: To what extent does an NVT member's level of greed shape his or her unethical behavior contingent on his or her trust in the team?

Understanding the link between greed and unethical behavior is important because managerial behavior shapes strategic decisions and firm-level outcomes (e.g., Hoskisson, Chirico, Zyung, & Gambeta, 2016), particularly under conditions of high discretion and high job demands (Hambrick, Finkelstein, & Mooney, 2005), which are typical of the NVT context (Klotz et al., 2014). To address our research question, we draw on one of the main theoretical foundations of greed—namely, self-interest (Haynes, Hitt et al., 2015; Smith, 1904)—and on situational strength theory (Cooper & Withey, 2009; Meyer et al., 2010). Specifically, we develop a model in which two distinct dimensions of an NVT member's trust—affective trust and cognitive trust in the team (De Jong, Dirks, & Gillespie, 2016; McAllister, 1995)—can differentially moderate the extent to which greed triggers the NVT member's unethical pro-organizational behavior (UPB). UPB refers to breaking social norms, values, and laws to support one's organization, the individuals within the organization, or both (Umphress & Bingham, 2011). We focus on UPB as a potential outcome of NVT members’ greedy desires because, in contrast to greedy managers of established organizations, who may personally benefit from taking (unethical) actions that harm their organizations, NVT members’ interests are usually closely aligned with those of their organizations (Haynes, Josefy et al., 2015). New venture managers “pursue their own goals, dreams, and desires in new firm creation” (Fauchart & Gruber, 2011: 935) and “envision their venture[s] strategically as an extension of themselves and their needs” (Ruvio, Rosenblatt, & Hertz-Lazarowitz, 2010: 145). Therefore, if the situation allows for it, acting unethically to promote their ventures’ development (i.e., UPB) might be a central way for NVT members to pursue their greedy desires. We test our model using data from 233 German NVT members nested in 111 NVTs, which we surveyed twice with a time lag of 3 months. We make the following three main contributions to the management and entrepreneurship literatures.

First, we add to the emerging theoretical understanding of how managerial greed can impact different types of firms in different ways (Haynes, Hitt et al., 2015). Studies on established organizations have shown that managers high in greed tend to sacrifice organizational goals to pursue their greedy desires (Haynes, Josefy et al., 2015; Haynes et al. 2017). In contrast, our study indicates that under certain circumstances, greedy NVT members are more likely to engage in unethical behavior that benefits their ventures. Therefore, our study emphasizes the theoretical distinctiveness between NVTs and top management teams of established firms when studying managerial greed and potentially other personality traits that may harm others.

Second, we add to the literature on trust in the NVT context (e.g., Chen & Wang, 2008), which has pointed to the important role of trust for enhancing a team's ability to overcome liabilities of novelty (Blatt, 2009), increasing team member commitment (Wang & Wu, 2012), and improving venture performance (De Jong et al., 2016). By documenting the differential moderating effects of affective and cognitive trust (e.g., Holste & Fields, 2010; Samian, Riantoputra, & Budihardjo, 2021) on the relationship between NVT members’ greed and their UPB, our study indicates that the predominant unidimensional conceptualization of trust in NVT research may be insufficient to capture how trust in their teammates influences entrepreneurs’ decisions and behavior.

Finally, our study contributes to the emerging research on the origins of UPB in the new venture context. In particular, recent theoretical work has suggested that entrepreneurs may lie to stakeholders, such as investors, to move their ventures forward (Pollack & Bosse, 2014; Theoharakis, Voliotis, & Pollack, 2021). While these studies have suggested that “dark” personality traits (e.g., psychopathy, Machiavellianism) and material desires can induce UPB (Theoharakis et al., 2021), we add to this body of research by identifying greed as a potential trigger of UPB in this context. Importantly, our study suggests that the potential effects of personality traits and material desires on UPB can particularly be observed when entrepreneurs have high levels of affective trust and low levels of cognitive trust in their NVT members. This finding indicates the importance of considering the NVT environment for understanding the emergence of UPB in new ventures.

Theory and Hypotheses

This study aims at understanding the contingencies of the relationship between an NVT member's greed and his or her UPB. Therefore, in what follows, we first discuss the definition of greed alongside its theoretical underpinnings. Next, we develop theoretical arguments on how an NVT member's affective and cognitive trust in the NVT can affect the extent to which greed manifests in his or her tendency to engage in UPB. Figure 1 includes a model of our constructs and their relationships. We elaborate on this model below.

Figure 1

Model Variables and Levels

Greed and Self-Interest

Greed refers to an insatiability causing individuals to focus on their material and immaterial desires (Seuntjens, Zeelenberg, Breugelmans et al., 2015). While some authors have considered greed a state (e.g., Cozzolino, Staples, Meyers, & Samboceti, 2004), others have studied it as a trait (e.g., Krekels & Pandelaere, 2015). Another stream of research (e.g., Haynes, Hitt et al., 2015; Haynes et al. 2017) has taken a more nuanced view, indicating that the extent to which individuals express their greed may be contingent on context-specific mechanisms. Recently, a consensus has emerged that greed is best understood as a stable personality trait that manifests in an individual's behavior depending on his or her situational context (Lambie & Haugen, 2019; Mussel et al., 2015).

With respect to the consequences of greed, the literature has documented both negative and positive outcomes so far (Oka & Kuijt, 2014; Seuntjens, Zeelenberg, van de Ven et al., 2015). For individuals, greed acts as a motivation for being productive (Seuntjens, Zeelenberg, Breugelmans et al., 2015), which may benefit them in terms of higher incomes (Seuntjens, van de Ven, Zeelenberg, & van der Schors, 2016). However, the permanent feeling of dissatisfaction from greed can also harm individuals (Seuntjens, Zeelenberg, Breugelmans et al., 2015). Regarding the effects of greed on others, some have argued that managers’ greed can lead to adverse economic consequences, such as diminished stockholder returns, because greedy managers tend to direct organizational resources toward their personal benefits instead of high firm performance (Haynes et al., 2017). Others, however, have considered greed the driving force behind organizational improvements (Bruhn & Lowrey, 2012) as well as behind economic growth and employment, thereby benefiting society as a whole (Greenfeld, 2001).

From a theoretical perspective, greed has often been related to self-interest, which goes back to Adam Smith's (1904) notion of self-interested economic actors. In traditional rational-actor economic models, self-interest has been considered the driving force of human well-being and economic behavior (Haynes, Hitt et al., 2015; Haynes et al., 2017; Wang & Murnighan, 2011). Although greed and self-interest are closely connected, they are also distinct in that greed is considered “hyper-self-interest” or “self-interest taken to the excess” (Haynes et al., 2017: 559). Therefore, greed originates at the point where (rational) self-interest ends and the individual engages in the excessive pursuit of material and immaterial wealth (Haynes et al., 2017). While self-interest thus refers to moderate desires, greed refers to a lack of moderation (Haynes et al., 2017). Because greed includes a focus on individuals’ desires “above all else” (Lambie & Haugen, 2019: 34), it involves little or no concern for the well-being of others (Haynes, Hitt et al., 2015; Mussel et al., 2018; Wang & Murnighan, 2011). People high in greed tend to only focus on their own needs to acquire more while neglecting how their actions can potentially harm others (Seuntjens, Zeelenberg, van de Ven et al., 2015). Thus, those high in greed are willing to take excessive risks to increase their chances of receiving high gains, thereby accepting the possibility of large losses (Mussel et al., 2015), including their future reputation and social status.

Greed in the Entrepreneurship Context and NVT Members’ UPB

Few studies have considered the role of greed in the NVT context. An exception is the conceptual study by Haynes, Hitt et al. (2015), which suggests that higher levels of greed may, for instance, prevent NVT members from distributing a fair portion of equity to stakeholders, such as investors or other top managers. The authors propose that such behavior may have particularly detrimental consequences in the context of new ventures. Specifically, it may limit a venture's human and social capital because fewer stakeholders can be attracted and maintained, which will eventually hamper the venture's performance and survival (Yang & Aldrich, 2017).

Despite these challenges that NVT members’ greed might impose on new ventures, greedy NVT members are likely to aim to advance their ventures to fulfill their greedy desires. While excessive self-interest may lead salaried CEOs, managers of established firms, and employees who are high in greed to sacrifice organizational goals for their own desires (Bruhn & Lowrey, 2012; Haynes, Josefy et al., 2015; Haynes et al., 2017) and thus harm organizations, we expect that greed triggers different behavior in NVT members, who usually own a substantial part of their firms (Hall & Woodward, 2010). Greedy top managers in established organizations (i.e., agents) create high agency costs, which reduce shareholders’ (i.e., principals) wealth (Haynes et al., 2017; Shapiro, 2005). However, this may not be the case in the entrepreneurship context, in which agents and principals are at least partially the same. Indeed, in this context, NVT members’ compensation is primarily provided in the form of shares (Hall & Woodward, 2010), and the expectation of high financial rewards in the future (e.g., from a trade sale or an initial public offering) is a key incentive for founders to successfully develop their ventures (Kagan, Leider, & Lovejoy, 2020). Therefore, NVT members’ personal gains are directly related to their ventures’ performance, and ventures thus represent a means for NVT members to attain their desires. Specifically, acting in favor of one's organization can result in higher future equity-stake valuations (Hamilton, 2000), which may be instrumental in addressing NVT members’ excessive material desires. Moreover, pro-organizational behavior might contribute to fulfilling NVT members’ immaterial desires, such as independence, prestige, and fame (Hamilton, 2000), and to developing a reputation as a highly successful entrepreneur (Schwienbacher, 2007).

While pro-organizational behavior can be highly desirable, such as investing high levels of effort into one's venture (Foo, Uy, & Baron, 2009), greedy NVT members might also be tempted to engage in UPB to advance their ventures. UPB refers to “actions that are intended to promote the effective functioning of the organization or its members (e.g., leaders) and violate core societal values, mores, laws, or standards of proper conduct” (Umphress & Bingham, 2011: 622). Through these actions, individuals intend to support their organizations, organizational members, or both. UPB is different from other forms of unethical behavior, such as counterproductive or deviant behavior, which can harm organizations. As such, UPB is a paradoxical construct as it is considered unethical by society but is likely conducted with the intention to obtain benefits for an organization (Lee, Schwarz, Newman, & Legood, 2019). Specifically, NVT members may engage in legitimacy lies, which refer to intentional misrepresentations of facts to investors or other stakeholders (Pollack & Bosse, 2014; Theoharakis et al., 2021), for instance by presenting dummy products that imitate an actual but not yet sufficiently developed product (Kuratko, Holt, & Neubert, 2020). Alternatively, NVT members may engage in bribery to gain access to resources from governmental programs (Baron, Tang, Tang, & Zhang, 2018), manipulate accounts to overstate financial performance, sell to vulnerable customers who do not understand nor need the focal products or services, and/or conceal information on the harmful effects of products and services from customers or regulators (Castille, Buckner, & Thoroughgood, 2018; Sheedy et al., 2021). Thus, UPB has also been described as a risky behavior that entails short-term benefits but involves the risk of being caught and irritating important stakeholders (Fehr et al., 2019).

We argue that NVT members high in greed might be tempted to engage in UPB because these team members are driven by excessive self-interest and are willing to take excessive risks to receive high gains. Therefore, they are inclined to engage in actions that society deems malfeasant or wrong (Haynes, Hitt et al., 2015). At the same time, UPB represents a direct way for NVT members to address their material and immaterial desires. Importantly, however, we do not expect the relationship between greed and NVT members’ UPB to always be positive and therefore purposefully do not hypothesize for a direct relationship between NVT members’ greed and UPB. Indeed, prior research on greed has found that despite its trait-like character, greed has a situational component (Bruhn & Lowrey, 2012; Lambie & Haugen, 2019; Mussel et al., 2015; Wang & Murnighan, 2011). For instance, studies in the context of established firms have found that employees’ greed is unlikely to manifest in greedy behavior when employees work closely with others (Bruhn & Lowrey, 2012) or when their superiors do not display greedy behavior themselves (Haynes, Josefy et al., 2015). Prior work has also found that top managers in established organizations are unlikely to display greedy behavior when under high supervision, when they have longer job tenure, and when they have little freedom in decision making (Haynes et al., 2017).

Therefore, in what follows, we argue that the extent to which an NVT member's greed is expressed in UPB is contingent on the specific situation he or she is in. Specifically, we build upon situational strength theory (Cooper & Withey, 2009; Meyer et al., 2010), suggesting that individuals do not always act in line with their personality traits; rather, the “strength” of the situation they are in determines the extent to which they express their traits in the form of behavior.

NVT Members’ Trust in Their Teams as a Contingency Factor in the Greed–UPB Relationship

Situational strength theory (Cooper & Withey, 2009; Meyer et al., 2010) explains why an individual's personality sometimes manifests more or less intensely in behavior by considering the situation the individual is in (Meyer et al., 2010). Situational strength refers to the degree to which situational constraints exist in the environment (Cooper & Withey, 2009). The theory distinguishes between two types of situations that individuals can experience—namely, strong and weak situations. While strong situations present clear indicators for how an individual is expected to behave, weak situations provide more freedom, allowing for the more intense expression of an individual's personality (Judge & Zapata, 2015; Meyer et al., 2010). Thus, whereas strong situations “limit the expression of individual personalities” (Cooper & Withey, 2009: 62), weak situations allow individuals to behave more consistently with their personalities (Judge & Zapata, 2015; Meyer et al., 2010).

For members of an NVT, the NVT constitutes an important element of their situation—that is, their work context—because NVT members work in close contact and highly interdependently (De Jong et al., 2013), making the members’ behavior observable within the team (Breugst et al., 2015). Therefore, we expect the NVT to substantially influence a member's perceived freedom to engage in behavior consistent with his or her personality. A key construct that shapes the collaboration and interaction in NVTs is trust between team members (Zheng, 2012). Generally, a team member's trust in his or her team refers to “a belief in the dependability and trustworthiness of team members” (Tsai et al., 2012: 639) and is based on the willingness of this team member to be vulnerable to the actions of other team members (Mayer et al., 1995).

Importantly, although many studies in the NVT context have treated trust as a unidimensional construct, the trust literature has established a two-dimensional model of trust (De Jong et al., 2016; McAllister, 1995). While affective trust refers to “individuals’ feelings of emotional involvement and others’ genuine care and concern for their welfare,” cognitive trust includes perceptions of “the reliability, integrity, and competence of others” (De Jong et al., 2016: 15). This distinction builds on the assumption that individuals assess others based on two dimensions—namely, warmth and competence (Casciaro & Lobo, 2008). For example, the literature on social networks has connected ties based on friendship with affective trust and ties based on instrumental support with cognitive trust (Chua, Ingram, & Morris, 2008). In the context of leadership, Tomlinson, Schnackenberg, Dawley, and Ash (2020) show that a follower's perception of his or her supervisor's benevolence is important for developing affective trust in the supervisor, while a supervisor's perceived ability and behavioral integrity give rise to cognitive trust. Moreover, research has found that cognitive trust in a leader emerges before affective trust in him or her, suggesting that emotional ties need more time to develop (Schaubroeck, Lam, & Peng, 2011). Importantly, affective and cognitive trust also differ with respect to behavioral outcomes (e.g., Schaubroeck et al., 2011; Zhu & Akhtar, 2014). For example, in the leadership context, affective trust in a leader helps a work team develop the belief of being in a safe environment, whereas cognitive trust in a leader generates favorable perceptions of the focal team's capabilities (Schaubroeck et al., 2011). At the level of peers, affective trust influences individuals’ willingness to share knowledge with a coworker, while cognitive trust impacts individuals’ willingness to use tacit knowledge provided by a co-worker (Holste & Fields, 2010).

In sum, there are substantial differences between affective and cognitive trust, and these differences are likely to shape individuals’ behavior in their work environments in different ways. Based on this two-dimensional model of trust, we expect that an NVT member's affective and cognitive trust in the rest of the NVT shape his or her perception of situational strength, mitigating or enhancing the freedom to act according to his or her personality and thus affecting the greed–UPB relationship.

Affective trust

NVT members who perceive higher levels of affective trust in their teammates are likely to feel accepted by them (Zhu & Akhtar, 2014). Specifically, when affective trust is high, team members tend to interpret their teammates’ behavior in a positive way and reframe potentially negative behavior as constructive and helpful (Samian et al., 2021). Thus, team members high in affective trust may believe that their teammates generally judge them and their activities in a favorable way (De Jong et al., 2016; Schaubroeck et al., 2011). Moreover, NVT members high in affective trust tend to believe that they are in a sufficiently safe environment and have the freedom to try out different types of behavior (Schaubroeck et al., 2011). When perceiving high levels of affective trust, an NVT member does not expect penalties or punishments to arise from his or her behavior because affective trust is connected to forgiveness (Samian et al., 2021).

Thus, a situation characterized by a high level of affective trust can be considered a weak situation for an NVT member as it involves no or little supervision and thus the freedom to make decisions (Judge & Zapata, 2015). Such a situation leads to “behavioral expressions that are in line with one's basic personal tendencies” (Judge & Zapata, 2015: 1151). Therefore, greedy NVT members who perceive higher levels of affective trust in their teams are more likely to act in line with their greedy desires and to engage in UPB. When they perceive high levels of affective trust, greedy NVT members might be particularly tempted to “tweak the chances” of their ventures in the hopes of enhancing their prospects of financial rewards—that is, more valuable shares (Haynes, Hitt et al., 2015)—as well as increasing their status and fame as successful entrepreneurs (Hamilton, 2000; Schwienbacher, 2007). Thus, we expect that the relationship between an NVT member's greed and UPB is particularly strong when this team member perceives a high level of affective trust in the NVT.

In contrast, a low level of affective trust in the NVT is likely to reduce an NVT member's perceived freedom to express his or her personality. An NVT member who perceives lower affective trust in his or her NVT is less likely to feel accepted and judged favorably by his or her teammates (De Jong et al., 2016; Schaubroeck et al., 2011). Thus, he or she might worry that some forms of behavior will be rejected by the other team members or have negative implications for him or her (Samian et al., 2021). As such, situations characterized by low affective trust represent strong situations in which situational barriers reduce the expression of NVT members’ greed. Based on these arguments, we propose the following:

Hypothesis 1: The relationship between an NVT member's level of greed and his or her tendency to engage in UPB is moderated by the member's affective trust in the NVT such that this relationship is more positive at higher levels of affective trust compared to lower levels of affective trust.

Cognitive trust

An NVT member who perceives high levels of cognitive trust in the NVT is aware of his or her teammates’ capabilities and skills (McAllister, 1995). This member is likely to be convinced that the NVT is able to achieve high levels of performance (Schaubroeck et al., 2011). Based on this perceived competence, the NVT member is likely to believe that his or her teammates can understand how each team member performs his or her tasks. Thus, he or she might anticipate that these teammates are able to monitor his or her own behavior and likely feels accountable toward his or her teammates; that is, the member expects that he or she “may be called on to justify [his or her] beliefs, feelings, and actions” to the other team members (Lerner & Tetlock, 1999: 255). Importantly, these feelings of accountability are likely to shape the NVT member's behavior independent of the teammates’ actual engagement in monitoring behavior (Lerner & Tetlock, 1999). Perceived accountability reduces individuals’ tendency to self-enhance (Sedikides, Herbst, Hardin, & Dardis, 2002), increases task performance (Mero, Guidice, & Werner, 2012), and facilitates careful information processing within teams (Scholten, van Knippenberg, Nijstad, & De Dreu, 2007). Specifically, Mero et al. (2012: 1629) explain these positive outcomes by arguing that people “react to accountability in ways that optimize their position within the social system,” which triggers behavior that is socially desirable. Thus, an NVT member's accountability toward his or her teammates triggered by cognitive trust is likely to limit his or her perceived freedom to act (Judge & Zapata, 2015; Meyer et al., 2010). Consequently, high levels of cognitive trust in the NVT present a strong situation involving high levels of supervision and no or little freedom to make decisions, thus potentially reducing behavioral expressions of the NVT member's personality. Therefore, a greedy NVT member who has a higher level of cognitive trust in the NVT is less likely to act in accordance with his or her greedy desires and, as a consequence, is less likely to engage in UPB to pursue these desires.

In contrast, if an NVT member experiences a lower level of cognitive trust in the NVT, he or she is less convinced about other team members’ capabilities and skills (McAllister, 1995). Thus, this member does not expect teammates to be able to understand and monitor his or her behavior in venture-related tasks. This lack of (anticipated) monitoring likely reduces the NVT member's perceived accountability and creates a rather weak situation in which the NVT member feels free to engage in behavior he or she wants to. In such a weak situation, the NVT member's personality is likely to be intensely expressed in his or her behavior (Judge & Zapata, 2015). As a consequence, an NVT member high in greed who has a lower level of cognitive trust in the NVT is more likely to act in accordance with his or her greedy desires and is thus more likely to engage in UPB. Based on these arguments, we propose the following:

Hypothesis 2: The relationship between an NVT member's level of greed and his or her tendency to engage in UPB is moderated by the member's cognitive trust in the NVT such that this relationship is less positive at higher levels of cognitive trust compared to lower levels of cognitive trust.

Methodology

Sample and Data Collection

Our sampling frame consisted of members of NVTs—that is, groups of at least two individuals who worked on operational and strategic tasks in their young firms (i.e., firms that are maximum 6 years old; Amason, Shrader, & Tompson, 2006). To build our sampling frame, we screened data from online venture databases (i.e., German entrepreneurship databases), attended events, visited incubators and accelerators (e.g., from local universities or private programs), and contacted entrepreneurs from our network in the metropolitan region of the first author's university. Overall, we identified 2,183 ventures and tried to contact them to ask for their participation, not knowing if all were still in business at this time. We were able to get in touch with 574 of these ventures personally via phone and e-mail. Using leaflets, a website, and business cards with a QR code, we informed them about the purpose of our study. To address potential threats of overly low response rates sometimes reported in survey research, we offered incentives for participation. Particularly, we offered participants the opportunity to publish job postings and student projects on our university's website. Furthermore, we informed prospective participants that they would be invited to an event at our institute and would receive individualized reports of our study. Lastly, we offered a certificate of participation to every team.

Overall, 285 NVTs registered to participate on the project's website. We decided to exclude corporate spinoffs as NVT members’ autonomy to make decisions (and thus to engage in UPB) may have been constrained by the parent companies (Koster, 2004). Further, some of the ventures that were listed as team-based ventures turned out to be managed by single entrepreneurs, so we also excluded them. Following Ensley, Pearson, and Amason (2002), we considered an individual to be part of an NVT and thus a legitimate participant if he or she met at least two out of three following criteria: the individual (1) founded the focal company, (2) currently held at least a 10% equity share, and/or (3) actively participated in strategic decision making for his or her venture. In total, 128 NVTs consisting of 281 members participated in our study.

One potential limitation from survey methodology is common method variance (see below). To reduce the effect of common method variance, we sent our participants two surveys at different timepoints (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003) 3 months apart. The first survey included the independent and moderating variables (greed, cognitive trust, and affective trust). The second survey captured entrepreneurs’ UPB. Participants who did not complete both surveys were excluded from our analysis, resulting in a final dataset of 233 NVT members nested in 111 teams. Based on the 574 ventures personally contacted, our venture-level response rate was 19.34%. Another limitation of the survey methodology is that missing answers from nonrespondents could bias our results and limit their generalizability. Thus, we follow the recommendations by Werner, Praxedes, and Kim (2007) to test for a potential nonresponse bias. As data from the nonrespondents were not available, we compared early versus late respondents following the assumption that late respondents are more similar to nonrespondents (Armstrong & Overton, 1979). We counted the days until the participants responded to the surveys and split the sample based on the median in early and late respondents. Then, we tested if these groups significantly differed in any of the variables included in our study, which was not the case (all p values > .25). As such, we are confident that nonresponse bias does not substantially impact our findings.

On average, the NVT members in our sample were 35.23 years old (SD = 7.83), and 12.45% were female. ¹ Regarding participants’ highest level of education achieved, 21.89% held a bachelor's degree, 52.36% held a master's degree (or equivalent), and 12.02% held a doctoral degree. The NVT members varied in their educational background: 36.05% had a background in business or economics, 25.32% in engineering, 14.16% in natural science or mathematics, 9.87% in information technology, 4.29% in social sciences (including languages, culture, teaching, law), 4.72% in creative arts (e.g., media studies, design, art, or literature), 3.86% in medicine or health care, and 1.72% were not specialized. The average team size was 2.88 (SD = 1.18) and ranged from two to eight members. This size is consistent with previous studies on early ventures (e.g., 2.8 in Hellmann & Wasserman, 2017; or 2.3 in Nuscheler, Engelen, & Zahra, 2019) and roughly matches the team size of 2.4 members reported in a representative survey of startups in Germany (Ripsas & Tröger, 2015). On average, the teams had worked together for 2.92 years (SD = 1.28) in their current compositions. The ventures were 2.64 years old (SD = 1.30) at the start of our data collection and employed 5.48 full-time employees (SD = 6.15) on average. The ventures were active in diverse sectors, with 54.51% operating in the computer hardware and software industry, 21.89% in the services (professional and others) industry, 3.86% in e-commerce, 8.58% in consumer products, 9.01% in life sciences, and 2.15% in science (materials and physical).

Variables

Appendix 1 includes a summary of all the items for the dependent, independent, and moderator variables of our study. Figure 1 presents our variables and their measurement levels.

Dependent variable

We measured UPB, with the six-item scale by Umphress, Bingham, and Mitchell (2010). The scale captures to what extent individuals would act unethically to help their organizations. We adapted the scale slightly so that it better fits the entrepreneurial context by replacing “organization” in the original scale with “venture.” We measured participants’ agreement with the six statements on a 1 (“not at all”) to 7 (“completely”) Likert-type scale. The Cronbach's alpha for UPB is .80, indicating high reliability of the scale (Hair, Anderson, Babin, & Black, 2010). We measured this variable in the second questionnaire of our study.

Independent and moderator variables

We captured greed using the seven-item Dispositional Greed Scale from Seuntjens, Zeelenberg, van de Ven et al. (2015), which we administered to participants in the first questionnaire. This scale has shown good internal consistency (Liu et al., 2019) and covers immaterial and material desires that play an important role in managerial and entrepreneurial decisions (Haynes et al., 2017). We asked our participants to indicate their agreement with the scale's statements on a Likert scale ranging from 1 (“strongly disagree”) to 7 (“strongly agree”). The Cronbach's alpha is .82, indicating high reliability (Hair et al., 2010).

We measured affective and cognitive trust in the first survey with five and six items, respectively. The scales developed by McAllister (1995) are well established and reliable (e.g., Ng & Chua, 2006; Wang, Tomlinson, & Noe, 2010). We made some slight adaptations to account for the NVT context. These adjustments included replacing “us” or “we” in the original scale with the term “our founding team.” Furthermore, when the original scale referred to coworkers as “individuals,” “persons,” or “coworkers,” we replaced the term with “my team members” or “our team members.” We captured our participants’ responses on Likert-type scales ranging from 1 (“not at all”) to 7 (“completely”). The Cronbach's alphas are .80 for both types of trust.

Control variables

At the individual level, we controlled for participants’ tendency to answer in a socially desirable way because of potential impression-management bias when answering questions about their UPB (Umphress et al., 2010). Specifically, we used a short form of the Marlow-Crowne's social desirability scale developed by Strahan and Gerbasi (1972). Second, we relied on the short scales by Gosling, Rentfrow, and Swann (2003) to measure the Big Five personality traits: extraversion, emotional stability, openness, conscientiousness, and agreeableness. Given that the scales measure broad personality traits with only two items per trait (Judge, Higgins, Thoresen, & Barrick, 1999), the Cronbach's alphas are acceptable (extraversion: .73; emotional stability: .55; openness: .34; conscientiousness: .59; agreeableness: .33). By controlling for the Big Five personality traits, we ensured that the observed relationship between greed and UPB was greed-specific and not confounded by major NVT members’ personalty traits (Judge, LePine, & Rich, 2006). Third, we controlled for founding experience as a major driver of an entrepreneur's behavior (Ucbasaran, Westhead, & Wright, 2009). For instance, experienced entrepreneurs tend to be more overconfident (Forbes, 2005), which might motivate them to behave unethically. Specifically, we captured the number of ventures the NVT members had started prior to their current ventures. Fourth, we controlled for industry experience as industry norms and rules may affect entrepreneurs’ behavior (Dimov, 2010). We asked the participants to indicate how many years they had worked in their ventures’ current industry. Fifth, we controlled for business/economics education as individuals with these educational backgrounds tend to be higher in greed than individuals with other backgrounds (Wang, Malhotra, & Murnighan, 2011). We coded participants with educational backgrounds in business or economics as 1 and others as 0. Sixth, we accounted for participants’ age in years as older individuals tend to act more ethically in business contexts (Ruegger & King, 1992). Seventh, we controlled for gender as men tend to act more unethically than women (Lee, Pitesa, Pillutla, & Thau, 2017). Men were coded as 0, and women were coded as 1. Eighth, we controlled for equity ownership because entrepreneurs’ tendency to engage in UPB may critically depend on whether they own (part of) their ventures or not. Equity ownership was coded as 1, whereas no equity ownership was coded as 0.

We also controlled for potentially important constructs at the venture and team levels—that is, Level 2 in our hierarchical linear model (see below). If the answers between members of the same team diverged, we either checked secondary data sources, such as official documents and the venture's website, or we reached out to the team to resolve any disagreement. Specifically, we controlled for team age. ² Team age is defined as the number of years that a team has worked together in its current composition (Breugst & Shepherd, 2017). As both forms of trust need time to develop (Schaubroeck, Peng, & Hannah, 2013), we considered team age to be an important control variable. Moreover, we controlled for venture size, measured as the number of employees. Venture size can impact resource availability (Nason & Wiklund, 2015), which may affect NVT members’ behavior (Zimmerman & Zeitz, 2002). Moreover, we included NVT size as team size may impact the development of trust (Pinjani & Palvia, 2013) and monitoring among team members (Liang, Rajan, & Ray, 2008). We also controlled for members leaving an NVT (“No. people left”) during the time frame of our study. We did so by asking participants in the first and second questionnaires to provide the names of all NVT members and by counting how many people left each venture. ³ Changes in NVT composition may impact a member's perceived trust in the rest of the team. ⁴ Finally, we controlled for venture industry as entrepreneurs are more likely to act unethically in hostile environments (Shepherd, Patzelt, & Baron, 2013). We asked participants to indicate the industry their ventures primarily operated in and included dummy variables (the reference category was the computer hardware and software industry).

Statistical Analysis

Although our conceptual model focuses on individual-level constructs, NVT members are nested within teams in our data. Importantly, “team contexts are inherently multilevel” (Carter, Carter, & DeChurch, 2015: 1408) because members of one team are more similar to each other than to members of different teams, which violates the nonindependence assumption of an ordinary least squares regression and is likely to bias regression estimates (Hox, Moerbeek, & Van de Schoot, 2017). Consistent with this idea, there was substantial within-team agreement on UPB among the members in each team (median rwg(j) = .83). To take these intrateam similarities into account, we used hierarchical linear modelling (HLM; Hox et al., 2017). Based on Aguinis, Gottfredson, and Culpepper (2013), we centered individual-level variables at the group mean and venture-level variables at the grand mean.

Common Method Variance

Following the recommendations by Podsakoff et al. (2003), we accounted for common method variance, first, by the nature of our data collection procedure. Specifically, we temporally separated the measurements of the independent/moderator variables and the dependent variable by measuring greed, cognitive trust, and affective trust in the first questionnaire and UPB in the second questionnaire (Podsakoff et al., 2003). Second, we followed the instructions provided by Williams, Hartman, and Cavazotte (2010) and applied confirmatory factor analysis (CFA) with a marker variable, which is a suitable way to control for the influence of common method variance in our model (Lindell & Whitney, 2001; Malhotra, Kim, & Patil, 2006). A marker variable needs to be theoretically unrelated to substantial variables; that is, the independent and dependent variables as well as the moderating variables should not be correlated with the substantial variables. We chose individuals’ resilience as a marker variable (measured with a three-item scale reported by Luthans, Avolio, Avey, & Norman, 2007) as resilience is theoretically unrelated to the substantial variables and as the correlations between resilience and the substantial variables are almost 0 (i.e., r = .04 for greed, r = −.03 for UPB, r = .01 for cognitive trust, r = −.03 for affective trust) and not statistically significant. Applying the procedure as outlined by Williams et al. (2010) indicated that common method variance is unlikely an issue in our data. For instance, there is no evidence of shared common method variance as the Method-U model ⁵ does not fit the data significantly better than the baseline model (Δχ² = .05, df = 1, p = .8166). Furthermore, the Method-R model is not statistically significantly different than the Method-U model (Δχ² = .01, df = 6, p = 1).

Results

Descriptive Statistics and Correlations

Table 1 contains the mean values, standard deviations, minimum and maximum values, medians, and modes of our variables. Table 2 contains the correlations between our variables. While the means for affective trust (6.22) and cognitive trust (6.07) seem high, they are in line with prior studies (McAllister, 1995; Webber, 2008), especially in the entrepreneurship context (e.g., Khan, Breitenecker, Gustafsson, & Schwarz, 2015), in which members self-select into NVTs (Lazar et al., 2020). We note that despite these high means, the minimum and maximum values of these variables indicate substantial variance in our data (affective trust: var = .63, min = 2.80, max = 7.00; cognitive trust: var = .64, min = 3.33, max = 7.00). The relatively high correlation between these two variables of .64 is as expected (e.g., Schaubroeck et al., 2013). We calculated the variance inflation factors to check for multicollinearity. As the highest variance inflation factor is 1.91, multicollinearity is unlikely to represent a major issue in our data (Hair et al., 2010).

Table 1

Descriptive Statistics

#	Variable	M	SD	Min.	Max.	Median	Mode
1	UPB	3.26	1.19	1.00	7.00	3.17	3.00
2	Greed	3.02	1.16	1.00	7.00	2.86	2.71
3	Affective trust	6.22	.80	2.80	7.00	6.40	7.00
4	Cognitive trust	6.07	.80	3.33	7.00	6.17	7.00
5	Social desirability	.54	.21	.00	1.00	.57	.57
6	Extraversion	5.09	1.38	1.00	7.00	5.50	6.00
7	Emotional stability	5.71	1.05	2.00	7.00	6.00	6.00
8	Openness	6.08	.82	3.50	7.00	6.00	6.50
9	Conscientiousness	5.82	.97	2.50	7.00	6.00	6.00
10	Agreeableness	4.81	1.21	1.00	7.00	5.00	5.00
11	Founding experience	1.11	1.81	.00	14.00	.00	.00
12	Industry experience	5.59	5.72	.00	40.00	4.00	2.00
13	Business education	.36	.48	.00	1.00
14	Age participant	35.23	7.83	21.00	64.00	34.00	32.00
15	Gender	.12	.33	.00	1.00
16	Equity ownership	.99	.11	.00	1.00
17	Team age	2.92	1.28	.50	6.00	2.83	1.16
18	Venture size	5.48	6.15	.00	35.00	3.00	2.00
19	Team size	2.88	1.18	2.00	8.00	3.00	2.00
20	No. people left	.06	.30	.00	2.00	.00	.00
21	Industry (services)	.22	.41	.00	1.00
22	Industry (E-commerce)	.04	.19	.00	1.00
23	Industry (consumer prod.)	.09	.28	.00	1.00
24	Industry (life sciences)	.09	.29	.00	1.00
25	Industry (science)	.02	.15	.00	1.00

Table 2

Correlation Table

#	Variable	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18	19	20	21	22	23	24
1	UPB
2	Greed	.12†
3	Affective trust	−.10	−.17*
4	Cognitive trust	−.13*	−.16*	.64**
5	Social desirability	−.22**	−.03	.01	.01
6	Extraversion	.16*	−.05	.16*	.18**	.00
7	Emotional stability	−.09	−.18**	.07	.09	.21**	.06
8	Openness	−.03	.05	.08	.16*	.10	.26**	.11†
9	Conscientiousness	−.13*	−.06	.04	.07	.20**	−.01	.22**	.09
10	Agreeableness	−.12*	−.10	.13*	.12†	.26**	.08	.19**	.14*	.21**
11	Founding experience	−.03	.05	−.07	−.01	−.06	.10	.04	−.04	−.09	.04
12	Industry experience	−.03	−.14*	−.01	.14*	−.10	−.05	−.04	−.02	−.02	.03	.16*
13	Business education	.00	.06	.02	−.03	.04	.00	−.01	−.09	−.04	.05	.09	.03
14	Age participant	−.07	−.16*	.01	.20**	.02	.04	.01	.02	.06	.15*	.33**	.57**	.07
15	Gender	−.03	−.11	.02	.01	.01	.06	−.08	.05	.08	.12†	−.10	−.05	.04	.06
16	Equity ownership	.04	.02	−.06	−.09	−.10	.02	−.03	.03	−.12†	−.11†	.07	.07	.01	.11†	−.30**
17	Team age	−.08	−.18**	.20**	.14*	−.08	−.07	−.08	−.06	−.11	.06	.00	.21**	−.02	.19**	.00	.07
18	Venture size	.04	−.12†	.07	.12†	−.07	.01	.06	−.10	−.02	−.05	.15*	.11†	−.02	.11†	−.14*	.04	.28**
19	Team size	.04	.18**	−.13*	−.17**	.00	.01	.13*	−.12†	−.02	−.06	−.09	−.23**	.02	−.38**	−.05	−.17**	−.32**	−.09
20	No. people left	.11†	.08	−.15*	−.13*	−.05	.02	.08	−.04	.04	.05	−.07	−.11†	−.02	−.12†	.19**	−.49**	−.26**	−.09	.30**
21	Industry (services)	−.03	.04	−.03	.03	−.08	.04	.11	.05	.03	.03	.08	.15*	.12†	.05	.02	−.03	−.09	.05	.07	−.06
22	Industry (E-commerce)	.05	.06	.14*	.01	.03	.05	.08	.08	−.05	.19**	−.01	−.09	−.01	−.03	−.01	.02	.09	.03	−.04	−.04	−.11
23	Industry (consumer prod.)	−.04	−.16*	.04	.04	−.01	.06	.00	−.04	−.06	.00	.02	−.06	−.04	.06	.12†	.03	.02	−.04	−.15*	.05	−.16*	−.06
24	Industry (life sciences)	−.07	.08	.01	.05	.01	−.04	−.10	.06	.02	−.03	−.06	.00	−.05	.13*	.06	.04	.01	−.05	−.06	−.01	−.17*	−.06	−.10
25	Industry (science)	.03	−.08	.08	.06	.02	.00	.00	−.03	.00	−.04	−.09	−.04	−.05	−.05	.03	.02	.19**	.03	−.03	−.03	−.08	−.03	−.05	−.05

Note. ^†p < .10. *p < .05. **p < .01.

Hypothesis Testing

We present the results of our HLM analysis in Table 3. We provide three main models. Model 1 contains the control variables. In Model 2, the main effect of greed is added. Model 3 is the full model including the interaction effects of greed with affective trust and greed with cognitive trust. Models 4, 5, and 6 are further added as robustness checks (see below).

Table 3

HLM Results

	Model 1			Model 2			Model 3 (Full Model)			Model 4 (Test)			Model 5 (Test)			Model 6 (test)
	Coeff.	Robust SE	p	Coeff.	Robust SE	p	Coeff.	Robust SE	p	Coeff.	Robust SE	p	Coeff.	Robust SE	p	Coeff.	Robust SE	p
Constant	3.334	.087	.000	3.334	.087	.000	3.362	.089	.000	3.256	.076	.000	3.325	.093	.000	3.371	.095	.000
Control variables
Social desirability GC	−.449	.596	.451	−.511	.594	.390	−.386	.587	.512				−.336	.597	.574	−.255	.574	.656
Extraversion GC	.153	.086	.074	.162	.083	.052	.147	.073	.044				.153	.077	.046	.160	.077	.038
Emotional stability GC	−.014	.080	.866	.014	.084	.864	−.009	.078	.907				−.002	.083	.977	−.028	.080	.724
Openness GC	.028	.148	.850	.026	.148	.860	.056	.141	.689				.018	.153	.907	.034	.149	.818
Conscientiousness GC	−.085	.096	.378	−.074	.094	.431	−.092	.098	.347				−.085	.098	.385	−.095	.103	.355
Agreeableness GC	−.111	.088	.210	−.108	.085	.208	−.098	.086	.257				−.107	.089	.226	−.116	.097	.232
Founding experience GC	−.030	.060	.620	−.035	.063	.574	−.028	.060	.644				−.034	.061	.582	−.014	.060	.811
Industry experience GC	−.040	.024	.087	−.041	.024	.082	−.033	.022	.141				−.032	.023	.154	−.027	.022	.223
Business education GC	.010	.201	.961	−.026	.203	.899	.060	.211	.775				.062	.214	.773	.019	.232	.936
Age participant GC	.017	.016	.291	.019	.017	.257	.026	.015	.080				.026	.015	.087	.026	.015	.073
Gender GC	.004	.386	.992	.030	.387	.938	.032	.367	.930				.063	.381	.869	.052	.369	.889
Equity ownership GC	−.080	.465	.864	−.163	.502	.745	.156	.456	.733				−.146	.524	.780	.336	.651	.606
Team age C	−.075	.066	.260	−.074	.066	.263	−.054	.067	.416				−.043	.066	.511	−.018	.067	.784
Venture size C	.013	.010	.182	.013	.010	.187	.009	.009	.319				.010	.009	.271	.009	.009	.351
Team size C	−.014	.052	.782	−.016	.053	.766	−.044	.057	.434				−.044	.060	.458	−.014	.112	.903
No. People left C	.468	.266	.078	.467	.273	.087	.711	.248	.004				-	-	-	1.452	.623	.020
Industry (services)	−.134	.218	.539	−.131	.218	.548	−.154	.213	.470				−.172	.216	.425	−.083	.239	.728
Industry (e-commerce)	.276	.730	.706	.271	.728	.710	−.252	.517	.626				−.264	.523	.614	−.266	.518	.608
Industry (consumer prod.)	−.199	.314	.526	−.186	.318	.559	−.249	.311	.423				−.461	.267	.084	−.295	.294	.316
Industry (life sciences)	−.309	.163	.058	−.310	.164	.058	−.302	.156	.053				−.326	.163	.045	−.284	.159	.074
Industry (science)	.301	.398	.450	.299	.399	.454	.256	.374	.494				.224	.382	.558	.273	.373	.464
Main effects
Greed GC				.090	.114	.430	.020	.108	.856	.025	.098	.796	.036	.112	.745	−.064	.096	.504
Affective trust GC							.265	.170	.119	.277	.154	.072	.248	.192	.197	.363	.178	.041
Cognitive trust GC							−.244	.200	.221	−.195	.192	.309	−.175	.223	.433	−.279	.221	.207
Interaction effects
Greed GC × Aff. Trust GC							.542	.233	.020	.454	.258	.079	.526	.238	.027	.421	.241	.081
Greed GC × Cog. Trust GC							−.892	.228	.000	−.803	.235	.001	−.878	.238	.000	−1.102	.269	.000
Observations	233			233			233			233			224			219
Number of groups	111			111			111			111			105			108
Snijders/Bosker R-squared Level 1	.080			.083			.162			.079			.158			.195
Snijders/Bosker R-squared Level 2	.080			.078			.159			.064			.158			.195

Note: GC = centered by the group mean; C = centered by the grand mean; prod = products; aff. = affective; cog. = cognitive.

In line with the theoretical foundations of this paper, we do not find the main effect of greed on UPB to be statistically significant (beta = .020, p = .108 in Model 3). Hypothesis 1 postulates that the relationship between an NVT member's level of greed and his or her tendency to engage in UPB is moderated by the member's affective trust in the NVT such that this relationship is more positive at higher levels of affective trust compared to lower levels of affective trust. In Model 3, we observe a statistically significant interaction effect of greed and affective trust (beta = .542, p = .020). In Figure 2, we plot the relationship between NVT members’ greed and UPB for different levels of affective trust in their teams. The x axis represents greed, and the y axis represents UPB. The two lines illustrate the relationships between NVT members’ greed and UPB under comparably low affective trust in their teams (solid line, 1 SD below the mean) and comparably high affective trust in their teams (dashed line, 1 SD above the mean), including the 90% confidence intervals (two sided) across all possible values for greed. Simple slope analysis (Aiken, West, & Reno, 1991) reveals a positive relationship between greed and UPB when affective trust is comparably high (beta = .291, p = .050); thus, with increasing greed, NVT members with comparably high affective trust in their teams act more unethically in favor of their ventures. For comparably low values of affective trust, the slope does not significantly differ from zero (beta = −.251, p = .137). We also relied on the Johnson-Neyman technique (Bauer & Curran, 2005) to identify the range of values of affective trust for which the relationship between greed and UPB is significantly different from zero. We found that the relationship is negative for low values of affective trust (centered values below −1.25) and positive for high values of affective trust (centered values above .45). ⁶ These patterns are consistent with Hypothesis 1.

Figure 2

Relationship between Greed and Entrepreneurs’ Unethical Behavior Contingent on Affective Trust

Hypothesis 2 states that the relationship between an NVT member's level of greed and his or her tendency to engage in UPB is moderated by the member's cognitive trust in the NVT such that this relationship is less positive at higher levels of cognitive trust compared to lower levels of cognitive trust. Model 3 indicates that the coefficient of the interaction between greed and cognitive trust is statistically significant (beta = −.892, p < .001). Figure 3 illustrates the interaction effect. We plot the lines for NVT members with lower (solid line, 1 SD below the mean) and higher cognitive trust in their teams (dashed line, 1 SD above the mean), including the 90% confidence intervals (two sided) across all values for greed. Simple slope analysis (Aiken et al., 1991) shows that the positive relationship between greed and UPB is significantly different from zero when NVT members’ cognitive trust is low (beta = .459, p < .001); that is, with increasing greed, NVT members with comparably low cognitive trust in their team members act more unethically in favor of their ventures. For NVT members with high cognitive trust, the negative slope is also different from zero (beta = −.433, p = .023); thus, with increasing greed, these NVT members act less unethically in favor of their ventures. The Johnson-Neyman technique reveals that the relationship between greed and UPB is positive for low values of cognitive trust (centered values below −.15) and negative for high cognitive trust (centered values above .40). ⁷ These analyses support Hypothesis 2.

Figure 3

Relationship between Greed and Entrepreneurs’ Unethical Behavior Contingent on Cognitive Trust

Robustness Tests

In Table 3, we also report robustness tests that largely support our findings. Model 4 excludes the control variables. We observe statistically significant interaction effects at the 10% and 1% significance levels for the interactions of greed with affective trust (beta = .454, p = .079) and cognitive trust (beta = −.803, p = .001), respectively. In Model 5, all control variables of the main model are included, but all ventures that experienced changes in founding team composition during the course of our data collection are excluded. The interaction effects are statistically significant at the 5% significance level for affective trust (beta = .526, p = .027) and at the 1% significance level for cognitive trust (beta = −.878, p < .001), respectively. In Model 6, all ventures with teams consisting of more than four entrepreneurs are excluded. Similar to Model 4, we observe statistically significant interaction effects at the 10% and 1% significance levels for affective trust (beta = .421; p = .081) and cognitive trust (beta = −1.102, p < .001), respectively.

Endogeneity Tests

We also wanted to ensure our results do not suffer from endogeneity problems, such as simultaneity or selection biases (Baum, 2006; Clougherty, Duso, & Muck, 2015). Based on the available data, we identified two potential instruments for greed. First, we considered the NVT members’ self-interested values, which we captured with a three-item scale developed by Agle, Mitchell, and Sonnenfeld (1999) asking participants to indicate the importance of values like leading a comfortable/prosperous life. Second, we captured participants’ tendency to engage in self-promotion—that is, stressing “their abilities or accomplishments in order to be seen as competent by observers” (Bolino & Turnley, 1999: 190)—with a four-item scale asking them questions like how frequently in interactions with strangers they “talk proudly about [their] experience or education” (Bolino & Turnley, 1999: 196). Theoretically, both constructs should be related to greed as an excessive form of self-interest.

We relied on Stata's ivreg and ivendog commands to test for potential endogeneity issues. In a first step, we checked if our instruments are appropriate—that is, if they are, first, highly correlated with greed and, second, uncorrelated with the error term from our model (Baum, 2006). The first condition is met because both constructs correlate significantly with greed (r = .280, p < .001 for self-interested values and r = .179, p < .01 for self-promotion), and they are also significant predictors in a first-stage model predicting greed and including all control variables of our original model (beta = .274; p < .001 for self-interested values and beta = .129; p = .010 for self-promotion). To check for the second condition, we relied on the Sargan and Basman tests, χ²(1) = 1.734, p = .188 and χ²(1) = 1.552, p = .213, respectively, which were not significant, indicating that the instruments are uncorrelated with the error term. Thus, self-interested values and self-promotion indeed represent appropriate instruments for greed. Then, we conducted a Durbin-Wu-Hausman test, which was not significant, χ²(1) = .265, p = .607, suggesting that greed is likely to be an exogeneous variable (Baum, 2006). Moreover, to control for potential selection effects (Clougherty et al., 2015), we relied on the inverse Mills ratio based on the two instruments, self-interested values (Agle et al., 1999) and self-promotion tendencies (Bolino & Turnley, 1999). First, we predicted greed based on the NVT members’ self-interested values, their self-promotion tendencies, and the control variables included in our original model. Using the predicted greed values of this first-stage model, we calculated the inverse Mills ratio, which we added as a control variable in the second-stage model estimating UPB consistent with our original model. The inverse Mills ratio is not significant, and our results do not change when it is included. These results indicate that endogeneity is unlikely to be a major problem in our data.

Discussion

Greed has been studied in various fields but has remained rather unaddressed in entrepreneurship studies. This is surprising as greed is considered a particularly widespread trait among NVT members (Djankov, Yingyi, Roland, & Zhuravskaya, 2006). This study set out to understand how the NVT context shapes the manifestation of greed in UPB. In line with situational strength theory, we found that higher levels of greed are connected to more UPB when an NVT member’s affective trust in the rest of the NVT is high and when his or her cognitive trust in the NVT is low.

Theoretical and Practical Implications

Our study contributes to the emerging theoretical understanding on how managerial greed influences different types of firms (Haynes, Hitt et al., 2015). Specifically, research on established organizations has indicated that executives high in greed typically sacrifice organizational goals for their own desires (Haynes et al., 2017; Haynes, Josefy et al., 2015). In contrast, the current study builds on the idea that entrepreneurs, who often see their ventures as “an extension of themselves” (Ruvio et al., 2010: 145), would rather engage in pro-organizational behavior than (intentionally) harming their ventures to fulfill their greedy desires. Thus, our study accentuates the important theoretical distinctiveness of the new venture versus the established firm context when studying personality traits.

Furthermore, our study contributes to work on managerial greed by identifying the team context as an important facet of the work environment that influences the extent to which greed translates into managerial behavior. Prior studies have found that social cohesion within an organization (Bruhn & Lowrey, 2012) as well as board power and managerial discretion (Haynes et al., 2017) are important factors triggering the manifestation of greed in established firms. In new ventures, however, the NVT context is particularly important as NVT members engage in frequent, interdependent, and intense interactions (Blatt, 2009; Breugst et al., 2015; De Jong et al., 2013). Unraveling the impact of greed in the NVT context is important because in entrepreneurship, NVTs are common (Klotz et al., 2014), and they represent a specific context, thereby presenting the opportunity for a unique contribution to the entrepreneurship and strategic leadership literatures (Haynes, Josefy et al., 2015). Our study suggests that NVT members’ trust in their teams is key to understanding when greed manifests in behavior, but importantly, this manifestation occurs differently for different types of trust. Future theorizing and empirical work on entrepreneurs’ personality traits should consider the potential role of the NVT environment and, specifically, the role of cognitive and affective trust in one's teammates because these trust types represent weak and strong situations and can thus be important contingencies linking traits to behavior.

Our study is also of interest to entrepreneurship scholars because it adds to the conversation on the role of personality in entrepreneurial behavior. This literature has studied a range of entrepreneurship-related personality traits, such as innovativeness (Mueller & Thomas, 2001), resilience (Korber & McNaughton, 2018), and grit (Mueller, Wolfe, & Syed, 2017), but also broader concepts, such as the Big Five (for a review, see Frese & Gielnik, 2014). Surprisingly, this literature has largely neglected the concept of greed (for an exception, see Haynes, Hitt et al., 2015). While Haynes, Hitt et al. (2015) theorize on how an entrepreneurial leader's greed impacts his or her venture's human and social capital, our study suggests that considering entrepreneurs’ social context is necessary to understand how greed manifests in the NVT setting.

Second, while trust has often been considered a valuable characteristic for NVTs (Blatt, 2009; Wang & Wu, 2012), entrepreneurship scholars have also pointed to its potentially negative implications, for instance, through overtrust or blind trust (Goel & Karri, 2006; Kautonen, Zolin, Kuckertz, & Viljamaa, 2010). We add to this discussion on the potential positive and negative sides of trust in entrepreneurship by unraveling a potential mechanism through which trust may facilitate or prevent UPB. In particular, it appears that the two forms of trust can represent weak or strong situations and thus play an important role in the expression of individuals’ personalities. Furthermore, by showing how cognitive and affective trust affect the relationship between greed and UPB in different ways, our study points to the importance of distinguishing between trust dimensions in the NVT context. This distinction is particularly important because prior studies have often used unidimensional operationalizations of trust that do not account for potential differences in entrepreneurs’ affect-driven versus cognition-driven behavior.

Third, we add to the emerging stream of research on UPB. So far, this literature has focused on understanding UPB in established enterprises. In particular, it has explored how individuals’ characteristics, such as Machiavellianism (Castille et al., 2018) or psychological entitlement (Lee et al., 2019); their identification with their organizations (Chen, Chen, & Sheldon, 2016; Umphress et al., 2010); and their work environment, such as superiors’ leadership style (Effelsberg, Solga, & Gurt, 2014; Miao, Newman, Yu, & Xu, 2013) or job security (Ghosh, 2017), relate to UPB. We add to this stream of literature by investigating UPB in a special setting—namely, that of new ventures. Entrepreneurship represents a theoretically interesting context as entrepreneurs’ interests are usually closely entwined with their organizations, which makes this context different from the setting of executives in established firms. Thus, it appears that the literature on UPB can gain new theoretical insights by exploring the entrepreneurial context and its unique characteristics.

Indeed, while entrepreneurship research has not explicitly studied the concept of UPB, it points to the unethical behavior of founders and new ventures. For example, entrepreneurs’ unethical behavior has been explored in terms of bribery to ensure access to governmental support (Baron et al., 2018), lying to stakeholders (Pollack & Bosse, 2014; Theoharakis et al., 2021), and the exploitation of opportunities that cause environmental harm (Shepherd et al., 2013). These studies suggest that entrepreneurs’ unethical behavior may be based on their lack of resources and the wish to present themselves and their ventures in a more favorable light (Pollack & Bosse, 2014) or on complex psychological processes, such as moral disengagement (Shepherd et al., 2013). Our study complements insights into entrepreneurs’ unethical behavior by establishing entrepreneurs’ greed as a potential explanation for such behavior. A future focus for research on entrepreneurial personality traits (for a review, see Frese & Gielnik, 2014) may contribute to understanding why some entrepreneurs behave unethically while others do not as well as under what conditions (i.e., situational strength) this is more or less the case. Such future research endeavors into unethical behavior could purposefully integrate NVT characteristics alongside personality traits.

Our study also has practical implications for NVTs, their members, and external stakeholders of new ventures. While engaging in UPB may, under some circumstances, look like a promising way forward for new ventures, prior studies indicate that UPB can have detrimental effects on organizations in the long run. Specifically, it may erode stakeholder interests, harm organizational reputation, and diminish public trust in the organization (Graham et al., 2020; Umphress & Bingham, 2011). Our study provides insights into the development of UPB in the NVT context, pointing to the role of greed and trust within NVTs. As the antecedents of affective and cognitive trust have been well documented (Schaubroeck et al., 2011; Tomlinson et al., 2020), practitioners working in or with an NVT could be cautious about the potentially detrimental effects of affective trust if NVT members tend to be greedy. The weak situation created by high levels of affective trust could be counteracted, for example, by including more structured processes, formal control mechanisms, and some peer supervision in the NVT (Judge & Zapata, 2015). On a positive side, our work also suggests that the development of cognitive trust can create strong situations and thus prevent unethical behavior by greedy entrepreneurs. It might be beneficial to foster the development of cognitive trust, for example, by highlighting team members’ reliability and fairness (McAllister, 1995) as well as their responsibility (Schaubroeck et al., 2011).

Limitations

In line with our research focus, we opted to conduct our research in a homogeneous setting—namely, that of early-stage ventures in the same geographic region. However, as firms mature, entrepreneurs’ shares may get diluted, and entrepreneurs’ self-interest may be less aligned with their ventures’ goals, potentially leading to more self-focused and less venture-focused (unethical) behavior. Moreover, different cultural values (House, Hanges, Javidan, Dorfman, & Gupta, 2004) may also influence entrepreneurs’ propensity to engage in UPB and NVTs’ facilitating or preventing role in this context. Thus, our study represents an important first step to understand the relationship between greed and UPB but also calls for more research in different contexts.

We focused on understanding the role of different trust dimensions as contingencies of the greed–UPB relationship. This focus was particularly warranted given the potential impact of an NVT member's affective and cognitive trust in the rest of the team on his or her freedom to act in line with his or her personality. However, also other relevant constructs describing the NVT context could shape the relationship between greed and UPB, such as perceptions of power distribution within NVTs (Xie, Feng, & Hu, 2020) or team climate, which includes shared perceptions and goals (Anderson & West, 1998). Future studies may incorporate these constructs.

Avenues for Future Research

Beyond the theoretical implications described above, our study also provides guidance towards theoretical perspectives opening up directions for future research. First, social contagion theory (Burt, 1987) suggests that behavior can be contagious within social entities. For example, effort can be contagious among members of NVTs under certain conditions (Breugst, Patzelt, & Shepherd, 2020). By consequence, UPB may also be contagious within the NVT. Future research could investigate if the development of UPB in NVTs can also be based on social contagion. Based on our findings, it would also be interesting to understand the role of trust in these social contagion processes. For example, it might be possible that higher levels of affective trust and lower levels of cognitive trust can intensify the social contagion of UPB within the NVT.

Our findings can also be considered from the perspective of stakeholder theory (Freeman, 1984). While we focused on the role of NVTs for understanding the greed–UPB relationship, also stakeholders external to the venture, such as investors or mentors, may affect the strength of entrepreneurs’ situations, depending on the freedom they give entrepreneurs to engage in specific behaviors, including unethical behavior. Future research could therefore assess how such stakeholders can affect the greed–UPB relationship by shaping an NVT member's perceptions of situational strength.

Conclusion

Entrepreneurs are sometimes depicted as greedy individuals who use their ventures to fulfill their insatiable desires. In this study, we investigated contingencies stemming from the NVT context that create weak or strong situations and thus either facilitate or inhibit an NVT member's greed turning into UPB. Particularly, we highlight the importance of affective and cognitive trust toward fellow NVT members in this context. Thus, we advance theory and sensitize practitioners to account for greed as a potential personality trait of entrepreneurs and for the role of the NVT environment in channeling the actions arising from entrepreneurs’ greedy desires.