ABSORPTIVE CAPACITY
The influence of connectedness and the governance
mechanisms on interfirm absorptive capacity
Claudio Gubitta
Master Thesis, Marketing Management
University of Groningen, Faculty of Economics and Business
1
ABSORPTIVE CAPACITY
The influence of connectedness and the governance mechanisms on interfirm
absorptive capacity
Claudio Gubitta
University of Groningen, Faculty of Economics and Business
First Supervisor:
Dr. Hans (J.) Berger
Second Supervisor:
Dr. Martijn Keizer
Neptunusstraat 15
9742JK, Groningen, the Netherlands
(+31) 0650411856
2
Table of Contents
1.Introduction ... 4 2.Literature review ... 8 2.1 Absorptive Capacity ... 8 2.2 Governance mechanisms ... 9 2.3 Connectedness ... 10 3.Hypotheses ... 11 3.1 Connectedness ... 11 3.2 Relational governance ... 13 3.3 Contractual governance ... 153.4 The relationship between the interplay of relational and contractual governance on ACAP ... 16
3.5 Conceptual model... 18 4.Methodology ... 18 4.1 Data collection ... 18 4.2 Measures ... 19 4.3 Statistical procedures ... 21 5. Results ... 22
5.1 Data validity and reliability ... 22
5.2 Reflective scales ... 22
5.3 Formative scales ... 24
5.4 Hypothesis testing ... 26
5.4.1 Buyer database ... 26
5.4.2 Supplier database ... 28
5.4.3 Summary of data analysis ... 29
6.Discussion, limitations and furthure research ... 29
6.1 Discussion ... 29
6.2 Limitations and suggestions for further research ... 34
6.3 Conclusion ... 35
3
ABSTRACT. Knowledge is one of the most important sources of competitive advantage. In interfirm relationships, the exchange of knowledge between partners determines their ability to refine their processes and improve innovation. In order to balance the risks associated with the cooperation and the need of flexibility that the relationship requires, the partners agree upon a relationship governance. In turn, the governance of the relationship impacts the ability to store new knowledge, as well as the ability to subsequently make use of it.
In this study, we investigate the relationship between the degree of recurrent exchange between partners (i.e., connectedness) and the absorptive capacity. By drawing from previous research, we will investigate the influence of relational-based governance on the relationship between connectedness and absorptive capacity. Moreover, we empirically contribute in evaluating the effect of contracts and relational norms on absorptive capacity. We will also study the interaction between the two types of governance in improving absorptive capacity.
For investigating these phenomena, our analysis adopts the knowledge-based and the transaction economics views. The analysis will be carried with a PLS-SEM analysis on database of 166 buyer-supplier dyadic relationships.
4
1. INTRODUCTION
In the last decades, a rich stream of research has focused on the way organizations can reach a competitive advantage. Particularly, knowledge has been indicated as a key competitive advantage because of the crucial role it has in boosting the firm’s innovating activities and the design of knowledge-based strategies (Grant, 1996; Laursen and Salter, 2006). In this panorama, absorptive capacity (ACAP) has gained importance since it was firstly treated in the famous Cohen and Levinthal (1990) relevant article. They defined absorptive capacity as “the ability of a firm to recognize value of new, external information, assimilate it, and apply it to commercial ends”.
Successively, the paper of Zahra and George (2002) conceptualized the ACAP as the combination of potential absorptive capacity (i.e., PACAP) and the realized absorptive capacity (i.e., RACAP); in fact, absorptive capacity refers to the ability to recognize and assimilate external knowledge (PACAP), as well as to the ability to transform and exploit it (RACAP).
Many scholars found a clear relationship between absorptive capacity and a firm’s ability to innovate or to outperform its competitors (Cepeda-Carrion et al., 2015; Tzokas et al., 2015).
5
The concept of trust has been widely accepted as the main construct behind the relational-based governance, following the common assertion in literature that trust-based governance serves as a self-enforcement mechanism (ibid.). Interestingly, Adler (2001) discusses how the classical control mechanisms of the relationship (that is, market/cost and hierarchy/authority) should be integrated with the mechanism of trust/community -precisely, the trust in a community of people-, because this way of managing the relationship is more effective in contexts where knowledge-based assets are involved. This assumes that in the governance of the relationship, the personal relations (that is, relational-based governance) and the degree of formal and informal personal interaction between partners of a community (that is, connectedness) are interrelated. Our focus is on the relational governance and on the personal contacts between people of two organizations (i.e., connectedness) in the buyer-supplier relationships. We assess the importance of the relational dimension of the relationships (vs. hierarchical/contractual) as claimed by e.g. Adler and Kwon (2002).
The relationship between connectedness and trust-based governance seems well-grounded in literature. For example, Rowley et al. (2000) recognize explicitly that “strong ties [and] dense networks serve as a trust-based governance mechanism in interfirm alliance”1. Surprisingly, Day et al. (2013) and Gargiulo and Ertug (2006) discussed the drawbacks of excessive embeddedness as a problem of excessive trust (i.e., the dark side of trust), thus substantially using trust and connectedness interchangeably. In discussing the results of the role of connectedness on governance mechanisms, also Gu et al. (2008) referred to the “hidden cost of trust” as a problem of excessive connectedness. Analyzing the organizational mechanisms associated with knowledge utilization, Jansen et al. (2009) explicitly refer to informal social relations or connectedness interchangeably.
Therefore, the way trust-based governance and connectedness were treated implies an intimate relationship between the two concepts. Levin and Cross (2004), though, partially treated this topic by underlying that while trust is a relational variable and that tie strength is a structural variable, the two concepts are clearly distinct one another, and we embrace this view. Nonetheless, through further empirical investigation we can contribute in at least one way. As also Levin and Cross (2004) explicitly recognize in their study, a rich stream of literature makes use of the two concepts interchangeably, so further empirical analysis is needed for reaching a clearer conclusion on what extent many authors properly considered them as synonyms; a clearer investigation may clarify some characteristics they share or diverge on, so that future research will be able to compare and
6
apply the two terms properly in the different contingencies of the research at hand. In this case, we will speculate on the two variables in the context of absorptive capacity. Particularly, we are going to investigate this aspect by evaluating the influence of relational governance on the relationship between connectedness and ACAP.
To sum up, although the empirical connection between relational-based governance and trust is apparent (e.g., Zaheer and Venkatraman, 1995) and that trust is seen the cornerstone of relational-based governance (Zoppo and Zenger, 2002) as well as an outcome of connectedness, we did not find any study that empirically investigated the impact of relational-based governance mechanisms on connectedness and ACAP. As said, the inclusion of trust/community into the classical transaction cost economics is particularly important for the management of knowledge-based assets (Adler, 2001). Due to the above-mentioned importance of knowledge in gaining a competitive advantage (Dyer and Singh, 1998), we investigate the mentioned interrelation of these variables in the peculiar context of absorptive capacity.
Differently stated, we address the following research question:” What is the influence of connectedness on the buyer-supplier absorptive capacity, and is that relationship moderated by the relational governance?”. By embracing an interfirm perspective (Dyer and Singh, 1998), we address this question through the analysis of 166 dyadic buyer-supplier relationships. Thus, as previously stated, the first contribution of this research is an examination of the impact of trust-based governance on the relationship that connectedness has on the capacity that firms have in order to absorb and to exploit knowledge (ACAP) in a buyer-supplier relationship.
The rationale to test the impact of relational-based governance -and not directly trust- on the effect of connectedness on absorptive capacity is twofold. In first instance, as Gargiulo and Ertug (2006) posit, “the definition of trust has occupied a considerable part of the literature on trust [itself]”2, so that trust is still a blanc concept that lacks an ultimate definition; secondly, many scholars (e.g., Day et al., 2013) called for an examination of the effect of trust-based organizational determinants (in this case, relational governance) on the (strategic) cooperation between firms (i.e., connectedness); considering the relational governance rather than trust permits to test the impact of a cooperative configuration (the relational governance) rather than a less actionable construct such as trust. The analysis of the organizational dynamics that promote absorptive capacity are particularly important because, in cooperative strategies, the diffusion of knowledge that is often
7
tacit, sticky, difficult to codify -thus difficult to spread- is crucial (Berger, 2015; Cohen and Levinthal, 1990; Dyer and Singh, 1998), and in such a way relevant in the examination of absorptive capacity as a phenomenon.
Unlike Levin and Cross (2004), we examine the impact of trust-based governance on the effect of tie strength on knowledge exchange (ACAP); In their work, Levin and Cross (2004) showed that the relationship between connectedness and knowledge transfer is mediated by two types of trust (i.e., competence- and benevolence-trust). Conversely, by embracing the call for an examination of the effect of trust-based organizational determinants (e.g., Day et al., 2013), we will broaden this perspective by using a more operationalized organizational determinant of trust (relational governance) and by showing the impact it has on the relationship between connectedness and the integration of new knowledge (ACAP) and its exploitation (RACAP), and not only on the more general knowledge transfer as in Levin and Cross (2004) study.
Secondly, we will contribute to establish whether contractual-based and relational-based governance serve as substitutes in promoting ACAP; this is in line with many scholars’ call for academic attention on the different conditions and organizational configurations that make relational-based and contracting substitutes rather than complements (e.g. Berger, 2015). Furthermore, since the focus of this paper is on the buyer-supplier relationship, and since buyers and suppliers have different perceptions on the ambiguous structural relationship dimensions (Oosterhuis et al., 2013), we expect that the parties score differently, for instance, on the perceptions of relational governance. Different perceptions about the relational norms would lead, in turn, to different perceptions of the balance between trust and contracting.
Therefore, the third contribution of this paper is to account for different perceptions that the buyers and the suppliers may have, and how these different perceptions are reflected in the different relationships among the constructs; in general, we will account for the differences between buyers and suppliers in explaining the results.
8
2.
ABSORPRTIVE CAPACITY, GOVERNANCE MECHANISM, ANDCONNECTEDNESS
2.1 ABSORPTIVE CAPACITY
The absorptive capacity (ACAP) was first defined by Cohen and Levinthal (1990) as “the ability to value, assimilate, and apply new knowledge”. As mentioned in the introduction, knowledge is a crucial intangible asset for a sustainable competitive advantage. When two partners cooperate, they share knowledge and benefit from each other’s spilled-over knowledge (Mesquita et al., 2008), so that it can be posited that “a firm’s critical resources may span firm boundaries and may be embedded in interfirm resources and routines” (Dyer and Singh, 1998). The proposed definition of ACAP stresses the multidimensionality of the construct. This multidimensionality was made clear with a famous work of Zahra and George (2002), where the authors identified four dimensions of ACAP. They defined ACAP as “a set of organizational routines and processes by which firms acquire, assimilate, transform, and exploit knowledge to produce a dynamic organizational capability”. Thus, the ACAP is comprised by four dynamic capabilities, that is, acquisition, assimilation, transformation, and exploitation. In particular, these dynamic capabilities contribute complementarily to the success of firms’ organizational learning, since the “capability to value and acquire external knowledge does not guarantee the [successive] exploitation of this knowledge”3
(Zahra and George, 2002): some capabilities comprise the potential absorptive capacity (PACAP), while others refer to the realized absorptive capacity (RACAP).
According to this conceptualization, the PACAP is composed by acquisition and assimilation. Acquisition refers to the capability to identify and acquire externally generated knowledge critical for a firm’s operations (Zahra and George, 2002). Assimilation is the set of a firm’s routine and processes that allow the firm to analyze, process, interpret, and understand the information obtained by external sources (Zahra and George, 2002). On the other hand, the RACAP is comprised by transformation and exploitation. While transformation refers to “a firm’s capability to develop and refine the routines that facilitate combining existing knowledge and the newly acquired and assimilated knowledge”, the exploitation capability is the “capacity of a firm to apply the newly acquired knowledge in product or services that it can get financial benefit from” (Zahra and George, 2002).
9
Because explorative learning comprises the departure from existing knowledge in favor of the search of new knowledge and competencies (Danneels, 2002), it thus corresponds to the PACAP (Berger, 2015); on the other hand, the explorative learning leverages on the external knowledge for the refinement of existing procedures and designs (Jansen et al., 2006, 2009; Danneels, 2002), thus corresponding with RACAP (Berger, 2015). Consequently, a firm should master the four capabilities of ACAP for reaching the desired organizational outcomes (Zahra and George, 2002), since they contribute complementarily to exploration (PACAP) and exploitation (RACAP).
The ACAP is context dependent, since its effectiveness is determined by the structures of communication used and by the actors’ capability to transfer information to other individuals (Cohen and Levinthal, 1990). It means that, to a certain extent, the ACAP depends on the direct and indirect contact among the organizations’ members, that is, the connectedness (Johnson and Sohi, 2001; Jarowski and Kohli, 1993). The connectedness promotes trust that produces, in turn, information and knowledge sharing (e.g., Larson, 1992). Although the connectedness promotes trust and serves as governance mechanism (Berger, 2015), none of the study we analyzed investigates empirically whether the relational governance has indeed an impact on the ability of connectedness of producing innovation.
2.2 GOVERNANCE MECHANISMS
According to the transaction cost economics (TCE), during the relational exchange buyers and suppliers may engage in opportunistic behavior. Accordingly, they try to routinize their transactions by establishing repeated practices that they apply to the relation exchange (Williamson, 1985; Provan, 1993). In order to lower these exchange hazards, two governance mechanisms can be adopted – that is, contractual or relational; Ghosh and John (2005) defined the governance as the extent to which the rules between parties are explicit (i.e., contractual) or implicit (i.e. relational). The more explicit rules, the higher the level of contract formalization. The latter can be described as the extent to which rules, processes, and communication are formally indicated in written contracts (Jansen et al., 2005).
10
governance) to agreements based on complex contracts characterized by high formalization (contractual-based governance). For example, Hill (1990) states that trust acts with its normative behavior, being so more effective than the costly contracts in safeguarding for opportunistic behaviors.
As indicated in the introduction, many scholars call for further investigation on the contingencies that make contracts and relational norms work as substitutes or complements (Berger, 2015; Poppo and Zenger, 2002). Therefore, with the present thesis we want to further contribute in exploring the relationship between relational-based and contractual-based governance in contributing to absorptive capacity in buyer-supplier relationship. We will furnish an interpretation of the results we found in literature on their role of substitutes or complements, and we will relate their interaction to absorptive capacity.
2.3 CONNECTEDNESS
As Dyer and Singh (1998) posit, “a firm’s critical resources may span firm boundaries and may be embedded in interfirm resources and routines”. Therefore, we can assert that when buyers and suppliers integrate and exchange part of their resources and share routines in a relationship, they give birth to a third distinct form of organization (Powell, 1990), which can be considered independent to those of buyer and supplier considered separately. In this organization, people of each firm who work together constitute a community (Adler, 2001). The extent to which the individuals of this community interact one another is thus the outcome of the interaction and the coordinating activities that take place in this new organization. From a network viewpoint, the degree to which two parties communicate and coordinate in their interfirm relationship is signaled by the partners’ or units’ connectedness (Johnson and Sohi, 2001); from an intrafirm perspective, Jaworski and Kohli (1993) define connectedness as the extent to which employees engage in formal or informal direct contact across departments.
11
As Berger (2015) posits, connectedness boosts trust and coordination, serving as a governance mechanism. As indicated in the introduction, Rowley et al. (2000) illustrate that the relational embeddedness serves as a trust-based control agent of alliances’ partners. Nevertheless, although scholars focused on how connectedness promotes trust between partners -serving so as a governance mechanism-, none of the authors investigated the relation between connectedness and trust-based governance on absorptive capacity.
3. HYPOTHESES
In this section, we will further elaborate on the relationships among the variables whose theoretical background was treated in the previous section.
3.1 CONNECTEDNESS
Connectedness between buyer and supplier is proven to promote relational benefits and knowledge sharing (Uzzi, 1997; Day et al., 2013); it also serves as an informal self-enforcing mechanism of the relationship, and improves trust (Uzzi, 1996; Adler and Know, 2002; Dyer and Singh, 1998). According to the transaction cost economics (Williamson, 1985), in interfirm relationship each firm attempts to reduce the relationship hazard – that is, the risk of opportunistic behaviors. In establishing a balance between trust and risk, partners can opt for different organization modes of economic activities, namely market or hierarchy (Williamson, 1979). Adler (2001) proposed that market and hierarchy are “ineffective” in coordinating knowledge-based assets, central in the theme of absorptive capacity; thus, he proposed a third coordinating mechanism (i.e. the community) which is based on trust.
12
high levels of interfirm connectedness will improve ACAP. Larson (1992) and Uzzi (1997) underlined the importance of strong ties (i.e., connectedness) for enhanced trust and interfirm exchange. Furthermore, Hansen (2002) found that the higher the network connections and the relatedness of knowledge between two systems, the higher the team absorption of external knowledge (PACAP) and the faster the performance. Consequently, if a performance is faster, we expect that behind the improvement of efficiency of effectiveness (or both) there is a certain kind of knowledge that has been applied. Therefore, connectedness -coupled with a minimum of shared base of knowledge- should positively be related with the subsequent implementation of previously-stored knowledge (RACAP) (Cohen and Levinthal, 1990).
Conversely, there is the general suggestion that low connectedness is positively related to access to different knowledge source (e.g., Granovetter, 1973). In fact, while weak ties lead to new information, networks that are too connected generate redundant information (Granovetter, 1976), damaging the benefits of diversity of knowledge (Cohen and Levinthal, 1990) and resulting in collective blindness (Uzzi et al., 1996, 1997; Rowley et al., 2000).
Unlike these findings, more recent results show different patterns. Informal coordination mechanisms (i.e., connectedness) are the most important antecedent in predict both explorative (PACAP) and exploitative innovation (RACAP) (Jansen et al., 2006). In a previous study, Jansen et al., (2005) found that, in line with their expectation, the socialization capabilities (i.e. connectedness) affected positively the RACAP. Jansen et al. (2005) found also that connectedness was positively related to assimilation of external knowledge (PACAP), contrasting the assertion that connectedness should be negatively related to PACAP because of the collective blindness; as the authors discussed in the paper, this finding may support the proposed notion that low connectedness impairs a regular and reliable knowledge exchange.
13
an excessive level of trust. Nevertheless, it is not said that the negative effect of excessive connectedness is because of the effect of trust overabundance. In fact, as we will see, although the drawbacks of excessive trust are proved in literature, Jansen et al. (2005, 2006) showed that the detrimental effect of connectedness on PACAP and RACAP was not found. With the present research, we are thus contributing to shed further light on this subtle interrelation between trust-based governance and connectedness.
Taking into account what we exposed so far, we can posit the following hypotheses:
𝐻1𝑎 : Connectedness will be positively related to PACAP (𝐻1𝑎1); this relationship is mitigated for high levels of Connectedness (𝐻1𝑎2).
𝐻1𝑏 : Connectedness will be positively related to RACAP.
3.2 RELATIONAL GOVERNANCE
Relationally governed relationships entail relational and social self-enforcing processes that promote flexibility, solidarity, and information exchange (Poppo and Zenger, 2002). As aforementioned, the role of trust as self-enforcing governance mechanism is prominent in literature (e.g., Day et al., 2013), and the empirical association between trust and relational-based governance is apparent (Zahaeer and Venkatraman, 1995; Saxton, 1997).
In turn, relationship characterized by high trustworthiness show a higher interfirm learning (Fawcett et al., 2012), and more allowance in sharing information among partners (Li et al., 2010; Zahra and George, 2002). Because of the improved information exchange and information sharing, we expect that a relational-based governance fosters the acquisition and the assimilation of external knowledge (PACAP). This is line with the finding of Selnes and Sallis (2003) that a reasonable level of trust improves the information sharing between partners (PACAP), but also a joint sense making, that is the usage of the information shared (RACAP).
14
type of integration as a governance structure that most of times -but not strictly- implies the classical hierarchical relationship proposed by transaction cost economics. Thus, in building their hypothesis (see page 379, description of hypotheses 4a and 4b) they state that “the presence of trust should be associated with a lower level of hierarchical governance since trust serves as a substitute for hierarchical control”. Therefore, unlike a first glace to the results of their analysis may suggest, a conclusively finding that trust does not serves as a substitute of more formal governance mechanism is absent in the paper. Therefore, in accordance to the analysis of their study, it appears that trust may serve as substitute to hierarchical (i.e. contractual) governance, but further evidence is needed. In any case, for benefit of our knowledge, they found a clear connection between trust and quasi-integration (the extent of recurrent exchange between partners, that is, connectedness).
On the other side, incurring into high level of trust-based governance may lead to the so-called hidden costs of trust. We have that Salnes and Sallis (2003) delineate three specific problems that partners may have incurring a learning relationship characterized by excessive trust: the trust may be a signal of friendship among individuals, who perhaps may prefer to avoid new pieces of information that might spoil the personal friendship between them; secondly, trust may lead to an excessive reliance on a partner, promoting opportunistic behaviors from the other; finally, if partners are too congruent one another, they can fall into the same thinking patterns (i.e., falling into a state of groupthink). While Gargiulo and Ertug (2006) individuated the same two first patterns - that they labeled with “relational inertia” and “malfeasance”-, the authors found also that high level of trust may lead the partners to invest into the relationship beyond the necessary (i.e., resource misallocation).
Given the description of the drawbacks of excessive trust, we may posit that some will be detrimental to PACAP while others to RACAP. In fact, if two partners fall into the same groupthink it is less likely that they will tap into different kinds of knowledge, so that this will be detrimental to PACAP. On the other hand, a resource misallocation is clearly detrimental for the refinement and the general improvement of an organization’s activities, since some resource will be wasted; consequently, an excessive trust is expected to be also detrimental to RACAP.
Taking into account the considerations made so far, we can hypothesize as it follows:
15
𝐻2𝑏 : Relational governance will moderate positively the relationship between connectedness and
RACAP.
𝐻3𝑎: Relational Governance has a positive impact on PACAP (𝐻3𝑎1); this positive relationship is
mitigated for high levels of Relational Governance (𝐻3𝑎2);
𝐻3𝑏: Relational Governance has a positive impact on RACAP (𝐻3𝑏1); this positive relationship is
mitigated for high levels of Relational Governance (𝐻3𝑏2).
3.3 CONTRACTUAL GOVERNANCE
In the attempt to balance risk and trust in a relationship, partners may opt for formal contracts (i.e., contractual-based governance). The formalization, and thus the degree this type of governance is adopted, is the extent to which rules, communication and practices are disciplined in formal documents (Jansen et al., 2005). Since exploration capabilities (PACAP) is comprised of searching and experimenting for innovations (Rowley et al., 2000), we expect formalization to be detrimental for PACAP. In fact, from a knowledge-based view formalization stifles the capability to fully explore the external environment, because makes it difficult to act and think “out of the box” (Weick, 1979). However, in the above-mentioned works, Jansen et al. (2005) tested the relationship between formalization and ACAP and reached interesting conclusions; although formalization improved RACAP, it was not detrimental for PACAP (see also Jansen et al., 2006). To a certain degree, thus, we expect detailed contracts to impact positively the PACAP; this is line with the finding that well-formalized procedure not only improved the implementation of new idea, but also their initial formulation and the subsequent diffusion (Zander and Kogut, 1995). Beyond a certain threshold, though, we expect a reverse effect of formalization on PACAP, since exploration is associated with expressions such as flexibility, experimentation, and risk taking (March, 1991), that presumably do not fit the need of conforming to established behaviors contained in contracts (Jansen et al., 2005; Weick, 1979).
Conversely, since exploitation entails improving efficiency and/or the effectiveness by using pre-store knowledge (Rowley et al., 2000), one would expect contractual governance to improve RACAP. In fact, formalization aids the retrieval of past knowledge, facilitating the causal understanding between the activities involved in a firm’s processes (Jansen et al., 2005). The external knowledge is thus more easily integrated in firms’ processes, so that also the actual usage of it is promoted (RACAP).
16
𝐻4𝑎: Contractual governance will have a positive relationship with PACAP (𝐻4𝑎1); this positive
relationship will be weakened for higher levels of contractual-based governance (𝐻4𝑎2);
𝐻4𝑏: Contractual governance will have a positive relationship with RACAP.
3.4 THE RELATIONSHIP BETWEEN THE INTERPLAY OF RELATIONAL AND CONTRACTUAL GOVERNANCE ON ACAP
The interplay between contractual and relational-based has been the object of many studies. In accordance with many authors, relational norms furnish a valid alternative to the more expensive contracts (Uzzi, 1997; Dyer and Singh, 1998; Wang et al., 2011), while for others the two types of governance can coexist (Poppo and Zanger, 2002; Cannon et al., 2000). Poppo and Zenger (2002) found a recursive relationship in which the use relational norms increases as the contractual complexity rises, and vice versa. Since contracts is the domain that limits the likelihood of opportunistic behavior, the authors proposed that, in turn, contracts improve relational norms because the parties can relax the worries about opportunistic behaviors. Cannon et al. (2000) reached the conclusion that high levels of transactional uncertainty make the contract design more detailed. For this reason, relational norms are particularly valued in presence of high uncertainty because the parties can trust one another even in case not all the possible contingencies are formally regulated.
Thus, even if there is evidence that contracts and trust-based governance can coexist and foster each other, but we can draw a novel interpretation of these results. Since contracts determine the domain for opportunistic behaviors, managers employ more relational norms as formalization increases (Poppo and Zanger, 2002) because the risk associated with higher formalization is smaller. Therefore, the adoption of highly formalized contracts arises only when trust between parties is not enough.
17
engage in discussions personally, without formal and rigid canals for cooperating and transmitting the information (Persaud et al., 2002).
Coupling the findings in literature, we therefore suggest that, in general, individuals prefer to perform the activities comprised by a certain relationship with as much flexibility as possible, and that contracts serve only to lower the perceived risk of opportunism and to establish a base for knowledge exchange. This is in line with Wang et al. (2011) conclusion that trust and contracts serve as substitute in promoting innovation.
Gulati and Nickerson (2008) argue that trust can substitute contracting in interfirm exchanges. The authors found that the higher the preexisting interorganizational trust, the higher the probability that trust-based governance overpowers contracting. Furthermore, more importantly, the authors found that trust is beneficial for interfirm exchange regardless the governance mode adopted. Given the centrality of trust in interfirm exchange, we would expect that its effect is preponderant in explaining ACAP.
In this study, hence, we want to analyze the interplay between the governance mechanisms in the context of absorptive capacity. The analysis of the interaction between contractual and relational governance per se would not be meaningful for a deeper understanding of ACAP. Consequently, we test if the governance mechanisms work as substitutes in promoting the ACAP.
Therefore, we can posit the following hypothesis:
18
3.5 CONCEPTUAL MODEL
Here is the representation of the variables and the relationships we consider in this study.
Figure 1: Conceptual model
4. METHODOLOGY
4.1 DATA COLLECTION
The data we used in this paper were collected by Berger (2015); the data are based on the answers given by both buyers and suppliers, and so they represent both of side of the dyadic relationship; this in necessary to embrace an interfirm perspective that gives a complete picture of the relationships at hand (Zahra and George, 2002), even though most of the studies consider methodologically only one side of the relationship (e.g. Claycomb and Frankwick, 2010). The referents who represented the companies were surveyed through a standardized survey. The informants of the customer organizations were contacted and surveyed in first instance. After that, the buyer informants indicated the contacts of the suppliers, who successively received the survey through mail, or email when located abroad. Because the responses are given without interacting with an interviewer and there is the promise of confidentiality, the bias to have answers that comply to certain expectations is reduced (Malhotra, 2009).
19
show peculiar knowledge strategies which are different one another (Linchtenthaler and Ernst, 2007); furthermore, the industries considered (that is, automotive, machinery, chemicals, pharmaceuticals, semiconductors and electronics) are among the most uncertain and volatile (Dyer et al., 2014).
Moreover, for avoiding a selection bias in favor of the most relevant relationships, a 2×2 design is developed; one dimension of the design indicates whether a certain relationship lasted shorter or longer than two year, while the other indicates if the relationship was either about an average or a crucial component or product.
4.2 MEASURES
All the items considered in this research are measured with 7 point likert scales that range from “Strongly disagree” through “Strongly agree”. Particularly, the questionnaire contains first- and second- order constructs. While a first-order construct “is a latent construct that has observed variables as indicators”, the second-order construct have other latent (first-order) constructs as their indicators” (Berger, 2015, p. 91). Furthermore, we distinguish between reflective and formative scales, based on the casualty between indicators and latent constructs. Formative indicators cause the latent construct, so that the latent construct changes if the values of the indicators change or some are left out (Berger, 2015); moreover, if a construct is composed of formative indicators, these measure different aspects and thus are not related (Hair et al., 2011). On the other hand, if a construct is made by reflective indicators, then changes in the values of the reflective indicators are provoked by changes in the value of latent variables. For this reason, high correlations among the reflective indicators are taken for granted (Berger, 2015).
20
routines and processes that allow it to analyze, process, interpret, and understand the information obtained from external sources” (Zahra and George, 2002); its measures are based on Jansen et al. (2005), and Camisòn and Forés (2010) work.
The realized absorptive capacity (RACAP) is comprised of transformation and exploitation, so it is a second-level construct. Transformation is a “firm’s capability to develop and refine the routines that facilitate combining existing knowledge and the newly acquired and assimilated knowledge” (Zahra and George, 2002). Its components are internalization and conversion, and its measures are grounded in Bontis et al. (2002), and Cadiz et al. (2009) work. Conversely, exploitation is “the capacity of a firm to apply the newly acquired knowledge in product or services that it can get financial benefit from” (Zahra and George, 2002). Its relevant components are use and implementation, while its measures are developed around Jansen et al. (2005), Cadiz et al. (2009), and Camisòn and Forés (2010) work. Both PACAP and RACAP are formative scales.
With regards to governance mechanisms, as previously treated, we have the contractual- and the relational-based governance. Contractual governance indicates the extent to which the diverse activities comprised in a certain relationship are written in formal documents (Jansen et al., 2005). It is a reflective scale that was measured following the measures of Deshpande and Zaltman (1982), Cannon and Perreault (1999), Buvik and Reve (2002), and Jansen et al. (2006).
21
4.3 STATISTICAL PROCEDURES
For testing the hypotheses involving moderation, we will follow the definition and the procedure of moderation as they are presented in Baron and Kenny (1986): we have a moderation when an independent variable affects the dependent variable, and that relationship is influenced at the same time by a third variable, called moderator; hence, the moderator has an impact on the strength of the effect of the independent variable on the dependent one. Clearly, this impact can be positive or negative, so that we have that the moderator positively or negatively affects not only the strength but also the direction of the relationship at hand. The total impact on the dependent variable is thus given by the combined effect of the independent variable and the moderator (Baron and Kenny, 1986).
For testing the hypotheses, we used a Partial Least Squares (PLS) analyses, that were performed through SmartPLS3. The PLS is a variance-based Structural Equation Model (SEM) that integrates both the factor analysis and the multiple regression analysis (Berger, 2015). It also permits to analyze complex models that contain higher-order constructs, multiple parallel mediation, and formative constructs, so the PLS fits the characteristics needed for our analysis. Given the sample size of 166, bootstrapping will be used as resampling methods for the calculation of the p-values relative to the betas in the structural model.
22
5. RESULTS
5.1 DATA VALIDITY AND RELIABILITY
In this section, we will carry a canonic PLS-SEM analysis for the estimation of the outer and the inner model; for the outer model, we evaluate the relationships between our latent constructs and their respective indicators. After deleting the indicators that do not satisfy our requirements, we will estimate the inner model that gives the estimates regarding the relationships between the latent constructs. This section will be closed with an overview on the results of the hypotheses tested to grasp the differences between the results referring to buyers- and suppliers-only databases.
As mentioned, our dataset is comprised of the dyadic relationships between buyers and suppliers. Thus, the hypotheses could be tested considering the whole dataset or considering the buyer- and supplier- only databases separately. Therefore, we conducted a paired sample t-test to establish if the buyer- and supplier-only datasets differ significantly. The t-values of the mean score differences are shown in Table 1. The differences between the means differ significantly for all the constructs (p<0.05), apart from contracting. We can conclude that we will analyze and test the hypotheses considering each party of the dyad singularly. With regards to the validity, we did not detect significant differences between buyers-only and suppliers-only databases. Consequently, the validity was assessed considering the aggregate dataset.
Table 1: Sample T-test between buyer and supplier scores.
5.2 REFLECTIVE SCALES
With regards to the reflective scales, the reliability of their indicators is based on the analysis of indicators’ loadings. The loadings between the indicators and the respective constructs should be higher than 0.5 each. Moreover, for assessing the internal consistency of the construct individuated –that is, the constructs’ reliability-, we will consider the Cronbach’s alpha, whose values should be higher of 0.7 for each of the construct considered. In case a value is lower than this threshold, it
23
means that that particular construct is not strong enough to be considered for further steps in our research.
Because the Cronbach’s alpha uses equal weights, all the indicators are assumed to contribute alike to the overall reliability. More precisely, we take into consideration the different loadings corresponding to each indicator by evaluating also the composite reliability (Gotz et al., 2010). Also for the case of composite reliability, the threshold is established on 0.7.
Generally speaking, as shown in table 2, all the conditions indicated so far are satisfied: the standard loadings are higher than 0.5 (p-values < 0.05), and the threshold of 0.7 is reached for all the values referring to the Cronbach’s alpha and to the composite reliability. The only standard loadings that did not comply to our requirements is Connectedness4 that has a standard loading of 0.229 (p=0.011), and Contracting5 (0.426, p=0.167). The latter two indicators were thus removed from the coming analysis.
Table 2: Construct reliability and convergent validity of reflective scales
Construct/Indicator
Standard Loadings Cronbach’s alpha Composite Reliability
24 RelGov8 .878 RelGov9 .897 RelGov10 .821 Contracting .827 .843 Contracting1 (Recoded) .629 Contracting2 .845 Contracting3 .880 Contracting4 .681 Contracting5 .463 Contracting6 .581 Relational Governance .876 .900 Flexibility .764 Information Exchange .881 Solidarity .881 Connectedness 0.755 .837 Connectedness1 .746 Connectedness2 .881 Connectedness3 .816 Connectedness4 .229 Connectedness5 .787 5.3 FORMATIVE SCALES
25
Table 3: Convergent validity of the formative indicators
Construct/Indicators
Outer Weights: p-value: VIF:
Acquisition Acquisition1 .296 <.001 1.321 Acquisition2 .134 .007 1.122 Acquisition3 .310 <.001 1.180 Acquisition4 .363 <.001 1.420 Acquisition5 .386 <.001 1.483 Transformation Tansformation1 .279 <.001 1.649 Transformation2 .122 <.001 1.274 Transformation4 .216 <.001 1.381 Transformation5 .099 .014 2.879 Transformation6 .236 <.001 2.414 Transformation7 .244 <.001 1.863 Transformation8 .212 <.001 2.216 PACAP Acquisition .458 <.001 1.000 Assimilation .639 <.001 1.000 RACAP Exploitation .431 <.001 1.000 Transformation .642 <.001 1.000
26 1. 2. 3. 4. 5. 6. 7. 8. 9. 1. Acquisition .649 2. Assimilation .652 .739 3. Connectedness .342 .457 .810 4. Contracting .109 .081 .063 .724 5. Exploitation .539 .650 .402 .100 .724 6. Flexibility .194 .347 .315 .023 .352 .811 7. Information Exchange .305 .411 .518 .050 .372 .502 .777 8. Solidarity .254 .380 .388 .101 .363 .534 .669 .866 9. Transformation .563 .649 .645 .094 .685 .305 .482 .394 .704 Table 4.a : Square roots of the average variance extracted (AVE) and correlation matrix (first-order constructs)
1. 2. 3.
1. PACAP .633
2. RACAP .535 .632
3. Relational Governance .427 .503 .690
Table 4.b: Square roots of the average variance extracted (AVE) and correlation matrix (second-order constructs)
5.4 HYPOTHESES TESTING
This paragraph contains the test of our hypotheses. The software used to carry the analysis is SmartPLS3. For increasing the precision of the estimates, the PLS was carried with a higher number of interactions than those proposed as the default option by the software (1000 vs. 300 respectively). With the same rationale, for increasing the reliability of the results the bootstrapping was carried selecting 2000 samples extracted rather than the default option of 500.
First, the hypotheses are tested first for the “Buyer-only” dataset, then they are tested for the “Supplier-only” dataset. In last instance, this section is concluded with an overview of the results, that will be broaden in the Discussion section.
5.4.1 “Buyer-only” Database
In line with the expectation, connectedness has a positive impact on PACAP (β=0.236, p=0.001) and on RACAP (β=0.402, p=0.000). Hence, 𝐻1𝑎 and 𝐻1𝑏 are supported. Also, the U-shaped
27
significant (β=-0.124, p=0.090)4, permitting to support 𝐻
1𝑎2 . With regards to the moderation
effects, the positive effects of relational governance on the relationship between connectedness and PACAP and RACAP are unexpectedly negative for both PACAP (β=-0.145, p=0.003) and RACAP (β=-0.137, p=0.016). Therefore, hypotheses 𝐻2𝑎 and 𝐻2𝑏 are not supported. The relationship
between Relational Governance and PACAP is significant and in the same direction as hypothesized (β=0.454, p=0.000), and the same can be stated for the relationship between relational governance and RACAP (β=0.433, p=0.000). Hence, hypotheses 𝐻3𝑎1 and 𝐻3𝑏1 were supported. With regards to the same relationships, the hypothesized U-shaped effects for high levels of Relational Governance were found for PACAP (β=-0.076, p=0.050) and for RACAP (β=-0.106, p=0.016), giving support to 𝐻3𝑎2 and 𝐻3𝑏2. Moreover, contracting was not found to have a positive effect on PACAP (β=0.074, p=0.546), nor on RACAP (β=0.018, p=0.613), missing to give support for 𝐻4𝑎1 and 𝐻4𝑏. The hypothesized U-shaped effect of Contracting on PACAP is not significant
(β=0.033, p=0.672). Hence, 𝐻4𝑎2 is not supported.
In last instance, we establish whether relational- and contractual-based governance serve as substitutes in promoting ACAP. The relationship between the interactive effect of the two types governance is insignificant for both PACAP (β=0.261, p=0.330) and RACAP (β=-0.109, p=0.521). Thus, 𝐻5𝑎 and 𝐻5𝑏are not supported by our data.
Table 5 gives an overview on the results of the hypothesized relationships for buyer-only database.
Table 5: Overview of the buyers-only hypotheses results
Hypotheses Supp./Rej. Dependent Independent β P-values
𝑯𝟏𝒂𝟏 Supported PACAP Connectedness 0.236 0.000
𝑯𝟏𝒂𝟐 Supported Quadratic -0.124 0.090
𝑯𝟏𝒃 Supported RACAP Connectedness 0.402 0.000
𝑯𝟐𝒂 Rejected Moderation1 -0.145 0.003
𝑯𝟐𝒃 Rejected Moderation2 -0.137 0.016
𝑯𝟑𝒂𝟏 Supported PACAP Relational Gov. 0.454 0.000
𝑯𝟑𝒂𝟐 Supported Quadratic -0.076 0.050
𝑯𝟑𝒃𝟏 Supported RACAP Relational Gov. 0.433 0.000
𝑯𝟑𝒃𝟐 Supported Quadratic -0.106 0.016
𝑯𝟒𝒂𝟏 Rejected PACAP Contracting 0.074 0.546
𝑯𝟒𝒂𝟐 Rejected Quadratic 0.033 0.672
𝑯𝟒𝒃 Rejected RACAP Contracting 0.018 0.613
𝑯𝟓𝒂 Rejected PACAP Interaction 0.261 0.330
𝑯𝟓𝒃 Rejected RACAP Interaction -0.109 0.521
28
5.4.2 “Supplier-only” Database
With regards to suppliers, the relationship between connectedness and PACAP is positive and significant (β=0.408, p=0.000), and the same can be said for the relationship between connectedness and RACAP (β=0.500, p=0.000), giving support for 𝐻1𝑎 and 𝐻2𝑏. Conversely, the U-shaped
relationship between Connectedness and PACAP is not significant (β=-0.034, p=0.642), so we reject 𝐻1𝑎2. The moderation effects of relational governance on the relationships between connectedness and PACAP and between Connectedness and RACAP are not significant (β=-0.037, p=0.599 and β=-0.017, p=0.819, respectively). Accordingly, we reject 𝐻2𝑎 and 𝐻2𝑏. Furthermore, the relationship between relational governance and PACAP is significant (β=0.210, p=0.032), and the relationship between the same independent variable and RACAP is significant at a 90% level (β=0.160, p=0.067). Hence, we have support for 𝐻3𝑎1 and for 𝐻3𝑏1. The hypothesized U-shaped
relationships between relational governance and PACAP and RACAP are not significant (β=-0.064, p=0.330 and β=-0.045, p=0.568 respectively). Hence, 𝐻3𝑎2 and 𝐻3𝑏2 are not supported. Moreover, the positive relationship between contracting and PACAP is not significant (β=-0.003, p=0.974), and the same goes for the relationship between Contracting and RACAP (β=-0.001, 0.993); hence, 𝐻4𝑎1 and 𝐻4𝑏 are rejected. The U-shaped relationship between contracting and PACAP is not significant (β=0.042, p=0.627), not supporting 𝐻4𝑎2. Lastly, the cross effect of relational and contractual governance is not significant on PACAP (β=-0.109, p=0.521), nor on RACAP (β=0.126, p=0.638). Accordingly, we can reject 𝐻5𝑎 and 𝐻5𝑏.Table 6 furnishes an overview of the results for
the buyer-only database.
Table 6: Overview of the suppliers-only hypotheses results
Hypotheses Supp./Rej. Independent Independent β P-values
𝑯𝟏𝒂𝟏 Supported PACAP Connectedness 0.408 0.000
𝑯𝟏𝒂𝟐 Rejected Quadratic -0.034 0.642
𝑯𝟏𝒃 Supported RACAP Connectedness 0.500 0.000
𝑯𝟐𝒂 Rejected Moderation1 -0.037 0.599
𝑯𝟐𝒃 Rejected Moderation2 -0.017 0.819
𝑯𝟑𝒂𝟏 Supported PACAP Relational Gov. 0.210 0.032
𝑯𝟑𝒂𝟐 Rejected Quadratic -0.064 0.330
𝑯𝟑𝒃𝟏 Supported RACAP Relational Gov. 0.160 0.067
𝑯𝟑𝒃𝟐 Rejected Quadratic -0.045 0.568
𝑯𝟒𝒂𝟏 Rejected PACAP Contracting -0.003 0.974
𝑯𝟒𝒂𝟐 Rejected Quadratic 0.042 0.627
𝑯𝟒𝒃 Rejected RACAP Contracting -0.001 0.993
𝑯𝟓𝒂 Rejected PACAP Interaction -0.109 0.521
29
5.4.3 SUMMARY OF DATA ANALYSIS
Based on our results we can conclude that, for buyers and suppliers, connectedness has a positive effect on PACAP and on RACAP; the diminishing positive effect of Connectedness on PACAP is significant for buyers only. Secondly, the positive moderation effect of relational norms on the relationships between connectedness PACAP and between connectedness and RACAP were found interestingly and unexpectedly negative for buyer and insignificant for suppliers. Third, the positive relationships between relational governance and ACAP are significant for buyers and suppliers.
With regards to the U-shaped effects of relational governance on ACAP, they were detected as significant only for buyers. Referring to the relationships between contracting and both PACAP and RACAP, the results are insignificant for buyers and for suppliers. Moreover, in both of cases the U-shaped effects of Contracting on ACAP are not significant. Lastly, contracting and relational governance seem to not work as substitutes for promoting ACAP.
To sum up, the hypotheses 𝐻1𝑎1, 𝐻1𝑏 , 𝐻3𝑎1 , 𝐻3𝑏1 are supported for buyers and suppliers. Conversely, hypotheses 𝐻1𝑎2, 𝐻2𝑎, 𝐻2𝑏, 𝐻3𝑎2, 𝐻3𝑏2 are supported only for buyers. Those relationships that are supported only for buyers are instead insignificant for suppliers.
Based on these results, we will account for these differences in the next section of this thesis, the discussion. Accordingly, the limitations, the suggestions for further research and the conclusions will follow.
6. DISCUSSION, LIMITATIONS, AND FURTHER RESEARCH
6.1 DISCUSSION30
levels of connectedness do not seem to be detrimental for PACAP. Our results are coherent with the findings of Jansen et al. (2006) for which connectedness is one of the most important antecedents of explorative innovation (PACAP); like in their study, we did not detect the undermining effect of collective blindness found by other authors (e.g. Uzzi, 1997), for suppliers. The reason behind the difference between buyers and suppliers on this nonlinear effect can be explained by the fact that suppliers engage in a constant effort to accommodate buyers’ specification of the delivered products (Johnston et al., 2004, Jap, 1999); for this reason, suppliers are perhaps more receptive of the flow of information that comes from the buyers, so that even for high levels of interaction they do not have “diminishing returns” from the ongoing interaction.
With regards to the positive moderation effect of relational norms on the relationship between Connectedness and ACAP, our expectations were not satisfied: we found a negative moderating effect for buyers, while for suppliers we found an insignificant effect. This goes against the findings of the only study we found on the interrelationships among trust (trust-based governance), ties strength and knowledge flow of Levin and Cross (2004); they found that the relationship between strong ties and the receipt of useful knowledge was mediated by competence- and benevolence- trust, thus reaching an opposite conclusion. As they underline in their work, although the concepts of tie strength and trust are used interchangeably by many authors -in the paper, they mention a lot of examples- the two concept are distinct.
Accordingly, our results permit to show that indeed the two concepts are empirically distinct. In case they were substantially synonyms, we would expect at least no interaction between them. Conversely, the moderation is negative for buyers and insignificant for suppliers.
31
Furthermore, Corsten and Felde (2004) found that the collaboration between buyer and supplier improves innovation (ACAP) and it is positively influenced by trust, and thus presumably by trust-based governance. Again, Zaheer and Venkatraman (1995) reached the conclusion that there is a positive relationship between trust and the “extent of recurrent exchange” between the parties, as we would expect considering the nature of the two construct. In the limitations of this study, we will indicate different possible explanations that can be investigated in future works.
As expected, we found a positive relationship between Relational Governance and ACAP. As indicated in the hypotheses section, this is in line with the finding that a reasonable level of trust improves the information sharing among partners (PACAP), and the joint effort to make sense of the generated information flow (RACAP) (Fawcett et al., 2012; Zahra and George, 2002; Selnes and Sallis, 2003). Also, the quadratic effects of relational governance on ACAP were found for buyers, but not for suppliers. The quadratic effect we found can refer, as hypothesized, to the hidden costs of trust: the avoidance of information that potentially could spoil the personal relationship between the two parties (Selnes and Sallis, 2003), or lead to overinvest in the relationship in virtue of personal relational bonds rather than the interests of the firm (i.e. resource misallocation) (Gargiulo and Ertug, 2006). We will attempt to furnish a reason for the unexpected results that we have for suppliers.
32
the quality of the relationship is also signaled by the fact that indeed suppliers tend to reduce opportunism with a strong communication with buyers (Ambrose et al., 2010).
Therefore, even for high levels of relational governance the suppliers may not perceive a diminishing impact of it on PACAP and RACAP; this is reflected also by the fact that suppliers were found to perceive fewer task and relational conflicts than buyers do (Oosterhuis et al., 2013), so they might be less sensitive also to negative relational outcomes of the relationship.
Moreover, the hypothesized relationships between contracting and ACAP are generally not supported. For buyers and suppliers, the relationship between contracting and PACAP is not significant. This can be explained by the fact that PACAP is associated with concepts such as thinking out of the box and experimentation, so that rigid contracts may impair the ability to derive from established behavior (March, 1991; Jansen et al., 2005). This is in line with Cannon et al. (2000) conclusion that contracts are ineffective in contexts where the information to spread is implicit and sticky; for such reason, the exchange of knowledge and skills is better promoted when the individuals of the organizations share their information in an informal way, without formal canals (Persaud et al., 2002).
With regards to the effect of contracting on RACAP, it was found insignificant for buyers and for suppliers. These unexpected results give support to the assertion that contracts can be a constraint in environment characterized by high uncertainty, where the contracts are redundant and not feasible to the contingencies at hand, particularly when implicit information is involved (Williamson, 1985; Cannon, 2000); after all, this is in line with the reasoning that many times contracts become less and less important as the relationship builds up and the parties thus prefer to operate with more flexibility (Wang et al., 2011), making contract not so important in the management of the relationship (Berger, 2015).
In last instance, we did not find that the two types of governance serve as substitutes in promoting ACAP. In line with Wing et al. (2011), we showed that relational governance is more important than contracting in explaining ACAP. Unlike them, though, we did not detect that the two types of governance work as substitutes in promoting firm’s innovation.
33
These results indicate that the two types of governance have a geometry-variable relationship. It means that in some relationships the relational norms prevail on contracting, while in others the opposite hold. We found a great deal of studies to support this view.
In a recent important article, Cao and Lumineau (2015) carried a meta-analysis to reach a conclusion on the interplay of the two governances in interorganizational relationships. They found that the relationship length positively -and not negatively- moderates trust-based and contractual governance. The reason is that as the relationship consolidates, contracting becomes more efficient and less costly because the parties take advantage of the trust built in earlier stage for contracting more openly with the partners. Moreover, revising a contract may signal commitment to continue the relationship. In this way, contracts are not only seen as controlling and coordinating tools, but also a tangible evidence of the bonds between the parties (Talay and Akdeniz, 2014; Zheng et al., 2008). This is in line with Poppo and Zenger (2002) assertion that there is a recursive interaction between formal and informal governance. In fact, contracts restrict the domain of opportunistic behaviors. In turn, the reduction of the domain of undesired behaviors improves trust.
Also, Gulati and Nickerson (2008) underline that preexisting trust may offset contracting in the governance of the relationship, but also that trust-based governance is beneficial and can be high even for high levels of contracting. In other words, as they argue in their paper, the central question to address is not on whether they work as substitutes or complements. Rather, the question is on “when” and “how” the substitution and the complementary views coexist.
Taken together, these recent articles contribute to a geometry-variable view of the two types of governance. In this view, the relative importance of each governance mode is determined by organizational contingencies, so it varies along different dimensions. One of the dimension, for example, can be the type of interorganizational relationships. Cao and Lumineau (2015) argue that in cross-border relationships the governance modes might be independent, while they might be complementary in vertical relationships. One of the reason they furnished for these differences is the degree of interdependence and the geographical distance.
34
In any case, the analysis we made of these results on the interplay between contractual and relational governance allow us to refuse the extreme views of “substitution” and “complementary” that we found in literature.
6.2 LIMITATIONS AND SUGGESTIONS FOR FURTHER RESEARCH
Herewith we will propose some limitations of the present study that might be taken into account for the development of future paper research.
First and foremost, a clear definition and operationalization of trust is needed. As many authors underline (e.g. Adler, 2001), trust has been used in different contexts in social sciences and thus can refer to different concepts accordingly the phenomena at hand. For example, we embrace the view of Poppo and Zenger (2002) that trust is the cornerstone of relational-based governance; it seems, tough, that the variety of the contexts where trust is called into play makes impossible to adopt the same scale validity for all the possible utilizations of trust. A further clarification at least of the general and common characteristics of trust is needed.
Secondly, in investigating on whether the relational- and contractual-based governance work as substitutes or as complements, it would be interesting to further consider what peculiar characteristics of the buyer-supplier relationship impact the adoption of the two governance. For example, since buyers and suppliers have different perceptions of the ambiguous aspects of the relationship, it would be interesting to evaluate if these differences reflect themselves in some patterns relative to the buyer-supplier relationship governance. As we mentioned in the discussion, further research should focus on the different conditions that impact the relative importance of one governance with respect to the other.
35
evaluate the relative impact of inherent characteristics of the parties – for instance, as said, the role they have into the relationship and the cultural background.
Fourth, future research should further study the empirical and the conceptual relation between trust, relational-based governance, and trust. The surprising negative interaction of relational-based governance and connectedness on ACAP opens some possibilities for further research. Since there is no evidence in literature that can justify a negative effect of trust on cooperation, this raises the possibility that trust itself is not the only antecedent and cornerstone of relational-based governance. Relational governance might be a mere organizational outcome of trust. In other words, it is not said that relational norms can serve as a proxy of trust. It would explain the unexpected interaction we found between relational norms and connectedness.
Lastly, longitudinal data would shed light on the casual relationships between our variables. For example, as Poppo and Zanger (2002) and Cao and Lumineau (2015) mention in their paper, this would be beneficial for studying the recursive relationship between contracting and relational norms.
6.3 CONCLUSION
This study shows that the extent of formal and informal personal contacts between personnel of a buyer-supplier relationship -that is, connectedness- impacts positively the capacity to explore (PACAP) and to exploit (RACAP) newly acquire external knowledge. For buyers, an excessive level of connectedness leads to a collective blindness that impairs the ability to explore new possibilities, having so a negative effect on PACAP. Conversely, suppliers are more concerned to establish an ongoing communication with buyers in order to adjust their offering. Thus, for suppliers, even for high levels of connectedness there is no detrimental effect on PACAP.
36
Moreover, the negative moderation opens up for the possibility that trust cannot be considered as the only cornerstone of relational-based governance.
Conversely, contracting was shown to be far less effective than expected, giving way for the assumption that the parties prefer to govern their relationship with a much as freedom as possible, so that the contingencies at hand are properly addressed. This view is strengthened by the fact that relational governance is instead positively related to ACAP, for buyers and suppliers.
37
BIBLIOGRAPHY
Adler P.S. (2001). Market, hierarchy, and trust: The knowledge economy and the future of capitalism. Organization Science. March/April, Vol. 12 - Issue 2 (215-234)
Adler P.S., SW Kwon (2002). Social capital: Prospects for a new concept. Academy of management review. Academy of Management Review. Vol. 27 - Issue 1, (17-40).
Arbussà, A. & Coenders, G. 2007. Innovation activities, use of appropriation instruments and absorptive capacity: Evidence from Spanish firms. Research Policy, 36, p. 1545-1558.
Baron R.M., Kenny D.a. (1986). The moderator–mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, Vol 51(6), Dec 1986, 1173-1182
Berger, J. (2015). Essays on the Governance of Buyer-Supplier Relationships. Groningen: University of Groningen.
Bontis, N., Crossan, M.M. & Hulland, J. 2002. Managing an organizational learning system by aligning stocks and flows. Journal of Management Studies, 39 (4), p. 437-469
Buvik, A., & Reve, T. (2002). Inter-firm governance and structural power in industrial relationships: The moderating effect of bargaining power on the contractual safeguarding of specific assets. Scandinavian Journal of Management, 18(3), 261.
Cadiz, D., Sawyer, J.E. & Griffith, T.L. 2009. Developing and validating field measurement scales for absorptive capacity and experienced community of practice. Education and Psychological Measurements, 69, p. 1035-1058.
Camisón, C. & Forés, B. 2010. Knowledge absorptive capacity: New insights for its conceptualization and measurement. Journal of Business Research, 63 (7), p. 707-715.
