1
‘Governance Mechanisms in Inter-firm Absorptive Capacity
the effect of the distribution of benefits’
Master thesis, Marketing Management
University of Groningen, Faculty of Economics and Business
22nd of June 2015
MARISKA KAPELLE
Contact details: Mauritsstraat 45 9724 BJ Groningen Student number: 1686690 Email: mariska.kapelle@gmail.comSupervisor
Dr. J. Berger
Second assessor
Dr. J. van Doorn
2
‘Governance Mechanisms in Inter-firm Absorptive Capacity
the effect of the distribution of benefits’
-MARISKA KAPELLE
Abstract.
Inter-firm absorptive capacity (ACAP) is the capacity to gain and sustain a competitive advantage by knowledge creation and utilization between firms. This paper distinguishes between potential absorptive capacity (PACAP) and realized absorptive capacity (RACAP). As knowledge is sticky, tacit and not easily transferable between firms, governance mechanisms are established to control for opportunistic behavior of the partnering firm and to ensure that value is, and remains to be added in inter-firm learning. This paper researches the effect of contractual-based and relational-based governance mechanisms on PACAP and RACAP and investigates to what extent these effects are influenced by the distribution of relational benefits between firms. Data is collected from buyers and suppliers (i.e. dyadic data) in vertical inter-firm relationships in the manufacturing and high technology industry.Results have shown both contractual-based and relational-based governance to positively affect PACAP. RACAP however, is only significantly positively influenced by relational-based governance. The results found no support for the impact of the distribution of benefits on the effects of contractual-based governance mechanisms on PACAP and RACAP. The influence of the distribution of benefits on the effects of relational-based governance on PACAP and RACAP is supported by this paper’s results. The effect of relational-based governance on PACAP and RACAP in both cases diminishes in case benefits are unequally distributed amongst buyer and supplier. Which means more relational-based governance is needed to equally stimulate PACAP and RACAP in case benefits are unequally distributed, compared to an equal distribution of benefits. The last chapters of this paper discusses and concludes based on these outcomes. Finally, recommendations for future research on the impact of governance mechanisms on ACAP are presented.
TABLE OF CONTENTS 3
TABLE OF CONTENTS
1. INTRODUCTION ... 5
2. THEORETICAL BACKGROUND... 7
2.1 Introduction in inter-firm absorptive capacity ... 7
2.1.1 Theoretical overview of research on knowledge-based inter-firm learning ... 7
2.1.2 Absorptive capacity and its dimensions ... 7
2.2 Governance mechanisms in inter-firm relationships ... 8
2.2.1 Types of governance mechanisms ... 9
2.2.2 Governance mechanisms in an inter-firm knowledge-based context ... 10
2.3 Distribution of benefits ... 10
2.3.1 Potential benefits in inter-firm absorptive capacity ... 10
2.3.2 Distribution of potential benefits in inter-firm absorptive capacity ... 11
3. HYPOTHESES ... 12
3.1 Contractual-based governance mechanisms ... 12
3.2 Relational-based governance mechanisms ... 13
3.3 Distribution of benefits ... 14
3.3.1 Impact of distribution of benefits on contractual-based governance ... 15
3.3.2 Impact of distribution of benefits on relational-based governance ... 16
4. METHODOLOGY ... 18
4.1 Research design ... 18
4.2 Sample and data collection ... 18
4.3 Measurement development ... 19
4.4 Data analyses ... 20
4.4.1 Step 1: Preparing the data for analysis ... 21
4.4.2 Step 2: Analysing the structural model ... 21
4.4.3 Step 3: Analysing the multi-group model ... 21
5. RESULTS ... 22
5.1 Data validity ... 22
5.1.1 Reflective scales ... 22
5.1.2 Formative scales ... 24
5.2 Results of the structural model ... 24
5.3 Results of the multi-group model ... 26
6. DISCUSSION AND CONCLUSION... 29
6.1 Impact of governance mechanisms in inter-firm learning ... 29
6.2 Impact of distribution of benefits on governance mechanisms in inter-firm learning ... 30
6.3 Conclusion ... 33
4
8. REFERENCES ... 37
APPENDIX A ... 43
APPENDIX B ... 45
INTRODUCTION 5
1. INTRODUCTION
Inter-firm relationships are increasingly viewed as valuable learning opportunities (Dyer & Singh, 1998; Inkpen, 1998; Muthusamy & White, 2005) and firm’s need to expend their knowledge scope, as ties with the outside world foster learning and knowledge as a competitive advantage (Berger, 2015; Dyer & Singh, 1998; Jap, 1999; Uzzi & Lancaster, 2003; Volberda, Foss, & Lyles, 2010). Inter-firm learning requires idiosyncratic resources, which are resources that are unique to the relationship and which may have little, or no value if they were redeployed outside the relationship or for another purpose (Anderson, Chu & Weitz, 1987; Berger, 2015; Jap, 1999; Wittmann, Hunt & Arnett, 2009). Moreover, knowledge is a complex competitive asset as it is difficult to imitate. Which is primarily caused by the characteristics of knowledge-based resources being tacit, sticky, difficult to codify and not easily transferable between partners (Cohen & Levinthal, 1990; Dyer & Singh, 1998; Jap, 1999; Szulanski, Cappetta & Jensen, 2004; Szulanski, 1996). The combination of the investment in idiosyncratic resources and the stickiness of knowledge in inter-firm relationships makes that both partners in this relationship have a natural need for incorporating safeguards against opportunistic behavior and ensuring initiatives of the partnering firm (Heide, Wathne & Rokkan, 2007; Mohr & Sengupta, 2002). This paper therefore focuses on the relationship between these safeguards (i.e. governance mechanisms) and knowledge absorptive capacity in inter-firm (i.e. buyer-supplier) relationships.
A firm’s absorptive capacity (ACAP) is the ‘dynamic capacity pertaining to knowledge creation and utilization that enhances a firm’s ability to gain and sustain a competitive advantage’ (Zahra & George, 2002: 185). Although this definition does not necessarily include the need for an inter-firm relationship, research has shown inter-organizational networks and inter-firm relationships to play an important role in innovation activities (Lane & Lubatkin, 1998; Rindfleisch & Moorman, 2001) and thus could stimulate ACAP (Lee & Cavusgil, 2006). As the aim of an inter-firm relationship in knowledge management is to minimize risk and maximize learning (Slater & Narver, 1995), firms establish governance mechanisms to control for opportunistic behavior and to create incentives for value-creation (Mellewigt, Madhok & Weibel, 2007; Roath, Miller & Cavusgil, 2002). These mechanisms could be either contractual-based or relational-based (Lee & Cavusgil, 2006). The first is characterized by formal, legally binding agreements, were the latter involves the role of mutual trust and commitment in the governance process.
6 benefits. Based on agency theory in inter-firm relationships (Misangyi & Acharya, 2014; Roath et al., 2002), one would expect the distribution of benefits from inter-firm learning to influence the need for, and type of, governance mechanism(s). The unique contribution of this paper is that it provides an insight into whether the effective governance mechanisms to stimulate inter-firm ACAP are affected by the distribution of benefits.
The paper therefore is a combination of knowledge-based competitive theory and agency theory. It aims to answer the research question: “To what extent does the distribution of benefits in a buyer-supplier relationship influence the effects of governance mechanisms on absorptive capacity?” Data is collected via a questionnaire distributed in dyadic buyer-supplier relationships in non-service companies.
Theoretically, the research contributes to the maturity and integrity (Carter, 2004; Kaynak & Hartley, 2008) of research on the effect of governance mechanisms on ACAP. Moreover, it contributes to theory by combining knowledge-based research and agency theory in researching whether the distribution of benefits between buyer and supplier is of influence on the use of governance mechanisms. From a managerial point of view, the paper provides an insight in the effect of different types of governance mechanism on ACAP, which enables managers to recognize new external knowledge, assimilate it, and apply it to commercial ends (Cohen & Levinthal, 1990). Secondly, it provides a first exploratory explanation of whether the effectiveness of governance mechanisms is influenced by whether benefits are equally or unequally distributed in inter-firm relationships.
THEORETICAL BACKGROUND 7
2. THEORETICAL BACKGROUND
This theoretical background describes and provides the research context of this paper. It starts with an introduction on inter-firm absorptive capacity (ACAP) and its current state of research, linked to knowledge-based theory. Hereafter, governance mechanisms will be explained. Lastly, the distribution of benefits between partners will be introduced. The following chapter then combines these theoretical elements and includes the hypotheses of this paper.
2.1 Introduction in inter-firm absorptive capacity
2.1.1 Theoretical overview of research on knowledge-based inter-firm learning
The resource-based view (RBV) was a rather common literature stream in especially the nineties. Which, in the beginning, focused primarily on physical resources (i.e. static resources) as the type of resource that could increase a firm’s position in the market and its competitive advantage (Lockett, Thompson & Morgenstern, 2009). A later assumption of the RBV is that not only physical resources, but the assets and competencies that are deeply embedded in the organization and therefore difficult to transfer across organizational boundaries make up for a competitive advantage (Mohr & Sengupta, 2002). This assumption in the RBV created a focus on knowledge-based resources. Because knowledge is characterized as rare, valuable, non-substitutable, and difficult to imitate (Barney, 1991; Dyer & Singh, 1998), recently, knowledge-based research increasingly addressed the importance of knowledge as an essential asset to achieve a competitive advantage (Eisenhardt & Santos, 2002; Grant, 1996; Jansen, van den Bosch & Volberda, 2005). Firms that are effective in developing and transferring knowledge-based resources are likely to achieve a competitive advantage over competitors who are not (Berger, 2015).
In order to enable a competitive advantage based on knowledge, it appears inter-firm relationships generate information and knowledge sharing that benefits both partners (Mohr & Sengupta, 2002) and that could not have been established by either one of the parties separately. More concrete, learning from others is one of the key requirements to generate new knowledge, and is therefore often one of the purposes of collaborative relationships (Mohr & Sengupta, 2002: 283). Inter-firm relationships increase the potential of a knowledge-based competitive advantage as these partnerships provide access to a greater extent of knowledge sources, increase the development of knowledge-sharing mechanisms, and offer alternative perceptions of the meaning of critical information that could lead to generative learning (Slater & Narver, 1995). Additionally, literature argues a firm’s competitive advantage to not only depend on the internal capabilities and resources of the firm, but also on the type and scope of its relationship with other companies (Lee & Cavusgil, 2006). Therefore, this paper focuses on knowledge-sharing and learning in an inter-firm relationship context.
2.1.2 Absorptive capacity and its dimensions
8 2002). More specifically, it is defined as absorptive capacity (ACAP), which is ‘a set of organizational routines and processes by which firms acquire, assimilate, transform, and exploit knowledge to produce a dynamic organizational capability’ (Zahra & George, 2002: 186). Literature on ACAP describes the importance of four different, but complementary dimensions of ACAP, which could be categorized as the potential absorptive capacity (PACAP, composed of acquisition and assimilation) and the realized absorptive capacity (RACAP, composed of transformation and exploitation) (Zahra & George, 2002). Separately, PACAP and RACAP do not necessarily improve performance or improve a firm’s competitive advantage. It is rather the effective combination of both dimensions that incorporates the potential (Cohen & Levinthal, 1990; Zahra & George, 2002). That is because firms focusing only on PACAP continuously renew their knowledge stock, however, may suffer from not gaining benefits from the exploitation of it. Secondly, firms focusing primarily on RACAP may well achieve short-term profits through exploitation, but may not be able to respond to environmental changes (Jansen et al., 2005). Therefore it is necessary to fully understand the dimensionality of ACAP, as all of its dimensions contribute to inter-firm learning performance.
The first dimension is acquisition, which is referred to as ‘a firm’s capability to identify and acquire externally generated knowledge that is critical to its operations’ (Zahra & George, 2002: 189). Assimilation is the second dimension and is referred to as ‘the firms routines and processes that allow it to analyze, process, interpret, and understand the information obtained from external sources’ (Zahra & George, 2002: 189). The combination of the first two dimensions define PACAP.
The third and fourth dimensions of ACAP jointly explain RACAP, which reflects the firm’s capacity to leverage the knowledge that has been absorbed (Cohen & Levinthal, 1990). Transformation denotes ‘a firm’s capability to develop and refine the routines that facilitate combining existing knowledge and newly acquired and assimilated knowledge’ (Zahra & George, 2002: 190). Lastly, exploitation is based on ‘the routines that allow firms to refine, extend, and leverage existing competencies or to create new ones by incorporating acquired and transformed knowledge into its operation’ (Zahra & George, 2002: 190).
When fully benefitting from PACAP and RACAP, an inter-firm relationship is generated which includes valuable knowledge-based resources. In this relationship, opportunistic behavior of either one of the parties is a potential risk (Lui, Wong, & Liu, 2009). Accordingly, both parties are expected to establish certain safeguards (Mohr & Sengupta, 2002) to minimize risk and maximize learning (Slater & Narver, 1995). The next section describes the most well-known types of ‘safeguards’, namely: contractual-based governance mechanisms and relational-based governance mechanisms.
2.2 Governance mechanisms in inter-firm relationships
THEORETICAL BACKGROUND 9 management is to minimize risk and maximize learning (Slater & Narver, 1995), both partners in the relationship have a natural need for incorporating safeguards against opportunistic behavior of the partnering firm (Heide et al., 2007; Mohr & Sengupta, 2002). Which is confirmed by Lui (2009: 333), which states: ‘to the extent that knowledge is sticky and partner opportunistic, exchange governance will play an important role in mitigating risks in inter-organizational learning’. The next section describes the types of governance included in this study.
2.2.1 Types of governance mechanisms
Literature has distinguished two types of governance mechanisms: contractual-based governance mechanisms and relational-based governance mechanisms. Although different names (e.g. contract and trust (Lui, 2009), legal and behavioral governance (Roath et al., 2002) and control and coordination (Mellewigt et al., 2007)) are used, the definitions of these two types of mechanisms are similar.
Contractual-based governance emphasizes the use of formalized, legally binding agreements or a contract to govern the partnership (Lee & Cavusgil, 2006) as well as the degree to which these are laid down in written documents or formal systems (Jansen et al., 2005). Relational-based governance, on the contrary, highlights the role of mutual trust and commitment in the alliance governance process (Lee & Cavusgil, 2006). It captures how inter-firm interactions develop self-enforcing safeguards (Heide & John, 1992; Roath et al., 2002). Both types of governance mechanisms are outlined in the formal structure as well as enforced in the inter-firm social interaction of the relationship (Zaheer & Venkatraman, 1995). However, they seem to be most effective given different circumstances. Contractual-based governance is supposed to be effective when information is highly explicit and uncertainty is low, relational-based governance mechanisms are supposed to be more effective in case information is ambiguous and organizationally embedded and in situations where uncertainty is high (Cannon, Achrol & Gundlach, 2000; Inkpen & Currall, 2004; Makhija & Ganesh, 1997; Uzzi & Lancaster, 2003).
10
2.2.2 Governance mechanisms in an inter-firm knowledge-based context
With respect to ACAP and knowledge-based research, governance mechanisms could be used to resolve the paradoxical nature of inter-firm learning. On the one hand, learning from one’s partner should be stimulated in order to maximize the effectiveness and efficiency of the relationship. However, on the other hand, transparency and sharing information should be limited in order to make sure the firm’s sources of competitive advantage will not be diluted (Mohr & Sengupta, 2002). Although governance mechanisms provide firms the security that the terms of inter-firm transactions are enforceable (Lee & Cavusgil, 2006), the potential benefits of inter-firm organizational learning remains a paradox. As argued by Mohr & Sengupta (2002: 283): ‘With respect to inter-firm learning, possibly the greatest risk comes in the potential loss of tacit knowledge to a partner (Dutta & Weiss, 1997), in which a firm’s source of competitive advantage is diluted when its partner acquires or internalizes its knowledge and skills’. This risk can cause disparities in inter-firm learning and result in an unequal distribution of benefits and risks in the relationship (Mohr & Sengupta, 2002). The following section describes the importance of the distribution of benefits in inter-firm relationships.
2.3 Distribution of benefits
As previously discussed, firms with knowledge resources that are rare, valuable, non-substitutable, and difficult to imitate are more likely to achieve a competitive advantage (Barney, 1991; Dyer & Singh, 1998). Additionally, the complementarity of knowledge resources among partnering firms could collectively generate greater benefits than the sum of these resources obtained from either one of the two partners (Dyer & Singh, 1998). Therefore the benefits that could potentially be drawn from an inter-firm relationship are important for both parties to enter and invest in the relationship as well as to determine effective governance mechanisms (Dyer & Singh, 1998).
2.3.1 Potential benefits in inter-firm absorptive capacity
Buyer-supplier interactions are seen as interactions between partners based on the reciprocity of giving and receiving relational benefits (Blau, 1964; Pulles, Veldman, Schiele & Sierksma, 2014). These relational benefits (i.e. relational rents) are ‘supernormal profits jointly generated in an exchange relationship that cannot be generated by either firm in isolation and can only be created through the joint idiosyncratic contributions of the specific alliance partners’ (Dyer & Singh, 1998: 662). An example of a relational benefit in inter-firm ACAP is the joint knowledge that contributes to research and development and innovation initiatives (Cohen & Levinthal, 1990). Potential benefits are based on interdependence between buyer and supplier that leads to a situation in which the outcomes or the payoff for both firms depends upon the actions of each other (Roath et al., 2002).
THEORETICAL BACKGROUND 11 which results in the joint creation of unique new products, services, or technologies; and (4) lower transaction costs than competitor alliances, owing to more effective governance mechanisms.
An essential element in the above mentioned research and characteristics of potential benefits in firm ACAP is that partners need each other to outperform the competitive advantage of an inter-firm relationship. In addition, they need to be willing to make relation-specific investments and both inter-firms should be open to sharing knowledge with their partnering firm.
2.3.2 Distribution of potential benefits in inter-firm absorptive capacity
Though on the one hand, inter-firm learning offers potential benefits and research is available on what is needed to attain these benefits, critics explain the risk of inter-firm learning as well as the risk of an unequal distribution of these benefits (Mohr & Sengupta, 2002). Inter-firm learning can for instance lead to unintended and undesirable skills and knowledge transfer, which results in the potential dilution of the competitive advantage involved (Mohr & Sengupta, 2002: 282). This potentially results in not attaining benefits of the inter-firm relationship by both of the parties, or by only one of the involved parties, which results in an unequal distribution of benefits (Mohr & Sengupta, 2002). As Dyer & Singh (1998: 666) found in one of their paper’s examples: ‘suppliers rationally refuse to engage in costly knowledge sharing activities, since they do not expect to receive some benefit in return’. Additionally, opportunistic behavior in inter-firm ACAP could potentially disrupt an equal distribution of benefits (Lui et al., 2009).
The risk of not being able to enhance fully, or equally from the relational benefits in an inter-firm context is severe. That is because knowledge is valuable and a first-mover advantage in ACAP is regularly found. Once a competitor has already exploited the first-mover advantage of specific knowledge in the market, this knowledge is no longer to the same extent valuable to the firm that did not benefitted from this ‘first-mover advantage’ (Cohen & Levinthal, 1990). More specific, if only the buyer, or only the supplier benefitted from the unique value of their joint knowledge and ACAP in the first place, the extent to which this benefit is equally contributing to the competitive advantage of the other firm, is minimal. Additionally, partner scarcity (Dyer & Singh, 1998) is a risk, which makes it even possible to not ever again benefit from this specific inter-firm ACAP and the competitive advantage as a result because the combination of both partners and the external knowledge sources might have been a unique opportunity.
12
3. HYPOTHESES
The previous section described and explained the theoretical background of this study. This chapter combines these previous findings and will hypothesize on the relationships between the constructs of the paper. Firstly, hypotheses (H1a-H2b) will be presented on the effect of contractual-based and relational-based governance mechanisms on potential and realized absorptive capacity (PACAP and RACAP). Which adds to the maturity (Carter, 2004) of research on inter-firm ACAP (e.g. Jansen et al., 2005; Lui, 2009). Hereafter, the expected influence of the distribution of benefits on the use of governance mechanisms in inter-firm learning will be hypothesized (H3a-H4b).
3.1 Contractual-based governance mechanisms
Formal agreements or contracts involve promises or obligations to perform particular actions in the future (Lee & Cavusgil, 2006). In case of inter-firm relationships, these promises and obligations are formally agreed upon by two firms. These future promises are not expected to stimulate the flexibility and speed that should be present in order to increase PACAP (Zahra & George, 2002). Moreover, contractual-based governance mechanisms are primarily designed to minimize risk and maximize value for the firms involved in this specific agreement (Mohr & Sengupta, 2002). It does not necessarily include restrictions, possibilities and potential benefits for the external knowledge sources outside the scope of this specific firm contract. Therefore, it is not likely that contractual-based governance in an inter-firm relationship includes external knowledge sourcing and external generation of knowledge, which are essential characteristics of both acquisition and assimilation (Zahra & George, 2002), or PACAP. Additionally, formalization reduces the likelihood of individuals to deviate from established behavior (Jansen et al., 2005), which in turn discourages individuals in either the buyer or the supplier firm to be flexible in acquiring and assimilating external knowledge. Lastly, knowledge is especially positively related to a firm’s learning capability when new knowledge is related, but at the same time different from the knowledge of contacts in the firm’s network (Zahra & George, 2002). In case of high contractual-based governance in the inter-firm relationship, this positive effect is expected to diminish as knowledge most likely is shared with and therefore not different, from the knowledge of contacts in the firm’s network. Based on these arguments, contractual-based governance is expected to be negatively related to PACAP.
H1a: Contractual-based governance will be negatively related to PACAP
HYPOTHESES 13 opportunistic behavior by the partnering firm as it protects the firm against self-interested behavior of the other party (Mellewigt et al., 2007). The minimization of this risk encourages the firms involved in the relationship to fully invest in, and benefit from, the transformation and exploitation of inter-firm knowledge sharing, and thus contractual-based governance is expected to positively relate to RACAP (Berger, 2015).
H1b: Contractual-based governance will be positively related to RACAP
3.2 Relational-based governance mechanisms
Relational-based governance in inter-firm relationships involve a certain level of mutual trust and commitment between parties (Lee & Cavusgil, 2006). Therefore, it is not expected that either one of the two parties (in this case buyer and supplier) which established relational-based governance mechanisms will take advantage of the other. More specifically, when relational-based governance is established, the fear of opportunistic behavior is limited (Gulati, 1995). Relational-based governance thus is expected to be positively related to both the acquisition and assimilation of external knowledge (PACAP), as both parties have trust in the external acquired knowledge to not be used as competitive resource against them (Dyer & Singh, 1998; Mohr & Sengupta, 2002). Additionally, research has concluded that the presence of relational-based governance and the trustworthiness of a partner lowers barriers for information sharing (Zahra & George, 2002) between parties. Lower barriers for knowledge sharing are related to PACAP and provides arguments for the expected positive relationship as new knowledge is more easily acquisitioned from partnering firms. Lastly, relational-based governance mechanisms are known to be rather effective in case of uncertainty (Cannon et al., 2000; Inkpen & Currall, 2004; Makhija & Ganesh, 1997; Uzzi & Lancaster, 2003). In addition these governance mechanisms could be relatively easily adjusted (as compared to contractual-based governance mechanisms) to a changing external environment (Zahra & George, 2002). The acquisition and assimilation of knowledge is especially concerned with possibilities in external knowledge sources, which are rather uncertain and unpredictable. Therefore, relational-based governance mechanisms are expected to stimulate the full potential of the acquisition and assimilation of knowledge, and accordingly, be positively related to PACAP.
H2a: Relational-based governance will be positively related to PACAP
14 to RACAP. Finally, previous scholars have concluded relational-based governance to foster learning curve effects (Kale, Singh & Perlmutter, 2010) and to ensure the sharing and distribution of knowledge in and amongst organizations in inter-firm relationships (Lane, Koka & Pathak, 2006). Based on these characteristics of relational-based governance and RACAP, this study hypotheses a positive effect between relational-based governance and RACAP. Figure 1 represents the hypotheses of the structural model of this paper.
H2b: Relational-based governance will be positively related to RACAP
Figure 1: Structural model of this paper
3.3 Distribution of benefits
The above mentioned hypotheses test for the effect of contractual-based and relational-based governance on PACAP and RACAP in the structural model (H1a-H2b), without taking into consideration specific characteristics of the relationship or the environment of the inter-firm relationship. This section will hypothesize on the impact of the distribution of benefits between partners in the structural model. As research on this topic is relatively limited, its contribution is exploratory in nature (Malhotra, 2009). First, hypotheses on contractual-based governance (H3a-H3b) will be formulated, after which hypotheses on the impact of the distribution of benefits on relational-based governance mechanisms (H4a-H4b) will follow.
HYPOTHESES 15 1998) as well as to provide incentives for value-creating initiatives (Lee & Cavusgil, 2006). Therefore, besides governing and controlling, governance mechanisms could be established in case of unequally distributed benefits as an encouraging incentive.
3.3.1 Impact of distribution of benefits on contractual-based governance
Contractual-based governance is characterized by its ability to ‘enforce’ agreements (Lee & Cavusgil, 2006) and to protect against opportunistic behaviour of the partnering firm in inter-firm relationships (Mellewigt et al., 2007). It could be used as an incentive to value-creating activities as well as a safeguard. Even so, in situations where one of the firms is more powerful than the other, contractual-based governance could set boundaries to what the more powerful party can do (Oosterhuis, Molleman & Van Der Vaart, 2013).
In order for PACAP to be efficient, both parties should be allowed to attain knowledge from the external environment as well as to flexibly respond to opportunities of knowledge acquisition in the external environment (Zahra & George, 2002). In case of an unequal benefit distribution, one of the firms will expect to have to invest equally in PACAP, without benefitting equally from the benefits. The initiatives to acquire and assimilate knowledge in the inter-firm relationship will therefore be either limited (Dyer & Singh, 1998), or will be based on contractual-based governance in order to ‘enforce’ initiatives by both parties and safeguard for at least a minimal benefit (Buvik & Reve, 2002). In case of an unequal benefit distribution, contractual-based governance is thus needed as a formality to ensure initiatives of both firms. However, the more contractual-based governance, the less flexibility and speed (Zahra & George, 2002), and thus the lower PACAP. The negative effect of contractual-based governance on PACAP (see Hypothesis 1a) is strengthened by an unequal distribution of benefits and therefore hypothesis H3a is formulated as follows:
H3a: The negative effect of contractual-based governance on PACAP will be positively affected (i.e. strengthened) by an unequal distribution of benefits (compared to an equal distribution of benefits) as less contractual-based governance is needed to ensure the same level of PACAP.
16 benefit distribution, in case of an unequal benefit distribution, more contractual-based governance is expected to be needed in order to fully exploit RACAP, in which this type of governance functions as a controlling device (Mellewigt et al., 2007). Therefore, it is hypothesized that an unequal distribution of benefits negatively influences the effect of contractual-based governance on RACAP.
H3b: The positive effect of contractual-based governance on RACAP will be negatively affected (i.e. diminished) by an unequal distribution of benefits (compared to an equal distribution of benefits) as more contractual-based governance is needed to ensure the same level of RACAP.
3.3.2 Impact of distribution of benefits on relational-based governance
Relational-based governance is based on mutual trust and commitment (Lee & Cavusgil, 2006) between firms and has shown to be effective in case of uncertainty (e.g. Cannon et al., 2000). This type of governance is described as a self-enforcing safeguard in an inter-firm relationship (Poppo & Zenger, 2002) and is attractive in case benefits are unequally distributed, as they are a relatively non-expensive governance mechanism (Hill, 1990; Uzi 1997) and could be adjusted continuously (Dyer & Singh, 1998). When acquisitioning and assimilating external knowledge, high uncertainty is involved, as the available knowledge from the external environment is supposed to be dynamic instead of static. The presence of commitment and trust in the relationship (i.e. relational norms) suggests a degree of willingness of both parties to invest in the relationship (Roath et al., 2002) and in acquisitioning and assimilating external knowledge to the benefit of the relationship instead of the individual firm. In case of an unequal benefit distribution, the presence of relational-based governance could be a safeguard to at least benefit to a minimal extent from the knowledge that is acquisitioned and assimilated. The more flexibility (Roath et al., 2002) and willingness and trust (Lee & Cavusgil, 2006), the less likely the party that benefits more will take advantage of the other firm (Dyer & Singh, 1998). In case of an unequal distribution of benefits, this paper however hypothesizes more relational-based governance to be needed in order to stimulate PACAP equally between partners, as the less benefitting firm needs relational-based governance as an incentive for knowledge sharing.
H4a: The positive effect of relational-based governance on PACAP will be negatively affected (i.e. diminished) by an unequal distribution of benefits (compared to an equal distribution of benefits) as more relational-based governance is needed to ensure the same level of PACAP.
HYPOTHESES 17 H4b: The positive effect of relational-based governance on RACAP will be negatively affected (i.e. diminished) by an unequal distribution of benefits (compared to an equal distribution of benefits) as more relational-based governance is needed to ensure the same level of RACAP.
Table 1 represent the hypothesized relationships of this study. The next chapter will elaborate on the methodology which is used for collecting data as well as the analyses.
Table 1: Hypotheses Hypotheses (structural model) Hypotheses (multi-group model) Contractual-based governance PACAP H1a: Contractual-based
governance will be negatively related to PACAP
H3a: The negative effect of
contractual-based governance on PACAP will be positively affected (i.e. strengthened) by an unequal distribution of benefits (compared to an equal distribution of benefits) as less contractual-based governance is needed to ensure the same level of PACAP
RACAP
H1b: Contractual-based
governance will be positively related to RACAP
H3b: The positive effect of
contractual-based governance on RACAP will be negatively affected (i.e. diminished) by an unequal distribution of benefits
(compared to an equal distribution of benefits) as more contractual-based governance is needed to ensure the same level of RACAP
Relational-based governance
PACAP
H2a: Relational-based
governance will be positively related to PACAP
H4a: The positive effect of
relational-based governance on PACAP will be negatively affected (i.e. diminished) by an unequal distribution of benefits
(compared to an equal distribution of benefits) as more relational-based governance is needed to ensure the same level of PACAP
RACAP
H2b: Relational-based
governance will be positively related to RACAP
H4b: The positive effect of
relational-based governance on RACAP will be negatively affected (i.e. diminished) by an unequal distribution of benefits
18
4. METHODOLOGY
4.1 Research design
This paper focuses on the effect of governance mechanisms on absorptive capacity (ACAP) in vertical inter-firm relationships. On the one hand, the paper has a conclusive character by testing the effect of governance mechanisms on ACAP, which previous research has focused on in different forms (e.g. Berger, 2015; Jansen et al., 2005). On the other hand, by including the potential influence of the distribution of benefits between buyer and supplier, the main purpose of this research is exploratory (Malhotra, 2009). It provides an answer to the research question: “To what extent does the distribution of benefits in a buyer-supplier relationship influence the effects of governance mechanisms on absorptive capacity?” Given the exploratory nature of this research, concluding on how, and why the distribution of benefits could influence the structural model is out of scope in this paper (Yin, 2003) and recommended for future research.
In exploratory research, it is most appropriate to base results on primary data via a questionnaire in order to answer the research question (Malhotra, 2009). A questionnaire provides several advantages that support and enable the aim of this paper. First, it allows for analysis on (the significance of) the effect of governance mechanisms on ACAP as well as a comparison between-groups. Secondly, compared to other research methods, a questionnaire provides the potential for a relative large sample size (Yin, 2003), which contributes to the robustness and reliability of results. Lastly, as respondents will be informed on the absolute confidentiality of the questionnaire and there is no direct contact between the researcher and respondents, the social desirability of the answers provided in the questionnaire is expected to be minimal (Malhotra, 2009).
4.2 Sample and data collection
METHODOLOGY 19 The distribution of the questionnaire is performed by firstly contacting buyer companies, which were all located in the Netherlands. After the recruitment of the buyer-firm informants, they were requested for contacts in their supplier organization. The questionnaire was then distributed either physically or via email to the key informants of both the buyer and supplier, which were informed on the confidentiality of their responses. This resulted in a data set of 166 matched-pair relationships (Berger, 2015). However, as the unit of analysis of this study is the firm-level and not the dyadic-relationship-level, the total amount of cases is 332 (i.e. buyer and supplier data).
4.3 Measurement development
As mentioned in the previous section, the questionnaire that is used to collect data has been developed in previous research on ACAP in an inter-firm context (Berger, 2015). All items that are used in the structural model of this research are constructed based on measurement items on a 7-point Likert scale ranging either from ‘strongly disagree (1)’ to ‘strongly agree’ (7) (Jap, 1999). The distribution of benefits is measured on a 7-point Likert scale ranging from ‘seller (1)’ to ‘buyer (7)’1 in order to determine whether the benefits are equally distributed among buyer and seller (i.e. supplier).
The questionnaire and the data of this paper contains both first-order as well as second-order constructs. A first-order construct is a latent variable that has (multiple) observed variables as indicators. A second-order construct, or higher-order latent variable is not directly connected to any measurement item and has other (first-order) constructs as indicators (Berger, 2015). Potential absorptive capacity (PACAP) for instance is a second-order construct, of which acquisition and assimilation (first-order constructs) are indicators. Moreover, the data differentiates reflective and formative scales (Henseler, Ringle, & Sinkovics, 2009). Reflective measures are expected to be correlated, and should be internally consistent as dropping one indicator should not alter the meaning of the construct (Jarvis, MacKenzie & Podsakoff, 2003). Formative indicators, however, are causing the latent construct, and are not interchangeable per se (Jarvis et al., 2003). Both types of variables should be analysed on validity differently, which will be explained later.
The constructs of the study are measured by several measurement items, which are displayed in more detail in appendix A. Contractual-based governance is measured as ‘the degree of formal contractual control in the buyer-supplier relationship’ (Geringer & Hebert, 1989). The measurement items of this construct are derived and obtained from previous research of Buvik & Reve (2002), Cannon & Perreault (1999), Deshpande & Zaltman (1982) and Jansen, Van Den Bosch & Volberda (2006). Relational-based governance is a second-order construct, based on the measurement items of flexibility norms, information exchange norms and solidarity norms (Heide & John, 1992). Flexibility norms are described as ‘the parties’ expectations of making adjustments in the ongoing relationship in accordance with changing circumstances’. Information exchange norms are ‘the expectation that particular pieces of information that might help the other party will be provided’. Lastly solidarity norms are measured as ‘the expected efforts of the parties toward preserving the relation’ (Heide & John, 1992: 38).
1 An example: “Who has benefited most from the following performance measures?
20 The potential absorptive capacity (PACAP) as well as the realized absorptive capacity (RACAP) are second-order constructs which are measured by first-order constructs, based on the four phases of ACAP (Zahra & George, 2002). Acquisition is included in explaining PACAP and refers to ‘a relationship’s capability to identify and acquire externally generated knowledge that is critical to its operations’ (Berger, 2015: 80). The second dimension that is used in measuring PACAP is assimilation which is measured as ‘the firm’s routines and processes that allow it to analyse, process, interpret and understand the information obtained from external sources’ (Berger, 2015: 80). The measurement items are constructed in the research of Berger (2015), based on previous work of Camisón & Forés (2010) and Szulanski (1996).
Transformation, which is a first-order construct in measuring RACAP, is defined as ‘a firm’s capability to develop and refine the routines that facilitate the combination of existing knowledge and the newly acquired and assimilated knowledge’ (Berger, 2015: 80). The measurement items are based on a combination of items previously constructed by Bontis, Crossan, Mary & Hulland (2002) and Cadiz, Sawyer & Griffith (2009). Exploitation (the second construct of which RACAP is composed) are ‘routines that allow firms to refine, extend, and leverage existing competencies or to create new ones by incorporating acquired and transformed knowledge into its operations’ (Berger, 2015: 80). Cadiz et al. (2009), Camisón & Forés (2010) and Jansen, van den Bosch & Volberda (2005) provide the basis for the measurement items of this construct.
Lastly, relationship performance (i.e. relationship benefit) is measured as a combination of explorative and exploitative learning performance (Berger, 2015). In our research, however, the emphasis is not on the actual performance of the relationship, but rather on the distribution of these benefits between buyer and supplier. Therefore, we have calculated the distribution of benefits as one variable that is composed of whether the buyer or the seller benefits more from explorative and exploitative performance, based on the measurement items of Lane, Salk & Lyles, (2001) and Selnes & Sallis (2003). Both types of performance have not been split, but have been combined in order to explain which party benefits more from the overall relational benefits of the inter-firm relationship. It is not expected that the measurement items of the construct are correlated and each item is expected to individually cause the latent construct. Therefore, the distribution of benefits is a formative scale, as explorative and exploitative learning performance were in the research of Berger (2015). Data validity will be explained in the following chapter and the developed constructs are presented in Appendix A.
4.4 Data analyses
METHODOLOGY 21 constructs. Given that these characteristics are all included in this paper, PLS is used for analyses. The analyses have been performed using a step-wise approach which will be briefly explained below.
4.4.1 Step 1: Preparing the data for analysis
In order to analyse the structural model as well as the multi-group model, the first and second-order constructs needed to be developed in SmartPLS. This has been done by calculating the latent variable scores for each first-order construct. These are then used as indicators for the second-order constructs. After calculating these scores, the second-order constructs PACAP, RACAP and relational-based governance were created.
Afterwards, the distribution of benefits is calculated and included based on all questions related to which party benefits most from the explorative and exploitative learning performances in the inter-firm relationship. See appendix A for a detailed description of all measurement items of this variable.
4.4.2 Step 2: Analysing the structural model
In this step, the structural conceptual model is tested using the PLS Algorithm in SmartPLS in order to determine the path coefficients (Beta = β) of the hypothesized effects. Additionally, bootstrapping will provide results on the significance of the found effects.
4.4.3 Step 3: Analysing the multi-group model
22
5. RESULTS
This section includes the results of the analyses. Firstly, the results on data validity will be presented. Second, the effects of contractual-based and relational-based governance mechanisms on the potential and realized absorptive capacity (PACAP and RACAP) will be examined, resulting in the outcomes for hypotheses H1a-H2b. Lastly, a first, exploratory, insight on the influence of the distribution of benefits on the structural model is presented, providing results for hypotheses H3a-H4b.
5.1 Data validity
Before assessing the validity of the data, a check for missing data was performed. Out of 332 cases, 18 cases contained missing data on the items included in the structural model of the paper. The missing values were coded and case-wise excluded in order to allow for analyses based on unique cases and to not cause any error in the results (Malhotra, 2009).
Secondly, an overall factor analysis (including all variables and measurement items) has been performed to test whether the data is likely to factor well. All Kaiser-Meyer-Olkin Measures of Sampling Adequacy (KMO values) well exceeded the threshold of 0.5, which supports that the data is likely to factor well, based on correlations and partial correlation (Malhotra, 2009). Additionally, the Bartlett’s test of sphericity showed significant support, which means the measurement items correlate significantly.
As mentioned previously, formative and reflective scales require a different approach in testing for their construct and convergent validity (Hair, Black, Babin & Anderson, 2010). The results of the reflective scales will firstly be presented, after which the formative scales are described.
5.1.1 Reflective scales
RESULTS 23
Table 2: Construct reliability and convergent validity of reflective scales
Constructs Buyer and supplier data
Measurement items Loadings Cronbach’s
alpha
Composite reliability
Assimilation 0,79 0,85
Assimilation 1 (Reversed coded) 0,67
Assimilation 2 0,77 Assimilation 3 0,81 Assimilation 4 0,75 Assimilation 5 0,66 Exploitation 0,74 0,83 Exploitation 1 0,66 Exploitation 2 0,65 Exploitation 3 0,78 Exploitation 4 0,85 Exploitation 5 0,54 Contractual-based governance 0,82 0,87 Contract. 1 0,61 Contract. 2 0,80 Contract. 3 0,87 Contract. 4 0,76 Contract. 5 0,57 Contract. 6 0,66 Flexibility norms 0,74 0,85 Flexibility 1 0,81 Flexibility 2 0,84 Flexibility 3 0,78
Information exchange norms 0,78 0,86
Inform. 1 0,74 Inform. 2 0,79 Inform. 3 0,78 Inform. 4 0,81 Solidarity norms 0,83 0,90 Solidarity 1 0,88 Solidarity 2 0,90 Solidarity 3 0,82 Relational-based governance (2nd order construct) 0,88 0,90 Flexibility norms 0,76
Information exchange norms 0,88
24
5.1.2 Formative scales
Formative items are not supposed to correlate with each other, as they are related to different dimensions of a construct. Although traditional assessments of convergent and discriminant validity are irrelevant for formative measures (Johnson, Cullen, Sakano & Takenouchi, 1996; Hulland, 1999) and validity is largely based on content validity (Tsang, 2002), there are some measurements that could provide evidence on the appropriateness of the formative scales in this paper. Convergent validity is measured by assessing the p-values and variance inflation factors (VIF’s) of indicator weights. Where the VIF score in reflective scales should not exceed the cut-off point of 10.0, in case of formative scales, the VIF value should not exceed 3.3 (Diamantopoulos & Winklhofer, 2001) and the lower the VIF score, the better. These measures are displayed in table 3. Based on inappropriate measures, ‘Acquisition 2’, ‘Transformation 2’ and ‘Transformation 3’ were excluded. However, given the characteristics of a formative construct, excluding these measurement items could potentially influence the content of the construct as all items are related to different dimensions of the construct. According to the argumentation of Berger (2015: 99) the exclusion of these three items does not affect the theoretical domain of the construct which is still captured adequately by the remaining items of the construct. Additionally ‘Distribution of benefits 9’ showed a rather low indicator loading. However, the VIF value is appropriate and thus ‘Distribution of benefits 9’ remains included in the analysis. All other constructs and indicators show appropriate results that support the convergent validity of the formative scales as presented in table 3.
5.2 Results of the structural model
RESULTS 25
Table 3: Convergent validity of formative scales
Constructs Buyer and supplier data
Measurement items Indicator weights p – value VIF
Acquisition Acquisition 1 0,31 < 0,001 1,27 Acquisition 3 0,31 < 0,001 1,18 Acquisition 4 0,38 < 0,001 1,44 Acquisition 5 0,38 < 0,001 1,47 Transformation Transformation 1 0,30 < 0,001 1,41 Transformation 4 0,21 < 0,001 1,21 Transformation 5 0,11 < 0,001 2,82 Transformation 6 0,26 < 0,001 2,47 Transformation 7 0,26 < 0,001 1,84 Transformation 8 0,22 < 0,001 2,17 Distribution of benefits Distribution of benefits 1 0,30 < 0,001 1,70 Distribution of benefits 2 0,16 < 0,001 1,87 Distribution of benefits 3 0,08 < 0,001 1,69 Distribution of benefits 4 0,32 < 0,001 2,15 Distribution of benefits 5 0,25 < 0,001 2,27 Distribution of benefits 6 0,73 0,03 1,80 Distribution of benefits 7 0,43 < 0,001 1,68 Distribution of benefits 8 0,53 0,04 1,49 Distribution of benefits 9 0,03 < 0,001 1,77 Distribution of benefits 10 0,28 < 0,001 1,94 Distribution of benefits 11 0,19 < 0,001 1,90 Distribution of benefits 12 0,70 0,02 1,42
Potential absorptive capacity
(PACAP) (2nd order construct)
Acquisition 0,46 < 0,001 1,65
Assimilation 0,65 < 0,001 1,87
Realized absorptive capacity
(RACAP) (2nd order construct)
Transformation 0,65 < 0,001 2,08
Exploitation 0,44 < 0,001 1,94
26
Table 4: Results of the structural model
Independent variable Path-coefficient (Beta β) T-statistic (*significance) Dependent variable Contractual-based governance PACAP 0,110 1,795 * RACAP 0,080 1,178 Relational-based governance PACAP 0,350 5,955 **** RACAP 0,450 7,389 ****
* 90% confidence level, ** 95% confidence level, *** 98% confidence level, **** 99% confidence level
5.3 Results of the multi-group model
In order to test for the results on the multi-group model, the sample is distributed in three groups, based on the scores on the variable ‘Distribution of benefits’, which measures whether either one of the two firms benefitted more from the outcomes of inter-firm learning. As this was measured on a 7-point Likert scale ranging from ‘seller (1)’ to ‘buyer (7)’, the groups are divided as follows: (1) ‘seller (i.e. supplier) benefit’, all measurements < 3.5; (2) ‘equal benefit’, all measurements ≥ 3.5 and < 4.5; (3) ‘buyer benefit’, all measurements ≥ 4.5. Table 5 visually represents our multi-groups as well as the number of cases for each group.
Table 5: Multi-group samples
All data
Seller (i.e. supplier)
benefit
Equal benefit Buyer
benefit
Likert-scale values range: 1 - 7 < 3.5 ≥ 3.5 and < 4.5 ≥ 4.5
Number of cases total: 304 *
missing: 28 * 18 * 181 * 105*
* The total number of cases equals 304, which is composed of the original sample (n = 332) and case-wise exclusion of missing value cases (n = 28)
As shown in table 5, the number of cases in which the supplier benefits most from the inter-firm relationship is relatively small (n = 18). This small sample size could cause problems and error in the results, as it does not comply with the minimum sample size which is recommended in PLS studies by Chin (1998). This rule is that sample sizes should either be at least ten times the amount of indicators of the scale with the largest number of formative indicators (in this paper: six indicators), or ten times the largest number of structural paths directed to a dependent variable (in this paper: two structural paths).
RESULTS 27 benefit’ group. If both groups do not differ significantly, these groups will be recoded as one, namely the ‘unequal benefit’ group. Table 6 shows the results of the PLS Algorithm, Bootstrapping and the Chi-square difference test between ‘supplier benefit’ and ‘buyer benefit’.
Table 6: PLS Algorithm, Bootstrapping and Chi-square difference test between 'supplier benefit' and 'buyer benefit'
Supplier benefit (n = 18) Buyer benefit (n = 105) Chi-square difference T-test Independent variable Path-coefficient (Beta β) T-statistic (*significance) Path-coefficient (Beta β) T-statistic (*significance) T-statistic (*significance) Dependent variable Contractual-based governance PACAP 0,450 1,204 0,129 0,953 0,543 RACAP 0,490 0,981 0,075 0,448 0,446 Relational-based governance PACAP 0,513 2,790 **** 0,138 1,305 1,082 RACAP 0,450 2,026 ** 0,210 1,996 ** 0,454
* 90% confidence level, ** 95% confidence level, *** 98% confidence level, **** 99% confidence level
As shown by the Chi-square T-statistics, although some differences are found, the ‘supplier benefit’ and ‘buyer benefit’ group do not differ significantly. This could be explained by the small sample of ‘supplier benefit’ which is insufficient to provide significant statistic support for the difference between groups. Therefore, these two groups are recoded into a new group for multiple-group analysis, namely ‘unequal benefit’ group. This adds to the robustness and reliability of results in the multi-group analysis by reducing the possibility of error which could be caused by the small sample of ‘supplier benefit’. The ‘unequal benefit’ group includes 123 cases, which well exceeds the minimal requirements as stated by Chin (1998).
28
Table 7: PLS Algorithm, Bootstrapping and Chi-square difference test between 'equal benefit' and 'uequal benefit'
Equal benefit (n = 181) Unequal benefit (n = 123) Chi-square difference T-test Independent variable Path-coefficient (Beta β) T-statistic (*significance) Path-coefficient (Beta β) T-statistic (*significance) T-statistic (*significance) Dependent variable Contractual-based governance PACAP 0,115 0,949 0,196 1,431 0,555 RACAP 0,075 0,783 0,144 0,886 0,502 Relational-based governance PACAP 0,432 6,381 **** 0,217 2,354 *** 1,793 * RACAP 0,594 9,600 **** 0,280 2,618 **** 2,657 ****
* 90% confidence level, ** 95% confidence level, *** 98% confidence level, **** 99% confidence level
The multi-group analysis showed no significant support for the effect of contractual-based governance on PACAP, neither in the ‘equal benefit’ nor in the ‘unequal benefit’ group. Moreover, there has been no support found for the statistical difference between the effects of contractual-based governance on PACAP between these groups. Therefore, no evidence is found to support hypothesis 3a. Hypothesis 3b, which predicted the effect of contractual-based governance on RACAP to be negatively influenced in case benefits were unequally (compared to equally) distributed, showed no significant support. No support was found in the separate multi-group models and the Chi-square test provided no significant evidence for the difference between-groups. These results provide no evidence for the hypothesized impact of the distribution of benefits on the effect of contractual-based governance in inter-firm ACAP.
DISCUSSION AND CONCLUSION 29
6. DISCUSSION AND CONCLUSION
This paper researches the impact of governance mechanisms on the absorptive capacity (ACAP) of inter-firm relationships and to what extent these will be influenced by the distribution of benefits in this relationship. This chapter discusses the findings and results of the paper in light of existing literature and concludes on the contribution of the paper in the conclusion. The first section includes the discussion of the findings in the structural model, the second section discusses the impact of the distribution of benefits.
6.1 Impact of governance mechanisms in inter-firm learning
The structural model, testing for the effect of governance mechanisms on inter-firm ACAP is a contribution to the integrity and maturity of research on ACAP (Carter, 2004; Kaynak & Hartley, 2008). Although most papers (Berger, 2015; Dyer & Singh, 1998; Jansen et al., 2005) hypothesize a negative effect of contractual-based governance on PACAP, hardly any results significantly supported this effect. The positive hypothesized effects of contractual-based governance on RACAP and relational-based governance on PACAP and RACAP have already been supported by previous scholars (e.g. Berger, 2015; Jansen et al., 2005).
30 separate dimensions and support has been found for the positive relationship between
contractual-based governance and knowledge acquisition. The paper of Lui (2009) found no significant support on the effect of contractual-based governance and knowledge assimilation.
The positive effect of contractual-based governance on RACAP could not be significantly supported based on the results from this paper. A significant support for this effect however has been found in research of Berger (2015) which split the analysis based on data from buyer and supplier separately. As buyer and supplier perceptions of inter-firm relationships could differ (Oosterhuis et al., 2013), we have split our results based on data from buyer and supplier separately (see Appendix B) in order to delve deeper into this non-significant hypothesis. As we can see from this analysis, data of both buyer and supplier separately, as well as combined lacks to significantly confirm the positive effect of contractual-based governance on RACAP. We therefore cannot conclude that formalization and routines from contractual-based governance contributes to the effective transformation and exploitation of knowledge in inter-firm relationships (Zahra & George, 2002). Additionally, the separate buyer and supplier data reveals that no significant support is found for contractual-based governance on PACAP in the supplier-only data. This might suggest why the confidence level (i.e. 90%) of the significant effect of contractual-based governance on PACAP in the structural model (including all data) is relatively low.
The results of the effect of relational-governance on PACAP as well as RACAP were both positive and strongly significantly supported in this paper. Relational-based governance, as hypothesized, has a positive impact on both dimensions of ACAP. Relational-governance is expected to indeed minimize the fear of opportunistic behaviour as well as to stimulate knowledge sharing in an inter-firm context. These characteristics, as supported by this paper, stimulate the acquisition and assimilation of knowledge from external sources as well as the transformation and exploitation of knowledge within and between firms. By using relational-based governance in inter-firm relationships, firms stimulate PACAP and RACAP that could not have been generated by one firm in isolation (Dyer & Singh, 1998). Results of relational-based governance on both PACAP and RACAP in the buyer and supplier data separately (Appendix B) support the findings in our structural model and do not statistically differ based on the Chi-square difference test. This finding adds to the robustness and reliability of the results in the structural model.
6.2 Impact of distribution of benefits on governance mechanisms in inter-firm
learning
DISCUSSION AND CONCLUSION 31 Based on this study, the impact of the distribution of benefits unfortunately cannot be concluded on this level of detail. Future research is recommended to strive for this level of detail as it might explain why either the buyer or supplier is more willing to establish and initiate governance mechanisms. As mentioned in the results chapter, this paper therefore concludes based on the analysis of the group which ‘equally benefits’ from relational benefits compared to the group which ‘unequally benefits’.
The multi-group analysis did not support the hypotheses on the effect of contractual-based governance on either PACAP or RACAP, whether in case of an equal or unequal distribution of benefits. Additionally, the Chi-square test did not show any significant difference between both groups in case of contractual-based governance. Therefore, we again split these results based on buyer and supplier data separately, in order to conclude on the robustness of the results found. These findings are presented in Appendix C. In these findings, we see that a significant effect of contractual-based governance on PACAP is found in case of an unequal distribution of benefits, in the buyer-only data. This is potentially caused by the large sample of ‘buyer benefit’ compared to ‘supplier benefit’ (table 5) and indicates that buyers are more willing to ensure and enforce the cooperation of suppliers by contractual-based governance in case benefits are unequally distributed. However, this finding does not statistically differ from the non-significant finding in the supplier-only data. Which means we have not find enough statistical evidence to conclude on whether the impact of the distribution of benefits on governance mechanisms in ACAP differs between buyer and supplier data. Therefore, the results of our structural model (i.e. including buyer and supplier data) seem robust and reliable.
This again raises the issue to what extent contractual-based governance is an effective governance mechanism on ensuring PACAP in inter-firm learning. Given the exploratory nature of Hypothesis H3a-H4b, and as most of the outcomes of contractual-based governance on PACAP are not significantly supported, the conclusion that contractual-based governance might not be effective in stimulation PACAP should be cautiously interpreted. The effect of contractual-based governance mechanisms on RACAP was expected to be higher in case benefits were unequally distributed, as this was supposed to be initiated by the firm with the least benefit to be able to enforce a minimal benefit in the inter-firm relationship. Additionally, in case of an unequal distribution of benefits (compared to an equal distribution), contractual-based governance was supposed to ensure routines and formalization to support the transformation and exploitation of knowledge within and between-firms. Based on the previous discussion, it can be concluded that the distribution of benefits has no significant impact on the effect of contractual-based governance on RACAP. Future research is needed to examine whether this finding is significantly supported for instance in a different research context, or when taking into account the development of the inter-firm relationship2.
Both hypotheses H4a and H4b expected the effect of relational-based governance on PACAP and RACAP to decrease in case benefits were unequally distributed. Which suggests more relational-based governance is needed when benefits are unequally distributed. The main arguments for these hypotheses were that relational-based governance have shown to be effective in case of uncertainties
32 (e.g. Cannon et al., 2000) which is argued to be the case if benefits are unequally distributed. In addition, relational-based governance could be used to minimize opportunistic behavior as well as firms taking advantage of each other (Dyer & Singh, 1998; Roath et al., 2002) but more relational-governance would be needed to identically stimulate PACAP and RACAP in case of an unequal benefit distribution compared to an equal benefit distribution. Results have shown for PACAP as well as for RACAP, the effect of relational-based governance in case of an unequal distribution of benefits was significantly lower than in case benefits were equally distributed. The outcomes of both models (equal and unequal distribution of benefits) as well as the Chi-square difference test significantly supported this outcome.
There are some potential additional explanations for this outcome. Firstly, the nature and intent of the inter-firm relationship could be of influence as Mohr & Sengupta (2002: 289) argue effective knowledge transfer to not only require effective governance mechanisms. They argue effective ACAP to be a combination of appropriate governance, learning risks, the type of knowledge, the learning intent of the partner as well as the duration that partnering firms have in mind for their relationship. In case the governance mechanisms do not fit or comply with the other requirements, the effect of governance mechanisms on the development of PACAP and RACAP might be limited. When benefits are unequally distributed amongst buyer and supplier, disparities are expected to occur in the relationship for instance based on the duration of the relationship. The partner that benefits most might be willing to continue the relationship, while the firm with the least benefit is expected to either change, or exit the relationship based on the reciprocity of giving and receiving relational benefits (Blau, 1964; Pulles et al., 2014). Secondly, if the nature of the relationship (arms-length versus an extension of arms-length) differs in the perception of both firms (Dyer & Singh, 1998), it might be the case that only one of the two parties invests in relational-based governance (i.e. mutual trust and commitment). This is most likely if the distribution of benefits is unequally divided, as one of the firms has a more transactional approach and is primarily focused on individual benefits instead of joint benefits (Whipple, Lynch & Nyaga, 2010). In this situation, relational-based governance will probably not be an effective governance mechanism to stimulate ACAP as this type of governance does not ensure the alignment of intentions. Lastly, if one of the firms structurally benefits less from the inter-firm relationship, there is no intention to establish relational-based governance. This suggestion is supported by research of Dyer & Singh (1998) which state that firms refuse to engage in knowledge-sharing activities in case they never expect to receive benefits in return.
