Programme: MSc Business Administration – Strategy
Master Thesis – Final version
Dedicated Alliance Function: Governance and Performance
LUCA PIAZZA 13812963
Thesis supervisor: Dr. Elko Klijn Second reader: Dr. Marten Stienstra
Submission date: 24th
2 Statement of originality
This document is written by Student Luca Piazza who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
I would like to express my thanks to my supervisor Dr. Elko Klijn for his guidance in this research. Furthermore, I would like to thank all the people who supported me during these university years.
3 Table of Contents
1. Introduction ... 5
2. Theoretical Background ... 8
2.1 Dedicated Alliance Function ... 10
2.1.1 Benefit 1: Improving knowledge management ... 11
2.1.2 Benefit 2: Increasing external visibility ... 11
2.1.3 Benefit 3: Providing internal coordination ... 12
2.1.4 Benefit 4: Facilitating intervention and accountability ... 12
2.2 Alliance Governance ... 13
2.2.1 Alliance contracts ... 14
2.2.2 Contractual completeness ... 15
2.2.3 Board Composition ... 16
2.2.4 Board size ... 17
3. Hypothesis development ... 18
3.1 Dedicated alliance function and performance ... 18
3.2 Dedicated alliance function and governance ... 19
3.3 Alliance governance and alliance performance ... 21
4. Methodology ... 23
4.1 Sample and data collection ... 23
4.2 Variables and measurement ... 24
4.2.1 Dependent variable: Alliance performance ... 24
4.2.2 Independent variable: The dedicated alliance function ... 25
4.2.3 Mediating variable: Alliance governance ... 25
4.2.4 Control variables ... 27
5. Results ... 28
6. Discussion ... 36
6.1 Contributions and implications ... 36
6.2 Limitations and future research directions ... 40
7. Conclusion ... 42
List of References ... 45
APPENDIX 1 – Measurement Scales ... 52
Despite decades of research demonstrating the potential benefits of alliances (JVs), the strategic literature has yet to define a way to maximise the likelihood of success of a Joint Venture.
While previous studies have provided many insights regarding the importance of the due diligence phase as a basis for alliance success, more recent findings have shown that many companies that are successful in forming alliances achieve this result through the establishment of a dedicated alliance function. However, in the literature on alliances, the dedicated alliance function (DAF) is still largely treated as a black box. To shed light on this and elucidate how the dedicated alliance function has a positive impact on alliance performance, this research has devoted significant attention to the exploration of the potential influence of a DAF on the governance of the alliance, a crucial determinant of alliance success. Based on data from a 2014 dataset generated by researchers from a Dutch business school, the mediation model of this thesis has been analysed by multiplying ‘alliance governance’ into two sub-elements:
‘board size’ and ‘contract completeness’. The main results indicate that the dedicated alliance function has a significant impact on alliance performance and that this relationship is mediated by the effect the DAF has on the size of the board of directors. This finding expands the literature on JVs, finally providing a method to address the low success rate that alliances have experienced in recent years.
5 1. Introduction
No company can go it alone (Doz and Hamel, 1998, page. IX). For industry leaders and ambitious start-ups alike, strategic partnerships have become a fundamental business strategy to achieve a competitive advantage in fast-changing global markets (Das and Teng, 2000; Doz and Hamel, 1998; Dyer and Singh, 1998; Heimeriks 2015; Zajac, 1998). More than ever, many of the resources and capabilities critical to a company’s future success lie outside the level of the individual firm and can be found at the level of relationships (Dyer and Singh 1998; Gulati, 1995; Hoffmann, 2005). In this new world of networks, strategic alliances are not more an option but a necessity (Doz et al., 1998, page. IX). To fully exploit the opportunities derived from them, firms nowadays must have the capacity to form, steer, and sustain a multitude of alliances. Indeed, today, a wide variety of partnerships is transforming industries from healthcare to information technology, media, education, entertainment, space, and beyond (Russo et. al, 2019; Niesten, 2015). Thus, the development of the ‘capacity to collaborate’ has become an essential and determining aspect for the future prosperity of any company.
Following this trend, in recent years, interfirm alliances have entered the agendas of most organizations and have received considerable research attention (e.g., Child, Faulkner, and Tallman, 2005; Contractor and Lorange 2002; Dussauge and Garrette, 1999; Hennart, 2006).
Among the numerous studies on the subject, most research has shown that alliances present a paradox for firms. Strategic alliances are difficult to manage, and firms generally fail with about half of the alliances they form (Alliance Analyst, 1996; Bleeke and Ernst, 1993; Kogut, 1988). Given these surprisingly low success rates, academics and managers have become extremely interested in studying the design of strategic alliances and understanding how these organizational forms should be governed to increase the probability of alliance success (Hoetker et al, 2009; Lumineau, 2014; Williamson, 1985).
6 Unfortunately, due to the extreme specificity of each alliance, researchers have not been able to identify a single type of governance structure capable of handling all types of alliances, indicating that there are multiple best practices available (Mellewgit et al., 2007; Hoetker et al., 2009; Dyer and Singh, 1998). These differ according to the levels of three main dimensions:
centralization, formalization, and specialization (Albers, 2010, Mintzberg, 1979). Despite the number of governance structures that can be created by combining the different levels of these three components, many studies have confirmed that just a few are successful in managing alliance processes (Lumineau, 2014). Further studies confirm that, showing that firms exhibit significant heterogeneity in terms of their overall alliance success and highlighting that some firms are much more successful at managing alliances, or creating value from them, than other firms (Anand and Khanna, 2000; Kale, Dyer, and Singh, 2002). Companies that are more successful in alliances are presumed to have a better dimension of specialization, which, by centralizing processes, would help them to develop alliance capability and have a greater learning capacity (Findikoglu et al. 2019; Kale et al., 2002). Hence, recently, researchers have shifted their attention to designing dedicated alliance functions, evaluating the impact these functions can potentially have on the governance structure of alliances and, therefore, on the overall alliance performance (e.g., Russo and Vurro, 2019).
Extant research has shown that the institutionalization of formal and centralized mechanisms such as the dedicated alliance function (DAF), contributes to the performance of the alliance portfolio, creating many potential benefits for allied organizations (Findikoglu et al. 2019; Kale et al., 2002). Establishing a DAF, responsible for capturing prior alliance experience, could help to bundle all alliance-related knowledge within a firm, providing the organization with an effective tool to create and develop an alliance capability (Hoffmann, 2005). The creation of an infrastructure that collects, shares, and diffuses experience in
7 managing alliances throughout the organization can improve internal coordination and resource support, as well as monitor and evaluate alliance performance (Dyer et al., 2001).
But if there are all these benefits, why do not all organizations have this specific function?
As statistics teaches us, we must be careful not to get confused between correlation and causation. Simply setting up a dedicated alliance function is not sufficient to achieve alliance success. The strategic literature has suggested that the success of an alliance depends primarily on its management (Kale et al., 2002; Niesten et al., 2015; Schreiner et al., 2009). Effective corporate governance has been identified as the fundamental requirement for sustained inter- firm collaboration, together with the creation of a solid business plan backed up by a detailed contract (Alliance Analyst, 1996; Contractor et al., 2000; Mellewigt et al., 2007) As such, the dedicated alliance function could only have a positive impact on alliance performance if it can positively influence alliance governance.
Therefore, to elucidate the value of the dedicated alliance function on the company’s ability to build an effective alliance governance structure and inform the debate in the literature about its merits on alliance performance, this thesis will answer the following research question: “Does the dedicated alliance function have a positive impact on the governance of the alliance, which, in turn, results in higher alliance performance?”. This will be determined by how the DAF influences, specifically, the composition of the board of directors and the details of the alliance contract, two important governance mechanisms.
This study has been carried out looking precisely at Joint Ventures (JVs). These have been chosen given their increasing number in recent years, which have made this type of alliance a fundamental feature of the global business landscape and one of the most common entry mode options to penetrate new territories (Cakoghirou et al., 2003; Whitelock, 1993).
Building on the above theoretical arguments, this paper has analysed the relationships between dedicated alliance function, alliance governance, and alliance performance. After
8 careful empirical analysis, the results of this thesis indicate that the dedicated alliance function has a significant impact on alliance performance and that this relationship can be explained by the effect that the DAF has on the board of directors. Contrary to the effect on the board of directors, this thesis did not support the hypotheses concerning the other governance mechanism, contract completeness, therefore concluding that the positive impact of the dedicated alliance function on alliance performance does not result from the mediation of increased contract completeness.
In sum, the findings of this study contribute to the emerging stream of research in three ways. First, it advances the literature on JVs by analysing the relationship between a dedicated alliance function and alliance performance. This relationship has been analysed by considering the impact of the dedicated alliance function on a critical determinant of alliance success not yet considered in the dedicated alliance function literature, alliance governance. Second, it shows that the dedicated alliance function has a positive direct impact on the improvement of the overall alliance performance, therefore clarifying the role of the establishment of a dedicated alliance function on both alliance governance and alliance performance. Third, broadly speaking, this research contributes to (dynamic) capability research in general by indirectly showing that an alliance learning process helps build up the creation of an alliance capability, which, in turn, is positively related to a firm’s overall alliance performance. Overall, this study opens the ‘black box’ between the dedicated alliance function and the overall alliance success in firms, contributing considerably to the strategy literature.
2. Theoretical background
Over recent decades, the world economy has become increasingly globalised in response to competitive pressures and a constant search for innovation. Disruptive technologies and accelerating rates of change have led to a growing interdependence of firms as a decisive tool to create flexible options in an uncertain future environment (Gulati 1995; Das and Teng, 2000;
9 Doz and Hamel, 1998; Dyer and Singh, 1998; Zajac, 1998). Consequently, strategic alliances have become a popular business strategy to achieve efficiency gains, exploit synergies, and respond to macroeconomic changes (e.g., Child et al., 2005; Contractor and Lorange 2002;
Dussauge and Garrette, 1999; Hennart, 2006).
Strategic alliances (also called ‘strategic partnerships’) can briefly be defined as
“purposive relationships between two or more independent firms that involve the exchange, sharing, or co-development of resources and capabilities to achieve mutually beneficial outcomes” (Gulati, 1995, page 86) The term ‘strategic’ emphasises the fundamental purpose of corporate alliances: achieve a strategic goal. Whereas the term ‘partnership’ or ’alliance’
describes the cooperation between two or more companies.
Depending on the level of resource commitment and knowledge integration, strategic alliances can be of various types (Schildt et al., 2005). This thesis will focus on just one of them, Joint Ventures (JV), hereafter also referred to as ‘alliance’. Joint Ventures refer to the
“partnership in which a new legal entity and organization are formed by two companies to pursue a business opportunity together” (Schildt et al., 2005, page 497). The incentives for companies to engage in JVs can be multiple, ranging from stock-market gains to improved firms’ operations. By combining knowledge and business activities, firms can gain access to complementary resources with which to build bundles of valuable resources necessary to gain a sustainable competitive advantage (Barney, 1991).
However, JVs pose a significant challenge for managers, given their complexity and uncertainty. (Rothaermel and Deeds, 2006). Managers find them difficult to form and hard to manage. Consequently, it is not surprising that many alliances do not live up to performance expectations (Schilke and Goerzen, 2010; Kale et al., 2001). To stop this trend and make alliances work, organizations are now dedicating significant efforts to the design of governance mechanisms that mitigate the risks of opportunistic behaviour (Williamson, 1985) and support
10 the coordination of the optimal combination of productive resources across partners (Hoetker et al., 2009). Two widely acknowledged governance mechanisms are formal (contractual) governance mechanisms and relational governance mechanisms (Dyer and Singh, 1998).
Contractual governance is expressed in explicit, formal, and usually written contracts (Vandaele et al., 2007, p.240). These are very detailed and legally binding agreements that specify the roles and obligations of contracting parties (Lyons and Metha 1997; Wang et al., 2021). Relational governance refers to more emergent governance mechanisms governed by social relations and shared norms (Poppo et al., 2008).
Moving from the strategic importance of these configurations, researchers have already investigated how formal and relational governance mechanisms influence JV performance and to what extent they substitute or complement each other (e.g., Cao and Lumineau, 2015; Poppo and Zenger, 2002). Yet, evidence is still anecdotal on how to influence these configurations to achieve superior performance. Therefore, in an attempt to address the limitations of the literature, this paper will focus on formal governance mechanisms to investigate whether the establishment of a dedicated alliance function can influence alliance governance and, thus, the overall alliance performance. This analysis will particularly be focused on the potential effects of the dedicated alliance function on the composition of the alliance board and the details of the alliance contract.
But what specifically is a dedicated alliance function?
2.1 Dedicated alliance function
“The dedicated alliance function (DAF) is an organizational unit that is directly responsible for developing and disseminating alliance management know-how and overseeing alliances in the firm’s alliance portfolio” (Findikoglu and Lavie 2019, page 177). Companies with this function centralise certain aspects of alliance management in their organizational structure, aiming to organize the alliance around key strategic parameters and facilitate the
11 exchange of knowledge and information (Russo and Vurro, 2019). “A dedicated alliance function acts as a focal point for learning and for leveraging lessons and feedback from prior and ongoing alliances” (Dyer et al., 2001, page 38), potentially generating value for the firm by promoting standardization, formalization, and centralization of alliance management practices (Findikoglu and Lavie 2019). By establishing a dedicated alliance function, an organization can potentially improve its knowledge management, external visibility, internal coordination, and accountability; thereby achieving a higher alliance success rate from improved practices, higher abnormal stock-market gains, and a greater ability to form more alliances and attract better partners (Dyer et al., 2001). These benefits will be analysed further in the following sections.
2.1.1 Benefit 1: Improving knowledge management
Companies with dedicated alliance functions have codified explicit knowledge of alliance management by creating guidelines and manuals to help them manage specific aspects of their subsequent alliances, including partner selection, negotiation, governance, coordination, performance assessment, and termination of alliances (Findikoglu and Lavie 2019). By learning from their interaction with partners, firms with dedicated alliance functions develop a set of best practices over time that make alliance management easier and more efficient for them (Liu and Ravichandran, 2011; Russo and Vurro, 2019). This often results in a greater ability for these organizations to successfully manage multiple alliances simultaneously, making these companies extremely competitive (Dyer et al., 2001).
2.1.2 Benefit 2: Increasing external visibility
A dedicated alliance function can play a significant role in keeping the market informed about both the formation of new alliances and the successes of current alliances (Hoffman, 2005). This external visibility tool can improve a company’s reputation in the market and
12 strengthen the perception that partnerships are creating value (Dyer et al., 2001; Liu and Ravichandran, 2011). Forming a dedicating alliance function signals to the market and potential partners that the organization is committed both to its alliances and to managing them effectively. Furthermore, it offers a simple and highly visible point of contact for prospective partners wishing to connect with the organization to discuss an attractive partnership (Dyer et al., 2001). In essence, a dedicated alliance function provides a clear sign of increased partnership maturity in an organization and offers a place to screen potential partners and attract specific ones if the potential partnership looks profitable.
2.1.3 Benefit 3: Providing internal coordination
One of the main reasons alliances fail is that the goals of the alliance no longer match the strategic priorities of one or both partners (Bleeke and Ernst, 1993; Kogut, 1988). Companies need to have a mechanism for communicating which alliance initiatives are most important for achieving the overall strategy, and they need to be sure they are able to mobilize the right internal resources to support these initiatives (Liu and Ravichandran, 2011). The dedicated alliance function helps organizations achieve this, by acting as an information channel as well as by constantly ensuring the availability of resources (Dyer et al., 2001). A dedicated alliance function provides internal coordination for the organization's strategic priorities by ensuring that these issues are continuously addressed by partner companies (Kale and Singh, 2007;
Russo and Vurro, 2019). Furthermore, in case of a lack of resources, the dedicated alliance function leverages corporate ties to connect with divisions and business functions to request the necessary resources to support the company's alliance initiatives (Dyer et al., 2001).
2.1.4 Benefit 4: Facilitating intervention and accountability
Finally, the establishment of a dedicated alliance function compels an organization to develop alliance metrics and systematically evaluate the performance of its alliances (Dyer et
13 al., 2001). Ongoing evaluation of alliance performance is a critical aspect that is often underestimated by organizations that do not have a DAF (Mjoen and Tallman, 1997; Hoffman, 2005). These organizations often do not implement formal evaluation metrics, with the result that most are unaware of the alliance's true performance and therefore unprepared in case of problems. Conversely, constantly diagnosing alliance performance is beneficial for informing senior managers of the actual state of the alliance’s portfolio and having them intervene when there is a value outside of specifications (Dyer et al., 2001). Consequently, having a dedicated alliance function that constantly monitors the progress of the joint venture could be beneficial for companies to achieve their alliance goals (Kale and Singh, 2007).
In sum, the alliance literature suggests that companies can potentially reap substantial rewards by establishing a dedicated alliance function. However, it remains questionable how a dedicated alliance function manages to secure such benefits. It is the purpose of this thesis to resolve this controversy by statistically demonstrating whether a dedicated alliance function significantly contributes to the performance of the alliance portfolio, analysing its impact on alliance governance. For this reason, the topic of alliance governance will be analysed in more detail in the following paragraph.
2.2 Alliance Governance
To be successful in alliances, companies must coordinate the optimal combination of resources between partners, limiting the risks of opportunistic behaviour (Hoetker et al., 2009).
In this context, alliance governance plays a key role (e.g., Ryall and Sampson, 2009). Through governance mechanisms, companies can resolve issues of safeguard, cooperation, and coordination that arise in most alliances, by specifying the behaviour required by the partner or, more generally, by monitoring that the actions of the companies involved in the alliance are in line with what has been established. To achieve this goal, two common approaches are used:
formal governance mechanisms and/or relational governance mechanisms (Mellewgit et al.,
14 2007; Hoetker et al., 2009; Dyer and Singh, 1998). Each of these mechanisms enables the parties to solve alliance-related issues and to effectively manage the alliance (e.g., Poppo and Zenger, 2002). The former through contract drafting or board monitoring and the latter by building trust among partners. Despite the importance of both mechanisms, this thesis, recognising that the success of an alliance depends on the effective management of the relationship (Lee et al., 2006), focused on the formal ones, analysing ‘alliance contract’ and
‘board composition’ as the main elements of it (e.g., Ryall and Sampson, 2009). Consequently, the following sections, in addition to elaborating on the concept of formal governance mechanism, introduce how the establishment of the dedicated alliance function could influence alliance contracts and board composition, thereby increasing the likelihood of alliance success.
2.2.1 Alliance contracts
Formal organizational structures, such as alliance contracts, have become fundamental features of any collaboration. They serve to establish the rights and obligations of partner firms and provide protection against uncooperative behaviour (Ryall and Sampson, 2009). In many industries (i.e., insurance), firms face considerable moral hazard problems, since partner behaviour is often unobservable, and the costs of opportunism are potentially high (Crocker and Reynolds, 1993). Anticipating these difficulties and crafting formal governance to address these issues is essential. The assumption is that the threat of legal enforcement will keep partners from behaving opportunistically (Poppo and Zenger, 2002). Consequently, more detailed, and better-structured contracts are expected to lead to better outcomes and collaboration between partners, as they set clear expectations for behaviour or provide means to identify and curtail opportunistic behaviour (Ryall and Sampson, 2009; Luo, 2002). Clearly, however, building detailed contracts is difficult and requires a great deal of effort on the part of allied companies. Given the cost of drafting more detailed contracts (Crocker and Reynolds, 1993), the impact of the dedicated alliance function on alliance contracting could be significant.
15 The dedicated alliance function may serve to develop trust, leading to less need for formal governance, or may help organizations to draft more detailed and complete contracts.
2.2.2 Contract completeness
“Contractual completeness is a legally binding framework that codifies alliance partners’
rights, duties, and responsibilities as well as procedures and policies involving joint activities, which is generally reflected by the quantity, integrity, clarity, and adaptability of the contract terms” (Wang et al., 2021, page 993). A complete contract provides a framework for alliance cooperation, enables partners to conform to expectations and beliefs about the future, and contributes to shared goals (Wang et al., 2021; Luo, 2002; Luo, 2005). In addition, a complete contract safeguards each party’s interests and regulates each party’s behaviour, commitment, and responsibility. In other words, a complete contract constitutes one of the most effective mechanisms for inhibiting opportunism and contractual hazards in the context of joint ventures (Poppo and Zenger, 2002; Luo, 2002; Luo, 2005).
In the literature on alliances, it has been repeatedly demonstrated that organizations that were better able to draft complete contracts benefited from higher alliance performance (e.g., Ryall and Sampson, 2009). In line with this view, in 1987, Joskow demonstrated that contract completeness reduces transaction costs, contractual hazards, and operational risks, which in turn boosts business performance. Yet, the difficulty of codifying a complete venture contract and securing inter-partner cooperation is substantial. And this difficulty increases as the number of partners increases (Luo et al., 2007). The governance complexity caused by multiple parties further constrains the possibility of attaining a more complete contract for organizations that do not possess a dedicated alliance function for two reasons. The first is linked to the rising diversity of interests that comes with an increase in the number of partners. The uncertainty that each partner faces grows as the number of partners grows, as does the chance of incompatible objectives and the number of potential contingencies that must be addressed
16 (Elfenbein and Lerner, 2003; Williamson, 1985). As a result, contractual governance of partner relationships and firm operations becomes more complex, and attaining a complete contract becomes more difficult. Second, the more partners there are, the greater the uncertainty that the focal partner has about the other partners’ intentions and behaviors (Gibbons, 1992), making a complete contract harder to establish and implement (Luo et al., 2007).
Overall, writing a complete contract is often a difficult undertaking that requires time and commitment from both or multiple partners. However, as has been demonstrated by previous scholars, the benefits of drafting complete contracts are extremely significant (e.g., Luo, 2002;
Luo, 2005, Ryall and Sampson, 2009). Under contract completeness, the negative influence of multi-party-related complexity on JV performance is curtailed, information asymmetry is reduced, and cooperation is easier and more cost-effective (e.g., Wang et al., 2021; Luo, 2005).
Therefore, the importance that the dedicated alliance function could have for drafting more complete contracts could be considerable, especially when considering the resulting effect on alliance performance.
2.2.3 Board composition
The other component of successful alliance governance is the composition of the board.
“The board is a ‘seat of authority’ that enables partners to exert their rights through their representatives on the board and detect opportunism by monitoring the activities of the JV”
(Klijn et al., 2021, page 8). In the strategy literature, the board of directors has been identified as an important aspect determining organizational performance and strategic decisions (Baysinger and Hoskisson, 1990). A significant body of empirical research has shown that corporate boards possess a lot of useful business contacts and are, therefore, recognised as powerful vehicles for dealing with the external environment (Pfeffer, 1972). Boards serve as an ex-post governance mechanism that enables partners to monitor and control the collaboration’s performance, engage in private ordering by addressing eventual conflicts, and
17 align the actions of the venture (e.g., Geringer and Hebert, 1989; Kumar and Seth, 1998; Ravasi and Zattoni, 2006). Furthermore, boards offer counsel to management, help partners respond to unanticipated contingencies in a coordinated manner, and enable JV partners to align and coordinate their actions and interests (Klijn et al., 2021). Yet, like other governance mechanisms, boards also have their potential pitfalls. For instance, research has called into question the value and efficient functioning of large boards (e.g., Raheja, 2005; Linck et al., 2008). In particular, larger boards are expected to face more free riding issues and take longer to reach a consensus (Lipton and Lorsch, 1992). For this reason, some scholars have argued that the size of boards should not exceed eight members (Jensen, 1993; Lipton and Lorsch, 1992). An analogous claim has been raised regarding board involvement. Costs to partner companies are higher when boards engage in more monitoring and coordination than necessary.
Previous studies suggest that the value of board involvement depends on the complexity of a firm’s operations (Boone et al., 2007). With limited complexity, shareholders face less agency risk that must be mitigated by greater board involvement (Fama and Jensen, 1983). In light of these arguments, this thesis focused on an important aspect of board composition, ‘board size’.
2.2.4 Board size
A number of empirical papers have examined the advantages and disadvantages of different board sizes on alliance performance (e.g., Guest, 2009; Lipton and Lorsch, 1992;
Fama and Jensen, 1983; Klijn et al., 2013; Judge and Talaulicar, 2017). Evidence from these studies showed that there exists a negative relation between board size and alliance success, indicating that smaller boards are preferred over larger boards (e.g., Guest, 2009; Judge and Talaulicar, 2017). The majority of empirical research has documented that although larger boards are expected to provide a wider range of services, they have several disadvantages in the form of coordination costs and free rider problems, which lead to slower and less efficient decision-making (Jensen, 1993), greater difficulty in reaching consensus and cohesion (Lipton
18 and Lorsch 1992), and greater inefficiencies in communication (Yemarck, 1996; Krause and Bruton, 2014), which, in turn, lead to a lower level of alliance performance. These arguments suggest that smaller boards are necessary for a good board structure and that the number of directors should be limited to a certain amount. Jensen (1993) and Lipton and Lorsch (1992) argue that a board size should not exceed 8 or 9 directors to be optimally effective and that organizations should adapt this number according to firms’ characteristics.
In summary, understanding how to influence board size, while having a positive impact on performance, is essential. It is for this reason that, in recent years, researchers have been questioning the actual effectiveness of establishing a dedicated alliance function that could, theoretically, through the creation of a stronger relationship with partner companies, make boards more effective and, through complete contract drafting, make boards more efficient, thereby reducing the need for their involvement.
3. Hypotheses development
In this section the hypotheses are developed, and the discussed literature streams are combined. This leads to the development of a new theoretical field where it is investigated whether the DAF positively influences alliance governance and, thus, alliance performance.
3.1 Dedicated alliance function and performance
The theoretical background presented in the previous section suggests to us that a company can potentially increase its likelihood of alliance success through the establishment of a dedicated alliance function (e.g., Dyer et al., 2002). As we noted earlier, prior studies suggest that organizations that possess an alliance function can achieve a greater alliance performance through the multiple benefits this function delivers (e.g., Liu and Ravichandran, 2011; Russo and Vurro, 2019). According to these studies, one of the main benefits of the DAF is that it serves as a focal point for developing or disseminating the learning and accumulation
19 of alliance management know-how within a firm (Dyer et al., 2002). The process of learning and absorbing these management practices is very often a difficult challenge (Kale and Singh, 2007). Given the complexity of this activity, it appears that companies would be better served by having a single entity that is explicitly responsible for and capable of regularly implementing and managing this organization-wide effort. The alliance function, which is a dedicated organizational unit responsible for managing alliance-related initiatives and processes at the enterprise level, appears undoubtedly best positioned to lead and coordinate this critical task in organizations (e.g., Findikoglu and Lavie 2019; Kale and Singh, 2007). As a result, it appears that companies that possess a DAF may exhibit a stronger alliance learning process that allows them to enjoy a better overall alliance performance. Contrary, companies that lack such a function may have relatively poor or lesser developed processes to learn and accumulate alliance management know-how and thereby enjoy a relatively lower alliance success rate (Kale and Singh, 2007).
Based on these arguments, this thesis suggests that a dedicated alliance function plays an important role in explaining a firm’s overall alliance success. Consequently, this paper argues that firms with dedicated alliance functions will enjoy better alliance performance than firms without them. This implies that firms that have higher alliance success, and thus alliance capability, do so by having a stronger alliance learning process in place due to the presence of a dedicated alliance function. Therefore, a first hypothesis is presented as follows:
Hypothesis 1: The dedicated alliance function is positively related to overall alliance performance.
3.2 Dedicated alliance function and performance mediated by governance
Despite the potential benefits of the dedicated alliance function, the alliance literature has not yet managed to define how companies could achieve greater alliance performance through
20 its establishment and, for this reason, the DAF is still largely treated as a black box (e.g., Liu and Ravichandran, 2011; Russo and Vurro, 2019). In an attempt to open this black box, many researchers have tried to understand the dynamics through which DAF could positively influence the overall alliance performance (Kale and Singh, 2007; Dyer et al., 2002). The latest insights indicate that the dedicated alliance function could have a positive impact on alliance governance (e.g., Findikoglu and Lavie 2019). Alliance governance has been repeatedly identified in the alliance literature as the key to a successful alliance, given its role in minimising opportunistic behaviour and its contribution to optimising the combination of shared resources and capabilities (Mellewgit et al., 2007; Hoetker et al., 2009). For these reasons, this thesis will attempt to analyse whether DAF has a positive effect on board size and contract completeness, two important factors of formal alliance governance.
The availability of clearly defined techniques and actions to be performed during the various phases of the alliance life cycle might guide shaping the best formal governance configuration, especially in conditions of high uncertainty (Williamson, 1985). In this context, the DAF may represent the best way to gather experience and develop alliance management expertise, thus determining board composition and contract details according to the knowledge learnt from previous cooperative ventures (e.g., Kale and Singh, 2007; Lumineau et al., 2011).
As such, centralization of practices through the DAF may facilitate the development of standardized methods on how to efficiently perform alliance monitoring tasks, making boards more efficient and thus reducing the number of board of directors needed (Cheng et al., 2008).
At the same time, the DAF may improve the drafting of contracts, by collecting important information from both allied companies regarding their alliance histories and expected behavior, with a potentially positive impact on their completeness (Mayer and Argyres, 2004;
Vanneste and Punaram, 2010; Lumineau et al., 2011).
21 In sum, the dedicated alliance function may facilitate alliance-related learning processes through experience, potentially resulting in more effective management and control of current and future alliances, as well as improved board size and contract completeness (Russo et al., 2019). These arguments lead to the following second hypotheses:
Hypothesis 2a: The DAF influences alliance performance by making boards more efficient, thus reducing board size.
Hypothesis 2b: The DAF influences alliance performance by drafting more complete contracts.
3.3 Alliance governance and alliance performance
In the context of managing alliances, the combination of a small-size board, that enhances the ability to reach consensus on what to monitor and coordinate, with the presence of a complete contract, that minimizes the likelihood of opportunistic behavior, enforces formal governance, and strengthens the contribution of knowledge, capabilities, and collaboration, could be extremely beneficial for the alliance (Judge and Talaulicar, 2017; Wang et al., 2020;
Joskow, 1987; Luo et al., 2007). The literature on alliance governance offers an excellent perspective on this. Numerous researchers have found that, especially for high market overlap joint ventures, superior performance in alliances (almost) cannot be achieved without small to medium-sized board conditions, and while this factor is not sufficient on its own to make this happen, it is extremely necessary to achieve high alliance performance (Castro et al., 2016;
Klijn et al., 2021). Evidence from these studies has shown that, although larger boards initially facilitate key boards functions, larger boards suffer from substantial coordination, communication, and free-riding costs (Cheng et al., 2008; Hermalin and Weibasch, 2003;
Lipton and Lorsch, 1992). For this reason, several authors have suggested that boards of directors should not exceed seven or eight people (Lipton and Lorsch, 1992; Jensen, 1993).
22 When board size exceeds this number, the board becomes less likely to function effectively, making more difficult for board members to arrange meetings, express their ideas, and react quickly (Guest, 2009; Krause and Bruton, 2014). The board of directors should, therefore, be limited in number in order to ensure efficiency and effectiveness. Taking this negative board- size effect into account, this thesis focuses on the effect that board size has on the performance of alliances and, consequently, hypothesizes the following:
Hypothesis 3a: Board size is negatively related to overall alliance performance
The other critical component of alliance governance in JV activities is the contract between the firms. An extensive literature on contractual governance has emphasized the role of contractual completeness in JV performance. Several studies (e.g., Killing 1983, Luo 2002, 2005, 2007) have described the importance of clearly specifying and codifying in a venture's contract all the issues or clauses relevant to the construction and management of a joint venture, given their potential effect on alliance performance (Poppo and Zenger, 2002). According to Killing 1983, a contract is complete if each of these issues is included in a joint venture contract: (1) the responsibilities, rights, and benefits of each parent company; (2) solutions to disagreements between the parent companies; (3) rules for appointing personnel to the venture;
and (4) the goals of the parent companies and the joint venture (Killing, 1983). Failure to address these issues may make it impossible to draft a complete contract that dictates or delineates for each situation the responsibilities and appropriate actions of the companies involved and undermine the creation of a governing mechanism that mitigates each party’s ability to act opportunistically (Luo 2007). As a consequence, the likelihood of alliance success in the case of a complete contract, which lays out in detail the roles and responsibilities of each party, may be greater than in the case of alliances governed by a contract that leaves out important issues (Ryall and Sampson, 2009). In light of these arguments, this thesis proposes the following hypothesis.
23 Hypothesis 3b: More complete contracts are positively related to overall alliance performance
Overall, this thesis theorizes that the dedicated alliance function facilitates alliance- related learning processes fostered by experience, leading to an increased ability of the focal firm to form more complete contracts and making boards more efficient by reducing their size, thus resulting in higher alliance performance.
4.1 Sample and data collection
In the empirical investigation, the impact of the DAF on board size and contract details is measured. To examine this relationship, this paper used secondary data. The used dataset was completed in 2014 by a group of researchers at a Dutch business school and has been chosen because it fits perfectly with the research question of this thesis. It includes all variables of interest and has been shown to be of high quality in a previously conducted study (Reuer et al., 2014). The dataset was built by surveying influential managers of both domestic (Dutch) and international JVs. The sample population has been chosen using the non-probability sample technique of purposive sampling. To minimize sampling bias, researchers increased calibration, by targeting managers of Dutch JVs firms through the Thomson Financial Security Data Corporation (SDC) and managers of international JVs through an alumni database of a European Business School (Reuer et al., 2014). The selected executives had at least ten years of work experience and were either familiar with (5% of all respondents) or involved in (95%
of all respondents) joint venture governance.
To investigate and ensure the validity of the measurement instrument, the research team conducted pre-test interviews with both senior managers responsible for establishing or managing international joint ventures and alliance scholars, implementing adjustments based
24 on their feedback (Reuer et al., 2014). Out of over 600 surveys sent out, the survey collected over 200 respondents. The dataset was then analyzed for the purposes of this thesis and all variables used in the following regression models were controlled for missing data. After cleaning out responses that were answered completely to less than 75% of all questions, the remaining dataset consisted of 191 observations.
To check for normality and outliers, a skewness and kurtosis normality test was conducted. After this analysis, it was found that two variables ‘IJV size’ and ‘Board size’ were not normally distributed. Therefore, they were logarithmically transformed to ensure normality. To enable the use of a dedicated alliance function (‘DAF’) as an independent binary variable, this thesis adapted the scale of it from ‘0’ (no DAF), ‘1’ (just my firm has a DAF), and ‘2’ (both my firm and partner firms have DAF) to ‘0’ (neither my firm nor my partners have a DAF), and ‘1’ (either/both my firm or/and my partner has a DAF). In addition, a control variable ‘Alliance function partner’ was created to account for the influence of the partner’s DAF on alliance performance, coding it ‘0’ if the partner does not have a DAF, ‘1’ if the partner has a DAF.
Finally, in order to examine the quality of the data, a number of additional analyses were performed. The Harman’s-single-factor test (1976) and the Podsakoff test (2012) showed that the dataset created could not be influenced by common method bias and that the data had high external validity. Therefore, in light of these validity analyses, this thesis used the aforementioned dataset to conduct further analyses on the subject.
4.2 Variables and measurement
4.2.1 Dependent variable: Alliance performance
Previous literature on alliances has emphasized the need to understand how a company can increase the likelihood of alliance success (e.g., Bleeke and Ernst, 1993; Kogut, 1988; Kale
25 et al., 2002). ‘IJV performance’ is, therefore, the dependent variable (Y) of this analysis. IJV performance was constructed according to Mjoen and Tallman (1997) performance measure (i.e., mt_perf). Thus, three main indicators were used to establish the level of performance of an alliance: (1) we are satisfied with the performance of the JV, (2) the JV has met the objectives for which it was established, and (3) the JV has been a profitable investment.
Together, these three items create a well-structured approach to evaluating the health of any alliance (Mjoen and Tallman, 1997), and are therefore an excellent measure of alliance performance.
4.2.2 Independent variable: The dedicated alliance function
This paper theorizes that the performance of an alliance can be positively influenced by the establishment of a dedicated alliance function. The dedicated alliance function is, therefore, the main independent variable (X) of this analysis. To assess the impact of a firm’s alliance function, two variables were created using information from the dataset: an independent variable and a control variable. The independent variable was named ‘DAF’ and indicates the presence or absence of a dedicated alliance function by the focal company. The control variable was instead named 'Alliance function partner' and was constructed to control the impact of the presence of a dedicated alliance function by the partner company. Both are dummy variables and have been coded with ‘1’ if the firm had an alliance function and ‘0’ otherwise.
4.2.3 Mediating variable: Alliance governance
The idea on which this thesis is based is whether the dedicated alliance function influences overall alliance performance by positively influencing alliance governance. As a result, to analyze the relationship between the dedicated alliance function and alliance governance, this thesis has multiplied alliance governance into two main sub-characteristics (1) board composition (i.e., board size) and (2) contract details (i.e., contract completeness).
26 Board Size
The first sub-factor ‘Board size’ consists of the effect of different compositions of managers in the alliance board on the dependent variable ‘Alliance performance’. This thesis used board size as a mediator as the size of the JV board of directors can theoretically fill contract gaps, thereby decreasing the need for renegotiation and increasing alliance performance (e.g., Reuer 2014, Williamson, 1985). Board size is determined by the total number of board members. Therefore, respondents were asked to provide information on the total number of executive and non-executive directors that serve on the JV board.
The second sub-factor ‘Contract completeness’ consists of the effect of different alliance contract safeguards on the dependent variable ‘Alliance performance’. A complete contract provides a framework for alliance cooperation, enables partners to conform to expectations and beliefs about the future, and contributes to shared goals (Luo, 2002; Luo, 2005; Wang et al., 2021). For these reasons, contract completeness is often regarded as a factor that can highly enhance the performance of an alliance. Therefore, to measure this variable, eight types of contractual provisions were identified and, in line with Reuer et al. (2014) and Parkhe (1993), participants were asked to indicate which of the following eight types of contractual provisions appeared in the formal agreement of their JVs: (1) periodic written reports of all relevant transactions, (2) prompt written notice of any departures from the agreement, (3) the right to examine and audit all relevant records through a firm of CPAs, (4) designation of certain information as proprietary and subject to confidentiality provisions of the contract, (5) nonuse of proprietary information even after termination of agreement, (6) termination of agreement, (7) arbitration clauses, and (8) lawsuit provisions (Parkhe, 1993). In order to test the robustness of this scale, a counting measure for contractual safeguard was used. The contractual safeguards measure suggested by Parkhe (1993) was chosen, as the dimensions are ordered by
27 increasing strength or severity. In this way, the measure takes into account the stringency of the contractual provisions and provides an accurate scale. Therefore, the following formula was used:
𝐶𝑜𝑛𝑡𝑟𝑎𝑐𝑡𝑢𝑎𝑙 𝐶𝑜𝑚𝑝𝑙𝑒𝑡𝑒𝑛𝑒𝑠𝑠 = ∑ 𝐷𝑖
where Di equals ‘i’ if the ith contractual provision was employed, and 0 otherwise.
4.2.4 Control variables
Several controls were included in the analysis in order to account for other factors that can potentially affect alliance performance or alliance governance. At the joint venture level, this thesis statistically controlled for the nature of the IJV (Greenfield JV). Specifically, it has been distinguished between Greenfields and Acquisitive Joint Ventures, as these JV configurations require, by their nature, different compositions of boards of directors and details of contracts to establish various operating policies and routines and to create a controlled collaborative environment. Thus, a value of ‘1’ was given if the JV was created by acquisition of another company and a value of ‘2’ was given if the JV was created by establishing a new company.
Second, this thesis controlled for the number of partners involved in the IJV (Number of partners). Controlling for the number of partners in a JV is a common practice, since potential conflicts and coordination needs tend to increase as the number of partners increases (Reuer et al., 2014). In order to create this variable, it was asked how many other companies were involved in the joint venture and the given number was then reported in the dataset without the use of intervals.
Third, to assess the scale of JV operations, a further variable that has been controlled is the size of the IJV (IJV size). Therefore, respondents were asked to provide information on the total number of employees working in the venture, and this number was subsequently reported
28 in the dataset. Due to the significant positive skewness, this variable was then logarithmically transformed to ensure normality.
At the interrelationship level, on the other hand, this thesis controlled for the presence of a partner's dedicated alliance function (Alliance function partner). As previously mentioned, it is important to isolate the contribution of the dedicated function of the company from that of the partner in order to obtain a result with strong validity. Indeed, the partner function can facilitate communication, commitment, and coordination in the alliance with the focal firm, which can contribute to the gain of the firm in the alliance with the partner (Findikoglu et al.
The following tables summarize the results of this study. Firstly, Table 1 provides some descriptive statistics and a correlation matrix for all variables considered in the regression analyses. Secondly, Table 2 shows the results of the linear regression analysis for the determinants of IJV performance. Two different statistical models were developed and tested for significance. Model 1 is a baseline model that simply includes the control variables, whereas Model 2 examines all linear effects. Model 2 shows that there is a close relationship between the DAF and IJV performance variables with a relevant level of significance (β=1.001, p<0.05). It also shows that the composition of the board of directors, particularly its size, in contrast to the contractual governance mechanism, has a significant influence on the performance of the alliance (β= -0.533, p<0.05). Thirdly, Table 3 summarizes the results of the interaction between the DAF and the two governance mechanisms. The results of this table suggest that while the DAF influences board size (β= 0.305, p-value<0.05), it does not have a significant effect on contractual completeness (β= 1.008, p-value>0.05). Further analyses were conducted to confirm these results. Table 4 presents an additional analysis to that carried out in Table 3 that tests the relationship between the independent variable ‘DAF’ and the mediator
29 variable ‘alliance governance’. This table confirms the results found in Table 3, confirming their robustness. Finally, Table 5 summarizes the regression models conducted in this thesis using Baron and Kenny's mediation table (Baron and Kenny, 1986).
In Table 1, it can be observed that the variable DAF, in line with previous literature (e.g., Russo and Vurro, 2019), shows a positive and significant correlation with the dependent variable IJV performance. Furthermore, the table shows that while contract completeness is positively slightly correlated with alliance performance, board size is negatively slightly correlated, which justifies further statistical tests. The result of the negative correlation between board size and alliance performance confirms the results already found in the article by Klijn et al. (2021), whose authors had already identified that, especially for joint ventures with high market overlap, the smaller the board, the higher the alliance performance. Another noteworthy result is the positive and significant correlation between DAF and board size. This relationship might suggest that in the presence of a DAF, board size might tend to increase. As one of the fundamental relationships of this thesis, this relationship will be further studied in a subsequent regression. Also worthy of attention is the positive correlation between DAF and contract completeness. This correlation might suggest that in the presence of a DAF, the completeness of the contracts might be greater. This will be analyzed in a later regression to determine whether the presence of a DAF increases alliance performance by creating more complete contracts. Finally, Table 1 shows a positive and significant correlation between JV size and board size. This finding is consistent with research demonstrating that larger JVs may have greater operational complexity (Reuer et al., 2014), which in turn may increase the need for more formal board monitoring (Boone et al., 2007).
Certainly, these initial indications of potential relationships need further analysis, which can be carried out by means of regression analyses. Therefore, the results of the regression analyses are provided below.
31 To test the first and third hypotheses of this study, a regression analysis was conducted.
This regression examined the linear relationship between the independent variable DAF and the dependent variable alliance performance. Furthermore, it also tested the relationship between the mediator variable alliance governance and the dependent variable alliance performance. The results of this model are exhibited above in Table 2. As it can be deduced from the table, the regression analysis (see Model 2) suggests that there exists a direct positive relationship between the independent variable DAF and the dependent variable alliance performance (β=1.001, p-value<0.05), therefore supporting the first hypothesis. Moreover, Model 2 shows that, while board size was found to have a negative effect on alliance performance (β= -0.533, p-value<0.05), thus confirming Hypothesis 3a, contract completeness did not obtain a significant p-value to confirm Hypothesis 3b (p-value>0.05). These results indicate that if the dedicated alliance function were to have a significant positive effect on
32 alliance governance, it would increase the likelihood of alliance success through a positive impact on board composition.
Consequently, to test the hypothesis that the dedicated alliance function positively influences the two governance mechanisms, a seemingly unrelated regression was performed.
As can be seen in Table 3, it has been found that the dedicated alliance function influences just one component of alliance governance, board size. However, contrary to Hypothesis 2a, the DAF appears to increase board size, rather than decrease it, thus making it impossible to prove that the DAF influences alliance performance through a reduction in board size (see Model 4a, β= 0.305, p-value<0.05). Furthermore, Table 3 shows that the DAF has no significant effect on contract completeness (see Model 4b, p-value>0.05) and that, therefore, the impact of the DAF on the success of alliances cannot be explained by this factor either.
In addition, a particular test to analyze the extent to which the two governance mechanisms are independent, the Breusch-Pagan-Godfrey test (Breusch & Pagan, 1979), was performed (see Table 3.2). The Breusch-Pagan independence test (Breusch & Pagan, 1979) revealed a small chi-square value (χ2=0.009, see Table 3.2) and showed that the two governance
33 mechanisms are not significantly independent from each other (p=0.926, see Table 3.2).
Consequently, due to the insignificance of the result, it was not possible to determine whether board size and contract completeness are substitutable or complementary.
Having identified the outcomes of the relationships between the dedicated alliance function (X), alliance governance (M), and alliance performance (Y), it is now possible to visually summarize the results of this thesis, using its conceptual model (see Figure 1).
Thesis conceptual model – Final results
H2a: β= 0.352**, s.e.= 0.131 H3a: β= -0.533*, s.e.= 0.237
H2b: β= 1.866, s.e.= 2.414 H3b: β= 0.008, s.e.= 0.112
H1: β= 1.001*, s.e.= 0.415
To confirm that these results are all correct, a supplemental analysis to that proposed in Table 3 was conducted to verify in particular the impact of the DAF on the two governance mechanisms (see Table 4). This analysis showed that the DAF has a significant positive effect on board size (see Model 5a, β=0.315, p-value<0.05) and that it has not a significant effect on
Board composition Contract details
Function Alliance Performance
34 the completeness of the contract (see Model 6, p-value>0.05). Consequently, these results suggest that both Hypothesis 2a and Hypothesis 2b cannot be supported.
To ensure that the result of the first relationship (DAF and board size) is correct, a negative binomial regression was also performed (see Model 5b). The negative binomial model was considered the most appropriate for an additional analysis, as the dependent variable is over-dispersed and does not have an excessive number of zeros. The negative binomial regression confirmed the result obtained from the previous regression (see Model 5b and Model 4a), thus finally rejecting Hypothesis 2a (see Model 4a, β=0.305, p-value<0.05 and Model 5b, β=0.189, p-value<0.01).
In light of these results, it can be concluded that the DAF has a positive impact on the performance of an alliance, but this positive effect cannot be explained by the reduction in board size or the impact the DAF has on contractual completeness. Even though, in line with
35 other authors (e.g., Klijn et al., 2021; Reuer et al., 2014), this thesis found that smaller boards have a positive impact on alliance performance, thus supporting Hypothesis 3a, it has not been found that the DAF has a reducing effect on the number of board members. As a result, the hypothesis concerning the negative effect of the dedicated alliance function on board size could not be supported.
Moreover, this thesis also did not support the hypotheses concerning the other governance mechanism, contract completeness. As opposed to the results of previous authors (e.g., Luo et al., 2007; Luo et al., 2002), the analyses did not find a significant positive relationship between contract completeness and alliance performance. In addition, the establishment of a dedicated alliance function was not found to be significantly relevant to the drafting of more complete contracts.
Table 5 summarizes the regression models conducted using Baron and Kenny's mediation table (Baron and Kenny, 1986). Therefore, for each hypothesis, the reference model and the results obtained by each were reported. This makes the results of the mediation between dedicated alliance function, alliance governance, and alliance performance clearer and more visible.
36 6. Discussion
6.1 Contribution and implications
Why are some companies more successful in forming alliances than others? In this study, we provided a theoretical account that potentially explains firms’ differences in alliance performance. Drawing on the literature on alliances (e.g., Killing 1983; Caloghirou et al., 2003;
Mellewigt et al., 2007), this paper has demonstrated that the dedicated alliance function contributes significantly to enhancing alliance success, improving the likelihood of creating a fruitful alliance. By uncovering this relationship, this thesis has shed light on the controversies that have arisen over the past decades concerning the role of the dedicated alliance function, resolving the mixed and inconsistent findings obtained to date (e.g., Heimeriks et al., 2007;
Dyer et al., 2001; Findikoglu et al. 2019). Previous research did not find a statistically significant contribution of the DAF to the performance of the alliance portfolio, maintaining the division between those who believed that dedicated functions created value and those who did not (e.g., Russo and Vurro, 2019; Heimeriks and Duysters, 2007). Until today, various authors suggested that the DAF had no impact on alliance performance and that, on the contrary, having one would have led to significant disadvantages. Proponents of this view believed that the DAF limited flexibility and did not improve alliance success (e.g., Heimeriks et al., 2007, 2015). According to these authors, the formalization and institutionalization of practices obtained by means of the DAF would have led to ‘superstitious learning’ (Hoang and Rothaermel, 2005; Zollo, 2009), and the establishment of codified and consistent procedures would have inappropriately led to emphasize process over speed in decision making (Dyer et al., 2001), therefore negatively impacting alliance performance. The results of the empirical analysis findings of this thesis confirm, however, that the dedicated alliance function has a significant positive impact on alliance performance, implying that firms that establish a dedicated alliance function can potentially have greater alliance success. This conclusion puts
37 an end to the vague and mixed evidence on the potential value of the DAF that has characterised the literature of alliances to date, clarifying this dispute once and for all (e.g., Heimeriks et al., 2015; Dyer et al., 2001; Russo and Vurro, 2019).
But how would DAF increase the performance of an alliance? By using a firm’s alliance performance as the dependent variable and the dedicated alliance function as the independent variable, this thesis showed that when a DAF is present, the overall alliance performance tends to increase. This can be explained by the numerous benefits that the dedicated alliance function would bring to the JV. Prior research found that “by establishing a dedicated alliance function, a firm can learn practices for selecting partners, negotiating agreements, and managing alliances, which in turn create value in the firm’s new alliances” (Findikoglu et al. 2019, page 177). Thus, in line with this thinking, by owning a dedicated function, an organization can potentially improve its knowledge management, external visibility, internal coordination, and accountability; thereby achieving a higher alliance success rate from improved practices, higher abnormal stock-market gains, and a greater ability to form more alliances and attract better partners (Dyer et al., 2001). Aiming to extend and explain this set of benefits already identified in the previous literature, this thesis focused on the impact the DAF has on alliance governance, a factor not yet considered in the existing literature. However, contrary to expectations, the results of this study revealed that the positive relationship between the dedicated alliance function and alliance performance cannot be explained by the effect that the DAF has on the development of smaller board sizes or the drafting of more complete contracts.
In line with earlier studies (e.g., Guest, 2009; Judge and Talaulicar, 2017; Reuer, 2014), this paper found that board size is an important antecedent for alliance success and that larger boards are expected to jeopardize alliance performance (Cheng et al., 2008; Hermalin and Weibasch, 2003; Lipton and Lorsch, 1992). However, analyses have shown that the DAF, instead of increasing alliance performance by reducing the number of boards of directors, can