Corporate Hub as a Governance Structure for Coupled Open Innovation in Large Firms

This paper addresses how the open innovation concept can be applied in large corporations aiming to work with small innovative firms. More specifically, we report an in&#8208;depth qualitative study of how AstraZeneca, a global bio&#8208;pharmaceutical corporation, designed and enacted a new mode of organizing for coupled open innovation, with an initiative called the BioVentureHub. The governance structure behind this way of organizing is explored and conceptualized, and we also analyse how this structure shaped the relations and the value being collaboratively generated and shared. Theoretically, the paper connects the governance literature with open innovation. From the analysis of the empirical case, we propose a distinct governance structure that we label as a corporate hub, and we distinguish it from other &#8220;open&#8221; governance forms, e.g. market, network and bazaar governance, and explain the potential value it brings for coupled open innovation between large and small firms.

However, the field of research is still relatively new and in need of further inquiries (Bogers et al., 2018;Chesbrough, 2017), especially since the concept can be used in many different ways (Dahlander & Gann, 2010). Open innovation has, for instance, been linked to industry-university collaboration (Perkmann & Walsh, 2007), communities (Dahlander, Frederiksen, & Rullani, 2011), crowdsourcing (Poetz & Schreier, 2012), user-centric innovation (von Hippel, 2005), platforms (Bullinger, Rass, Adamczyk, Moeslein, & Sohn, 2012) and intellectual property markets (Arora & Gambardella, 2010), all of which arguably work under specific conditions, applying distinctive organizing mechanisms and emphasizing different modes of governance (cf. Demil & Lecocq, 2006;Nambisan & Sawhney, 2011). The inclusiveness of the concept is allowing actors to adopt it in their own ways, but it also poses challenges when navigating the motley landscape precisely because different meanings are being ascribed to the same notion. Hsieh and Tidd (2012) state that the theorizing of open innovation tends to be general rather than specific to particular contexts and contingencies, but naturally it is very different to design an open business model for a small app developer than for a large biopharmaceutical firm. Moreover, there may be significant differences and challenges in enacting open innovation initiatives based on controlled transactions similar to "business-as-usual", such as buying or selling intellectual properties (IP) on the market, compared to facilitating ongoing, serendipitous interactions where the results and the value are unknown a priori. Often, the greatest value potential for large corporations is through what Gassmann and Enkel (2004)  In the paper, we link the governance structure to the consequences for the open innovation initiative, in terms of who are invited to participate, and how are they being coordinated and controlled. We do so by empirically reporting a study from Sweden on how AstraZeneca, a global bio-pharmaceutical corporation, enacted an open innovation initiative called AZ BioVentureHub (BVH). It was based on the idea of creating a "hub" inside the corporate walls, where small external innovative firms are invited to locate their development work, with the possibility to interact also with the large firm.
The study's unit of analysis is AZ BioVentureHub, which is a corporate initiative and thus mainly rests on an organizational level of analysis.
However, as we also include the participation of small hub companies in the discussion, the study also to some degree engages in an interorganizational level of analysis. We will in the paper explain and analyse AstraZeneca's approach, in order to identify the specific governance structure that characterized this specific design. BVH, we argue, moves away from the more traditional view that firms are best governed through either the hierarchy of the firm or the price mechanisms of the market, depending on the context (Williamson, 1998).
We propose that an attempt such as the BVH is better described as a corporate hub, and we analytically distinguish this governance structure in relation to markets vs. hierarchy (Coase, 1937;Williamson, 1975;Williamson, 1985;Williamson, 1998), but also against network (Powell, 1990) and bazaar (Demil & Lecocq, 2006) governance.
This study directly links open innovation research to the governance discussion within the broader management and economics fields, and specifically directs attention to the governance of interorganizational dynamic relationships, which has been pointed out as a central question for open and distributed innovation research Tiwana, Konsynski, & Bush, 2010). The governance structure of a corporate hub is grounded in the resources of a large corporation, but where informal and relational ties are allowed to evolve among carefully selected participants. The BVH is an interesting example of how large corporations within a local environment can maintain their internal control over key assets and IP while simultaneously sharing and utilizing certain resource and R&D spillovers emerging over time. The empirical and theoretical conclusions advance the knowledge about how the governance structure of coupled open innovation between large and small firms can be designed and implemented, and the study sheds light on how such governance structure shapes both the content and the form of open innovation in practice.
The reminder of the paper is organized as follows. First, we present an overview of the literature on open innovation in relation to collaboration between large and small firms and linking it to the governance literature. After a method section, the case of the BVH is introduced and analysed from a governance perspective, ending up in concluding remarks.

| OPEN INNOVATION BETWEEN LARGE AND SMALL FIRMS
Open innovation is defined by Chesbrough and Bogers (2014) as "a distributed innovation process based on purposively managed knowledge across organizational boundaries, using pecuniary and nonpecuniary mechanisms in line with each organization's business model." In recent years, various definitions and conceptualizations of open innovation have been proposed (e.g., Bogers et al., 2017;Bogers & West, 2012;Dahlander & Gann, 2010;Gassmann & Enkel, 2004;Huizingh, 2011;Remneland Wikhamn, 2013), which have furthered our understanding of open innovation, and advanced open innovation as an academic field as well as in managerial practice. For instance, based on an analysis of 43 cross-sector firms, Mortara and Minshall (2011) argue that organizations tend to adopt open innovation differently depending on why and when the implementation occurs, and what culture exists in the focal firm. They suggest that firms may need innovation just to support their current innovation pipelines, but they may sometimes also have a need for organizational transformation and enhanced ambidexterity. If the motive is the latter, Mortara and Minshall (2011) propose that firms need to pursue both inbound and outbound activities and continuously interact with external actors to challenge and renew the internal organization. Put differently, apart from advancing the innovative capabilities, open innovation has also been argued to enhance the firm's dynamic capabilities and absorptive capacity (Remneland Wikhamn & Styhre, 2017;Teece, Peteraf, & Leih, 2016;Zobel, 2017).
Engaging with small firms, such as start-ups or scale-ups, is a specific open innovation approach to enhance both innovation and business transformation simultaneously in large corporations (Richter, Jackson, & Schildhauer, 2018;Weiblen & Chesbrough, 2015). Synergies can emerge when large and small firms interact, where the small firms can gain access to knowledge, resources and legitimacy, while the large firm can become inspired by the small firms' agility and entrepreneurial spirit (Weiblen & Chesbrough, 2015). Examples of open innovation between large and small firms involve, for instance, innovation contests (Bullinger, Neyer, Rass, & Moeslein, 2010), corporate accelerators (Kohler, 2016;Richter et al., 2018) and the engagement in broader innovation ecosystems (Rohrbeck, Hölzle, & Gemünden, 2009). Also, the BVH initiative explored in this paper can be placed under this open innovation category. Following the argumentation of Mortara and Minshall (2011), all these different forms of initiatives can differ in motivations, implementations and consequences for inter-organizational innovation work, thus warranting a closer examination of the designs and effects of governance structure.

| OPEN INNOVATION AND GOVERNANCE
In his seminal text The nature of the firm, Coase (1937) laid the grounds for the transaction cost theory, which was later popularized by Williamson (1975Williamson ( , 1985 and has grown to a substantial area of research on rationalizing strategic firm behaviour. The theory explicates when and why firms decide to make business operations inhouse or to trade them on the market-i.e. the classical question of "make vs. buy" (Walker & Weber, 1984). This question is naturally of greatest centrality for scholars of open innovation, as it involves the demarcation of firm boundaries.
The theorizing of governance is a central theme in transaction cost theory (Coase, 1937;Williamson, 1975), where a governance structure develops and maintains "a good order and workable arrangements" (Williamson, 2005) when different actors need to coordinate their work efforts. A core argument for why governance matters lies in that interactions among humans involve uncertainty of the outcomes due to, e.g., bounded rationality and opportunism (Williamson, 1998), which is why such interactions also bear transaction costs.
Such costs can, for instance, involve the costs of searching and evaluating information, bargaining and acquiring it, and controlling and policing external counterparts (Demil & Lecocq, 2006).
In short, transaction cost theory suggests that firms should organize business operations internally, governed by hierarchical orders and management supervision, unless the transaction costs of utilizing the market are lower than the gains. When innovation occurs within the hierarchy of the firm, the value that is created is also captured by the same entity, and the corporation facilitates the innovation work through conventional human resource management. When extending the boundaries of the firm, large corporations have traditionally primarily solved the governance issue through the market function, However, many forms of transaction costs have arguably been reduced and altered in contemporary business practice, through the introduction of new ways for communication, knowledge diffusion and project collaboration across traditional organizational boundaries (Benkler, 2006). Within the academic field of open innovation, attention has thus also been increasingly drawn to open as libre (Remneland Wikhamn, 2013), which is approaching knowledge more as relational and ongoing. The driver for such value creation is generativity (Remneland Wikhamn, Ljungberg, Bergquist, & Kuschel, 2011;Zittrain, 2008), in which voluntaristic and spontaneous ongoing interactions create positive feedback loops to leverage knowledge generation.
Informal ties like these open up possibilities for serendipity (Murayama, Nirei, & Shimizu, 2015) and radical innovation, and scholars have suggested that the firm's innovation performance is enhanced when more relational approaches are utilized (Felin & Zenger, 2014;Fey & Birkinshaw, 2005). Conversely, such open innovation activities may need new modes of governance, different from both hierarchy and market mechanisms.
When analysing governance structures, several parameters have been proposed as being of importance; for example, the forms of selection in which external actors are allowed to participate; the forms of interactions that the initiative is designed for; the control mechanisms established to avoid opportunism and unethical behaviour, and the forms of value generated and shared among the participants (c.f. Demil & Lecocq, 2006;Felin & Zenger, 2014;Pisano & Verganti, 2008;Powell, 1990;Williamson, 1975, Williamson, 1985, Williamson, 2005. In Table 1, these parameters are used for distinguishing hierarchy vs. market, based on the above discussion, but we have also added two other governance structures that are related to open innovation-network and bazaar governance. Powell (1990) explains networks as "more dependent on relationships, mutual interests, and reputation [than in markets]-as well as less guided by a formal structure of authority [than in hierarchy]." The relations developed in a network are indefinite, sequential, long-term and complex, where sanctions are normative rather than legal. The different actors' complementarity and reciprocity help them agree to abandon self-maximization and opportunism in favour of network maximization. In short, networks are non-market, and non-hierarchical modes of exchange that represent a distinct governance structure of collective action.
In similar vein, Demil and Lecocq (2006) posit the emergence of bazaar governance as a separate governance structure, drawing on the open source phenomenon in software development. Bazaar governance differs from network governance in that no long-term relations are required, and that it is more open to access and enrolment due to its anonymity, absence of selection processes and no requisite for long-term relations. Furthermore, it utilizes a legal contract-the open license-that prevents users from appropriating the end-result for their own profit maximization, but also facilitates strong network externalities and generative knowledge transactions.
The BVH initiative that will be analysed in this paper is governed neither through the hierarchy of the firm, nor through market mechanisms. It shows more resemblance with networks of learning (Powell, 1998;Powell, Koput, & Smith-Doerr, 1996) and bazaar governance (Demil & Lecocq, 2006), as it shares the relational view (Dyer & Singh, 1998) rather than the transactional view (Williamson, 1998). We will, however, argue that the BVH characterizes a distinct governance structure that we explain in terms of a corporate hub structure-a governance structure that will lead to specific consequences in relation to whom is invited to participate, how the interactions are coordinated, and how they are being controlled.

| Life sciences in Sweden
Sweden has a long tradition of developing innovations in the life sciences, with world-leading university and industry expertise acting as contributors of medicines (e.g., Bricanyl, Losec/Prilosec) and medical technologies (e.g., pacemaker, respirator). In 2012, the life sciences industry in Sweden encompassed 40,764 employees in 1,487 companies (Sandström, 2014). Among these firms, 791 were active in research and development, product development, consulting or manufacturing (Sandström, 2014). Until the late 1990s, Sweden had two multinational pharmaceutical firms in Astra and Pharmacia. Both
The study of the BVH was part of a research programme that explored innovative forms of organizing in the Swedish life sciences industry, including biotech, pharmaceutical and medtech areas. The

BVH emerged as an interesting attempt from a large corporation to
create an open business model in practice, acting within an institutional setting based on closeness, regulatory control and intellectual property orientation. While similar initiatives have been launched by universities and various public or regional actors, such as incubators and science parks, the case of the BVH differs in that it is facilitated by a for-profit Big Pharma corporation, with unique skills, experiences and resources for disposal. For sure, this is not the first time that big corporations have launched incubators, science parks, accelerators and open innovation initiatives, but the BVH is a forerunner in its specific design and content within the bio-pharmaceutical industry.
Although a single case study is by definition unique and bounded to its empirical specifics, it is a well-established method for generating  Williamson (1975Williamson ( , 1985 Powell (1990) Demil and Lecocq (2006) theoretical insights in early stages of theory building (Eisenhardt & Graebner, 2007) and to provide analytical generalizations in transforming empirical data to theory, rather than to population (Flyvbjerg, 2006;Yin, 1994). That being said, the BVH has gained much attention among policy-makers as well as from other companies in both the life sciences industry and in other industries.

| Data collection
In total for this study, 74 interviews were conducted with 54 individ-

| Data analysis
The interviews were either summarized, including quotes that were perceived as interesting, or fully transcribed (Poland, 2001). The transcribed texts were continuously coded in NVivo by one of the two senior researchers conducting the interviews, following the recommendations of Clarke (2005), Corbin and Strauss (2014) and Miles and Huberman (1984) Table 2 and visually presented in Figure 1 in the empirical section of the paper. Host decides who is allowed to participate Forms of control mechanisms "We do not have any own interests invested in these hub companies. We do not own them or have legal contracts to have the right to acquire them later. Therefore, we can speak from our hearts when we guide them, and they can trust that we do not say things just because we want to sell them something" (COO, BVH, 25 May 2015). "There is a sort of social contract. There is an understanding that even if something could be seen as confidential, in this hub we trust each other. Contracts and legal documents have their purpose, but it is not that we will sue each other. What is important is that we have shared-you and me as human beings" (COO, BVH, 1 April 2016). "I think that it is very brave of AstraZeneca to create this hub. Firstly, that they give us facilities in the heart of the R&D centre and not somewhere in the outskirts, and secondly that they give us the same access as anyone else in AstraZeneca. I can go around anywhere I want, and that shows a lot of trust. The . "This is the beauty of sitting here. Everyone has the same keycards. When we eat lunch together, we just say; 'Okay, which project should we discuss this time?' So it is really good for building relations with AstraZeneca, but we also discover the other hub companies and start to interact with them" (Head of science, hub firm, 28 May 2015).

| THE AZ BIOVENTUREHUB
Ongoing, informal and facilitated through formal process Forms of value "We have formed a specific collaboration project with one of the hub firms. We have realized that we learn a lot from them, and at the same time, we do stuff for them that they are not able to do. So one plus one equals three. One of our guys works about 60% for them, with their stuff, but with a technology that we really would like to learn about. This is stone age economy, in a sense, but it works" (CEO, BVH, 28 October 2015). "Just as an example, I just came from a lunch where I have talked with two guys from AstraZeneca who are experts in Phase I studies, and we discussed various important challenges that we face at the moment. This lunch costed us 250 SEK, and it gave us more knowledge than we could ever buy from an expensive consultant" (CEO, hub firm, 10 October 2014).
Value is based on reciprocal and ongoing exchange for generating synergies F I G U R E 1 The governance structure of BVH and its consequences for value creation formalize and lock relations in contracts. An AstraZeneca manager described such partnerships in the following terms: It becomes much more formal. You do not as easily enter collaborations just because you want to explore interesting things without knowing where it will lead.
You have to negotiate everything from A to Z.

| The governance structure of AZ BioVentureHub
The BVH governance mode is built on a formal structure of the firm and the protection of IP, but it also constitutes the facilitation of infor-  Figure 1 summarizes the governance structure of BVH, and each area will be discussed in greater detail below.

| Form of selection
In the beginning, the mobilization process of hub companies was based mainly on personal connections and accidental meetings, in a search of a mutual companionship. The companies needed to fit into the general profiles set up by AstraZeneca, but they also themselves needed to be convinced that this was the right move for them. Hence, a lot of interviews and discussions took place. In this initial phase, AstraZeneca located three companies that agreed to be somewhat test-pilots. This was even before the hub was officially announced and before the lab facilities and offices had been renovated. As the hub received more attention in the press and became known to the public, requests to join from small and mid-sized enterprises (SMEs) increased. Over time, new companies entered one by one. In February 2019, more than five years after its launch, the BVH hosted 29 companies and one academic group and the plan was to continue the growth.
Already from the beginning, some unofficial criteria were devel-

| Form of control
The basis for participation in the hub is a standard rental agreement, defining the rights and obligations of the small firms. From this contract, the tenants gain access not only to their offices and research labs, but also to the vast knowledge that AstraZeneca has incorpo- Early on, a second manager was hired to the hub, a Chief Operating Officer (COO), whose task was more focused on operational issues, for example to develop processes and infrastructure as well as to act as a bridge and gatekeeper to the internal experts of AstraZeneca. In general, the interactions between AstraZeneca and the hub companies are informal and based on trust and a shared goodwill, but when in need of more advanced advice from experts, a rather structured and facilitated process, coordinated by the COO, has been established in order to not disclose any secret information by either side, and to clearly document all discussions. For these meetings, AstraZeneca receives a consultancy fee to cover the direct salary costs involved with the SME spending time with the AZ expert, but without profit margin (as is the same with the SMEs' utilization of AstraZeneca's equipment and resources). Hence, the model is designed so that AstraZeneca does not have its own direct profit incentive to contribute to the hub companies in these initial

| Form of interactions
All the hub companies' employees receive keycards to the whole

| The value generated from the governance structure
BVH has a business rationale from AstraZeneca's viewpoint, or as the CEO puts it: "This is not philanthropy, it is just an alternative use of resources" (CEO, BVH, 12 December 2014). By carving out the specifically tailored governance structure described above, the BVH can act as a lever for many of the challenges that the large corporation faces, while at the same time offering extremely positive benefits for the involved SMEs within the environment.
The equipment utilized in drug development is rather cyclic, with heavy peaks and periods of downtime, and, from this perspective, BVH can be seen as a proactive way to avoid downsizing. However, while the corporation needs to cut cost, it also needs also to attract competence, scout for new drug candidates, build a scientific reputation, rejuvenate the brand and the corporate culture, and to form new productive partnerships. With BVH, AstraZeneca has gained insights in new scientific areas and practices, as well as being inspired by the hub firms' entrepreneurial cultures. Furthermore, the BVH has gained a lot of media attention that positively influenced the corporate brand, and received interest and visits from policy-makers and industry representatives. For sure, it is difficult to estimate the total value from this initiative in quantifiable measures, especially as the drug development process is lengthy and complex in nature. But at least the nonpecuniary value streams have been highly appreciated by AstraZeneca and all the stakeholders connected to the hub.
The benefits generated to the small firms in BVH were plenty.
SMEs that earn the entrance ticket to the BVH gain access to the large smorgasbord of knowledge, experiences and resources of AstraZeneca without losing their strategic independence and ownership of intellectual property. They gain access to expertise in regulatory issues, which can be extremely labyrinthine and complex to learn. They also gain access to expertise, processes and expensive equipment for clinical research work, which can save them a substantial amount of money if done in a clever manner. The collaboration with the BVH, moreover, gives them an implicit operative license that can attract and convince future investors. In other words, the SMEs in the hub can utilize the knowledge spillovers distributed among the hub companies as well as world-class expertise from AstraZeneca, for example in regulatory issues and advanced clinical research work.
Unlike clinical research organizations and external consultants, the advice is not as much coloured by the willingness to sell specific ser-   (Agrawal & Cockburn, 2003;Feldman, 2003), suggesting that large corporations, such as AstraZeneca, act as anchors within local environments, to attract skilled labour pools and provide knowledge spillovers that benefit smaller firms within the area. Geographical proximity creates benefits in terms of establishing trust and collaborations between parties (Maskell, 2001) and it has been argued that "tacit" knowledge (e.g., related to how to develop drugs) is favourably transferred through informal communication (Saxenian & Hsu, 2001).

| CORPORATE HUB AS A SPECIFIC GOVERNANCE STRUCTURE
The so-called "Valley of Death" (e.g., Butler, 2008) has been metaphorically illustrating the challenges that small firms face in the transition from research to development, when they encounter huge problems in accessing new capital and new competences. These firms are generally vulnerable to what has been coined as the "liability of newness" (Stinchcombe, 1965) and/or the "liability of smallness" (Freeman, Carroll, & Hannan, 1983), which suggests that smaller firms face challenges for survival and growth just because of legitimacy and resource issues related to their size and age. With the BVH, AstraZeneca infuses their own competences and assets earlier in the value chain to aid the SMEs in progressing across the otherwise deadly valley. The involvement with AstraZeneca and the other hub companies also reduces the SMEs "liability of unconnectedness" (Powell et al., 1996), which is critical for staying tuned in to the knowledge race of the rapidly developing life sciences industry, and continuously opens up possibilities for further collaborations. In a sense, the Valley of Death is thus being replaced by the friendlier Valley of AstraZeneca.
We argue that a corporate hub is a governance structure in its own right, one that differs from the traditional modes of hierarchy, market, network or bazaar (Demil & Lecocq, 2006;Powell, 1990)  in line with a long-term relational and trust-based view (Dyer & Singh, 1998). Furthermore, it differs from network organizing (Powell, 1990) in that it is somewhat firm-centric and with a carefully selected and limited number of actors which are geographically co-located on the same site. Contracts are important as a basis, for the invitation as well as for potential later outcomes in the form of rather traditionally organized partnerships. Hence, it clearly differs from the bazaar governance (Demil & Lecocq, 2006) in that intellectual property and pecuniary motives are still at the centre of attention for many of the involved stakeholders. Freeriding is not as much in evidence as in bazaar governance, characterized by high anonymity and low restrictions to access. The differences between corporate hubs and other collaborative governance forms are summarized in Table 3.

Much of the previous open innovation literature has focused on
the distinction between transactional governance, emphasizing IP as the building block for knowledge exchange and collaboration (Arora & Gambardella, 2010;Gans & Stern, 2003), and relational governance, highlighting legitimacy, interactions, norms and trust as means for successful long-term shared value creation (Dahlander & Frederiksen, 2012;Fauchart & von Hippel, 2008;Powell et al., 1996;von Hippel, 2005 innovation processes by improving their innovation scouting (Rohrbeck, 2010), absorptive capacities (Cohen & Levinthal, 1990) and dynamic capabilities (Teece, 2007 openness and proactivity to existing and potential employees, partners, shareholders and other key stakeholders. In short, within the corporate hub, the SMEs still have the responsibility for their development processes and are exposed to the risk of failure, and all the above-mentioned advantages for the large corporation are in reach without the need for overly bureaucratic arrangements, project staffing or new venture investments. We have contributed theoretically to the field of open innovation by connecting the concept to the governance literature from the broader management field. We have theorized the rationale behind the BVH initiative, and more specifically its governance structure, in terms of who is invited to participate, how they are coordinated and how they are controlled. We have characterized the BVH as a corporate hub, and differentiated it in relation to other "open" governance modes, that is, market, network and bazaar. In the bio-pharmaceutical industry, market governance (e.g., trading with IP) as well as network and bazaar governance (e.g., building collaborative knowledge in mostly early phases of R&D) have all become well-established ways of conducting drug development (c.f. Barbanti, Gambardella, & Orsenigo, 2004;Chiaroni, Chiesa, & Frattini, 2008;Powell, 1998). The corporate hub structure goes beyond the more common transactional activities of selling and acquiring IP (Dutfield, 2014) and the coupled, relational process is applied not only in earlier phases of R&D (Kar, 2010;Trouiller et al., 2002) but also in the later, more competitive stages of drug development.

| CONCLUSION AND IMPLICATIONS
As a managerial contribution, the in-depth study can provide inspiration and direction to other large corporations aiming to enact coupled open innovation with small firms. We have detailed a governance structure which allows the SMEs to bridge the "liability of smallness" and the large corporations to bridge the "liability of oldness" into a productive symbiosis with reciprocal learning through ongoing day-to-day relations while also facilitating the effective use of otherwise abundant resources and opening up possibilities for new future revenue streams through protected IP routes. AstraZeneca has shown with the BVH that open innovation has a potential for large corporations even in the bio-pharmaceutical industry, which is an environment that Teece (1986) would call a tight appropriability regime. With the emphasis on the governance structure, we also draw attention to the design options that managers need to consider, pointing out that the governance structure has a direct link to potential consequences for the open innovation initiative, in terms of which forms of relations will emerge, and what type of values are being facilitated.