The governance of formal university–industry interactions: understanding the rationales for alternative models

Formal university–industry interactions: context and characteristics

Since the 1980s, industrial research and development (R&D) processes have become more open and distributed (Chesbrough, 2003; Powell and Grodal, 2005; Nesta and Saviotti, 2005) and the involvement of universities as innovation partners has increased. Several processes have combined to make interaction between firms and universities more appealing for both parties. From the perspective of firms, the growing complexity of products and processes and their scientific and technological content (Arora and Gambardella, 1994) increases the costs of vertically integrating all the competences needed for their development and makes it more convenient for firms to look for complementary competences outside their boundaries. Interaction with universities allows firms access to a wide international networks of scientists with heterogeneous competences and opportunities to establish relationships with the potential to generate innovations (Antonelli, 2008).

The increased pace of organizational and technological change generates uncertainties about the economic context in general and the likely development of technological trajectories and the emergence of dominant designs in particular. By building relationships with other organizations, especially universities, firms can hedge against the risk of backing the ‘wrong’ technology by engaging in several innovation processes at the same time (Powell et al., 1996; Wang and von Tunzelmann, 2000) and at a fraction of the cost of fully vertically integrated R&D activities. They can keep up to date with scientific developments (Meyer-Kramer and Schmoch, 1998) and enhance their learning and research opportunities by accessing advanced infrastructures and well-qualified human resources.

Interactions with universities are particularly cost effective for firms. The economics of knowledge shows that the costs of knowledge production in academia are lower than in the private research system because of the split structure of academic salaries (Dasgupta and David, 1994). University researchers’ fixed costs are covered by the payments received for their teaching activities, so that ‘the compensation schemes practiced in the academic system allow the supply side to operate on a variable cost base’ (Antonelli, 2008, p. 12). Also, a university affiliation signals quality and competence, based on the institution’s reputation in the open science system, which is an independent system that confirms the ability of academic researchers, lowers firms’ search costs for high-quality competences, and reduces the agency problems inherent in collaborations with knowledge workers whose skills are difficult to assess (Antonelli, 2008).

From the university’s perspective, developments in science-based technologies make industrial collaborations important for individual academic scientists to test models and to access funds and production and testing facilities (Koumpis and Pavitt, 1999; Lee, 2000). University institutions have become more proactive in seeking collaborations with firms because political trends are forcing them to find ways to reduce their dependence on public grants. These trends include reduced funding for university defence research and reduced government intervention in the economy (Geuna and Muscio, 2009). Policy interventions are encouraging universities to engage in third stream activities and are highlighting the role and importance of institutional interactions with industry and supporting the creation of an institutional infrastructure for knowledge transfer between universities and firms (Macdonald, 2011). Examples of such interventions include the creation of publicly funded regional knowledge transfer organizations in Germany, joint university–business competence centres in Sweden (Sellenthin, 2006) and support for the adoption of model contracts for university–industry collaboration in the UK (HM Treasury, 2003). As a result, third stream activities have become increasingly important for universities as a direct source of funds derived from commercial transactions and as a means of acquiring visibility and legitimacy and providing benchmarks for policymakers to measure their effectiveness (see HM Treasury, 2003).

The greater engagement of universities in third stream activities may bring some problems. The literature highlights such issues as the possible effects of increased dependence on private funding on the direction and content of research activity and the autonomy of the scientific enterprise and on the functioning of the open science system. Universities’ engagement in commercial transactions may have negative effects on firms’ innovation processes by putting the parties into direct competition, which may result in universities restricting access to scientific information, thereby making it more difficult for firms to appropriate the results of collaborative research with universities.

A possible side effect of increased participation in third stream activities could be the replacement of more traditional forms of interaction with industry, such as direct personal contractual relationships with academic researchers, with interactions mediated by an institutional knowledge transfer infrastructure. While some (Gibbons et al., 1994; Etzkowitz and Leydesdorf, 2000) argue that the supposedly more efficient new institutional knowledge transfer model is substituting for the older model and should be developed further, it could be argued that there are important differences between the two models that are not related to whether an individual academic or a representative of the university is signing the contract. The two models of governance have different histories, characteristics and economic implications (Geuna and Muscio, 2009) and, potentially, play different roles in the knowledge transfer process.

Personal contractual interactions involve an official contract between an academic and a firm. The individual scientist is hired to work as an external consultant on a firm project. The firm organizes and monitors project activities, which means that the firm retains control over the scope and organization of the project, but also has to bear the project coordination costs. Also, because the scientist working on the project is a self-employed external consultant, monitoring is relatively costly for the firm because of high agency costs.

Much of the empirical research overlooks this type of university–industry interaction, concentrating instead on analysing the interactions, even in consulting activities, mediated by the university structure. In part, this is because it is generally easier to collect this type of data, as we discuss in the next section, but it is also because the nature of personal contractual interactions is often misunderstood. It is often assumed that most interactions with individual academics are informal and, hence, difficult to measure and that, even if they are formalized, they involve ‘soft’ topics, such as the application of business methods or the solution of simple day-to-day business problems (see, for example, HM Treasury, 2003; Bercovitz and Feldmann, 2005). However, studies that explicitly consider formal academic consulting activity as a distinct knowledge transfer channel highlight its importance (see Rebne, 1989; Cohen et al., 2002; Beath et al., 2003; Perkmann and Walsh, 2008; Jensen et al., 2010) and suggest that consulting includes a wide range of activities linked to the exploitation of existing knowledge, the commercialization of research results and the performance of original research (Dechenaux et al., 2007; Perkmann and Walsh, 2008).

Personal contractual interactions traditionally played a major part in knowledge transfer from academia to industry. There is evidence of significant interactions between companies and university scientists in the late nineteenth century in Germany and England and in the early twentieth century in America. In Europe, these interactions continued to constitute the main share of university–industry collaborations throughout the twentieth century. The absence of formal business networks involving industry and academia meant that these interactions were usually built on social networks (often based on a common educational background, as in the case of alumni associations) involving academic researchers and their industry counterparts, and were characterized by high levels of interpersonal trust (Colyvas et al., 2002). The freedom of academics to contract with external organizations enabled these personal contractual interactions, which, in most contexts, were tolerated or regulated by specifying the number of hours per month that an academic could devote to external consulting activities.³

As universities came under increasing pressure to intensify their knowledge transfer activity in order to procure funding and to demonstrate excellence in knowledge dissemination (a criterion increasingly used by public funding agencies to assess university performance), academics were encouraged by their institutions to abandon personal contracts in favour of university-mediated arrangements. One of the arguments used by the university is that the university’s involvement insures the academic involved for any damage resulting from the performance of the contracted work. Institutional contracts are signed by the firm and a university representative. The scope and content of the project and the rules regarding the assignment of rights over the intellectual property emerging from the project are often negotiated by the parties, but the coordination costs of organizing and monitoring project activities are shared. Also, since the scientist works on the project as a university employee, monitoring the work is cheap for the firm. Although many institutional interactions are based on personal friendships between academics and commissioning firms, knowledge transfer offices are becoming proactive in looking for potential partner firms.

Conceptual and measurement problems in the analysis of university–industry interactions

Although there is a large body of research on the characteristics of university–industry interactions,⁴ there are very few investigations seeking to identify which governance form is most effective to mediate specific knowledge transfer interactions. There is also a lack of consensus on the most appropriate mechanisms for knowledge transfer between universities and industry. One of the reasons for this lack of agreement is that most studies rely on measures and data that are not directly comparable to capture the complexity of university–industry relationships.

Several empirical studies use data collected through surveys of academics and firms that take account of a wide range of alternative knowledge transfer channels. There is a recognition that several channels may be exploited simultaneously and that formal channels involving the commercialization of university knowledge (i.e. spin-offs, licences, patents) are among the least frequently used (Schartinger et al., 2001; Cohen et al., 2002; D’Este and Patel, 2007; Bekkers and Bodas Freitas, 2008). According to Bruneel et al. (2009), conference attendance and graduate recruitment constitute the main forms of firm interaction with universities, while Abreu et al. (2008) suggest that the most frequent interaction occurs within collaborative research networks. D’Este and Perkmann (2007) find that collaborative research projects, including consultancies, are more important sources of income than licensing in the UK. Schartinger et al. (2001) highlight crucial inter-sectoral and inter-disciplinary differences in the intensity with which the different channels are used. Because of their broad focus, these studies provide little evidence about the governance of formal interactions and the relative importance of difference governance forms.

In addition, knowledge transfer channels are categorized in different ways. For example, Perkmann and Walsh (2006) propose a distinction between socialized and non-socialized interactions; that is, between interactions that involve the establishment of social relationships (sponsored research projects, research consortia, collaborative joint ventures, research centres) and those that are purely contractual (licensing, specific ad hoc consultancy). Others indicate that all knowledge transfer channels – including those based on the sharing of codified knowledge, such as access to scientific publications and university patent licensing – are accompanied by the establishment of social relationships (Meyer-Kramer and Schmoch, 1998; Bozeman et al., 1995).

Most studies are based on one-off survey data or internal university information that is not standardized across universities. Evidence from different surveys is not always comparable because of respondent and sample biases. The results from comparing the responses from firm managers, R&D managers and inventors are often divergent (e.g. company researchers are usually more likely than firm managers to consider university research to be an important source of knowledge for firms). The survey design can also introduce bias. For example, the Community Innovation Survey (CIS) divides sources of knowledge into universities, scientific publications and conferences. However, this classification not only neglects many channels of knowledge transfer from universities that are discussed in the literature, it also introduces a downward bias in the importance of university knowledge (the overall importance of university knowledge should be measured as the sum of the knowledge directly obtained from universities and the knowledge obtained from scientific publications and conferences). Moreover, since the CIS and similar surveys focus on capturing innovation-related activities, respondents focus on business-related, accountable types of activities and sources, which usually results in comparative bias (respondents tend to rank the most concrete sources of knowledge as most important and understate the importance of interactions, such as personal contractual arrangements, that involve individual collaborators rather than partner organizations).

Finally, when (as is most often the case) studies rely on data made available by university technology/knowledge transfer offices, they capture only the set of interactions managed directly by the university (see Joly and Mangematin, 1996; Thursby et al., 2001). Policy often considers only statistics related to university and other higher education institutions, overlooking the fact that individual academic researchers may be active players in interactions with industry. Hence, although work on identifying knowledge transfer channels has become quite sophisticated, it has several limitations when the focus is on comparing different modes of governance of university–industry interactions.

A conceptual framework to explain the rationale for different forms of interactions

Interaction arrangements seem to reflect the concerns about and motivations for collaborating (Foray and Steinmueller, 2003). There is evidence suggesting that the governance of an interaction is associated with different aims of collaboration and different perceived transactions costs and levels of uncertainty (Artz and Brush, 2000; Zang et al., 2007). Several authors consider that different modes of governance of organizational interactions should apply to different levels of transaction costs. Gulati and Sing (1998) show that coordination costs, which depend on the expected complexity of the activities undertaken and the expected interdependence of tasks across organizational boundaries, can be high. Efforts to reduce these costs may influence the firm’s choice of governance form. Interactions in which project control and coordination costs are shared are argued to be more appropriate where the need for coordination, integration and processing of diverse information relative to diverse subtasks is high. Gulati and Sing suggest that this applies particularly in complex technological projects. On the other hand, if coordination costs are low, they can be more easily managed by one or other of the partners. Appropriability considerations also play a role in the choice of governance form, with shared governance structures being more relevant in conditions of weak appropriability, which carry a higher risk of opportunistic behaviour, such as free riding. Other things being equal, trust between the partners reduces the risks of opportunism (Williamson, 1975) and the need for shared coordination.

Another type of transactions cost that affects interaction is the cost of monitoring the degree of commitment of partners. According to Lacetera (2009), outsourcing to an academic institution increases the commitment of research scientists to the project because it is conducted according to the norms of the scientific community and within realistic timescales, which reduces the need for monitoring. The higher the potential monitoring costs, the more important it becomes to select a form of governance that allows scientists to operate within the framework of their academic institution.

These insights apply to the choice of governance mode for university–industry interaction. Since the content of the project around which the interaction is developed influences the extent of the transactions costs (especially coordination and monitoring costs), we would argue that certain governance modes are more appropriate for certain types of projects. Projects related to basic scientific knowledge have high coordination costs for two reasons. First, they involve high fixed costs and require the contribution of teams of researchers rather than of individuals. Second, they are characterized by high levels of codification of results, uncertainty in terms of possible application and externalities related to the range of possible beneficiaries (Nelson, 1959), all of which leads to low appropriability (Arrow, 1962). These types of projects also tend to involve high monitoring costs because the research is complex and open ended and monitoring the efforts and commitment of scientists is difficult. Based on these arguments, such projects are suited to institutional contracts that allow the firm to participate in the governance of the collaboration, which reduces coordination costs and ensures researchers’ commitment because they are operating within the framework of their scientific institution. Applied research projects, on the other hand, produce knowledge that is more immediately appropriable by the firm. Because the research is short term and closer to the firm’s activities, monitoring of university scientists’ commitment is easier for the firm. According to the arguments presented above, these types of projects are better suited to personal contractual governance.

In a survey of professors in different academic fields in Germany, Meyer-Krahmer and Schmoch (1998) found that collaborative research (typically involving the university institution) is important in microelectronics, software and biotechnology, where research tends to be basic, and that contract research (which includes academic consulting) is important in production technology, a field strongly oriented towards applied R&D. Because firms tend to be involved in a range of innovation projects, we expect firms to be involved in both modes of governance. In a companion paper (Bodas Freitas et al., 2010), we analyse the firm characteristics associated with choice of governance mode. Firms with high levels of absorptive capacity (Cohen and Levinthal, 1990) are likely to benefit from basic research projects, which are uncertain but enhance research productivity and allow unexpected technological spill-overs (Fontana et al., 2006; Laursen et al., 2011). Other things being equal, we expect firms with high levels of absorptive capacity to engage in institutional interactions.

The form of governance that firms favour depends on the resources available for cooperative activities. Small firms generally do not have spare resources (financial, managerial skills, cognitive abilities) to deal with such cognitively and socially distant institutions as universities and find it difficult to initiate and organize university–firm collaboration. We would expect small firms to be more likely to interact through personal contracts with individual academics.

Firms that rely on sourcing technology from external organizations (via collaborations, licensing of intellectual property, etc.) are more likely to have the capabilities required to search for knowledge providers and their search costs will be lower for personal contractual interactions. They may be part of a network of trusted academics with whom they have collaborated in the past, which reduces coordination and monitoring costs. They may have the technological and codification capabilities to enable them to write water-tight contracts. We would expect such firms to be more likely to engage in personal contractual interactions.

The policy framework also matters for firms’ choice of governance models for university–industry interactions. For example, public funds to support university to industry knowledge transfer that are restricted to university-mediated interactions may be an important determinant of the firm’s choice to set up an institutional collaboration. In countries where policies to support the institutional model are recent (e.g. Italy), we would expect both models of governance of university–industry relationships to exist and respond to different knowledge exchange needs. There is evidence of the co-existence of these two modes of interaction in the Piedmont region of Italy.