Appropriation strategies and open innovation in SMEs

Abstract

Drawing upon data from the fifth UK Innovation Survey, this article sheds light on how management choices on the nature of appropriation relate to management choices on the degree of openness within small and medium-sized enterprises (SMEs). To this end, our findings indicate a threshold effect of both informal and formal appropriation mechanisms on the likelihood of engaging in both coupled and inbound open innovation. That is, an emphasis on appropriation appears to be important in shifting firms from a closed to an open strategy. There is, however, little evidence that either approach to appropriation increases the extent of open innovation. In this, only informal intellectual property (IP) protection mechanisms associate with an increasing extent of inbound open innovation.

Keywords

cooperation innovation intellectual property search small firms

Introduction

While one might wonder about the extent to which it represents old wine on new bottles (Oakey, 2013; Trott and Hartmann, 2009), there is little doubt that the concept of ‘open innovation’ (Chesbrough, 2003a, 2003b) has enjoyed growing popularity over the last decade. In short, open innovation holds that businesses must increasingly utilise both internal and external sources of innovation – rather than relying solely on internal research and development efforts. Moreover, firms must recognise that both internal and external paths to the market exist (in the latter case, for example, through technology licensing). This, we were told, is ‘a new logic of open innovation that embraces external ideas and knowledge in conjunction with internal research and development (R&D). This change offers novel ways to create value – along with new opportunities to claim portions of that value’ (Chesbrough, 2003b: 41; emphasis added).

The ability to capture value from one’s innovation is called appropriability – and is typically concerned with monetising intellectual property (IP) rights. However, to the extent that openness entails disclosure this implies a tension with appropriability. Moreover, a key issue is that both open innovation and appropriation present particular challenges to smaller firms. As Gassmann et al. (2010: 219) observe, ‘SMEs are the largest number of companies in an economy, but they are under-researched in the open innovation literature’ (see also Wynarczyk et al., 2013). While most of the literature regarding open innovation has been concerned with large firms, recent empirical evidence suggests that open strategies are increasingly widespread in small and medium-sized enterprises (SMEs) (Van De Vrande et al., 2009) and associate with improved innovation performance (Parida et al., 2012). Given internal resource constraints to innovation in small firms (Rothwell, 1991), the potential offered by openness ought not to be too surprising. Similarly, recent studies have shown the IP strategies of smaller firms to be different from their larger peers, at least in part because of the lack of resources to defend more formal protection mechanisms (Leiponen and Byma, 2009). Rather, SMEs are likely to rely on looser forms of IP protection, such as secrecy, complexity and speed.

This, then, raises a paradox: open innovation strategies seem particularly well suited to the pursuit of innovation in resource-constrained SMEs, with open innovation enabling firms to overcome the liability of smallness (Gassmann et al., 2010). However, openness entails disclosure and resource constraints limit the ability of SMEs to establish and enforce protection. This apparent paradox is the focus of this article. Specifically, we draw upon data from the fifth UK Innovation Survey (IS), which allows us to construct measures for both the inbound open innovation and coupled open innovation strategies. These measures have considerable recent precedent in the empirical literature on open innovation (Classen et al., 2012; Laursen and Salter, 2006; Leiponen and Helfat, 2010; Love et al., 2014). The survey also provides data on the use of a variety of both formal and informal mechanisms for the protection of IP, narrowly, and innovations, broadly (as well as on a variety of variables which one might anticipate moderating the relationship between the organisation of innovation and the pursuit of particular appropriation strategies). Our goal is to better understand how management choices on the nature of appropriation relate to management choices on the degree of openness.

The article is structured as follows: The opening section provides a brief elaboration of the concept of open innovation, as it appears in the literature, and the dimensions of openness which are the focus of the analyses presented in this article. In addition, this section reflects upon the apparent tensions between openness and appropriability and presents three linked hypotheses, which our subsequent analyses test. The next section describes our data followed by the methods and results section which outlines our choice of econometric method and details the results of our estimations. Finally, we offer a concluding discussion with initial reflections on business and public policy.

Open innovation

Open innovation as a concept is most obviously associated with the work of Henry Chesbrough. Accordingly, Chesbrough’s initial text (2003a) is where we turn for an overview. To this end, Chesbrough (2003a) observes that

valuable ideas can come from inside or out of the company and can go to the market from inside or outside the company as well. This approach [i.e. “open innovation”] places external ideas and external paths to market on the same level of importance as that reserved for internal ideas and paths in the Closed Innovation model.

To students of innovation studies, from a wide variety of disciplinary backgrounds, this may seem rather ‘old hat’. Interactive and network models of innovation have achieved something of a normative status over the last 30 years. And, while there has been the occasional recent related defence of the linear model (Balconi et al., 2010), Chesbrough’s closed model has much of the appearance of a straw man. Nevertheless, the open innovation concept has use as a focusing device and, in according outflows the same importance as inflows, adds something to standard empirical accounts of innovation networking (Bougrain and Haudeville, 2002; Nieto and Santamaria, 2007).

Importantly, as Chesbrough (2003a, 2003b) stressed, openness is not a binary phenomenon. Yet, much of the empirical work has sought to contrast open and closed strategies (Barge-Gil, 2010). Such a stark distinction offers simple clarity, but it does little to capture the essential more or less character of open innovation. Recognising Chesbrough’s (2003b) continuum ‘from essentially closed to completely open’ (p. 37) is a crucial precursor to ‘understanding the mechanisms, inside and outside of the organisation, when and how to fully profit from [open innovation]’ (Enkel et al., 2009: 312). As outlined below, the approach taken here follows recent empirical work (Garriga et al., 2013; Laursen and Salter, 2006) in adopting measures of open innovation that are grounded in a recognition that firms may be more and less open, not simply open and closed.

Modes of openness

Much of the work on innovation systems and innovation networks has stressed the importance of external sources of inspiration, ideas, knowledge and resources for innovation by firms (or regions, or nations) (Pittaway et al., 2004). The perspective adopted is that of the firm (or region, or nation), and SMEs are important in many expositions, with networks viewed as a means to ameliorate resource constraints while preserving behavioural advantages. The emphasis is on external search and subsequent absorption. Improving absorptive capacity (Cohen and Levinthal, 1990) is the frequent managerial exhortation (Spithoven et al., 2011) or policy advice (Giuliani, 2005). This is Chesbrough’s (2003a) inbound innovation, and it is likely to be considerably more commonplace than outbound innovation (Huizingh, 2011). That is, if firms acquire and source from many places, but sell and reveal to only a few (Dahlander and Gann, 2010), then an asymmetry is likely to be apparent from the firm’s perspective. Moreover, to the extent that outbound innovation entails that the firm’s IP (the bases of its innovation) be commercialised by others (e.g. through the sale of licences or spin-off activity), for smaller firms, in particular, this mode of open innovation is likely to be relatively infrequent. Given Arrow’s classic information paradox,¹ firms are likely to require both identifiable and legally protectable rights to their IP and the wherewithal to protect those rights.

One immediate consequence is that micro data generated directly from firms are likely to better represent inbound innovation than outbound innovation. This is true of the Community Innovation Survey (CIS) data used here and is manifest in the focus of open innovation studies that draw from it (Laursen and Salter, 2006). Certainly, in SMEs that dominate business populations (and survey samples), a tendency towards inbound innovation over outbound innovation may be an inevitable consequence of resource constraints (Van De Vrande et al., 2009). The sorts of knowledge-sourcing activities that are commonly measured as inbound open innovation (see Garriga et al., 2013; Leiponen and Helfat, 2010) may often require little procedural formality and limited resource commitments.

Given a predominant focus on inbound open innovation, a popular approach adopted in the empirical literature follows Laursen and Salter (2006) in distinguishing between broad and deep forms of open innovation (see also Garriga et al., 2013; Tether and Tajar, 2008). Here, an open innovation strategy is conceived of as a search strategy. Openness breadth is typically measured as a simple count of the number of sources of information and knowledge used for innovation, while depth is measured as either the intensity of use of the various sources of information and knowledge or as a simple count of the number of innovation-related cooperative partners. This approach raises some concerns: In the first instance, as simple counts of sources or partners, both are essentially concerned with breadth. Certainly, intensity of use or cooperation may imply greater reciprocity and resource commitment, but they do not fundamentally measure depth. For instance, Leiponen and Helfat (2010) study the impact of breadth in knowledge sourcing on the innovativeness of a sample of Finnish firms and measure breadth as a count of the number of sources used intensively. Similarly, Love et al. (2013) explore the relationship between past openness to current openness and measure breadth as simple count of innovation partners. Indeed, in their more recent work, Laursen and Salter (2014) appear to recognise the limitations of their prior measures of breadth and depth, distinguishing instead between external search breadth and collaborative breadth. The former is measured to capture some element of intensity, while the latter is measured, as before, as a simple count of cooperative partners.

Of course, the measures themselves are not problematic (besides the inconsistent use of measures and terminology). Rather, the problem is with what they purport to measure. Here our concern is with the conception of cooperation or partnering as inbound open innovation. To this end, there is a nice parallel in network theory’s distinction between flow and bond models of networks (Borgatti and Halgin, 2011). The flow model denotes the transfer of tangible and intangible resources (e.g. information) between actors in a network. In contrast, the bond model represents the alignment of interests and coordination of action. As Glückler (2013) summarises, ‘As an economic outcome, innovation can either be explained through flows of information or through the coordinated creation of collective novelty’ (p. 884).

To the extent that cooperation entails complementary partners working towards the development and/or commercialisation of new technologies, it is essentially concerned with co-creation (Enkel et al., 2009). Cooperation necessarily entails reciprocity, that is, ‘give and take’. To that end, its conflation with inbound open innovation seems misplaced and is likely to muddle the empirical landscape. Here, we prefer the approach adopted by Gassmann and Enkel (2004), which is consistent with recent network theorising. These authors distinguish three core open innovation archetypes: outside-in, inside-out and coupled (Gassmann et al., 2010). The first two are, respectively, the familiar inbound and outbound open innovation processes. The coupling process, however, aligns with the bond model of networking and acknowledges the likelihood that, in some instance, open innovation is a more intensively collaborative process.

Unfortunately, the available data limit us to exploring only two of these three types: outside-in (or inbound) open innovation and coupled open innovation. Regardless, inbound and coupled models of open innovation represent quite different logics of organising innovation. These differing logics are likely to be manifest in, inter alia, differences in resource commitments, managerial commitment, reciprocity and the importance of trust. We anticipate that these, in turn, will be reflected in different attitudes to appropriation and in the use of different protection mechanisms. That is, different forms of open innovation, entailing different degrees of exposure, will be influenced by how firms perceive the problem of appropriability.

Openness and appropriation

As Geroski (1995: 91) observes, ‘the feature of inventive and innovative activity that most clearly sets it apart from other strategic investments made by firms is the problem of appropriability’. Both the creation of value and the appropriation of value are central to competitive advantage (Lawson et al., 2012). Yet, to the extent that openness entails disclosure (or exposure), this implies a tension with appropriation. Almirall and Casadesus-Masanell (2010) characterise this tension in terms of a trade-off between divergence and discovery. Divergence relates to the idea that the goals of network actors are unlikely to be perfectly aligned, while discovery rests on the notion that openness may stimulate innovation and improve value creation by combining multiple complementary perspectives. Divergence and discovery work in opposite directions, such that there is a trade-off between the cost of losing control of the technology and the benefits of aggregating knowledge from other actors to enhance innovation. As Almirall and Casadesus-Masanell (2010) suggest, opening up ‘often hampers the strength of property rights and impairs the developers ability to capture value’ (p. 27). This friction may retard open innovation if it discourages information search or collaboration (West, 2006). In their study of European firms, Enkel et al. (2009), for instance, note ‘loss of knowledge’ and ‘loss of control’ as among the principal risks associated with open innovation activities from the perspective of firms. In the case of technology-based small firms, Oakey (2013) argues that concerns over confidentiality and the degradation of internally generated IPs provide good reasons for firms to be closed.

What is clear is that varying use of appropriation mechanisms and their inconsistent ability to protect technologies from imitation have long been considered an important determinant of variety in innovation modes (Pavitt, 1984). For instance, in his classic essay, Teece (1986) observed that the nature of appropriability regimes influenced the organisation of innovation most suitable for profiting from innovation. In loose regimes, integration is suggested, while tight regimes with secure legal protection allow firms to pursue partnering strategies in search of specialisation economies. In short, the ability of a firm to pursue open innovation strategies will be tied, at least in part, to their IP strategies.

Reflecting on Teece’s (1986) classic work, Pisano (2006) notes that

In order to help innovators specialize (safely), markets for know-how must work effectively. Networks of innovation thus depend partly on intellectual property regimes. Strong intellectual property regimes would support broader and more diffuse networks of innovation (p. 1124).

Certainly, there is some supporting empirical evidence for this view. For instance, using data from the first Dutch CIS, Brouwer and Kleinknecht (1999) observed that both the propensity to patent and patenting intensity correlated with participation in collaborative R&D. From this, it is argued that patents, as a formal and well-defined appropriation mechanism, play an important role in clarifying the ownership of the intellectual output of collaborative innovation and limit the scope for disagreement (Arundel, 2001). It may also signal their quality to potential partners (Laursen and Salter, 2014). In his study of 154 industrial firms, Lichtenthaler (2010) similarly concludes that ‘a firm’s intellectual property portfolio constitutes an important driver of open innovation’ (p. 387). ‘At first glance’ (West, 2006: 32), it would seem that stronger IP regimes are likely to associate with more open innovation.

Yet, patents are not the only appropriation mechanism available to firms and appropriation strategies may take various forms, employing a variety of isolating mechanisms – ranging from patents and copyright to secrecy and speed-to-market. In recent work, Laursen and Salter (2014) use UK IS data to construct a measure of appropriation strategy which acknowledges this broader array of protection mechanisms. In this, they simply sum responses on the perceived importance of all appropriation mechanisms listed (a total of six). The implication is that higher scores on this measure equate with stronger appropriation strategies. The findings suggest a concave relationship between the strength of appropriation strategy and openness. Openness is measured in the same fashion as in this article and the concavity is observed to be stronger for coupled openness than simpler inbound openness. Yet, in both cases, the point of inflection is high and the general conclusion is that stronger appropriation is associated with greater openness.

However, just as the concept of open innovation is not uniform, it seems rather crude to treat the various appropriation mechanisms as simply additive. Indeed, Laursen and Salter (2014) recognise this limitation, noting that future research may advance their discussion by exploring the relationship between choices among protection mechanisms and modes of openness. Greater sensitivity in the treatment of different protection mechanisms may go some way to resolving the apparent contradictions in the literature, with openness and appropriation seen as both conflicting and complementary. Contrast, for example, Chesbrough’s (2003a) strong managerial injunction to open up the firm and move away from an excessive focus on protecting ideas and Oakey’s (2013: 330) suggestion that for technology-based firms ‘internal development of new products … must remain confidential’, with consistent empirical evidence that the employment of (certain types) of protection mechanisms associates with openness (Laursen and Salter, 2014).

To this end, it may be useful to distinguish between types of protection mechanism. Our data list six mechanisms: patents, trademarks, registration of design, secrecy, lead-time and complexity. While some of these are inevitably mutually exclusive (e.g. patents, which require disclosure, are inevitably incompatible with an appropriation strategy grounded in secrecy) (Hurmelinna-Laukkanen and Puumalainen, 2007), evidence suggests broad complementarity across innovation protection mechanisms (Thomä and Bizer, 2013). In this vein, it is common to distinguish two broad types of mechanisms: on one hand, are those mechanisms founded on legal protection (e.g. patents, trademarks and registration of design). These may be characterised as institutional forms of protection and be contrasted with more the strategic forms of protection afforded by secrecy, complexity and speed (Arbussa and Coenders, 2007).

In other words, this is the familiar distinction between formal and informal IP protection (Amara et al., 2008; Thomä and Bizer, 2013). Formal mechanisms are those which have a legal basis (patents, design registration and trademarks) and informal mechanisms are extra-legal (secrecy, complexity and speed-to-markets). Empirical evidence has shown informal mechanisms to be more widely used than formal mechanisms. Aggregating data from the third, fourth and fifth waves of the UK IS, Hall et al. (2011) show that only 22% of firms use any formal IP protections (with only 16% using patents), while almost 34% used at least one informal mechanism (around 32% used secrecy). These differences are starker still among small firms (Leiponen and Byma, 2009). This is perhaps unremarkable. The strength of the protection afforded by patents, for example, rests on the availability of resources to credibly threaten litigation in the event of infringement. Smaller firms rarely have the necessary resources (West, 2006). Moreover, as Thomä and Bizer (2013) argue, cost and complexity issues are likely to see smaller firms more reliant upon informal methods of protecting their intellectual assets rather than formal methods.

Our interest, however, is not in the relative frequency of the use of formal and informal protection mechanisms (although our data are consistent with prior evidence that informal mechanisms are predominant). Rather, it is in their relationship with our two forms of openness – inbound and coupled open innovation. Following the arguments above, one might associate formal protection mechanisms most clearly with Teece’s tight IP regimes and anticipate a positive relationship with coupled openness. Vested in collaboration, coupled openness requires partners to observe an agreed structure for the relationship. As a partnership, it implies reciprocal knowledge flows which are likely to entail more revealing (Henkel et al., 2013) than in the inbound open innovation model. Importantly, in contrast to the use of informal protection mechanisms, a firm’s use of formal protection mechanisms is likely to be highly visible to external actors (Hall et al., 2011). In this way, it sends quality signals to potential partners, signalling the possession of valuable technology and the capability to develop valuable technology (Laursen and Salter, 2014). It is also likely to signal a familiarity with the formal intellectual property rights (IPR) system, which may lessen concerns over IP conflicts. This leads us to hypothesise the following:

H1. The use of formal innovation protection mechanisms is positively associated with a coupled open innovation strategy.

With respect to the relationship between the use of informal protection mechanisms and coupled open innovation, we are somewhat ambivalent. On one hand, one might anticipate that the extent of disclosure required would render informal methods insufficient in protecting the firm’s prior art. In contrast, however, Leiponen and Byma (2009) find that small firms that engage in horizontal innovation-related cooperation show a preference for protecting innovations through speed-to-market rather than patents. Given resource consideration which may preclude formal methods for all but a peculiar subset of SMEs, an emphasis on informal protection methods may imply confidence in a firm’s complementary assets. This, in turn, may support a belief in the firm’s ability to appropriate a larger share of value from a coupled open innovation process. However, Leiponen and Byma (2009) find no relationship between the use of informal protection mechanisms and vertical cooperation. In their sample, it would appear, only cooperation with rivals requires a particular emphasis on appropriation. On balance, we do not anticipate any relationship between a coupled innovation process and the use of informal protection mechanisms.

In the case of inbound open innovation, where the extent of disclosure is limited, the use of informal mechanisms may be sufficient. Here, the information and knowledge flows are predominantly one way. The limited revealing required on the part of the focal firm is likely to be adequately protected by secrecy, complexity or lead-time advantages. Moreover, confidence in the efficacy of these methods is likely to provide a platform for searching outside the firm, trusting that external actors will have limited incentive or ability to imitate. This leads us to hypothesise the following:

H2. The use of informal innovation protection mechanisms is positively associated with an inbound open innovation strategy.

We also anticipate the relationship between an inbound open innovation strategy and the use of formal appropriation mechanisms to be positive. Partly this reflects observed complementarities in the use of various protection mechanisms (Thomä and Bizer, 2013). However, it may also reflect the quality signalled by formal mechanisms, with higher quality firms better able to recognise valuable external sources of knowledge and information. In addition, the use of formal mechanisms has been shown to be higher in competitive environments, rich in technology opportunity and where codified knowledge plays a greater role (Hall et al., 2011). These features are likely to both induce and admit greater external search in the form of inbound open innovation. This leads us to hypothesise the following:

H3. The use of formal innovation protection mechanisms is positively associated with an inbound open innovation strategy.

In summary, our interest is in the association between the relative emphases on formal and informal innovation protection mechanisms and two forms of open innovation: inbound open innovation and coupled open innovation. Despite recent contributions (Laursen and Salter, 2014; Mina et al., 2014), relatively little is known about how approaches of SMEs to appropriation relate to their open innovation practices. Given the varying strengths of the relationships implied in inbound and coupled modes of openness, we anticipate variations in the quantity and quality of resource and knowledge flows and, importantly, the extent to which the firm (and its partners) are exposed. Varying exposure (or disclosure) places varying pressure on IP and its protection. This paradox of openness (Laursen and Salter, 2014), in relation to appropriation, is likely to be particularly acute for smaller firms. Such firms possess limited resources to enforce formal, institutional, protection mechanisms and, given smaller innovation portfolios, are likely to be more selective in the protection mechanisms they employ (Thomä and Bizer, 2013).Yet, it is exactly these resource constraints that may see them garner the greatest relative returns from openness (Parida et al., 2012; Van De Vrande et al., 2009).

Data and measures

Data

To test our hypotheses, we employ data from the 2007 UK IS, which covers the period from 2004 to 2006.² The UK ISs are part of the wider European CIS project. The CISs derive their motivation and guidance from the Organisation for Economic Co-operation and Development’s (OECD) Oslo Manual (OECD, 2005) and are a subject (rather than object) oriented approach to collecting information on innovation activities (Arundel and Smith, 2013). The European ISs have been extensively used as the basis for empirical studies of innovation.

The 2007 survey is the fifth UK IS and was sent to 28,000 UK establishments (here, we call them firms, which may be the source of confusion in the case of multiplant enterprises, but is less likely to be a concern in the SMEs that are our particular interest) with 10 or more employees. While the sample incorporated a census of firms employing more than 250 workers, the SME component of the sample was stratified to ensure adequate regional and industrial coverage. The survey was funded by the Department of Business, Innovation and Skills (BIS), and it was administered by the Office for National Statistics (ONS). Survey questionnaires were distributed in a single wave on 31 March 2007. A 53% response rate was achieved. The survey includes both service and manufacturing firms. As noted, our interest is in SMEs. In line with European Union (EU) convention, we take that to mean firms employing fewer than 250 workers (and, in this case, more than nine). This particular focus results in a study sample of 5781 firms.

Measures

Our dependent variables measure the extent of inbound open innovation activities and coupled open innovation activities, respectively. In doing so, they fall in line with a recent body of empirical work on open innovation that draws on CIS data (or similarly constructed data) (see, for example, Laursen and Salter, 2006; Leiponen and Helfat, 2010; Love et al., 2013). Specifically, we construct measures for inbound open innovation and coupled open innovation.

Our measure of inbound openness of innovation employs a survey question which asked respondents about the use and importance of a variety of information sources for the firm’s innovation activities. In total, 11 potential information sources (see Appendix 1 for a list of the sources) were listed and firms were directed to indicate whether each source was of high, medium or low importance, or was not used. To create our variable, we gave each high response a value of 1; all other responses were given a value of 0. These scores were summed. In this way, our measure of inbound open innovation in a count of the number of high use sources of information for innovation. This has the benefit of measuring both the extent of inbound openness and capturing, to some degree, the intensity of use. That is, firms may be more or less open, not simply open or closed.

Our measure of coupled open innovation follows a similar logic. Firms were asked whether they had cooperated with a variety of potential innovation partners over the period covered by the survey. In total, seven partner types were listed (see Appendix 1). While firms were also given scope to indicate the location of partners, this information is not used here. Rather, where firms indicated cooperation with a given partner type, located at one or more spatial scale, this was coded as 1, while no cooperative relation with that type of partner was coded as 0. The seven types were then simply summed to arrive at our measure of coupled open innovation. This measure does not necessarily capture the total number of cooperative partners (since firms may have cooperated with more than one supplier, customer, etc.). However, it does afford an indication of the breadth of cooperation and, albeit imperfectly, the extent of coupled open innovation. Although framed in terms of the breadth or depth of inbound open innovation, this measure enjoys recent empirical precedent (cf. Laursen and Salter, 2006; Love et al., 2014).

Our independent variables measure the use of an array of formal and informal innovation protection mechanisms. In total, six means of protecting innovation were listed, and given the mixed sectoral composition of our sample, all six are used in constructing our variables. Firms were asked to indicate whether they had used each mechanism for the protection of innovations during the period covered by the survey.³ From these responses we construct two variables which simply sum the use of the various mechanisms: formal (registration of design, trademarks and patents) and informal (secrecy, complexity of design and lead-time advantage on competitors) to create indices of the use of formal and informal appropriation strategies. Our formal methods are public and have a legal basis. Our informal methods are strategic, extra-legal and imperfectly observable outside of the firm.

In addition to the dependent and independent variables, the survey collects data on a variety of firm level characteristics that past research has shown to influence openness. Following the oft-repeated aphorism that ‘firms must have resources to get resources’ (Eisenhardt and Schoonhoven, 1996: 137), we include firm size⁴ in 2006 to capture broad resource stocks. We also include measures of innovation-specific training and the conduct of internal R&D (both binary variable indicating incidence) to proxy resource qualities. These variables are intended to control for our expectations that access to resource and resource sophistication may associate with both the extent of openness and the extent of appropriation strategies. Recognising that openness may vary across industries (Aslesen and Freel, 2012), we include a series of dummy variables for industry using Standard Industrial Classification (SIC) at the 2-digit level. Finally, we include controls for market location. In part, this is in recognition of the variety in systems and enforcement across IP jurisdictions, but also reflects the notion that a more local market focus may make it easier to develop high trust relationships which lessen the need for protection (Thomä and Bizer, 2013).

Finally, as outlined below, we ultimately estimate zero-inflated negative binomial (ZINB) models in recognition of the high number of ‘zero’ responses in the dependent variables (some 57.14% of sample firms used no sources of information ‘intensively’ and 45.01% reported no cooperative partners). ZINB estimations assume that the zero outcomes and the non-zero outcomes are a result of two different processes. Here, the former is concerned with modelling the likelihood of being ‘open’ at all (inbound and coupled) – that is, a binary outcome modelled with a logit. The group of ‘closed’ firms is likely to be a mix of non-innovators and ‘closed’ innovators. To the extent that innovation outputs are an imperfect proxy for innovation intent, the relative distribution of these is difficult to estimate.⁵ The latter is concerned with modelling the extent of openness – that is, a count outcome modelled with a negative binomial model. While our primary concern is with extent, recognising the two processes is important in generating robust estimates. A similar approach to modelling openness was followed by Xia and Roper (2008) and Drechsler and Natter (2012).⁶ Unfortunately, Xia and Roper (2008) do not report the components of their logit (the ‘zero’ openness model). Drechsler and Natter (2012) note that ‘[t]he underlying reasons for these large fraction of zeros might be unobserved heterogeneity across the firms and/or a process with separate mechanisms for generating zeros (closed innovation) and non-zeros (high levels of openness)’ (p. 442). They do not speculate on the separate mechanisms and include the same variables in both the binary and count components of their ZINB, rationalising that a similar vector of variables will explain both openness (as a binary phenomenon) and its extent. We take a similar approach here.

Table 1 reports mean statistics and a correlation matrix for our independent and control variables. The correlation matrix offers some indication of the extent to which multicollinearity may be a problem in our subsequent analyses. As can be seen, none of the reported values are high enough to raise concern over the model specifications. In addition to the correlation matrices, variance inflation factor (VIF) scores were calculated for each of the sets of independent variables in each of the models reported in this article. Again, none of the VIF scores were sufficiently high to cause concern. Another concern in studies of this type (i.e. single respondent surveys) concerns bias arising from common methods variance (CMV). While the IS was designed to minimise the potential for CMV by varying the response type (mixing Likert scale questions, yes/no answers, percentages and absolute numbers), we formally tested for CMV using Harman’s one-factor test (for recent examples in the open innovation literature, see Leiponen and Helfat, 2010 and Love et al., 2013). We included all independent and control variables (excepting industry dummies) in the exploratory factor analysis. This identified a number of factors, with the first factor accounting for less than the majority of the variance. Accordingly, we are confident that the bias arising from CMV is not a major concern in our study. Both the VIF scores and the CMV test results are available from the authors on request.

Table 1.

Mean scores and correlation matrix.

	Mean	1.	2.	3.	4.	5.	6.	7.	8.	9.
1. Coupled innovation	0.542	1.00
2. Inbound innovation	0.558	0.46^a	1.00
3. Log (of employment)	3.611	0.18^a	0.20^a	1.00
4. Innovation-specific training	0.342	0.33^a	0.37^a	0.19^a	1.00
5. Internal R&D	0.294	0.36^a	0.45^a	0.13^a	0.41^a	1.00
6. Formal IP (0–3)	1.056	0.30^a	0.39^a	0.23^a	0.35^a	0.40^a	1.00
7. Informal IP (0–3)	1.630	0.35^a	0.41^a	0.24^a	0.41^a	0.38^a	0.41^a	1.00
8. International market – Europe	0.328	0.20^a	0.20^a	0.15^a	0.19^a	0.17^a	0.30^a	0.37^a	1.00^a
9. International market – non-EU	0.223	0.23^a	0.22^a	0.14^a	0.15^a	0.30^a	0.32^a	0.35^a	0.64^a	1.00

R&D: research and development; IP: intellectual property; EU: European Union.

N = 5781.

Note: The superscript a is intended to indicate significance at the 0.01 level (i.e. p<0.01).

Methods and results

As noted, given the count nature of our dependent variables, count regression methods are suggested. While Poisson models have a number of attractions, over-dispersion in the dependent variables renders them unreliable for both inbound and coupled innovation models.

Following this, we estimated ZINB models for inbound and coupled open innovation. In all instances, the likelihood ratio test that alpha = 0 is significantly different from zero. This suggests that the data are over-dispersed and that a ZINB model is more appropriate than a zero-inflated Poisson model. Furthermore, the Vuong test results in both models suggest that the ZINB model is a better technique to use compared to the standard negative binomial model.

Coupled open innovation regression results

Table 2 presents the results of the model testing the relationships between the extent of coupled open innovation strategies and respective use of formal and informal appropriation mechanisms. Column 1 contains the results of the logit model intended to explain the zero outcome (i.e. being closed), while column 2 contains the results of the negative binomial model concerned with the degree, or extent, of coupled open innovation. As indicated, in modelling our relationships, we control for a variety of factors which have previously been hypothesised to influence openness. In large part, these controls act as anticipated. In both the logit and the negative binomial models, resources matter. Both the volume of resources (proxied by the number of employees) and the quality of resources (investment in innovation-specific training and investment in R&D) are positively associated with coupled open innovation and with higher rates of coupled open innovation. Clearly, the role of internal resource and capability in allowing the pursuit of cooperative models of innovation ought not to be underestimated when delivering open innovation imperatives to small firms (Fritsch and Lukas, 2001; Tether, 2002). However, contrary to expectations, the geography of markets does not appear to have any influence on the extent of coupled open innovation, although it does appear that engagement in distant (i.e. non-EU) markets associates with a greater likelihood of having at least one cooperative partner.

Table 2.

ZINB regression models of the extent of coupled open innovation.

	‘Closed’ innovation (i.e. zero outcome)		Degree of openness (i.e. >0 outcome)
	Coefficient	Standard errors	Irr	Standard errors
Log (employees)	−0.168^a	0.041	1.054^a	0.017
Training	−1.337^a	0.118	1.198^b	0.081
Internal R&D	−1.766^a	0.121	1.050^a	0.026
Formal innovation protection (0–3)	−0.160^b	0.066	1.022	0.047
Informal innovation protection (0–3)	−0.681^a	0.060	1.018	0.055
International market – European	0.036	0.186	0.964	0.244
International market – non- European	−0.546^a	0.182	0.915	0.388
Constant	4.903^a	0.219	1.621^a	0.283
Ln alpha	−2.304^a	0.250
Alpha	0.125	0.032
Vuong test	13.44^a
Log likelihood	−2418.92
Likelihood ratio	94.52^a
N	5781

ZINB: zero-inflated binomial; R&D: research and development; Irr: incident rate ratio.

Thirty-eight industry dummy variables were included in the model with construction (SIC45 as the reference).

p < 0.1; ^bp < 0.5. The superscript ^a is intended to indicate significance at the 0.01 level (i.e. p<0.1).

Turning to our variables of interest, we hypothesised (H1) that the use of formal IP protection mechanisms would associate with a coupled open innovation. However, our data provide only partial support. The use of formal mechanisms does not appear to be related to the extent of engagement in coupled open innovation. The rationalisation underpinning our hypothesis rested partly on empirical precedent related to patents (Brouwer and Kleinknecht, 1999), but largely on the anticipated dual effects of the greater revealing associated with partnerships (Henkel et al., 2013) and the quality signals that visible IP sends to potential partners (Laursen and Salter, 2014). However, in our SMEs, these influences do not appear to dominate beyond a threshold effect (see below). This may reflect potential countervailing forces. For instance, given evidence that smaller firms tend to be highly selective in their use of protection mechanisms (Thomä and Bizer, 2013), it may be that an excessive emphasis on appropriability signalled by the use of multiple formal protection mechanisms is sometimes ‘interpreted as signalling that collaboration with the focal firm will be difficult, and that conflicts over control and ownership of knowledge might ensue’ (Laursen and Salter, 2014: 870). Yet, it may also reflect anxiety over exposure for firms with limited resource to prosecute infringement.

In contrast to formal protection mechanisms, we were ambivalent about the likely association between coupled open innovation and the use of informal protection mechanisms. For instance, we speculated that the extent of disclosure likely to attend coupled open innovation would render the protection afforded by informal mechanisms insufficient. However, past empirical work had suggested a positive relationship between the use of informal IP protection and cooperation with competitors (Leiponen and Byma, 2009). Our SME data reflect this ambivalence. Again, there is no evidence that a focus on IP protection associates with greater openness.

Strikingly, IP protection (both formal and informal) is significantly associated with the zero outcomes. In other words, the extent of IP protection reduces the likelihood of being closed to coupled open innovation. While IP protection does not associate with the extent of openness, it appears to have the flavour of a ‘ticket to the game’. A similar finding is noted by Drechsler and Natter (2012) on the basis of German IS data. Their concern was also with a cooperation-based openness measure, although their sample included large firms. From their results, they conclude that ‘an effective appropriability regime is important to move from closeness to openness in innovation’ (p. 443). There is, however, no evidence in our data that it subsequently associates with increasing levels of openness.

Inbound open innovation estimation results

Turning to H2 and H3, Table 3 follows a similar logic to Table 2, with respect to our model of inbound open innovation. As before, column 1 contains the output from the logit model to explain ‘zero’ outcomes, and column 2 contains the negative binomial model of the extent of inbound open innovation.

Table 3.

ZINB regression models of the extent of inbound open innovation.

	‘Closed’ innovation (i.e. zero outcome)		Degree of openness (i.e. >0 outcome)
	Coefficient	Standard errors	Irr	Standard errors
Log (employees)	−0.080^a	0.024	1.028^a	0.003
Training	−0.679^a	0.077	1.336^a	0.057
Internal R&D	−1.181^a	0.088	1.234^a	0.042
Formal innovation protection (0–3)	−0.229^a	0.033	1.027	0.022
Informal innovation protection (0–3)	−0.250^a	0.039	1.091^a	0.023
International market – Europe	−0.251^b	0.103	0.811	0.234
International market – non-European	−0.300^a	0.101	1.001	0.238
Constant	3.779^a	0.132	1.860^a	0.226
Ln alpha	−7.102^a	0.706
Alpha	0.071	0.024
Vuong test	9.88^a
Log likelihood	−4138.027
Likelihood ratio	169.68^a
N	5781

ZINB: zero-inflated binomial; R&D: research and development.

Thirty-eight industry dummy variables were included in the model with construction (SIC45 as the reference).

p < 0.1; ^bp < 0.5. The superscript ^a is intended to indicate significance at the 0.01 level (i.e. p<0.1).

Again, taking our control variables in the first instance, as before, having resources appears to be associated with both the likelihood of being open and with the extent of openness. That is, resource quantities and qualities associate with the likelihood of using at least one information source intensively and using more sources. Inbound open innovation (here, conceived of as external search) may be thought of as a soft form of openness (Laursen and Salter, 2014). It frequently involves sourcing knowledge from external actors without the requirement for formal, often legally binding, agreements. In this sense, it may be thought to require less resource. As Van De Vrande et al. (2009) observe, the most common open innovation practices in smaller firms include customer involvement and external networking. These are typically ‘informal, unstructured practices which do not necessarily require substantial investments’ (p. 434). Accordingly, one may think of these forms of open innovation as accessible for all firms. However, firms must still have resources to get resources: capable employees and internal innovation efforts (both common measures of absorptive capacity) are strongly associated with external search activities.

As was the case for coupled open innovation, the geography of markets does not appear to influence the extent of openness. However, again, the data suggest that it associates with the likelihood of being open or not. In this case, being active in any international market increases the probability of a non-zero result.

Turning to our appropriation variables, the results are broadly the same as those for coupled open innovation, but differ in one important respect. Again, use of IP protection mechanisms appears to act as a ticket to the game. Increasing use of both informal and formal mechanisms is negatively associated with ‘closedness’ – providing partial support for both H2 and H3. Further support for H2 is provided by the observation that the use of informal protection mechanisms positively associates with the extent of inbound innovation. The use of one additional informal protection mechanism (secrecy, complexity or speed-to-market) is associated with, on average, a 9% higher rate of inbound open innovation. It is clear that the use of strategic methods of protecting IP is strongly related to external search in our sample of SMEs. As Van de Vrande et al. (2009) observe, due to limited resources, smaller firms will be confronted with the boundaries of their firm sooner rather than later. This, in turn, requires them to draw extensively on their networks to find innovation resources. Confidence in their abilities to protect the own IP may encourage broader and more intensive search. Similarly, higher managerial quality (Laursen and Salter, 2014), signalled by a clearly articulated appropriation strategy, may also enable greater inbound open innovation.

There is, however, no evidence of a relationship between formal IP protection and the extent of inbound open innovation. Formal IP protection is not statistically significant related to the number of sources of information used (>1). Taken together with the earlier observations on coupled open innovation, it is tempting to see our results as an extension of West’s (2006) observation that it is easier to see the relationship between IP protection and openness when it is outbound rather than inbound. Yet, an important caveat would be to note the threshold effect, echoing Drechsler and Natter (2012), that a focus on IP helps firms shift from closed to open, although it may not induce greater openness thereafter. In the case of inbound and coupled open innovation, formal IP may indicate more complex or sophisticated innovations, which are less well suited to extensive openness (Almirall and Casadesus-Masanell, 2010).

In summary, in a similar vein to Thomä and Bizer (2013), our results suggest that the extent of innovation cooperation and networking in small firms is strongly related to the kind of appropriation strategies chosen. SMEs emphasising informal, strategic methods of protection record higher rates of both coupled and inbound open innovation. In contrast, the use of formal protection mechanisms has limited association with the extent of either inbound or coupled openness. Crucially, however, a focus on both formal and informal IP protection predicts the likelihood of being open at all.

Concluding remarks

Both value creation and value capture are important elements of successful innovation (Lawson et al., 2012). Openness can stimulate value creation by allowing firms to tap into a larger and more diverse pool of complementary knowledge and related resources (Almirall and Casadesus-Masanell, 2010). This is likely to be particularly important for smaller innovators, who confront the boundaries of their expertise rather sooner than their larger counterparts. Opening up innovation processes may enable SMEs to overcome their liability of smallness, and there is growing evidence that smaller firms are engaging in open innovation practices at an increasing rate (Gassmann et al., 2010; Van De Vrande et al., 2009). Indeed, recent work has suggested that ‘SMEs are more dependent on OI than large companies’ (Spithoven et al., 2013: 555).

Yet, openness entails various extents of revealing; and revealing may limit value capture. In this vein, there is evidence that firms identify loss of knowledge as a key open innovation concern (Enkel et al., 2009). Indeed, Laursen and Salter (2014) identify this as the paradox of openness, and West (2006) suggests that the apparent friction may retard open innovation processes if it discourages information search or cumulative innovation.

To date, there is limited empirical evidence on the relationship between open innovation strategies and their appropriation strategies of SMEs. While recent work has looked at the relationship between openness and appropriation strategies in a sample of small and large firms (Laursen and Salter, 2014), the link between different choices of appropriation mechanisms and openness in SMEs is underexplored. The current study attempted to begin filling this gap.

Drawing upon data from the fifth UK IS, we find that there is little evidence of a relationship between the extent of openness and the use of formal protection mechanisms. Formal protection mechanisms offer legal exclusion rights and are frequently argued to benefit (and even be necessary in) outbound and coupled open innovation modes (Henkel et al., 2013). For our SMEs, and for inbound and coupled open innovation, there is simply no evidence that this is the case. This may reflect a number of factors. For instance, legal exclusion rights are meaningful only to the extent that the firm is both willing and able to prosecute infringement. Smaller firms rarely have the resources necessary to credibly threaten litigation (West, 2006). In this way, owning extensive patent portfolios, for instance, is unlikely to either deter or encourage open innovation. Of course, other factors may also be important. For instance, while networks and collaboration allow smaller firms to mitigate resource and knowledge constraints, a declared focus on formal IPRs may signal potentially complex negotiations or reluctant reciprocity to potential partners. It may also be that possession of valuable IP leads some firms to be more inward-looking (Lichtenthaler, 2010). If technology rather than market advantages are central to competitive advantage, such firms may be reluctant to rely on technology that can only be developed with external assistance.

This is not to say that there is no relationship between appropriation strategies and open innovation. Rather, our data suggest that both forms of IP protection significantly distinguish between open and closed firms. Moreover, the use of informal mechanisms to protect a firm’s IP is strongly associated with the extent of inbound open innovation in small firms. Even in small firms, managers can make their firms more open by engaging in innovation-related cooperation or by drawing broadly from a network of actors. Those that are able to identify a clear strategy for protecting their IP seem better placed to engage in open innovation. This may reflect either managerial quality or a more specific confidence in the firm’s ability to appropriate value – and the relative valence of these explanations remains an open empirical question. Regardless, appropriating the benefits from an innovation requires significant managerial effort (Laursen and Salter, 2014) and the varying extent of that effort appears to associate strongly with varying modes of innovation (Thomä and Bizer, 2013).

It is also likely that the role of IP strategy may differ depending upon the degree or ambition of firm innovation activities. To that end, an earlier version of the study (results not reported here, but available from the authors on request) estimated the models for novel innovators and incremental innovators separately. Here, novel innovations (in either products or processes) are those that were new to the industry, while incremental innovations were new to the firm only. Our tentative results suggest little difference in the role of IP strategies for inbound open innovation, with only an emphasis on informal protection mechanisms associating with a higher degree of openness in both novel and incremental innovators. In the case of coupled open innovation, there is some evidence that attention to both forms of IP associates with the degree of openness in incremental innovators, in this way providing support for H1. In contrast, there is no evidence of an association between formal IP and degree of openness in novel innovators. Rationalising this finding is difficult. One may speculate that prior formal IP in incremental innovators signals their quality as potential collaborators and provides a ‘ticket to the game’, while, in novel innovators, concerns about the centrality of IP to the current innovation act against the presumed benefits of collaboration. Given our apprehension over interpretability, we do not fully report the results here, but believe that they hint at an important avenue for further research.

In closing, we recognise limitations to our work. In the first instance, the data we draw upon are cross-sectional. Accordingly, we have consciously avoided strong causal statements, although, to some extent, our argument has the flavour of appropriation strategies influencing innovation strategies. One may speculate that the casual mechanisms at play are likely to require richer data to fully elucidate. Technically, our models contain a potential endogeneity grounded in reverse causality: one may plausibly argue that experience with open innovation will shape firms’ attitudes to appropriation. We recognise this possibility and, indeed, believe that it is likely to be the case. However, echoing Laursen and Salter (2014), we believe that any relationship between appropriation efforts and openness is likely to be circular rather than linear. In this way, having access to panel data, while attractive, is unlikely to be conclusive. Rather, in contrast to calls for more large-scale data analysis and fewer case studies, we believe that detailed case studies of SMEs in different technology settings are likely to be required.

In addition, it may be that the relationship between appropriation and open innovation differs according to partner type or information source.⁷ We have treated partners and sources as simply additive. Our goal was to explore the relationship between IP strategies and broad measures of open innovation common in the empirical literature. However, future work may be well served trying to disentangle the relationship between espoused IP strategies of smaller firms and, for instance, cooperation with particular partners – extending the work of (Cassiman and Veugelers, 2002).

The issue of firm size is also not straightforward. Data considerations have constrained us to look at SMEs as a group. The focus on SMEs is consistent with recent empirical work in field specific journals (e.g. Parida et al., 2012; Spithoven et al., 2013; Verbano et al., 2015); however, it seems likely that as resources diminish (with diminishing firm size), the tension between openness and appropriability will become more acute. Studies focusing on exclusively the smallest firms may yield additional insights.

As an aside, the data in the reported correlation matrix indicate a potential complementarity between inbound and coupled open innovation (a Pearson’s r of 0.46). While these appear to be independent constructs, explaining around 21% of the variance in each other, they may be frequently pursued together. The extent of this complementarity is beyond the scope of the current analysis, but may present opportunities for future research.⁸ For instance, one may speculate that the extent of complementarity is related to the nature of the innovation pursued, with more novel innovation projects drawing upon both inbound and coupled innovation, while less novel projects more frequently relying on inbound open innovation alone. Unfortunately, we are constrained to examine them separately here.

Footnotes

Appendix

Appendix 1

Potential sources of information	Potential collaborative partners
Within your enterprise or enterprise group	Clients of customers
Suppliers	Suppliers
Clients or customers	Competitors
Competitors	Within your enterprise or enterprise group
Consultants, commercial labs and private R&D institutes	Consultants, commercial labs and private R&D institutes
Universities or other higher education institute (HEI)	Universities or other HEI
Government or other public research institutes	Government or other public research institutes
Conferences, trade fairs and exhibitions
Scientific journals
Professional and industry associations
Technical, industry or service standards

R&D: research and development.

Acknowledgements

We are grateful for the considered and constructive comments from two anonymous reviewers and from the editor during the development of this manuscript. We are also grateful to the UK Data Archive and to the Virtual Microdata Laboratory at the ONS for access to the data used in this paper. The usual caveat applies.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

Author biographies

Mark Freel holds the Royal Bank of Canada Professorship in the Commercialisation of Innovations in the Telfer School of Management, University of Ottawa. He is also a professor of entrepreneurship in the Department of Entrepreneurship, Strategy and Innovation at Lancaster University Management School.

Paul J Robson, is a professor of entrepreneurship in the School of Management, Royal Holloway, University of London. His research interests are concerned with various aspects of the economics and management of smaller enterprises.

References

Almirall

Casadesus-Masanell

(2010) Open versus closed innovation: A model of discovery and divergence. Academy of Management Review 35(1): 27–47.

Amara

Landry

Traoré

(2008) Managing the protection of innovations in knowledge-intensive business services. Research Policy 37(9): 1530–1547.

Arbussa

Coenders

(2007) Innovation activities, use of appropriation instruments and absorptive capacity: Evidence from Spanish firms. Research Policy 36(10): 1545–1558.

Arundel

(2001) The relative effectiveness of patents and secrecy for appropriation. Research Policy 30(4): 611–624.

Arundel

Smith

(2013) “History of the Community Innovation Survey.” In: Gault

(ed.) Handbook of Innovation Indicators and Measurement. London: Edward Elgar, pp.60–86.

Aslesen

Freel

(2012) Industrial knowledge bases as drivers of open innovation? Industry and Innovation 19(7): 563–584.

Balconi

Brusoni

Orsenigo

(2010) “In Defence of the Linear Model: An Essay.” Research Policy 39(1): 1–13.

Barge-Gil

(2010) Open, semi-open and closed innovators: Towards an explanation of degree of openness. Industry and Innovation 17(6): 577–607.

Bougrain

Haudeville

(2002) “Innovation, Collaboration and SMEs Internal Research Capacities.” Research Policy 31(5): 735–747.

10.

Borgatti

Halgin

(2011) On network theory. Organization Science 22(5): 1168–1181.

11.

Brouwer

Kleinknecht

(1999) Innovative output, and a firm’s propensity to patent.: An exploration of CIS micro data. Research Policy 28(6): 615–624.

12.

Cassiman

Veugelers

(2002) Complementarity in the innovation strategy: Internal R&D, external technology acquisition and cooperation. Available at: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=308601

13.

Chesbrough

(2003a) Open Innovation. Boston, MA: Harvard Business Press.

14.

Chesbrough

(2003b) The era of open innovation. MIT Sloan Management Review 44(3): 35–41.

15.

Classen

Van Gils

Bammens

et al . (2012) Accessing resources from innovation partners: The search breadth of family SMEs. Journal of Small Business Management 50(2): 191–215.

16.

Cohen

Levinthal

(1990) “Absorptive Capacity: A New Perspective on Learning and Innovation.” Administrative Science Quarterly 35(1): 128–152.

17.

Dahlander

Gann

(2010) “How Open Is Innovation?” Research Policy 39(6): 699–709.

18.

Drechsler

Natter

(2012) Understanding a firm’s openness decisions in innovation. Journal of Business Research 65(3): 438–445.

19.

Eisenhardt

Schoonhoven

(1996) “Resource-Based View of Strategic Alliance Formation: Strategic and Social Effects in Entrepreneurial Firms.” Organization Science 7(2): 136–150.

20.

Enkel

Gassmann

Chesbrough

(2009) Open R&D and open innovation: Exploring the phenomenon. R&D Management 39(4): 311–316.

21.

Fritsch

Lukas

(2001) “Who Cooperates on R&D?” Research Policy 30(2): 297–312.

22.

Garriga

Von Krogh

Spaeth

(2013) How constraints and knowledge impact open innovation. Strategic Management Journal. Available at: http://onlinelibrary.wiley.com/doi/10.1002/smj.2049/full

23.

Gassmann

Enkel

(2004) Towards a theory of open innovation: Three core process archetypes. In: R&D management conference, Lisabon, 6 July, pp.1–18. Available at: https://www.alexandria.unisg.ch/274/

24.

Gassmann

Enkel

Chesbrough

(2010) The future of open innovation. R&D Management 40(3): 213–221.

25.

Geroski

(1995) Markets for technology: Knowledge, innovation and appropriability. In: Stoneman

(ed.) Handbook of the Economics of Innovation and Technological Change. Oxford: Blackwell Publishers, pp.90–131.

26.

Giuliani

(2005) Cluster absorptive capacity why do some clusters forge ahead and others lag behind? European Urban and Regional Studies 12(3): 269–288.

27.

Glückler

(2013) Knowledge, networks and space: Connectivity and the problem of non-interactive learning. Regional Studies 47(6): 880–894.

28.

Hall

Helmers

Rogers

M, et al.

(2011) The choice between formal and informal intellectual property: A review. Available at: http://eml.berkeley.edu/~bhhall/papers/Halletal14_formal_informal_IP_JEL.pdf

29.

Henkel

Schöberl

Alexy

(2013) The emergence of openness: How and why firms adopt selective revealing in open innovation. Research Policy 43: 879–890. Available at: http://www.sciencedirect.com/science/article/pii/S0048733313001595

30.

Huizingh

EKRE

(2011) Open innovation: State of the art and future perspectives. Technovation 31(1): 2–9.

31.

Hurmelinna-Laukkanen

Puumalainen

(2007) Nature and dynamics of appropriability: Strategies for appropriating returns on innovation. R&D Management 37(2): 95–112.

32.

Laursen

Salter

(2006) Open for innovation: The role of openness in explaining innovation performance among UK manufacturing firms. Strategic Management Journal 27(2): 131–150.

33.

Laursen

Salter

(2014) The paradox of openness: Appropriability, external search and collaboration. Research Policy 43: 867–878. Available at: http://www.sciencedirect.com/science/article/pii/S0048733313001832

34.

Lawson

Samson

Roden

(2012) Appropriating the value from innovation: Inimitability and the effectiveness of isolating mechanisms. R&D Management 42(5): 420–434.

35.

Leiponen

Byma

(2009) If you cannot block, you better run: Small firms, cooperative innovation, and appropriation strategies. Research Policy 38(9): 1478–1488.

36.

Leiponen

Helfat

(2010) Innovation objectives, knowledge sources, and the benefits of breadth. Strategic Management Journal 31(2): 224–236.

37.

Lichtenthaler

(2010) Intellectual property and open innovation: An empirical analysis. International Journal of Technology Management 52(3): 372–391.

38.

Love

Roper

Vahter

(2013) Learning from openness: The dynamics of breadth in external innovation linkages. Strategic Management Journal 35: 1703–1716. Available at: http://onlinelibrary.wiley.com/doi/10.1002/smj.2170/full

39.

Mina

Bascavusoglu-Moreau

Hughes

(2014) Open service innovation and the firm’s search for external knowledge. Research Policy 43: 853–866. Available at: http://www.sciencedirect.com/science/article/pii/S0048733313001200

40.

Nieto

Santamaría

(2007) “The Importance of Diverse Collaborative Networks for the Novelty of Product Innovation.” Technovation 27(6): 367–377.

41.

Oakey

(2013) Open innovation and its relevance to industrial research and development: The case of high-technology small firms. International Small Business Journal 31(3): 319–336.

42.

OECD (2005) The Oslo Manual: Proposed Guidelines for Collecting and Interpreting Innovation Data, OECD, Paris.

43.

Parida

Westerberg

Frishammar

(2012) Inbound open innovation activities in high-tech SMEs: The impact on innovation performance. Journal of Small Business Management 50(2): 283–309.

44.

Pavitt

(1984) Sectoral patterns of technical change: Towards a taxonomy and a theory. Research Policy 13(6): 343–373.

45.

Pisano

(2006) Profiting from innovation and the intellectual property revolution. Research Policy 35(8): 1122–1130.

46.

Pittaway

Robertson

Munir

et al . (2004) Networking and innovation: A systematic review of the evidence. International Journal of Management Reviews 5(3–4): 137–168.

47.

Rothwell

(1991) External networking and innovation in small and medium-sized manufacturing firms in Europe. Technovation 11(2): 93–112.

48.

Spithoven

Clarysse

Knockaert

(2011) Building absorptive capacity to organise inbound open innovation in traditional industries. Technovation 31(1): 10–21.

49.

Spithoven

Vanhaverbeke

Roijakkers

(2013) Open innovation practices in SMEs and large enterprises. Small Business Economics 41(3): 537–562.

50.

Teece

(1986) Profiting from technological innovation: Implications for integration, collaboration, licensing and public policy. Research Policy 15(6): 285–305.

51.

Tether

(2002) “Who Co-Operates for Innovation, and Why:: An Empirical Analysis.” Research Policy 31(6): 947–967.

52.

Tether

Tajar

(2008) “Beyond Industry-University Links: Sourcing Knowledge for Innovation from Consultants, Private Research Organisations and the Public Science-Base.” Research Policy 37(6–7): 1079–1095.

53.

Thomä

Bizer

(2013) To protect or not to protect? Modes of appropriability in the small enterprise sector. Research Policy 42(1): 35–49.

54.

Trott

Hartmann

(2009) Why ‘open innovation’ is old wine in new bottles. International Journal of Innovation Management 13(4): 715–736.

55.

Van

Vrande

De Jong

JPJ

Vanhaverbeke

et al . (2009) Open innovation in SMEs: Trends, motives and management challenges. Technovation 29(6): 423–437.

56.

Verbano

Crema

Venturini

(2015) The identification and characterization of open innovation profiles in Italian small and medium-sized enterprises. Journal of Small Business Management 53(4): 1052–1075.

57.

West

(2006) Does appropriability enable or retard open innovation. In: Chesbrough

Vanhaverbeke

West

(eds) Open Innovation: Researching a New Paradigm. Oxford: Oxford University Press, pp.109–133.

58.

Wynarczyk

Piperopoulos

McAdam

(2013) Open innovation in small and medium-sized enterprises: An overview. International Small Business Journal 31: 240–255.

59.

Xia

Roper

(2008) From capability to connectivity – Absorptive capacity and exploratory alliances in biopharmaceutical firms: A US–Europe comparison. Technovation 28(11): 776–785.