Master Thesis MSc BA – Strategic Innovation Management
Lennert Schuttenbeld S2378795
l.j.schuttenbeld@student.rug.nl
University of Groningen Faculty of Economics and Business
Supervisor: Dr. Thijs Broekhuizen Co-assessor: Dr. Killian McCarthy
24th of June 2019 Word count: 9190
2
ABSTRACT
To improve environmental innovation performance, firms increasingly rely on external knowledge sources. This thesis examines the effects of collaboration breadth and collaboration depth on a firm’s environmental innovation performance. Innovation literature shows collaboration breadth and depth to have a significant impact on a firm’s overall innovativeness, but empirical evidence for this relationship is lacking for environmental innovation performance. Based on absorptive capacity theory, I propose a firm’s internal R&D to positively influence these relationships. Using survey data of 218 Dutch SMEs, I test the conceptual model and find important implications for research on eco-innovations and SMEs. The results show that both collaboration breadth and depth have a positive effect on a firm’s environmental innovation performance – with the first having a stronger impact – but no evidence is found for a moderating effect of a firm’s internal R&D.
Keywords: SME, environmental innovation, eco-innovation, collaboration breadth, collaboration depth, internal R&D, absorptive capacity
3
TABLE OF CONTENT
1. INTRODUCTION... 4
2. THEORETICAL BACKGROUND AND HYPOTHESES ... 6
2.1 Conceptual Framework ... 6
2.2 Environmental Innovation ... 7
2.3 Collaboration as Antecedents of Environmental Innovation Performance ... 8
2.4 Moderating Influence of Internal R&D ... 11
3. METHODOLOGY ... 12
3.1 Data and Sample... 12
3.2 Measurements... 13
3.2.1 Dependent Variable ... 13
3.2.2 Independent Variables ... 13
3.2.3 Moderating Variable ... 14
3.2.4 Control Variables ... 14
4. RESULTS ... 15
4.1 Descriptive Statistics and Correlations ... 15
4.2 Hypotheses Testing ... 17
4.3 Robustness Checks and Additional Tests... 18
5. CONCLUSION AND DISCUSSION ... 19
5.1 Conclusion ... 19
5.2 Theoretical Implications ... 20
5.3 Managerial Implications ... 21
5.4 Limitations and Future Research... 21
6. REFERENCES ... 23
4
1. INTRODUCTION
Ever since the WCED, in 1987, gave the first clear definition of sustainable development ‘seeking to meet the needs and aspirations of the present, without compromising the ability to meet those of the future’, much scholarly work has been published on sustainable innovation (Bos-Brouwers, 2010; Elkington, 2007). However, much of this scholarly work focuses on large companies and neglects efforts made by small- and medium sized enterprises (SMEs) (Bos-Brouwers, 2010, Uhlaner et al., 2012).
This tendency to focus on large firms is understandable, as the environmental impact of those firms is the most visible and easiest to identify (Brammer, Hoejmose and Marchant, 2012). It is also important though to research environmental or ecological innovations (short: eco-innovation) in an SME context. It is acknowledged that SMEs can significantly influence the environment (Gadenne, Kennedy and McKeiver, 2009), as in many countries they represent more than 95% of all enterprises (OECD, 2000). A large body of literature recognizes significant differences between the innovation process of SMEs and large firms (Tether, 1998), stressing that SMEs have different values and norms, organizational structures and access to resources (Uhlaner et al., 2012).
Hence, recommended practices prescribed for large companies do not easily translate to an SME context (Jenkins, 2006; Williamson, Lynch-Wood and Ramsay, 2006).
Compared to large firms, SMEs face considerable resource constraints and often do not possess the knowledge and human capital required for innovation (Nieto and Santamaría, 2010; Rogers, 2004). In line with the resource-based view, which considers a firm’s resources as a key aspect for innovation and sustained competitive advantage, collaborations are regarded to positively influence a firm’s innovation performance (Faems, Van Looy and Debackere, 2005). External knowledge sources serve as an important innovation input for SMEs to overcome the lack of internal resources (Nieto and Santamaría, 2010; Rogers, 2004). By pooling knowledge resources and sharing risks, SMEs are able to innovate and stay competitive (Love and Roper, 2015).
These arguments also apply to eco-innovations, as collaborating with external partners provides SMEs with the often specific knowledge and skills required for such innovations and reduces the increased risk and uncertainty inhibiting eco-innovations (Bos-Brouwers, 2010; De Marchi, 2012; Klewitz and Hansen, 2014).
When considering collaboration, past research has demonstrated that a firm’s collaboration breadth (the number of different external collaborations) and collaboration depth (the intensity of a collaborative relationship) are influential in shaping a firm’s innovative performance (Katila and Ahuja, 2002; Laursen and Salter, 2006). The way firms externally search for knowledge and skills, having more ‘broad’ or ‘deep’ relationships, can significantly impact innovation outcomes (Flor, Cooper and Oltra, 2018; Laursen and Salter, 2006).
Despite wide recognition for the positive effect of collaboration on innovation outcomes for SMEs in non-environmental innovation literature (Cassiman and Veugelers, 2006; Cohen and Levinthal, 1990), research on the influence of collaboration breadth and depth on an SME’s environmental innovation performance - measured in the number of eco-innovations - is still lacking. Determinants for eco-innovations can differ considerably from general innovations (Horbach, 2008), as they often involve more complexity (Brammer et al., 2012), lack real market demand (Beise and Rennings, 2005) and face higher institutional and regulatory pressures (Rennings, 2000). Therefore, the aim of this paper is to shed some light on this gap in literature by researching the effects of collaboration breadth and depth on a firm’s environmental innovation performance.
5
A firm’s internal R&D is assumed to moderate these relationships of collaboration breadth and depth with environmental innovation performance. Based on absorptive capacity theory, internal R&D does not only generate new knowledge but also increases the marginal return to external knowledge sourcing strategies (Cassiman and Veugelers, 2006). It explains why some firms are better at taking advantage of external knowledge sources than others (Ferreras-Méndez, Fernández-Mesa and Alegre, 2016). Previous innovation research shows that a firm’s ability to absorb knowledge largely explains why there are significant differences in the extent to which firms can benefit from greater collaboration breadth and depth (Cruz-González et al, 2015; Ferreras-Méndez et al., 2015;
Flor et al., 2018). As eco-innovations are knowledge intensive, requiring much outside knowledge (De Marchi, 2012; Husted and De Sousa-Filho, 2017), a firm’s internal R&D likely influences the assimilation of externally sourced knowledge. Hence, besides researching the direct effect of collaboration breadth and depth on a firm’s environmental innovation performance, this research also focuses on the possible moderating effect of a firm’s internal R&D on these relationships.
To test the hypothesized effects, I use an existing survey dataset from the Innovation Benchmark North Netherlands of the year 2017, which encompasses 218 unique SMEs. Empirical findings indicate collaboration breadth and depth to both have a positive effect on a firm’s environmental innovation performance – with breadth having a stronger impact than depth. No support is however present for the proposed interaction effect of internal R&D, which implies that a firm’s internal R&D does not have an effect on the strengths of the relationship between collaboration breadth and depth with environmental innovation performance.
Given that the field of environmental innovation is currently relatively underdeveloped with respect to SMEs (Klewitz and Hansen, 2014), this study contributes to literature on eco-innovations and SMEs in two ways.
First, it is one of the first studies to research the differential effects of collaboration breadth and depth on a firm’s environmental innovation performance. Showing similar results to research on general innovations, it confirms the importance of seeking a diversity of partners and maintaining in-depth relationships with them in a new setting.
Remarkably, the effect of collaboration breadth is more pronounced, which suggests that having a wide diversity of partners is more effective in improving a firm’s eco-innovation performance. It can be argued that SMEs context of this study helps to explain this finding, as literature suggests that SMEs tend to focus more on incremental innovations (Oke, Burke and Myers, 2007; Woschke, Haas and Kratzer, 2015). The additional test performed also demonstrates the increased attention of SMEs on incremental eco-innovations. This indicates a confirmation of the findings of Laursen and Salter (2006), who suggested collaboration breadth to be more effective for incremental innovations. Secondly, this study reveals that the environmental setting is less affected by a firm’s degree of absorptive capacity. Contrary to findings in non-environmental innovation studies, a firm’s internal R&D does not seem to influence these relationships. This raises the question of whether complementary effects of internal R&D and external knowledge sourcing, found in non-eco-innovation studies, also hold for eco-innovations (in an SME context). Additionally, these findings show that the collaboration benefits hold across the board, as it does not matter if a firm is R&D intensive or not; the effects of collaboration breadth and depth on a firm’s eco- innovativeness remain the same.
The paper is structured as follows: in the next section I will present a literature review that provides the theoretical background to ground the hypotheses. The subsequent section will describe the data and methodology of this study, followed by a section which presents the empirical results and additional robustness checks.
6
Finally, I will conclude this study and discuss the theoretical and managerial implications as well as limitations and suggestions for future research.
2. THEORETICAL BACKGROUND AND HYPOTHESES
2.1 Conceptual Framework
This study intends to explain environmental innovation performance by introducing the concepts of collaboration breadth (H1) and depth (H2). Based on the resource-based view, I expect broader (scope of partners) or deep (collaboration intensity) relationships to provide a firm with the additional resources required to improve a firm’s environmental innovation performance. A firm’s realized environmental innovation output is taken as a proxy to represent environmental innovation performance, as many other regularly used proxies (e.g. eco-patents or environmental R&D) fail to include ‘unintentional environmental innovations’ (Arundel and Kemp, 2009). In addition, based on absorptive capacity literature, I expect a firm’s internal R&D to positively moderate the effects of collaboration breadth (H3) and depth (H4) on a firm’s environmental innovation performance, as a firm needs some internal knowledge to be able to assimilate and use externally sourced knowledge (Cassiman and Veugelers, 2006).
Figure 1 - Conceptual framework
7 2.2 Environmental Innovation
Environmental or ecological innovations (short: eco-innovations) are defined as: ‘innovations which represent new or enhanced processes, organizational forms, as well as products or technologies that are beneficial to the environment in that they reduce or avoid negative environmental impacts’ (Klewitz and Hansen, 2014 p. 58). In line with traditional understanding of innovation as defined by the OECD (1997) it distinguishes between process, product and organizational innovations, but adds direction by including ecological aspects as additional innovation targets (Klewitz and Hansen, 2014). It is important to clearly define eco-innovations, since many related definitions such as sustainability-oriented innovation (Adams et al., 2016) or CSR-driven innovation (Mishra, 2017) exist, which incorporate the environmental dimension, but also social aspects. To clarify, this study’s interest lies solely on the environmental dimension, following the definition of Klewitz and Hansen (2014).
To measure environmental innovation performance, many studies use environmental R&D or eco-patents as proxies (Jaffe and Palmer, 1997; Nameroff, Garant and Albert, 2004). However, these measurements receive much criticism as they are not specifically designed to measure eco-innovation performance and are extremely limited in scope (Arundel and Kemp, 2009). Many organizational or process innovations do not lead to patent introductions, whereby using eco-patents as a proxy insufficiently captures a firm’s environmental innovativeness (Jaffe and Palmer, 1997; Arundel and Kemp, 2009). Furthermore, although it does not belong to its primary objectives, non-environmental innovations can also produce environmental benefits (Horbach and Rennings, 2013). These ecological side-effects of general innovations are not measured when using environmental R&D or eco-patents as proxies (Arundel and Kemp, 2009). Therefore, to fully capture a firm’s environmental innovation performance, this study considers an innovation as an eco-innovation when it ‘creates environmental benefits compared to existing alternatives’ (De Marchi and Grandinetti, 2015, p. 573). By including this distinction to the environmental innovation definition, it incorporates all and only innovations which produced some sort of environmental benefits, also innovations without a specific ecological target. This is in line with the Community Innovation Survey (2008).
A comparison between eco-innovations and general innovations reveals some important differences, relating for instance to the so called ‘double-externality problem’ (Rennings, 2000). This double externality problem makes it somewhat hard to generate eco-innovations, due to increased uncertainty and risk (De Marchi, 2012; Hall and Helmers, 2013). Besides the usual knowledge externalities generated by R&D activities - the creation of knowledge spillovers which benefits other firms and prevents a firm to fully appropriate the created value - eco-innovators can also generate environmental specific externalities (De Marchi, 2012). Society could partly appropriate the value created by eco-innovators, as it enables consumers to reduce its environmental damage, without having to bear the costs which are incurred by the eco-innovating firms. To charge a premium for these additional costs though is generally not an option, as consumers are often unwilling to pay extra for environmentally friendly products (Brammer el al., 2012; Rennings 2000).
In addition, the relative inexperience of firms regarding eco-innovations prevents them seeing the economic potential of eco-innovations (Horbach, 2008). Many companies often fail to realize that eco-innovations may be costly in the short run, but can enable them to reap long-term benefits (Horbach, 2008; Triguero, Moreno- Mondéjar and Davia, 2013). Introducing eco-innovations can lead to ‘win-win’ situations, as the reduction of environmental pollution often goes hand in hand with a decreased need for production inputs, generating cost-
8
savings (Gadenne et al., 2009; Triguero et al., 2013). Nevertheless, as firms often use the rule of thumb principle in their innovation approach, they tend to miss out on the potential of eco-innovations (Horbach, 2008). This particularly concerns SMEs, as they often have limited knowledge on environmental management, making it difficult to estimate costs and benefits of eco-innovations (Brammer et al., 2012).
Due to this double externality problem and the inability of companies to see the potential of eco- innovations, firms frequently under-invest in environmental R&D and innovation (Horbach, Rammer and Rennings, 2012). Therefore, policy interventions and regulatory frameworks are of greater importance in stimulating eco-innovations (De Marchi, 2012; Rennings, 2000). Rennings (2000) refers to this as the ‘regulatory push-pull’ effect, as technology push and market pull factors do not seem to be sufficient alone to encourage eco- innovations (Rennings, 2000). As SMEs often possess less knowledge on environmental management, the influence of this ‘regulatory push-pull’ effect is emphasized in a SME context (Gadenne et al., 2009). Legislation acts as guideline to SMEs, clearly stating what is required of them (Gadenne et al., 2009).
Furthermore, De Marchi (2012) suggests eco-innovations to be more systemic, complex and radical.
Because firms are still relatively inexperienced in developing eco-innovations, they involve a lot of risk and uncertainty and often require knowledge more distant from a firm’s traditional knowledge pool (De Marchi, 2012;
Husted and De Sousa-Filho, 2017).This increased risk and complexity make it especially difficult for SMEs to improve their environmental innovation performance, as they require additional resources which SMEs are often unable to deliver (Biondi and Iraldo, 2002). Compared to large firms, SMEs lack financial resources and often cannot provide the required ex ante investment capital necessary to develop eco-innovations (Biondi and Iraldo, 2002). SMEs generally do not possess the more specific, expert knowledge necessary to develop eco-innovations (Brammer et al., 2012) and they often face human resource constraints as eco-innovations require increased time and effort (Del Brío and Junquera, 2003).
Finally, as environmental issues are global problems, stretching beyond the scope of smaller firms (Johannson, 1997), SMEs generally think they have little environmental impact (Brammer et al., 2012). This lower salience of environmental issues reduces their need to introduce environmental innovations (Gadenne et al., 2009).
2.3 Collaboration as Antecedents of Environmental Innovation Performance
The stream of research on open innovation has emphasized the importance of external knowledge sources for a firm’s innovation process (Lichtenhaler, 2011). As Chesbrough (2006, p.1) termed it ‘open innovation is the purposive inflows and outflows of knowledge to accelerate internal innovation, and expand the markets for external use of innovation, respectively’. Due to the difficulty of controlling proprietary ideas and expertise (Chesbrough, 2003) and in order to reduce uncertainty (Cainelli, De Marchi and Grandinetti, 2015), firms increasingly interact with their environment (Lichtenhaler, 2011). Firms rarely innovate on their own but tend to bond together in teams and coalitions to improve innovation (Laursen and Salter, 2006).
Due to the more systemic and radical character of eco-innovations, collaboration is considered to be especially important for the introduction of eco-innovations. It is seen as a key mechanism to understand which firms are more likely to succeed in introducing these eco-innovations (De Marchi, 2012; Klewitz and Hansen, 2014). To generate such complex and radical eco-innovations, firms often need knowledge and skills located outside the boundaries of the firm (De Marchi, 2012; Husted and De Sousa-Filho, 2017). Therefore, besides
9
depending on a firm’s existing knowledge base, a firm’s capacity to innovate also depends on its ability to access and use external sources of knowledge (Ferreras-Méndez et al., 2015), which is in line with the resource-based view. By collaborating with external partners firms access new and specific knowledge and skills enabling them to create new, unique resources (Das and Teng, 2000), necessary to develop eco-innovations.
In an SME context, collaborations are particularly important to develop eco-innovations. Compared to large firms, SMEs are usually at a disadvantage when it comes to having access to critical knowledge and skills (Faems et al., 2005; Nieto and Santamaría; 2010). Due to these resource constraints, SMEs are often unable to internalize all elements of the innovation process (Biondi and Iraldo, 2002). By collaborating with external partners, SMEs gain access to complex (complementary) technologies and technological expertise (Bos-Brouwers, 2010). As eco-innovations are more complex, they require more specific resources and capabilities increasing the incentive for SMEs to collaborate (Bos-Brouwers, 2010; Brammer et al., 2012; De Marchi, 2012).
Furthermore, because of this more systemic, complex and radical character, eco-innovations involve technologies in which firms are generally still inexperienced, making them a risky, uncertain business (DeMarchi and Grandinetti, 2015). The double externality problem and absence of widespread accepted measures to evaluate environmental performance contributes to even more uncertainty (De Marchi and Grandinetti, 2015). To reduce these uncertainties involving eco-innovations, collaborations are a suitable and often used strategy (Klewitz and Hansen, 2014). Instead of bearing all risks and uncertainties of eco-innovations alone, collaborating with external partners can help spread the risks and costs of eco-innovations and help establish industry standards (Fischer and Pasucci, 2017), increasing the incentive for firms to invest in eco-innovations (Bos-Brouwers, 2010).
Based on these arguments, I hypothesize for collaboration breadth and depth to have a positive effect on a firm’s environmental innovation performance. In general innovation literature, the effects of collaboration breadth and depth have been empirically researched and found to significantly affect a firm’s innovation performance1 (Ferreras-Méndez, 2015; Laursen and Salter, 2006; Leiponen and Helfat, 2010) While collaboration breadth is defined as ‘the number of external sources or search channels that firms rely upon in their innovative activities’, collaboration depth is defined as ‘the extent to which firms draw deeply from the different external sources or search channels’ (Laursen and Salter, 2006, p. 134 - 135).
1 Existing literature has also suggested and found curvilinear (inverted-U) relationships for collaboration breadth and depth and innovation performance, suggesting that their effects to become disadvantageous to a firm’s innovation performance after a certain threshold (Kobarg, Stumpf-Wollersheim and Welpe, 2019; Laursen and Salter, 2006). Most of these studies take ‘the fraction of the firm’s turnover from new products’ (Laursen and Salter, 2006, p. 140) to measure innovation performance, and find diminishing returns to collaboration efforts. In contrast, this study only looks at the environmental performance by looking at the number of types of eco-innovations realized and does not assume decreasing returns. A firm might naturally experience financially decreasing returns, but still reap valuable learning outcomes from greater depth and breadth of collaboration that help to realize the introduction of more and different types of innovations (Davila, Epstein and Shelton, 2012). For comprehension purposes I do test for this non-linear relationship as an additional analysis (see robustness checks and additional tests).
10
Arguments explaining the positive influence of collaboration breadth on a firm’s innovation performance mainly focuses on the economies of scale linked to accessing a large number of different external knowledge sources. Accessing a greater number of external knowledge sources increases the likelihood of firms obtaining knowledge that will lead to successful innovation (Leiponen and Helfat, 2010). Drawing upon a greater number of knowledge sources enriches a firm’s knowledge pool, in that it increases a firm’s choice options to solve problems (Katila and Ahuja, 2002). Therefore, by increasing a firm’s collaboration breadth it increases the probability to access the valuable knowledge needed for eco-innovations, reducing the increased uncertainty inherent in eco-innovations.
As eco-innovations often involve new technologies (Horbach et al., 2012) and require creating new markets (Beise and Rennings, 2005) building legitimacy is essential for eco-innovating firms (Kishna et al., 2017).
To enhance legitimacy, collaborations can serve as a powerful tool (Eisenhardt and Schoonhoven, 1996). When many actors perceive the technology as beneficial, this signals enhanced status to possible future users increasing the chances of successful diffusion (Eisenhardt and Schoonhoven, 1996; Kishna et al., 2017).
Also, as many firms, especially SMEs, are still unable to see the ‘win-win’ potential of eco-innovations (Gadenne et al. 2009; Triguero et al., 2013), gaining access to a broader knowledge base enables firms to find new market opportunities (Flor et al., 2018). Besides an increased awareness of environmental problems, collaborating with a broader range of external partners enables firms to recognize the opportunities that come along with these environmental problems (Cainelli, et al., 2015).
Furthermore, since most innovations are the result of knowledge recombination (Kaplan and Vakili, 2014) the number of innovations which can be created with the same pool of knowledge is limited (Katila and Ahuja, 2002). By adding new sources of knowledge to the set, it increases the chances of finding new, useful recombinations (Katila and Ahuja, 2002). This is especially important to eco-innovations as they require knowledge more distant from a firm’s traditional pool of knowledge (De Marchi, 2012; Husted and De Sousa- Filho, 2017).
In addition, just as individuals, firms tend to focus and search narrowly for new knowledge (Leiponen and Helfat, 2010). Based on social psychology, decision heuristics such as ‘availability heuristic’ and
‘representative heuristic’ are also likely to influence firms (Nooteboom, 2004). This narrows the cognitive scope of a firm, increasing the risks of myopia (Nooteboom, 1992). By enriching a firm’s knowledge pool with more, new external knowledge sources, it increases the cognitive scope of the firm, reducing the risks of myopia (Nooteboom, 1992). Therefore, I hypothesize the following:
Hypothesis 1: Collaboration breadth positively influences a firm’s environmental innovation performance.
When looking at collaboration depth, arguments explaining a positive relationship with a firm’s innovation performance mainly build on the concepts of trust and experience (Flor et al., 2018; Laursen and Salter, 2006). By drawing deeply from an external knowledge source, and by making repeated use of this source, a stable platform for collaboration is created, whereby a firm builds confidence in a partner’s reliability and capability to deliver.
Based on the social exchange theory, managers prefer to interact and communicate on the basis of trust (Nooteboom, Berger and Noorderhaven, 1997). Repeated interactions increase partners’ shared understanding and
11
leads to the forming of habits (Laursen and Salter, 2006; Nooteboom et al., 1997), also reducing uncertainty (Jones and George, 1998) which is especially present in eco-innovations (De Marchi, 2012). Trust and confidence lower the barriers for knowledge protection and foster the sharing of tacit knowledge (Hardwick, Anderson and Cruickshank 2013), stimulating cooperative behavior (Jones and George, 1998). This increased interaction leads to significantly deeper understanding of the given knowledge elements (Katila and Ahuja, 2002). It enables a firm to find new combinations and connections between knowledge elements which were not apparent before, increasing a firm’s chances to successfully develop innovations (Katila and Ahuja, 2002). This is especially important for the development of eco-innovations given their complex and radical nature (Brammer et al., 2012).
As more radical innovations involve knowledge and skills which are either new or not well-understood by a firm (Flor et al., 2018), intense collaboration can help to gather such knowledge, as deeper relationships allow for the exchange of non-superficial and deep knowledge. While collaboration breadth results in more superficial knowledge, intense relationships enable firms to deepen the understanding of the knowledge possessed by collaboration partners and access the expert know-how necessary to develop radical innovations (Cruz-González et al., 2015; Flor et al., 2018). In sum, these arguments suggest the following:
Hypothesis 2: Collaboration depth positively influences a firm’s environmental innovation performance.
2.4 Moderating Influence of Internal R&D
To effectively utilize the know-how from external knowledge sourcing strategies, a firm needs to already possess a certain level of internal expertise (Cassiman and Veugelers, 2006; Cohen and Levinthal, 1990). This dependency between internal and external knowledge is an important part of the absorptive-capacity theory introduced by Cohen and Levinthal (1990) who stress that a firm’s internal knowledge is an important aspect enabling a firm to
‘effectively scan, screen, and absorb external know-how’ (Cassiman and Veugelers, 2006, p. 68). As an important indicator of absorptive capacity, internal R&D not only generates new knowledge, but also increases the innovative returns from external knowledge sourcing strategies (Cassiman and Veugelers, 2006; Cohen and Levinthal, 1990).
In non-environmental innovation studies, the strengthening effect of absorptive capacity has been shown to exist for broad and deep collaboration networks (Flor et al., 2018). Firms collaborating with a diverse, broad range of different partners often incur high external search and management costs (Cruz-Gonzalez et al., 2015).
To become more selective and make sense out of the diverse partners in broad networks, a firm’s absorptive capacity can help to find the most promising collaborations to complement a firm’s core activities, reducing the costs from a broad external knowledge search (Flor et al, 2018). Concerning collaboration depth, greater internal knowledge and absorptive capacity (as gained from internal R&D) may help to decode and integrate the more specific, expert knowledge gained from deep collaborations (Cruz-Gonzalez et al., 2015; Flor et al., 2018).
Therefore, a firm needs a certain level of internal expertise to be able to assimilate and use this specific, expert knowledge (Flor et al., 2018; Lane et al., 2006).
When considering eco-innovations, the concept of absorptive capacity is still relatively underdeveloped.
While Mazzanti and Zoboli (2005, p. 24) indicate a possible complementarity between ‘networking relationships’
and ‘firm structural characteristics’ for the development of eco-innovations, De Marchi (2012) found contradictory results, suggesting that a substitution effect exists between a firm’s internal R&D and external
12
knowledge sourcing for the development of eco-innovations. On what both papers do agree though is that this ‘key new topic’ (Mazzanti and Zoboli, 2005, p. 24) still needs to be ‘further verified’ (De Marchi, 2012, p. 621).
According to Lichtenhaler (2009) the importance of a firm’s absorptive capacity is emphasized in contexts characterized by high technological and market turbulence. When operating in technological and market turbulent environments, firms are facing a high degree of technological change and market uncertainty (Jaworski and Kohli, 1993). Eco-innovations are a good example such technological and market turbulent environments. They are characterized by a relatively new field of technology and often new markets need to be created for environmental products, processes and services (Beise and Rennings, 2005). In such turbulent environments, firms increasingly rely on experience and tap into their prior internal related knowledge (Lichtenhaler, 2009). Therefore, I expect internal R&D to increase the benefits gained from external knowledge sourcing strategies for eco-innovations.
With regard to collaboration breadth, a firm’s internal expertise and experience will allow it to be more selective with whom to collaborate, lowering the costs (Flor et al., 2018) for the development of eco-innovations. Following the logic of absorptive capacity theory, internal R&D will enable a firm to understand the more specific, expert knowledge gathered from deep collaborations (Flor et al., 2018; Lane et al., 2006), which is essential for the development of eco-innovations (De Marchi, 2012). Given the limited availability of resources in SMEs this effect of internal R&D is especially important, as it enables them to make the right choices and avoid wasting valuable resources (Valentim, Lisboa and Franco, 2015). Thus, I hypothesize:
Hypothesis 3: A firm’s internal R&D positively moderates the relationship between collaboration breadth and environmental innovation performance, in such a way that the positive relationship gets stronger with higher levels of internal R&D.
Hypothesis 4: A firm’s internal R&D positively moderates the relationship between collaboration depth and the environmental innovation performance, in such a way that the positive relationship gets stronger with higher levels of internal R&D.
3. METHODOLOGY
3.1 Data and Sample
For this research, I will be using an existing dataset from the Innovation Monitor North Netherlands. This Innovation Monitor is a collaboration between the University of Groningen and SNN, which is a cooperative partnership aimed at stimulating and facilitating economic development for SMEs in the north of the Netherlands.
The Innovation Monitor is an annually conducted research to benchmark and monitor the innovation activities and performances of companies located in the three northern provinces: Drenthe, Groningen and Friesland. It specifically focuses on SMEs and covers innovation topics such as innovation investment, subsidies, creativity, collaboration and disruptive technologies.
The data are collected from a firm-level perspective, and respondents are (mostly) executives of SME firms located in either Drenthe, Groningen or Friesland. More than 5000 SMEs currently cooperating with the SNN are invited by e-mail to participate. To increase the response rate, reminders are sent and non-responding
13
firms are contacted by phone. As the questions included in the Innovation Monitor vary from year to year, I will be only using data from 2017, since this is the only year which properly measured collaboration breadth and depth.
Only firms which indicated that they introduced an innovation during the years 2014-2016 are included in the sample, as these were the only firms which were asked to answer questions about their collaboration for innovation.
Other relevant variables for this sample are: internal R&D, firm size, age, family influence, subsidy and intellectual property protection. By including only observations which provide complete information, I come to a final sample of 218 firms.
3.2 Measurements
3.2.1 Dependent Variable
Environmental Innovation Performance - Environmental innovation performance is tested using a question from the Innovation Benchmark which asks respondents whether or not they have introduced a product, process or organizational innovation during the year 2017 which resulted in one of the following environmental advantages:
(1) lower material use (2) lower energy use (3) lower CO₂ footprint of company (4) replacement of materials with less dangerous substitutes (5) less pollution of soil, water and/or noise (6) recycling of waste, water or other materials. For each item, I create a dummy variable, as the questions could only be answered with either a ‘no’
(score 0) or a ‘yes’ (score 1). I transform the 6 dummy variables into an overall score to represent environmental innovation performance, which ranges from a minimum score of ‘0’ (all 6 items indicated as ‘no’) to a maximum score of ‘6’ (all items indicated as ‘yes’).
3.2.2 Independent Variables
Collaboration Breadth - To measure collaboration breadth, the research of Laursen and Salter (2006) is used as guidance. In the Innovation Monitor, firms are asked to score the importance of each of the 8 possible external knowledge sources (as listed in table 1), during the year 2017 with regard to their innovation activities.
Respondents could answer this question on a 4-point Likert scale ranging from very important (1) to not important (4). In line with Laursen and Salter (2006), each of the 8 sources are then recoded as binary variable, where a ‘0’
represents no use and a ‘1’ represents use of the given knowledge source. After this, the scores are simply added up creating a new variable ‘Breadth’ in which a ‘0’ represents a firm which made no use of external knowledge sources, while an ‘8’ represents a firm which made use of all of the possible external knowledge sources.
Collaboration Depth - Again, the research of Laursen and Salter (2006) is used as guidance to measure collaboration depth. The same question from the Innovation Benchmark to measure collaboration breadth is used to measure collaboration depth. However, in this case each of the 8 possible external knowledge sources are recoded with a ‘1’ when the firm in question reported the knowledge source to be of high importance, while firms which reported the knowledge source to be of average, low or no importance are recoded with a ‘0’. A new variable
‘Depth’ is then created by simply adding up the scores in which a ‘0’ represents a firm which scored none of the possible external knowledge sources to be of high importance, while an ‘8’ represents a firm which scored all of the possible external knowledge sources to be of high importance to the firm.
14
Table 1: External information and knowledge sources for innovation activities of SMEs in Drenthe, Groningen and Friesland. (N = 218)
3.2.3 Moderating Variable
Internal R&D - In order to measure a firm’s internal R&D intensity I use the ratio between a firm’s personnel devoted to carry out R&D activities measured in Full Time Equivalents (FTEs) and a firm’s total personnel (in FTEs).
3.2.4 Control Variables
Firm size – According to research of Uhlaner et al. (2012) firm size can significantly affect, even within the SME context, a firm’s environmental innovativeness, due to resource constraints faced by smaller firms. To provide a stronger test this study will control for firm size by using the number of total employees in FTEs. As the distribution of firm size appears to be positively skewed, I create a natural logarithm of the firm size variable.
Firm age – General and environmental innovation literature indicate a significant relationship between the age of a firm and its (environmental) innovativeness (Jiménez-Jiménez and Sanz Valle, 2011; Wagner, 2007). Most of this literature indicate a positive relationship as older firms are more experienced and able to develop their operations more efficiently (Jiménez-Jiménez and Sanz Valle, 2011). Hence, I control for age. As the firm’s age distribution appears to be positively skewed, I create a natural logarithm for this variable.
Family influence – Much quantitative research indicates a positive, significant relationship between family ownership and environmental performance (Uhlaner et al., 2012). To control for this, I create a dummy variable in which a ‘1’ represents a firm which is family owned, while a ‘0’ represents a firm which is not family owned.
15
Intellectual Property Protection (IP-protection) – As literature indicates that protection mechanisms such as patents or secrecy significantly influence a firm’s innovative behavior (Chesbrough, 2006), I control for IP- protection. To do this, I use a question from the Innovation Benchmark in which firms are asked whether or not it has used one of the following protection mechanisms with regard to its innovations: (1) patents (2) design rights (3) trademarks (4) copyrights (5) secrecy. I create a dummy variable for each item, as the questions could only be answered with either a ‘no’ (score 0) or a ‘yes’ (score 1). I transform the 5 dummy variables into an overall score to represent IP-protection, which ranges from a minimum score of ‘0’ (all 5 items indicated as ‘no’) to a maximum score of ‘5’ (all items indicated as ‘yes’).
Subsidy – Subsidies trigger environmental innovation (Horbach, 2008). As such, I control for firm orientation and success with subsidies by using a question from the Innovation Benchmark which asks whether or not respondents received a so-called ‘VIA-subsidy’ (Accelerator Innovation Ambitions). This subsidy aims to stimulate innovations which contribute to CO₂ reduction, specifically targeting SME firms. To account for this, I create a dummy variable in which ‘1’ represents a firm that did receive a VIA- subsidy and the ‘0’ represents a firm that did not receive a VIA-subsidy.
4. RESULTS
4.1 Descriptive Statistics and Correlations
Table 2 reports the descriptive statistics and correlations of the sample of 218 firms. The mean of the variable environmental innovation performance shows an average firm score of 2.30 out of a total 6. However, the standard deviation of 2.0 indicates a relatively high variance in relation to a firm’s environmental innovation performance.
Another interesting finding is the mean of collaboration breadth which shows that on average firms collaborated, at least to some extent, with 7.14 out of 9 possible partners, which is fairly high. Yet, when looking at the mean of collaboration depth, the average partners with whom firms collaborate deeply is only 1.76. The mean of family influence (0.54) indicates that over half of the firms included in the sample are family owned. Furthermore, for both collaboration breadth (r = 0.290) and collaboration depth (r = 0.149) there is a positive, significant correlation with environmental innovation performance. This provides initial support for the hypotheses regarding the effects of collaboration breadth and depth. In addition, especially for age and internal R&D I find (highly) significant results. Interesting are the negative, significant correlations of firm size (r = -0.429) and age (r = -0.395) with internal R&D. This suggests that larger and older firms have less internal R&D. Because none of the variables highly correlate (r < 0.6), I suspect no multicollinearity issues. But to fully exclude any suspect of multicollinearity, I also conducted a Variance Inflection Factors (VIF) test. The variable firm size shows the highest VIF-value (tolerance = 0.376; VIF = 2.656), which is significantly below the tolerance level of 10, indicated by Hair, Tatham and Anderson (1998). It confirms that there are no issues with multicollinearity.
16
Table 2: Descriptive statistics and correlations
17 4.2 Hypotheses Testing
Table 3 displays the results of the linear regression. I conducted a series of regression analysis, which increased the complexity of the model in each stage.
Model 1 – baseline model – only includes all the control variables and their effect on environmental innovation performance. Except firm size, all control variables show a positive, significant relationship. However, while the variables firm age (β = 0.388; p < 0.05), subsidy (β = 0.769; p < 0.05) and family influence (β = 0.771;
p < 0.01) remain positively significant throughout all models, the variable IP-protection (β = 0.170; p < 0.1) is only significant in the baseline model.
Model 2 – includes the independent variables collaboration breadth and collaboration depth and the moderator variable internal R&D. Supporting hypothesis 1, the results show a positive and significant relationship between collaboration breadth (β = 0.509; p < 0.01) and environmental innovation performance. Also, the results show collaboration depth (β = 0.272; p < 0.1) to significantly and positively influence a firm’s environmental innovation performance, supporting hypothesis 2.
Model 3 – full model – this model includes the interaction term of the moderator internal R&D. It shows that 17.8% of the variance in the dependent variable, environmental innovation performance, can be explained by this model. Despite this, the results show no significant relationships, and hypotheses 3 and 4 are thus not supported.
Table 3: Results of the linear regression for environmental innovation performance.
18 4.3 Robustness Checks and Additional Tests
To strengthen the model and see if it also holds under different conditions, I conducted several robustness checks.
Besides asking firms to score the importance of 8 possible external knowledge sources, a different question in the Innovation Monitor survey asked whether or not firms collaborated with any of the six possible collaboration partners. As this question could only be answered with either ‘yes’ or ‘no’, it does not allow to differentiate between collaboration breadth and depth, but it helps to measure firms’ collaboration in general. I created a measurement similar to the dependent variable, environmental innovation performance, using the sum of the 6 dummy variables to represent a similar scale that ranges from ‘0’ (all 6 items indicated as ‘no’) to a maximum score of ‘6’ (all items indicated as ‘yes’). Using this new scale, the analysis shows no significant changes compared to the original test. The relationship between collaboration and a firm’s environmental innovation performance is still positive and significant while no evidence is found for a moderating effect of internal R&D.
This shows that the model holds under a different measurement for collaboration.
In addition, I take an alternative measurement for the moderating variable, internal R&D. Instead of using the ratio between a between a firm’s personnel devoted to carrying out R&D activities measured in FTEs and a firm’s total personnel (in FTEs), I use the total percentage of turnover spent on internal R&D. As the distribution of this variable appears to be positively skewed, I create a natural logarithm of the total percentage turnover spent on internal R&D. The results are significantly different from the original analysis. While the positive relationship of collaboration breadth remains present, the results show a negative, significant direct effect of a firm’s internal R&D on its environmental innovation performance (β = -0.147; p < 0.1). This could be caused by the generic content of the internal R&D (Cainelli et al., 2012). Eco-innovations need specific environmental R&D (Borghesi, Cainelli and Mazzanti, 2012); conversely, R&D which is not designed for eco-innovations could potentially even slow down the innovation process.
Because of the inability of firms to see the economic potential of eco-innovations (see paragraph 2.2), I also want to test if there is any difference between collaborating for eco-innovations which reap either direct or indirect financial benefits. I combined the first two environmental advantages used to measure the dependent variable (i.e. lower material and lower energy use) to create a new variable direct eco-innovation, represented by a scale that ranges from ‘0’ (both items indicated as ‘no’) to a maximum score of ‘2’ (both items indicated as ‘yes’).
The remaining four environmental advantages used to measure the dependent variable are combined to create the new variable indirect eco-innovations, represented by a scale that ranges from ‘0’ (all 4 items indicated as ‘no’) to a maximum score of ‘4’ (all 4 items indicated as ‘yes’). Separate linear regressions show that collaboration breadth remains significant and positive for both new dependent variables, but collaboration depth is only significant and positive for indirect eco-innovations. This suggests that the specific, expert knowledge acquired from deep collaborations only contributes to the development of eco-innovations that produce indirect financial benefits but not for eco-innovations that produce direct financial benefits. It indicates that the indirect innovations are probably more complex and radical. This is supported by research of Laursen and Salter (2006) which showed for collaboration depth to be positively related with innovation radicalness (Laursen and Salter, 2006).
As prior literature indicates a curvilinear relationship of collaboration breadth and depth with a firm’s innovation performance using financial metrics, I also test for this possible curvilinear relationship with a firm’s
19
environmental innovation performance. To test for this, I add a quadratic term for the independent and moderator variables. Results of this test show no indication of such a curvilinear relationship.
Finally, with regard to the control variable, subsidy, instead of controlling only for the ‘VIA-subsidy’, I also control for the total effect of all subsidies. As this question could only be answered with either ‘yes’ or ‘no’ a measurement similar to the dependent variable, environmental innovation performance, is created. The sum of the nominal variables of all the five possible subsidies is taken to represent a scale ranging from 0 to 5. The analysis shows no significant changes compared to the original model. Hypotheses 1 and 2 thus remain significant and positive, while again no significant effect is found for hypotheses 3 and 4. This adds to the robustness of the findings.
5. CONCLUSION AND DISCUSSION
5.1 Conclusion
On the road to a more sustainable economy, firms increasingly rely on external knowledge sources (Beattie and Smith, 2013; Bocken et al., 2014; Lowitt, 2013). Instead of acting alone, they collaborate with external partners to become more successful in developing environmental innovations. Based on the concepts of collaboration breadth and collaboration depth, the aim of this study was to deepen our understanding of how firms collaborate for environmental innovations. In addition, a great deal of literature on absorptive capacity theory indicates a firm’s internal R&D to be complementary to a firm’s external knowledge sourcing strategy (Cassiman and Veugelers, 2006; Cohen and Levinthal, 1990). I also analyzed, therefore, the possible interaction effects of a firm’s internal R&D on the relationships between collaboration breadth and depth, on the one hand, with a firm’s environmental innovation performance, on the other hand. It specifically focused on SMEs, as although it is acknowledged that SMEs have significant influence on the environment, research in this field is still underdeveloped.
Regarding the direct effects of collaboration breadth and depth, the empirical results confirm the hypotheses. Broad or deep collaborations with external partners result in more environmental innovation output - with the first having a bigger impact. This is in line with the resource-based view and research on open innovation.
Collaborating with external partners enables firms to collect outside ideas and knowledge that increase a firm’s opportunities to develop innovations (Chesbrough, 2003). It can serve as a powerful tool for SMEs to overcome the resource constraints they face compared to large firms (Nieto and Santamaría, 2010; Rogers, 2004). While the advantages of collaboration breadth mainly lie in increased scale of resources accessed (Cainelli et al., 2015; Flor et al., 2018), collaboration depth has a positive effect due to increased partner trust and experience (Flor et al., 2018; Laursen and Salter, 2006). With respect to eco-innovations, broad or deep collaboration helps to overcome the specific difficulties inherent in eco-innovations such as greater complexity (Brammer et al., 2012), lack of market demand (Beise and Rennings, 2005), and higher institutional and regulatory pressures (Gadenne et al., 2009; Rennings, 2000).
Yet, the results show no support for the proposed moderating effects of a firm's internal R&D. This contradicts the complementary effects of internal R&D and external knowledge sourcing, found in non-eco- innovation studies (Cassiman and Veugelers, 2006; Cohen and Levinthal, 1990). It raises the question of whether
20
similar effects hold for eco-innovations. Specifically, because of the SME context of this study, these findings can have important theoretical and managerial implications and offer opportunities for future research.
5.2 Theoretical Implications
The analysis shows the effect of collaboration breadth to be more pronounced, suggesting that diversity of knowledge is more relevant than depth of knowledge when it comes to the realization of eco-innovations. While some research has found collaboration breadth to be more effective for radical innovations (Chiang and Hung, 2010). This study’s findings indicate instead support for the findings of Laursen and Salter (2006), who suggested external search breadth to be more effective for incremental innovations. The SME context of this study helps to explain this indication, as previous research has shown SMEs to focus more on incremental than radical innovations (Oke et al., 2007; Woschke et al., 2015). The bigger impact of collaboration breadth therefore suggests an ‘overrepresentation of incremental eco-innovations’. It seems reasonable to accept this explanation, as most SMEs are unable to deal with the higher uncertainty involved in environmental innovation. They do not possess the required resources necessary to develop radical eco-innovations, but tend to take a follower approach and only implement eco-innovations which have already been proven to be successful (Biondi and Iraldo, 2002; Brammer et al., 2012). This brings a certain nuance to what is stated before about the radicalness of eco-innovations (De Marchi, 2012). Yes, eco-innovations are probably ‘more’ radical than non-environmental innovations, but this study’s findings suggest that SMEs are more likely to realize incremental than radical eco-innovations.
The additional test, which distinguished between direct and indirect effects of eco-innovations on a firm’s financial performance, also indicates support to confirm the findings of Laursen and Salter (2006). Generally, the aim of incremental innovations is to reap short-term benefits, while radical innovations have a more long-term orientation (McDermott; 1999). Therefore, I assume eco-innovations which produce immediate, direct financial benefits for a firm to be considered more as incremental innovations and eco-innovations which have indirect effects, to be considered more as radical innovations. Since the findings indicate both collaboration breadth and depth to relate with eco-innovations which have indirect financial effects, but show no significant effect of collaboration depth on eco-innovations which have a direct financial effect, this suggests that only having a wide diversity of partners collaborations is effective for the development of incremental eco-innovations. Again, this is in line with the findings of Laursen and Salter (2006).
Furthermore, contrary to what would be expected based on absorptive capacity theory, I do not find support for the proposed moderating effect (and direct effect) of internal R&D. This non-finding can most likely be explained by the relative inexperience of firms regarding eco-innovations and the SME focus of this study. It contributes to the assumption, indicated in many other studies that SMEs have limited internal resources (Nieto and Santamaría, 2010; Rogers, 2004; Biondi and Iraldo, 2002) and shows that this likely holds true also for eco- innovations. Firms often do not possess the required specific environmental R&D (Borghesi et al., 2012; Cainelli et al., 2012). Based on absorptive capacity theory, a firm’s internal R&D can increase the marginal returns from external collaboration (Cassiman and Veugelers, 2006). But as this internal R&D for eco-innovations is often lacking, there is no knowledge to combine and connect, which would explain why internal R&D does not moderate the relationship between collaboration breadth and depth and a firm’s environmental innovation performance. The content of the internal R&D which firms do possess is probably too generic (Cainelli et al., 2012), in that it lacks the ability to enhance a firm’s environmental innovation performance (Borghesi et al., 2012). Moreover, this
21
finding raises the question of whether the complementary effects of internal R&D and external knowledge sourcing, found in non-eco-innovation studies (Cassiman and Veugelers, 2006; Cohen and Levinthal, 1990), also hold for eco-innovations (in an SME context). Assuming that the innovation process is highly path dependent, other technological and organizational factors may be far more important in determining an SMEs eco- innovativeness and specific environmental R&D is required to significantly impact its environmental innovation performance (Borghesi et al., 2012; Cainelli et al., 2012).
As the direct effects of collaboration breadth and depth are significant and positive, this shows that external knowledge sourcing is an important way for SMEs to overcome the general absence of environmental internal R&D. This also supports previous research which indicated external knowledge sourcing as an important strategy enabling SMEs to innovate and to stay competitive (Freel, 2005; Love and Roper, 2015). Specifically, this study contributes to literature as it is one of the first to show that both collaboration breadth and depth can be used by SMEs to improve environmental innovation performance and compensate shortcomings. In addition, the non-finding of internal R&D shows that the collaboration benefits hold across the board, as it does not matter if a firm is R&D intensive or not, the effects of collaboration breadth and depth on a firm’s eco-innovativeness remain the same.
5.3 Managerial Implications
This study’s findings suggest firms should actively engage in broad as well as deep collaborations to increase their environmental performance. Because of the non-significant findings regarding the effects of a firm's internal R&D, managers of SMEs should not focus strongly on improving internal knowledge creation (as it does not have any direct effect or moderating effect), but concentrate on acquiring knowledge needed to enhance a firm’s eco- innovativeness from external knowledge sources. This implication is particularly important to SMEs, as these firms often possess less internal knowledge creation resources. Also, what managers should realize is that engaging in either broad or deep collaborations influences the type of knowledge that gets acquired. While broad collaboration leads to mainly superficial knowledge as it focuses on obtaining a high quantity of new knowledge, deep collaboration focuses more on the quality of knowledge, resulting in less, but more expert knowledge (Cruz- González et al., 2015; Flor et al., 2018). Managers should therefore take into account the type of eco-innovation pursued – incremental or radical – when deciding in which collaborations to engage. As, the relationship of the control variables firm age and family influence show a positive, significant relationship with a firm’s eco- innovativeness it would be best to collaborate with older firms, preferably family owned.
For policymakers, these results indicate that broad or deep collaboration should be encouraged. As much of the knowledge required for eco-innovations lies outside the boundaries of a firm’s traditional knowledge base, policy makers should play an active role in enabling firms to access such knowledge (Wagner and Llerena, 2011).
For instance, by organizing networking initiatives to accelerate inter-firm collaboration (Rosenfeld, 1996).
5.4 Limitations and Future Research
This study is subject to several limitations, which in turn offers opportunities for future research. First, generalizability may be an important limitation of this study. As it only focuses on SMEs in the three northern provinces of the Netherlands, it is uncertain if analysis would show the same results in other countries or regions.
