Sincerity is a dangerous thing: On how appropriability regimes shape innovation strategies

Abstract

Since the pioneering work of David J. Teece (“Profiting from technological innovation: Implications for integration, collaboration, licensing and public policy”, Research Policy 15, 1986), the concept of “appropriability regime” – defined as the level of threat stemming from potential imitators in the firm’s environment– has been useful in describing specific sectoral patterns at the aggregate level. However, there have been few attempts to formally prove the impact of different regimes on knowledge management strategies at the firm-level. Using firm-level longitudinal data from software and hardware sectors in Spain, we show that differences in appropriability regimes condition the profitability of open innovation strategies, and also induce specific patterns in firms’ knowledge management strategies. More specifically, we show that the appearance of diminishing returns from an open innovation strategy is more accelerated in firms immersed in weaker regimes (i.e. firms facing more threats from potential imitators). We also show that this is due to the fact that such firms must resort to more complex (and costly) practices that go far beyond the legal protection of intellectual property, internalizing other management practices related to productive capacity and the commercialization of innovations.

Keywords

Innovation knowledge management appropriability regime imitation threat

1 Introduction

Many scholars have studied the effects of the environment on innovation strategies. In the management area, various studies have shown the importance of hostility – defined as the threat to the firm posed by competitive intensity and the speed of technological change– in the shaping of management practices [6 , 25]. As noted in Teece’s [30] seminal contribution, in the case of innovative firms, one of the main sources of “hostility” is the threat stemming from potential imitators. This threat may be explained, at least partially, by environmental characteristics such as the legal and technological difficulties faced by imitators, summarized under the concept of “appropriability regime”. An interesting conclusion of this work is that the weakness or strength of appropriability regimes affect the structure of the company, since it conditions its decisions on integrating or outsourcing the different types of activities leading to the creation and profitable use of new products or processes. Teece’s contribution merged two previously disconnected research lines: innovation and strategic management [27]. In this sense, the appropriation concerns of innovative firms take the shape of special management challenges that absorb time and resources from the firm.

Initially, research on the mechanisms through which firms face appropriation concerns was almost exclusively reduced to intellectual property rights (IPR), including patents, copyrights, trademarks and utility models [11]. However, recent studies have expanded the spectrum of analysis to include other legal mechanisms such as contracts and changes in labour legislation, as well as strategic mechanisms such as secrecy, lead-time, complexity and other complementary resources and capabilities such as manufacturing capacity, control of distribution networks, and relationships with customers and suppliers, among others [5 , 24].

Recently, the intersection between innovation studies and strategic management has led to the emergence of an increasingly popular research agenda: open innovation studies. The literature on open innovation has shown a direct relation between open innovation strategy and firm performance, as a consequence of the effect stemming from access to external resources. This research stream, however, has been characterised by the adoption of a generalist, context-free view on the openness-performance relationship that may be unrealistic given the presence of management costs associated with openness [22]. However, the empirical evidence substantiating this view might be affected by certain methodological choices, since a large part of the work in this area does not explicitly contemplate the possibility of a cost-benefit relation in open innovation strategies [13].

The resurgence of a line of research initiated by Laursen and Salter [19] – but with little attention until recently [2 , 28]– allowed the inclusion of this cost-benefit analysis, acknowledging that appropriation concerns play a very relevant role. In particular, in this paper we are interested in the costs raised by appropriation concerns, since “the creation of innovations often requires openness, but the commercialization of innovations requires protection” [20, p. 867].

Meanwhile, the digital transformation has created new challenges. In particular, the emergence of digital platforms and ecosystems has led to new business models aimed at capturing the value of innovations [31]. This is also the case with intelligent systems, which despite being “intelligent” still require at least the ability to sense the environment, make decisions and control action [1].

In this paper, we analyse how appropriability regimes are helpful to predict specific patterns of knowledge management strategies in innovative firms, proposing an analytical framework where differences in strategies followed by firms are explained by differences in the environments they face, characterized by disparate levels of threats posed by potential imitators. We show that such strategies in weak appropriability regimes, in contrast with stronger ones, are characterized by more complexity (regarding the use of multiple protection mechanisms), and also by decreasing returns from the use of an open innovation strategy.

Our results have practical implications for R&D managers and policy-makers, since they provide a method to visualize the variety of context-dependent mechanisms used by firms to profit from their innovations.

2 Knowledge management strategy and the appropriability problem

Schumpeter [29] and Arrow [4] introduced the “appropriability problem” in the theoretical discussion of innovation, explaining that a certain degree of monopoly might be desirable to foster innovation. This theoretical problem has been typically used to justify the introduction of IP rights to guarantee the appropriation of benefits from innovations. However, as Teece [30] showed, the legal protection of IP is not a perfect strategy in certain contexts, and the appropriation strategy of innovators might be quite complex, involving managerial decisions conditioned by the environment. In particular, innovators’ strategies will be shaped by what he called the “appropriability regime”. This regime refers to environmental factors, different from firm and market structure, that govern an innovator’s ability to capture the profits generated by an innovation. According to Teece, the most important dimensions of this regime are the complexity of the innovation and the efficacy of legal mechanisms of protection. These two dimensions are used to conceptualize the thread posed by potential imitators.

Elaborating on the argument of Teece, there are at least two types of activities needed to profit from innovations:

inventive activities, to produce (or acquire) the “core knowledge” embedded in a potential innovation

appropriation activities, to profit from the creation of this core knowledge

It is obvious that not all firms engage in inventive activities. And as stated by Teece a successful innovation cannot be assured if firms focus solely on inventive activities, given the need for managing appropriation activities.

According to recent studies [5 , 24], the different mechanisms to manage appropriation problems can be classified in two groups:

legal mechanisms, such as patents, utility models, copyright, trademarks, contracts and changes in labour legislation

strategic mechanisms such as secrecy, lead-time, complexity and other complementary resources and capabilities such as competitive manufacturing capacity, control of distribution networks and relationships with customers and suppliers, among others.

There are very good reasons to expect that innovation strategies might be context-dependent, since the threats and opportunities for innovative firms are not the same everywhere. Castellacci [8] proposed a sectoral taxonomy based on differences in appropriability regimes. Although he attempts to discuss the variety of patterns at the aggregate level, it is easy to see that these differences might have profound implications for management practices at the firm-level. For example, relying on legal protection of IP might be a good strategy for certain types of codified knowledge. However, this kind of strategy is far from perfect in other contexts. Take, for instance, the case of tacit knowledge flowing with the mobility of workers. In this case the management of human capital in the company might be essential to guarantee the appropriation of profits from R&D [7 , 17]. The creation of networks also affects the ability of firms to appropriate the profits of innovation. As Henttonen et al. [14] observe, collaboration with other partners for R&D activities might generate risks of spreading knowledge and misappropriation of the value generated. This risk can be managed using different strategies, alongside the legal mechanisms of IP protection, such as contracts, lead-time, etc. [26].

In this paper we suggest that the study of knowledge management activities cannot be approached from a context-free perspective, since these types of activities have an organic relation with the specific context faced by the firm. In this sense, we propose that not all knowledge management activities are equally desirable in all contexts, given the fact that some of them will have little effect on the general strategy of the innovative firm under certain conditions. Thus, we address the following research questions: does the appropriability regime faced by the firm influence the impact of management strategy on innovation activities? And, if this is the case, is it possible to find clear patterns for knowledge management responses from companies, which can be explained by differences in appropriability regimes?

3 Research method and data

3.1 The model

Most analyses relating knowledge management with innovation processes use “output variables” such as innovation performance or sales. In this paper, we have chosen a different strategy studying the role of such practices on the innovation process through its impact on R&D investment.

There are at least two good reasons for this choice. Firstly, the advantage of using input variables is that they differ from output variables in the sense that they reflect an active attitude towards the attainment of innovations, and are not directly mixed with demand conditions for innovation, as is the case of variables such as innovation performance [3]. In this sense, most analyses involving performance measures might be using incorrectly specified models, since the explanation of innovation performance might require, besides the input variables, other measures representing demand conditions that are usually omitted in such models.

Secondly, based on Teece [30], recent studies have highlighted the relevance of complementary knowledge and capabilities affecting the profitability of R&D investment, but the study of knowledge management as a complementary intangible asset in innovation processes has been overlooked in empirical analyses. Indeed, some management practices are expected to have an intimate (maybe trivial) relation with the profitability of R&D investment since they are directly connected with the development of the core knowledge embedded in potential innovations (e.g., R&D outsourcing, collaboration for R&D projects, acquisition of external knowledge such as patents, licenses or other technical knowledge, etc.). But there are other management practices that are not related with this core knowledge, and therefore their role in innovation processes (if any) is not straightforward and might show how R&D profitability depends on other types of knowledge and capabilities, such as manufacturing capacities, managerial capabilities, marketing activities, and so on.

As stated before, from an innovation perspective we might split the firm’s activities into two different groups: inventive and appropriation activities, that is to say, activities that produce the core-knowledge embedded in innovations and activities that appropriate the benefits of such innovations, respectively. From an empirical point of view, when analyzing knowledge management as a determinant of private R&D investment, we must be aware of the distinction between these two kinds of activities. On the one hand, the relation of inventive activities with R&D profitability might be somehow trivial: inventive activities are expected to be complementary or substitutes to R&D activities, because R&D is one of them. On the other hand, appropriation activities might be related (but not necessarily) with R&D profitability because they might be a part of the firm’s strategy to profit from R&D activities though the commercialization of R&D results.

The basic modelling approach to R&D investment from a micro-level perspective is well developed [10, 15]. The starting point is the assumption that the firm sorts the existing R&D projects hierarchically according to their expected returns, in order to start investing in those considered more profitable. The marginal profit of R&D activities can then be represented as a function involving marginal returns (MR) and marginal costs (MC), as presented in the following system $F_{i} = {MR}_{i} (R_{i}, I_{i}, X_{i}, Z_{i}) - {MC}_{i} (R_{i}, I_{i}, X_{i}, Z_{i}),$ (1) where R_i is a scalar measure of internal R&D intensity for i-th firm– as it is usual in conventional analysis, each project is implicitly taken as being finely divisible for expositional convenience. The vector I _i contains measures for inventive activities (different from internal R&D); vector X _i captures appropriation activities, such as IP protection, production/manufacturing capacity, marketing activities, up-to-date commercial practices, and collaboration agreements with costumers; Z _i are other factors possibly related with the profitability of R&D activities.

A competitive equilibrium requires that F_i = 0, for all i. Regarding R_i, the “shift factors”, I _i, X _i, and Z _i, are assumed to displace the marginal profit function. Figure 1 illustrates this case, assuming that $F : ℝ^{2} \to ℝ$ . A shift in the use of input X from X’ to X” generates new incentives to invest in R&D. The result is an increase of R&D activities of R”-R’.

Fig. 1

Shifts in R&D intensity.

Assuming that the output of R&D activities is a function with strictly decreasing returns to scale for each input, and the profit maximizing behavior, then we can interpret the relation between R&D activities and other inputs as a function^rmbox1: $R_{i} = R_{i} (I_{i}, X_{i}, Z_{i}), for i = 1, \dots, n$ (2)

The dependent variable in equation (2) is explained by other variables divided in two groups: inventive activities ( I _i) and appropriation activities ( X _i). Given the fact that our dependent variable is also an “inventive activity”, its relation with other activities of this type is somehow straightforward, showing the scale of the invention plan, its openness to other sources of knowledge, and the substitutability between internal R&D and external sources of knowledge. On the other hand, the explanation provided by appropriation activities captures the non-trivial relation of the activities to appropriate profits from R&D efforts. Finally, among the remaining shift factors ( Z _i), we will pay attention to the degree of openness to other sources of knowledge – as studied in Laursen and Salter [19], Chen et al. [9], Garriga et al. [12], Salter et al. [28], Ardito and Petruzzelli [2], Huang et al. [18]– in order to verify the profitability of the open innovation strategy.

With this model, the empirical strategy to answer our research questions is to test the relation of knowledge management strategies with R&D profitability under different appropriability regimes, following the sectoral taxonomy proposed in Castellacci [8]. In order to simplify the exposition of the results, we shall focus on two sectors of activity: Software and Hardware. The former is classified as a knowledge intensive business service having a weak appropriability regime, as opposed to the latter one which is part of the science-based manufacturing sector with a stronger appropriability regime. In this sense, our analysis can be seen as a comparative case study, where empirical evidence is intended to show the plausibility of the relation between the environment of the firm and the knowledge management strategy. The justification for the choice of these two sectors is that they appear to be based on very different strategies regarding IP protection; therefore we expect to find not only that patenting is a profitable strategy in hardware, but also to prove that the strategy in Software activities is typically more complex, making use of other practices besides legal protection of IP. Additionally, given these differences in appropriability regimes in these sectors, we will show that the threat from imitators in weaker regimes negatively impacts the profitability of open innovation strategies.

3.2 Data and measures

The specialized literature usually expresses equation (2) as a linear function. We also follow this traditional approach, using panel data techniques for the period 2008– 2015 with data from the Technological Innovation Panel (PITEC, Spain), which is the result of a joint effort of the National Institute of Statistics (INE) and the Spanish Foundation for Science and Technology (FECYT), along with the advice of a group of academic experts. The sample is selected from the Community Innovation Survey (CIS-Spain) and consists mainly of firms with intramural R&D expenditures in some points of period covered by the project (but not necessarily all) and firms of more than 200 employees. In contrast with other European experiences with CIS, PITEC is designed as a panel survey, i.e. it gathers information annually from a stable sample with the possibility of observing a specific firm in several points in time.

Given the fact that the aim of the study is to analyze the role of certain inputs in the profitability of R&D devoted to the generation of new products, the sample only includes companies trying to innovate in at least 4 years out of the 8 years included in the analysis. As explained in the previous section, we focus our comparison on two sectors of activity: hardware and software. In the former case, we use a sample of 170 firms, and in the latter case a sample of 387 firms.

R&D intensity, our dependent variable, is measured by R&D expenditure per employee. Given the fact that this variable is left-censored in the value of zero (since we have several innovative firms that do not invest in R&D activities, in both sectors), we use a random-effects tobit regression for longitudinal data. On the other hand, our independent variables are divided into three different sets. The first one comprises the inventive activities: R&D outsourcing; acquisition of other external knowledge, such as the purchase or licensing of patents and non-patented inventions, know-how, and other type of knowledge; and training for the firm’s personnel specifically for the development of new products or processes.

The second set of variables contains measures for appropriation activities: legal mechanisms such as patents, utility models and copyright; production capacity, measured by the product per employee; marketing activities for the introductions of new products, including market research and launch advertising; up-to-date commercial activities, including changes to the design of products, new methods for product placement, new media or techniques for product promotion, and new methods of pricing goods or services; and, finally, cooperation with clients or customers for the development of new products or processes.

The third set of variables include the strategy of open innovation, measured by the openness depth (i.e. the number of external sources of knowledge that are highly important for the firm); and other control variables such as the number of employees, participation in international markets, membership to a group of firms, and time effects. All variables are presented in Table 1.

Table 1
Definition of variables

Variable Description

Internal R&D intensity Internal R&D expenditure per employee (natural log.)

Openness depth Number of external sources of knowledge that are highly important. Ranging from 0 to 11. Possible sources: firms within the group of the firm, costumers, providers, competitors, consultants, universities, public research centers, technology centers, conferences, scientific papers, and professional associations

Openness depth (sq.) Openness depth squared

Legal protection Use of patents, utility models or copyrights (dummy)

Porduction capacity Sales per employee (natural log.)

Marketing Marketing activities for the introductions of new products, including market research and launch advertising (dummy)

Commercial innovations Up-to-date commercial activities, including changes to the design of products, new methods for product placement, new media or techniques for product promotion, and new methods of pricing goods or services (dummy)

Colaboration with costumers Formal agreements with costumers to collaborate in innovative projects

Num. of employees Number of employees (natural log.)

Num. of employees (sq.) Num. of employees squared

R&D outsourcing External R&D activities (dummy)

External knowledge Purchase or licensing of patents and non-patented inventions, know-how, and other type of knowledge (dummy)

Training Internal or external training for firm’s personnel for the development of new products or processes (dummy)

Group of firms Membership to a group of firms (dummy)

International markets Participation in international markets (dummy)

Time effects 7 dummy variables for each year of activity (excluding 2008)

Variable	Description
Internal R&D intensity	Internal R&D expenditure per employee (natural log.)
Openness depth	Number of external sources of knowledge that are highly important. Ranging from 0 to 11. Possible sources: firms within the group of the firm, costumers, providers, competitors, consultants, universities, public research centers, technology centers, conferences, scientific papers, and professional associations
Openness depth (sq.)	Openness depth squared
Legal protection	Use of patents, utility models or copyrights (dummy)
Porduction capacity	Sales per employee (natural log.)
Marketing	Marketing activities for the introductions of new products, including market research and launch advertising (dummy)
Commercial innovations	Up-to-date commercial activities, including changes to the design of products, new methods for product placement, new media or techniques for product promotion, and new methods of pricing goods or services (dummy)
Colaboration with costumers	Formal agreements with costumers to collaborate in innovative projects
Num. of employees	Number of employees (natural log.)
Num. of employees (sq.)	Num. of employees squared
R&D outsourcing	External R&D activities (dummy)
External knowledge	Purchase or licensing of patents and non-patented inventions, know-how, and other type of knowledge (dummy)
Training	Internal or external training for firm’s personnel for the development of new products or processes (dummy)
Group of firms	Membership to a group of firms (dummy)
International markets	Participation in international markets (dummy)
Time effects	7 dummy variables for each year of activity (excluding 2008)

4 Results

In this section we summarize the results obtained in this research. The evidence confirms that appropriability regimes explain the differences in knowledge management strategies. Firms in the software sector typically make use of a more complex strategy that goes beyond IP protection. On the other hand, the profitability of open innovation strategy exhibits a U-inverted shape, and the effect of decreasing returns is higher in the software sector, consistently with the hypothesis of higher threats in weaker appropriability regimes.

In Table 2 we present descriptive statistics of the variables used in the econometric analysis. Summarizing, the hardware sector exhibits a higher R&D intensity, uses a more open strategy, relies more on legal protection of IP, shows a higher propensity to engage in marketing activities and R&D outsourcing, and participates more in international markets.

Table 2
Descriptive statistics

Explanatory variables Hardware Software

Internal R&D intensity (log.) 8.3 6.9

Openness depth 2.2 1.9

Legal protection 0.20 0.11

Porduction capacity (log.) 11.7 11.1

Marketing 0.40 0.27

Commercial innovations 0.49 0.41

Colaboration with costumers 0.23 0.19

Num. of employees (log.) 3.6 3.6

R&D outsourcing 0.40 0.26

External knowledge 0.03 0.03

Training 0.22 0.19

Group of firms 0.34 0.34

International markets 0.76 0.38

Explanatory variables	Hardware	Software
Internal R&D intensity (log.)	8.3	6.9
Openness depth	2.2	1.9
Legal protection	0.20	0.11
Porduction capacity (log.)	11.7	11.1
Marketing	0.40	0.27
Commercial innovations	0.49	0.41
Colaboration with costumers	0.23	0.19
Num. of employees (log.)	3.6	3.6
R&D outsourcing	0.40	0.26
External knowledge	0.03	0.03
Training	0.22	0.19
Group of firms	0.34	0.34
International markets	0.76	0.38

The results of econometric analysis are presented in Table 3. The first important result is the positive coefficient related to the variable openness depth, and the negative value of the coefficient related to the squared value of this same variable, consistent with a curvilineal (U-inverted) relation between the strategies of open innovation and the profitability of R&D.

Table 3

Econometric results

Explanatory variables	Hardware		Software
	Coef.	Pr>t	Coef.	Pr>t
Openness depth	0.571	0.00	1.261	0.00
Openness depth (sq.)	– 0.047	0.01	– 0.133	0.00
Legal protection	0.514	0.02	0.504	0.06
Porduction capacity	– 0.332	0.05	0.072	0.66
Marketing	0.138	0.40	0.349	0.04
Commercial innovations	0.280	0.11	0.738	0.00
Colaboration with costumers	0.078	0.73	0.607	0.01
Num. of employees	2.218	0.00	2.524	0.00
Num. of employees (sq.)	– 0.289	0.00	– 0.299	0.00
R&D outsourcing	0.643	0.00	1.621	0.00
External knowledge	– 1.278	0.01	– 2.198	0.00
Training	– 0.010	0.96	0.064	0.76
Group of firms	0.348	0.20	– 0.176	0.52
International markets	0.447	0.10	0.574	0.01
Cons.	6.802	0.00	– 0.937	0.65
Num. of obs.	1,320		2,838
Num. of firms	170		387
Pseudo R-squared	0.125		0.171

In Fig. 2 we isolate the relation between openness depth and R&D profitability, taking the coefficients of the model and assuming that the optimum degree of openness is equal in both sectors.

Fig. 2

Openness depth and R&D profitability.

As shown, the profitability of an open innovation strategy, measured by the openness depth, falls faster in the software sector. This result is consistent with the hypothesis that the threats of sharing knowledge with other agents are bigger in weaker appropriability regimes.

Regarding the complexity of knowledge management strategies, results show that firms from the software sector rely to a larger extent on practices that go beyond the legal protection of IP. In Fig. 3 we compare the coefficients of appropriation activities (weighted by 1 minus the p-values, as a measure of the probability of the coefficient being equal to zero).

Fig. 3

Differences in appropriation mechanisms.

Firms from the hardware sector clearly rely on legal protection of IP, over other appropriation practices. Additionally, those firms profiting more from their R&D activities in this sector are characterized by low productive performance, consistently with Teece’s idea on the lower necessity of developing manufacturing capacities in stronger appropriability regimes. On the other hand, although companies in the software sector also rely on the legal protection of IP, it is very clear that the companies that make the most profitable investment in R&D in this sector use more complex marketing strategies.

5 Concluding remarks

Management practices across firms are highly conditioned by the environment they face. This is particularly true in the case of innovative companies, which face different levels of threat from potential imitators. Since the pioneering work of Teece [30], the concept of appropriability regime has been useful in describing specific sectoral patterns in the appearance of such threats. However, there have been few efforts to formally prove the impact of different regimes on firms’ knowledge management strategies.

In this article, we demonstrate the profound effect that appropriability regimes can have on the strategies of innovative companies. Our results show that, although there are benefits associated with open innovation strategies, the costs of appropriation tend to grow rapidly in sectors with relatively weak appropriability regimes. As our empirical evidence suggests, this is due to the fact that companies immersed in such regimes must resort to more complex (and costly) practices that go far beyond the legal protection of IP, requiring the firm to integrate many other management practices related to productive capacity and commercialization.

These results have direct implications for the managers of R&D activities in firms, since they highlight many of the costs and efforts that their companies could face in order to profit from their innovative activities. This conclusion is not trivial, given the fact that such costs and efforts might be associated with activities that are relatively independent from the knowledge and capacities needed to develop a new product or process.

On the other hand, our results are also useful to inform innovation policy, since they could help to move the focus of attention from inventive activities (which are traditionally the focus of innovation policies) to appropriation activities. In this sense, many of the policies designed to promote more efficient production and commercialization practices could be framed within a holistic vision of innovation policy.

Footnotes

Acknowledgments

Support to our project entitled “Policies to encourage business innovation in Chile”, from Inter-University Cooperation Projects UAM-Santander with Latin America is gratefully acknowledged.

By the implicit function theorem, these functions will exist if ∂F/∂R ≠ 0. It is easy to show that decreasing returns of scale guarantee that ∂F/∂R < 0, since ∂MR/∂R < 0, and ∂MC/∂R > 0. It would be possible to obtain the relation between R and the other input variables using this theorem, but as far as we know there are no reasonable assumptions surrounding the sign for each ∂F/∂X_i. These are exactly the relations estimated by the model proposed in this paper.

References

Albus

J.S.

and Meystel

, Behavior Generation in Intelligent Systems. US Department of Commerce, Technology Administration, National Institute of Standards and Technology, 1997.

Ardito

and Petruzzelli

A.M.

, Breadth of external knowledge sourcing and product innovation: The moderating role of strategic human resource practices, European Management Journal35(2) (2017), 261–272.

Arqué-Castells

, Persistence in R&D performance and its implications for the granting of subsidies, Review of Industrial Organization43(3) (2013), 193–220.

Arrow

K.J.

, Economic Welfare and the Allocation of Resources for Invention, In: Nelson, R. (Ed.). The Rate and Direction of Inventive Activity. Princeton University Press (1962), 609–625.

Barros

H.M.

, How changes in the market for ideas impact appropriability behaviour? In ISPIM Conference Proceedings. The International Society for Professional Innovation Management (ISPIM), 2015.

Calantone

R.J.

, Schmidt

J.B.

and Di Benedetto

C.A.

, New product activities and performance: The moderating role of environmental hostility, Journal of Product Innovation Management14(3) (1997), 179–189.

Casper

and Whitley

, Managing competences in entrepreneurial technology firms: A comparative institutional analysis of Germany, Sweden and the UK, Research Policy33(1) (2004), 89–106.

Castellacci

, Technological paradigms, regimes and trajectories: Manufacturing and service industries in a new taxonomy of sectoral patterns of innovation, Research Policy37(6-7) (2008), 978–994.

Chen

, Chen

and Vanhaverbeke

, The influence of scope, depth, and orientation of external technology sources on the innovative performance of Chinese firms, Technovation31(8) (2011), 362–373.

10.

David

P.A.

, Hall

B.H.

and Toole

A.A.

, Is public R&D a complement or substitute for private R&D? A review of the econometric evidence, Research Policy29(4-5) (2000), 497–529.

11.

Dosi

, Marengo

and Pasquali

, How much should society fuel the greed of innovators? On the relations between appropriability, opportunities and rates of innovation, Research Policy35(8) (2006), 1110–1121.

12.

Garriga

, Von Krogh

and Spaeth

, How constraints and knowledge impact open innovation, Strategic Management Journal34(9) (2013), 1134–1144.

13.

Greco

, Grimaldi

and Cricelli

, An analysis of the open innovation effect on firm performance, European Management Journal34(5) (2016), 501–516.

14.

Henttonen

, Hurmelinna-Laukkanen

and Ritala

, Managing the appropriability of R&D collaboration, R&D Management46(S1) (2016), 145–158.

15.

Howe

J.D.

and McFetridge

D.G.

, The determinants of R&D expenditures, Canadian Journal of Economics 9(1) (1976), 57–71.

16.

Hurmelinna

, Kyläheiko

and Jauhiainen

, The Janus face of the appropriability regime in the protection of innovations: Theoretical re-appraisal and empirical analysis, Technovation27(3) (2007), 133–144.

17.

Hurmelinna-Laukkanen

and Puumalainen

, Nature and dynamics of appropriability: Strategies for appropriating returns on innovation, R&D Management37(2) (2007), 95–112.

18.

Huang

, Chen

and Liang

, How open innovation performance responds to partner heterogeneity in China, Management Decision56(1) (2018), 26–46.

19.

Laursen

and Salter

, Open for innovation: The role of openness in explaining innovation performance among UK manufacturing firms, Strategic Management Journal27(2) (2006), 131–150.

20.

Laursen

and Salter

A.J.

, The paradox of openness: Appropriability, external search and collaboration, Research Policy43(5) (2014), 867–878.

21.

Liao

, Liu

and Jia

, How Does A Firm Build Its Capabilities through Its Network and Under Environmental Hostility? In Academy of Management Proceedings Briarcliff Manor, NY 10510: Academy of Management, 2016.

22.

Lin

J.Y.

, Effects on diversity of R&D sources and human capital on industrial performance, Technological Forecasting and Social Change85 (2014), 168–184.

23.

Löfsten

and Lindelöf

, Environmental hostility, strategic orientation and the importance of management accounting— an empirical analysis of new technology-based firms, Technovation25(7) (2005), 725–738.

24.

Milesi

, Petelski

and Verre

, Innovation and appropriation mechanisms: Evidence from Argentine microdata, Technovation33(2-3) (2013), 78–87.

25.

Miller

and Friesen

P.H.

, Strategy-Making and Environment. The Third Link, Strategic Management Journal4 (1983), 221–235.

26.

Pérez-Cano

, Firm size and appropriability of the results of innovation, Journal of Engineering and Technology Management30(3) (2013), 209–226.

27.

Pisano

, Profiting from innovation and the intellectual property revolution, Research Policy35(8) (2006), 1122–1130.

28.

Salter

, Ter Wal

A.L.

, Criscuolo

and Alexy

, Open for ideation: Individual-level openness and idea generation in R&D, Journal of Product Innovation Management32(4) (2015), 488–504.

29.

Schumpeter

J.A.

, Capitalism Socialism and Democracy. Harper and Row, New York, 1950.

30.

Teece

D.J.

, Profiting from technological innovation: Implications for integration, collaboration, licensing and public policy, Research Policy15(6) (1986), 285–305.

31.

Teece

D.J.

, Profiting from innovation in the digital economy: Enabling technologies, standards, and licensing models in the wireless world, Research Policy47 (2014), 1367–1387.