Abstract
This study has the purpose of conducting a dynamic comparative analysis of interorganizational innovation networks of Brazilian and Spanish biotechnology companies. It aims to explore the differences between the network characteristics and their actors’ positions and types as well as to analyze the temporal evolution of these networks. Although analyses of the evolution of networks are relevant in supporting comprehension of paths for innovation, they are incipient, particularly as cross-country studies. Considering that, this comparison between Brazil and Spain, which are at different levels of biotechnology development, may contribute to the advancement of this area. Social network analysis techniques are employed to draw and measure the network characteristics constructed under the information of patent applications by biotechnology companies located in Brazil and Spain in the period of 1990–2012. The results demonstrate an impressive growth of innovation networks from both countries but show an inferior position for Brazilian data in terms of constancy, volume of partnerships, partner diversity, and main actor types. Thus, this article reveals patterns of evolution over time for each country, which allows for the determination of the implications for public policy and managerial experts.
Introduction
Considering that the necessary sources of innovation are widely dispersed around the world, organizations need to identify and connect to new external sources in order to develop new technologies. This may be done through the establishment of formal partnerships with external actors to conduct research and development (R&D), which is also considered as a means of forming interorganizational networks for innovation. They are formed by actors with different characteristics that pool their resources and share risks, knowledge, personnel, and infrastructure in complex projects by complementing the competences of each one involved in the network. “These collaborations offer the possibility of exploiting potential cognitive synergies and accessing knowledge wherever located within the network.” 1 This articulation is especially relevant in sectors such as biotechnology, which is characterized by multidisciplinary and complexity of knowledge. It is noteworthy that the development of the biotechnology industry is directly related to the establishment of interorganizational partnerships and to the development of innovation networks, which enable companies to access and combine the resources and capabilities they internally lack. 2
Powell et al. 2 also demonstrated the importance of the different types of organizations that are part of an innovation network in biotechnology to innovation performance as well as the different types of agreements, or relationships, that these organizations can establish. The study of Demirkan and Demirkan 3 demonstrated that the network’s knowledge heterogeneity, through the different characteristics of actors involved and the strength of their relations, directly affects the innovation performance of biotechnology companies.
In addition to diversity of actors, characteristics of networks and actor positions have also been focused on. Choi et al. 4 show the relevance of central actors as linkers for providing learning by players. This behavior of network has impacts on another recent path of study about innovation networks, that is, the analysis of the evolution of the networks. 5 For example, Abbasi 6 notes that the attraction of new entrants during the network evolution is explained by the capacity of intermediation by central players.
The innovation networks’ characteristics in the biotechnology sector have been exhaustively investigated in the recent past. 2,3,7 However, analyses of the evolution of networks are incipient, particularly if we consider cross-country studies. This type of research enables us to systematically describe the structure and dynamics of the networks in order to plot the past and to predict the possible future paths for innovation. 5,7
Considering that Spain initiated the development of biotechnology in a prior period to that of Brazil, 8 this study has the purpose of conducting a dynamic comparative analysis of innovation networks of Brazilian and Spanish biotechnology companies. The social network analysis (SNA) methods were applied to compare the differences among the network structures and their main actors’ positions and characteristics as well as to analyze the temporal evolution of these networks. This comparison is also motivated by the lack of comparative studies on innovation networks of companies operating in countries at different levels of development.
As a contribution to the research area, the main addition of this study is to comparatively analyze the evolution of innovation networks of biotechnology companies operating in environments with different development levels, revealing the corresponding stage of each country and patterns of evolution over time. These comparative studies are rarely found in the literature and may be considered for predicting different innovative routes though the entrance of new countries to a particular industry.
Literature review
Innovation networks
Since the middle of the last century, evidence of the importance of establishing technological innovation networks has been reported. Through innovation networks, organizations could access the external resources and capabilities that, combined with internal research and development sources, could collaborate to the success of innovation. 9 This generates some advantages such as accelerating the development of new products and services, anticipating the products and services’ introduction to market, reducing and sharing R&D spending, and improving the success rate and innovation levels. 10
By being part of an interorganizational network for innovation, organizations could join and recombine key resources, tangibles and intangibles resources like knowledge, expertise, know-how, infrastructure, personal, and others, thus allowing the value creation and appropriation by the members. 11,12 In these relationships for innovation prevails the process of knowledge creation and distribution. 1 “The relationships between network members can be understood as deriving from their autonomy and interdependence, the coexistence of co-operation and competition, as well as reciprocity and stability.” 13
Organizations rely on various means to access external sources to innovate, including the establishment of interorganizational formal partnerships for R&D. Formalization is established through contracts that constitute the guidelines for resources and knowledge exchange and appropriation, promoting confidence among partners. 11,14 Establishing formal partnerships for R&D is also considered as a means for forming interorganizational networks for innovation.
These interorganizational networks are composed of multiple organizations with distinct characteristics that establish relationships with the purpose of innovate. Therefore, the SNA perspective is a research tool very suitable for study and visualize these relationships, because it deals with relations between actors, allowing the analysis of the structural characteristics in a set of complex relations by means of mathematical tools. Besides, providing a theoretical framework, by which it is possible to investigate the effects between the structures of a network and the innovation output, as well as the impact of a particular position occupied by an actor on the network and the individual innovation performance. 12,15,16
The effects of network characteristics exert a direct impact on firms’ innovative performance. 4,17 In highly dense and clustered networks, the resource flows are more easily accessible to the players. However, in the long term, resources may become redundant to players embedded in highly connected communities, so links between different communities, such as a bridge (linking distant communities), can enrich the available knowledge to players and increase innovation results. 18,19
The organizations embedded in networks highly clustered with many shortest paths present better innovation results than companies in networks that do not present such characteristics 17 or in highly connected networks with central actors linking distant communities, which allow non-redundant knowledge access and increase learning of players. However, it is necessary to create new arrangements and hub formation to prevent redundancy in the long term. 4
The network position occupied by an actor is also important, as an actor in central positions can benefit from superior opportunities to access and mobilize many of the resources. Being close to other players in a network gives an actor the advantage of being able to disclose resources across the network. 20 On the other hand, the intermediating position of an actor is important for non-redundant knowledge access, what may conduce to superior creativity. 19 The intermediation position could also be considered as source of power, which enables the control of resource flows across the network, particularly if it is the only actor linking different communities. The node in this position strategically operates like a bridge, controlling the access of resources to other players in the network. 18
Partner diversity
An interorganizational network for innovation is formed by actors who have different characteristics, abilities, knowledge, and objectives. They will share their resources, knowledge, personnel, and infrastructure in complex projects by identifying the competences of each actor involved in the network. Thus, the selection of a suitable partner is very important for organizations because knowledge heterogeneity and the partnerships’ continuity influence the innovation performance. 1,3,21
Powell et al. demonstrated the importance of the different types of organizations that are part of the network to the innovation performance and different types of agreements, or relationships, that organizations can establish. The study was conducted with 225 biotechnology companies in the United States in the 1990–1994 period and focused on formal agreements between actors. It showed that these innovation networks are becoming increasingly denser and that the diversity of types of actors has increased.
The study of Demirkan and Demirkan sought to analyze the characteristics of the organizations involved in an innovation network in biotechnology to assess the innovation performance. The study demonstrated that the network’s knowledge heterogeneity, through the different characteristics of the actors in it, directly affects the innovation performance of these companies. The different partners were classified as other companies, universities, R&D centers, and pharmaceutical institutions. It was also shown by analyzing the number of partnerships established among network members in a given period of time that the relationship strength affects the innovation performance. 3
In addition to the characteristics of the actors of a network and to the relationships among them, some aspects, such as the geographical location of partners, can also affect network performance. That is because, upon establishing partnerships with actors located in other countries or regions, organizations are accessing new knowledge sources that are not locally available.
Van Beers and Zand 22 suggest that functional diversity can lead to a variety of knowledge and synergy effects needed to develop new products, having further effect on radical innovations. However, the partners’ geographic diversity affects the performance of the incremental innovation, which suggests that, to introduce products or services with incremental improvements, it is important to have partners from various geographic locations, which is related to the adjustments and customizations offered specifically for local preferences or regulations and standards of different countries.
According to Patel et al., 23 the development of products in innovation networks that are geographically balanced among local and foreign actors accelerates the insertion time of a product in the international market. They argue that foreign partners enable the company to have access to key knowledge to innovate and internationalize, demanding, however, extensive resources and efforts to establish the partnership.
Morescalchi et al. 24 affirm that in the last decade, the intensity of cross-country partnerships among countries of the European Union (EU) increased at a pace greater than that for countries outside the EU. The authors also emphasize the emergence and constant growth of hubs that attract partners from different regions and with different characteristics.
Network evolution
The studies of network evolution have been of growing interest because neither a network structure nor an actor position is stable over time, being in constant evolution. Thus, the dynamic studies can reveal patterns or characteristics over time (e.g. how new links are created, the attachment propensity of the new entrants, and how bridges emerge between unconnected communities). 5
To explain the evolution of cooperative networks in biotechnology, some rules of attachment were tested, such as cumulative advantage, homophily, follow-the-trend, and multiconnectivity. The cumulative advantage follows a recognized rule in social sciences, the “rich get richer” phenomenon, in which nodes with higher degree tend to attract the new connections, this characteristic was identified in many real networks. In turn, for homophily rule, new connections were explained by similarities in actors’ attributes. On the other hand, follow-the-trend rule the actors tend to follow the dominant behavior of entire population. Multiconnectivity rule suggests the search for diversity and the interactions with heterogeneous actors. 25
Another view, recently proposed in the study conducted by Abbasi, 6 in which the factors that affect the attachment behavior of nodes in a collaboration network were investigated. The author demonstrates that, during the evolution process of a network, structural position is the most effective process that exposes the attachment behavior of nodes. The results demonstrated that the preferential attraction of new entrants during network evolution is better explained by actors with great capacity for intermediation.
Diffusion of innovation in biotechnology
Modern biotechnology is characterized by multidisciplinary and complexity of knowledge and by the high dependency on basic research in addition to the high risk involved in developing a new discovery. Thus, the development of the biotechnology industry is directly related to the establishment of interorganizational partnerships and the formation of cooperation networks, which enable companies to have access to sources of knowledge that are not internally available. 2
According to Niosi et al., 8 the development of the biotechnology sector, both scientific and technological with commercial application, started in the United States. The development of biotechnology was rapidly disseminated to a group of countries, the EU, Canada, and Japan, that was considered by the authors as the second biotechnology diffusion wave. The main developing countries of Asia and Latin America, including Brazil, are considered to be the countries of the third biotechnology diffusion wave and began their development with nearly two decades of delay in relation to pioneering countries.
The diffusion of biotechnology to developing countries, or of the third wave, occurred mainly in the area of science. The study conducted by the authors showed that, in the 1996–2007 period, scientific publications related to biotechnology from third-wave countries increased worldwide participation by 64%. Publications coauthored with developed countries also increased, and the United States was the main coauthor in publications with developing countries.
However, technological production, or patents granted to countries in the third wave, evolved slightly and still little represented in relation to developed countries, thus demonstrating that the countries of the third wave adopted biotechnology primarily in the area of science but with little representation in technology.
Niosi
26
highlights the continuous US leadership in biotechnology. In this recent study, the author summarized some reasons for this permanent leadership based on the Prevezer
27
as follows: There are substantially more academic research funds in the United States than in Europe. It is easier for US academics to found start-ups while retaining their academic posts. US start-ups are concentrated in human health and agricultural biotechnology, where commercial potential is higher. Financing and managerial conditions are easier in the United States in terms of venture capital, stock market admissibility, and access to managerial expertise. There is greater access for US dedicated biotechnology firms to alliances with large domestic and foreign pharmaceutical corporations.
Another fact added by Niosi based on Cockburn and Stern 28 about the intellectual property system is that it is easier and less costly to patent in the United States than it is in Europe.
Yagüe-Perales et al. 29 conducted another study comparing how similar and different are the national system of Spanish biotechnology with those of other countries such as Canada, the United Kingdom, or the United States. This study shows that Spain is emerging in biotechnology. The Spanish system is based on university publication and reduced patenting, few academic spin-offs and little venture capital support. However, public funding of biotechnology research is higher than European standards but low by comparison with the Anglo-Saxon system. They suggested that governments have to provide additional incentives for private firms to invest in Dedicated Biotechnology Firms, and universities have to be more commercially oriented and find ways to convert more novel ideas into patents instead of scientific publication.
They highlighted that Spanish biotechnology tends to follow a service orientation rather than a production approach. However, Spain has some companies that are reaching worldwide leadership in specific R&D such as that for anticancer drugs and nutraceutics. These problems are quite similar if we regard the biotechnology sector in Brazil. Thus, it is very interesting to compare these two countries for how they are catching-up and overcoming these challenges in order not to lose the biotechnology industry race.
Methodological aspects
Data collection
This research used data of patents’ applications under co-ownership by biotechnology companies located in Brazil and Spain. The filing of patents in co-ownership is an indicator of formal collaboration for R&D. According to the Organization for Economic Cooperation and Development (OECD), 30 patents are often linked to R&D and can be considered a good indicator because it is the result of companies’ innovation process, covering all fields of technology and offering global geographical coverage, as most countries have a patent system.
In Brazil, samples were selected from directories of companies provided by studies representing the biotechnology sector, such as Biominas Brazil and PWC, 31 and the Brazilian Centre for Analysis & Planning and Brazilian Association of Biotechnology 32 ; additionally, we selected some companies that are incubated or installed in technology parks related to biotechnology such as the Business Incubator of the Biotechnology Center (IE-CBiot, BIO-RIO, the Technology Park of Viçosa, CIATEC, CIETEC, SUPERA, HABITAT, and CIAEM). A total of 229 Brazilian biotechnology companies were selected for patent research.
In Spain, we selected companies associated with the Spanish Bioindustry Association and the biotechnology companies installed or incubated in the Scientific and Technological Park of Barcelona, Science Park of Madrid, and the Scientific Park of the University of Valencia. After the consolidation of companies’ information and removal of duplicated companies, 200 Spanish biotechnology companies were selected for patent research.
For collection of patent information, we used the search platform Thomson Innovation, provided by Thomson Reuters, which is one of the largest and most comprehensive databases of patents that exists, allowing simultaneous access to most of the world’s patent offices. The patents were searched for with the name of the companies selected as applicants. We considered patent applications between 1990 and 2012. The type of patent information available for this study is the publication number, filing date, assignee names, and address. The data of patent families were filtered to select patent applications under co-ownership (two or more owners).
Data analysis was conducted first in a descriptive manner for both countries intending to determine the time evolution of patent applications and classification of types of partner; then, SNA was applied to compare both networks’ characteristics.
Network analysis
Innovation network relates to interactions between a set of actors with the aim of innovation. In this context, the perspective of SNA is an ideal tool to analyze that type of relationship. 33 The SNA is a set of mathematical algorithms that are applied for representation and analysis of networks with a sociological approach. 16,34 In recent innovation network studies, the SNA techniques have been applied extensively. 15
The SNA is based on the relations of a set “nodes” linked by a set of “edges.” 16,34 In this research, the nodes of networks represent patents’ owners (organizations; it is worth mentioning that co-ownership with individuals (natural persons) is not considered in this research because this study aims to analyze the innovation networks at the interorganizational level), and the undirected connections between nodes will be established by a co-ownership of a patent. The strength between actors is represented by a number of patents applied under co-ownership over time. As the focus of this research is on biotechnology company innovation networks, at least one assignee of each patent is a biotechnology company, and the other nodes that share the ownership of the patent are considered the “partners,” which were classified by type and the country that they are located as universities, R&D centers (public and private), companies (any legal firm as hospitals, pharmaceutical institutions, other biotechnology companies, and others), and government (public agencies or funding institutions).
SNA features a set of mathematical algorithms to explore the networks’ structural characteristics (macro-level) and the actors’ positions (microlevel). The networks’ structural characteristics, in this study, are analyzed under the metrics of average degree, density, diameter, number of components, and size of the giant component. The average degree metric allows for understanding the average connection of nodes; this measure could also be weighted when it takes into account the strength of links (frequency of connections between two nodes over time). The density describes how connected the nodes are in the network in comparison to all possible connections that the network could have. The diameter measures the longest geodesic distance (geodesic distance represents the minimal number of connections necessary for a node to reach another node in the network) in the network or the distance between a pair of the farthest nodes that exhibit the network extent. A component refers to a set of nodes directly or indirectly connected. A network may have many components, but the giant component is the network’s largest group, where the majority of nodes are connected.
The main metrics useful for analyses of the actor position are the centralities (degree and betweenness). The degree centrality defines that central actors are those who have more connections. For the betweenness centrality measure, the best positioned nodes are those that intermediate the most important paths in the network, connecting components or distant nodes and revealing the node capacity to intermediate connections in the network. 16
To compare the dynamic evolution of the networks, the characteristics were analyzed for three separated periods of 5 years each: 1995–2000, 2001–2006, and 2007–2012. The choice of range period was based on the dynamic methodology applied by Hu et al. 35 and takes into consideration that a patent co-ownership represents a collaboration by two or more business entities or a specific R&D agreement between multiple parties to share the ownership of the patents which could take long to appear. During the evolution through each period, the network metrics were compared between Brazilian and Spanish networks and also observed were the actors who remained in the network and the new entrants, as the remained links and the new establish connections, the formation and growth of the giant component, and the main actors who attracted new connections influencing the growth of the entire network and the giant component.
The graph representation of the network was designed for better data understanding. It distinguishes the diversity of partner by means of shape and colors of the nodes. The node size represents the degree measure (largest nodes represent highest coefficient). The size of the edges represents the strength among actors. Gephi 8.2 software was used to design the R&D cooperation networks of the companies and to calculate the network metrics.
Results
The results are presented in two sections: first, the data are described, and the partners of biotechnology companies are classified. Then, in the second section, Brazilian and Spanish innovation networks are analyzed and compared in terms of their characteristics, dynamic evolution, and main actor positions.
Descriptive data and partner diversity
A total of 701 patents were collected, and they were applied for by 108 biotechnology companies from both countries. In Brazil, there were 179 patent applications, 25% of the sample, belonging to 42 domestic companies. In Spain, there were 522 patent applications, 75% of the sample, belonging to 66 Spanish companies. Regarding the number of patent applications, the participation of Spain was three times that of Brazil.
Companies from both countries have experienced evolution in patent applications in recent decades, but the relationship variable between countries. Spain, from 1995 to 2012, multiplied the number of applications per year by 12. Brazil, however, started its evolution more recently, from 2000 onward, and it experienced peaks of growth and decline rather than a continuous increase. Figure 1 represents the temporal evolution of patent applications by both countries in the 1990–2012 period, demonstrating the proportion of applications under co-ownership.
Temporal analysis of patent applications by biotechnology companies (number of patents)—comparison between Brazil and Spain.
In Brazil, from 1995 to 2012, 18 companies filed for patents in co-ownership, which resulted in 42 patent applications in partnership, or 23.46% of the Brazilian sample, as demonstrated in Figure 2. Companies with patents under co-ownership established relationships with 28 different partners. Universities were identified as the main partners of Brazilian companies, representing more than 50% of the partnerships, of which three were foreign universities. In Spain, 22 companies filed for 82 patents under co-ownership or 15.71% of the Spanish sample. These partnerships were established with 43 different partners.
Comparison of partners classification—Brazil and Spain.
The biotechnology companies from Spain, compared to Brazil, held partnerships with a greater diversity of companies and R&D centers. In addition to the number of different partners, the number of interactions among them was also higher, compared to Brazil. This diversity of actors and the strength of the relationships between them directly influence the capacity for biotechnology companies’ innovation. 2,3
Regarding international partnerships, only four Brazilian companies sought foreign partners to share R&D, which resulted in seven patents filed. A total of four international actors were identified; three of which are universities and only one is a company. All foreign partners are located in developed countries.
Spanish biotechnology companies currently seek more international knowledge sources for R&D, which is evidenced by the number and diversity of foreign partners and by the number of patent applications under co-ownership with foreign partners. However, in general, international partnerships (measured by patents) for companies from both countries are of little significance.
Network dynamic analysis
The results of network analysis are presented in this section. First, the network characteristics are dynamically compared. Then, the positions of the main actors are presented.
During the three periods of analysis (T1: 1995–2000, T2: 2001–2006, and T3: 2007–2012), innovation networks of both countries showed impressive evolution in the volume of actors and relationships, but this growth differed among countries, with Spanish companies’ network standing out in volume and pattern of growth. Table 1 presents a comparison between the structures of Brazilian and Spanish dynamic innovation networks.
Comparative structures of dynamic innovation networks.
B: Brazil network; S: Spain network.
Considering the period of study, two basic measures describe the networks’ evolution in terms of the number of nodes (N): new entrants in the networks (actors that come to the networks) and actors that remained in the networks. Both networks experienced evolution in the number of new entrants, mainly during the T3; however, the Spanish network exceeded the number of entrants since the second period. The pattern of actors that remained in the networks for two or three periods differs among countries; in the Spanish network, from T1 to T2, half of the actors remained, and from T2 to T3, more than one-half stayed in the network. In the Brazilian network, only two actors remained from T1 to T3 and were absent in T2, but from T2 to T3, almost half of the actors remained. This shows that networks from both countries have developed distinctly, with the Spanish network evolving more constantly during the periods of study.
In a similar analysis, the numbers of new and remaining links were observed in order to describe the evolution of the networks’ relationships. However, in this case, it has been observed that few links remained in the network during the three periods; none of them remained for more than two periods. The Spanish network had one link that lasted from T1 to T2 and two links from T2 to T3. The Brazilian network has only two links that lasted from T2 to T3. From this, it is seen that, in terms of links, there is no difference between the two networks as few links remain for more than one period and considering the number of nodes that remained. This highlights a pattern for the networks: they do not sustain their links for more than one period; at the same time, they create links with new partners.
The analysis of the number of components reveals that both networks are highly fragmented, with many disconnected communities, particularly in the third period, when there were increased quantities of links and nodes (Figure 3 and 4). This situation creates a network with many structural holes, with potential to be linked. As Burt posits, the bridge between different communities could represent access of non-redundant resources, and it is considered to be the source of innovation. 19
Innovation networks of Brazilian biotechnology companies.
Innovation networks of Spanish biotechnology companies.
The size and evolution of the giant component show other fundamental differences among countries’ networks. In the Brazilian network, the actors inside the giant component change at each period, so the giant component did not evolve through the three periods, grouping 20% of the nodes in the third period. In the Spanish network, the giant component arose in the first period, and it had increased the number of nodes and links during the three analyzed periods, representing at the third period with 29% of all nodes and connected by 40% of all links. It can be observed that during this evolution, while some actors remained in the giant component, others left, but new links were created, bringing new actors inside the giant component. The size of the giant component directly affected the network’s diameter, which is another measure that is higher in the Spanish networks than that in Brazilian networks. The diameter of the network influences the actors’ capacity to reach resources from distant members in the network, and it could be associated with knowledge diffusion. 18
The last measure used to characterize the two countries’ network evolution was the average degree (average number of partners) and the weighted average degree (average number of partners considering the frequency of relationships). The data reveal similarities in the two countries’ average number of partners per actor at T1 and T3; however, the average strength between the partners was higher in Spain’s network in T1 and T2.
To analyze the actors’ structural positions in the networks, we used measures of degree centrality and betweenness centrality. The degree centrality describes the importance of the actors by the number of partners directly connected, and the measure of betweenness centrality indicates status of actors by the intermediation position. The weighted degree reveals the strength of relations between partners, what contributes to the confidence that exists between them and the experience that they accumulate in managing partnerships. 3
Brazil’s main actors’ metrics are presented in Table 2, with Ourofino Agronegócios presenting the highest measure of betweenness centrality; this company entered into the network in the second period and remained in the third period, is part of the giant component, and has established relationships with six different partners, including private companies and universities. Ourofino’s most recurring partner was Núcleo de Pesquisas Aplicadas, with a total of five relationships, two during the second period and three during the third period.
Main actor’s structural metrics—Brazil.
T1: period 1995–2000; T2: period of 2001–2006; T3: period of 2007–2012; G: giant component; D: degree; W: weighted degree; B: betweenness centrality.
The main actors of Spanish networks are presented in Table 3. We highlight that the highest metrics of betweenness centrality belong to the Spanish National Research Council (CSIC) and to the biotechnology company Newbiotechnic. The main network path is the connection between CSIC and Newbiotechnic, representing the link of two different sets of actors and contributing to the diameter and size of the giant component. CSIC, a state-owned R&D center located in Madrid, is a prominent partner of companies from Spain. It has established relationships with 12 partners, among them 7 different biotechnology companies, which were co-owners of 15 patents.
Main actors’ structural metrics—Spain.
T1: period 1995–2000; T2: period of 2001–2006; T3: period of 2007–2012; G: giant component; D: degree; W: weighted degree; B: betweenness centrality.
The Spanish company Newbiotechnic stands out compared to others. It filed 28 patents, 20 of which were under co-ownership, which reflects partnerships with other companies, universities, R&D centers, and funding agencies. There were nine different partners, three of which were foreign, highlighting the University of São Paulo, which appears as co-owner in a patent along with other universities from Spain and the United States.
Discussion of results
Biotechnology companies of both Brazil and Spain are establishing innovation-driven partnerships, observed by patent co-ownership in a joint application. In Spain, biotechnology companies have established more partnerships and with a greater diversity of actors, both national and foreign, which promotes access to useful knowledge and improves the innovation performance. 2,3 In Brazil, companies are very dependent on the knowledge generated by universities, which are the main partners of companies. In Spain, however, R&D centers, public or private, play a major role, while companies in Brazil lack that type of partner.
The aim of this study is to dynamically compare Spanish and Brazilian innovation networks of biotechnology companies. From the results, we see that both networks have evolved during the period of analysis, mainly in the third period. However, the results demonstrate some differences among the characteristics of networks from the two countries, highlighting a better performance of Spanish network evolution in relationship to volume and constancy. A basic difference concerning this evolution relates first to the increased number of new entrants during the period of analyses and second to the presence of some key actors in the network during two or more periods. The presence of them in the Spanish network not only attracted many other actors for this innovation network but also contributed to formation and growth of a giant component, interconnecting a set of different actors, thus contributing to knowledge diffusion among actors. 18,19 The analyses of actors’ position shed light on the type of key actors that influenced the giant component formation of the Spanish network, mainly to CSIC, an R&D Center, which impressively connected many of the other actors. The Brazilian network lacks the presence of actors with such a role. The results corroborate the findings of the recent study by Abbasi, 6 which shows that during an evolution process of a collaboration network, actors in intermediating positions (such as CSIC) tend to explain the preferential attachment of new entrants.
Conclusion
The study concludes that innovation networks of companies operating in different levels of industrial development, presenting not only different characteristics and patterns of evolution but also different types of actors in the network composition. The innovation network of biotechnology companies from Brazil demonstrates the recent evolution in the volume of participants and connections, although this evolution was not sufficient for the network to become more attractive to knowledge diffusion compared to the Spanish network, which initiated the biotechnology development in the early period. 8 The Brazilian innovation network presents many opportunities to linking unconnected communities, increasing the diffusion, and reducing the redundancy of knowledge.A recommendation for Brazilian policy makers is to strengthen actors with the capacity to connect different actors, thus intermediating relationships in the network, and there is also a need to find mechanisms to encourage knowledge exchange through the establishment of partnerships with greater partners’ diversity, including those of foreign origin. Access to knowledge, which is different from that which is locally found, can improve the capacity for innovation of biotechnology companies. Partnerships with foreign organizations can be an alternative to obtain access to the technological production of developed countries and should be encouraged.
Thus, this study has implications not only for public policy but also for managers who wish to improve access to resources for innovation. Stimulating learning through partnerships is critical for companies to develop capabilities to better compete. This is particularly true for dynamic industries, such as biotechnology, which depend to a greater extent on increasing innovation capacity. Comparing innovation networks of biotechnology companies from countries in different stages of development enables the understanding that new entrants (third-generation countries) must plan to be able to reach the same standards of developed countries as soon as possible.
As a contribution to the research area, the main addition of this study is to show that SNA applied to compare the evolution of companies’ interorganizational innovation networks from countries in different levels of development reveals patterns and differences over time, which may subsidize decisions toward innovation network improvement and show the corresponding stage of each country. These comparative studies are rarely found in the literature.
Our study focused on formal partnerships for R&D and resulted in patent applications under co-ownership. The use of these data may be seen as a limitation of this research. Further studies should focus on comparatively investigating the interorganizational partnerships and analyzing the formation and evolution of biotechnology companies’ innovation networks through qualitative research, seeking to identify other partnerships, including informal ones, thus expanding this study.
Footnotes
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by São Paulo Research Foundation (FAPESP) under the research project no. 2012/22686-9.