Modular Innovations and Distributed Processes The Case of Genetically Engineered Vaccines Antoine Bureth — Julien Pénin BETA, CNRS-UMR 7522, Université Louis Pasteur Strasbourg I 61 avenue de la Forêt Noire,F-67085 Strasbourg Cedex
[email protected] [email protected] The conception and the production of genetically engineered vaccines (GEV) constitute an highlighting example of innovations based on knowledge re-combination. We show that the development of a GEV is achieved through the combination at various levels (cognitive, organisational and technical) of sub-products or modules. Such modular configuration offers an important potential for innovation, but also raises the issue of the coordination of the different actors in charge of the modules. In particular we stress here the role of the patent system, which is essential to improve the interactions among firms and the assembling of heterogeneous pieces of knowledge. We therefore present the development of GEV as being a process distributed over a wide range of actors, in which patents are used as interfaces at the organizational level and as management devices to align heterogeneous incentives. ABSTRACT.
KEYWORDS: Intellectual Property Rights, Biotechnology, Modularity, Collaboration, Collective Innovation.
DOI:10.3166/EJESS.20.251-274 © 2007 Lavoisier, Paris
EJESS – 20/2007. Knowledge and Innovative Activities, pages 251 to 274
252
EJESS – 20/2007. Knowledge and Innovative Activities
1. Introduction This paper describes genetically engineered vaccines (GEV) as a specific example of innovation through knowledge re/combination. We show that the development of a GEV is achieved through the combination at various levels (cognitive, organisational and technical) of sub-products or modules. Such combination process is actually the most promising approach to generate new vaccines, and is made possible by the patent system, which plays the role of the interface among the modules. Patents enable each actors to specialise into one module and then to combine the different modules in order to generate innovation and to create new economic value. GEV use DNA manipulation to trigger an immune response. For instance, instead of introducing the complete antigen into the human body, it is the genetic material of the antigen (the DNA fragment that will code for this antigen) that is introduced into the cells of the patient. But genomic technologies are also mobilized for the production of traditional vaccines: for instance, cells of culture are engineered to produce the desired vaccine proteins. In the same perspective, the gene manipulation can be done directly on the pathogenic micro-organisms, in order to make them harmless and / or to use them as carriers. We consider here as GEVs all kinds of vaccines that at a level or another implies DNA manipulation. The field of GEV is worth to investigate since vaccines might renew interest, and offer promising applications for genomic approaches, especially for gene and cell therapies. Those therapies have raised huge hopes from a theoretical perspective but, due to several negative and unexpected results during clinical trials, gene therapy in peculiar now raises some reluctance. Scientific and technological unresolved problems as well as a negative perception by the general public, prevent financers from investing in the sector. Yet, it is widely acknowledged among experts that the development of vaccines using DNA techniques should upsurge the interest for those approaches. It is thus important to understand the main characteristics of the innovation process in the case of GEV. Specifically, we show here that those new vaccines exhibit a modular structure. Three types of modularity are taken into consideration: cognitive, organizational and technical modularity. The production of a vaccine requires combination of elements at each of these three levels. At the cognitive level, we stress that actors involved in the development of GEV use a mosaic of different pieces of knowledge. At the organisational level, we show that this dissemination of the knowledge goes along with a high fragmentation of the institutions and actors, each of them being involved only on a small segment of the development of the GEV. Finally, at the technical level we show that the product itself is technically decomposable. Several elementary bricks are combined in order to bring a reliable vaccine to the market: one antigen, one vector and one adjuvant.
Modular Innovation and Distributed Processes
253
Considering the production of innovations, the concept of modularity has been developed in industrial approaches, mainly in the field of activities of conception (Simon, 1962, Langlois and Robertson, 1992, Brusoni, Prencipe and Pavitt, 2001, Sako and Murray, 2002). Broadly speaking, modularity allows to improve and to valorize cumulative learning processes, innovations being incorporated in a product of an increasing complexity. Innovating in Live Science is somewhat different, in that it requires a kind of reverse engineering of natural systems: the objective is not to improve and to sophisticate the technical architecture of a product, but rather to master directly complex systems. In other words, the matter is not to trigger specialization in design and production activities, but much more to articulate heterogeneous (and already constituted) competencies. In that perspective, modularity differs from the notion of division of labor. The main issue thus becomes the coordination among the modules, i.e. the question of the interface. The quality of the interactions among the different modules is indeed central since it may foster or prevent the development of a project. With respect to GEV several types of interfaces exist: At the technical level, interactions among the different modules are ensured by molecular and biochemical technologies (Bureth et al., 2007a); at the cognitive level, articulation is ensured by the existence of standardised models, paradigms, procedures (guidelines, good practices), languages and by the codification of the existing knowledge (through publications, patents, etc.); at the organisational level, human capital is essential to establish informal interfaces, whereas contracts and agreements “connect” modules in a more tangible way. It is worthwhile to notice that in each of those types of interfaces, patents are mobilized. They enable information, communication, negotiation, value creation, rent extraction, alliances. As such, they are crucial in a field in which players are usually specialised into one specific module. Even if the ideal business model is to master all the modules, few firms are able to manage the complete structure (antigen – adjuvant – vector), which means that few firms are able to develop a GEV in an integrated way. Collaborations among firms are therefore essential and the development of new vaccines is typically a process of distributed innovation (Coombs et al., 2003). Consequently, an important issue (on which we do not insist in this paper) deals with the existence or not of an integrator in such a process. The big-pharmaceutical companies have the potential to be such integrators. Although they are rarely directly involved at the early stage of the process, they finance, control and orient noticeably the innovation process (Bureth et al., 2007b). But in many cases, innovation networks act without being submitted to a central authority, and firms (especially start-ups) have to evolve within a loosely coupled system of contracts. The central concern of this paper thus bears on the role of patents in a modular perspective. They are ideal instruments to ensure the interface among the different modules: patents preserve the autonomy of the partners and are flexible enough to adapt to various type of collaboration. Despite inevitable conflicts, we show that they crucially improve the interactions among actors belonging to different
254
EJESS – 20/2007. Knowledge and Innovative Activities
technological trajectories, different fields of knowledge and different organizations (Levin et al., 1987, Cohen et al., 2000, Jaffe, 2000, Bureth et al., 2005 , Cohendet et al., 2006). As such, patents become architectural elements of modular innovations. Without patents firms would find it much more difficult to protect their knowledge but also to collaborate and to articulate their competences, making the vaccine production much harder. In other words, we propose a vision of patents as devices of co-opetition. On the one hand patents allow their owners to compete within each module. On the other hand patents improve the interactions and the collaborations among the different modules. In both cases they “flag” the innovation development paths and are instruments of incentives and interaction. Our analysis of the modular structure of GEV is supported by an empirical study that relies on two complementary methodologies: (1) 28 in depth interviews of French actors in the field of GEV (Start-ups managers, public researchers, capital venture executives, etc.). (2) A network analysis based on the data provided by the Recap database on interfirm collaborations. Those works were realized in the frame of a project entitled MIDeV (Modularity and Incentives in the DEvelopment of genetically engineered Vaccines). In this paper we will only provide a short description of the dataset. The detailed results of the empirical analysis are described elsewhere (Bureth et al., 2007a and 2007b). Section 2 introduces the basics and the main economic features of GEV. Section 3 focuses on modularity and stresses why the development of GEV can be described as a modular process. It also contains a short description of the data on which our contribution relies. Section 4 studies the role –positive and negative- of the patent system as an interface of the system at the organisational level. Finally the last section concludes and analyses the impact of our work on the industrial organisation in the field of vaccines. In particular, we argue that the development of GEV is a case of distributed entrepreneurship (McKelvey, 1998). 2. Economic and technical aspects of GEV 2.1. Definition: GEV vs. traditional vaccines A vaccine is an antigenic preparation used to produce active immunity to a disease1. In other words, a vaccine stimulates the production of antibodies and 1. The principle of vaccination dates back to E. Jenner (1796) and the vaccine against the deadly smallpox virus. Jenner observed that milkmaids are often infected with cowpox, which is a mild relative of smallpox, through their interactions with dairy cows’ udders. Jenner had thus the idea to take infectious fluid from the hand of a milkmaid and to insert this fluid into the arm of a healthy boy. The later then showed symptoms of cowpox infection. Forty-eight days later, after the boy had fully recovered from cowpox, Jenner injected some smallpoxinfected matter into him again. The boy did not show signs of smallpox infection. Jenner concluded that the boy had developed himself his own protection against the disease.
Modular Innovation and Distributed Processes
255
therefore enhances the immune response of the organism. Vaccination techniques use the natural tendency of our organism to destroy unknown foreign agents. The immune system recognizes vaccine agents as foreign, destroys them, and memorizes them. Then, when the pathogenic agent appears again, the body is prepared to react. The main characteristic of vaccines is thus to induce an endogenous protection. What is brought from the outside is just the substance (a killed pathogenic organism, or an attenuated form of that organism) that helps the body to improve its defence. In that line, GEV use DNA manipulation to trigger the immune response. For instance, in the case of pure DNA vaccines, instead of introducing directly the antigen within the human body, the genetic material is placed into a cell in order to produce within the human body the antigen that will then stimulate the immune system of the patient. In short, it is the patient’s own cells that produce the vaccine. Another possibility offered by DNA techniques is to use transfection techniques in order to produce the antigen ex vivo, using specialized cell (bacteria, but also plants or animals). The antigenic protein is thereafter isolated by purification, and administrated to the patient, in order to activate an immune response. As stated in the introduction, we consider as GEV all kind of vaccines that implies DNA manipulation for its production or its administration. Peculiar types of GEVs are therapeutic vaccines. In the case of cancer, for instance, current research try to develop processes that would allow a person's immune system recognizing and destroying malignant cells without harming normal cells. Such cancer vaccines are considered as an immunotherapy insofar as they rely on the stimulation of the immune system to cure an already contracted disease. Unlike prophylactic vaccines against diseases such as polio and tuberculosis, they are not preventive but must be administered after cancerous cells start to develop. Compared to traditional vaccines, GEV are attractive due to their simplicity (in the theoretical principle), robustness and wide scope of application. They use relatively standard genomic techniques that can theoretically be applied to the vaccination of many diseases, and thus support product diversification. Second, some GEV (especially pure DNA vaccines) can be stored over long period of time. Conversely to traditional vaccines, they do not need refrigeration, making them potentially attractive for developing countries. Finally one of the main advantages of GEV as compared with traditional vaccines is the diminution of the risk for the patient, since there is no need to inject the pathogen agent within the organism in order to develop a protection. Yet, although quite promising the GEV approach faces several problems. One of the main issues is the carriage and delivery of genetic material into the cell’s nucleus. This problem can be decomposed into several steps, each of them corresponding to the crossing of a specific “cellular barrier”: The first difficulty is to enter the cell, i.e. cross the membrane, which envelopes the cell and controls what moves in and out. The second difficulty is to carry the genetic material within the cell in order to reach the nucleus. The third difficulty is to enter the nucleus and to
256
EJESS – 20/2007. Knowledge and Innovative Activities
maintain the genetic material inside the nucleus. At this step two alternatives are possible: either the genetic material introduced into the cell does not integrate with the genome of the host cell and merely orders the production of the needed protein; or the DNA integrates within the cell’s genome and may conduct to permanent hereditary consequences. The crossing of these different borders is further made complex by the huge quantity of DNA needed for inducing immune responses. The production and purification of large quantity of DNA is not only costly but also increases the risks of side effects. These problems may explain why, to date, most DNA immunization studies have been carried out on mice and not yet on human beings. 2.2. The market for vaccines: opportunities for GEV? The market for vaccines is segmented (by countries and by diseases) and concentrated, with only few big players that are all divisions of global pharmaceutical houses (Merck, GSK, Aventis, Wyeth and Novartis). Those major vaccine producers generate around 85% of the sector’s sales2. To compare, in 1955, 18 firms shared the market. The principal focus of the actors in this field is paediatric vaccines, which represent about 70% of the market3. To understand the peculiar industrial structure that is actually observed (specialization of one or two producers per vaccine/per country) it is important to remember some specific features of the vaccine market (Danzon et al., 2005): – Vaccines markets are small “by definition” due to the winner curse: contrary to what is observed for the treatment of chronic illness, the longer the efficacy, the smaller the demand. Furthermore, cultural habits and national regulation mechanisms keep the vaccine markets country- (and sometimes regional-) specific. – The public intervention, which exists in many countries concerning vaccines against the principal contagious diseases, causes a concentration of the demand and a reduction of the prices. – Vaccines aim at treating large number of healthy people. Consequently the risks of liability are largely increased compared to therapeutic drugs. The trials must be conducted on a very large scale to demonstrate absence of rare events, and the financial consequences of a failure are higher than in standard therapeutic approaches. Costs of development of new vaccines are therefore increasing, especially due to safety requirements from the public authorities4.
2. Datamonitor, Strategic Perspectives: Vaccines (2003). 3. Scrip Reports, The World Vaccines Market (2002). 4. New constraints on the production of vaccines have contributed to increase the cost. For instance, the case of Thimerosal, a conservative derived from mercury, whose utilisation has been forbidden by public authorities obliged vaccines companies to fraction their lots, which
Modular Innovation and Distributed Processes
257
In short, vaccine producers must deal with a strong scissor effect, due to high fixed costs on one hand and to low prices on the other hand – explained by limited markets, public price regulation, low-income vaccines users, etc Two alternatives are possible to solve that tension: the first is to cover the innovation costs by prolonging as much as possible the years of sales. Under that perspective, the vaccines field stagnates for decades, mobilizing few new technologies and generating weak revenues. The second is obviously to focus on more profitable vaccines and to reduce the production costs of existing vaccines, or better said, to modify the production constraints attached to the product. As we will see below, biotechnologies offer such new opportunities on this second path of development. Indeed, even if it has long been considered as a small segment of the pharmaceutical market, the economic status of vaccines is changing. The bioterrorist threat, the threat of new pandemics (chicken flu for instance), recent progress in cancer research and the impulse given by the non profit sector (the Gate foundation for instance) contribute to restore the image of vaccines and to improve the perspective of this market (Orsenigo et al., 2006). Annual vaccine sales have grown from $2 billion in 1990 to an estimated $11 billion today. This remains low as compared to the $550 billion in 2004 for the whole pharmaceutical market. But the market for vaccine, in opposition to the classical medicine market is expected to grow significantly in the next years5. Much of this predicted growth is expected to come from the introduction of new vaccines, either against diseases for which no vaccine currently exists or as second-generation products to replace existing vaccines. The economic potential for those new types of vaccines is huge since they could pretend to higher prices due to their efficacy to treat important diseases in developed countries (cancer for instance)6. This evolution of the vaccine market gives opportunity to new biotech firms based on scientific excellence to enter this market. One can already see the emergence of new players (Baxter, Acambis, MedImmune) which, despite their
increase the cost of production and may trigger problems in the management of stocks (penuries). 5. Merck, who will market soon a vaccine against cancer of the neck of the womb, assesses that the overall vaccine market should reach $18bn in 2009. Similarly, Novartis, who has recently penetrated the vaccine market through the buyout of Chiron, assesses the market to reach $20bn in 2009. As an example of the new economic status of vaccines, the market has experienced its first blockbuster in 2005, with the sales of the Prevenar® (produced by Wyeth and put on the market in 2000) that have passed 1$bn sales. 6. Yet, it is difficult to establish precisely how many GEV are in the pipeline of the industry. If we stick to pure DNA vaccines (using plasmid to trigger the production of an antigen by the patient’s immune system), a dozen of projects are in human development worldwide, but none has gone beyond phase 2 trials (Forde, 2005). Tacking into account the use of other vectors, or new modes of production of the vaccinal compound, many potential vaccines are in late stages of development (clinical trials phases 2 or 3) and could enter the market soon (Braunagel and Das, 2003).
258
EJESS – 20/2007. Knowledge and Innovative Activities
small size, may expect strong growth based on innovative products, thus modifying the industrial structure of the market. The genomic revolution encourages this evolution of the industrial landscape. Orsenigo et al. (2001) stressed how the new biotechnologies impacted drug design and even modified the nature of the search space. In the case of vaccines, genomic techniques even might, to a certain extent, reduce barriers to entry. We observe a form of product differentiation and generation of variety that was not possible under the old pharmaceutical paradigm. Those strategies of differentiation are made possible by the modular feature of GEV, which triggers in depth changes in the organisation of the sector. Thanks to modularity, new biotech firms can now enter the market by focusing into one module and then partnering with big incumbents. 3. GEV and modularity 3.1. What is a modular system? Since modularity is discussed extensively elsewhere in this issue, let us briefly remind here the main outlines of the concept. A modular system can be defined as a complex system, each parts of which conceived independently but functioning together in a homogenous way (Langlois and Robertson, 1992, Brusoni et al., 2001; Sako and Murray, 2002). Modularity refers to the existence of modules (sub-part of the whole system) that when combined together through well-defined interfaces fulfil a global function. The opposite of a modular system is a perfectly integrated system, which cannot easily be decomposed in autonomous sub-systems. Such a conception of modularity is clearly derived from the “loosely-coupling” approach developed by Simon (1962). A modular system is therefore made up of two things: Modules (subparts of the system) and an interface that ensures the connection of the modules. The aim of a modular system is to achieve an interface that minimizes interdependencies between modules performing different functions. This central point is raised by Koppl and Langlois (2000, p. 18): “Modularity is not about cutting a system into parts. All systems are already made up of parts. Modularity is about how parts are grouped together and about how groups of parts interact and communicate with one another”. Modularity can be considered at three levels: technical, organisational and cognitive. First, modularity can deal with the technical architecture of products. This perspective, already present into the pioneer work of Simon (1962), has been the most explored by scholars. Technical modularity refers therefore to the possibility to decompose a product into independent sub-products that can then be recomposed in multiple ways. Second, organisational modularity refers to a specific kind of division of labour and specialisation. It supposes horizontal relations among workers connected to a standardised platform, as opposed to hierarchical organisations that are based on vertical relationships. Third, cognitive modularity deals with the
Modular Innovation and Distributed Processes
259
division of knowledge. It refers to the possibility of decomposing the knowledge necessary to conceive, build and market a product into independent sub-parts. Cognitive modularity raises the issues of communication and language to transfer knowledge. In case of stabilized technologies, those three kinds of modularity frequently converge since, for instance, technical modularity may induce organisational and cognitive modularity. Yet, considering GEV, nor this overlap neither the link of causality are clearly established and strong gaps between the different partitions can even generate tensions and hinder the development of innovations. The benefits of achieving modularity are mainly three-folds: – First, a modular system may spare resources on management by enabling the decomposition of a complex system into several simple sub-systems. Yet, too much modularity may increase the cost of management since it may also increase the cost of communication among modules. This suggests that there exist an optimal level of modularity for a system. – Second, since each module can be conceived independently from the others, a modular system will not be affected in its totality by external perturbations and unforeseen events. – Third, the overall efficiency to fulfill a global goal will benefit from improved learning process. Indeed, modularity leads to separate the learning about the architecture of the system, from the learning about the features of the modules. This dichotomy allows an increase of diversity generated by recombination of the modules, and, along with the previous point, reinforces the capabilities of the whole system to absorb, to integrate and to valorize innovations and ruptures generated locally and incrementally. In the production of GEV, modularity may help to manage complexity, uncertainty and the generation of variety. In other words, a modular structure improves performance along three dimensions, in terms of efficiency, adaptation and creation. Modularity allows rationalizing the management of a set of differentiated activities. The development of GEV relies on an important variety of scientific and technical fields. The actors mobilized are extremely heterogeneous. In other words, the collective innovation process is costly in terms of coordination and transaction costs, and it is almost impossible for a single actor to master it globally. Modularity – i.e. autonomous sub-parts interconnected through stabilized interfaces – is a necessary condition to allow the establishment of leadership, and to align incentives. A modular system exhibits stronger adaptation capabilities. Indeed, modularity provides an answer to the attrition constraints and reduces uncertainty, by enriching the portfolio strategy. It limits the risk of dead ends and unexpected perturbations, insofar as recombination of modules offers a greater variety in the potential paths of development.
260
EJESS – 20/2007. Knowledge and Innovative Activities
Modularity improves the creation of novelty. In terms of economies of variety, generation of diversity in vaccines can be achieved by declining a platform. The possibility of recombining the different modules at the technical and cognitive level considerably lowers the cost of incremental innovations. Such a possibility of late differentiation is crucial in a sector in which the high fixed costs combined with regulated prices limit profitability. Strategies of rationalization, adaptation or creation are adopted depending on the co-evolution of knowledge, market conditions and players positioning. They are not specific to one type of actor, or one stage of development, and can be implemented in a decentralized way, at different decisional levels within an innovation network. 3.2. The case of GEV These advantages suggest that there is a strong incentive to benefit from a modular structure, which allows economies of variety through the recombination of the modules while increasing learning through specialisation within each module. Now, it makes sense to think of the development of a GEV as a modular system at the three levels –technical, organisational and cognitive. First of all, at the technical level it is possible to decompose the product into three main quasi-independent subparts: antigen, vector and adjuvant. To develop a vaccine one needs first an antigen, which is the substance that stimulates the production of antibodies. Antigens mark foreign agents in the organism. It is the reconnaissance of the antigen by immuno-competent cells that active the immune reaction. Our interviews suggest without ambiguity that the module antigen is the most important to develop a GEV (Bureth et al., 2007a). It has consequently a strategic position. Second, to produce a GEV it is also necessary to be able to carry naked DNA within the human cells. This requires the use of sophisticated transfection techniques in order to cross the frontiers of the cell that were described above (to enter the cell, to enter the nucleus etc.). Transfection involves opening holes in cells to allow the entry of molecules. Transfection differs from transformation since the DNA is not generally incorporated into the cells’ genome but is only transiently expressed7. There exist various methods of introducing foreign DNA into a cell: electroporation, heat shock, gene gun or the use of viruses as carriers. Third, developing a GEV also requires using an adjuvant, which is a substance that enhances the immune response. In immunology, adjuvants stimulate the immune system, thus increasing the response to a vaccine, while not exhibiting antigenic effect when given by themselves.
7. This distinction is central. Nowadays, transfection is allowed but transformation is not for ethical and safety reasons.
Modular Innovation and Distributed Processes
261
To summarise, at the technical level the development of GEV requires combining three main elements. We have therefore the following simple equation: GEV = one antigen + one vector + one adjuvant This decomposition of the product is linked to the progress in genomics, which facilitates the autonomisation of each module. The first traditional vaccines were conceived and produced in an integrated way without decomposing the product, the knowledge or the organisation of labour. Conversely, nowadays, actors in the field of GEV are very heterogeneous and deeply specialised in a given module. Those three technical modules are relatively autonomous, i.e. can until a certain point be considered independently one of each other and, more important, can be recomposed in different ways to generate new products (at least in theory). Broadly speaking, a vector can serve for many antigens and one antigen can be carried out by many vectors. Similarly, one adjuvant may enhance the performance of several different antigens. From a given amount of existing antigens, vectors and adjuvants, this combination process can therefore lead to many new vaccines and at reasonable costs. Yet, in practice this modular property is coming up against the fact that modules are not perfectly independent. Each module needs specific adjustments before being combined with another one. Technical modularity in the architecture of GEV goes along with a form of organisational modularity. Division of labour and specialisation is indeed greatly facilitated by the product decomposition. New biotech firms, that are too small to undertake all the steps to market a vaccine can specialise into one module. Modularity fosters therefore strategies of product differentiation which, as explained earlier, lowers the costs of entry and favours the emergence of a new market structure in the field of vaccines. Empirical analysis strongly supports this organisational decomposability in the field of vaccines (Bureth et al., 2007a and 2007b). Only big-pharmaceutical firms can undertake in an integrated way all the steps to bring a vaccine on the market (from fundamental research to the distribution and marketing of the vaccine). New biotech firms, which rely on scientific excellence, can enter the market only through specialisation on a given activity (transfection, adjuvants or antigens). Strategies of development for those new firms are clearly based upon modules rather than on therapeutic applications, as it was usually the case in the traditional pharmaceutical paradigm. Yet, big-pharmaceutical companies manage to keep their leadership on the market by locking the access to the final market. New biotech firms do not control all the components needed to put a vaccine on the market and are usually obliged to partner with big-pharmaceutical companies. We find here a model widely observed in life sciences: start-ups grow on the basis of a small scientific discovery and then interact with more mature firms. Genentech, Amgen, Celera, or Millenium, Incyte and Affymetrix are famous examples of this kind of strategy.
262
EJESS – 20/2007. Knowledge and Innovative Activities
Finally, with respect to GEV modularity is also relevant at the cognitive level. Developing a GEV involves a mosaic of different knowledge in various domains. The following scientific disciplines are involved in the development of GEV: immunology, molecular biology, genomic, virology, bacteriology, oncology, toxicology, pharmacology, biochemistry, etc. Obviously it is impossible for a single actor to control all these fields of knowledge, i.e. the ability to articulate and coordinate scientific disciplines becomes strategic. The dispersion of the knowledge mobilised requires complex coordination mechanisms to develop a GEV. An approach in term of modularity may therefore help to deal with this complexity. Advances in scientific research allows to build elementary bricks of knowledge quasi autonomous. Then, with the help of legal instruments such as intellectual property rights, those bricks can be valorised on a market. It is worthwhile to underline that this modular representation goes beyond the specialization trajectories and localized learning effects that are often referred to in the literature on innovation processes. Emphasize is put on the capabilities to build, to manage and to valorise interfaces, rather than on the efficiency of isolated learning processes taking place within pre-defined borders. In other words, the way knowledge is exchanged/technical components are coupled/organizations are interacting dominates the way knowledge is produced/products are manufactured / organizations are designed. Whereas specialization provides an understanding in a micro perspective, the concept of modularity is better suited to grasp the collective dimension of the innovation process. To conclude, in the case of GEV we envisage modularity at the three levels, technical, organisational and cognitive. Yet, only modularity at the organisational level can be clearly observed empirically, with a wide dispersion of actors specialised in specific modules. Cognitive modularity is hardly observable in reality and technical modularity, although strongly supported by our interviews, cannot be observed yet, due to the relative youth of the sector. Too few GEV are on the market to validate this hypothesis. 3.3. Our empirical study: The MIDeV project Our analysis of the modular structure of GEV is supported by an empirical study that relies on two complementary methodologies: A campaign of interviews and a network analysis based on the alliances of French vaccine’s firms. Those works were realized in the frame of a project entitled MIDeV (Modularity and Incentives in the DEvelopment of genetically engineered Vaccines). In this section we provide a description of the dataset and of the methodology we used to build it. The detailed
Modular Innovation and Distributed Processes
263
results of the empirical analysis are described elsewhere (Bureth et al., 2007a and 2007b)8. In a first stage, 28 interviews of major French actors involved in the field of GEV have been conducted. The panel of specialists we interviewed contains public researchers, start-up founders, capital venture executives, regulation offices, etc. We believe this panel is representative of the French population active in the field of GEV, due to the small size of this population. In France, there are only two or three key players with respect to GEV and hardly more than two dozens of start-ups. Although prepared on a directive basis, some parts of the questionnaire were largely open in order to let room for discussions and to be able to gather more informal information. The interviews were all conducted on a face to face basis, with sometimes a further contact by telephone with another executive of the company. Different questionnaire was designed with respect to the organization that was interviewed. Basically, three different questionnaires have been made up: one for public research organizations, one for firms (start-ups and big-pharmas) and one for regulation offices. For each version we collected information on the main characteristics of the organization, on the perspective of the field of GEV (both from a technological and economic point of view), on collaborating activities, on patenting activities, on the financing of innovation, etc. Thanks to this campaign of interviews, we have therefore at our disposal qualitative detailed information on the reasons why firms patent or collaborate, on the organization of the innovation process in GEV (our interviews confirm strongly the modular aspects of innovation), on the main modules, on the role of science, etc. The main results of these interviews are threefold: The actors have all validated the technical decomposability of a GEV in three main subparts or modules: one antigen, one adjuvant and one vector. Furthermore, it comes out from the interview that the module antigen is the most important. Second, the actors in their majority emphasize the role of the patent system to bring some cohesion into the innovation process. Finally, our interviews stress the major role played by the academia in the field of GEV. Most start-ups are academic spin-offs and firms could hardly exist without scientific collaboration scientific labs. In complement to these interviews, we have used several databases on alliances in order to complete an analysis of the collaborations of French firms involved in the field of GEV. The main database we have used to extract data is Recap (Recombinant Capital), which contains information on 13 000 alliances in biotech since 1978. This database contains very rich information: the name of the participants, the year, the objective of the alliance, the nature of the alliance (licensing, joint venture, etc.), a short description of the alliance, the therapeutical
8. For sake of confidentiality, we do not mention firm’s name when an information was gathered during an interview. Yet, when the information was found on the web or was drawn from public databases there is no reason not to mention firms by name.
264
EJESS – 20/2007. Knowledge and Innovative Activities
axis, etc. We have further enriched these data by several research on the Internet, especially on the website of 150 biotech firms, and by extractions from the database of scientific publications Business Source Premier. The search on firm website allowed us, among others, to find firms’ activity and to put each firm into one of the three following groups: antigen, vector or adjuvant. At the end we have identified 241 alliances that involve 30 French GEV firms (the complete network, which is obtained by adding the alliances among foreign partners, is bigger). The main results of the network analysis can be summarised as follow: First, most French firms are highly specialised in one specific module (antigen, vector or adjuvant), which is compatible with the hypothesis of modularity at the organisational level. Second, most of the alliances that involved French firms contain licensing agreements, which suggests a central role played by patents in order to contract and structure alliances. Third, there is less alliance (and less licensing agreements) within modules than between modules, which sustains the hypothesis of patents as instruments of co-opetition. Fourth, the connectivity of the network is more important when it includes big-pharmaceutical companies. This is compatible with the hypothesis that those companies play a central role of integrator in the development of GEV. 4. Interface at the organisational level: The role of the patent system The modular aspect of GEV provides a strong basis for innovation generation just through combination and recombination of existing knowledge and subparts of the products (antigens, vectors, adjuvants). This modularity raises the question of the interaction between the different modules. How players combine to arrive to produce a vaccine? Which kinds of arrangements are implemented? We focus here on the role of patents to ensure the coordination among the different modules at the organisational level. Patents are the ideal interface in the case of GEV because they can play on the two faces of protection and coordination. On the one hand they protect the modules, which makes it feasible for firms to valorise independently the production in a given module, and on the other hand they facilitate the interactions between modules through inter firms’ collaboration and technology trading. 4.1. Patents as instruments of co-opetition With respect to GEV it is possible to present patents as instruments of coopetition in the sense that they simultaneously favour competition and collaboration: Within each module they are used as tools of exclusion, but between the different modules they are used as instruments of coordination.
Modular Innovation and Distributed Processes
265
4.1.1. Competition First, patents are a tool of competition within each module in which firms are rivals. Inside modules we can therefore observe a traditional use of patents as tools of exclusion and protection against imitators. Firms involved in vector production, for instance, usually do not collaborate with other firms producing vectors. They rather use their patent portfolios to exclude other vector producers from imitating them, so that they enjoy a monopoly position on their vectors and can pretend to higher prices. Our interviews confirm this aggressive use of patents and stress that without patents most firms would be reluctant to invest in R&D (Bureth et al., 2007a)9.
Figure 1. Co-opetition in the development of GEV
Two important remarks may nevertheless put into perspective this traditional view of patents as devices to exclude: First, vaccine development does not involve only codified knowledge. Vaccines are far from corresponding to the traditional chemical paradigm that occurs in the pharmaceutical industry. To develop a vaccine firms must also possess the know-how to consistently produce a safe and effective
9. This view of patents is conform to the traditional use of patents in life science in which it is well-known that patents are highly important to appropriate the returns of an innovation and to enhance incentives to do research (Levin et al., 1987; Cohen et al., 2000). Firms acting specifically in biotechnologies are small firms faced with high competitive pressures. Consequently they strongly rely on patents, which are their only tangible asset. Patents are the only element to valorise in front of potential partners and financers.
266
EJESS – 20/2007. Knowledge and Innovative Activities
biological product. Hence a vaccine cannot be as easily imitated as a chemical component. This reliance on know-how, which is of course not covered by patents, may attenuate the role of patents as being essentially a device to exclude. Firms in vaccine may not need patents’ protection since they are already protected by their secret know-how. Second, financial constraints on small biotech firms may encourage the latter, even within the same module, to cooperate to some extent. There is indeed a convergence of firms’ interests because the success of one firm is likely to distress the financial constraints that are exerted on all biotech firms. Financers are currently in a waiting position. Yet, as soon as one firm will commercialise a product, it is likely that investors will again inject funds in the sector, which will therefore benefit to the whole industry, including competitors. However, those two features may introduce some collaboration but overall they do not counterbalance the competitive situation within modules in which patents are mostly use as tool of exclusion, as described by the traditional arrovian vision. 4.1.2. Coordination Completely different is the situation among modules. Here firms are not rivals but clearly allies that must collaborate to achieve a common goal. Firms in different modules must combine their competences in order to produce a vaccine. Those interactions among firms in separate modules are facilitated by the existence of the patent system. Patents ease coordination and even collaboration between firms because they hold concurrently two important properties: They both disclose and protect an innovation. Whereas the protection dimension of patents is widely acknowledged, their signalling dimension is less known. Yet, patents signal competences to the world. When an innovator applies for a patent he must provide a description of the innovation, which must enable its reproduction by any person knowing the “state of the art”. Eighteen months after the application this description is published, i.e. becomes public. The patent system contributes therefore to the creation of a public database that contains most of the technological competences in a wide area of technological fields. Beyond the technical information they deliver, patent databases also provide information about the actors present in the field. This information concerning the “know-who” is often essential to find partners or to avoid entering technological fields that are already explored by competitors. Patents thus enable firms to establish contacts in a field in which the multiplicity of small and heterogeneous actors may complicate the identification of partners. The signalling dimension of patents also enables patentees to collect funds or to hire bright students more easily (Pénin, 2005). The patent system is therefore a key instrument to facilitate contacts among actors within a sector in which highly differentiated domains of activities imply high transaction costs.
Modular Innovation and Distributed Processes
267
As showed in previous studies, it is the coupling of these two properties – disclosure of the knowledge underlying the invention and protection of this invention- that makes patents central elements to ensure coordination within an industry (Bureth et al. 2005). Patents can help the interactions among modules at two different levels: (i) They enable technology trading, through licensing agreements for instance; (ii) They are central to frame inter-firms collaborations. (i) Technology trading. First, patents help technology and knowledge trading. Firms specialized in a given module can produce knowledge, patent their results and then sell them through licensing contracts that specify the price and the terms of the transaction. Such exchanges could hardly emerge without the existence of the patent system. Patents both signal where competences are located and protect those competences, thus preventing free riding from occurring. The reconciliation of the appropriation and the revelation of information solves the Arrow’s paradox (1962). It enables innovators to sell their innovation and prevent them to be “hijacked”. Therefore, paradoxically, property rights may often favour knowledge transfer. In a sense, the patent system allows the creation of a market for technologies and highly codified knowledge. Yet, patents may also sometimes allow the transfer of tacit knowledge by including in licensing contracts clauses of assistance, of exchange of employees, etc. As explained by Foray (2004, p. 136): “Patents create transferable rights and can therefore help to structure a complex transaction that also concerns unpatented knowledge”. In the case of GEV, this role of patents is essential. New biotech firms can specialise into one module, patent their results and then valorise them on a market for technology. Empirically we indeed observe many patent licensing agreements among firms from different modules. Our interviews all confirm this feature. Furthermore, our network analysis reveals the importance of patent licensing as being the predominant mode of collaboration in the vaccine domain. It also suggests that firms are significantly more likely to license in and out to firms in other modules than to firms in their own module (Bureth et al., 2007a and 2007b). Notice that patents may not always be exchanged against money. They are also often bartered against other patents. In this way patents are used defensively in order to prevent their holders from uncertain and risky lawsuits (Grindley and Teece, 1997, Rivette and Kline, 2000). In sectors in which technologies are overlapping and in which innovations are most of the time incremental, which is typically the case of vaccine, firms are likely to be blocked during their research by other firms’ patents. Expecting such situations, firms are therefore induced to gather important patent portfolios that will serve as “legal bargaining chips” and will be bartered when firms need to use technologies that are protected by patents held by other firms. (ii) Interfirms collaborations. Beyond facilitating technology trading through licensing, patents also ease more integrated inter-firms collaborations such as research joint venture. Indeed, R&D cooperation is a risky process in the sense that
268
EJESS – 20/2007. Knowledge and Innovative Activities
participants must often share parts of their most important intellectual assets. Since patents protect the knowledge held by a firm from plundering by her partners, they decrease the risk of opportunistic behaviours and of hold up of competences. It follows that firms protected by patents may be more willing to be involved in R&D cooperation (Ordover, 1991). Patents are also useful bargaining devices to set up the terms of the interaction. They enable to assess the competences of each partner, i.e. they provide a benchmark that allows firms to compare their relative competences. Moreover, as they represent a credible threat to block the entente, they limit the effect of size or financial asymmetries. Finally, patents facilitate the coordination between sometimes very heterogeneous actors because they represent a common language that can be understood by all of them (public labs, big multinationals, consulting agencies, stockholders, etc). Our interviews confirm this role of the patent system in the field of GEV. For instance, one of the firms we interviewed is involved in the development and commercialisation of chemical based vectors that serve to transfer genes or other bio-molecules within cells in-vivo or in-vitro. This firm belongs therefore to the vector module. Now, the firm is currently engaged in an important project of vaccine against bladder cancer with US partners that belong to the antigen module. The firm clearly acknowledged that without patents this collaboration would hardly have been made possible. First it is by scanning patent databases that partners came to know each others. Second, patents played a central role during the bargaining of the term of the cooperation. According to the firm it is therefore likely that without the patent system, the collaboration would have been much more costly to implement. To summarise, we emphasised in this section that patents may support the formidable potential of innovation induced by the modular structure of GEV. New vaccines can be generated through recombination of existing knowledge and products and the patent system provides and adequate interface among the different modules. However, if patents can improve the interactions among modules, it is also possible that they have the opposite effect when they are used abusively. We explore this eventuality in the next section. 4.2. New organisation in the vaccine industry and tragedy of the anticommons This role of patents as sustaining a new division of labour was already emphasised during the biotech revolution at the end of the last century. The emergence of biotech led to a new organisation in drug production. Instead of an integrated production in which big pharmaceutical companies were screening new compounds, testing and commercialising them, the biotech revolution has led to a division of labour between the research of new active compounds on the one hand and the testing and commercialising of those compounds on the other hand. Based on scientific excellence new biotech firms discovers new compounds, apply for
Modular Innovation and Distributed Processes
269
patents and then license and sell them to big pharmaceutical companies, which have the ability (financial and organisational) to test and commercialise the drug on a marketplace, thereby earning some revenue. The role of the patent system is obvious in this process. Big pharmaceutical companies are the only actors to generate revenue from the market place. Without patent protection new biotech firms would never engage into costly research, which means that big pharmaceutical companies would have to undertake those researches by themselves, which would be more costly than to buy licence to new biotech firms. Patents here allow a division of labour enabling each actors of the innovation process to concentrate on the aspect in which it is the most efficient. They are the interface between new biotech firms and big-pharmaceutical companies. Within the development of GEV, the division of labour is still push forward in the sense that patents enable not only to distinguish between the research stage and the development and commercialisation stage but also to specialise during the research stage. New biotech firms are leading complementary research on different modules and then sell their research to firms that will have to develop and produce the finite vaccine. This means that firms at the end of the chain do not have to interact with only one research company but with several start-ups, each of them specialised in one single module. Yet, this modular structure, which fosters innovation by allowing division of labour and by easing knowledge and product recombination, may encounter one important problem due to agressive patenting behaviours. We are indeed here in the exact configuration described by Cournot in its seminal work in 1838 about the properties of models dealing with complementary intermediate goods. Cournot demonstrated what is nowadays called the “tragedy of the anticommons” (Heller and Eisenberg, 1998), an expression that relies on the notion of “tragedy of the commons” stressed by Hardin (1968). As stated by this biologist, the lack of property rights on a common good can induce to use this good above its regenerative capacities and thus can lead to its entire destruction, thus requiring public intervention to regulate the use of the common good. The idea of the anti-common tragedy fosters on the exact reverse problem, which is the consequence of the multiplication of property rights on a single product. In the case of complementary goods there is a risk of suboptimal use of resources linked to the addition of monopoly situations that may increase the overall price to use the resources. A tragedy of the anti-commons therefore means that a good is underexploited due to an excessive price induced by the addition of monopoly positions on its components. With respect to the case of GEV, the addition of monopoly position on each module that composes a vaccine may sharply increase the overall price of production of the vaccine and at the end, may jeopardize its production. There is therefore a risk that socially desirable vaccines are never developed. In this case it would be in the interest of society that policy makers implement mechanisms to
270
EJESS – 20/2007. Knowledge and Innovative Activities
correct this disfunctioning. For instance, in the case of complementary goods, it has been shown that patent pools are welfare increasing (Lerner and Tirole, 2004). Yet, up to now it seems that the risk of being trapped in a tragedy of the anticommons has been avoided by the actors in the field. None of the person we interviewed mentioned this problem, although some scientists acknowledged that they are sometimes bothered in their research by patents held by others. It seems therefore, as explained by Walsh, Arora and Cohen in a study on research tools in life science, that actors are able to set up “working solutions” to solve the problem of the anticommons: “We find that there has in fact been an increase in patents on the inputs to drug discovery (“research tools”). However, we find that drug discovery has not been substantially impeded by these changes. We do not observe as much breakdown or even restricted access to research tools as one might expect because firms and universities have been able to develop “working solutions” that allow their research to proceed. These working solutions combine taking licenses, inventing around patents, infringement (often informally invoking a research exemption), developing and using public tools, and challenging patents in court” (Walsh et al., 2003, p. 286) However, the eventuality of an anticommons tragedy exists, especially in fields where many patents are granted. Indeed, if scientists usually do not question the utility of the patent system in the field of vaccine, they warn against abusive and too intensive use, which could render the development of “working solutions” too costly. We must therefore remain vigilant if we want to preserve the huge, but fragile, potential of innovation that is made possible by modularity. 5. Conclusion: towards distributed entrepreneurship in the vaccine industry In this paper we studied the development of genetically engineered vaccines. In particular we showed that GEV exhibit a modular structure at the technical, cognitive and organisational levels. This modularity entails a huge potential of innovation (mostly incremental) just by recombining existing elementary modules. To generate innovations the quality of the interactions among the different modules is therefore central. With this respect we then focused on the role of the patent system as providing an adequate interface between the modules. Patents are indeed ideal for this role since they manage to (i) protect and define the modules and (ii) ease the exchanges and collaborations among firms. A strong conclusion of the paper is therefore that, although there is a risk that patents lead to a tragedy of the anticommons and hamper innovation, they provide in the development of GEVs a considerable springboard for modular innovation by facilitating the interactions among modules. In a sense patents are the “organisational cement” of this process of recombination. However, this result must be limited to the vaccines sector. Further studies are obviously needed to generalise our conclusions.
Modular Innovation and Distributed Processes
271
This configuration of modular innovation induces a new form of industrial organisation that relies on competences and assets distributed on a wide scale of heterogeneous actors. The development of GEV builds on collective and decentralised organisation that relies on market, contractual as well as informal relationships. This pattern of innovation corresponds to processes that are more and more often observed in life sciences. Indeed, the phenomenon of entrepreneurship in biotech, with the multiplicity of small new ventures it has generated, has often been perceived as a renaissance (as compared with the big traditional pharmaceutical companies) of the entrepreneurhero that Schumpeter had in mind in his seminal book “Theory of Economic Development” (1911). To quote Radosevic (2005), the entrepreneur would be “a heroic almost Nietzschean type of lonely individual who faces resistance to innovation and who is driven by non utilitarian motives” (Radosevic, 2005, p. 5). In line with this description, biotech start-ups are usually founded by a small number of persons (often a single individual) who personally assume the direction of the project. Motivations of those founders, which are often not only profit oriented, and the recourse to credit to launch the business are also in line with Schumpeter’s early view. Hence, according to this view, the revival of small business and selfemployment world-wide may announce a reversal of the trend concerning the nature of entrepreneurship, from Schumpeter mark II to Schumpeter mark I. Yet, this view of entrepreneurship ignores interdependences between the multiple actors of innovation. It comes from a deep misunderstanding of entrepreneurship in biotech, which is mainly a collective process that relies on the assemblage of competences distributed across a large number of agents. The entrepreneur is not a single agent but belongs to a network and has to interact with other members to succeed in his enterprise. In short, the locus of innovation has shifted from individual organisations to networks (Freeman and Perez, 1988, Powell, 1996, Baum et al., 2000). There is no reversal to Schumpeter mark I but rather the emergence of a new type of entrepreneurship that relies on the importance of networks and collaborations (Powell, Koput and Smith doerr, 1996). The entrepreneurial activity is not retained into the hand of a single individual but is distributed among across many institutions (McKelvey, 1998). With respect to GEV, this distributed process is largely decentralised. However, it is not completely, since big-pharmaceutical companies often intervene and play a central role of integrator. Their position at the end of the chain (and therefore in contact with the end consumer) and their financial power gives them the legitimacy to undertake this role (Bureth et al., 2007b). As emphasized by Staropoli (1998, p. 17): “In the pharmaceutical industry, the major determinant criteria for the delegation of authority are the financial constraints and the strategic and institutional position in the industry, especially in the distribution and marketing activities. Considering these criteria, a pharmaceutical company has the legitimacy to be the coordinator in charge of the authority in a network involving start-ups, universities and public research organizations”.
272
EJESS – 20/2007. Knowledge and Innovative Activities
Furthermore, with respect to vaccines, public private partnership may often play this role of integrator, as suggested by Orsenigo et al. (2006), who studied the case of IAVI (International AIDS Vaccine Initiative). They conclude the following: “Overall the evolution of the system over the last 20 years has increasingly led to a model in which innovation is carried out in a distributed way by networks of various players […] in the pharmaceutical industry the hub organisation is typically a large pharmaceutical firm. In the case of vaccine development, however, PPPs seem to be taking an increasingly central role in coordinating research activities” (Orsenigo et al., 2006, p. 6). Yet, if the interactions are sometimes undertaken by a dominant player, in most cases the different members of the network interact without being subjected to a given authority, individual or collective. The network gathers a multiplicity of decision centres, each having to do with the others while trying to adopt a specific position. The network of vaccine adopts in a sense a shifting and evolutionary structure, the objectives of which are impossible to define ex ante (or in an incomplete way), and emerge as the network becomes more mature. This view of innovation in the field of vaccine as following a pattern of distributed entrepreneurship is likely to be a promising area for future research. Acknowledgments The authors would like to thank Sandrine Wolff, Moritz Mueller, Rachel Levy, the editors of this special issue and one anonymous referee for their helpful comments. 6. Bibliography Arrow K. J., “Economic Welfare and the Allocation of Resources for Invention”, The Rate and Direction of Inventive Activity: Economic and Social Factors, Princeton university Press, 1962, p. 609-625. Baum J.A.C., Calabrese T., Silverman B.S., “Don’t go it alone: alliance network composition and start-ups’ performance in Canadian Biotechnology”, Strategic Management Journal, Vol. 21, 2000, p. 263-294. Braunagel M., Das R.C., “Promises and Perils of DNA vaccines”, GPT, 2003, p. 25-28. Brusoni S., Prencipe A., and Pavitt K. “Knowledge specialization, organizational coupling and the boundaries of the firm : why do firms know more than they make?”, Administrative Science Quaterly, Vol. 46, No. 4, 2001, p. 597-621. Bureth A. et al. “Modularité et incitations dans le développement de vaccins géniques”, final report for INSERM - MiRe – DREES in the frame of the program “Sciences biomédicales, santé et société”, 2007a.
Modular Innovation and Distributed Processes
273
Bureth A., Mueller M., Pénin J. et Wolff S. “Brevet, innovation modulaire et collaborations: Le cas des vaccins géniques”, Revue d’économie industrielle, No.° 120, 2007b, p. 135154. Bureth A., Levy R., Pénin J. et Wolff S., “Strategic Reasons for Patenting: Between Exclusion and coordination Rationales”, Rivista di Politica Economica, No. IX-X, September-October 2005, p. 19-46. Cohen W.M., Nelson R. R., Walsh J., “Protecting their Intellectual Assets: Appropriability Conditions and Why US Manufacturing Firms Patent (or not)”, NBER working paper 7552, 2000. Cohendet P., Farcot M. and Pénin J., “Entre incitation et coordination: Repenser le rôle du brevet d’invention dans une économie fondée sur la connaissance”, Management International, Vol. 10, 2006, p. 65-84. Coombs R., Harvey M. and Tether B.S., “Analysing distributed processes of provision and innovation”, Industrial and Corporate Change, Vol. 12, 2003, p. 1125-1155. Danzon P. et Sousa Pereira N., “Why sole-supplier vaccine markets may be here to stay”, Health Affairs, vol 24, N°3, May-June 2005, p. 694-96. Freeman C., Perez C., “Structural crisis of adjustment: business cycles and investment Behaviours”, Technical change and Economic Theory, Dosi, Freeman Nelson Silverberg and Soete (eds), London, 1988. Foray D., The Economics of Knowledge, Cambridge, MIT Press, 2004. Forde G/ M., Rapid-response vaccines - does DNA offer a solution?, Nature Biotechnology, vol. 23, n°9, 2005, p. 1059-62 Grindley P., Teece D., “Managing Intellectual Capital: Licensing and Cross-Licensing in semi-conductors and electronics”, California Management Review, Vol. 39, 1997, p. 8-41. Hardin G., “The Tragedy of the Commons”, Science, Vol. 162, 1968, p. 243-1248. Heller M.A., Eisenberg R.S., “Can Patents Deter Innovation? The Anticommons in Biomedical Research”, Science, Vol. 280, 1998, p. 698-701. Jaffe A., “The US Patent System in Transition: Policy Innovation and the Innovation Process”, Research Policy, Vol. 29, 2000, p. 531-557. Kaddar M., “Vaccine market trends”, international vaccinology training course, Seoul, 2004. Koppl R., Langlois R., “Organizations and Language Games”, Journal of Management and Governance, Vol. 5, No. 3-4, 2001, p. 287-305. Langlois R., Robertson P.L., “Networks and innovation in a modular system: lessons from the microcomputer and stereo components industries”, Research Policy, Vol. 21, 1992, p. 297-313 Lerner J., Tirole J., “Efficient Patent Pools”, American Economic Review, Vol. 94, 2004, p. 691-711.
274
EJESS – 20/2007. Knowledge and Innovative Activities
Levin R.C., Klevorick K., Nelson R.R., Winter S., “Appropriating the Returns from Industrial Research and Development”, Brooking Papers on Economic Activity, Vol. 3, 1987, p. 783-820. McKelvey M., “Evolutionary innovations: Learning, entrepreneurship and the dynamics of the Firm”, Journal of Evolutionary Economics, Vol. 8, 1998, p. 157-175. Ordover J. A., “A Patent System for Both Diffusion and Exclusion”, Journal of Economic Perspectives, Vol. 5, 1991, p. 43-60. Orsenigo L., Cacciatori E., Brusoni, S., “The changing landscape of vaccine development: Public Private partnerships between brokerage and knowledge integration”, presented at the SPRU 40th anniversary conference, september 11-13, 2006. Orsenigo L., Pamolli F. and Riccaboni M., “Technical change and network dynamics: lessons from the pharmaceutical industry”, Research Policy, Vol. 30, 2001, p. 485-508 Pénin J., “Open knowledge disclosure, incomplete information and collective innovations”, working paper BETA 2005-10, 2005. Powell W.W., “Inter-organizational collaboration in the biotechnology industry”, Journal of Institutional and Theoretical Economics, Vol. 152, 1996, p. 198-215. Powell W.W., Koput K.W., Smith doerr L., “Interorganizational collaboration and the locus of innovation: Networks of learning in biotech”, Administrative Science Quarterly, Vol. 41, No. 1, 1996, p. 116-145. Radosevic S., “National systems of innovation and entrepreneurship: in search for a missing link”, paper prepared for the Keins meeting held in Lisbon, 12-13 october 2005. Rivette K. G., Kline D., “Discovering New Value in Intellectual Property”, Harvard Business Review, Vol. 78, 2000, p. 54-66. Sako M., Murray F., “Modules in Design, Production and Use: Implications for the Global Automotive Industry”, GERPISA international colloquium, 2000. Schumpeter J., Théorie de l’évolution économique, 1911, Edition Dalloz, 1999. Simon H.A., “The Architecture of Complexity”, Proceedings of the American Philosophical Society, Vol. 106, No. 6, December 1962, p. 941-73. Staropoli C., “Cooperation in R&D in the Pharmaceutical Industry - The network as an organizational innovation governing technological innovation”, Technovation, Vol. 18, No. 1, 1998, p. 13-23. Walsh J.P., Arora A., Cohen W.M., “Research Tool Patenting and Licensing andBiomedical Innovation”, in Patents in the Knowledge Based Economics, 2003, p. 285-340.
