3.2.1.1. Patents

A patent is an exclusive right granted over an invention. Broadly, the patent confers on its holder the right to decide whether – and how – an invention may be used by third parties. In return, the patentee discloses specific technical information to the general public by publishing invention-related information on the corresponding patent document. Thanks to this mechanism, patents help determine the date, place and author of an invention. At the same time, the PCT¹ (Patent Cooperation Treaty) system provides users with the possibility of applying for patent protection of an invention in several countries, through the filing of a single international application.

3.2.1.2. Copyrights

Copyrights is a legal term referring to the rights that creators enjoy over literary and artistic works. Work protected by copyrights range from books, musical pieces, paintings, sculptures and films, to software, databases, advertising creations, maps and technical design.

3.2.1.3. Industrial designs

Depending on the national regulation considered and the type of industrial designs, these can be protected under industrial design law as a “registered design” or under patent law as “design patents”. Industrial designs may also be protected as works of art under copyright law (WIPO 2014). In certain countries, industrial design protection and copyrights are cumulative. In other countries, they are mutually exclusive, meaning that from the moment the patent holder has chosen a type of protection, he/she is not entitled to claim another one.

3.2.1.4. Trademark

This is a sign that distinguishes the products or services of a company from those of other companies. Trademarks are protected intellectual property rights. At a national or regional level, protection of a trademark may be obtained by registration, by filing an application for registration with the national or regional trademark office and by paying the required fees. At an international level, there are two possibilities: filing an application for registration with the trademark office of each country where protection is sought, or using the Madrid system of WIPO. The Madrid system is a complete solution for the registration and management of brands worldwide. By filing a single application in one language and paying a single set of fees, we can ensure the protection of a trademark in the territories covered by 114 members.

According to the most recent statistics of the WIPO (2016), all forms of intellectual property are rapidly increasing (Table 3.2).

Nonetheless, patents still remain one of the most widely used technological indicators. The strong and continuous growth of international applications for registered patents is a phenomenon that dates back to the 1990s and has been observed not only in several European countries, but also in the United States and Japan. In 2015, China played a major role in the record increase of patent applications worldwide, totaling 2.9 million. According to WIPO figures (2016):

“In total, innovators filed some 2.9 million patent applications worldwide in 2015, up 7.8% from 2014, higher than the 4.5% growth rate in 2014. Resident filings, where innovators filed for protection in their home economy, accounted for around two-thirds of the 2015 total.

China’s patent office received 1,101,864 filings in 2015, making it the first office to receive more than a million applications in a single year – including both filings from residents in China as well as from overseas innovators seeking patent protection inside China. This totaled almost as many applications as the next three offices combined: the U.S. (589,410), Japan (318,721) and the Republic of Korea (213,694).”

As shown in Table 3.3, the distribution per company reflects the general trends already observed at a national level:

3.4.1.1. The concept of technology and innovation

The “traditional” theory of technology is based on the writings of [ARR 62]. It assimilates technology to the production of knowledge, as if it were a merchandise with three fundamental characteristics: indivisibility, uncertainty and weak appropriability, pretty much in the same way that informational goods are assimilated to uncertainty: “which makes technology […] a commodity which can neither be divided nor appropriated. For this reason, there is no incentive to allocate resources to R&D, because the real value of its output (for example, information) cannot be determined by the market forces” [GAF 92].

Information can be seen as the result of long and complex processes of accumulation and appropriation of technology [CHE 86]:

“In particular, weak appropriability means that there is a divorce between the costs of research and the revenues derived from the exploitation of research results, because it is always possible for a firm who did not provide research efforts, to take advantage of the research results produced by others at a very low cost” [GAF 89].

The imperfection of appropriation engenders a consequence: on the one hand, it requires the implementation of organizational coordination to remediate market failure mechanisms [BEH 85, COH 92] and, on the other hand, it leads to the need for reinforcing legal regulations that safeguard property rights.

3.4.1.2. Appropriability of innovation and externalities

According to [DEM 88]:

“The first function of property rights is to serve as a guide that encourages a greater internalization of externalities […]. Changes in knowledge result from changes in production functions and individual aspirations. New techniques, new ways of producing the same goods or conceiving new ones engender beneficial or perverse effects that society is simply not used to […]. In fact, the emergence of new property rights occurs in response to people’s desires to adapt to new cost-benefit opportunities.”

Now, innovation is a type of production, which, in certain cases, makes it difficult to assign property rights to the value it creates, and this also provokes the emergence of problems related to externalities.

“As it is widely recognized, an optimal allocation of resources cannot be achieved through a market system if they are confronted with technological externalities. There are goods for which no market can be created” [ARR 84].

Delving into the works of [GRI 79, GRI 92 and GRI 95] and [MOH 91 and MOH 96], we may find that there are two categories of externalities related to technology. First, there are the productivity spillovers, innovations that provide benefits to not only the industry concerned but also downstream industries. The second category of externalities is related to the innovative industry (these externalities are also referred to as “R&D spillovers” or “carry-over effects”). They are rooted in the informational nature of innovation and its degree of appropriability.

[TEE 86b] described an appropriability regime of assets created by alliance. He distinguished between “codified” and easily transmissible knowledge (highly appropriable) from tacit knowledge, which is difficult to transmit and which leads to a regime of strong appropriability. This can even be reinforced if the innovation benefits from protection (patent).

Now let us consider the case where innovation efforts performed by a firm may benefit to other firms, and this at a negligible cost. These externalities lead to opportunistic behavior on the part of firms in the measure that R&D efforts made by other firms are profitable to them. We can then observe certain “passivity” from the part of firms that do not engage in a race for innovation. [HAL 04a, HAL 04b].

3.4.2.1. Cooperative and non-cooperative R&D, with externalities

[DAS 88, DAS 90] studied the impact of cooperation on R&D efforts from a symmetric two-phase duopoly model (a “precompetitive” R&D phase and a production phase). During the first phase, companies choose a level of investment in R&D that will affect the criteria for the next phase in which prices and quantities will be determined.

This model allows them to compare cooperation with competition in R&D, and to be faced with several concepts of equilibrium: a situation of total cooperation, of partial cooperation, of non-cooperation and of welfare (defined as the scenario which results from adding consumer surplus to producer surplus). Consequently, the firms must choose between three options:

• – to cooperate during the two phases (R&D and production);
• – to cooperate only for R&D activities;
• – not to cooperate at all (neither during R&D nor during the production phase).

Expenditure levels on R&D (x1 = x2 = x) and production (Q = q1+q2) are calculated for each of the equilibrium scenarios described above. Thus, R&D expenses (x**, x’’, x*, x’) and production costs (x**, x’’, x*) correspond to welfare situations, non-cooperation and partial cooperation.

Firms have a production function with externalities. The presence of positive externalities (as measured by the ß parameter) is considered in terms of a decrease in the rival firm’s production costs.

D’Aspremont and Jacquemin considered a duopoly with an opposite demand function: D–1 (Q), where Q = q + q2 is the produced amount. Each firm faces a Ci (qi, xi, xj) cost of production, which is a function of their own qi production, of the expenses for the xi research undertaken and the xj research expenses of the rival. In this scheme, D–1 and C functions are supposedly linear:

where:

According to the value of ß parameter, results differ:

• – in the presence of important externalities (on their model ß > 0,5) over R&D profits, we obtain: x**> x’’> x’> x* and Q**> Q’> Q*> Q’’;
• – firms which cooperate during the R&D phase, but not during the production phase, run into higher R&D expenses than non-cooperative firms in both phases and, on the contrary, reach a higher level of production (the closest to “socially optimal” level);
• – in the presence of weak externalities (i.e. in the context of ß < 0,4 model), firms which cooperate neither at the R&D nor during the production phase spend more on R&D and produce higher amounts than the firms that cooperate. The second best balance for R&D is obtained thanks to non-cooperative behavior during both phases. Then, we get:

and:

These results were generalized by [SUZ 89, SUZ 92], and the conditions of stability for the solutions were studied in depth. For instance, [HEN 90a] confirmed the explorations of Jacquemin and d’Aspremont in the case of strong externalities. Furthermore, [SUZ 89, SUZ 92] proved that the introduction of externalities in the context of the non-cooperative model tends to favor stability. However, in the case of the cooperative model, when the level of externalities is higher, the equilibrium ceases to exist.

As we can infer, the value of ß parameter plays an essential role. According to Jacquemin and d’Aspremont, numerous factors may influence the value of ß. First, there is the nature of research. A priori, the results of pre-competitive, generic research are less easily appropriable and lead to more externalities than those regarding specifically applied development activities.

Moreover, the nature of the contract and the degree of perfection in information also affect the rate of externalities. On the one hand, these externalities are higher for the different partners of a cooperative agreement than for firms that do not cooperate with one another. On the other hand, within the cooperative group itself, ß can vary according to the type of organizational arrangement, where ß = 1 corresponds to perfect communication and information use, as in the case of integrated research laboratories.

Ultimately, regardless of whether there is a patent or not, positive externalities make it impossible to fully protect the innovative firm. Cooperative behavior is privileged by the firms wishing to internalize these externalities.

3.4.2.2. Multi-firm model generalization

[SUZ 92] also looked into the question of cooperative and non-cooperative R&D with externalities, but this time from an oligopolistic point of view. The author compared the effects of cooperative R&D against non-cooperative R&D, particularly focusing on cooperative R&D at a pre-competitive stage, at the moment when firms become rivals in the market. Furthermore, [SUZ 92] intended to draw normative conclusions for economic policies.

In the presence of sufficiently large externalities, neither the non-cooperative equilibrium nor the cooperative equilibrium makes it possible to reach levels of second best in R&D. In the absence of externalities, however, while the level of cooperative R&D remains socially insufficient, the non-cooperative level should surpass the first and second best R&D levels.

Suzumura’s analysis is conducted in the context of an oligopolistic competition model. During the first stage, firms decide to reduce the cost of R&D either cooperatively or non-cooperatively, whereas in the second stage, they engage in competition on the basis of the amounts of the product market.

The author considers that the soundness of the results offered by Jacquemin and d’Aspremont can be questioned; he intends to resume their conclusions and generalize them. As we have previously observed, Jacquemin and d’Aspremont concluded that cooperative R&D agreements between otherwise rival firms should increase the level of R&D expenditure related to the totally uncooperative case. Thus, it should be remembered that R&D externalities are large enough and that cooperative R&D does not reach the socially first best level.

Suzumura reminded us that [DAS 88] presented their results as if they were astounding on the basis that cooperation should reduce excessive duplication of R&D efforts in the face of significant externalities. However, he pointed out that the motivation for R&D of a single firm depends on the appropriability of R&D profits. In such a case, the presence of significant R&D externalities should drastically reduce incentives to reduce costs, with the result that voluntary R&D undertaken by a firm tends to be socially weak. From this point of view, an applicable agreement on cooperative R&D efforts seems to encourage more commitments. The result of this net effect of R&D cooperation can be explained by the relative strength of these antagonistic effects.

The objective of Suzumura is to prove not only that the second effect dominates over the first one in the duopoly example, with a linear inverse demand function and a linear marginal cost function, as described by [DAS 88, DAS 90], but also, in a wider category of oligopolistic industries, what allows them to generalize the scope of their results.

In the model, firms are engaged in two-phase competition. During the first phase, they make an irreversible commitment to R&D. In the second phase, the strategic variable is the level of production.

Suzumura then examined two types of equilibrium. The first one is non-cooperative for the phases. Under these conditions, the balance of the second stage is the Cournot–Nash equilibrium, leading to perfect equilibrium in subsets for the overall game. The second equilibrium integrates both cooperative and non-cooperative dimensions: firms are supposed to coordinate their R&D during the first stage in order to maximize their joint profits and to become rivals in the second stage. As we will see, the results are summarized in Tables 3.4 and 3.5.

Suzumura reached a number of conclusions concerning the merits (or not) of technological policies. For a start, if we compare the first and second lines in each of the tables, it appears that the conclusions are identical for the two types of welfare. In Table 3.4, the comparison of the first and second columns shows that, in the presence of large externalities, the equilibrium levels of non-cooperative R&D and of cooperative R&D are socially insufficient.

Thus, a technological policy that facilitates an additional investment in R&D marginally improves welfare (regardless of whether the firms cooperate or not). In Table 3.5, the first and second columns show that in the absence of externalities, the level of equilibrium of non-cooperative R&D turns out to be excessive, whereas the level of equilibrium of cooperative R&D is insufficient. Then, if firms cooperate, technology policy should encourage investment in R&D.

Suzumura recommended taking into consideration different elements for a deeper analysis of cooperative R&D:

• – First, R&D undertaken by firms outside the industry should influence the marginal cost of a firm. Such effects of inter-industry externalities should be taken into account with the effects of intra-industry externalities, in order to obtain a balanced assessment of the effects of R&D externalities;
• – Second, one of the functions of cooperative R&D is precisely to generate synergy effects by bringing complementary resources together, such as information research and experience, teams of researchers and technological expertise. From this point of view, Suzumura disagreed with the idea that R&D externalities in terms of average variable cost function should remain the same whether firms cooperate or not. For a deeper analysis, the function of externalities should probably be endogenized;
• – Third, the potential benefits of cooperative R&D are often linked to an increase in the speed of invention and innovation, in particular by fairly distributing risks. An element of uncertainty should be introduced in the analysis.

All this work on technological externalities conceives cooperation as a particular modality of internalization: the R&D alliance is integrated into a continuum of secrecy, patent and temporal asymmetry. Therefore, what criteria make it possible to choose the alliance in R&D as a mode of appropriation instead of secrecy, patent or temporal asymmetry? Apart from the contribution of [KAT 90] on patent arbitration (ex post cooperation) and the R&D alliance (ex ante cooperation), the theory of externalities does not provide an answer and does not propose a typology, helping us to clearly classify the different modalities for internalizing technological externalities.

3.5.2.1. Degree of openness

This notion of openness reflects the complexity of the phenomenon of open innovation. Open innovation should not be interpreted as unrestrained access without obligations or control. There are several degrees of openness in open innovation. In one extreme form, accessibility is total and access is completely unrestricted. In this case, open innovation is accompanied by a promise of a public good:

“Our notion of openness is defined as the pooling of knowledge for innovative purposes where the contributors have access to the inputs of others and cannot exert exclusive rights over the resultant innovation. In its purest form, the value created through an open process would approach that of a public good” [CHE 07].

As pointed out by Pénin [PEN 13], these characteristics are mainly found in the practices of open access, open data or in open source software, but are not part of open innovation. In most cases, open innovation can be accompanied by strong protection.

“In the end, for Chesbrough, contracts and intellectual property rights are always present in open innovation approaches, which are most often facilitated by strong patents or the possibility of formal contracts. A company whose technology is protected by strong patents will be better predisposed to agree to collaborate or to provide licenses, because betrayal and technological looting risks are reduced; conversely, if the company is not protected, collaboration is risky, considering that the partner can always take possession of the company’s manufacturing secrets free of charge” [PEN 13].

3.5.2.2. Outside-in and inside-out

The term outside-in involves internalizing external skills so as to develop innovation in one’s business.

For a company, the notion of inside-out is associated with knowledge-sharing and outsourcing. This practice enhances the company’s external intellectual property by creating structures detached from the main firm (subsidiaries or joint ventures), while still having access to resources (patents) and existing competencies at the heart of the “parent company”, and particularly its R&D activities.

As explained by [BIN 15], open innovation is one model among others that requires tailor-made answers in terms of intellectual property:

“How does Intellectual Property Law understand the open innovation model?” The open innovation model is above all rhetoric. It is more a synthesis of practices than a revolution in creative methods. There is no conflict between intellectual property and open innovation. The idea is that – in order to innovate - the company must mobilize a plurality of models and skills. It cannot be satisfied with a single model. There is a real difficulty in organizing work. We do not innovate uniquely in an open or closed environment, either alone or collectively. It is necessary to simultaneously mobilize internal innovation, external acquisition, collaborative research, open innovation, the open model, the closed model. Success depends on the articulation of all these models; it is difficult because they are often presented as antinomic when in truth, they are complementary.”