4.2.1. Innovation awards, or how to rehabilitate an old approach

It is important to distinguish two categories of contest prizes that function according to different logics. The first category is “inducement prizes,” which generate competition between inventors ex ante, in the goal of obtaining an innovation. The second category is “recognition awards” which put inventions in competition ex post and reward the best contribution.

Inducement prizes are based on identifying a technological barrier that needs to be overcome. They may either set a deadline at which time competing solutions will be compared on the basis of pre-determined objective criteria and reward the most performing invention, or they may set project specifications that stipulate the expected performances for various criteria and in what circumstances they must be obtained, and reward the first solution that demonstrates that it satisfies those criteria. In both cases, the reward amount is generally pre-defined. Some of the examples described in historical studies include the longitude prize established by the British authorities in 1714¹, the substitute food prize launched in 1772 by the Académie des Sciences de Besançon in France², and the Orteig prize created in 1919 for the first flight across the Atlantic by plane. As early as 1944, Polanvyi [POL 44] suggested reforming the patent system and replacing it with an award system as a means to encourage creativity and innovation, without incurring the inefficiencies of dead losses induced by the monopoly power conferred by a patent³, The renewed interest in prizes and more generally in promoting policies not directed by the market as complements or alternatives to Intellectual Property Rights (IPRs) is discussed in Stiglitz [STI 06] who also advocates for the introduction of a prize system as “an idea whose time has come”⁴, From a practical perspective, the renewed attention for prizes comes on the heels of the resurgence of innovation contests in the United States, especially the contests hosted by the X Prize Foundation. Although this foundation is known for its audacious initiatives in private financing for space projects, it also distributes prizes in other areas like medicine, automobiles and the environment⁵. The enthusiasm for prizes also extends to public authorities. The America COMPETES (Creating Opportunities to Meaningfully Promote Excellence in Technology, Education and Science) Reauthorization Act signed in 2010 by the Obama administration allows all American federal agencies to easily make use of the contest prize system to stimulate innovation⁶. On the European side, in March 2014, the European Commission awarded the first prize for the Horizon innovation award program.

The basic idea is that inducement prizes would be a good leverage to overcome technology bottlenecks. Confronted by technological hurdles, the private actors or group of actors such as public authorities can tap into very diverse areas of research efforts. For public authorities, of course, the motivations are to direct research efforts to respond to challenges that are specific and relatively urgent (as in the case of the Académie des Sciences de Besançon prize in response to the famine), strategic (the longitude prize also made it possible to establish the British dominance over the oceans) or major long-term challenges⁷. The Horizon prizes from the European Commission notably include six “low-carbon” challenges out of twelve: reuse of carbon dioxide, low-carbon hospital, photovoltaics, engines retrofit for clean air, clean engines for the future and materials for clean air. For some private actors, the motivation can reside in the aura that they can draw from the financing of such prizes or the visionary image that they want to give to their company, with the economic benefits coming in the long term⁸. For others, profit-seeking is the motivation. Newell and Wilson [NEW 05] give the example of the editor of the Chicago Times-Herald who, in 1895, sponsored a series of automobile races with different prizes for speed, endurance and even aesthetics. The public enthusiasm for these races created an audience for a specialized press. More recently, the NRG COSIA Carbon X Prize is a prize funded in the context of the X Prize foundation by Canada’s Oil Sands Innovation Alliance (COSIA) to stimulate and accelerate industrial uses of carbon dioxide. In doing so, the consortium aims to change perceptions about the environmental and economic consequences of its activity by making it so that carbon dioxide is no longer a pollutant emitted into the air that contributes to global warming, but a by-product that is captured and reused profitably. This example is indicative of the exploitation of skills and knowledge that are complementary to those in the profession of financer. William [WIL 12] specifies that if the prize is awarded, some financers require either the intellectual property on the result directly or an exclusive license in exchange.

Khan [KHA 13] analyzed innovation outside of the patent system, especially the impact of international exhibitions on inventive activity. More specifically, Khan [KHA 13, KHA 14] compared the systems that award recognition prizes to invention patents. A first series of works by Khan [KHA 14] focused on a sample of technologies exhibited and awarded during industrial exhibitions organized by the Massachusetts Charitable Mechanic Association between 1837 and 1874. A second series of works by Khan [KHA 13] focused on comparing the experience of British and American inventors between 1750 and 1930 in terms of incentives to innovate. The results show the factors that led inventors to choose the protection conferred by intellectual property rights or, a contrario, seek out alternatives such as prizes awarded at industrial exhibitions. One of the first lessons of these works is the existence of a degree of cognitive bias toward elites. In fact, the winners of contest prizes generally tend to belong to the most privileged classes, compared to the population of patent holders, as measured by their wealth and professional status at the time of the exhibitions. Khan [KHA 13] showed that one of the most important factors when awarding prizes to British inventors was whether or not they had attended Oxford or Cambridge for their studies. The second lesson is that the awarding of prizes tends to be largely unpredictable and unrelated to proxies related to the productivity of the innovation such as inventive capital or the commercial success of the invention (Khan [KHA 13, KHA 14]). The lack of transparency about the awarding rules seems to have encouraged idiosyncratic and inconsistent decisions. This lack of method in the prizeawarding procedures is obviously questionable. It encouraged the quest for rents and reduced incentives for inventors who realized that the prizes, in many cases, were disconnected from inherent inventive merits. In fact, contest prizes were sometimes awarded on the basis of arbitrary factors such as the personal aura of inventors rather than the quality of the technological contribution itself, and were subject to the possible partiality of judges, even corruption and risk of capture. Statistical analyses show, for example, that the quasi-totality of the variation between prizes in gold and silver is not explained, which suggests that these awards were based on fairly random foundations (Khan [KHA 13]). In addition, Khan [KHA 14] emphasized that industrial exhibitions are not representative of a country’s inventive capital given that the selection of objects introduces a bias that may not correlate to their technological capacities. Not all inventions are patented and likewise not all inventions are presented at fairs. Khan [KHA 14] concluded by showing that, with all these elements in mind, the enthusiasm for prizes and a system administered through prizes in Europe and the United States fizzled out in the 19th Century. The growing disillusionment around prizes as an incentivizing mechanism to generate innovations became obvious when the Royal Society of Arts (1901) recognized the lack of social value in this practice (Royal Society of Arts [ROY 01]). The contemporary change of tack in favor of prizes may only be a pendulum swing in a world that, through privileging patents and seeing their limits, has forgotten past alternative experiences and their own limits.

4.2.2. Could innovation awards replace patents?

Innovation prizes, or at least inducement prizes and patents have in common that they offer ex post compensation to inventors. They also reward a result and not an effort, unlike innovation aids such as R&D subsidies, research tax credit, preferential loan rates or reimbursable advances for innovative projects that support R&D effort financially and ex ante. As mentioned in Chapter 1, innovation is an inherently random process. This raises a problem of moral hazard in cases of ex ante support. The impossibility of discriminating between, on the one hand, R&D projects that are not successful because the problem to be solved is more complex than imagined and, on the other hand, projects that are not successful due to the lack of effort by the searchers, is a disincentive for researchers’ efforts. By construction, mechanisms based on an ex post reward avoid this pitfall. Leaving aside for the moment the issue of whether they have faults that ex ante support mechanisms do not, we can legitimately ask if innovation prizes are not interchangeable, even preferable, to patents.

The question of the conditions in which intellectual property rights are the best incentive system or not in terms of innovation has been the subject of abundant literature in economics. A first synthesis is proposed by Gallini and Scotchmer [GAL 02]. This literature is located within the context of a binary choice between either a system based on patents or a system based on innovation prizes financed by public funds. A first argument in favor of innovation prizes is that, unlike patents, the reward that they offer is not limited in value. In the case of patents, the compensation is restricted by the monopoly profit. Aside from the extreme case of a monopoly that is perfectly discriminating and therefore succeeds in extracting the entirety of the consumers’ willingness to pay, generally the profit of a monopoly only represents a fraction of the exchanged gains on a market. Without a complementary support system, there is almost always a lack of incentive in a patent system compared to what is socially optimal. Patent boxes⁹ are a means to partially circumvent the problem through the introduction of a fiscal advantage. However, patent boxes are critized due to the distortion of competition they generate between domestic and foreign firms (Bradley Dauchy and Robinson [BRA 15], Merrill [MER 16], Sullivan [SUL 15], OECD [OEC 15]). On the other hand, the amount of compensation is not a priori restricted in a prize system. It is even conceivable to reward “spillovers” induced by an innovation with a prize system. A second argument in favor of prizes as an alternative to patents is that they are not subject to the effects of distortion related to the exercise of market power conferred by a patent. In fact, they are generally presented as a system where, once the innovation has been rewarded monetarily, it is widely distributed. The intensity of the competition that results on the new market avoids the exercise of all market power in principle. However, these results disregard the problems of information asymmetry between inventors and public authorities. Wright [WRI 83] showed how information asymmetries that prevail ex ante between inventors and the government influence the choice of the optimal innovation incentive mechanism. More precisely, these works show that in the absence of an information asymmetry (or in other words, if the costs and benefits of research projects are known to both researchers and the government) then prizes are preferable to the temporary monopoly associated with patents (Maurer and Scotchmer [MAU 04])¹⁰. However, one of the conditions is that the needs must be clearly identified (“well-known needs”). On the other hand, if there is an information asymmetry between researchers and the government about the costs and advantages and if the terms of compensation are set before this informational asymmetry is resolved, then prizes are not socially more efficient than patents¹¹. By focusing on the information asymmetries and the difficulties in determining the optimal level of compensation, Shavell and Van Ypersele [SHA 01] emphasize the weakness of incentives generated by the contest prize system¹². Not only Wright [WRI 83] but also Gallini and Scotchmer [GAL 02] showed that in the end, intellectual property is probably the best mechanism to filter the innovation projects when the value and costs are not observable by the government. The key idea is that with IPRs, the government does not need to know in advance what projects are socially beneficial. On the other hand, context prizes are an optimal solution if the probability of success is moderately high, if the elasticity of the supply of inventions is low and if prices can be adjusted ex post.

Concerning recognition prizes, the argument is often raised that they offer social (rather than pecuniary) incentives and compensation. The basic idea is that members of the “special class of genius” respond better to honors and prizes than to simple material incentives. In other words, scientists are mainly motivated by recognition from their peers and the satisfaction of solving problems that were long considered unsolvable. The question is whether or not inventions really can be stimulated by invitations to compete for essentially honorific prizes. According to Sidney [SID 62], recognition prizes or “medals” would be ineffective and improving prospects on the demand side, as well as the stimulus of competition, is a much more effective incentive to stimulate inventive activity. The only effect of recognition prizes would be to improve the reputation capital of the firm in question, and in doing so improve the commercialization prospects of the promoted technologies¹³.

Levine [LEV 09] summarized the choice between inducement prizes and patents by stating that, if a government knew how much it would cost to develop a given technology and what its value was, then it would be relatively easy to design a compensation system that was appropriate for inventors. The problem is that this is obviously not the case, because in practice, the asymmetry of information regarding both costs and the value of the new technologies renders this kind of exercise inoperative. One consequence is that prizes and patents cannot be strictly considered alternatives. Although they both constitute ex post reward mechanisms based on results, they both have advantages and disadvantages. Some authors, such as Shavell and Van Ypersele [SHA 01] followed by Abramowicz [ABR 03], suggest introducing a certain flexibility in the choice between these instruments. More specifically, they suggest leaving the choice up to inventors. Shavell and Van Ypersele [SHA 01] explain that inventors who predict that the monopoly gains associated with a patent will be low will opt for a prize and in doing so, protect the community from losses due to the exercise of market power. The self-selection mechanism, even if it only avoids losses inherent to market power where it is weakest, is still more efficient than both of the systems applied exclusively. Inspired by the example of the patent on the daguerreotype bought by the French state in 1839, Kremer [KRE 98] suggests leaving the choice to patent holders to either exploit their patent or sell it to the public authorities through an auction mechanism. In practice, a few rare examples of à la carte choice have been applied. The sterilized food canning method, invented by Nicolas Appert, is one example¹⁴. Instead, a system of complementary use of prizes and patents emerges in practice. Patents apply in an undifferentiated way to all inventions. They offer the advantage of scaling the reward for the success of the invention fairly automatically. They therefore direct R&D toward innovations whose gains, appropriable thanks to the conferred right to exclude, are the highest in relation to mobilized means. As long as these gains are aligned with the social gains of the innovation, patents can be considered as a satisfying incentive mechanism. In this sense, patents are a generic mode of support for innovation. If, on the other hand, the appropriable gains are not correlated with the social gains drawn from the innovations, and they are much lower than for the average invention, then the incentive supplied by a patent must be supplemented by another mechanism. Inducement prizes can typically allow for this. In fact, it targets an innovation due to its specifications. According to this logic, the prize is added to the patent. In practice, the overlap of incentive mechanisms is often observed and many inventors receive both patents and other forms of compensation¹⁵. Contrary to many academic works that treat them as alternatives, the prizes awarded by the State or other organizations do not prevent inventors from also profiting from other means, including patent protection¹⁶.

4.2.3. Complementarity with support for R&D efforts

Not only can innovation prizes and patents be complementary to each other, they can also be complementary to ex ante aid mechanisms for innovation. Two shortcomings of ex post compensation mechanisms cause this complementarity (Newell and Wilson [NEW 05]): the risk aversion of inventors, and the financial constraints they face.

Because the risk associated with each innovation is primarily borne by the inventor, reward mechanisms do not offer the advantage of risk mutualization. Under the relatively acceptable assumption of the absence of correlation between the success of different inventions, the law of large numbers dictates that a fairly certain number of projects among a large set of projects will succeed. The community can therefore somewhat neutralize risk by playing on the large number of projects that it finances. On the other hand, inventors concentrate their investments on a single project and are therefore faced with a non-diversified risk. If, like most economic agents, they have an aversion to risk, they will be less incentivized to embark on their project than the community. This effect can only be partially offset by the limited responsibility of the inventor. By directly financing the R&D effort of inventors, the community can incentivize them to take risks that they would not take alone but that would be socially desirable to take. Through its financial support, the community takes responsibility of a part of the risks. This taking of responsibility is often considered to be all the more desirable if the invention is in upstream phase and if the technological and market uncertainties are high. The design of aids also makes it possible to adjust the distribution of risks and prevent a crowding out phenomenon of private effort through public support¹⁷.

Reward mechanisms also encounter limits related to the financial constraints of inventors. They require that the inventor commits to R&D spending before eventually receiving the reward, and sometimes only after several years. Although this solves the problem of moral hazard evoked above, the counterpart is that only the inventors that have sufficient financial resources or easy access to external funding can actually launch their innovation projects. The system therefore favors large firms over small ones. Small and Medium Enterprises (SMEs) often have more limited financial means and, as their survival depends more directly on the success of the project, the risk premium required when making recourse to external financing is much greater. Some patent offices like the US Patent and Trademark Office (USPTO) partially remedy this by facilitating the filing process and lowering renewal costs for patents filed by small- or medium-sized companies. This solution remains very partial because it only comes into effect once an invention is at the patent-filing stage. Direct aids for R&D effort mitigate this problem. Again, they make it possible to realize projects that are socially desirable but that would not see the light of day if they relied exclusively on private economic calculations. It is often considered that small companies, particularly start-ups, are more inclined to launch breakthrough innovation projects than large companies who can find themselves locked in routines and because of this, only proceed with incremental innovations. Once the research stage has ended and a patent has been filed, and a fortiori once a patent has been obtained, venture capital takes over for public aid for the development component¹⁸.

4.2.4. An example of complementarity between instruments: low-carbon innovation

“Low-carbon” innovation is supposed to respond to two market failures. One, which is environmental, concerns the absence of a carbon pricing and therefore an excessive emission of this gas, and more generally of greenhouse gases. The other, in matters of innovation, results when knowledge has the characteristics of a public good, as detailed in Chapter 1. A wide range of economic policy instruments have been proposed by economists to remedy both of these failures. Intuition suggests that by applying the right economic policy instruments to correct the environmental problem, there would no longer be a low-carbon innovation issue. More specifically, the innovation problem would be generic and would not be regulated by other types of innovation policy instruments than the ones usually used, notably patents. In other words, once the “good” carbon price signal is sent, innovation would not need to be specifically directed toward low-carbon technologies. This is the cornerstone of “price fundamentalism” defended especially by Nordhaus [NOR 11].

Price fundamentalism only appears to be simple. In fact, economic policies intended to remedy a market failure are often only analyzed by economists by assuming the absence of other market failures. One notable exception is the notion of the “double dividend” (Goulder [GOU 95]) which starts from the observation of practical or limited incapacity to implement non-distortionary taxation to study the question of the use of the product of environmental taxation¹⁹. Starting from the same assessment, we can argue that all direct and public financial support policies require first collecting funds using a distortionary tax system. Such a policy immediately breaks with the purely theoretical framework of the absence of a market failure other than the ones related to innovation and to greenhouse gas emissions. In his argument in favor of price fundamentalism, Nordhaus [NOR 11] does not elude this problem, but his presentation tends to underestimate the consequences. The public policies that should be implemented are often contingent on the persistence of market failures beside the ones that the public authorities seek to correct. Aside from the question of persistent failures, price fundamentalism skips an essential point: as we saw above, for informational reasons among others, there are no totally efficient public policies to address the question of incentives for innovation. Patents, especially, are only an imperfect system.

In contrast with price fundamentalism, other economists have developed the idea of a technical progress that must necessarily be directed in favor of low-carbon innovation. An article by Acemoglu et al. [ACE 12] formally set out the elements of this reasoning. Two key hypotheses underpin their conclusion when it comes to the necessity for an innovation policy specifically dedicated to green innovation to promote green growth. The first hypothesis is that knowledge accumulates over time. It is inscribed in a well-established school of thought in the economics of innovation that every inventor is “standing on the shoulders of giants”²⁰ in the sense that they pursue, by improving on the margin, the creation of a common knowledge capital. It is this capital, and not just new knowledge, that makes it possible to innovate. The second hypothesis is that “green innovations” require a different capital of knowledge than the one that makes it possible to develop other innovations known as “gray innovations.” In other words, it is not possible to mobilize the capital of accumulated knowledge in an economy that would have privileged only “gray innovations” in order to begin to develop “green innovations” overnight. It is, however, necessary to construct a capital of knowledge specific to green innovation. This perspective is certainly radical, but it has the advantage of emphasizing that by simply setting a price for greenhouse gas emissions, we cannot also bring back the stock of gas accumulated over time on a socially optimal trajectory and do the same for the stock of knowledge dedicated to “green innovation.”

This is a case of a well-known principle according to which we cannot control two trajectories with a single instrument. One can illustrate this by comparing the economy to either a boat or a plane. If the innovations were only resulting from the influx of new knowledge, whatever it is, the economy would be similar to a boat: the environmental policy would serve as a rudder and make it possible to entirely control the trajectory. The environmental policy only needs to send the right carbon price signal to redirect the innovation effort to “green innovations” (the model does not account for non-rivalry and the impossibility of excluding in the use of knowledge). By introducing the role of the stock of capital and by assuming that “gray” knowledge is not substitutable for “green” knowledge, the economy becomes more similar to a plane rather than a boat: in addition to controlling the direction, the height must also be controlled. In other words, another instrument of control in the form of a support in favor of “green innovations” must be added to the rudder which is the environmental policy. Without this second instrument, it is impossible to set both the correct trajectory for greenhouse gas emissions and the correct trajectory to constitute the stock of “green knowledge”.

Although the works of Acemoglu et al. [ACE 12] establish the theoretical elements of an innovation policy specifically dedicated to low-carbon technologies, the degree of abstraction of their approach leaves it vague as to the form of the low-carbon innovation support instruments. After them, other works sought to specify the outlines of these instruments (Olmos et al. [OLM 12], Veugelers [VEU 12], Aalbers et al. [AAL]). They all underscore the multiplicity of low-carbon technologies and consequently the very heterogeneous character of the stock of “green knowledge” to create. They conclude the necessity of adapting the choice of instruments to each technology, case by case. The degree of maturity between, for example, onshore wind energy and wave energy converter systems determines the choice of appropriate instruments. The instruments available to the public authorities to promote a targeted low-carbon innovation, on top of the generic support instrument represented by the patent system, are often grouped in a few categories which are worth briefly reviewing.

First of all, we traditionally distinguish “demand pull” instruments from “technology push” instruments. Demand pull instruments act downstream by attempting to stimulate and sustain a viable demand for innovations by anticipating, in return, that the development of a market will stimulate the production of innovations upstream. The mechanism of “feed-in tariffs” mobilized in several countries to support renewable energies is typical of this approach. This mechanism establishes a guaranteed sale price, greater than the wholesale price for electricity, for electricity producers using renewable resources and a purchase obligation in the context of a long-term contract (often in the order of 15 years) with the incumbent. Technology push instruments act upstream by promoting the production of innovations at the source. We can sub-divide them into, on the one hand, the ones that support R&D efforts, in which we include public loans with preferential rates, and on the other hand, the ones that reward the results of R&D, in which we include innovation prizes. The inducement prizes of the Horizon program by the European Commission have already been mentioned above. In the United States, an often-cited example is the Super-Efficient Refrigerator Program (SERP) by the US Environmental Protection Agency implemented in the 1990s²¹. Most developed countries combine these three types of instruments to guide innovation toward low-carbon more than the generic patent system does, even in the presence of a carbon tax.

4.3.1. Patent pools as a premise for open innovation

The tragedy of the anti-commons, presented in Chapter 1 and discussed as part of the more specific context of standard-essential patents in Chapter 3, is an essential limit to the efficiency of the patent system. This inefficiency is related to the supposed absence of coordination between complementary patent holders in setting amounts for their licenses. This lack of coordination can cause concern about royalty-stacking, in the sense that a licensee must pay a sum of royalties., the amount of which exceeds what would have been asked if a single entity had owned all of the licenses in question. Consequently, the adoption of the innovation can be slowed and so can the revenue that patent holders would obtain from their rights. A natural response to this flaw of patents seems to be the formation of structures coordinating the action of patent holders by holders themselves. Economic literature documents the emergence of such structures, primarily in the United States during the 19th Century and at the start of the 20th Century. These structures, although not totally homogeneous in their rules of functioning, are today commonly known as patent pools. The literature highlights that instead of a desire to coordinate between patent holders to increase their efficiency in terms of the valorization of their intellectual property assets, it is above all the desire to resolve fratricidal patent wars that motivated the emergence of patent pools²³. Analyzing a few representative cases in the history of industrialization allows us to better understand the logic of cooperation and better appreciate to what extent we can see the premises of open innovation.

Lampe and Moser [LAM 10] propose an analysis of one of the first patent pools, which was for sewing machines in the United States in the second half of the 19th Century. The authors begin by reviewing the facts, namely the legal disputes that developed between the companies Singer and Elias Howe following the latter obtaining US patent No. 4750 (“Improvement on the sewing machines”). Following these legal battles, the three main American manufacturers of sewing machines joined with Howe in October 1856 to form a pool that initially grouped together nine complementary patents essential for manufacturing sewing machines. The authors then engaged in a quantified analysis of the effects of this patent pool on the pace of innovation. They noted that members of the pool and non-members saw their number of patent publications decrease after the implementation of the pool and then increase again after its dissolution in 1887 following the expiration of the patents that composed it. One explanation proposed is that manufacturers engaged in a patent race before the creation of the pool and then non-members would have restricted themselves in their patent requests for fear of litigation with pool members who had engaged in an aggressive patent assertion strategy for pool patents rather than reinforcing their own innovation. This can be seen as proof that patents were instrumentalized in a logic of judiciarization without a direct connection with the pace of innovation. The interest of the work by Lamp and Moser [LAM 10] is that they are able to propose an objective measure of the rhythm of innovation through the speed of machines as documented by technical records from that era. However, the speed gains stagnated during the period of the existence of the pool, confirming a positive correlation between issuing patents and the pace of innovation. This rather negative evaluation of the effects of the pool is supported by a work of cliometrics based on a diff-in-diff evaluation method of patents obtained by members and non-members of the pool during the years covering and not covering the existence of the pool. The authors complete their quantitative study by comparing, on the one hand, grants for patents in the rest of the American industry for the same period and, on the other hand, grants on patents for sewing machines in the United Kingdom. In the first case, they showed that the portion of patents concerning sewing machines in the total patents granted in the United States decreased. Similarly, the authors showed that the United Kingdom did not see the same downturn in sewing machine patents. They also insist on the fact that at the time, British patents were known to be much more expensive and administratively complex to obtain than American patents.

The two other examples of patent pools are less negative regarding the effect on innovation of this type of agreement. Before presenting them, we can note that the term “agreement” suggests a violation of the rules of competition. This aspect did not escape the competition authorities who, notably in the middle of the 20th Century, were suspicious of patent pools, seeing them as a mechanism that facilitated collusion, in light of a reinforced market power, rather than as an attempt to coordinate in favor of the development and distribution of innovations²⁴. The case of sewing machines detailed above provides valid arguments. The American patent pool for aeronautics followed the patent war already noted in Chapter 1. Due to the very large American patent that they obtained for flight control systems for airplanes, and although this patent initially concerned a system that was rapidly surpassed by others developed in Europe at the same time, the Wright brothers kicked off a period of exacerbated litigation at the start of the 20th Century in the United States²⁵. This litigation opposed them to Glenn Curtiss starting in 1908. It was not until the time when, observing the difficulty for the American industry to provide the planes needed following their involvement in World War I in April 1917, that the American authorities put pressure on them and the industrialists implemented a patent pool. The pool made it possible to put an end to the blockage in the American aeronautic industry. The case of the automobile, documented by Welsh [WEL 48] among others and presented in detail in the work of Greenleaf [GRE 61], was contemporary to the aeronautics case. At the source of this case is the American patent by George Selden originally filed in 1879²⁶ but only published in 1895, due to the many modifications made to the request during the examination phase. The patent concerned an automobile vehicle using an internal combustion engine with a hydrocarbon base. Although neither Selden nor the Electric Vehicle Company, which obtained an exclusive license in 1899, commercialized such a vehicle, the latter adopted an aggressive patent assertion policy against American automobile manufacturers at the time. Despite the presumed fragility of the patent, these manufacturers had to join with the Electric Vehicle Company to form the Association of Licensed Automobile Manufacturers (ALAM). Their main reason was to negotiate a “wholesale” price for the license that would be less than the cost of litigation. Henry Ford, who founded his eponymous automobile company in 1903, Ford Motor Company, refused to follow this logic and in 1911, won his case against ALAM. Although it dissolved after it lost that case, the Association was reformed under different names to become from 1934 to 1977 the Automobile Manufacturers Association (AMA) which, as of 1939, regrouped almost all of the manufacturers, with the notable exception of the Ford Motor Company. Although it had a wider scope of abilities, the AMA de facto exercised the role of a patent pool. Joining the AMA involved subscribing to a cross-license agreement with all other members. As noted by Welsh [WEL 48], this contributed to raising concerns of litigation and constituted a de facto patent abrogation system between AMA members. It is remarkable that in the two cases of the aeronautics and automobiles, the initial blockage that pushed for the implementation of a pool was attributable to the aggressive assertion of a patent whose technical interest was not self-evident and was not commercially exploited. Additionally, we can consider that these patents were both excessively large. The Wright brothers’ patent covered all flight control systems even though the holders had really only developed a wing torsion system. George Selden’s patent concerned the very principle of motorization of a vehicle from internal combustion of hydrocarbon, even though there was only one particular solution for this type of combustion that was really proposed. It is symptomatic that, at the same time, countries in western Europe also granted patents in these two sectors and succeeded in positioning their industries on the cutting edge without encountering the blockage problems that appear to have been uniquely American. The literature tends to suggest that, since this time and even previously, the American patent system operated with too little discrimination, if not that it was lax.

Although, unlike the sewing machine case, there does not seem to have been a quantitative study to date assessing the effect of American patent pools for aeronautics and automobiles at the start of the 20th Century, the theoretical works published tend to establish the conditions in which a patent pool not only promotes innovation, but also guarantees a general positive effect on welfare. It was Lerner and Tirole [LER 04] who showed what these conditions are. These authors demonstrated, more specifically, that a pool of complementary patents acts in the name of and on the account of, its members in matters of licenses granted, favoring innovation in a socially optimal way. Also, and above all, this pool is maintained even though members of the pool are allowed the freedom to issue licenses separately to non-members. Innovation continues to be favored in the case of a pool of substitutable patents, but it is not guaranteed that its effect will be socially optimal. In fact, an excess of innovation may result in wasting resources dedicated to producing innovations that are too similar. Above all, the authors emphasize that the pool can be unstable in the sense that it disappears when, in parallel, the members are authorized to grant individual licenses at the same time. Consequently, Lerner and Tirole [LER 04] interpret the fact that certain pools authorize their members to issue individual licenses to be a credible sign of social interest in the pool intended for regulation authorities.

Layne-Farrar and Lerner [LAY 11b] refine the understanding of rules of sharing revenue drawn from a pool between its members. In particular, they study the decision whether or not to join a pool for firms holding patents that are essential for information communication technologies (ICTs) standards. The interest in working on such patents is that firms holding SEPs are easily identifiable through the organization in charge of the standard. However, not necessarily all SEP-holding firms are part of a pool, or several pools can form on a portion of all SEPs concerning the same standard. Layne-Farrar and Lerner [LAY 11] showed that R&D intensive firms, a fortiori start-ups for whom innovation is the core of their work, privilege revenue-sharing mechanisms over the pro rata of the supposed value of patents. An alternative rule, privileged in the case where members are symmetrical enough in terms of size and business model, is the rule of pro rata sharing of the number of patents. The rule discussed by Bittlingmayer [BIT 88] for the American aeronautics pool appears relatively similar to the extent that it established the leadership of the Wright-Martin Company (which succeeded the firm created by the Wright brothers) and the Curtiss Company at the origin of the litigation over the patents. The rule provides for adjustments at the time of the expiration of the patents of either firm as well as revenue caps. Finally, Layne-Farmer and Lerner [LAY 11b] also address another rule, that of using a pool to relinquish all licensing rights and giving freedom of access to inventions on which it holds a patent. The rationale for such pools put forward by the authors is that members could thus promote the spread of innovations for which they sell complementary products. The authors take up the argument advanced by Allen [ALL 83] to explain the phenomenon of collective invention, or in other words, the a priori free contribution of a community of actors to develop an invention. The idea of complementarity was notably advanced by Allen [ALL 83] in the case of blast furnaces in the district of Cleveland in the mid-19th Century in the United Kingdom. The blast furnace owners were often owners of ore deposits. By graciously contributing to the collective improvement of blast furnaces, these owners saw the demand for ore stimulated by the decrease in iron production costs and in the end, earned a monetary gain. The financial aim of a priori altruistic behavior is at the heart of the debate about open innovation. We can therefore see patent pools as not only a cooperative approach between members in the valorization of their patents, but also as a free and open (although nevertheless involved) approach to their patent portfolio, especially for members granting open licenses. In other words, the premises of an open innovation strategy.

4.3.2. From R&D cooperation to open innovation

Although open innovation is based on a larger strategy than R&D cooperation between two firms, the latter aspect is nevertheless a link. Cooperation agreements were observed and analyzed earlier than other characteristic elements of an open innovation strategy, often even before the concept of open innovation had been formulated. Nevertheless, the literature on R&D cooperation agreements provides a first series of highlights that can be extended to open innovation.

It was notably Mariti and Smiley [MAR 83] who proposed one of the first quantitative studies on the subject. To do this, the authors relied on an examination of European financial press articles addressing such agreements, completed by interviews. Somewhat more recently, Arora and Gambardella [ARO 90] discussed the emergence of R&D agreements in the chemical and pharmaceutical sectors as of the 1970s in Europe, the United States and Japan. Starting from the observation that such agreements break with the traditional approach to innovation that corresponds to a vertically integrated model and functions independently, these authors looked at the possible causes for their emergence. They argue that the increasing complexity of innovation in these sectors, especially the upsurge in biotechnologies, required a diversity of assets that traditional actors in the sector did not necessarily have entirely. The rapprochement between actors made it possible to play on a complementarity of the assets provided by each of them. The typical example is that of a start-up which, thanks to its direct links with the academic world, provides a firm that is well-established in the industry with a skill which the firm did not have, for instance in the production of new molecules and the investigation of new associated markets. The start-up, which hitherto did not have the necessary know-how for mass production, gains this from the firm.

Three alternative choices are presented to bring together the complementary assets of the start-up and the large firm. The first is a merger between the two entities through the purchase of the start-up by the firm. Apart from its potentially elevated financial cost, this purchase can generate integration issues for the human resources of the start-up and the firm, out of distrust for the academic partners of the start-up vis-à-vis a major private actor with which they are not familiar or cohabitation problems of two business cultures that are too radically different. The second choice corresponds to the opposite case, or in other words, a minimal rapprochement, reduced to a transfer of patents or issuing a license on a patent. Although this solution has been described as a major characteristic of the evolution toward a technologies market in Chapter 1 as well as by the same authors in two distinct publications (Arora et al. [ARO 01], Arora and Gambardella [ARO 10]) it is not imposed systematically. The knowledge to be transferred can prove to be too tacit to be effectively transmitted through a patent document, regardless of how complete and well-written it is. Between these two opposite cases, an R&D cooperation agreement can be a good compromise to transfer tacit knowledge without needing to re-think the organization of the two entities in-depth. Mariti and Smiley [MAR 83] add a fourth possible choice to these three, which is the creation of a “joint venture” between the two entities. In the panel of cases that they had constituted by 1980, this fourth choice corresponded to 55% of cases, compared to 29% of cases passing through a specific agreement for the project. Likewise, they complete the argument based on the complementarity of assets (which are involved in 41% of their cases), by arguments founded on the tacit nature of the knowledge (29% of their cases), on the exploitation of economies of scale (16% of their cases) and risk sharing (14% of their cases). Mariti and Smiley [MAR 83] also emphasize the pro-competition nature of the R&D agreements in the case where they determine the survival of one of the actors and thus prevent a larger concentration in the sector in question.

The development in Europe of the Community Innovation Survey (CIS), with its different national variations, led to a series of empirical works about cooperation between German (Becker and Dietz [BEC 04]), French (Miotti and Sachwald [MIO 03], Negassi [NEG 04]) Spanish (Bayona et al. [BAY 01], Arranz and Fdez. De Arroyabe [ARR 08]), British (Tether [TET 02]) and Dutch (Belderbos et al. [BEL 04]) firms. While taking up the argument about the complementarity of assets and resituating it in the context of a resources-based approach of the firm, most of the works focus on the study of the factors for the choice of partner in cases of cooperation. They confirm that these complementarities favor vertical cooperation with suppliers and clients, rather than horizontal cooperation between competitors. Vertical competition also appears to be a relatively effective way to internalize the knowledge loss inherent to contacts with upstream and downstream partners. In other words, vertical cooperation starts from the observation of the existence of knowledge losses, known as “spillovers” and seeks to organize them to better take advantage of them. Spillovers between competitors also exist, if only through the mobility of personnel between companies, but are more delicate to internalize with horizontal cooperation agreements. The risk of transferring to a competitor any elements that may constitute a competitive advantage is increased in this case. The existence of economies of scale in R&D activity and the perspective of sharing (while diversifying) the risks associated with new innovations for the market (and not only for the firm), are likely to offset the problem of knowledge loss in the case of horizontal cooperation. Whether it is vertical or horizontal, R&D cooperation seems to be facilitated by the absorption abilities of the firm, measured especially by the intensity of their R&D activity.

The increased desire for vertical R&D agreements compared to horizontal R&D echoes the concept of user-driven innovation. This concept refers to situations where users of a good or a process create a formal or informal network in order to exchange information about their user experience and improvements that could be made. The phenomenon is old. Allen [ALL 83] notably exposed the case of blast furnaces in the district of Cleveland in the United Kingdom. Owners and engineers of these blast furnaces proceeded by trial and error to optimize their production, making modifications to their design and communicating their results. This process of trial and error did not involve a financial effort in terms of R&D properly speaking, but resulted in a series of incremental innovations which, combined, lead to a substantial improvement in the effectiveness of blast furnaces. All interest in the cooperation by exchange of information is that, not only does each party take advantage of the feedback of others’ experience and learn more quickly than if he was isolated, but he can also hope that another party will take up his experience and contribute an element that he had not initially imagined. This is the idea of the complementarity of knowledge and skills. Meyer [MEY 03] offers four other historical examples: the steam engine in Cornwall in the 19th Century; steel production in the United States during the second half of the 19th Century; the development of microcomputing by the Homebrew club in the United States in the 1970s; and finally, the development of the Linux operating system. In these examples, we find the role of user experience feedback and its contribution to the improvement of the good or system in question. Meyer [MEY 03] as well as Bessen and Nuvolari [BES 11a] note the connection between some of these cases with considerations of intellectual property. In the case of steel production in the United States, it was the difficulties of implementing a process patented by Henry Bessemer that pushed users of this process to exchange technical information, first in the context of a patent pool created by Alexander Holley and then by regular technical publications organized by Holley. The principle of these exchanges in the form of technical notes published and distributed within a network can be found in the example of the steam engine in Cornwall. Meyer [MEY 03] proposed the first modelling of experience exchanges, based on individual economic rationality, more specifically search theory: economic agents proceed by trial and error in order to reduce technological uncertainty, and they gain from doing it by exchanging if each agent can complete the experience of the other, until this uncertainty is sufficiently reduced and they consider that they either have enough visibility to find that a purely individual exploitation would be profitable, or enough negative messages to abandon it. The two other examples concern the development of computer science and the information communication technologies (ICTs) in the 20th Century. This sector was the subject of specific studies because of its novelty and its mediatization through the open source movement. Henkel [HEN 06] and then Henkel et al. [HEN 14], by respectively studying the case of objects functioning on the embedded Linux operating system and the case of drivers using Linux for computer components or accessories, insisted on the cost-benefit approach of making the codes public and free. The trade-off is not only between publishing the entirety of the code free or publishing nothing. As they specify, the rules of the Linux community (notably the General Public Licence) are such that it is possible to delay (while waiting for an explicit request for publication to be formulated, for instance) and exploiting a first-mover advantage in order to make only a fraction of the code public and free. On average, only half of the code is published for free. The costs associated with the production of the code are fairly obvious and are represented by the risk of being copied by competitors. The advantages highlighted by the interviews and surveys conducted by Henkel [HEN 06] and Henkel et al. [HEN 14] fall under three categories. The first category is associated with a question of reputation. Not communicating its code even though Linux is an open system is poorly viewed by the community of users and risks being counterproductive in terms of market opportunities. In other words, the openness calls for even more openness. The second category concerns the improvement of the code by users if it is free access, and the resolution of bugs which strengthens the quality of the product for free, this time for the firm. Also entering into this category is a standardization approach on their solution on the part of pioneer companies and, as a consequence, a first-mover advantage. The third category refers to the phenomenon of hold-up discussed in Chapter 3. By relinquishing its intellectual property rights on a code, a firm credibly eliminates concerns for hold-up because there is open access to the code irreversibly.

Studies on historical cases and those on quantitative analysis in the biotechnology and ICT sectors allow for perspective and mitigate the image of open innovation devoid of individual interest²⁷. On the contrary, we see an economic logic emerging that is completely in line with the methodological individualism on which a large part of economic analysis is based. The approach used by Lerner and Tirole [LER 02] is very explicit on this point. Without necessarily rejecting alternative explanations, Lerner and Tirole [LER 02] clearly announce their desire to show that they are not required to explain the open source approach in the ICT industry and that the economic analysis based on individualism is not inconsistent with the observation of this approach. Using four examples (Apache, Linux, Perl, Sendmail), they explain that by contributing freely to the development and improvement of a product, the members of the community send a credible signal concerning their skills. This credibility is assured precisely by the possibility that any third party, especially a potential recruiter, could observe and verify the result. It sometimes also passes through an accreditation procedure to publish the contribution, fairly similar to what happens for academic careers. This signal effect is particularly important for young graduates (and even more so for non-graduates that still have strong skills) wishing to begin a promising career. Harhoff et al. [HAR 03b] give a broader overview of the pecuniary incentives to get firms to reveal private information about their R&D voluntarily. Aside from the incentives already noted, they emphasize, by relying on a work document from 2000 later published by Lim [LIM 09] that describes the case of IBM with semiconductors using copper instead of aluminum, that a firm can have an interest in quickly spreading an innovation to its competitors. As contradictory as it may seem, and although innovation is costly to implement for suppliers, the fact of offering them a larger market by spreading a new technology to competitors of the firm also allows suppliers to better amortize their investments and therefore accept the technology change. However, this requires benefitting from a lead time in the development of technology. This is what IBM did by partially disclosing its new method of producing copper-based semiconductors starting in 1997. The new manufacturing process was not intuitive and, rather than risk a blockage with its manufacturers, IBM preferred to incite them by imposing the new manufacturing method as a standard for the industry. Harhoff et al. [HAR 03] also note by referring to the theoretical works of De Fraja [DEF 93] that it can be strategically optimal for firms engaged in a patent race to reveal knowledge acquired by its R&D services. The idea, which is somewhat intuitive, is that if the gains of innovation do not go entirely to the winner in a patent race (which occurs when the breadth of patents, as defined and discussed in Chapter 1, is not maximized) it can be preferable to be a second inventor, but quickly, rather than a first inventor, but belatedly²⁸. This result shows that a strategy likely to enter into the context of open innovation is not necessarily incompatible with the use of patents.

4.3.3. Why is open innovation so “patent-compatible?”

If firms make recourse to cooperation for their R&D, then we can expect that this allows them to increase their performances in that area and that a surplus of innovation would result. However, although R&D spending is a measure of R&D effort, patents are often used as an indicator approximating the result of this effort²⁹. We could expect a positive causal relationship of R&D cooperation agreements on patent filing. Inversely, however, the number of patents held by a firm being an indicator of its performance in matters of innovation, we could expect that, all other things being equal, a company that has many patents would more easily find partners with which to cooperate in matters of R&D. Using a database of start-ups specialized in biotechnologies, Shan et al. [SHA 94] conducted a first exploratory work by attempting to test the direction of the causality. By considering the potential simultaneity of the two variables, they found a significant causal effect of the number of cooperation agreements on the number of patents, but not the inverse. In line with this initial result, Cassiman and Veugelers [CAS 02] then explored other paths than the one focusing on intellectual property to explain which company behaviors could go hand in hand with a cooperation strategy. To do this they used information from the Belgian chapter of the Community Innovation Survey. The corresponding questionnaire included questions about sources of information as well as protection methods for firms. In their empirical study, Cassiman and Veugelers [CAS 02] were able to distinguish the companies based on two dimensions, in addition to the one relative to R&D cooperation. The first dimension corresponds to a firm’s absorption capacity for incoming spillovers. This absorption capacity is understood through the accumulation of the types of sources of external information that a company uses, including the monitoring and the reading of patents filed by others; participating in specialized conferences; access to technical or academic publications and participation in professional trade shows. The second dimension corresponds to the appropriation capacity of product knowledge, or in other words, the ability of a company to control outgoing spillovers. This is partly based on legal protection, especially through patents filed or trademark and copyright, but also through secrecy, complexity or lead time. Cassiman and Veugelers [CAS 02] distinguish the effect of these two capacities on cooperation by whether the cooperation is vertical or horizontal and if it occurs with a company or research organization. The authors conclude from their empirical study that a strong absorption capacity significantly increases the probability of cooperation with research institutions while a strong appropriation capacity increases the probability of vertical cooperation. Inversely, vertical R&D cooperation leads to a lower appropriation capacity while strengthening cooperation with research organizations implies a stronger absorption capacity. R&D cooperation and appropriation therefore seem to interact in a contradictory way, which is confirmed by reading works that sometimes seem to contradict each other. Jensen and Webster [JEN 09], for their sample of Australian companies, highlight that firms privileging a more closed approach to innovation tend to protect themselves more, especially with patents and the reverse for firms that opt for a more open approach. However, Zobel et al. [ZOB 16] showed that in the American solar industry, the patent portfolio of new entrants tends on average to increase the number of cooperation agreements in which they engage.

The ambiguous relationship between an open innovation strategy and recourse to patents has been summarized by Laursen and Salter [LAU 14] under the term of the paradox of openness, taken up by Arora et al. [ARO 16]. In essence, openness requires disclosing knowledge to the outside but is only beneficial on the condition of profiting more strongly from incoming spillovers than suffering from outgoing spillovers. Patents actually make it possible to disclose information about inventions while protecting them. In doing so, they facilitate an open innovation strategy because they limit the negative consequences of outgoing spillovers. As noted by Hagedoorn and Ridder [HAG 12], they also play a role as a credible signal to potential partners fairly similar to the one discussed in Chapter 1 on the topic of external funding for innovation through venture capital. It is in this sense that open innovation is not in opposition with patents, but even partly built on patents. However, this effect has its limits. Excessive use of patent protection complicates the contractual relations with R&D partners and can cause concern about opportunistic behavior. The efficient mix vis-à-vis partners between the stimulating role of the patent portfolio held by a company and its dissuasive role, corresponds to the inverted U effect demonstrated by Laursen and Salter [LAU 14] concerning the relationship between the intensity of the appropriation strategy by patents and the openness strategy.

An important element of the connection between openness and patents that remains little-explored is that of the property regime (or not) on the result of R&D cooperation. A “natural” regime seems to be one that includes the co-ownership of inventions that result from the cooperation. This co-ownership regime exists with most patent offices allowing several co-filers for the same patents. The rights and obligations inherent to the co-ownership of patents varies rather strongly from one patent office to another. In 2007, the International Association for the Protection of Intellectual Property (AIPPI) published a report summarizing the differences in rights and obligation associated with the co-ownership of patents in different offices [AIP 07]³⁰. Significant differences appeared in at least five areas. The first area concerns exploitation. In the case of exploitation by only one party of the co-owners, this party must pay compensation to the others in the case of French patents. Inversely, exploitation is totally free for co-owners in the United States, Germany and Korea. The second area concerns licenses. Issuing licenses generally requires the agreement of all co-owners, with the notable exception of the United States, where a co-owner can freely issue licenses without requesting the others’ agreement. In the American case, conflicts can arise if two co-owners separately issue exclusive licenses to different third parties. Elsewhere, each co-owner’s freedom to act is much more hampered. The third area concerns the transfer of property rights to a third party. It requires the agreement of co-owners in some countries (Australia, Brazil, Finland, Germany, Italy and the United States) but only an obligation of information for co-owners in other countries (the Netherlands, Iceland). For a third series of countries (France, Spain, Norway and China) there is a pre-emptive right for co-owners. The fourth area concerns the possibility of mortgaging. Often free, filing a mortgage sometimes requires the agreement of co-owners in some countries, such as Singapore. Finally, legal action with the goal of asserting a patent is at the discretion of each co-owner in Switzerland or subject to the obligation to inform the other co-owners in other countries (France, Iceland, China and Spain). In the United States on the other hand, it is done in the name of all co-owners and requires their agreement. Hagedoorn [HAG 03] stresses that this last characteristic can strongly affect the capacity to assert American patents³¹. The generally accepted opinion is that the patent co-ownership regime is not desirable because it strongly reduces the freedom of action of each co-owner. This result was confirmed by Belderbos et al. [BEL 14] who showed that co-patents (co-owned patents) in patent portfolios of publicly traded firms are negatively valued by the financial markets. However, these same authors showed that the part of co-patents in a set of patents is stable. Co-patents therefore remain an option, even if it is by default³². Alternative rules consist of predetermining which partner in the cooperation will be the exclusive owner of the patent based on the technology field of the patent and/or the importance of the mobilized R&D means and of contractually providing for the license obligations of other partners. However, these solutions can generate asymmetrical situations between partners which can complicate ex ante the implementation of a cooperation agreement. The agreement around the principle of the co-patent can therefore reduce ex ante the negotiation costs for the agreement and facilitate it, even if it means deferring the costs ex post. On the other hand, given that filing a co-patent involves trusting co-filers and relinquishing a freedom to act and a form of legal security, committing to such a system can be seen as a credible signal of a mutual desire for openness and cooperation³³. In any event, the question of intellectual property on the results of an R&D cooperation is therefore a key element of an open innovation strategy but remains very little explored in economic literature.

4.3.4. Patents at the center of intermediate innovation

The sharing of knowledge in an open innovation strategy implies technology exchanges. In addition to the contractual forms between partners described above, the exchange can pass through models of “distributed innovation” where the innovation is designed, developed and improved by a virtual “community” made up of volunteers who do not know each other³⁴. These models involve the implementation of new structures of intermediation intended to accelerate the intake of external knowledge.

These hybrid models of innovation, mid-way between the classic model of innovation and open innovation, call on crowd sourcing to fuel the innovation process (Howe [HOW 06]). Through an open access platform, companies or organizations confronted with unresolved difficulties in the deployment of their innovation policy can anonymously expose the problems they encounter and propose an award for each challenge in case of receiving a solution that is deemed appropriate³⁵. Innovative organizations act as intermediaries³⁶. Their role can be similar to that of a broker to the extent that they connect companies or organizations confronted with a breakdown of ideas about questions of innovation (seekers), with firms or individuals who can think about and find solutions (solvers). The use of services by solvers is free. This is an illustration of the typical function of two-sided markets, which offer an intermediation service to two distinct groups of users for whom the cost of the search for a partner would be much higher than the membership cost³⁷. The business model of this type of platform comes from the fee-for-service payment for the published call to tender (in the order of $10,000, this corresponds to the formulation of the challenge by the seeker) and from a commission collected on the total amount of the financial compensation in the order of 40%, applicable only if the challenge is awarded.

R&D subjects externalized in this way are generally very specific. They concern the questions on which internal teams are stumped or dormant subjects to which the company does not wish to commit internal resources. The problems posted are sufficiently broken down to allow different individuals to solve them. This granularity also contributes to ensuring that solvers, who self-select, do not dedicate too much time to searching for unsolvable problems. Solvers and seekers remain anonymous throughout the process in order to mitigate the risk of revealing exclusive knowledge about internal scientific programs. The solutions provided for the submitted problems are the subject, every time possible, of the filing of an intellectual property right³⁸. This is the case whether or not the solution is retained by the company that submitted the problem. The company that finds a solution adapted to its problem acquires the solution of the solver in exchange for the agreed price. The platform acts for solvers as a guarantee of their intellectual property rights. IP management is innovative because the license contract, signed ex ante, is standardized in the sense that the clauses are uniform from one solver to another. This type of IP management, where the negotiation is completely removed, is opposed to traditional forms of licenses for which the parties enter into negotiation ex post and individually on the conditions of the technology transfer, price and duration. The interest of this approach is to reduce as much as possible the uncertainties and problems posed by knowledge and rights transfers. The platform guarantees to solvers whose solutions were filed and examined but not retained that the client companies will not retrieve them. In other words, the platform keeps watch for patent assertion. This can generate “sleeping patents” in the sense that they are not exploited, which makes part of the activity of the platforms similar to that of patent assertion entities (PAEs) studied in Chapter 3, more commonly known as “patent trolls”.

The major interest for companies who submit problems is access to a global base of solvers and a decrease in R&D costs³⁹. The company commits a modest sum compared to the usual cost of R&D to obtain a solution, notably because it does not finance the cycles of trial and error inherent to an innovation process. The probability of obtaining a solution is even greater because the platform attracts a large number of solvers from diverse backgrounds and, inversely, a platform is all the more attractive for solvers when it proposes a wide variety of problems and increases their probability of finding a problem that coincides with their skill set. The fact that both sides of the platform (the seeker side and the solver side) gain when the other side includes a large number of actors is characteristic of two-sided markets. More specifically, each new actor for one side increases the expected gain of the platform on the other side, but does not incorporate this effect in its decision to join the market. Otherwise, there are cross-network externalities and the platform must have a certain critical mass for each side to work. The use of this type of intermediation also allows companies who submit their problems to avoid implementing mechanisms to filter and monitor solvers because the solver will self-select by responding to the challenge or not⁴⁰. This also allows them to benchmark solutions that are proposed to them and to not be stuck on a particular approach to the problem, like it might be if it only counted on its own services. Beyond the pecuniary advantages, reputation and recognition are also key driving forces behind the motivation of individuals, often scientists, to participate in the challenges (Lakhani [LAK 07]), followed by a desire to share their knowledge with the community. Participation in this type of challenge is a signal on the job market and makes it possible to build skill and reputation with potential employers.

The impact of this model of platform on innovation is limited at the present time, however. First of all, we can observe a relatively high failure rate. Data from intermediation platforms reveal, for example, that most attempts to solve submitted problems fail. InnoCentive’s resolution rate is 30%, based on about one satisfying solution out of 10 submitted for each problem. There is a potentially strong bias in this failure rate for distributed innovation systems. It is probable that companies submit their problems as a last resort, after noting the defeat of their own R&D services. In other words, the problems submitted are likely to be the problems that are the most difficult to solve from a seeker point of view. Then, there is a strong random element connected to distributed innovation systems that is such that they cannot be used in the same way as traditional R&D organizations. Internal R&D services function on the basis of an annual or multi-annual plan that predetermines what they work on. A company can increase the resources in order to reach a result more or less quickly, even if the innovation process remains inherently random. It does not have this leverage in the case of a distributed innovation system and is therefore confronted with a potentially stronger risk regarding obtaining a solution⁴¹. In addition, distributed innovation is not a cure-all because, although its virtual platforms make it possible to generate answers for precise problems, up to now they have not permitted the development of major innovations. Finally, within traditional organizations, there is a lot of resistance internally to embrace distributed innovation systems (which can be perceived as a challenge of their abilities), especially because over-the-counter contracts are not easy to implement and often require time to arrive at a satisfying arrangement for the two parties.