- Ex ante R&D Subsidies
A first alternative to the patent system could be R&D subsidies (this idea goes back to Pigou, 1920). The granting of subsidies, proportionally to the amount of R&D investment, to the firms who invest in R&D can help to restore incentives. In this way, firms are urged to invest the quantity of R&D that maximises the social surplus and not only their own profits. However, several shortcomings can be pointed out: This method works only in a perfect world with no informational problems. The State must know ex ante not only the future private value of the invention which, most of the time, the inventor himself does not know, but it must know the economic value of the spillovers generated by the invention too. Moreover, a method of ex ante subsidies will generate moral hazard problems, due to strong asymmetries of information between the firms and the State. The latter holds the role of the principal in the principal-agent problem. Empirical studies (cf. for instance, Bernstein & Nadiri, 1988, 1991) showed that private R&D done with public subsidies is less productive than the R&D financed by the firm itself. Arrow (1962, b) speaks about a moral factor: Subsidies decrease the incentive of success because, like insurance, they decrease the risk associated with failure. Somehow, subsidies are an incentive to do R&D but not to do it well.
- Ex post rewards (or patent buyouts) to preserve incentives while ensuring full diffusion.
A system of ex post rewards, as compared with ex ante subsidies, would allow the State to benefit from more information about the social value of the invention and to solve the moral hazard problem. Moreover, when compared to patents, Kremer (1998) explains that such patent buyouts are attractive since they offer the opportunity to eliminate monopoly pricing distortions and incentives for duplicate research, while raising the incentives for original research. Thus, ex post rewards, or patent buyouts, would have the potential to perform better than both ex ante subsidies and patents. The idea of patent buyouts comes from the purchase of the daguerreotype (from the name of its inventor, Daguerre) by the French government in 1839.
A reward system may not only allow a better diffusion of the produced knowledge, but increase the incentives to do R&D as well. Rewarded inventions would be put in the public domain, thus society would earn huge benefits in the form of more learning by using the now free technology and from increased pecuniary spillovers. As argued by Polanyi (1943), a sufficient condition for rewards to perform better than patents is to warrant to the inventor a reward at least equal to the expected monopoly profit he would have realised with a patent.
We must never forget that our aim is merely to distribute prizes for successful inventions and that if the new system succeeds only in making these prizes as ample as, and not markedly less fair than, the rewards which are earned by patentees today, while maintaining practically complete freedom of the use of inventions by everybody, it may be said to achieve its purpose to the full. If the government were to fix the total sum allocated for public rewards at a level, which would just suffice to induce inventors to be as eager to obtain patents as rewards, the general public would be left with a handsome balance (Polanyi, 1943).
Even smaller rewards than the expected monopoly rent would be sufficient to induce the same level of incentives because the monopoly holder faces a risk in exploiting its monopoly and usually economic agents are perceived as risk averse. The reward system would especially be profitable for small firms, which do not have the financial capabilities to use the patent system. It would secure a fixed remuneration for them while the patent system is too risky for them.
As exposed by Kremer (1998), a major challenge for any system of patent buyouts is determining the price. From an economic efficiency point of view, the price that maximises the social surplus must take into account all the spillovers generated by the invention. Indeed, only this price can ensure that all the socially profitable innovations are implemented and only them. But how could this ideal price be computed. Lately, Kremer introduced an original and interesting method using an auction mechanism.
To sum up, its method contains three steps:
-- At the time t0 a firm decides to announce publicly the sale of one of its patents.
-- Interested firms reveal their bids in order to buy this patent and at the time t1, fixed in advance, the State computes the private value of the innovation according to the different bids.
-- From this private value, the State deduces the social value of the innovation, pays this price to buy the innovation and put it in the public domain.
During the second step, the State would face problems of collusion and of crazy bidders. To solve these problems, the State could choose, for instance, to fix the private value of the innovation equal to the third highest bid or to compute an average. Moreover, in order to deter crazy bidders, some offers would be randomly accepted and sold to the winner of the auction. Within the third step, Kremer proposes to compute the social value of the invention in doubling, at least, the private value obtained in the second step. Finally, as a safeguard against confiscation of inventions, patent holders could choose whether to sell their patent or not.
This system of auction may seem rather difficult to implement in practice. Thus, if the idea that among a particular industry, there exists a consensus about the value of an innovation (at least about the private value) is accepted, why not just set up a committee of experts, composed of several members considered competent in each industry, who would have the responsibility to estimate the private and the social value of each innovation? In this way, Polanyi (1943) proposed to let the market reveal the value of the innovation. Time after time, the State would reward the innovative firms at the end of each period according to the benefits due to the innovation, made by the society during this spell of time. According to Polanyi, the benefit arising from a particular invention could be traced with a certain degree of precision to a definite set of persons. Even if this is a very ambitious and strong guess denied by almost every econometric specialist, for the modest purpose of Polanyi, to perform better than patents, it may contain some truth. As the State would dispose of more and more information about the benefits of the innovation it would be able to reward it at its most exact value. Moreover, this solution would have the advantage to split the reward over time and so, to avoid giving too much financial power to one firm.
Beside the problem of determining the ideal amount of the reward, four other shortcomings can be identified:
-- To ensure big firms with important amount of cash is equivalent to displacing the problem, not to solving it: Firms would benefit from a financial power to build monopolies rather than a legal power.
-- In allowing a fast and universal diffusion, rewards could lead to network effects and lock-in phenomena under efficient technology.
-- If it is assumed that rewards create at least as many incentives as patents, the problem of duplication of R&D ex ante because of reward races, will be amplified.
-- Maybe the most important problem faced by rewards is that the State has, ex post, an incentive to under-estimate the value of the innovation. As the innovator may know this tendency, he has less incentive to invest in R&D.
Thus, a reward system can only succeed in a climate of mutual trust between the State and the firms. But, as explained by Taylor (1995), since contract courts seldom possess the ability or expertise necessary to evaluate technical research projects, contracts that base compensation on research output are usually impractical. It must be noticed that most of these problems occur with the patent system too. Thus, ex post rewards may be especially interesting to use in specific industries where the static loss of patents is important.
- Research tournament or research race: A complementary solution to rewards.
The reward system may not provide enough incentives in sectors where the social value of an innovation is high but not the private value. In such cases, firms may not be able to evaluate the social benefits of an innovation. A solution for the State to orient firms to this kind of sectors could be the sponsoring of research tournaments or research races. Basically this kind of sponsoring can be seen as a reward system directed to specific sectors. Taylor (1995) distinguished research race and research tournament. He compared the first with the 500 miles of Indianapolis, in the sense that the goal (the specific characteristic of the desired innovation) is fixed but the length of the race (the time needed to meet the fixed criteria) can vary, and the second with the 24 hours of the Mans because the time is fixed but the final performance of the output may change; the goal is to do the best you can in a limited time. In one case the sponsor controls the time to reach a specific goal, in function of the amount of money he is willing to invest and in the other, he controls the quality of the desired output. The other tools at the disposition of the State are the number of participants and the participation fee (which may be negative if there are important fixed costs).
One of the biggest advantages of this system of tournament or race is that it can solve the problem of trust between the contractor and the inventor. In the case of a race all that a court will have to do is to compare the ex ante well specified criteria on, for instance, the global performance wanted for the invention, with the ex post propositions of the candidate (however, in the case of a tournament the court will still have to compare different candidates, maybe very close). As it is very easy to check if a reward has been paid or not to one of the participants, this kind of sponsoring is equivalent, for the sponsor, to gaining some credibility. Moreover if the sponsor is rational and does not have a favourite participant, he will have all the incentives to reward the best invention. Thus, this system may obtain good results in specific low market incentives.