Final Report Summary - INCPATSYS (Incentives in the Patent System)
The patent system constitutes the centrepiece of many countries’ innovation policy. It is intended to provide incentives to innovate and improve the diffusion of innovative knowledge. A crucial requirement for the patent system to play this beneficial role is that patents are granted only for novel and non-obvious inventions. Making sure that only true innovators obtain patents requires some sort of watchdog, a role that is played in practice by the patent office and courts. This research project studies the design of these institutions and their role in the innovation process. It develops a theoretical framework in which questions of optimal policy design can be addressed, and tests some of the resulting theoretical predictions empirically. The findings of the research project can help improve the design of the patent system, thereby leading to more innovation and a more efficient allocation of resources.
The project has resulted in three research papers, the first two of which deal with the patent office whereas the third one extends the analysis to include judicial review of patents. In “Patent quality and incentives at the patent office” (published in the RAND Journal of Economics) I study the design of incentives for patent examiners. Patent examination is modeled as a moral-hazard problem followed by an adverse-selection problem: the examiner’s incentives have to be structured so as to make him exert effort searching for evidence to reject (prior art), but also to make him truthfully reveal whatever evidence (or lack thereof) he finds. I assume that examiners are motivated by explicit incentives, such as performance pay, but also by implicit incentives within the organization (promotion, dismissal, etc.) inducing a desire to avoid mistakes. I show that these implicit incentives are crucial for effort provision. When implicit incentives are weak, the examiner’s effort is low, and the adverse-selection component forces the principal to structure the explicit incentives such that they reward the examiner for granting patents. When implicit incentives are stronger, explicit incentives can be used more effectively, both of which combine to raise effort. I argue that the model can explain the compensation scheme in use at the U.S. Patent and Trademark Office (USPTO), where examiners are essentially rewarded for granting, as well as variation in compensation schemes and patent quality across patent offices. In particular, high attrition in the examiner workforce at the USPTO means that implicit incentives, by definition long-term in nature, are largely inexistent. At the European Patent Office (EPO), by contrast, examiners typically stay for much longer parts of their career, implying stronger implicit incentives.
In “Inventors and impostors: An analysis of patent examination with self-selection of firms into R&D” (forthcoming in the Journal of Industrial Economics), I provide micro-foundations for the behavior of patent applicants. The paper develops a model in which firms choose between more or less ambitious research projects. Ambitious projects lead to social gains, while, if patented, unambitious projects lead to social losses. To obtain patent protection, firms have to file an application with the patent office. The patent office maximizes welfare and wields two instruments: an application fee and the intensity with which it examines applications; the combination of the two is what I define as patent policy. I first study how patent policy affects investment in R&D and characterize the optimal policy when firms are not financially constrained, so that the patent office can set application fees at any desired level. The optimal patent policy in this benchmark case involves full deterrence of bad projects. I then introduce a subset of small firms that have wealth constraints, so that they cannot afford high fees. I show that the patent office will examine applications from wealth constrained firms more rigorously than those from other firms. Interpreting wealth constraints as being correlated with firm size, I derive predictions on differences in grant rates, patent quality, and the private value of patents between small and large firms. To test the theory, I look at the 1982 reform that introduced differentiated fees in the U.S. The reform was associated with a substantial increase in application fees for large firms and only to a lesser extent for small firms. Using a difference-in-differences approach, I provide evidence that renewal rates and fee-adjusted renewal rates of patents issued to large firms increased relative to those of small firms, which is consistent with the predictions of the model.
In “Patent screening, renewal fees, and the courts” (joint with Mark Schankerman, London School of Economics; soon to be published in Tilburg University’s working paper series), we take into account the role of the courts. When a patent holder asserts a patent, alleged infringers often counterattack by challenging the patent’s validity, which can lead to the revocation of the patent. This means that legal rules, such as the evidentiary standard used in determining which of the litigants prevails, are crucial for the functioning of the system of patent review as a whole. We develop a model in which inventors have to first discover an idea through research and then turn the idea into a marketable innovation through development. They apply for a patent after discovery and then privately learn about the cost of development. The development cost represents a noisy signal of whether or not the invention is obvious. From a social perspective, obvious inventions should not be patent protected because society would receive their benefits even in the absence of patent protection, which causes deadweight loss. By contrast, non-obvious inventions must be developed by the inventor himself in order for society to receive their benefits. Because inventors have information about the likely obviousness of their inventions, a benevolent social planner would like to condition the post-grant scrutiny to which the patent is subjected on the inventor’s information. (One determinant of the level of scrutiny is the standard of proof applied by the courts in the event of litigation.) Inventions that are more likely to be obvious should face a greater level of scrutiny. We show, however, that a mechanism in which scrutiny increases with the probability of obviousness is not incentive compatible. These incentive problems imply that the social planner may have to apply a uniform level of scrutiny and forego some socially worthwhile inventions. Yet, the planner may sometimes still want to screen inventors in order to extract higher renewal fees.
The results of all three papers are of direct relevance for both patent offices and legislators. In particular, the research on patent examination has implications for the compensation of patent examiners, for the structure of application fees, and the optimal examination intensity. The research on the legal rules used in patent litigation has implications for the design of the European Union’s unified patent court.
The project has resulted in three research papers, the first two of which deal with the patent office whereas the third one extends the analysis to include judicial review of patents. In “Patent quality and incentives at the patent office” (published in the RAND Journal of Economics) I study the design of incentives for patent examiners. Patent examination is modeled as a moral-hazard problem followed by an adverse-selection problem: the examiner’s incentives have to be structured so as to make him exert effort searching for evidence to reject (prior art), but also to make him truthfully reveal whatever evidence (or lack thereof) he finds. I assume that examiners are motivated by explicit incentives, such as performance pay, but also by implicit incentives within the organization (promotion, dismissal, etc.) inducing a desire to avoid mistakes. I show that these implicit incentives are crucial for effort provision. When implicit incentives are weak, the examiner’s effort is low, and the adverse-selection component forces the principal to structure the explicit incentives such that they reward the examiner for granting patents. When implicit incentives are stronger, explicit incentives can be used more effectively, both of which combine to raise effort. I argue that the model can explain the compensation scheme in use at the U.S. Patent and Trademark Office (USPTO), where examiners are essentially rewarded for granting, as well as variation in compensation schemes and patent quality across patent offices. In particular, high attrition in the examiner workforce at the USPTO means that implicit incentives, by definition long-term in nature, are largely inexistent. At the European Patent Office (EPO), by contrast, examiners typically stay for much longer parts of their career, implying stronger implicit incentives.
In “Inventors and impostors: An analysis of patent examination with self-selection of firms into R&D” (forthcoming in the Journal of Industrial Economics), I provide micro-foundations for the behavior of patent applicants. The paper develops a model in which firms choose between more or less ambitious research projects. Ambitious projects lead to social gains, while, if patented, unambitious projects lead to social losses. To obtain patent protection, firms have to file an application with the patent office. The patent office maximizes welfare and wields two instruments: an application fee and the intensity with which it examines applications; the combination of the two is what I define as patent policy. I first study how patent policy affects investment in R&D and characterize the optimal policy when firms are not financially constrained, so that the patent office can set application fees at any desired level. The optimal patent policy in this benchmark case involves full deterrence of bad projects. I then introduce a subset of small firms that have wealth constraints, so that they cannot afford high fees. I show that the patent office will examine applications from wealth constrained firms more rigorously than those from other firms. Interpreting wealth constraints as being correlated with firm size, I derive predictions on differences in grant rates, patent quality, and the private value of patents between small and large firms. To test the theory, I look at the 1982 reform that introduced differentiated fees in the U.S. The reform was associated with a substantial increase in application fees for large firms and only to a lesser extent for small firms. Using a difference-in-differences approach, I provide evidence that renewal rates and fee-adjusted renewal rates of patents issued to large firms increased relative to those of small firms, which is consistent with the predictions of the model.
In “Patent screening, renewal fees, and the courts” (joint with Mark Schankerman, London School of Economics; soon to be published in Tilburg University’s working paper series), we take into account the role of the courts. When a patent holder asserts a patent, alleged infringers often counterattack by challenging the patent’s validity, which can lead to the revocation of the patent. This means that legal rules, such as the evidentiary standard used in determining which of the litigants prevails, are crucial for the functioning of the system of patent review as a whole. We develop a model in which inventors have to first discover an idea through research and then turn the idea into a marketable innovation through development. They apply for a patent after discovery and then privately learn about the cost of development. The development cost represents a noisy signal of whether or not the invention is obvious. From a social perspective, obvious inventions should not be patent protected because society would receive their benefits even in the absence of patent protection, which causes deadweight loss. By contrast, non-obvious inventions must be developed by the inventor himself in order for society to receive their benefits. Because inventors have information about the likely obviousness of their inventions, a benevolent social planner would like to condition the post-grant scrutiny to which the patent is subjected on the inventor’s information. (One determinant of the level of scrutiny is the standard of proof applied by the courts in the event of litigation.) Inventions that are more likely to be obvious should face a greater level of scrutiny. We show, however, that a mechanism in which scrutiny increases with the probability of obviousness is not incentive compatible. These incentive problems imply that the social planner may have to apply a uniform level of scrutiny and forego some socially worthwhile inventions. Yet, the planner may sometimes still want to screen inventors in order to extract higher renewal fees.
The results of all three papers are of direct relevance for both patent offices and legislators. In particular, the research on patent examination has implications for the compensation of patent examiners, for the structure of application fees, and the optimal examination intensity. The research on the legal rules used in patent litigation has implications for the design of the European Union’s unified patent court.