Alexander Tabarrok, of the Marginal Revolution and George Mason University, writes this paper, Patent Theory versus Patent Law about, you guessed it, how patent theory compares to patent law as it is practiced. Refer to this blog post on how to get the PDF.
The abstract, verbatim:
According to the economic theory of patents, patents are needed so that pioneer firm have time to recoup their sunk costs of research and development. The key element in the economic theory is that pioneer firms have large, hard to recoup, sunk costs. Yet patents are not awarded on the basis of a firm’s sunk costs. Patent law, in fact, ignores costs. The disconnect between patent law and patent theory suggests either that modifying patent law so that it better fits with patent theory would reduce the costs and inefficiencies associated with current patent practice or that the standard economic theory of patents is wrong.
Alexander first explains the economic theory briefly and then puts forth the shape of his argument:
Although the economic theory is well accepted, I argue that it does not fit well with the actual patent system. A patent system designed around the idea of recouping sunk costs would look quite different than the current system. In particular, as I will explain below, the current system ties returns to innovation to the benefits of a patented idea, i.e., as the value of the idea increases monopoly profits increase. Yet the economic theory of patents implies that returns should be tied to the sunk costs of researching and developing the patented idea.
The poor fit between the economic theory of patents and the actual patent system suggests two alternative hypotheses. Either the theory is correct and the patent system is poorly designed, or the patent system is well designed but not for the purposes of recouping sunk costs. Which of the two hypotheses one prefers depends in part upon one’s priors about therelativeefficiency of theory versus practice. My primary approach is to assume that the economic theory is correct. Thus, I argue for reforms that would bring the patent system more into accordance with the economic theory.
This section explains the theory of patents and briefly mentions that the current patent system is not connected to the economic theory because it ignores the costs that innovates face that imitators do not. Those are sunk costs that the economic theory is concerned with.
An interesting observation of this section is that when a product’s value is high, it should not be protected as much as a product whose value is low. This is because since the sunk costs are not related to the benefits, they are repaid easier.
Practice: Examples of Patented Products with Low Innovation to Imitation Costs
Amazon’s “one-click purchasing” is a patent that was granted for something with little to no R&D costs. Another is the discovery of a medical procedure, which by definitions means there were not R&D costs.
Edison famously said that “Genius is one percent inspiration, ninetynine percent perspiration.”8 A patent system should reward the ninetynine percent perspiration, not the one percent inspiration. In inventing the lightbulb, for example, Edison laboriously experimented with some 3000 possible materials for the filament, before hitting upon carbonized cotton thread (Shulman 1999). If Edison were to patent the lightbulb today he would not need to go to such lengths. Instead, Edison could patent the use of an “electrical resistor for the production of electro-magnetic radiation,” a patent that would have covered oven elements as well as lightbulbs.9
The key thing that needs to be done is that the sunk costs of an patented product must be integrated into patent policy, somehow. Alexander proposes multiple ways of doing this:
- Changing what products can be patented, preferably by looking at standard innovation/imitation cost ratios in the industry.
- Changing the probability that a patent is granted (or enforced.)
- Changing the length of the patent.
- Changing the breadth of the patent.
An interesting suggestion is to register innovations but not award patents, instead you wait twenty years and estimate how much the innovation was worth and award that with interest. This solves the problem of not being able to estimate correctly the benefits of the innovation before hand.
The “realistic” proposal that does not strive for complete perfection is to allow firms to apply for patents of varying lengths, and as the length increases the scrutiny of the patent office. But, Alexander tries to point out many changes that may not require legislation and would instead modify the discretion of patent examiners and infringement judges.
Objections and Implementation
“Perhaps the strongest objection against a patent system that takes into account sunk costs is that measuring sunk costs is difficult (Scotchmer 1988).21 If the sunk costs are born by the patent applicant then this problem does not seem overly difficult.”
And ever the economist:
In order to improve the current system it is not necessary to estimate sunk costs precisely so long as they can be estimated well enough to reasonably assign patent duration to a limited number of categories. If sunk costs justify a patent of 20 years, for example, it’s unimportant whether they are two, three or four times greater than necessary. The thrust of this proposal is to make marginal changes in the current patent system rather than to replace that system with something entirely new.
(Note: There’s a typo in note 23, “The higher sunk costs are” not “The higher are sunk costs”–I think.)
There is Little to Lose from Patent Reform
If it were clearly true that on net the patent system increased economic growth and technological advancement then on precautionary grounds alone there would be a good case against reform. But the consensus from many studies of innovation is that most innovations would occur without patents.
Alexander then discusses the proof of this that has been studied and documented by others. He also cites in the footnotes someone, Cohen et al. (2000), who found that “much patenting is simply generated by the patent system itself as firms patent in order to protect themselves from the patenting of other firms.”
But, Tabarrok recommends merely weak patents in cases where there are not useful, because they are useful in some instances. He is not completely against patents for any ideological reason.
Other Theories of Patents
In this section he addresses the concern that perhaps the economic theory of patents is wrong and that either another theory is necessary, whether it is another defined one (e.g. the disclosure theory) or a theory implicit in the current patent system.
Another brief summary like the abstract.