Independent invention as a defense to patent infringement.

• Author: Vermont, Samson

Under current law, independent invention is no defense to patent infringement. This Article argues that independent invention should be a defense, provided the independent inventor creates the invention before receiving actual or constructive notice that someone else already created it. The defense reduces wasteful duplication of effort and enhances dissemination of inventions without lowering the incentive to invent below the necessary minimum. To be sure, the defense lowers the incentive for inventions that face significant odds of being invented by more than one inventor. By enabling a second inventor to compete with a first inventor, the defense essentially breaks up the first inventor's patent monopoly into a duopoly. Monopoly profits exceed the collective profits of duopoly. Thus, from the perspective of inventors ex ante the defense reduces the expected profit for inventions that could be invented by more than one inventor. This Article argues, however, that the reduction in expected profit is moderate and that the reduced expected profit is generally sufficient. Per Bayes' theorem, the fact that an invention could be invented by more than one inventor is itself evidence that a moderately reduced expected profit will still motivate at least one inventor to create the invention without inefficient delay.

TABLE OF CONTENTS INTRODUCTION I. PREVIOUS COMMENTARY A. Long's Theory of Information Costs B. Reinvention Defense for Concealable Inventions C. User Rights II. FEATURES OF THE REINVENTION DEFENSE A. The Thoughts and Acts That Constitute Reinvention B. One Patent per Invention C. When Reinvention Must Occur D. First Inventor as Reinventor E. Transferability III. FRAMEWORK FOR EVALUATING THE REINVENTION DEFENSE A. System Costs 1. Monopoly Loss 2. Rent Dissipation 3. Miscellaneous Costs B. Kaplow's Test IV. THE REINVENTION DEFENSE LOWERS SYSTEM COSTS WITHOUT DESTROYING NECESSARY INCENTIVE A. The Reinvention Defense Lowers System Costs 1. The Reinvention Defense Reduces Monopoly Loss 2. The Reinvention Defense Reduces Rent Dissipation 3. The Reinvention Defense Reduces Miscellaneous Costs B. The Reinvention Defense Seldom Lowers the Reward below Break-Even Point 1. Reward R Is Only Moderately Lower Than Reward C 2. Reward C Is Too High for Reinventables V. OBJECTIONS TO THE REINVENTION DEFENSE A. Inventors Will Learn to Issue Notice Immediately B. The Reinvention Defense Could Exacerbate Some Patent Races C. Reinvention Can Be Faked D. Patent Monopolies Can Be Reinstated through Bargaining CONCLUSION INTRODUCTION

Many scholars who study intellectual property (IP) argue that we should tailor rights in IP so that they provide protection only when output of IP would clearly be too low without protection. (1) This argument rests on a variant of one of the following three claims.

First, the last bit of supra-competitive pricing by a monopolist provides disproportionately small profit to the monopolist in comparison to its social cost. (2) For example, lowering by 10% the price of an invention charged under a patent monopoly (3) might reduce the patentee's profit by only 1% while reducing the social costs of the patent monopoly by 19%. (4) This suggests that the optimal patent monopoly falls short of providing complete exclusivity over the invention. (5)

Second, property rights serve two related functions: they provide (i)a way of determining who gets what when; and (ii) an incentive to create or improve the what. (6) For rights in IP, however, function (i) is perverse--because IP is largely non-rivalrous. We cannot both drive the same car to different places, but we can both use the same invention to build different machines and we can both simultaneously read different copies of the same book. This suggests that the case for rights in IP is weaker than the case for rights in tangible property. (7)

Third, IP engenders higher rent dissipation and transaction costs than tangible property does. (8) This too suggests that the case for rights in IP is weaker than the case for rights in tangible property.

These three claims are more or less sound. By themselves, they counsel in favor of erring on the side of under-rewarding invention. But they do not tell the whole story. The other part of the story is that once in a while an invention provides net benefits to society that astronomically exceed the net benefits to the inventor. Consider the polio vaccine. The benefits to the inventor, Jonas Salk, consisted of fame, gratification, and relatively modest financial gain. The social benefits consisted of millions of saved lives. No particular piece of tangible property can compete with the polio vaccine when it comes to the size of the gap between the net social benefits and the net private benefits to the owner or creator. Indeed, it is tempting to conclude that the social benefits of the polio vaccine exceed the combined social costs of every invention and IP right that has ever existed. This may also be true for inventions such as the smallpox vaccine, penicillin, and the pesticide DDT. (By preventing malaria, DDT reputedly saved hundreds of millions of lives. (9)) Nor are such great gaps between social and private benefit exclusive to inventions in health and medicine. Other super socially valuable inventions include the light bulb, telephone, combustion engine, airplane, television, transistor, and integrated circuit.

Would these super socially valuable inventions have ever been created and put to use had they been entitled to weaker patent protection? The answer is almost certainly yes. Super socially valuable inventions are precisely the inventions that are likely to be produced even in the total absence of patent protection. Weaker patent protection, however, might have delayed some of them. Weaker patent protection implies lower expected revenue for the inventor. An inventor will not pursue an invention unless her expected revenue exceeds her expected costs of invention. The costs of invention fall with time. (Inventing the polio vaccine today, for example, would not require the Herculean effort it required in the 1940s.) Thus, to weaken patent protection is to increase the risk that inventors will postpone invention.

Many IP scholars seem to believe that we can often get something for next to nothing--that by scaling back patent protection we can often markedly decrease its social costs without markedly increasing the risk of missing out on or postponing invention. Actually, we can get something for next to nothing, but the opportunities to do so are scarce. To find these opportunities, we must identify the inventions for which scaling back patent protection would avoid disproportionately high social cost or would pose disproportionately low risk of under-rewarding inventors. This paper identifies a class of inventions that can satisfy this requirement: inventions that face significant odds of being invented by more than one inventor at roughly the same time.

Under current U.S. law, the second to invent generally comes up empty handed. Suppose inventor Smith and inventor Jones toil away for years in separate efforts to build a better mousetrap. They conceive of essentially the same design but Smith does so a few months after Jones. Smith and Jones separately perfect the design, file for patents, and then start commercializing the mousetrap. Under U.S. law, only inventor Jones will receive a valid patent. As soon as Jones's patent issues (usually two to three years after its filing date), Smith can no longer make, use, or sell the mousetrap unless Smith acquires a license from Jones.

This type of neck-and-neck finish is common. Researchers frequently converge on the same idea at roughly the same time. (10) Famous examples include the light bulb (Edison and Swann), the telephone (Bell and Gray), the integrated circuit (Kilby and Noyce), calculus (Newton and Leibniz), the periodic table (Mendeleyev and Meyer), the telegraph (Morse, Henry, and Cooke and Wheatstone), the telescope (Hans Lippershey, Drebbel, Fontana, Jansen, Metius, and Galileo--each claiming they invented it in 1608 or 1609), (11) and certain facets of the theory of relativity (Einstein and Poincare). (12) Some historians and philosophers of science believe convergence is the rule rather than the exception. (13)

A crucial claim of this paper is that convergence is evidence that a smaller reward would have been sufficient to incentivize the invention. This claim rests on Bayes' theorem, which tells us how to revise an estimated probability in light of new information. (14) For our purposes, the estimated probability in question is the probability that complete patent protection, meaning the standard protection conferred under current law, provides an excessive reward for a given invention. A reward is excessive when it exceeds the minimum necessary to incentivize timely creation of the invention. The new information in question--the information we use to revise the estimated probability--is the number of inventors who create the invention, that is, the number of inventors who invent the same invention. If no inventor ever creates the invention, the probability is very low that complete protection provides an excessive reward for the invention. If exactly one inventor creates the invention, the probability is higher but still not high overall. If two or more inventors independently create the invention at about the same time (convergence), the probability of an excessive reward is high overall.

Therefore, I argue, in a case in which two or more independent inventors converge on an invention, the patent protection available for that invention should be ratcheted down moderately. A good way to ratchet down protection moderately is to automatically bestow a defense to patent infringement on the independent inventor(s) not entitled to the patent.

Hereinafter, this proposed defense is referred to as the "reinvention defense" or simply the...