Patent Races with No Entrants

Abstract

It is generally said that patent races cause overinvestment, because when several entities are racing to create and patent an invention, only one can win the race and acquire the patent. As a result, the investments of the losers in the race are wasted. In this article, I address the opposite problem: underinvestment in patent races. Underinvestment can occur, I show, when a participant that would ultimately win a patent race has reason to believe that it will lose the race, and therefore drops out. This can occur when the participant suffers some early setbacks in its research, but does not know whether its competitors have had similar problems. In such circumstances, the participant may believe that it has fallen behind and quit the race, when it would have won if it had continued. In some circumstances, all the participants in a race can believe that they will lose the race, and all may abandon the race.

Both overinvestment and underinvestment are caused by a lack of information. Overinvestment occurs when a participant incorrectly chooses to continue its research because it does not know that it will ultimately lose the race. Underinvestment occurs when a participant incorrectly chooses to discontinue its research because it does not know that it will ultimately win the race. Most previous commentators on patent races have focused on overinvestment because they have assumed that participants have complete information about their competitors' progress. Under that assumption, a participant that encounters early difficulties in its research knows if its competitors have encountered similar difficulties, and therefore knows that it continues to have a chance to win the race. In fact, though, participants in patent races rarely have knowledge of their competitors' progress. I propose several possible approaches to dealing with this information problem.

Although I show in the article that the underinvestment problem could occur with different numbers of race participants and different likelihoods of research success, it is difficult to know how significant the problem is in practice. I am in the process of seeking the relevant information on how, specifically, these decisions are made in research enterprises, and this information may shed some light on that question. Tentative information suggests that the problem could be a significant one in the pharmaceutical industry.