The alternative of trade secrets protection lowers overall innovation and research in genetics

In contrast with what one would infer from a naïve comparison of follow-on innovation on patented and non-patented genes, both of our quasi-experimental approaches suggest that gene patents have not had quantitatively important effects on either follow-on scientific research or follow-on commercial investments. The estimates from our first quasi-experimental approach—comparing follow-on innovation across genes claimed in successful versus unsuccessful patent applications— document estimates which are economically small and meaningfully precise. While the estimates from our second quasi-experimental approach—using the leniency of the assigned patent examiner as an instrument for which patent applications are granted patents—are less precise, the fact that these two approaches generate similar conclusions provides additional confidence in our estimates.

When interpreting these empirical results through the lens of the well-developed theoretical literature on cumulative innovation (e.g., Green and Scotchmer 1995), several points are relevant to highlight. First, as best we can measure, very little of the follow-on innovation in this market is done by the gene patent holders, suggesting that follow-on innovation in this market will often require licensing agreements. Second, essentially all (96 percent) of the patent applications in our sample are assigned to for-profit firms, and in the vast majority of clinical trials (86 percent) both the patent assignee and the follow-on innovator are for-profit entities. Taken together, these facts suggest that cross-firm licensing contracts in this market seem to have operated at least somewhat efficiently. Consistent with this interpretation, there has been relatively little patent litigation observed in this market.

Interpreting these empirical results in the context of existing empirical evidence, perhaps most closely related to this paper is work by one of the authors (Williams 2013), who found that a non-patent form of database protection held by the private firm Celera on their version of the sequenced human genome was associated with large declines in follow-on scientific research and commercial product development, on the order of 30 percent. Why did Celera’s intellectual property cause declines in follow-on innovation, while gene patents did not? Theoretical models tend to analyze stylized characterizations of intellectual property rights in which Celera’s intellectual property could reasonably be seen as practically identical to patent protection. However, Celera’s intellectual property differed from the gene patents we analyze in one key dimension. The sequenced genetic data in both the accepted and the rejected patent applications we analyze were disclosed in a way that enabled open access to the data for all prospective follow-on users. In contrast, Celera’s data were disclosed in a much more restrictive way. We argue that this difference in disclosure may plausibly account for the different observed effects of Celera’s intellectual property and gene patents on follow-on innovation outcomes. Importantly, this disclosure interpretation reinforces one important reason why our estimates may not generalize beyond the case of human gene patents: as we discuss in more detail in Section V, the USPTO’s specific (and more stringent) requirements for the disclosure of sequenced genetic data may have made the disclosure function particularly effective in our context.

From a policy perspective, one reason this disclosure feature of the patent system may be important is that it highlights a potential unintended consequence of a recent set of high-profile legal rulings on the case Association for Molecular Pathology v. Myriad Genetics.2 The firm Myriad Genetics was granted patents on human genes correlated with risks of breast and ovarian cancer, and in June 2013 the US Supreme Court unanimously ruled to invalidate a subset of Myriad’s gene patent claims, arguing that such patents “would ‘tie up’ … [genes] and … inhibit future innovation…” (p. 11). That is, the Court argued that gene patents had sufficiently strong negative effects on follow-on innovation that genes should be ineligible for patent protection. While, consistent with the Court’s view, there has been widespread concern that patents on human genes may hinder follow-on innovation, as argued by a 2006 National Academies report and by Caulfield et al. (2013) there was no empirical evidence available to either support or refute that assertion prior to this paper. From the disclosure perspective discussed above, underlying this ruling seems to be an assumption that if genes are not patented, they would be placed in the public domain. But at least in the case of Celera, when Celera’s patent applications were rejected the firm instead chose to rely on an alternative form of intellectual property which, taken at face value, resulted in both lower private returns and lower social returns to Celera’s research investments. Stated differently, our analysis corresponds to a partial equilibrium analysis which holds disclosure constant, but in general equilibrium firms may choose to protect their inventions with, e.g., trade secrecy or other strategies rather than with patents if patent protection is not available. This type of behavioral response would be critical to assessing the potential welfare consequences of the US Supreme Court’s ruling not just for human genes but also for other technologies which have recently been declared to be unpatentable.3