What can the human genome project teach us about intellectual property policy?

By Kyla Fisher, M.A. Economics, Queen’s University

Innovation is one of the primary drivers of economic growth and improvements in living standards. It often produces larger social benefits than private benefits, leading firms to under-invest in R&D compared to the socially-optimal level. One of the ways that the government works to overcome this gap is through offering intellectual property (IP) protections, giving firms a temporary monopoly on commercializing their ideas. In addition, many governments allocate significant funds directly towards research through public research institutions or universities. However, it is difficult to determine the impact of these public efforts to stimulate innovation as we are unable to know the counterfactual. This article reviews the findings from an innovative study by Heidi Williams (2013) on the use of IP during the sequencing of the human genome. The study exploits the discrete nature of gene sequencing and the fact that it was researched both publicly and privately to evaluate the impact of IP on innovation outcomes. Despite the importance of IP policy for technological innovation there are relatively few empirical studies in this area. For this reason, Williams’ study generated quite a bit of interest at the time of publication and has been cited in multiple U.S. Supreme Court briefings.

One of the potential social costs of IP protections is that it may hinder follow-on research and innovations, reducing and/or delaying technological change. This idea has influenced several recent US Supreme Court decisions to restrict the discoveries eligible for IP protection (Williams, 2016). To try and evaluate the impact of IP on subsequent innovation, Williams took the novel approach of analyzing data from the human genome project. This project has some features that make it a useful natural experiment for empirical analysis of this issue. The human genome was sequenced concurrently by both a private firm, Celera, and a public research group, the Human Genome Project. Parts of the genome were first sequenced by Celera and this information was then placed under a form of IP protection. Public researchers later sequenced this part of the genome as well. However, there remained a brief period from 2001 to 2003 when information on the sequencing of certain parts of the genome was held privately and was under IP protection. Genome sequencing is an input into research on the links between genes and physical traits/diseases. Once these links are established this information can be used to develop genetic diagnostic tests, which have commercial value. By comparing the amount of research on parts of the genome sequenced publicly (“public genes”) vs. privately (“Celera genes”), Williams is able to estimate the impact of IP protection.

Williams found that in 2009, over five years after the full genome sequencing became public, there was still a significant difference between the number of research papers published on genes that were first sequenced privately (1.2 on average per “Celera gene”) compared to publicly (2.1 on average per “public gene”). In addition, there was a difference in which genes were used in diagnostic tests, with 3% of Celera genes used in a diagnostic test, compared to 5.4% of public genes. These comparisons were based on genes sequenced in the same year, suggesting that research and genetic testing innovation was in fact hindered for Celera genes. Williams does find some evidence of a selection bias in the publicly sequenced genes, which she finds accounts for roughly one quarter to one half of this difference. When taking this into account she estimates that for genes with the same research potential the IP held on the Celera genes caused a roughly 20-30% reduction in the number of diagnostic tests produced from those genes by 2009. Her estimate of the effect on the number of papers published was similar. Williams also finds that within the group of Celera genes there was less research on the genes during the period they were under IP protection. She estimated this effect as a reduction of 0.11 publication per year during the years the Celera genes were under IP protection, a reduction of 45% from the mean of 0.244 publication per year for those genes. There is a small “catch up” increase in papers following the end of IP protection but that it is not enough to make up for the lost research during that time. The end result is a decrease in the overall literature about the Celera genes.

Williams suggests several potential reasons for this reduction in research output and product development. It is important to note that the IP protection used by Celera was not a patent but a different contract-law type of protection that allowed Celera to sell the data for private research and required the negotiation of licensing agreements with Celera for any resulting commercial discoveries. The IP allowed free use by academic researchers but, based on interviews with researchers, Williams says there was uncertainty around what could be done with the data. In addition she notes that the separation between academic and commercial research is not always well-defined, which could cause further reluctance to use the data. In a perfect contracting environment we might expect that other biotech firms would be able to negotiate with Celera to get a licensing agreement and proceed with researching the genes as they might normally. Williams suggests that firms were deterred from this route despite the commercial potential as it put them in a poor bargaining position. Firms had to negotiate ex-post, when they had already made a significant research investment to develop the idea for a genetic diagnostic test. This meant that there was an increased cost to other biotech firms of using the Celera genes. The fact that the difference in the level of research related to Celera genes persisted after IP protections were removed suggests some continued untapped commercial potential. Williams does not provide an explanation for this persistence, other than to note that public research often underpins research on commercial applications.

What are the implications of this study? One potential implication is that we should support public research in cases where there is a significant potential for follow-on research, such as the human genome project. In this case, we saw that genes first sequenced by Celera were less researched and have been used in fewer diagnostic tests even years after the IP protection has ended. This has potential health consequences for the public. There may be similar areas where public research programs make sense in order to get the information out to all firms as quickly as possible. In addition, Williams notes that both the uncertainty around the type of IP and the fact that negotiations for any commercial applications would occur ex-post were deterrents for other firms to work on Celera genes. This suggests that it is important to ensure IP protections are clearly defined and to consider any advantage that is being given to the IP holder in commercial negotiations. A more recent study by Sampat and Williams (2015) also supports the importance of IP design, as the authors find that standard patents on genes do not actually inhibit follow-on innovation. Interestingly, it seems that what may have made the difference in this case is the unusual nature of the IP protection involved. Due to the limited amount of work in this field there are still many remaining unknowns. However, given the importance of technological innovation for our society it will be worth following future research developments.



Sampat, B., & Williams, H. (2015). How Do Patents Affect Follow-On Innovation? Evidence from the Human Genome. NBER Working Paper No. 21666.

Williams, H. (2013). Intellectual Property Rights and Innovation: Evidence from the Human Genome. Journal of Political Economy, 121(1), 1-27.

Williams, H. (2016). Intellectual Property Rights and Innovation: Evidence from Health Care Markets. Innovation Policy and the Economy, 16, 53-87.