The Effects of Law Enforcement on Innovation

By Brian Silverman*

The session titled “Effects of Law Enforcement” included presentations of three fascinating papers, all related to the impact of legal or regulatory institutions on innovation. The session was held on Friday morning in a well-appointed classroom at the architecturally striking Stockholm School of Economics.

Albert Galasso (University of Toronto) led off the session with “When does product liability risk chill innovation? Evidence from medical implants” (co-authored with Hong Luo, Harvard Business School). Although there have been numerous studies of the impact of various legal/regulatory shocks on innovation, none of which I am aware has explored a potential link between product liability law and innovation. Alberto presented a model that demonstrates how such liability at one stage of production can actually affect innovation at subsequent stages. Empirically, Alberto’s study exploited a relevant shock: a wave of lawsuits in the early 1990s that targeted large producers of generic polymers, such as Dow Chemical company, for health problems stemming from implantable medical devices that used such polymers as an input. (Given that the medical device producers were small firms that went bankrupt, Dow etc. were the obvious deep pockets to pursue.) Given the cost of fighting these cases, Dow and all other major U.S. polymer producers immediately ceased selling the product to makers of implantable medical devices…but not to the makers of non-implantable medical devices, which carry far less product risk.

As economics-minded readers will already have anticipated, this shock provides an excellent setting for diff-in-diff analysis, comparing subsequent innovation rates (as measured by patents) for implantable vs. non-implantable medical devices. The results offered stark evidence that innovation in implantable medical devices, as measured by patents, was sharply suppressed by the increased product liability risk faced by upstream suppliers. Further analysis indicated that this was particularly true for domestic (U.S.) medical device companies, while foreign firms were less affected. To the extent that this reflects a greater willingness of foreign polymer producers to continue supplying foreign device manufacturers on the assumption that they were one step further removed from being dragged into U.S. litigation, this is additional support in favor of the Galasso-Luo-proposed mechanism. Of particular note, a conventional analysis of the impact of the increased liability risk on Dow’s innovation would find no impact, given the general-purpose nature of this plastic, thus missing the statistically and economically important suppression of innovation in a key downstream sector.

Bernhard Ganglmair ((ZEW, MaCCI, University of Mannheim) followed with “Visibility of technology and cumulative innovation: Evidence from trade secrets laws” (co-authored with Imke Reimers, Northeastern University). Whereas there exists a huge literature on the use and efficacy of patent protection, far less is known about alternative knowledge protection mechanisms such as trade secrets. Bernhard laid out a theoretical model demonstrating that variation in the strength of trade secret protection should influence the use of such a mechanism (as a substitute for patenting), but primarily for innovations that are not easily visible or otherwise susceptible to reverse-engineering – for example, for process innovations rather than for product innovations. Bernhard then presented empirical evidence exploiting variation in the timing of U.S. states’ adoption of the Uniform Trade Secrets Act, and variation in the extent to which this adoption changed the level of protection offered in a given state, to explore the impact of increased protection of trade secrets on the level of patenting. (One cannot directly observe the level of trade secrecy because it is, well, secret.) His results demonstrate that increased protection of trade secrets indeed reduces patenting, particularly for process patents, implying that firms substitute trade secrecy for codified IP protection. As a result, stronger trade secrecy encourages initial innovation by enhancing appropriability, but reduces follow-on innovation by restricting diffusion of information about the innovation. Using a sophisticated structural model, he also estimated welfare effects of stronger trade secrecy protection, finding plausible effects for a range of industry settings.

Finally, Stephen Salant (University of Maryland; University of Michigan) presented “Reducing drug prices without depressing innovation” (co-authored with William J. Adams (University of Michigan). Stephen began with the well-known empirical pattern that drug prices are substantially higher in the U.S. than in most other countries. He then described the common academic explanations for this, and noted that each fails to explain at least one stylized fact associated with this pattern. He then laid out a new model that accounts for these stylized facts, and used this model to assess four policy proposals to reduce U.S. drug prices. In particular, the Salant-Adams model indicates a policy of lowering entry barriers in the downstream (retail) channel would impose the least damage to innovation incentives while reducing drug prices. Stephen then noted additional government policies – related to targeted subsidization of innovation early in the innovation process – that might enable the U.S. to maintain existing innovation incentives while reducing drug prices, at the least cost to public coffers.

Each presentation was followed by a lively question-and-answer period. A good time was had by all, and then we adjourned to a sumptuous coffee break in the School’s lovely atrium.

* For the benefit of scholars who could not attend the conference or who attended different sessions, we publicize summaries of several sessions held at the SIOE 2019 meetings at the Stockholm School of Economics. The full overview of presented papers is here.