Friday, March 13, 2015

Are Patent Fees Effective at Weeding out Low-quality Patents?

GaƩtan de Rassenfosse & Adam B. Jaffe have posted Are Patent Fees Effective at Weeding out Low-quality Patents? (sorry, behind NBER paywall, though many universities have subscriptions). The abstract is here:
The paper investigates whether patent fees are an effective mechanism to deter the filing of low-quality patent applications. The study analyzes the effect of the Patent Law Amendment Act of 1982, which resulted in a substantial increase in patenting fees at the U.S. Patent and Trademark Office, on patent quality. Results from a series of difference-in-differences regressions suggest that the increase in fees led to a weeding out of low-quality patents. About 16–17 per cent of patents in the lowest quality decile were filtered out. The figure reaches 24–30 per cent for patents in the lowest quality quintile. However, the fee elasticity of quality decreased with the size of the patent portfolio held by applicants. The study is relevant to concerns about declines in patent quality and the financial vulnerability of patent offices.
Others have suggested that patent fees be used to weed out patents, for example to weed out low value or to deter patent trolls (who tend to assert later). This paper seems to support that notion, but I'm not sure it gets us all the way there, as I discuss after the jump.

The study is a differences-in-differences study, which looks a the relative changes between two groups after a shock. Here, the two groups were foreign and domestic patents. The shock was a 1982 increase in fees, which increased fees more for US applicants than foreign applicants. Thus, the goal is to see whether the higher fees for US companies changed the types of patents coming out.

So far, so good. But how do we test patent quality? The authors follow Lanjouw & Schankerman: "We use four such indicators: the number of citations received by the patent (Y1); the number of claims at grant (Y2); the size of the patent family (Y3); and the number of times the patent was renewed (Y4)." Of these, citations received and number of claims have been shown in multiple studies to have little bearing on patent validity. However, these quality measures are about the economic/technological value of the patent, and not about the validity of the patent (except that invalid patents often wind up with fewer renewals). What this means is that this study doesn't really tell us that raising costs might get us more valid patents, an important aspect of arguments that raising fees will reduce frivolous assertions.

One interesting finding, though not surprising, is that the effect is lessened on large portfolios. This means that larger companies with money will continue to file more patents that this study considers lower quality than small companies. This does not give me comfort. To me, it just shows that high fees are just another way that those with resources can squeeze out those without them. Indeed, this is exactly what other studies have shown - that to the extent that patents serve as a barrier to entry, it is primarily by big companies against little companies. Those who think raising fees will somehow help startups are -- in my view -- engaging in wishful thinking.