Should We Have Patents?

The ways of funding innovation

We have two principle methods of subsidizing innovation: patents and prizes. We can either grant an inventor the exclusive rights to market an invention for a period of time, or we can give them either a one time prize, or a subsidy to future units sold. We subsidize this research because, in the absence of government intervention, people will not fully capture the positive spillovers onto others, with the social value of invention being around three times greater than the private value. (See Bloom, Schankerman and Van Reenen (2013), and Jones and Williams (1998)).

These two ways have advantages and disadvantages. The patent has the advantage of reducing the amount of information which the government has to know. In order for prizes to be worthwhile, we must be able to specify what we want in advance, and we must know the value of the invention to society. Patents subject it to a market test – it is valuable only to the extent that people are willing to pay for it. However, it creates profits only to the extent that the monopolist is able to restrict production. In cases where a technology is built on top of other technological developments, the distortion is multiplied for each patent. In addition, so long as there is unextracted consumer surplus, there will be inventions that would make the world better to create, but are not made. It is not possible for patents to maximize consumer welfare, so long they are unable to perfectly price discriminate. In addition, as noted in BSVR (2013), firms which work in the same technological space don’t necessarily compete in the same product space, and vice versa. Two firms might find a cure for the same disease with very different methods, making their patent no monopoly at all. Patents are also biased against preventative medicine, both because it will reduce market demand for the product by making it harder to be transmitted, and because only treating the sick allows for price discrimination.

These two ways can overlap, of course. With public procurement of medicine, getting FDA approval (which allows you to sell the good) is not dissimilar to a prize, and we can combine patents and prizes. Michael Kremer proposes requiring patents to be put up for auction, and have the government some percentage of the time buy it and put it in the public domain, a proposal which I support. Both systems can and should mandate the design of the invention being made public in a way that is useful for other companies to understand – Hegde, Herkenhoff, and Zhu (2022) exploit a 1999 law which accelerated the publication of patents to find that firms more exposed to the disclosed patents increased R&D and follow on innovation.

While the theoretical case is obvious, the empirical case for patents is not particularly clear. In lots of industries, firms don’t even patent their inventions. Instead, they keep them secret. What they do might not be so much a discrete idea or invention, but a system of organization, which cannot plausibly be patent protected. Econometric tests of changes in patent laws do not seem to have had much effect on patenting – for example, Sakakibara and Branstetter (2001) find that a 1988 Japanese law which expanded the scope of patents seemed to have no effect whatsoever on patenting rates or R&D. I do not necessarily put all that much stock in the econometric tests, as they are old, the data is shaky, they deal with time series, and they don’t generally cover the level of variation which would correspond to going from patents to no patents at all – but it is still extremely important to note that there is simply no good evidence of patents increasing the rate of innovation. Instead, what we do have are studies relating profits in the pharmaceutical industry to research spending and medical discoveries. Patents increase profits, and so the case for patents increasing innovation in pharmaceuticals is rather soundly established.

Patents also protect current inventors, at the cost of hindering inventions later on. As I mentioned obliquely before, the losses from a system of patents will increase the more technologies something else relies on. Take a smartphone – it is not patented in its whole, but is rather an accumulation of thousands of patents for different processes and features, many of which the smartphone makers license. Each individual patent is a block point for future innovation, which is why patent trolls (companies which acquire vaguely worded patents in order to sue actually productive companies for “patent infringement”) can be harmful at all. In contrast to the null results of the effects of patents, there is a very clear literature that allowing licensing increases innovation. Galasso and Schankerman (2013) use the revocation of patent rights in the US Court of Appeal to find that this substantially increases later-on citations to that patent, indicating that people are innovating in ways that build off of the old patent. (The paper uses a judge IV – there are randomly assigned judges to cases, and those judges have different innate likelihoods to overrule on any given case). Heidi Williams (2013) has a really cool study with evidence from the Human Genome Project. Both a private company, Celera, and the HGP sequenced genes, with the HGP concluding in 2003. Celera came first, and its genes were protected by intellectual property laws, while any genes the HGP sequences (including the ones sequenced first by Celera) would be put into the public domain. Celera’s property rights substantially reduced innovation around the genes they sequenced. And in an historical example, the United States essentially seized all the patents of German companies in 1917. Moser and Voena (2012) find this increased domestic innovation by 20 percent.

So this brings me to the central point which I have been circling around for this whole article. There is one industry for which everyone agrees that patents work – pharmaceuticals. That is also the one industry which everyone agrees prizes would also work. So why have patents at all?

Patenting is so ubiquitous that I do not think we have actually had adequate opportunity to know what the world would be like without them. The theoretical case for some subsidization of invention is so strong that the lack of evidence is insufficient to sway me entirely from patents mattering. It is a bit like jumping out of an airplane without a parachute – no RCT on the matter has ever been conducted, but we are so certain that parachutes do matter that it would be considered madness to ever actually experiment by jumping without them. (The only prior work was marred by limited external validity).

Patenting also has the advantage of obscuring the costs and benefits to the taxpayer. It is unfortunate that we must speak of the voter as an idiotic child who must be protected from the consequences of their own decisions – if it makes you feel any better, I argued recently that voted preferences are not in any meaningful way real. We are shocked by large dollar figures attached to prizes for pharmaceuticals, even though sums in the billions of dollars would be eminently fair. Patents protect against penny pinching by the public.

Still, I am confident that there is a lot of latitude for actors both private and public to stimulate innovation by offering prizes, targeting diseases that plague the poorest parts of the world. Just recently, Bill Gates is giving a prize of $1 million for new work in Alzheimer’s disease. I think that, once someone’s fortune becomes sufficiently large, the most effective way for them to improve the world is not to fund the distribution of the medicine we have now, but to subsidize the finding of new cures.

What this all really convinces me of is that direct government funding of research and training is irreplaceable. Trying to subsidize innovation by rewarding the inventions themselves is like hammering on pudding. We should focus on equipping people with skills and interests that lead to innovations we could not possibly specify in advance. The present administration attacking the universities is something worth freaking out about – it’s only our future, after all.