One of the best empirical economists working on patents is Petra Moser of Stanford. She has an extraordinary ability to find unusual data that answers really hard questions. Her most recent paper Compulsory Licensing
with Allessandra Voena does just that.
There is - a frankly rather ridiculous view, one held only by lawyers for whom the distinction between 10 ^ -6 and 10 ^ +6 is invisible - that somehow compulsory licensing in a country will reduce innovation because it will be every so much more attractive to "steal" all those great foreign inventions. In this view - which crops up in these pages occasionally in anecdotal form - the great thing about patents is that it increases innovation by forcing people to "invent around" patents.
The "Compulsory Licensing" paper attacks this question directly - looking at an episode where the US "stole" a bunch of inventions by compulsory licensing after World War I. The consequent effect on innovation in the areas covered by the licenses? It went up by 20%.
In my view debunking the rather silly view that compulsory licensing would diminish innovation is not this papers main contribution. We recognize that patents have two effects (ignoring "invent around" and "revealing secrets" both of which are of at best minuscule significance): increasing innovation by increasing incentives to innovate, and decreasing innovation by making it more costly to innovate. This paper gives us a pretty clean measure of the latter effect: the benefit of being able to access existing ideas without negotiation or licensing raises innovation by around 20%.
The idea that forcing people to "invent around" patents leads to innovation is a classic broken window fallacy. Every hour you have to spend reinventing the wheel is an hour you can't spend inventing something new. At best, this has the effect of channeling present inventors' effort into duplicating past inventors' efforts, which is hardly beneficial when they could be working on unsolved problems instead.
Oh, that one is going to have the Patent Maximalists freaking in the streets.
>> the benefit of being able to access existing ideas without negotiation or licensing raises innovation by around 20%.
The effect should be that much more pronounced when we consider inventions in areas where there would be potentially very many more participating inventors if licensing were not an issue. This is the case in an area of development where a relatively large fraction (in comparison to the examples from the study) of the potential invention contributors have little negotiation leverage. I think this high percentage of vulnerable inventors would exist when we consider areas with a low bar to entry. An example would be open source software development. There is a very low bar to entry in this field. The "laboratory" and the "industrial plant" is the PC. The "distribution channel" is the Internet.
An interesting study would be analyzing how general were the patents from the study. If we approximate no patenting as patenting extremely narrowly. Then it might follow from that 20% figure that, the broader the patenting, the greater the hit on innovation.
It would also be interesting to consider how much more significant the effect might have been if the examples covered in the paper had been more influenced by the Internet. The Internet promotes collaboration, in some cases, significantly.
And I would not do this topic justice IMO if I fail to say that many many people are capable of inventing many things as the occasion requires but could not in 10 lifetimes actually sit down and list out all the things they were capable of inventing. The implication is that, in a patent system, if you aren't the first to list out some particular "invention" within your capabilities, you can potentially be forever barred in that area by the patent holder that beat you to the punch. This means even if an invention is within the capabilities of 1000 contemporaries (all of whom might eventually move into this area as clients demand it), one of the 1000 can get a monopoly on the invention if the other 999 were distracted in other areas of development (among the large set of things each might be capable of inventing or large set of things that might otherwise distract someone) or if the 999 were keeping implementations private instead of rushing out the patent application claims description. How reasonable is it that this one be given power over the other 999?
That 20% (or possibly much higher number) represents opportunity costs to society.
Short-term (ie, soon), it will hopefully be recognized by the courts (eg, by the SCOTUS in Bilski) that we are almost surely denying society (and individuals at large) whenever we let patents interfere with inventing that is accessible to ordinary people. [Eg, open source software, business methods, and a list that grows as technology allows individuals to become more efficient or creates new areas of development.]
Long-term, perhaps patenting as a whole will be downsized to the cases where it can first be proven that such a monopoly (for X years for that particular patent) would in fact promote the progress of science and useful arts. There are many alternatives to monopolies that the government could leverage in order to help society (eg, giving tax credits and monetary grants). Monopolies should not be doled out without a lot of care unless the monopoly has very narrow scope and/or falls within a very well-defined set of bounds.
Speaking of Bilski, what the heck is going on with it? When will the SCOTUS render its decision? Inquiring minds want to know.
Petra Moser wrote some other interesting papers on patent history, which can be found with a search, and I think accessed at her homepage. They are cited in Against Intellectual Monopoly.
And good call by Jesse that the inventing around a patent rationalization exemplifies the broken window fallacy. Our man Bastiat lives again!