Prolegomena to an evidence-based policy for software patents | Opensource.com
Prolegomena to an evidence-based policy for software patents
The Hargreaves Report issued last month has some thought-provoking things to say about the problems of software patents, lengthy copyright terms, and other intellectual property issues. It doesn't propose complete solutions, but its approach seems directionally sound. Officially titled "Digital Opportunity: A Review of Intellectual Property and Growth," the Report was commissioned by the British government and addresses the UK's IP system, but much of its discussion of the economics of intellectual property applies to the US.
The Report starts with the premise that in making decisions about IP, "evidence should drive policy." At first glance, this sounds boringly obvious. I was reminded of first hearing in the 1990s about the "evidence-based medicine" initiative, and wondering, what other kind of medicine would you want? But the fact is, our existing IP system has seldom, if ever, been driven by empirical analysis. Rather, it came into being piecemeal over hundreds of years through battles for markets and wealth. (And no doubt a few battles with purer motives, like claims of right and fairness, as described by my colleague Ruth Suehle.) Throughout this history, there have been theories to support policy decisions, but they have usually been more articles of faith than of science.
For example, a central premise of copyright and patent policy is that the exclusionary rights they confer are socially beneficial, because they create incentives for creativity and innovation. A corollary is that without the prospect of rewards from these exclusionary rights, authors and inventors would not author and invent. This theory sounds reasonable, and it's been repeated for so long it seems like a law of nature. But it hasn't been thoroughly tested. Remember, we're talking about legal doctrines that result in individuals and companies receiving monopoly rights, which can sometimes be enormously valuable to the recipients and costly to everyone else. What if we insisted on looking at the supporting economic evidence?
That's a big job. The Hargreaves Report notes that there are practical obstacles to basing policy on objective evidence, including the difficulty of collecting data, lack of data, and the influence of private interests (which it characterizes with the lovely/ugly coinage "lobbynomics"). But that doesn't mean that we have no information or that no conclusions are possible.
In discussing the repeated lengthening of the term for copyright, the Report is blunt: "Economic evidence is clear that the likely deadweight loss to the economy exceeds any additional incentivising effect which might result from the extension of copyright term beyond its present levels. This is doubly clear for retrospective extension to copy term, given the impossibility of incentivising the creation of already existing works, or works from artists already dead." Hear, hear.
Discussing patents, the Report is less direct. It quotes with approval submissions of Microsoft and Qualcomm to the effect that the patent system as a whole works quite well. Of course, this is debatable. (Full disclosure: I participated on behalf of Red Hat in a presentation to a working group for the Report, which presentation is noted in Annex B, but, I'm sorry to say, not quoted.) Fortunately, the Report recognizes that "some aspects of the way the system is currently working are a source of concern, because they appear to be causing barriers rather than incentives to innovation."
The Report says that "in some business sectors patent proliferation" is occurring, with "patents of uncertain validity represent[ing] a disproportionately high share of the increase in patent numbers." It also notes that the growth of patenting is "strongest in areas such as [information and communications technology], where patent scope and validity is inherently more uncertain than in other, less sequential technologies, such as pharmaceuticals." It describes the dangers of patent thickets, where a product "may well be covered by hundreds of patents owned by tens of rights holders," which can result in "a form of gridlock . . . in which firms underuse new knowledge because too many owners can block each other."
As for software patents, the Report finds that "the evidence that patenting supports innovation is weaker in computer technology and telecons than in other areas." Innovation in this area is "becoming more cumulative and collaborative in nature, building on previous inventions and innovations." In this area, "[t]he burdens of patent clearance and licensing, and the possibility of litigation associated with others' computer program patents often discourage innovation."
This is strong language. If correct, it's hard to see how to justify software patents. The Report does not follow its argument all the way to this conclusion, though. This may be in part because in the UK, unlike in the US, the patent office is not supposed to grant patents on "non-technical computer programs" -- a category that continues to elude a clear definition. In any case, the Report recommends only that patents not be extended into areas "they do not currently cover without clear evidence of benefit."
I wish the Report had followed its own logic further, but at least it's a step in the right direction. Its emphasis on evidence-based IP policy is timely and valuable.