David Leonhardt writes in the NYTimes today about another way to increase innovation offer prizes (link here). He takes off from Netflix' search for better algorithms to suggest movies to subscribers in which they would likely be interested, based on what they have already seen and rated highly. The prize is a million dollars, and there is a threshold to beat, a ten percent improvement over Netflix' present algorithm. The competition lasts five years and seems very risky to enter, but has attracted widespread participation.

Prize competitions were apparently common centuries ago; Leonhardt cites one example, the British Parliament's £20,000 prize for a way to determine longitude and thus a ship's position at sea.

Leonhardt argues that prizes are efficient in that they only pay for success, the cost is fixed in advance, and the competition is open to everyone. This contrasts with grants, an alternative way to get innovation, but one which is essential in the sciences where the setup costs are so great.

I would suggest that the lesson is, if you want innovation, create competition. The contrast is with patents and copyrights, where the payoffs are undefined, the costs to consumers are huge, and they are increasingly used to prevent competition; in the process, they end up slowing innovation.

The longitude prize does show that prizes are not completely unproblematic. The Board of Longitude, which was set up to award the prize, did their best to cheat the winner, John Harrison, out of the prize.

Good lord, I get the idea about this blog, of course, but you really lose credibility with inanity like this. See if this makes sense to you: There is competition for patents! All the moreso if they are as restrictive of competition as the bloggers on this site insist. You pretend like there are no incentive (or other) positive effects from patents, but that's just absurd. It may be that patents offer too much compensation, and it may be that there are institutional problems, but to pretend like there isn't competition for patents is both to contradict yourself and to deny reality. To pretend like there are no problems with the Pigouvian-tax-like preset prize versus the uncertain patent is silly. As I said--there may be good arguments here, but your rhetoric (and that of your co-bloggers) is so over-the-top that your valid arguments are hard to take seriously.

I'm not quite sure what this post is a response to. My "rhetoric" pointing out that there are problems with prizes as well as with patents? John's rhetoric in the original post? Rereading it doesn't seem especially strident to me. Has somebody here denied that there is competition for patents? Is it unclear that patents once awarded decrease competition - nobody can use the patented idea without the permission of the monopolist who owns the idea? Does someone here argue that there are no incentive effects from patents? here it is necessary to distinguish between individual patents and patent system. Obviously everything else equal, rewarding someone with a monopoly for inventing something gives more incentive to invent. Less obvious perhaps, is the fact that a patent system also discourages innovation by making it more difficult to to create new ideas (need to license existing ideas) and also by distorting incentives (me-too drugs). The striking empirical fact is that the evidence suggests that as far as encouraging innovation goes patents on the net do nothing: neither increase nor decrease innovation.

The reponse is to John. And you don't think this implies quite strongly that patents and competition are anathema?:

I would suggest that the lesson is, if you want innovation, create competition. The contrast is with patents and copyrights, where the payoffs are undefined, the costs to consumers are huge, and they are increasingly used to prevent competition; in the process, they end up slowing innovation.

Notice the juxtaposition of the phrase "if you want innovation, create competition" and the phrase immediately following, "The contrast is with patents and copyrights."

I get the "less obvious" point (and it's actually pretty obvious) about sequential innovation, but this post does nothing to advance that line of thought. The claim here is that prizes work and patents don't, but as your own comment points out, things are hardly this clear-cut. And along this very dimension--the dimension of fostering sequential innovation--it's not at all clear why prizes should be at all superior. It's a world of second bests. As my comment hints, if you get the prize value wrong (and I have no idea why, short of using some market mechanism, you think you'd get it right), you don't actually end up encouraging more innovation (or you encourage too much). What is the basis for the claim that these costs are smaller than the costs of the current patent system, imperfect as it may be?

Finally, your "striking empirical fact" is also less than clear-cut. I am familiar with the research to which you are allduing. Here's my question: Where is the paper that sets out the baseline--the level of innovation we would have absent patents? I've seen some historical analysis which is ambiguous, at best. I've seen no other work that actually compares the but-for world with our world in any persuasive way. And it's not surprising--I can't imagine how you'd go about collecting this data. So what, exactly, do you think the empirical results show so conclusively?

Well I should probably let John speak for himself. I'm not quite sure what he meant by "payoffs are undefined," but I take the operative statement to be "they [patents] are increasingly used to prevent competition." I agree with this - I think that there is a lot of evidence that patents are used to prevent competition, not in the trivial sense that you can't use the patented idea, but in the sense that there is an effort to bottleneck key technologies; prevent rivals from innovating and getting ahead of you and so forth. Whether that is "increasing" is hard to say, since patents have always been used this way - but they are certainly much easier to get and enforce in court.

On the evidence part - I meant "striking empirical fact" not in the conclusive sense, but in the inconclusive sense that there have been many studies - most by patent proponents - and they haven't been able to find evidence that patents have any important effect in increasing innovation. The points of comparison are generally between countries that have patents versus those that don't; Lerner - who is no enemy of patents - in his 150 years of evidence paper doesn't find any evidence over 150 years; Petra Moser doesn't find any evidence from her 19th century studies. In the 8th chapter of our book MIchele and review all the studies we have been able to find in an effort to see what evidence there might be that patents increase innovation - that would be a place to get references. If there are studies we have missed we'd obviously appreciate a pointer.