Saturday, 27 February 2010

The price of ideas: competing incentives for innovation

[I intended my next post to be a followup to yesterday's, but I need a bit more time to work on that. So you can have this one instead - it's quite good too]

Only someone with Nathan Myhrvold's Microsoft money - and reputation as a mad genius - could get away with starting Intellectual Ventures and still being taken seriously. But since he can, it makes it an interesting concept.

This article by Ryan McClafferty raises some tricky questions about whether the company's attempts to create a marketplace in ideas will encourage, or instead stifle, innovation.

Normally if we deepen the market for a commodity, more of it will be produced. It gives producers more chances to sell, consumers more places to buy and lets price discovery be more accurate. As a side-effect it will usually improve the quality of the good - imagine that instead of a competitive dealer network for buying and selling cars, the government simply issued one to every family. It's very unlikely we'd have the quality of cars that we do now (and let's not go too far with that Toyota example).

This is Myhrvold's argument for Intellectual Ventures - by putting a price on ideas, it will correct a market failure which currently leads to them being under-produced. This is a respectable and mainstream economic argument: if a good has unpriced positive externalities, the economy will not produce as much of it as people would like to consume.

But some people think ideas are different. Simple production of new ideas is not enough to make them useful - they must also be used as widely as possible. And despite the Coase theorem, which says that assigning property rights to a good will let it be allocated with maximum efficiency, that might not be true of so-called non-rival goods like ideas or inventions.

The biggest reason is transaction costs. The main problem with software patents - from my point of view as owner of a software company - is not the fact that we might need to pay a licence fee. We already pay a licence if we resell someone else's database software or other components alongside our products. The problem is the search cost. We have no effective way of knowing whether our software infringes any of the millions of patents out there. If we had to go through them all and find out, we would swiftly run out of money before ever creating a useful product.

I guess it wouldn't be so bad if we could just go out there and make something, then worry about paying the licence fee afterwards if someone proves their patent claim to a court's satisfaction. But that wouldn't really work either - because we might have sold products for $100 which the patent owner wants to charge a $10,000 licence on. The court might just say no, the patent holder can never have more than a 20% share of revenue - but that makes it too easy for us to exploit the patent by simply setting a very low price.

A theoretical fix for this is to contractually construct a claim over buyers of our software saying that patent owners could come after them for retrospective payment. But guess how many we'd sell on that basis? The level of uncertainty for buyers would kill the market.

And that leads us to another problem with this market: uncertainty. Those who use one idea to build another - the potential customers in this new market - often have huge uncertainty about whether their work will lead to anything. And even if there is a price on ideas, they're also uncertain about whether they'll be able to capture the benefits of the invention for themselves. Who knows what competing products will emerge in the five years it takes me to bring a new invention to market? Who knows if anyone will even want it? This uncertainty also causes under-production (and under-use) of ideas, and stricter enforcement of patents could increase the cost of that uncertainty, resulting in less production and use of inventions.

I don't know which of those effects is greater - the underpricing or the uncertainty - but I do know that transaction costs are likely to dominate either of them, at least until there are better methods of searching for inventions and using them. Maybe IV can produce a search engine which will let me describe my potential invention in an abstract way and present me with a list of available ideas that I can use. That would both save me a lot of time in redeveloping the same thing twice, and take away the major transaction cost - search - which stops this market working properly.

To do that, they would need to be less of a patent owner and more of a patent exchange. Is that in their gameplan?

Is there an alternative mechanism that would work better? Open source adherents propose a mechanism which I used to think would never work - but having tried it out in the last couple of years, I'm coming round to it.

That is the idea of rewarding one non-rival good with another. In return for generating and publishing my ideas, I get acknowledgment, admiration, fame or social status. I used to think this was all very well for Linus Torvalds who has become world-famous for creating Linux - and as a result is paid quite well to advise companies, speak, write books and so on. But for people in niche markets, I didn't think it could work.

I don't quite agree any more. Even the little recognition, in small niches, that I've had from writing this blog, is both gratifying and profitable. A couple of comments on the BBC, articles in trade magazines and a few speeches have led to contracts for my business and a raising of my profile and status in some communities. It doesn't yet rationally justify the thousands of hours I've put in, but that will no doubt come.

If I had to try to earn money directly by selling my ideas, it would have been a lot harder to get a return - and certainly much harder to get the useful feedback I've had from many well-informed readers - than it is when I give them away for free.

A related aspect of this market is that I get to read the ideas of Tyler Cowen, Scott Sumner and Rajiv Sethi in return for contributing mine; and my readers get to see my ideas in return for writing comments or sending me emails occasionally.

Because these knowledge goods are non-rival, thousands of people can benefit from a piece of writing that only one person has to produce. I get much more benefit from reading the thoughts of other economists than the effort of writing my own stuff. In a naive sense my net benefit would be even greater if I just read theirs and didn't bother with my own, but actually the cognitive benefits I get from reading are greatly magnified by the effort of writing my own pieces.

Maybe innovators' preferences for recognition, status and feedback are far stronger than their preference for excludable rival goods like, say, hamburgers. Maybe the structure of production is such that recognition and feedback actually allow people to buy more hamburgers than if they sell the ideas directly. If so, eliminating intellectual property rights and reducing the transaction costs which put a brake on the free exchange of ideas would result in more production and use of ideas, not less.

I'm afraid I don't have a clear answer to this question - if only economists could be so decisive - but economics does at least provide us with the tools to analyse the issue.

So to resolve this debate, we'd need to measure the strengths of those competing incentives and understand the production function in a lot more detail. Anyone have any ideas how to do that? If you do, name your price and maybe I'll buy it from you. Or you could just publish it and get famous.

3 comments:

Marco Novarese said...

I share all of your claims and doubts. The same Economics is a field in which we face the problem.
New ideas and paradigms are constrained by the need to be recognized and accepted within the community. Economists need to know Economics, and not to be able to understand economy!
Working on something new is really hard, also because you have to face a big risk: even in case your work lead to some good result, it can be neglected and go ununderstood. Peer review makes it safer investing time in research within a specific paradigm.
Yet, as you wrote, working in a niche allows to have a specific recognition that for someone may ne relevant. Probably innovators are motivated at least in part by this kind of goal, but it may be not sufficient, if their work is neglected and so there is no real feedback and discussion.

Leigh Caldwell said...

Hi Marco

Thanks for posting - I have found your work on cognitive economics very interesting and I would like to talk more about it with you.

If, as you say, an innovator's work is neglected by their peers, do you think that the existing intellectual property framework would provide any better incentives for them to create?

It is certainly possible for (say) a small software company to produce work that pays their living without being widely recognised and quoted. But in this case, do intellectual property laws make any difference or is their work unlikely to be infringed anyway?

My instinct is that copyright laws in this situation still increase the incentives to produce work; but patent laws do not. What do you think?

M. said...

Hi Leigh,
I'd like to discuss my work (and yours too).

A small enterprise's incentive to create may be mainly intrinsic to its founders and they need to be protected by the law, as later (even years and years later) someone may finally understand and recognize their work.
Intellectual property laws could be relevant in order to stimulate enterprises to put on the market their new products.
I'm not an expert, but yes, just as an instinct, I could say that copyright could be more effective than patent laws.