Improving science with markets

In case anyone hasn't seen it, here is Robin Hanson's classic paper on Idea Futures, a way of using markets to improve scientific inquiry and accuracy. In a similar but less ambitious vein, Kieran Barry writes about having editors, authors, or publishers of scientific papers put up financial rewards for errors. Even a small incentive would make people much more likely to send in errors. Grad students work cheap - and they'd accumulate some academic reputation.

Robin Hanson has a lot of clever ideas about using markets. In this paper, he points out that current medical systems have poor incentives. Your doctor sells health-care, not health, hence may recommend expensive but useless tests. If you simply bought health insurance *and* life insurance from the same company, the incentives become aligned. Your life insurance company wants you to live as long as possible (they get more premiums), hence they have incentive to spend money on your health efficiently.

Share this

Thanks for giving this idea

Thanks for giving this idea some airplay. It's good for my ego.

Of course, it would be even better for my ego if you could spell my name correctly.


Isn't this all too clever by

Isn't this all too clever by half? If you simply stopped public funding of science wouldn't the functional equivalent of Idea Futures spontaneously arise quite naturally in a free market?

For it to spontaneously

For it to spontaneously arise, someone has to have the idea. Here is the idea.

Of course, the hard part is stopping the public funding of science.

"For it to spontaneously

"For it to spontaneously arise, someone has to have the idea. Here is the idea."

I'm not convinced the idea is needed at all. Wouldn't existing markets, especially the stock market, largely do this if science were privatized? Scientists would really have no choice but to "stake their reputations" in a free market.

Stanz said it Venkman: "You don't know what it's like in the private sector: They expect results."

"Of course, the hard part is stopping the public funding of science."

Heh, ya.

Barry has an interesting

Barry has an interesting idea. The concern that comes to mind immediately, however, is that it seems that publishing high-impact work could become a very expensive process, both in terms of the money it would require to bond a paper, and the time required to respond to the deluge of criticisms.

Case in point--I just had a paper accepted in a moderately well-read journal. Being realistic, I would imagine that maybe 20 or 30 people worldwide will read it in depth in the first few months after its publication, and another few hundred hundred to a thousand will probably scan over it. It's a decent paper, it just happens to be in a relatively specialized and small field.

By contrast, I have colleagues who work with relatively famous PIs in HUGE fields. Their papers are much more widely read, simply because so many people study the same things. It seems that they would be facing many more challenges to their papers--not only because more potential challengers are reading their work, but presumably "killing" the work of a famous (relatively speaking, of course) scientist would be a significant accomplishment in and of itself.

That's what this grad student would try to do, anyway. :smile:

Brian - Note that in your

Brian - Note that in your analysis, the papers with the most cost for dealing with comments are the papers with the most worth (the most widely read). This suggests that it may not be a problem at all. If there is any link between "widely read" and "resources available", then those papers with the most criticisms to answer will have the most time to do so.

Alternatively, popular papers could only examine criticisms from people who have already established a reputation for accuracy. If a non-reputed person found an error, they would have to convince a reputed person to submit it - hence distributing the effort. Another mechanism would be to require a small payment with each criticism - to be refunded if the error was genuine. This would discourage frivolous errors.

I think there are plenty of solutions, if we just think about aligning incentives :grin:.