Marginal Revolution: Five open questions about prediction markets

December
S	M	T	W	Th	F	St
	1	2	3	4	5	6
7	8	9	10	11	12	13
14	15	16	17	18	19	20
21	22	23	24	25	26	27
28	29	30	31

« Prudie blows it again | Main | Dark Matter »

Five open questions about prediction markets

Here is the new version of the paper, for $5. Here are older versions of the paper. Here is a summary of the five questions. Of these five, I worry most about getting uninformed traders to participate in a game with zero-sum (at best, assuming no trading costs) pecuniary returns. I doubt if idea futures will be the sexiest form of gambling for most people. Insofar as traders look for fun, only the celebrity-obsessed and current events-obsessed will find idea futures more attractive than Las Vegas or Baden-Baden. (That is why contracts on politicians and Michael Jackson have been so popular.) If you are simply looking to earn money for the longer run, Wall Street, with its positive expected rates of return, offers better odds. That being said, as the fixed costs of market creation fall, a relatively small number of traders may suffice to keep most of the important idea futures going.

Posted by Tyler Cowen on February 21, 2006 at 07:50 AM in Economics | Permalink

TrackBack

TrackBack URL for this entry:
http://www.typepad.com/t/trackback/3576/4301229

Listed below are links to weblogs that reference Five open questions about prediction markets:

Comments

Hello professor Tyler Cowen,

1. The demographics of the TradeSports/InTrade traders are very different than the Las Vegas visitors.

(Male, young, Wall Street professionals, white, Republican, computer literate, etc.)

2. Sports/Horseracing prediction markets attract the most volume. Without them, TEN/BetFair would not be profitable.

3. Silly to sell an academic paper. Let's evangelize the free world with pertinent information. That's the principle on which the two Googlers built their fortune.

Posted by: Chris. F. Masse .COM at Feb 21, 2006 8:33:22 AM

I agree, selling academic papers is a non-starter. Anything relevant is going to put itself on the web for free (yes, there is a message here for the NYT editorial page).

Posted by: Matthew Cromer at Feb 21, 2006 9:13:49 AM

A chance at fame will fuel the market. When it gets big enough reality shows will be based on it.
Televised "agony of defeat."

Posted by: Huggy at Feb 21, 2006 9:35:05 AM

The biggest hope for attracting "uninformed" players seems to me to be hedging. There is an open question here, which is why companies aren't hedging on the various matters of public concern that show up in betting markets. Companies absolutely do have a hedgable interest in who is elected to Congress or whether bird flu turns up in their markets. So why isn't there more hedging going on?

Your argument in your paper -- that there is not enough volume in prediction markets to support hedging -- seems to put the cart before the horse. Hedging ought to dramatically increase the volume in prediction markets by skewing prices out of line with informed expectations. If there is not enough demand for mispriced contracts to soak up the hedging, we would expect to see these markets end up at absurd prices rather than at the apparently rational ones that they do produce.

I would suggest another alternative: Public exposure. No company wants to be seen to have bet against a political winner, for fear of retribution. No company wants to be seen to have bet on a disaster, however noble its reasons, for fear of both losing public sympathy and more importantly being denied emergency "bail-out" funds.

If this is true then the next step toward more successful prediction markets is to enable a certain amount of money-laundering -- anonymized betting that allows a party to hedge without risking public censure.

Posted by: Grant Gould at Feb 21, 2006 9:50:28 AM

Speculative markets have long aggregated information; what is new now is the idea of creating markets on purpose to take advantage of this effect. Not only can you pay to create a market on a topic where you want information, you can also pay to subsidize the market, to induce participation.

But you wanted the information on the cheap without paying for it? You might get lucky, but there is no fundamental reason to expect you to succeed.

Posted by: Robin Hanson at Feb 21, 2006 10:13:43 AM

How do we know the demographics of Tradesports traders? Where can I get that information? Tradesports has released it, I assume? If so, where can one get this?

Also, assuming you're right and traders are disproportionately white, Republican, computer literate, etc., might they be sorting into markets where their knowledge is nonetheless relatively accurate? It could be those outside the demographic you note are self-selecting out because they lack the information necessary to make a bet profitable.

Also, it'd be interesting if the markets perform less well in situtation where those not represented by the demographic you note would appear to have superior local knowledge about the probability of an event occurring. I'm not sure what that would look like, though, because if I did, I probably could make a lot of money taking advantage of those imperfections.

Posted by: scott cunningham at Feb 21, 2006 1:43:10 PM

Hi Scott,

Public information on demographics is only available for IEM:
http://www.biz.uiowa.edu/iem/archive/references.html

As for the questions you ask, look into the IEM papers what they call the "marginal traders" and what their function is.

Best regards,

Posted by: Chris. F. Masse .COM at Feb 21, 2006 3:37:13 PM

What Tyler said applies to _all_ futures markets, not just idea futures markets : they are bad investments and if you want to invest for the long run, you are better off putting your money in Wall Street (i.e. some combination of stocks and bonds). Pork bellies and oil futures are just as bad as idea futures in this respect. Actually, they may be worse as physically deliverable commodities have a greater risk of manipulation by the people who write the deliverability specs.

Grant Gould is exactly right that hedging is the big motivation for futures. Speculation (placing what you think are well informed bets) is the second motivation. Although, it is likely that companies already are hedging bird flu via currency and agricultural futures.

Posted by: DK at Feb 21, 2006 4:56:26 PM

Grant's "cart before the horse" comment was dead-on. That prediction markets have been relatively thin is an historical factoid and is in no way part of their definition. As I've harped on, hedging usefulness not only provides liquidity but is also conducive to favorable regulation, which of course is the #1 issue faced by these markets. However, such liquidity comes with at a cost -- it may then become more difficult to interpret prices as probabilities, although trading will probably dry-up before prices get too far out of whack.

It will be very interesting to watch the new CME housing futures which apparently are still planned to launch in the second quarter.

Also Grant, most exchanges would maintain anonymity unless there were a criminal investigation -- precisely to maximize liquidity/commissions.

Posted by: Jason Ruspini at Feb 21, 2006 6:19:26 PM

Tyler, the San Francisco Federal Reserve have just published the paper in their economic working papers series (2006-06). Available here:
http://www.frbsf.org/publications/economics/papers/index.php

Also worth looking at is the Wolfers and Zitzewitz guide to prediction markets prepared for the forthcoming new edition of the New Palgrave Dictionary of Economics. My 9 December 2005 post has the details:
http://neweconomist.blogs.com/new_economist/2005/12/a_hitchikers_gu.html

Posted by: New Economist at Feb 22, 2006 8:54:53 AM

hemels tiener ^^^ amateurs himmel ^^^ allievo ubriache in casa ^^^ asiatiche merda nel bagno ^^^ sentiment jeune du cul a la bouche ^^^ extraterrestre poupees amour ^^^ sofa quinze ^^^ papa quarante deux ^^^ pleasant fighette pompino ^^^ celeste amatoriali sesso ^^^ lojlig tonaring tillverkningen alska ^^^ overnaturlig tonaring samlag ^^^ piu bollente allievo inculate ^^^ lust agente di polizia azione ^^^ varm pikk ^^^ hetere nifs fatter ^^^ kald fyldig ^^^ sentimental ubetydelig ^^^ to pio kafto nosokoma ^^^ koritsia kolos sto spiti ^^^ chaude fils roue ^^^ salope trente quatre ^^^