Gambling and a New Approach to Regulating Information Markets
Paul Tetlock October 3rd, 2006
We address the question of what constitutes gambling in a recent article in the Journal of Regulatory Economics: “A New Approach to Regulating Information Markets.”
In brief, Congress should impose a requirement that prediction market contracts pass an “economic purpose test.” We would only allow prediction market contracts that are likely to provide significant financial hedging opportunities or valuable information for improving economic decisions. For example, contracts on expected annual rainfall could allow a farmer to reduce his losses in the event of a drought. Our test would not allow for Internet poker games or lotteries, but would definitely allow contracts involving GDP or housing starts.
For such regulation to work well, it needs to be light-handed. To maximize innovation, Congress should allow for federal preemption of other state and federal regulatory authorities for prediction market contracts that pass our economic purpose test. Such preemption would reduce uncertainty for innovators concerned about violating state gambling laws.
–Robert Hahn and Paul Tetlock
- All Best Posts Ever , All Guest Authors's Posts , Analysis (Industry) , Regulations
- Comments(6)
on 
Unfortunately, the CFTC will be reluctant to approve precisely the sort of contracts that most interest us - those on policy events. I was told by someone at an exchange that the CFTC said they would approve a presidential election market, but not a market tied to a legislative event. The fact that the latter has more direct hedging utility (of taxes or subsidies, for example) points to a structural flaw in the CFTC. The CFTC relies on Congress for funding and is an unusual “sunset” agency that must be reauthorized by Congress every five years. Thus they need to avoid crossing the legislature in any way.
But hopefully their outlook can be swayed over time. Contact from someone at AEI-Brookings certainly wouldn’t hurt.
This is a huge flaw in how the CFTC operates on this particular issue, I agree Jason. We’ve talked about the legislative event products before - while difficult to implement, they would be hugely beneficial. There might be just too much opportunity to cheat for these to be sacnctioned by any governmental body.
That said, I think the test proposed by Robert and Paul is a simple but effective test. It is good because it seperates trading from gambling - not that gambling is bad, but that gambling fullfills a different purpose. At the very least, it provides the first step to evaluation on how the particular product should be regulated. If it passes this economic test - its a CFTC or SEC style issue. From there you can make further distinctions about the viability of the proposed contract. If it doesn’t pass this economic test, gambling style regulations should be considered.
A semi-ironic effect of this test is that products that pass the test would in some ways be less regulated than those that fail this test.
From Mike Sankowski’s comment, I would further his point and would like to see a “predictive test”, which would distinguish between entertainment-only gambling (bookmakers) and speculative prediction markets.
Am I making sense?
(Maybe somebody should start a brand-new blog post on my proposal, because I’m carrying away Paul Tetlock’s original topic.)
Chris,
You are making sense. I think the overall test of economics is in some ways vague - but thats its a good guiding principle. I don’t think you’re going off topic that far - this is an area that simply needs more hashing out, and
“Is this for entertainment?” is an excellent additional question. It might actually be better opener than economic value. For example, many of the bets monitored by the Long Now foundation
http://www.longbets.org/predictions
address issues that are important (i.e. “you are immortal”), but are too vague or too powerful to have much real economic value to a single person/entity. I hesitate to call them ‘entertainment’ as the whole point is to encourage discussion on the topics listed. You could easily create interesting predictive markets around some of the ‘bets’. Long Now solves the ‘gambling’ part of this by having the winnings go to charitable foundations - but why should this be necessary?
I meant:
1. Entertainment-only gambling (bookmakers, casinos)
2. Speculative prediction markets (BetFair, TradeSports)
3. Prediction markets that has hedging capacity (HedgeStreet, etc.)
2. Predictive power = social utility #1
3. Hedging = social utility #2
If there is a social utility, it should be legal.
Tetlock and Hahn want to assess whether there’s “valuable information for improving economic decisions”. Isn’t it a bit too narrow? What’s useful for one person or business is useless for another, and conversely. What if sports prediction markets generate useful predictive power for my particular sports-related business? How could I demonstrate that to the two AIE-Brookings guys?
[...] - Gambling and a New Approach to Regulating Information Markets [...]