pageok
pageok
pageok
[Michael Abramowicz, guest-blogging, January 28, 2008 at 11:34am] Trackbacks
An Intro to Prediction Markets and the Liquidity Problem:

Thanks, Eugene! I am pleased to be a guest conspirator, and I'm looking forward to writing about Predictocracy. I imagine most readers here are familiar with prediction markets, so I'll start with only a brief explanation of how prediction markets usually work (a longer explanation from my book is here).

In a prediction market, traders can buy and sell contracts that will pay off should a particular event occur. For example, as of this writing, on Intrade.com, you can buy for approximately 58 cents on the dollar a contract that will pay off should Sen. John McCain win the Republican nomination. The current prices at which people are willing to buy and sell McCain shares translate into an estimate that McCain has approximately a 58% chance of being the nominee.

Most of the markets that the book imagines, however, differ in a fundamental way from the Intrade markets. While Intrade charges for its services, the markets that Predictocracy envisions would generally be subsidized by institutions (such as businesses or governments) willing to pay for the estimates that prediction markets produce.

Appropriately administered subsidies would respond to a common criticism of prediction markets: that on non-sexy topics, prediction markets have too little liquidity to be of much use. Even some Intrade markets attract too little attention, at least some of the time, to be useful; consider the current estimate that Imran Kahn has somewhere between a 10% and a 90% chance of being elected Prime Minister of Pakistan.

A subsidy can solve this problem by rewarding individuals who place aggressive offers to buy and sell contracts. This approach, which I describe here, helps reduce the "bid-ask spread," and thus indirectly increases the incentives of people to develop information and analysis that they might then be able to trade on. An alternative approach, such as Robin Hanson's "market scoring rule," is to create an automated market maker that is willing to buy or sell shares at prices based on the current prediction.

Duffy Pratt (mail):
What's the over/under on the number of days until Edwards drops out?
1.28.2008 11:46am
Chris 24601 (mail):
What's the best response to the argument that the nomination markets are manipulated by supporters of different candidates?
1.28.2008 11:51am
Chris 24601 (mail):
Intrade has a 10.0-56.9 bid-ask for Edwards dropping out by the end of February. Not as much liquidity as you might like there...
1.28.2008 12:44pm
byomtov (mail):
What's the best response to the argument that the nomination markets are manipulated by supporters of different candidates?

Or, more broadly, by traders who stand to gain substantially from certain outcomes.

Suppose a business uses a prediction market to decide which of a group of technologies to pursue. What stops those employees with particular expertise in one area from aggressively bidding up its price? Sure, others may do the same, but that sort of game of chicken hardly seems to be a sound basis for decision-making.

The overall point is that a decison made by a prediction market may have sufficient importance outside the market itself that individuals are willing to spend money to influence it. Large "extra-market" payoffs to the outcomes make the market itself much less useful as a decision-making mechanism.
1.28.2008 12:53pm
J. F. Thomas (mail):
Where is my jet pack, flying car and picture phone?

Of course the market for Kahn winning is going to be horribly distorted by his international celebrity.

And don't forget once Stephen Colbert hears about these markets, all kinds of strange outcomes will be predicted as the Colbert Nation throws in their $.02 worth.
1.28.2008 1:02pm
J. F. Thomas (mail):
And this is different than betting on a football or basketball game how?
1.28.2008 1:09pm
Duffy Pratt (mail):
J.F. Thomas:

That was pretty much the point I was trying to get at when asking about the over/under. I don't have anything against gambling, but it's hard to see how it gets converted to something else by renaming it. In the immortal words of Billy Ray Valentine: "Sounds like you alls a bunch of bookies."
1.28.2008 3:10pm
J. F. Thomas (mail):
Actually, isn't the average horse track pari-mutual pool much more sophisticated than what is proposed here since a typical horse race has 8 or more participants and the odds depend on how much is bet on each horse. And how much is bet depends on an odd mix of experienced bettors who scour the racing forms, desperate fools who think they have a system or inside knowledge, and recreational gamblers who might bet on a horse simply because they like the name, it is pretty, or they like the colors the jockey is wearing.
1.28.2008 3:30pm
zap:
According to economist Paul Tetlock, too much liquidity actually makes prediction markets less accurate, due to phenomena such as the favorite-longshot bias. Very liquid prediction markets tend to overestimate the favorite's chances of winning and underestimate the probability of a longshot victory.
1.28.2008 3:51pm
Elliot123 (mail):
1.If I am the high bid for some given length of time, do I get a subsidy for that length of time?
2.If I am hit at my bid, do I still get the subsidy.
3.If I cancel my bid, do I still get the subsidy?
4.Is the subsidy limited to the initial market maker, or do others who join the bid also accumulate subsidy minutes?
5. If I am outbid, do I keep the minutes I accumulated while high bidder?
6.If I enter a market order bid and immediately hit the ask do I get any subsidy at all?

All tbis obviously contemplates scalping for minutes. While most futures markets have scalpers, the scalpers and all other participants win and lose based on price differential. In this scheme both price differential and time are the rewards.

I suppose the best way to test these ideas is to get some curious billionaire to fund a test.
1.28.2008 4:05pm
byomtov (mail):
Imran Kahn has somewhere between a 10% and a 90% chance of being elected Prime Minister of Pakistan.

I doubt that any member of the Kahn family has any chance at all of holding political office in Pakistan. Imran Khan is a different story, however.
1.28.2008 4:23pm
Daniel Chapman (mail):
I'll ask the same question I asked in your intro post because it was never answered there...

If you use the results of the prediction market to make a decision, don't you turn the prediction market into a self-fulfilling prophecy? Does guessing with the majority always mean you win? Do you get rewarded for committing enough resources to the market to sway the results in your favor? How does this differ from a vote where people's votes are weighted by how much money (or whatever is at stake in the market) one is willing to spend?
1.28.2008 8:18pm
Deven Desai (mail) (www):
The book looks fascinating. For those looking for what I think is an opposing view, check Taleb's Black Swans. A short version of the argument is here.

I have not read Predictocracy but given the pitch, Taleb's article "You Can't Predict Who Will Change The World" in Forbes (link above) seems to disagree. Put differently the two views may be passing each other but they seem related though opposed.
1.28.2008 8:47pm
AnonLawStudent:
J.F. / Duffy,

Isn't the point of these artificial "prediction markets" to generate precisely the same type of data as Vegas odds? With sports betting, very smart people with extensive knowledge of the subject (or casinos willing to pay to hire those people) are putting their money behind a given outcome. A prediction market is trying to generate that same data, by encouraging similarly knowledgeable people to enter the market (gamble if you like). In my mind, the better view is that pure gambling, which involves nothing but chance, can be distinguished from sports betting which, while certainly involving an element of chance, also involves a significant element of knowledge.
1.28.2008 11:16pm
Elliot123 (mail):
We would also see hedgers in some of these markets. For example, suppose the contract was for Dokes to win the presidency. If Dokes wins, I will get a nice fat political job. If he doesn't I get to mass mail resumes. I think he will win, but want some insurance.

So, I might sell Dokes as a hedge. If he wins, I get the nice job and lose my bet in the futures market. If he loses, I don't get the job, but do get a profit in the futures market. It's a risk management tool.

On the other hand, I might not sell Dokes, but might buy Jones instead. The decision would depend on the prices of Dokes and Jones at the time I make my trade.
1.29.2008 1:04am
Daniel Chapman (mail):
Or you could actually try to sway the market by buying up all offered shares of Dokes, thereby causing the desired outcome, getting your job, AND making money in the futures market.

There's no risk in the market if the market determines the outcome directly... and if there's no risk, then the market doesn't even work properly. Isn't this a paradox? Why are we even discussing this?
1.29.2008 7:42am
Michael Abramowicz (mail):
Chris 24601 -- I'll speak about manipulation soon.

J.F. Thomas -- Pari-mutuel markets are fairly good, but prediction markets have some advantages over them. See here.

Zap -- Good point, but t's hard to extrapolate this finding to markets in which liquidity is provided by an automated market maker.

Elliot123 -- See the discussion in the book, but yes, in the decentralized subsidy scheme, payouts are proportional to the amount of time your offer is in the front of the queue and the size of your order.

Daniel -- I did (eventually) respond in the previous post.

Deven -- Thanks for the link -- I don't really disagree with the argument, but think that it doesn't undermine the case for prediction markets.
1.29.2008 11:15am
cathyf:
Along with Elliot, I'm curious about people using the markets to hedge. Suppose, for example, you think that presidential candidate Pop U. List is going to be a disaster economically. So you hedge your stock market positions by buying List on the predictions markets, and/or selling contracts of other candidate(s) who have economic planks which are not taken out of an ancient soviet 5-year plans.

This causes the prediction markets to overestimate the probability of the List winning, and underestimate the probability of the other candidates winning. This is a variation on the phenomenom where the index futures, index options and index futures options appear to "lead" the spot market because the markets function differently.

The other question is insider trading. For example, a candidate decides to pull out of the race, goes and sells his contracts up the wazoo on the predictions markets, and then pulls out. Financial markets and sports organizations have laws and regulations to prevent players from betting against themselves and then throwing the game. (Organized crime run gambling has similar mechanisms in that they'll break your legs if you try to steal from them.) If I'm not an insider, why should I go into the predictions markets and get ripped off by insiders?
1.29.2008 6:29pm