Prediction Markets Fail Again
GO FIGURE: Prediction markets -- those thinly traded games played by college kids and other traders who lack the capital to trade in deeper broader markets (equities, fixed income, commodities, currencies) -- are for shit:
"To the amazement of economists and online bettors, the answer has varied a great deal among betting Web sites.
Markets are not supposed to work that way, even online prediction markets, where bettors trade on the chances of a candidate’s winning an election in the same way that they might bet on pork bellies to go up in value.
In the last few weeks, Intrade.com, which is based in Dublin, had consistently given John McCain as much as a 10 percentage point edge in his chances to be elected president compared with other large online overseas betting sites. These include the British-based Betfair.com, as well as the Iowa Electronic Markets, a research project at the University of Iowa that allows bets of $500 on election results."
Actually, thinly traded markets such as this are supposed to work EXACTLY this way. What prevents the "real" markets from operating this way (most of the time) is the enormous amounts of money at stake, and the huge and diversified crowds of traders watching for aberrations. Little markets, small amounts of money (millions not trillions) and thin trading are prone to this sort of nonsense. (Don't say you weren't warned)
Here's, the question of the evening: How many legitimate market strategists, economists, and traders still believe in the Efficient Market Hypothesis? Anyone?
Previously:
Misunderstanding Prediction Market Failures (February 14, 2007)
http://bigpicture.typepad.com/comments/2007/02/misunderstandin.html
Why Prediction Markets Fail (January 11, 2008)
http://bigpicture.typepad.com/comments/2008/01/prediction-mark.html
Sources:
Trading Variance in Election Predictions Raises Questions
NOAM COHEN
NYT, October 19, 2008
http://www.nytimes.com/2008/10/20/business/20predict.html
Intrade Betting is Suspicious
Nate Silver,
Five Thirty Eight , September 24, 2008
http://www.fivethirtyeight.com/2008/09/intrade-betting-is-suspcious.html
Trader Drove Up Price of McCain ‘Stock’ in Online Market
Josh Rogin,
CQ Oct. 17, 2008 – 3:44 p.m.
http://www.cqpolitics.com/wmspage.cfm?docID=news-000002976265&
Monday, October 20, 2008 | 08:53 PM | Permalink
| Comments (47)
| TrackBack (0)
add to de.li.cious | 
digg this! | 
add to technorati | 
email this post
TrackBack
TrackBack URL for this entry:
http://www.typepad.com/services/trackback/6a00d8341c52a953ef0105359d994b970c
Listed below are links to weblogs that reference Prediction Markets Fail Again:
Comments
It is obvious from your blog postings that you do not believe EMH is a valid hypothesis regarding large, liquid financial markets but I don't believe I have ever read where you write what you think EMH means. How about a quick post about your take on EMH and why it is invalid?
~~~
BR: I've written extensively on this -- but try this for a starter:
http://bigpicture.typepad.com/comments/2004/11/the_mostlykinda.html
Posted by: Joe | Oct 20, 2008 9:14:38 PM
One individual was dumping hundreds of thousands of dollars to prop McCain's price up on intrade. Have no idea why though.
http://www.intrade.com/jsp/intrade/misc/blog/?initialBlogId=jd_1
Posted by: Joe | Oct 20, 2008 9:23:42 PM
BR,
Seriously. Put up or shut up on EMH. There's next-to-no chinks in the armor of weak or strong form, and the prevailing evidence is that semi-strong is only invalidated in cases where very very small stocks are under the auspices of good ole' boys doing what good ole' boys do.
Honestly. For the love of god. Please. Just shut up about it. You're a trader and a PM, not some Math wonk from MIT beavering away doing lemmas in your head most people can't do with years of calculus and a TI-89. [BR: I was an applied mathematics/Physics major at Stony Brook under some guy named James Simon, who now runs Renaissance Technologies] Either get an NSF grant and produce some non-anecdotal peer-reviewed evidence or be quiet already.
Yes, I am aware that EMH breaks down at a microcap level, falling under the purview of sites such as Intrade that have traded a *whopping* $20 mil on the elections so far. $20 mil is good for about 2 minutes worth of MSFT. Wow, shocker: illiquid, thinly traded markets get things wrong. Alert the press.
Sorry to be such a dickhead, but I happen to have a math and stat background, and I don't like having my work and the work of my colleagues impugned by people who, quite frankly, aren't especially well-versed in what they or we are talking about. I can change my oil; it doesn't mean I tell the mechanic how to fix my car. By the same token, I don't expect the un-inititated to tell me how to go about my business, especially when nearly all of the evidence suggests the opposite of your implication.
Doesn't mean markets aren't irrational in the short term. Doesn't mean markets don't misprice assets in the short term. Doesn't mean markets are not quite efficient and quite rational in the long term either.
Seriously, honestly, no-sarcasnm, have a great evening. I will still come to your site three times a day, because even though I am being an ass, I wholeheartedly agree with 90%+ of the content you post. Decimate me. Tear what I'm saying apart. Marshall every resource you can think of to demonstrate I'm wrong. Or don't. But do not continue to whale away at EMH with nothing more than anecdotes involving political markets representing less capital than your firm trades in a single day.
One last thing: How the West was Won is still on my Mp3 player 5 years later. Epic.
Posted by: Byno | Oct 20, 2008 9:47:56 PM
Trade you ten shares of McCain for 1K of Pets.com.
Posted by: leftback | Oct 20, 2008 10:01:27 PM
BR,
I am not sure if you are aware, and the chart you post does not make this clear, but IEM and Intrade contracts on president do not trade on the same event.
Intrade contracts are based on who wins the electoral vote. IEM contracts are based on who wins the popular vote. As we know from 2000 these are not the same things and in a year when a few weeks ago Obama had a statistical lead in the national polls but appeared to not be translating that into victories in swing states it makes sense that Intrade would given McCain a better chance to win than IEM since the odds of an electoral vote win at the time were higher than a popular vote win.
Posted by: RedOcean | Oct 20, 2008 10:03:40 PM
Although I am agnostic on the issue, Edgar Peters provides a somewhat convincing argument against the EMH in "Chaos and Order in the Capital Markets"
Posted by: hugh | Oct 20, 2008 10:08:06 PM
Byno, exactly what is your definition of "long term" in this sentence:
"Doesn't mean markets are not quite efficient and quite rational in the long term either."
I just wanna know, because in 2007, i watched efficient and rational blown to effing crap. And, anyone, that every tells me that markets are forward looking is also full of crap. Furthermore, the SEC and Fed have turned the markets into a crapshoot, while i'm on the subject of crap.
Posted by: karen | Oct 20, 2008 10:09:16 PM
The EMH is irrelevant, IMO. Why? Because things change. We don’t make money by exploiting inefficiencies in the market, but by anticipating those changes better than others.
Posted by: KJ Foehr | Oct 20, 2008 10:11:53 PM
Barry, yes you are correct that this spread has been in existence for a while now, and Intrade even posted on its website to alert traders to this arbitrage opportunity. However, to use this to bash all prediction markets is way out of proportion to the event. Since its inception in the mid 1980s, the Iowa Electronic Market has predicted the winner of every presidential election, and has been a more accurate predictor of share of popular vote won than any poll, aggregation of polls, or political pundit. Also, if you want another example of accurate prediction markets, look at a chart of Hollywood Stock Exchange predicted movie grosses vs. actual grosses. Drawing a best-fit line, there is hardly a difference. Yes, many prediction markets are small, illiquid, and thus have problems with prediction, but they can nonetheless be very valuable tools in predicting future events.
Posted by: Zach | Oct 20, 2008 10:12:15 PM
Barry, your posting is a little naive. If I wasn't capped by the $500 deposit limit that IEM imposes, I assure you that that 10% arbitrage would have shrunk to zero much earlier.
Posted by: Frank Tobin | Oct 20, 2008 10:13:23 PM
Sounds like Soros's reflexivity theory should get more of a boost here as opposed to the Efficient Market or Equilibrium hypotheses that have failed us miserably.
Posted by: Joseph McCann | Oct 20, 2008 10:27:58 PM
BR --
I saw two stocks that typically move in an inverse fashion, move in parallel today for an entire 7 minutes.
Therefore all previous evidence must be thrown to the wayside in favor of my new superior theory which I gleaned over 7 full minutes of anecdotal observation.
The EMH needs an EM to work. Are there non efficient markets? Of course.
-- Obio
Posted by: Obio | Oct 20, 2008 10:39:11 PM
Byno,
with this: "...but I happen to have a math and stat background, and I don't like having my work and the work of my colleagues impugned by people who, quite frankly, aren't especially well-versed in what they or we are talking about..."
Gee, maybe, You could provide an example of what you're talking about, you know, to bolster You're position?
past that,
@karen,
no kidding! one would think that these would be the 'glory days' of EMH-style fundz..
if EMH had the goods, it'd be PR-flack's dream-assignment to trumpet the 'good news' from every hilltop through ev'ry holler..
EMH is just more BS concocted to sell MutFunds, Index Funds, and the 'Buy 'n Die' approach to Di-vesting..
Posted by: Mark E Hoffer | Oct 20, 2008 10:43:47 PM
i love these emh people. they should try this equation: human existence = zero. and how do we get there????
Posted by: karen | Oct 20, 2008 10:48:36 PM
Byno,
I'm ignorant. But, here's what I think. Please correct me.
I don't think the argument against EMH has to be made mathematically. I look at the oil price boom/bust and the recent equities "Wile E. Coyote" moment as proof that wildly disseminated information (e.g., Level III mark-to-myth, demand destruction) can be widely ignored for a time, such that stock prices -- for quite a while -- don't price in reality.
What does EMH say about the idea of "open secrets" that 90% the market thinks "doesn't matter" until it does? Does it embrace/explain the psychology of markets and the abnormal returns of the couple of people who can see the emperor has no clothes? In those situations, everyone has access to the same information. But, not everyone "sees" things the same way. Is that a form of "luck"?
Actually, I could see an argument that it *is* luck, if you define that term loosely. But, what does EMH say?
Regarding "proof" of EMH, Wikipedia says:
"To test for strong-form efficiency, a market needs to exist where investors cannot consistently earn excess returns over a long period of time. Even if some money managers are consistently observed to beat the market, no refutation even of strong-form efficiency follows: with hundreds of thousands of fund managers worldwide, even a normal distribution of returns (as efficiency predicts) should be expected to produce a few dozen "star" performers."
Well, fund managers worry about more than just performance. They worry about relative underperformance and client defections. As a practical matter, I'm skeptical that it's possible to even conduct a plausible study to measure to prove/disprove EMH.
Posted by: wunsacon | Oct 20, 2008 11:11:23 PM
Byno,
Even Eugene Fama (the creator of EMH) has repudiated some of the theory.
There's extensive proof that trends have persistency, which seems to violate the Random Walk side of things.
You may want to read some of Andrew Lo's work as well.
Posted by: Barry Ritholtz | Oct 20, 2008 11:20:52 PM
Bar, would you rather check intrade once a day for 5 seconds, or watch CNN for more than 5 minutes.
The idea is that even the "thinly traded" markets give you better info than everything else (pundits, polls, etc) combined, quickly and without bias.
And yeah, they're about as efficient as the equity market, meaning not at all, just better than any other price setting mechanism.
Posted by: yt | Oct 20, 2008 11:30:20 PM
Byno: "I happen to have a math and stat background, and I don't like having my work and the work of my colleagues impugned by people who, quite frankly, aren't especially well-versed in what they or we are talking about..."
I imagine Merton and Scholes said something similar right before LTCM blew up. Luckily, the huge CDS mess currently unwinding has nothing to do with mathematical or statistical models created by arrogant PhD's.
Posted by: JD | Oct 20, 2008 11:34:16 PM
By the graph, then, considering the 10% handicap to McCain, and seeing the convergence happening, I'd say that Obama is on his way to having the vote locked up, unless the game is rigged, of course.
--mf
Posted by: Monkeyfister | Oct 20, 2008 11:45:46 PM
Byno: "I happen to have a math and stat background..."
Hence you've bought into the EMH sham. You believe your tools work because you know your tools and without your belief, you'd have to admit you don't know jack about the markets.
The EMH is based on rational behavior. Behavioral economics has put the lie to the presumption that people behave (economically) rationally. The bubbles and crashes we've been experiencing over the last 30 years put the lie to the presumption that markets behave rationally. The huge debacles (e.g. LTCM) show how far wrong one can go by following statistical models which assume this behavior. The current problem with MBSs and CDSs show how financial engineering constantly changes the game so that there is no "long run".
If that doesn't persuade you, go read Taleb for some philosophical meanderings over just how flawed your view of the market is.
If you need something more academic, go read the work by Stiglitz and Greenwald to see that efficient markets are the exception, not the norm.
Posted by: srg | Oct 21, 2008 12:05:40 AM
I should probably clarify my earlier post and make it less snarky. I misunderstood Barry's position to be that the arbitrage existed because it was inefficient, and that is somewhat true, since because I can't put more capital into the IEM, I couldn't full advantage of the inefficiencies. However, more directly, if simply the deposit limit wasn't imposed, the market would have become more efficient, eliminating arbitrage and averaging out the two markets' predictions.
As to whether or not the prediction markets are efficient at predicting outcomes, I would agree that they are inefficient. At IEM, long odds on rare events can be purchased cheaply (and without transaction cost!), and people appear to be severely underpricing their likelihood.
Posted by: Frank Tobin | Oct 21, 2008 12:16:23 AM
I can change my oil; it doesn't mean I tell the mechanic how to fix my car. -- Byno.
Byno --
Bad analogy. Knowing how to change the oil in your car and knowing when to change it and knowing how to know when it needs to be changed is a significant part of knowing how to fix a car.
Your faux supremacy is bit unconvincing.
Cheers.
Posted by: Douglas Watts | Oct 21, 2008 12:35:43 AM
Answer to your question: 27
How many surrealists does it take to change a light bulb?
Posted by: swimming | Oct 21, 2008 1:41:42 AM
Usually I like you Barry, but you don't know jack about these markets. First, the IEM and Intrade contracts are different. Even in a trillion dollar market you would not be able to arbitrage the difference. Second, you do not need a massive sample size to get very good statistical significance. That is, just because the market is relatively small does not mean that it is not a relatively accurate measure of likelihood. Third, prediciton markets have been very good at predicting outcomes historically. Not perfect, but often better than standard polling models. Fourth, the tone of your post is aggressive indicating that you are insecure about your underlying understanding of the topic you chose to write about(think Sarah Palin here).
If you want to tear something apart, do it with real data, not phony bluster. And remember, it's not the end of the world if McCain wins, just the end of the world as we know it.
Posted by: GaryD | Oct 21, 2008 1:49:10 AM
Byno,
How much is Renaissance down YTD? They seem to own all the value stocks I'm coming across (not having a PhD from StonyBrook - I'm not smart enough to run EMH models). The funny thing is, I think they bought in at 4-5x current prices.
Of course, they might be the ones selling out of those stocks, because that 77% return last year forgot to mention the 10-1 leverage.
Oh, right.
Posted by: s | Oct 21, 2008 1:50:15 AM
