- The Big Picture - http://www.ritholtz.com/blog -

Prediction Markets Redux

Posted By Barry Ritholtz On July 24, 2005 @ 11:20 pm In Markets,Psychology | Comments Disabled

Its like waving a red flag in front of a bull. And I knew it would engender a response  . . . I just was hoping for a more intellectually honest one — one that might stimulate a dialogue on the value and inherent limitations of prediction markets. 

Last month, I mentioned that the InTrade Prediction Market got not one but two high profile events wildly wrong: The first was the Michael Jackson verdict — Not Guilty on all counts [1]; The Second was the Morgan Stanley CEO Purcell resignation before June 30, 2005 [2].

And the betting crowd didn’t just barely miss, either — they totally $*%t the bed.

My pointing this out caused a fit of pique amongst the prediction market crowd. Now that the hyper-ventilating has subsided, lets revisit (again) some of the readily identifiable problems with endemic to these markets.

1) The inherently unknowable is, by definition (duh) inherently unknowable. The further into the future one peers, the less accuracy one can expect on forecasting singular events. Call this the rule of the unknown future.

2) As we saw in both the Jackson [1] and Purcell [2] cases (as well as Howard Dean’s Iowa Primary [3]), the crowd cannot create wisdom simply by pooling their collective ignorance.  Indeed, the less an event is dependent upon the actual behavior of crowds, the less likely it is that prediction markets will forecast a particular outcome correctly.

3) I use markets as a forecasting tool in my own work. The collective behavior of market participants can often provide insight into future market behavior. In other words, the present behavior of investors can sometimes be used to predict the future behavior of these same investors. The key is to use markets that are liquid, widely followed, with significant money at stake.

Example: Fed Futures have a very respectable track record of predicting Fed interest rate tightening and loosening. Not perfect, but good. That’s because fixed income managers use it to hedge their portfolios. They also have an idea of whether they will be buying or selling bonds, and therefore they collectively impact bond prices, yield, and therefore the Fed. There’s a feedback loop involved.

Speaking of the 30 trillion dollar U.S. Bond Market, low interest rates (set by collective bond traders and investors — not the Fed) have a good track record predicting future economic activity. Anyone who ignores the Long Bond’s low rates, and the implication of slowing economic activity, is inviting disaster.

Again, not perfect, but respectable. The depth, amount of money involved, and liquidity of this market is what makes its “predictions” worth following.

4) Where futures market do particularly well is where they track the behavior, beliefs and perspectives of their collective participants. For example, I do not believe political futures predict anything — instead, they are more like an alternative polling method that consolidates and synthesizes all the pre-existing tracking polls.

That’s why the last election cycle, as we saw in the Primaries, the Iowa Electronic Markets  totally missed Dean’s meltdown, and underestimated Kerry.  Further, on election day 2005, as soon as exit polls in Florida gave Kerry a lead, the futures traders did what I believe is now incontrovertible: they aped the polls!

As the data shows, by 5 pm, the Futures Markets were PREDICTING:

President.Kerry2004 trading at 67.0 by 67.9
President.GWBush2004 trading at 32.2 by 33.0

If that’s not prima facie proof that prediction traders merely track polls . . .

As Slate’s Dan Gross [4] noted:

“The furious and seemingly irrational Election Day market action stands as evidence that the traders are more poll-followers than poll-beaters.”

Somehow, this intraday swing gets omitted from all of the charts mustered to show how prescient prediction markets are.

That hardly demonstrates collective wisdom or forecasting acumen. I will reiterate that what makes these markets so susceptible to failure: too thin, too few participants, illiquid, with too little real money actually at stake.

Even The Capital Markets Get it Wrong

“The Stock Market has predicted 11 of the past 4 recessions”

That old joke is quite telling: Despite the liquidity and massive
dollar amounts involved, even the Stock and Bond markets are quite
fallible as predictors. 

After the 2000 crash, the market had 5 substantial rallies [5]
which many economists proclaimed as a sign of an imminent recovery.
They were wrong. Similarly, some people were insisting the first half
of the year’s weakness as a sign of an imminent recession (wrong again).

If these enormous, liquid, widely followed markets forecast wrong
(or were wrongly interpreted by the experts) — how accurate can we
expect the thinly traded futures markets to be?\

Less than Honorable Criticism
One final note: I am in favor of accountability. I have said repeatedly that forecasting is folly [6]; When I do make forecasts, my working assumption is that they will be wrong [7]

As a strategist, I like to wargame scenarios, taking them into whatever dark crevices they may lead. This leads me to, on occasion, I often stake out positions that are not mainstream. Offbeat, but intellectually honest.

Which brings us back to “fair & honest" criticism: Last year, A discussion I had in Business Week last year before the elections is now being used to (lets call it) inelegantly critique my criticism of prediction markets [8].

After a long discussion about some of the historical quantitative data [9] that typically corresponded to an incumbent defeat, I was asked who I thought would win:

“What makes the 2004 election such a challenge to forecast is that we have never seen a Presidential term with a burst market bubble, a recession, a major terrorist attack on U.S. soil, a big tax cut, and not one, but two, wars. So without an analogous comparable, making a prediction with a high degree of confidence becomes quite problematic — it’s just a crapshoot.”

That’s the context of my “prediction” — I called it a crapshoot. Hardly pounding the table, certainly not the same as putting on a 10 million dollars position (see the whole thing for yourself here [9])

Yet somehow, this article was called the basis for my “humbling” and was mustered as an example of the superiority of prediction markets.

As criticism, I find it wanting.

Here’s my criticism:  I make this challenge to Chris Masse & company [10]:

I believe the political futures traders don’t lead, but merely ape the polls. If that’s not the case, please explain (in a 1000 words or less) why once the exit polls were released as of 5 pm on election day 2004, Kerry futures shot up to 67 while G.W. Bush were trading at a mere 32.2.

I’ll pick the best one three and post them on The Big Picture [11].

Article printed from The Big Picture: http://www.ritholtz.com/blog

URL to article: http://www.ritholtz.com/blog/2005/07/prediction-markets-redux/

URLs in this post:

[1] Not Guilty on all counts: http://bigpicture.typepad.com/comments/2005/06/the_prediction_.html

[2] Purcell resignation before June 30, 2005: http://bigpicture.typepad.com/comments/2005/06/purcell_predict.html

[3] Howard Dean’s Iowa Primary: http://bigpicture.typepad.com/comments/2004/01/iowa_and_predec.html

[4] Slate’s Dan Gross: http://www.slate.com/id/2109137

[5] 5 substantial rallies: http://bigpicture.typepad.com/comments/2003/07/rally_compariso.html

[6] forecasting is folly: http://www.thestreet.com/_tscana/comment/barryritholtz/10226887.html

[7] they will be wrong: http://www.thestreet.com/comment/barryritholtz/10215965.html

[8] inelegantly critique my criticism of prediction markets: http://www.chrisfmasse.com/2/2005/20050618.html

[9] long discussion about some of the historical quantitative data: http://www.businessweek.com/bwdaily/dnflash/aug2004/nf20040830_5077_db049.htm

[10] Chris Masse & company: http://www.chrisfmasse.com/

[11] The Big Picture: http://bigpicture.typepad.com/

Copyright © 2008 The Big Picture. All rights reserved.