Iowa and Prediction Markets

Markets are not God.

To many people, this statement is a form of economic blasphemy.

I suggest those people should get over it. In the past, I’ve challenged the issue of how “predictive” markets actually are. I note that many people read what they want into short term jags and twists, despite the obvious limitations of such forecasting. Even longer term trends are less revealing of the future than they are of the recent past — other than to say that the trend is more likely to continue than not.

Consider the Democratic Caucus in Iowa. There’s been a solid year of campaigning leading up to the Iowa Caucuses, with the last 3 months totally intense; Its the first — and some will argue the most influential — primary of the entire process. If any prediction market should function correctly, this one was it. No excuses, just right or wrong forecasts. Otherwise, the markets would amount to little more than a polling mechanism of popular opinion.

Up until a few days before the primary, this market’s “prediction” was that Dean would win — and do so handily.

The Prediction Markets got it dead wrong:

Predict This!

The chart shows exactly where each candidates “futures” traded. The Dean futures had reached a peak of 76 cents late November, and maintained a fairly high value until just before the primary. In early January, they were still as high as 60 cents or so.

Meanwhile, Kerry has been trading no higher than 35 cents since 11 months prior in February 2003. Kerry went mostly sideways (with a slight negative bias) until August, when he started trending down more rapidly. In December, Kerry futures traded at just a few pennies. He had one last “dead cat bounce” to a dime, and then slumped back to 2 cents in early January 2004.

So much for the predictive value of futures markets.

The current Democratic National Convention Nomination Market shows a very different picture. Dean dropped some 75% to 15 cents or so. Meanwhile, Kerry is partying like he’s Yahoo! and its 1999: His futures rallied about 3,000%, from 2 cents to 60 cents in a week. Talk about overbought!

The stock market’s up since March — therefore it must be predicting a recovery, right? Well, not necessarily. We’ve had 4 previous bear market rallies since 2000, although this one has lasted the longest by a large margin. But ask yourself: What was the market predicting late February 2000, when the Nasdaq broke 5,000? What was it predicting in October 2003, at Nasdaq 1,100?

There are myriad problems with this form of interpretation. Financial markets act not as “predictors,” in my opinion, but rather as future discounting mechanisms. Think of them as racetracks: you get paid lower odds the better the horse looks before a race. When the nag appears ill, old or tired, the odds are highest, and buyers get the greatest potential payoff. When the steed starts to look healthier, the odds slide lower — they get “discounted” — as an economic recovery starts looking more and more likely.

If “Prediction Markets” do not actually predict the future, than what do they actually do? I suggest they merely reflect the majority opinion at a given moment. That does not imbue them with any special omniscience. I think of them as polls that avoid random spoofing ’cause the polled must pay an entry fee to participate. That generates more serious responses than other polling data — but the answers are just as potentially wrong as any other future guess. Like the majority, sometimes they are right, and sometimes they are wrong.

I’m not suggesting that market data can’t be informative. In the hands of thoughtful and skilled analyst, it can indeed impart some measure of knowledge. But interpreting market data is far more complex and significantly less predictive than is the common wisdom on Wall Street.

What, then, can we learn from futures markets? The smart answer is, “It depends.” Interpretating what markets are saying requires flexibility, skepticism, and an open mind. It behooves the market watcher to recognize that “the crowd” is what takes markets higher — or lower. One must also recognize that when the crowd turns into a thoughtless mob, a reversal of fortune becomes ever more likely.

The sooner people understand what market data suggests, the more financially secure they will be in their own futures. “Thy shall have no false idols before me,” exhorts the old testament. Markets are not God. People who think they are eventually succumb to a Hell — tragic losses — of their own making.

UPDATE: 01/28/04 3:48pm
Frontline has a Presidential Futures Market, albeit one where you do not have to put any money at risk, thus removing what little predictive powers it might have had. Hey everybody, its an internet poll!

Source:
via Junkie Wire

Iowa Electronic Markets

Site List

2000 – 2003 Rally Comparisons, Excel (July 2003) Download file