Joi Ito's Web

Joi Ito's conversation with the living web.

Graph of Bush vs Kerry on Iowa Electronic Markets
The Iowa Electronic Markets are real-money futures markets in which contract payoffs depend on economic and political events such as elections. These markets are operated by faculty at the University of Iowa Tippie College of Business as part of our research and teaching mission.
Jimbo told me about IEM when I met him in Linz. A paper (PDF) describes the past elections and how the markets have been amazingly good at predicting their outcomes. IEM has a current market quote which is updated every 15 minutes. As of this posting, it is 40.6%/59.4% Kerry/Bush.

On IRC, crw, just pointed out a blog post on The SaltwaterPizza that used Google to see how many people said they were voting for one candidate or the other. The sample size was 104,789. This gave 46.8%/53.2% Kerry/Bush.

The results are disappointing for those of us who are hoping Kerry will win, but these alternatives to traditional polls are very interesting none the less.


...and then they stop saying anything at all. :P

I was wondering about that back in June, and decided to be hopeful and take some of the votes away from Bush because I have no idea how many of the results are a carry-over from previous "bush" elections.

I hadn't checked in a while, but I see that the number of hits has really increased for both candidates since then. June averaged around 1,840 for Kerry and 2,370 for Bush. using only the "I'm voting for XXX" search

I guess the jump in bush's popularity of 2,330 hits since then is can not be from a the Bush / Gore race... and since June Kerry only improved by 1,800. Too bad.

Of, course, if we assume that all of Nader's 852 hits would otherwise go to Kerry, we can see that Nader will end up costing him the election

*sigh* Thank goodness, I didn't stop drinking like you did, Joi. I'll quit drinking after suffering four more years of crappy management.

Surowiecki talks aboutt his in Wisdom of the Crowds and I've been meaning to put further study into it. People put their money where their mouth is with spreads like this so this could be bad news.

The scuttlebutt from the camps, and you probably have better contacts than I do, is that the Kerry campaign has lost some momentum. the danger with lost momentum is that it is infectious. So we now believe Kerry has lost momentum, so he has.

If IEM is such a bellwether. And If we believe in a change of government. And If we need think our intentions are good. Can we skew the data in IEM so that we create our own virtuous circle? Ahhh the ethic of market-mandering.

Azeem, if you want to skew the IEM it probably isn't too hard because it's a pretty small academic market (although they put pretty low limits on how much money you can put in).

But, then you'd have to go against all of the other futures markets out there like or And if you're wrong, you lose all your money.

Interestingly there's an arbitrade opportunity right now between tradesports and IEM as bush has a 67-68% chance to win on TradeSports but just 60% on the IEM.

I wonder if George Soros is reading this.

There's an interesting (although quite technical) piece on the valuations of the candidates in this Crooked Timber posting

I should put the maximum buy into Bush winning then.

It'd be kinda a win-win thing on election day, plus providing my airfare on JAL to get me back to the land of the rising yen.

I first flew to Japan in Aug '92, coming back in May '00, so I missed all but the last few months of the clinton presidency.

I've been playing IEM for over ten years now, back since the days of Gopher and Telnet. It's actually quite interesting, and shifts quicker than the polls sometime. There are some skewers in it vs. a public interest poll. People will bet across party lines based on perception of victory, not necessarily their belief. And in some cases, IEM probably would have led to a bad result (take the VA governor race 10 years ago when an African-American was clobbering in the polls but faced a dramatically different result when people actually voted). Polls can be skewered on a variety of factors, and that leds to a greater shift in perception, which is in turned magnified by IEM type systems.

I'm happy as heck that Bush is leading (I'm a Republican) but I'm not buying him right now as I feel the poll race (and thus IEM) will shift in the next few months (polls always, always tighten in October) and thus, the price for a Bush win will drop before it rises in the final week of the race (they usually go into the 90s in the last week).

The votes are hidden in the anti-Bush slogans.

Jerome Armstrong found the missing Kerry votes.

The Iowa Markets predicted Dean big time.

2 TrackBacks

Listed below are links to blogs that reference this entry: Google and Iowa Electronic Markets say Bush.

TrackBack URL for this entry:

I used to play the IEM Joi Ito is talking about when I was on Capitol Hill, buying and selling election victories (and mainly defeats). I remember buying 100s of shares in the 1994 House of Representatives market in the... Read More

Hoje de manh Read More