The Iowa Electronic Markets are still going for Bush

Election-predicting traders are betting not just on a narrow Republican victory, but a landslide. What's behind the Bush bubble?

By Farhad Manjoo

Sep 29, 2004 | It might seem difficult to recall now, but just over a month ago, in the middle of August, John Kerry's star shone the brightest in the political firmament -- polls had had him up against George W. Bush in national races, he was ahead in key battleground states, and commentators, impressed by his acceptance speech at the Democratic Convention, were declaring the race "John Kerry's contest to lose." The zeitgeist was decidedly on Kerry's side; only one stubborn measure, the price of Kerry shares on the Iowa Electronic Markets, seemed to point to the possibility of the senator's loss in November.

The IEM is a real-money, Web-based system on which anonymous traders buy and sell contracts that offer payouts based on actual events. On the IEM's presidential election market -- which I've been playing since June -- you can exchange shares of Kerry and George W. Bush according to your guesses about how each man will do on Election Day; the price of each candidate's stock reflects the traders' collective forecast of who will win in November, a forecast that, in the past, has proven quite accurate. As I wrote in August, it seemed rather curious that Kerry's lead in the polls and among the pundits was not reflected in his stock price: Why were the IEM's traders not captivated by Kerry's speech, as was the rest of the nation? Did they know something that the pundits did not?

Looking back, it seems they did. Perhaps Kerry's low stock price did not point out the specific problems that led to his subsequent decline in the polls -- the numbers didn't lead us to watch out for attacks on Kerry's Vietnam service, for instance -- but you can make the case that the mid-August market indicated real uncertainty about Kerry's ability to run a winning campaign, uncertainty that now seems thoroughly justified. Traders on the IEM were not clouded by emotion, by a desire for a Kerry presidency, by pat political theories having to do with perceived Kerry "momentum" -- instead, when they looked at Kerry they saw a tough rough road ahead, and while the rest of the political establishment swooned, they stayed in their seats. And, so far, they've been right.

As I've written, I'm supporting Kerry for the White House, but on the IEM, I've been betting on Bush. This is not meant as a slight against Kerry -- it's just that I have doubts about his ability as a campaigner, and more important, I think that the Bushies are extremely good, not to mention that historically, incumbents running during a time of economic growth tend to win reelection. So far, alas, my bet is panning out, as measured not just by Kerry's shortfall in the polls but also by the current performance of Kerry shares on the market. As of Monday, traders on the IEM give Bush about a 70 percent chance of winning in November. To look at the graph of the IEM is to despair over the prospect of four more years -- a confident red line aims straight for the stratosphere, while a morose blue line nosedives.

For the first time since I started playing the market, though, I must say that I'm dubious of the IEM's high valuation of Bush. Here's why: On Sept. 21, the gurus who run the IEM split the two main share contracts -- one Bush contract, one Kerry contract -- into four share contracts. Now, instead of betting either that Kerry will win or that Bush will win, you've got to also guess what percentage of the popular vote the candidate will get. You can buy a Bush contract that will pay you off only if Bush receives more than 52 percent of what's called the "two-party popular vote," which is a measure of the popular vote that excludes third-party candidates such as Ralph Nader; or you can buy a contract that will pay you off only if Bush wins between 50 and 52 percent of the two-party popular vote. Similarly, you can purchase a Kerry contract that will pay you off if Kerry wins between 50 and 52 percent of the vote, or one that will pay off if he wins more than 52 percent. The IEM trader is thus asked to forecast both the winner of the election as well as the size of his mandate: Who will take the White House in November? And will he win big?

What's happened since the stock split seems strange. Share prices corresponding to either a small Bush win or small Kerry win have remained relatively low, while shares for contracts that say that Bush will win big have skyrocketed. On Monday, a Kerry small-win share was selling for about 22 cents; a share corresponding to a small Bush win was priced just a bit lower, about 21 cents. A share saying that Kerry would win big was selling for about 7 cents. And then the big one -- shares for contracts saying Bush would win by more than 52 percent of the vote were selling for 51 cents, and that price has been rising over the past few days. What this means is that according to the collective wisdom of traders on the IEM, there's a better than even chance that Bush will win in a landslide on Election Day.

To me, these numbers seem off. Everything we know about this election says it's going to be close. In the last two weeks, only Gallup's polls have shown Bush leading Kerry by more than a three- or four-point margin, and those surveys have been criticized for possibly oversampling Republicans in their calculation of how many voters are "likely" to come to the polls on Election Day. Just about every other national poll shows either a dead tie or Bush with a tiny lead.

Of course, it seems possible that, despite the polls, Bush could win more than 52 percent of the vote; nobody would be astonished if that occurred. But the IEM's traders aren't saying that this outcome is merely possible -- they're saying that Bush's winning big is the most likely outcome in the presidential race, more than twice as likely as the prospect of Bush's winning small. Bizarrely, the traders believe it is more likely that Kerry will win the race than that Bush will win with a small mandate. If Bush wins at all, they're saying, he will win big.

To figure out the IEM traders' reasons for betting on a Bush landslide, I thought it might be a good idea to look back at the history of presidential elections in America and count the number of elections in which candidates won big. Using raw data I found on the trusty Web, I constructed a spreadsheet to calculate the two-party popular-vote share of all the Republican and Democratic presidential candidates since 1900 -- that is, the vote share of the candidates that eliminates the effect of third-party candidates. What I found surprised me: It turns out that of the 26 presidential elections conducted in America since 1900, in the vast majority -- 21 -- the winning candidate has received more than 52 percent of the two-party vote. Bill Clinton did it in 1992 and 1996. George H.W. Bush did it in 1988. Ronald Reagan did it twice. Richard Nixon did it in 1972. Indeed, the only presidents since 1900 who failed to win more than 52 percent of the two-party vote were Woodrow Wilson in 1916, John F. Kennedy in 1960, Richard Nixon in 1968, Jimmy Carter in 1976, and George W. Bush in 2000. In other words, winning big is the norm, and close elections are the exception.

Next Page: Is the IEM biased toward Republicans?

• View blog reactions