Why a Romney Win Would Shock the Markets
Presumptive Republican presidential nominee and former Massachusetts Gov. Mitt Romney speaks during a campaign stop. Photo by Edward Linsmier via Getty Images.
As of yesterday, President Obama had already been re-elected. Or so it seemed to Paddy Power, the Irish online betting site that paid out more than £400,000 ($750,000) on Monday to Obama bettors. According to IrishCentral, the company admitted it was sticking its neck out. "Romney gave it a good shot and is doing well in the popular vote," said a spokeswomen, "but we suspect he's had his moment in the sun and is likely to be remembered more for his legendary gaffes than presidential potential."
The London Evening Standard, wrote: "Paddy Power -- which has made a habit of publicity-grabbing early payouts -- said: 'The overall betting trend has shown one-way traffic for Obama and punters seemed to have called it 100 percent correct.'" Still, despite its penchant for publicity, if Romney were to win Tuesday, Paddy Power would have to pay another half-a-million dollars to Romney bettors, so it is presumably confident.
And most other online bookies are similarly sure of an Obama victory. Scanning the list of UK bookmakers showed average odds for Mitt Romney late last night, when I went to bed, at less than 25 percent. To bet on President Obama, the most common odds were that you had to put down £9 to win £2.
This raises two questions. First, how come the polls are so close? Second, how come Intrade, the one major online market that takes US bets -- in dollars -- had Romney at fully 33 percent last night?
Let's consider the polls. I remind you, to start, that in polls you can be close, but still get no cigar. It's all or nothing. So an ironclad 51-49 percent advantage in polls would translate into an overwhelming likelihood of winning.
Moreover, national polls mean little. A president gets chosen by the electoral college and here, even in the polls, Obama seems decisively ahead in the key states, most notably Ohio. That's why Nate Silver of fivethirtyeight.com, who studies every poll, puts the odds of an Obama victory at 91.6 percent as of 7 a.m. this morning. Over at Professor Sam Wang's Princeton Election Consortium, the percentage has been above 98 percent for awhile. Wang's site proudly points out that in 2008, it came within one electoral vote of exactly predicting the outcome, a better result than any rival forecaster. Okay, but if prediction markets are more reliable than polls, as many an economist would argue, how come Intrade varies so significantly from every other market?
We have mentioned allegations of manipulation here and here -- charges that Romney backers have been boosting the odds for their candidate on Intrade, the one market where Americans can put down real money, in an effort to bolster enthusiasm for him. Nate Silver explored the issue on Oct. 23. The evidence is hardly conclusive, however. A very optimistic (and well-heeled) Romney supporter could simply keep thinking his man is going to win and bet that way.
For those of you interested in more venal matters, as in "can I make money from this market discrepancy?", here's Portuguese MIT MBA student Andre Gloria, whom we featured in our broadcast story on markets versus polls and in the "how to make money" Making Sen$e post linked to above. Andre sent this email yesterday when I asked if he were again betting on the arbitrage between Intrade and other markets:
"I put on two extra bets on Saturday because the spread did become quite wide again between non-U.S. sites and Intrade. It is even better now, and I am waiting for funds to clear before I can continue the party. Basically, non-U.S. sites are again underestimating Romney odds when compared to Intrade.
"As an example, 5dimes (a betting site) is offering Romney at +300 (a 25 percent implied probability) [Monday morning] with Intrade at 34 percent.
"Betting £500 on 5Dimes gives a £1500 profit if Romney wins and £500 loss if he loses. Betting £1000 on Obama on Intrade gives about a £500 profit if Obama wins or a £1000 loss if he loses. Overall this nets a £500 riskless profit if Romney wins. Let's call this part of the trade the 'emotional hedge': if Romney wins, at least I'll make some pocket money."
The problem with Andre's strategy thus far, of course, is that he makes nothing if President Obama is reelected. That is, he loses £500 pounds on his Romney bet, wins back £500 on Obama. So he has to put more money on Obama to make money on this "arbitrage." Accordingly, he continued:
"What I plan to do Election Day is start using the riskless profit I will make to bet on Obama. So, for example, I'll bet £370 on Obama to win approximately £123 if he wins. This will make a riskless profit either way: +£123 if Obama wins or +£130 if Romney does (Andre's £500 Romney winnings minus the £370 he has just bet on Obama).
"To put things in perspective, the net position is approximately 7 percent sure profit on a one-day bet (a profit of £123 on an £1870 investment) 9£500 on Romney, £1000 on Obama, and another £370 on Obama)."
Unfortunately for Andre, Obama has become even more of a favorite on most betting sites this morning. Andre may have waited too long to make a 2555 percent annual return (7 percent a day multiplied by 365 days). But he'll still make something.
We at Making Sen$e are working on a story to explain where the monthly unemployment numbers come from. To do so, we are looking for interviewees who have worked on the Current Population Survey (CPS; household survey) and/or the Current Employment Statistics survey (CES; establishment survey). Are you a former surveyor? Do you know one? If so, we want to hear from you! Please email us at firstname.lastname@example.org. Be sure to include your contact information. Very much obliged.