The Political Gamblers

Have you ever seen a basketball player make a head-fake so obvious that you couldn’t believe the opposing player was fooled by it?

That’s what happened to political gamblers Tuesday night at when the first returns reported from Texas gave Barack Obama a 55% to 45% lead over Hillary Clinton. The price of the bet on Obama to win the Texas primary jumped from 83 to nearly 90 even though those returns represented well under 1% of the precincts and despite exit polls showing Hillary likely to edge out a narrow victory.

Over the course of the evening, as Hillary’s position improved steadily, the price of the Obama bet dropped steadily (and the price of the Hillary bet rose accordingly). With about 40% of the vote counted, however, even though Hillary was showing about a 1.5% lead in the vote, bettors were still trading Obama at around 52% to win the contest. 

I had sold Obama as the nominee of the party at over 84% the prior day and didn’t really need to bet more, but that 52% seemed just silly given what was happening…almost like the same head-fake from earlier in the evening…so I sold a very small amount of Obama to win Texas at 51.7%. (Selling a position you don’t already own, with a view toward buying it back lower later is called short-selling or shorting. It’s simply the reverse order of most people’s investing strategy of buying low first and selling higher later.)

(The way the betting works is that 1 point in the price is worth 10 cents and contracts settle at 100, representing $10, or zero. So, selling/shorting one contract at 51.7 means I was risking $4.83 to try to make $5.17, before intrade’s commission which is a percentage of your winnings.)

Even late into the evening when Hillary’s victory seemed almost certain, Obama was still trading around 25%. Thinking about these apparently silly prices, i.e. Obama trading “too high” all night, I reached a fairly obvious but still important conclusion: Participants in internet political gambling are not representative of the population. In fact, one could have anticipated the “stupid” betting on Obama simply by remembering that Obama does much better than Clinton among younger people and wealthier people, both of whom are more likely than older or poorer people to be participating in such betting.


Now that we don’t have another major primary until Pennsylvania on April 22, the major betting action will be on who will be the Democratic nominee and which party will win the White House in November. In early Wednesday betting on the Pennsylvania primary itself, Clinton was trading about 70% and Obama about 33%.

[While it would be “free money” to sell the Obama-Clinton pair for over 100, since the maximum final value is 100 for the winner and zero for the loser, they add up to more than 100 for now because the markets are illiquid until more people start betting. I couldn’t pass up the opportunity, however, and was able to sell Obama and Clinton for a combined price of 113.2, meaning I’ll make 1.32 on each “pair.” That large an arbitrage opportunity usually doesn’t last long, and it didn’t.]


Here is what’s happened to the price of Obama to be eventual nominee over the 72 hours ending about 9 AM Eastern Wednesday morning. For several days going into yesterday’s primaries, Obama traded about 86%. When the exit polls started coming out of Ohio and especially Texas, his price dropped to about 80% going into the first reporting of results. As mentioned earlier, the first results were a head-fake, but gamblers bit on it, taking Obama as nominee (not just to win Texas) up to about 88%.  Then the reality of Texas set in. Over the course of the results coming in, Obama traded down to about 74% and has settled in around 72.5 after trading as low as 70. (I covered 3 of my 5 short Obama at 70.3.)

The other continuing interesting piece of betting on our election is which party will eventually win the White House. In the past 72 hours, the price of the GOP to win the general election has climbed steadily upwards from about 35.5% (where I bought some last week) to about 38%, after briefly trading about 39.5%. (Note, graph shows last trade of 37.1, but the contract is 37.1 bid, 38.9 offered, so 38 is “mid-market.”)

It is not surprising that the GOP’s overall chances are perceived to have improved with yesterday’s results. First, Hillary’s chances of being the nominee doubled, from 15% to 30%, and many people believe Hillary is more beatable than Obama (although that changes if you believe she would have Obama as her running mate.) Second, continued divisiveness within the Democratic Party increases the chance of a divided convention and general turmoil for them as well as keeping Hillary in the race, most likely all the way to the convention, therefore causing both Democrats to spend most of their energy and tremendous amounts of money attacking each other while John McCain sits back with his patented big grin, raising money and watching his two potential opponents drag each other down.

I bought the GOP at about 35.5% not necessarily because I believe McCain will win (though I believe he can), but because I think the price will trade at least in the mid-40s, and possibly over 50, before election day. Tuesday’s results make me even more confident in that bet.


There appears to be more “stupid money” in these markets than there has been in the past. As noted above, this sort of activity is not one whose participants are a representative sample of the overall population. In particular, they are younger than the average American and therefore maybe more likely not to have the wisdom that would keep them from making stupid bets based on less than 1% of returns coming in. This means that there can be opportunity for someone who can take a more sober view of things, taking the other side of short-term emotional bettors.

It also means that the academic and predictive value of internet political gambling might have decreased rather than increased in the past year or two despite the increased number of participants. If the early players (such as in the Iowa Electronic Markets) were just a few political junkies, the average of their opinions, as expressed in the prices of the futures contracts, tended to be better predictors of actual outcomes than polls were. 

But when the flood of new participants is dominated by people who bet their emotions, i.e. people who bought Ron Paul up to 9% (I sold him at 8.7%) or who bought Mike Huckabee up to over 20%, it becomes very hard to say that political gambling sites offer good predictive value. Instead, they offer what more experienced political observers might think to be silly pricing, and therefore some profit opportunity for those willing to make a bet.