The Market for Bush

I don't know what to make of all those contradictory polls, but Bush has been cleaning Kerry's clock in the Iowa Electronic Markets. Salon's Farhad Manjoo has an interesting analysis of why that might be so.

Comments

[Anon]

The fine folks at Crooked Timber have chewed over the IEM several times. I believe Daniel Davies remains unconvinced that these markets really reveal anything, and I must agree. This is one of the more technical posts, but I found his comments about Surowiecki and Hanson (even if Hanson is mischaracterized) useful. The question is really _if_ you believe information is being usefully aggregated as well as _how_ it is being aggregated.
http://www.crookedtimber.org/archives/002460.html

A more recent post on the matter discusses these issues some more:
http://www.crookedtimber.org/archives/002560.html

While Manjoo's thoughts are interesting (I haven't read the earlier posts linked in the article), it feels too much to me like a rationalization of the IEM rather than an explanation of what is happening there.

Anon

[thoreau]

I've thought about playing this market. I wouldn't be confident enough to buy commodities on the vote margin, but I am predicting a popular vote victory for Bush. (Some poster will now undoubtedly point out that the popular vote is irrelevant under the Constitution, and that is true, but the popular vote is what this market is paying out for.)

Face it, on the dominant issue of this campaign (the war in Iraq), Bush has more consistently and forcefully articulated his message, and his message is both popular and reassuring: This is a vital part of the war on terror, we're making America safer, and a very evil and dangerous man has been removed from power. You might not agree with that assessment, or you might think he's performed poorly in pursuing those objectives, but that's his message, and it's been articulated consistently without deviation. Kerry, on the other hand, has basically said that he'd try to do a better job, but he has failed to articulate a consistent critique (yes, joe, I know, when you put it all in context and iron out the nuances it is consistent, but the message certainly doesn't come across that way at first glance).

Maybe I'll buy a few shares of Bush.

[joe]

Isn't it possible that people who buy shares in a political futures market, being both interested in financial markets and interested in politics, are likely to be politically conservative, and thus prone to having their opinions about the election skewed towards the GOP candidate?

[kmw]

I would posit that money is more important to the futures investor than person ideologies. If datum indicates there?s money to me made, they will jump on it, whoever is ahead.

No matter how much you like the Tampa Bay Devilrays, you?re not going to put money on them in March for a World Series victory.

[Curious]

How good a job did this market do in the past few elections at predicting winners?

[kmw]

me=be.

Sorry.

[joe]

kmw, but this is a judgement call. I'm not suggesting that they're deliberately ignoring reality, but that they're perceiving that reality through filters, as we all do. Finding Bush's arguments more compelling, and Kerry's less, becasue they agree with Bush more. But in this case, there are probably more people with Republican filters than Democratic filters.

[thoreau]

joe-

It's certainly possible that they are biased. But keep in mind that this is REAL money at stake. That forces them to ask themselves whether they really think their favorite can pull it off or not. Speaking as a former Democrat, I know what it's like to look at my preferred candidate and think to myself "Man, we're screwed!"

Even now, I still think Kerry is the lesser evil, and I'm watching him campaign and thinking to myself "Man, we're screwed!" I realize that he's been a strong finisher in races in Massachusettes, but this is a very different audience that he has to appeal to now, and it's an audience biased in favor of the Republicans (look at the vast swaths of red on the 2000 map, and remember that states with small populations have disproportionate electoral votes).

I just don't see how Kerry can pull this one out. I hope I'm wrong, but I just don't see it.

[JAG]

Not only is Kerry toast, he's going to drag Democrats like Daschle down with him.

Kerry is an empty suit. Ever fiber of his being reeks of opportunism and virtually every step in his life can be accurately viewed through that prism.

Rather than nominate a Dean who at least had the virtue of being a representative Democrat (unapologetically liberal) Democrats chose Kerry. Why? Because he made the best "appearence" of being acceptible to mainstream America. Everyone knows Kerry is just as liberal as Dean but they bet on him being able to use the war hero act as cover.

Now that this ploy has been blown apart Kerry has to run on his record (nothing there) and ideas (none he can "dislose" before he's elected.

Its going to be a huge loss and one that the delusional left very richly deserves since they have orchestrated it entirely on their own.

[Jose Ortega y Gasset]

I agree with Thoreau. At best, you could speculate that wealthy Republicans were buying Bush shares with utter disregard for the financial consequences... simply to create positive press for Bush. Despite the cries of outrage, the terrorism futures market was actually an interesting idea. I trust speculators and traders in an active marketplace to develop better information than moribund government intelligence agencies.

[joe]

Exhibit A...

[kmw]

JAG,

As a former Dean man myself, I'm finding myself perplexed as to why the Democratic leadership/delegates insisted on putting Kerry in to face Bush. The American Prospect blog has an entry on this very thing today.

I'm quite certain that Dean wouldn't be doing any worse than Kerry at this point, and there's a slight chance he'd be doing better. But I guess we'll never really know now.

[Douglas Fletcher]

Okay, I'll bite:

"A" is for August
a terrible month of Kerry?

[c]

buying a future contract on somebody is not the same thing as voting for them (or telling a pollster that you will) -- voting for them says you want them to win, buying a future contract says you think they will. This is a point perhaps lost on JAG, who thinks arguments about who SHOULD win have anything to do with this conversation.

So, how to explain Bush's larger lead in the markets than in the polls? I dunno, maybe democrats are pessimists.

[c]

or, perhaps the hot market in Bush futures indicates a heavy hedge bet, to cover one's losses in case GWB ends up with the national checkbook for 4 more (expensive) years.

[Jason Ligon]

Bush Bubble?

[isildur]

Interestingly, the vote-share market at IEM is predicting a Bush popular-vote win at 51%. I guess the results market folks are expecting him to pick up a few more points before the election. The vote-share market seems likely to be more resistant to partisanship, given the very specific results one is investing in.

[M1EK]

kmw:

The problem with your theory is that Vegas knows better. Why do you think the Vegas line moves?

People are MORE likely to be realistic rather than idealistic when money is on the line, but they still tend to be more idealistic than a perfect market would predict.

[gaius marius]

people who believe markets are crystal balls are naked utopians. folks, any time someone tells you that you can predict the future with what you know today, they are trying to sell you a bill of goods (possibly one that they bought themselves and now desire your company in holding to sate their insecurity). that doesn't change just because they're trying to sell you market ideology.

markets are not deterministic; they are probablistic, and often not even that. they are guesswork based on extant information. there is high noise:signal ratio in their volatility. even if they are functioning relatively rationally, new information can come along at any moment that can dislocate the market completely, rendering all prior pricing null. but markets rarely rational -- they are highly emotional by nature, subject to manias and all the vagaries of crowd psychology.

[Rob]

I can't speak for the IEM population as a whole, but I know a fair number of financial people who play around on that exchange, and they're all pretty damn conservative. And they view it with an attitude somewhere between buying squares in a Superbowl pool, and running a fantasy baseball team. They are not performing heavy-duty analysis to try to get an edge like they would in a trading book with "real" money.

IMO, we can learn roughly the same amount about the likely outcome of the election by observing whether the bratwurst or the hotdog won the race during 7th inning stretch today.

If the exchange would drastically increase the account sizes, then big political contributors could try to hedge their bets in this market. That might tell us something interesting about the likely outcome of the election...

[Bill]

I think both Bush and Kerry are dickheads, and I won't be voting for either of them, but I'd prefer Bush to win. Why? Because is live in San Francisco and I enjoy schadenfreude--I want to see my local lefty population suffer. Childish? Yes, but since I think lefties are little better than enslavers, so my wanting them to suffer isn't so strange.

Bush is leading in the markets because, at this point, he's more likely to win. If some event causes him to lose support, it will be reflected in these same markets.

[isildur]

M1EK:

Vegas spends a lot of time, effort, and money convincing you to be irrational with your money. There's more practical psychology at work in one casino than in any given major university's psych department. Vegas doesn't really work as a counterexample, as a consequence. Only the savviest gamblers can resist the lure of comped drinks and cute girls at the roulette table, when they're winning at blackjack.

Which is not to say that I think this is a magical peek into the future. But at the very least, markets are demonstrably more reliable predictive tools than, say, an opinion columnist's latest rant.

[Steve in CO]

This is all horseshit. Here's why. There really isn't anyway to measure the "haven't ever voted before, but I'm gonna this year," category. I think the numbers are going to surprise even the most jaded pundits.

[Dan]

The markets don't seem that out of whack to me. Not because I'm convinced Bush will win the actual election -- if I was convinced of that, I'd invest in the IEM myself -- but because with his current poll numbers it seems perfectly reasonable to say that there's a 70+% chance of him winning the popular vote.

Polls measure popular opinion, popular opinion determines the popular vote, and it takes time for popular opinion to shift. Inasmuch as those three statements are true, the following is also true: each day that we get closer to the actual election, the odds of Bush winning the popular vote improve. This is true simply because Kerry has progressively less and less time to shift public opinion in his favor.

So we would expect Bush's "stock" to steadily rise for as long as he maintains a lead over Kerry in the polls, even the lead itself isn't increasing.

[thoreau]

Dan-

Good point. I suspect that on Nov. 1, or whenever they close trading, a contract for a Bush victory (irrespective of the margin) will be trading at close to $1, since the payout per contract is $1 if the candidate wins.

When people say that these markets are good predictors of outcomes, I wonder when they make the measurement. On the day before the election most people who read the newspaper can probably predict the outcome with decent accuracy. Sure, that sort of consensus would be wrong now and then, but we're talking about an accuracy rate measured over many elections. On the other hand, a month before a close election the outcome can be less clear. So I wonder where they take their baseline for saying that this market has good predictive power.

[Brian Cook]

joe, to address your question of potential bias in the IEM: I don't think that explanation flies. if you look at the chart it was relatively even until about a month ago when the lines start to diverge. So either there was a huge influx of extremely conservatively-biased money over the past month... or Kerry had an awful month. I'd venture to guess that almost all the money in the IEM now has been there for a while (especially given the relative inconvenience of acquiring shares). And I'm willing to assert that Kerry did have an awful month.

I don't know how much predictive value this thing has, but it's at least an interesting look at how people think the race will end up at any particular point in time. It will be interesting to see how it reacts tomorrow after the debate, and then in two days when the media dissects the debate.

[McClain]

CO Steve:
...'the "haven't ever voted before, but I'm gonna this year," category.'

I suspect you're imagining an unprecedented show of voting force from the anti-Bush alterna/slacker crowd.
But what about the "haven't ever voted REPUBLICAN before, but I'm gonna this year," category?
(that would include me, yeah)

And what's with all these so-called "Libertarians" who all-of-a-sudden have grave doubts about Markets? Hmmmn?

[Jason Bourne]

kmw,

People make boneheaded decisions about betting all the time.

[Gadfly]

Thoreau, here's an article from Nov 2, 2000 showing Bush with a 47 to 43 lead with 5% undecided, a 2% margin of error and Nader on almost every state ballot. Gore wound up winning the popular vote by a half million. True, Bush "won" but he lost his lead in actual votes.

A newspaper reader on that day got lucky if they predicted Bush winning on the basis of the popular vote.
http://www.cnn.com/2000/ALLPOLITICS/stories/11/02/tracking.poll/

[swami]

Look at http://www.tradesports.com, It is a more liquid market and larger than IEM. Bush is in the high 60's to win the election. They also have individual state markets, but much less volume in those. The market looks pretty accurate to me. I think it might tighten towards election day and Kerry might win, but the price of Kerry at around 2-1 seems about right.

[JDM]

"folks, any time someone tells you that you can predict the future with what you know today, they are trying to sell you a bill of goods "

I'll keep that in mind next time you explain how obvious it is that we are headingin to fascism.

[joe]

It is possible for a data source to accurately track trends, even if the underlying absolute numbers are bunk. If the market starts moving in one direction or the other, or if it fails to move over a period of time, that's probably a pretty reliable indicator that the movement is (or is not) happening in the real electorate.

[Dan]

When people say that these markets are good predictors of outcomes, I wonder when they make the measurement.

Well, judging from the chart linked above, the IEM has been "predicting" a Bush victory since well before it was conventional wisdom that Bush was likely to win. I'm skeptical of the predictive value of such a tiny, specialized market, though.

[gaius marius]

it pays to mention that, regardless of what actually happens, there's a measurable probability that the debate series will reverse the fortunes of the candidates -- as has happened in a number of close races past.

argue all you like that such probabilities are priced in -- but if dubya gaffes horribly or kerry comes away looking like solon, these futures are in for a hell of a price dislocation that would immediately render transparent any utopian dreaming about "crystal ball contracts" as the utopian market-religion nonsense it is.

[Patri Friedman]

The problem with IEM is that you can only deposit $500. So people's opinions are weighted equally, even though some people may have much higher knowledge and confidence, and would be willing to bet much larger amounts. For this reason, I think Tradesports is likely to have more accurate prices. I've blogged about this difference on Catallarchy.

[JAG]

"This is a point perhaps lost on JAG, who thinks arguments about who SHOULD win have anything to do with this conversation."

Uh, "c" sorry for offending you by digressing from this critical train of thought!

I'll never, ever, ever do that again massa, I promise!

[Dan]

it pays to mention that, regardless of what actually happens, there's a measurable probability that the debate series will reverse the fortunes of the candidates

Ok, then -- what's the probability? Currently the IEM estimates that Kerry has a 32% chance of managing to pull ahead of Bush and win the popular vote. Are you claiming that the chance of Kerry pulling ahead is more than 32%? Less than 32%? What is it?

argue all you like that such probabilities are priced in -- but if dubya gaffes horribly or kerry comes away looking like solon, these futures are in for a hell of a price dislocation

You obviously don't comprehend what "probability" is. The fact that we say Bush is the *probable* winner means that we acknowledge that something might happen to cost Bush the lead. Yes, Kerry might pull something off; Bush might screw up. That's why the IEM predicts a 32% chance of Kerry winning.

that would immediately render transparent any utopian dreaming about "crystal ball contracts" as the utopian market-religion nonsense it is

Let's say that I wanted to bet on what the chances were that the majority of three coin flips would be heads. The odds start at 50%. A coin is flipped, and lands heads. The odds of getting mostly heads are now 75%.

Whereupon some pseudonymous would-be philosopher proclaims "Ah, but you have ignored the possibility that you might flip TWO TAILS IN A ROW. If that happens, it will expose your utopian probabilistic caculations for the utopian mathematical dreaming that they are!" The coin is flipped again -- tails. The P.W-B.P sneeringly observes that the calculated odds are now 50%, which "obviously" puts the lie to the earlier 75% claim. Whereupon the mathematically educated among us enjoy a quiet chuckle at his expense.

You're acting as if the IEM is saying Bush is definitely going to win. If the IEM thought Bush was definitely going to win, Bush shares would cost $1 and Kerry shares would be free.

What the market is actually saying is that Bush is probably going to win -- specifically, that there's a 65% chance, as of today, of Bush winning in November. If new data becomes available, obviously the prices (and the probabilities) will shift. You're not saying anything here that we didn't already know.

Summary: the IEM says Bush will probably win. You say it's possible he won't. One word: duh.