Greetings. I see a lot of blog comments where people argue back and forth about Intrade prices for Obama and McCain. My diary, cross-posted from my blog Heartless and Brainless, explains why this is pointless. I'm having trouble getting some of the bells and whistles to work so please check out the original post for added content. There something for market rookies and savvy veterans. Join me below the fold, if you will.
Intrade provides a market for people to predict the presidential election. It operates on a scale of 0 to 100. You get nothing if your candidate loses and you profit if your candidate wins. For example, let's say Obama is priced at 60 and McCain is priced at 40. (We're assuming you put down $100 to make the math much easier.) If Obama wins, you would have a profit of 40. If the underdog McCain wins, he would gain 100 minus where you bought his "stock."
This is further complicated by profit-taking. Let's say you expect Obama to win a close election. You could buy McCain stock when it was down at 38% and sell right after his convention at 50%. That would give you a profit of $12.
It's based on the Iowa Electronic Markets (Go Hawks!) and the wisdom of crowds. Anyone who watches cable news knows that any buffoon can spout an opinion. They don't even have to believe what they're saying as long as they advance their agenda. Calvin Trillin had a point when he called them the Sabbath gasbags. The Iowa professors assumed that people would be accountable when they have money on the line. Prediction markets are also much more valuable in the long term, especially 100 days in advance or more. Other markets like Intrade went online with the same principle.
"What's your point, Guancous? Do you denigrate Intrade because they're competing with your alma mater?"
First of all, that's really creepy. I didn't type that. How do you carry a conversation in the middle of my blog post? Anyway, I have no problem with the market model. My problem is with blog commenters. (Not you, feel free to leave a comment. We won't bite unless requested and then you need to pay for the full hour.) You know what I'm talking about. You'll be on a great political site like FiveThirtyEight.com and someone will say, "Oh noes! Obama fell from 55 to 49 on Intrade!" or "McCain passed Obama on Intrade. Victory is ours, [homosexual slur redacted]." There are several reasons why this is wrong.
Fake Reason 1: I was going to say that Intrade has a self-selecting bias. I mean, of course Republicans are going to favor free markets! That's what their second religion is based on! Then I realized that Democrats were way, way more passionate about this race until a couple weeks ago. Obama held a large lead since the end of the primaries. Let me try that again.
Reason #1: Prediction markets are a tool like a Doppler radar. They are supposed to tell you if it will rain on Wednesday. Too many people use it instead of poking their head out the window. If a candidate is winning 51-49 with two months to go, this tells you nothing except that the race is close. The nature of these markets lead to blowouts since there is no profit from losing. Many rats leave the sinking ship at the first sign of trouble. Let's say for example that Sarah Palin tells a reporter that selling condoms should be a capitol offense. You would see prices like Obama 93 - McCain 7 in a couple of hours. The current leader isn't as important as the spread until some gamechangers like the debates. That leads me to my next point.
Reason #2: Intrade experiences severe jet lag. Due to some authoritarian gambling rules, there is a roundabout method to joining Intrade. It only takes a few minutes to register but you have to physically mail a check. Quaint, I know. (Remember when you could pay for pizza with a check? It makes me feel old to remember that and I'm not even 30 yet!) That means that's you'll waiting for the snail mail to reach another country. Let's say you're an evangelical Christian who planned on sitting Election Day out. You flip past the convention and Palin gets your heart racing like that Michelle Malkin picture you used to keep under your mattress. (Intern this, big boy!) You ship off a check to Intrade so that you can profit on the inevitable McCain victory. One week later you make your purchases and hop on the internet to crow about how the prices changed. Let's see, a week after the convention would be right about now, September 11th, 2008. McCain's rise in the market simply coincides with a delayed convention bounce. Now, allow me to adjust my tin foil hat as I move on to...
Reason #3: Shenanigans. You could swing a couple points in a close election by yourself. This New York Times article says there was $50 million in Intrade market volume all the way back in February. It seems insane that anyone would spend that type of cash to make a minor point. The glory of a tied election is that you don't have to. A campaign could buy $1,000,000 to move the prices a couple of points. Some Obama shareholders would sell to take a profit and some people would buy McCain to take advantage of the run. A million dollars seems risky but it's not like the McCain campaign would lose the money. Let's say they try to corner the market and move the price from 48% to 50%. Let's assume they fail and it moves to 45%. That means the campaign is only out $30,000 plus fees. That costs about one second of an Olympic ad, not counting the potential upside of changing the narrative. There is absolutely no proof this has ever happened and I don't quite buy it myself. That money would be much better spent on Diebold machines.
Reason #4: I'm absolutely full of shit. One of my favorite writers, Nassim Nicholas Taleb, wrote a book called Fooled By Randomness. (After reading this, check out The Black Swan. It will add another dimension to your thinking.) People tend to attribute random events to a variable and a narrative instead of realizing there can be a complex web of contradictory reasons guiding the outcome. This is the most likely. That's why I used the analogy of the Doppler Radar in Reason #1. You can have a million dollar state of the art machine but it will only tell you the most likely outcome. Someone could give a hog farm in Tulsa some spoiled feed and the resulting methane cloud could cause acid rain in Bolivia. That sounds silly because there are only one in a billion odds of that happening. Multiply that by the millions and millions of random-ass situations that happen every day and then you start carrying around an umbrella in your backpack, even when it's sunny. You can't know everything, no matter how much you study and prepare. That's why I love fantasy baseball. No matter how much research I do, current events always manage to find a way to keep me humble.
Prediction markets are much more accurate than polling the farther you get from the election. That's because more knowledgeable people are willing to bet on an election 24 months away than your average voter. They are also a lot more likely to bet on a sure thing. These predictions markets tell you the direction you're moving, not your likely destination. They're fun and more intuitive than a regular poll. Feel free to continue the discussion in the comments.