Saturday, March 08, 2008

The Future May Turn Out to Be Futarchy

Have you ever read the blog, “The Conspiracy to Keep You Poor and Stupid”?. If not, you’re missing a real treat.

Mr. Luskin published a book by the same name, and from there grew his fascinating economics website.

I’m particularly fond of his viewpoint since he regularly bashes Paul Krugman, an “economist” with The New York Times who is infamous for his incorrect and often silly predictions regarding the economic health of America. Mr. Krugman is one of the shills for the “Hate America First” lemmings. Definitely, on the A team; if Krugman predicts thunder showers you can safely leave your umbrella at home.

The following is Mr. Luskin’s essay, which he posted at Townhall, and cross-posted on his own site:

They say every politician has a price. So you want to buy John McCain? Hillary Clinton? You can do it, online.

You can buy a futures contract betting that McCain, or Clinton -- or any other candidate -- is going to be the next president. Or you can bet on whether the GOP will recapture the House of Representatives. Or whether the US will bomb Iraq. Or capture Osama bin Laden. You can do it all in the brave new world of online prediction markets.

That kind of gamble may not be for you. But the market prices of political candidates and current events have remarkable predictive power. Academic studies have shown that they have a far better track record than traditional polling. And they ask questions that traditional pollsters would never think of.

Here is the current chart on the sidebar of Mr. Luskin’s site. This is posted here as a plain HTML table, but at the Conspiracy website, it’s a clickable widget that will send you to the original site, here.

US recession 0865.569.1
US attack N.Ko by 6/081.54.5
US-I bomb Iran by 6/085.79.8
Osama captured by 6/086.06.7
GOP House 20086.09.0
GOP Senate 20087.013.3
DEM >60 Senate seats5.08.9
GOP president 200838.138.4
Paul GOP nominee1.01.1
Paul third party7.516.9
McCain GOP nominee96.497.0
McCain president37.537.7
Huckabee GOP nominee0.80.9
Huckabee VP nominee5.56.9
DEM president61.162.0
Obama DEM nominee72.772.8
Obama president46.346.6
Obama VP nominee14.115.3
Clinton DEM nominee27.027.1
Clinton president17.017.3
Clinton VP nominee12.614.8

- - - - - - - - -
Mr Luskin again:

The prediction markets nailed the GOP sweep in 2004, when the polls were showing that John Kerry was going to win and that Republicans were going to lose ground in congress. This year they predicted Barack Obama’s surprise win in the Iowa caucuses. And before Saddam Hussein was captured, no one was taking polls on when people thought it would happen -- but the online futures markets predicted it. The internet has made modern prediction markets possible. But before modern scientific polling techniques were introduced in the 1940s, betting on political candidates was widespread in America. Bets were traded on stock exchanges, and were reported in newspapers every day just like polls are now.

The New York Times wrote in 1924, “Wall Street betting odds are ‘never wrong.’“ In fact, from 1886 to 1940 they correctly forecasted the outcome in every presidential election with the exception of one -- Woodrow Wilson in 1916.

Here’s how modern online prediction markets work. Let’s say you want to be that John McCain will be the next president. To start, you’d get online and go to Intrade, the Dublin, Ireland- based website that dominates prediction markets the way Ebay dominates auctions. Intrade CEO John Delaney told me that last year his exchange “traded about 65,000 markets, 1,200 operating at any given time, with more than 500 on politics. There’s a trade done every 8 seconds from traders in 128 countries.”

You’d click on the link that takes you to the 2008 US presidential markets, and you’d check the market for futures contracts on John McCain to be the next president. And like any futures market, the first thing you’d see is the price. For the sake of a simple example, let’s say the price was 40. What exactly does that mean?

Each political futures contract is worth $10 if you win and worth zero if you lose. So if you buy the McCain contract and he wins the election in November, your contract pays you $10. But if he loses, it’s worth zero and you get nothing. The quoted price of 40 means you have to pay 40% of $10 -- which of course is $4 -- for the contract. If he wins and you collect $10, then your profit was $6 (the difference between your price of $4 and your payoff of $10). If he loses, you’re out the $4.

The person on the other side of the trade -- who sold you the contract for $4 -- takes the exact opposite side of the bet. He loses $6 if McCain wins and he pockets the $4 you paid him if McCain loses.

That price means that the market is saying -- in a continuous, ever-shifting global referendum drawing on the collective knowledge of the entire world -- that John McCain has a 40% chance of winning. A price of 60 would mean he has a 60% chance. A price of 20, a 20% chance, and so on.

A traditional poll just measure how a sample of voters say that they themselves will vote if the election were today. But the election isn’t today. And who’s to say that poll respondents are being honest in the answers they give to pollsters?

A prediction market tells you what people willing to put their money at risk think will occur in the future, taking every possibility into account for what could happen between now and then. There’s no motive to lie, or to indulge in wishful thinking. There’s money at stake. You may want Hillary Clinton to win, but if you think John McCain will win, you’ll buy the McCain contract and you’ll sell the Clinton contract.

The financial incentive to get it right, and the ability to draw on bettors from around the world -- anyone who might have any information on whatever proposition is being bet on -- is what gives these markets their uncanny predictive power. Who do you suppose made the bets that Saddam Hussein was about to be captured, just before he actually was? Somebody knew something -- and these markets provided an incentive structure that drew their knowledge into the public domain.

Prediction markets have potentially far-reaching applications. Some corporations are using them to make product development and marketing decisions.

Economics professor Robin Hanson, who has studied prediction markets extensively, told me he envisions a “futarchy” -- government by futures contracts traded in a prediction market. Instead of traditional democratic representative political processes, under futarchy the prediction markets would be how the electorate “decides what to believe.”

For example, you could have a betting market on whether or not a particular tax cut will stimulate economic growth. If the market ends up giving you a high enough probability, say 75%, you’d go ahead and implement the tax cut. If it works, and growth picks up, the people who bet on it would will win (and the people who bet against is would lose).

Maybe that makes more sense than letting congress make those decisions. At least the bettors in the prediction markets would be directly accountable, so they’d have a very tangible motive to do the right thing.

Today prediction markets are threatening to replace political polling -- they’re certainly doing a better job. Tomorrow, who knows? Prediction markets might replace politics itself. They might do a better job at that, too.

From your pen to God’s plans, sir.

Imagine there’s a future
Where no politicians abide
Imagine there are no lobbyists
And no one left to bribe…


Derailed Cluetrain said...

Problem: at 1 trade every 8 seconds - for all their bets, even - it's an extreme low-volume market compared to the stock markets. Maybe this will change in the future, but today any campaign willing to sink an small part of its total cash could totally dominate the political prediction market and set any predictors it felt like for less than a TV ad.

Afonso Henriques said...


heroyalwhyness said...

From the article:
==>Instead of traditional democratic representative political processes, under futarchy the prediction markets would be how the electorate “decides what to believe.”<==

In essence, the deep pocket$ of Soros or Heinz-Kerry can shift electoral opinion through futarchy $$ manipulation?


Dymphna said...

Derailed cluetrain and herroyal whyness:

I though the same thing in the last election round. Soros was determined to beat Bush and he sank a great deal of his billionaire money into the effort. But he was never able to nudge the stats. Bush stayed ahead.

I don't know if there is a way to avoid the machinations of someone like Soros. He certainly damaged the British economy...otoh, it may be that speed of communication kept him from tilting the game.

One thing is for sure: it would make government transparent.

I think this is a trend we should look for in the future -- at this point, books, monographs, etc. But in a generation or two, who knows?