Can markets trump election polling?

By Jason Abbruzzese

Who will you (or would you) vote for in the 2012 presidential election? Now let’s say you have $10 to bet on a winner of the election. Is the answer the same?

Maybe, maybe not. What matters is that those are two different questions, and the results from the answers give us two very different sets of data. Polling has long been used as an implicit indicator of the likelihood of a certain outcome.

(AP) President Harry S. Truman holds up an Election Day edition of the Chicago Daily Tribune, which, based on early results, mistakenly announced "Dewey Defeats Truman" on Nov. 4, 1948.

So polling is not perfect, but what’s the alternative? For people looking to hedge positions against the likelihood of one president over another, the wisdom of the market could be one answer.

The University of Iowa’s predictive markets, specifically the Iowa Electronic Markets 2012 Presidential Election Winner Take All contracts, (IEM) is one attempt to measure it.

The IEM is a futures market in which people can take positions based on certain outcomes. In this blog series we’ll be comparing the Real Clear Politics poll and the IEM Winner Take All to see how they deviate – or converge.
Thomas Rietz, professor of finance at the University of Iowa, explains:

“Suppose that the IEM WTA market has a Republican price of 75 cents. This is a prediction that the Republican will win 75 per cent of the time. It also means he should lose 25 per cent of the time. Suppose the Republican wins. You can’t conclude on the basis of that single observation alone that the IEM was ‘right’. Similarly, you cannot conclude that it was wrong if the Democrat wins. You have to compare the probability forecast with the actual frequency of winning across a lot of elections before you can conclude whether the forecasts are accurate. The evidence is that the actual winning frequencies match the market’s predictions closely.”

To give an idea of what we might see, here is a graph of the IEM Winner-Take-All market. As we can see, there is some relative stability until late September, when Obama pulled away from his competitor, Senator John McCain:

Here’s the RCP Average from 2008:

The polls follow a similar general trend – the race widened a bit in July, narrowed in September, before widening in November. It’s worth noting that in the RCP poll, Mr McCain jumped ahead of Mr Obama for a brief period while the IEM got closer, but never did quite crossover. After that, Mr Obama returned to the lead and the IEM widened considerably.

So, does the IEM’s market wisdom trump conventional opinion polls? Well, that’s what we’re intending to find out. Check back here every other week to see how the predictions are going.