Last night, the United Kingdom held elections. What was supposed to be a tightly contested race between the Tory (Conservatives) and the Labour Party turned into a convincing victory for the former. Voters took away 10 seats from the Labour Party, putting David Cameron, leader of the Conservative Party, in a firm position to run the new government. What is interesting is that it wasn't the pollsters who were able to predict the results.
Leading up to the election, the London School of Economics aggregated all of the national polling data and found that the Conservative Party would likely come out with 278 seats to Labour's 267. After election day, the Tories had shattered all projections, collecting a comfortable majority of 331 seats.
Party leaders and pundits were stunned when the first exit polls showed landslide victories for the Conservative Party. One Liberal Democrat, Paddy Ashdown, even offered to "eat his hat" if the exit polls proved to be accurate. Later that night, as the results became clear, Mr. Ashdown had to concede.
“I have been offered 10 hats on Twitter tonight,” he told Andrew Neil of the BBC, “not all of them politely, I have to say.”The result comes as a surprise to many politicos, especially in the age of
Nate Silver and statistic-driven analytics. However, there is one group that was able to accurately report the odds of each party's chances of taking a majority.
"Who is that group?" You ask. The bookies.
That's right, the bookies. The guys who set the spread on College Football Game Day are also very good at predicting geopolitics.
An article in the New York Times points out that while major pollsters ran off a list of reasons why their models turned out to be incorrect, the betting market had the election pinned, with the Conservative Party coming in at five-to-one favorites on Betfair.
While there are legitimate reasons for the sampling being skewed, including bias answers to surveys and the inability to reach a younger audience on cell phones, it is interesting that money markets were able to cut through noisy data and hone in on the correct pattern.