Is Google Trends an Election Indicator?

A personal favorite election indicator is the Halloween mask indicator, which holds that whichever presidential candidate sells more Halloween masks is the one that will eventually win the election.  The indicator has correctly predicted every presidential election since 1980.

The reason it's my favorite indicator is that it's somewhat counter-intuitive (wouldn't you think that you would buy the mask of the candidate that scares you more--the one you aren't voting for?).  It's also a favorite because of what it says about mind share, marketing and elections.  Whichever candidate holds the greater mindshare both at the ballot box and at the costume shop is the one that ends up winning.

Thanks to Google trends, measuring mind share is easier than ever.  The more that people are searching for a candidate, the more that they are thinking about him.  Google trends only goes back to 2004, but in 2004 and 2008, the candidate that was more frequently searched for on Google is the one who ended up winning.

Below are the google search trends for Bush/Kerry, Obama/Mccain and Obama/Romney.  Mccain never eclipsed Obama in searches in 2008, but in 2012, Romney has taken a slight lead out of the Republican convention.  As the election heats up in coming weeks, this could be an indicator to watch.

Note the scale is slightly different on each chart.

Clear Arbitrage Opportunity on Intrade

Even though intrade continues to gain in popularity as an indicator of the probability of a future event happening, it's interesting that the market seems to have some glaring inefficiencies.  Not only is volume relatively light, but also, there is a clear arbitrage opportunity in the site's most popular contract.

Right now the Obama and Romney presidential contracts sum to show that there is only a 97.6% chance that Obama or Romney will be elected president this year.  Seeing as how there is no viable 3rd party candidate in the race, the contracts seem to be discounting that there is a 2.4% probability that either Barack Obama or Mitt Romney will drop out of the race between now and November.  

If an investor were to buy both contracts today, he or she could stand to earn a nearly risk free 2.4% between now and November: an annualized return of 8.1%.