The 2010 Midterm Election, set for November 2, is shaping up to be one of the most interesting elections in recent times. Notwithstanding a few exceptions, a candidate’s profile this time around is not as big of a factor on the voters’ mind as the general mood of the country. Although most polls give the edge to Republicans, the Democrats have been able to close the gap somewhat. In this article, I will provide an empirical analysis of the 2010 Midterm Election and offer predictions as to which party will control Congress come next year.
In economics, there is a concept called Tiebout Hypothesis which contends that people reveal their personal preferences in their habits and behavior. This idea is applicable to voting in elections as well. As people are increasingly using the Internet to search for information on their preferred political party or candidate, an empirical analysis of web search data may allow us to capture the mood of the country and predict the voters’ choice come this November.
My analysis of Google search trends data (U.S. only) finds that it had correctly predicted the general outcome of the last two congressional elections. In 2006, average search trends data (Jan-Oct) for information on the Democratic Party was 0.92; the average for the Republican Party was 0.82. The Democrats ended up earning a net gain of 31 seats in the House and 5 seats in the Senate. In 2008, the average search trends data was 1.0 for the Democrats and 0.8 for the Republicans. The Democrats eventually gained 21 House seats and 8 Senate seats.
If the trend holds for the 2010 Midterm Election, the Democrats appear to be in serious trouble. The search trends data for 2010 shows a dramatic shift that clearly favors the GOP. The average search trends data (up to October) for information on the Republican Party is 1.32, compared to 1.0 for the Democratic Party. The difference of 0.32 is much larger than the ones enjoyed by Democrats in 2006 and 2008. This tells us that 2010 may be a repeat of 1994, when the Republicans gained a whooping 54 seats in the House at the expense of Democrats.
=>CONTINUE ON PAGE 2
Will the Republicans be as successful in Senate races? While a net gain of Senate seats is assured, how big it will be is still uncertain. Search trends data provides a moving snapshot of the mood of the country, which makes it very good at predicting House elections. For Senate races, a candidate’s profile is equally as important as the current political environment. There is a reason why some senators have been able keep their seats through many tough election cycles. These people possess strong name recognition and internal support that are difficult for their challengers to overcome. The best estimate would be a gain of 6-10 seats for the Republicans.
If the predictions hold for the 2010 Midterm Election, the Republicans would take over the House as the majority party and at worst, prevent the Democrats from retaining a clear majority in the Senate. This will inevitably put President Obama in the same situation former President Clinton was in following the 1994 Midterm Election. How the President deals with a Republican congress will ultimately decide whether he will be reelected two years from now.
Google search trends data for 2004-2010 are available for download (in CSV format) at the Google Trends website. This analysis is based on my interpretation of the trends data.