Shark Attacks and College Sporting Events Can Decide Elections

Leading up to each election, voters are called to assess the performance of their elected officials since the previous election and then decide whether their performance merits another term in office. Political talking heads seem to think that the majority of voters are deliberating over quarter-to-quarter changes in the gross domestic product, number of construction permits granted, and the consumer price index inflation. At some level, people may consider these indicators and use them to inform their vote. Yet, people consider a bunch of other factors that may surprise you.

Political scientists hypothesized that real-world events that evoke negative feelings, like shark attacks, droughts, and various "acts of God," would lead voters to punish incumbents on election day. Christopher Achen and Larry Bartels report that in counties where shark attacks occurred, President Woodrow Wilson lost as much as 8% of the vote he was projected to receive in such counties in 1912 and 1916. In other elections, droughts would dampen (pun intended) support for the incumbent by several percentage points. 

Informed of this, incumbent's political consultants may be starting their rain dances. Don't rain dance, too well or too much, though. Floods can also lead voters to wash incumbents right out of office. 

Well, the clever readers of this blog may be scratching their head asking the question, "Might voters be voting against the incumbent not because of these events, but rather because of how the incumbents failed to satisfactorily respond to these tragedies?" 

Stanford University professor Neil Maholtra asked the same question. He and his colleagues decided to test whether these effects existed for negative real-world events that are entirely outside of the control of the locally elected officials -- the outcomes of college basketball and football games in the 10 days before an election. 

Again,when a negative event happened people penalized the incumbents in their community. Specifically, when the local college football teams lost, the vote total received by the incumbent was 1.6% lower than if there was no college football game played or if the local college football team won. This effect was even larger when the team was popular and tended to sell out all of the seats in their stadium. 

In a follow-up test of this hypothesis, they looked at perceptions of the health of the country and President Obama's approval ratings in the counties which were the homes to the teams in the final 3 rounds of the annual March Madness college basketball tournament. This should be no surprise by now, but people living nearby colleges with basketball teams beating expectations and winning games that the Vegas odds-makers did not predict they would, led to more optimistic evaluations of the nation's health and of the job President Obama was doing.  

So, it appears that real-world events that are largely outside of the control of elected officials can have a surprising impact at the polls. These effects are small, generally a 1-2% of the vote. But, in competitive elections, 1-2% may be enough of a boost to shift the election in one candidate's favor.