The way the wind blows: The effects of Hurricane Sandy on the US elections

[Note: The research cited in this article was highlighted in a post on The Monkey Cage by Professor John Sides of George Washington University and in a comment responding to that post by Professor Neil Malhotra of Stanford University. While I am grateful to both of them for citing this work, they bear no responsibility for the opinions offered below, which are mine alone.]

As Hurricane Sandy barrels towards the east coast of the United States, our first thoughts are of course with all of the people who are already dealing with the storm and those who will be forced to deal with its aftermath. However, with the presidential election in the United States little more than a week away, legitimate questions are beginning to be asked about the potential effect of the hurricane on the election. 

While there are a myriad of ways to think about how the storm could affect the election, it is probably best to think about this in two different ways. First, there could be direct effects from the storm on people’s ability to vote. Alternatively, there could be indirect effects from the storm in terms of changing how voters evaluate the candidates. Let us consider each of these in turn.

Direct effects

For the storm to have direct effects on the election, we will have to assume that it has enough of a long lasting effect to actually make it more difficult for some voters to cast their vote in the election. With that in mind, three factors seem particularly important.

First, the storm is primarily located in a portion of the country that – in general – is more supportive of Democratic candidates than Republican candidates for national office. Thus the net effect of a decrease in turnout in the northeast part of the United States should be to decrease the total vote for Barack Obama more than the total vote for Mitt Romney. However, in the United States the president is not elected on the basis of total votes across the country; instead the candidate that wins a majority of votes in the Electoral College becomes president.

Electors for the Electoral College are determined on a state-by-state basis, so as long as the storm uniformly depresses turnout among supporters of Obama and Romney in any state where its effects are felt, it should not necessarily impact who wins more electoral votes. This could, however, increase the likelihood of an election result where one candidate (Romney) receives a plurality of the popular vote nationwide, but the other candidate (Obama) receives a majority of electoral votes and wins the election.

The second point is whether we think the above assumption – that a hurricane would uniformly depress turnout among Democrats and Republicans – is warranted, and there are important reasons to think this might not be the case. On the one hand, we know that turnout in elections can be a function of resources (eg, higher income, more education), and that Democrats in the United States draw more electoral support from lower resource individuals than Republicans do. If the storm were to make it harder for lower income Americans to participate in the election than middle and upper income Americans (eg, by knocking out public transportation), then we would expect this to hurt the vote for Obama. And indeed, in a paper published in The Journal of Politics in 2007 that relies on data from 14 US presidential elections, political scientists Brad T Gomez, Thomas G Hansford and George A Krause find that poor weather is associated with better performance by the Republican candidate.

Is there reason, though, to think that Hurricane Sandy might be different from a typical rainstorm? Jonathan Baron, a psychologist at the University of Pennsylvania, offered an interesting argument in a comment on a post on this topic at The Monkey Cage. He suggested that if the hurricane leads to widespread power outages and road closings, this might have a disproportionate effect on voters outside of cities, who will have a harder time getting to the polls. Since Obama and the Democrats poll better in cities relative to the suburbs and, especially, rural areas, this could have a net effect of helping the Obama vote. Additionally, according to Gallup, Romney has a non-trivial lead among voters 65 and older. If Hurricane Sandy should make it disproportionately more difficult for older voters to participate in the election, then this might depress the Romney vote more than the Obama vote.

The final factor to consider in terms of a direct effect on the election outcome is the extent to which the storm will effect early voting, which is possible in many US states. Two stories are possible. First, if voting on Election Day is suppressed, then early voting may be disproportionately influential in this election relative to what it would have been had everyone who wanted to vote on election day been able to do so. This might seem to help Obama, who by most accounts is doing better among early voters than Romney.

This, however, is a double-edged sword, because if the storm suppresses early voting but not voting on Election Day, then the same logic would suggest an advantage to Romney. Research by MIT political scientist Adam Berinsky, however, suggests that people who participate in early voting actually tend to be highly motivated voters, not casual voters, and thus are precisely the type of people who will make the effort to vote on Election Day. This would suggest that fewer opportunities for early voting alone – independent of serious barriers to voting on election day – might not have much of an effect on total vote figures.

Indirect effects

Of course, it is also possible that an event such as a massive storm, the likes of which Hurricane Sandy is projected to be, could change voters’ opinions of the candidates. For Romney, the stakes and opportunities are both low; as long as he is not perceived to be attempting to benefit politically from the suffering of US citizens dealing with the storm, his image is probably unlikely to change much.

For Obama, however, the stakes are higher. He has both the opportunity to appear presidential in helping lead the response to the storm, but also the potential of being blamed (eg, seen as incompetent, a poor leader) should things go badly.

There is actually quite a bit of political science research on the topic of how natural disasters affect support for sitting presidents, and there is evidence both of natural disasters hurting incumbents (see here and here) and disaster relief efforts helping incumbents (see here and here).

Although this is only speculation on my part, I would guess that given the fact that there is only a week before the election, the potential downside for Obama is greater in this regard than the opportunity; there may unfortunately be plenty of opportunities for images of suffering and problems associated with the storm in the coming week, but less so for the effects of relief efforts.

One final point, however, is worth noting. A great deal of the ideological conflict between the two candidates in this election has been about whether government is part of the problem (Romney) or part of the solution (Obama) for America’s future. Natural disasters of the type that Hurricane Sandy is possibly shaping up to be are the types of events to which governments – as opposed to any individual private actor – are uniquely positioned to respond.

It is possible that the events of the next few days may reinforce the message that government is indeed necessary to solve some problems facing US citizens; this may not be the last thing about which the Romney teams wants voters to be thinking before they enter the polling booths.

Joshua A Tucker is a professor of Politics at New York University, a National Security Fellow at the Truman National Security Project, and a co-author of the award-winning politics and policy blog The Monkey Cage.