How can a hurricane impact an election? Donald Trump had some thoughts back in 2012.
In a matter of days, Trump suggested that Hurricane Sandy in the Northeast could help both President Obama and Mitt Romney, for very different reasons.
Hurricane is good luck for Obama again- he will buy the election by handing out billions of dollars.
— Donald J. Trump (@realDonaldTrump) October 30, 2012
Could be the hurricane helps @MittRomney--people are rioting in the streets over gasoline
— Donald J. Trump (@realDonaldTrump) November 2, 2012
Hurricane Matthew is significantly earlier in the election than Sandy was — early October vs. late October — and we still don't know precisely how much it will affect Florida, Georgia and South Carolina. So it's very early to talk about political implications.
But given Florida's status as a hugely important swing state (and even Georgia's status as a surprising battleground), there will be plenty of debate about the political impact the storm could have come Nov. 8. And the political fight over it has already begun, with Florida Gov. Rick Scott (R) on Thursday declining the request of Democrats to extend voter registration in that state.
Here are three ways in which storms like this can affect elections — along with whether there's evidence they actually do.
1. Reducing turnout
The most obvious way in which storms can affect elections is by depressing turnout.
Given the states hardest hit by Sandy were all blue ones, a decline in turnout in 2012 wouldn't have had an effect on the actual outcome. But there was evidence Sandy kept at least some voters from voting.
A study by Rice University professor Robert M. Stein in 2015 found the following:
- “Turnout declined on average 2.8 percent between 2008 and 2012 in counties in which disaster declarations were issued for Hurricane Sandy. Voter turnout declined only 0.8 percent in all other U.S. counties.”
- “Early voting increased significantly in unaffected counties between 2008 (14 percent) and 2012 (16 percent), while also increasing by 2 percent in counties most adversely affected by Hurricane Sandy.”
- "As expected the number of polling places declined in counties most adversely affected by Hurricane Sandy. In 2008 these counties operated 1.1 polling places per 1,000 registered voters; in 2012 this figure declined to 1.0 polling places per 1,000 registered voters, a statistically significant change in the number of polling places.”
And then there is the matter of voter registration and Scott's decision not to extend it. “Everybody’s had a lot of time to register,” Scott said. “On top of that, we have lots of opportunities to vote, early voting and absentee voting, so I don't intend to make any changes.”
Florida professor Daniel A. Smith notes that 50,000 Floridians registered to vote in the final days of the 2012 election. Democratic registrations during this period far outpaced Republican ones, by about 73 percent. And people who register shortly before an election would seem pretty likely to want to vote.
Sorry, those previous figures I posted were for the final 8 days of VR in in FL 2012; final 5 days had roughly 50k newly registered voters
— daniel a. smith (@electionsmith) October 6, 2016
The Table on the left are new Dem registrations; Table on the right are new Rep registrations, 2012, but you probably knew that... pic.twitter.com/vbyzZ5bknY
— daniel a. smith (@electionsmith) October 6, 2016
Fifty thousand voters in a state where 8.5 million people voted in 2012 isn't a huge number, and a drop of a point or two in overall turnout doesn't seem like a massive shift. But this is the point at which I remind you that Florida in the 2000 presidential race was decided by just 537 votes — and it is a state that yet again is polling very competitively.
So really, any effect on turnout could theoretically swing the election — in Florida and nationally, if the overall margin of victory is 29 electoral votes or less.
Another important factor: Sandy's hardest-hit state, New Jersey, in 2012 made all kinds of arrangements for people affected by the storm to be able to vote, including allowing them to do so on the Internet. That surely helped keep turnout closer to normal levels, but it's not clear whether Scott will make similar arrangements should they be necessary. And his decision on the voter registration period — “I don't intend to make any changes” — suggests he'll be reluctant to do so.
2. The government's response
Storm responses can have massive impacts on how the politicians who handle them are viewed.
New Jersey Gov. Chris Christie became the most popular governor in the country in 2012 after his response to Sandy was praised. A poll conducted a month after the storm pegged the Republican's approval rating at an unreal 77 percent — up 21 points from just the month before. And it stayed in the stratosphere for months. (It has dropped steadily since then, of course, and now Christie ranks among the least-liked governors in the country.)
A contrasting example that everyone will remember is George W. Bush and Hurricane Katrina. Americans had just reelected Bush as president the year before and were about evenly split on him before the storm; afterward, his approval rating dropped to around 40 percent, and it spent much of 2006 in the 30s.
Going back further, Bush's dad, George H.W. Bush, in 1992 had to contend with the fallout from Hurricane Andrew, a Category 5 storm that hit Florida very hard in August of the election year. While there was some hand-wringing over the response, Bush largely got a passing grade and actually won Florida — despite losing reelection.
It bears noting in all of this that an incumbent president isn't on the ballot in 2016. Were President Obama facing reelection, the ratings of his response would certainly play into the 2016 election.
Neither Hillary Clinton nor Donald Trump are in charge of disaster response, but given the fact that Clinton has tied herself to Obama — and Obama is pretty popular right now — views of how his administration handles the response could theoretically filter down in some way.
3. Gamesmanship and the political response
Trump's tweets above — and everything about his campaign — show just how willing he might be to use the hurricane to cast doubt on the results of the election.
Trump has already suggested there is no way he could lose a state like Pennsylvania unless there is voter fraud and has called for his supporters to personally monitor the polls. He warned in North Carolina that people might vote "15 times” for Clinton. He has said the system is rigged against him.
Going back to Sandy again, a Rutgers Law School study in 2014 found that New Jersey's efforts to make it easier for displaced people to vote — including Internet voting — were illegal and opened the door to voter fraud at a time when the state was dealing with other things and ill-equipped to combat it. The Christie administration strongly disputed the findings.
Since Scott doesn't appear inclined to make adjustments to how his state will conduct its election — and since he's a Trump backer — it's less likely a similar controversy could arise in Florida. But Trump has shown a willingness to seize upon basically anything (even without facts to back him up) to call into question any potential loss. And it wouldn't be surprising to see him use a hurricane in the biggest swing state to make the same case.
And then there's just the overall responses of the candidates. Trump issued a very disciplined statement on Thursday ...
... but his responses to tragedies like the mass shooting in Orlando have been decidedly less diplomatic and presidential.
Appreciate the congrats for being right on radical Islamic terrorism, I don't want congrats, I want toughness & vigilance. We must be smart!
— Donald J. Trump (@realDonaldTrump) June 12, 2016
More recently, Trump earned some praise for being the first big-name politician to visit Louisiana after it was ravaged by floods.
For both Clinton and Trump, their words in the face of tragedy can have an impact on how they're viewed — even if they're not being graded on the actual response.