Hurricane Harvey has slammed into Texas, bringing power outages, destruction and extensive flooding that is expected to get worse. While personal harm and property damage resulting from Harvey are rightly the overriding concern, there is already speculation about how Donald Trump will handle the first major natural disaster of his presidency and whether it could significantly affect his reelection fortunes.
Based on our new research examining the impact of natural disasters on American presidential elections, we believe that Harvey’s effect will largely depend on the preexisting partisanship of the counties hit.
What the political science research has found
Several studies of how voters respond to natural disasters find that they punish incumbents at the ballot box. For example, Christopher Achen and Larry Bartels, who show that, after shark attacks off the New Jersey coast in 1916, voters in affected counties punished Woodrow Wilson more than voters in unaffected counties in that year’s presidential election. They also show that voters punish incumbent party candidates after droughts and floods in 26 presidential elections across the 20th century.
While Achen and Bartels’s results about shark attacks have been questioned, they do not stand alone. In a recent article, we show that voters in Southern counties affected by the great Mississippi Flood of 1927 punished Republican presidential candidate Herbert Hoover in the 1928 election. This is especially notable because Hoover — as a member of the Coolidge administration — was personally responsible for managing flood relief efforts, which were large-scale for that time.
On the other hand, other scholars have found that if incumbent politicians provide relief from the natural disaster, voters do not punish them. In fact, incumbents can even benefit electorally from the natural disaster. Any punishment comes only when incumbents fail to provide relief in the wake of a natural disaster.
What our new research shows
Our new research argues that voters are not simply considering the scale of the disaster and incumbents’ response. Instead, they are judging the actions of elected officials based on their partisanship. As is true in so many domains, voters who identify with the president’s party will assess his performance after a natural disaster much more positively than those who identify with the opposite party.
For example, consider Hurricane Sandy, which hit the northeastern United States days before the 2012 presidential election and was one of the deadliest hurricanes in recent history. President Obama’s handling of Sandy was praised at the time — including by New Jersey Gov. Chris Christie, a Republican.
Despite this, Obama did not do notably better or worse in counties affected by Sandy. Sandy’s political impact only becomes clear when you take account of the partisanship of counties. Obama did better in safely Democratic counties that were hit by Sandy, compared to similar counties that were not affected by the hurricane. In contrast, Obama did noticeably worse in safely Republican counties.
We found something similar when we examined a much larger set of natural disasters and relief efforts between 1972 and 2004. In disaster-affected counties that were safely in the incumbent party’s column, candidates of that party were rewarded in the wake of a natural disaster. In disaster-affected counties that were safely in the opposition party’s column, incumbent party candidates were punished severely — even if they provided relief efforts.
What does all of this mean for Trump?
Answering this last question entails many caveats, of course. Most obvious is that the next presidential election is still more than three years away — and we don’t know whether any effect of Hurricane Harvey on Trump will last. Moreover, our research examines the effect of natural disasters only in the counties that are directly affected. Poor relief operations could end up damaging Trump’s image elsewhere, as happened to George W. Bush after Hurricane Katrina.
Still, our research helps to estimate any effect on Trump in the 2020 election. The actual size of that effect depends on a number of factors, including the ultimate track of the storm, the extent of damage it ultimately causes, and the Trump administration’s relief efforts.
But examining the 16 coastal Texas counties between Corpus Christi and Houston indicates potential harm to Trump’s political support. Using a range of plausible assumptions about total storm damage, our models show that Trump would lose between 9,900 and 32,600 votes in these 16 counties combined. While most counties in Texas are Republican strongholds, the bulk of those lost votes come from Harris County, which includes Houston itself. Harris County went for Clinton by a comfortable margin in 2016.
Texas’s status as a safe Republican state means that this effect is unlikely to swing it to a Democratic presidential candidate. However, to the extent that Harvey’s destruction remains concentrated in Democratic-leaning urban areas, Trump may face significant punishment in 2020.
Boris Heersink is an assistant professor in the Department of Political Science at Fordham University. Brenton D. Peterson is a PhD candidate in the Woodrow Wilson Department of Politics at the University of Virginia and a research affiliate at Strathmore University. Jeffery A. Jenkins is the Provost Professor of Public Policy, Political Science and Law; the Judith & John Bedrosian Chair of Governance and the Public Enterprise; and director of the Bedrosian Center at the Sol Price School of Public Policy at the University of Southern California.