Why are there so many natural disasters during presidential election years?

Commentary about business and finance.
Feb. 29 2008 2:22 PM

Landslide!

Why are there so many natural disasters during presidential election years?

Tornado damage
Tornado damage in Arkansas

I predict with a high degree of confidence that this year, the nation will witness an unexpectedly large number of destructive acts of nature—floods of biblical proportions, out-of-control wildfires, destructive tornadoes, and bone-rattling earthquakes—that will require a prompt, compassionate, and expensive response from the federal government. No, I haven't just returned from a crash course in meteorology. Rather, I recently had lunch with Erwann Michel-Kerjan, an economist at the Wharton School who studies catastrophes. According to Michel-Kerjan, this uptick in natural disasters will likely have nothing to do with global warming and everything to do with the fact that 2008 is an election year.

As part of a large—and sobering—study on the risks posed by natural and man-made catastrophes to the global economy, Michel-Kerjan crunched data on the incidence of federally declared disasters since the early 1950s. The results can be seen in the chart below.

"Disasters" in the U.S. - Presidential Declarations per Year chart

Advertisement

(The raw data can be seen at FEMA's Web site, which charts disasters by year and by state going back to 1953.) Michel-Kerjan believes it is no coincidence that 1) the incidence of disasters has risen over time; and 2) there are some conspicuous spikes during presidential election years.

What gives? Disaster declarations are presidential acts, which are undertaken at the chief executive's discretion. While there are some quasi-official criteria—"You're not going to declare a major disaster just because the wind is running at 10 miles per hour more than average," Michel-Kerjan notes—the president has some flexibility. And it seems plausible that during an election year, a president seeking to be re-elected or who wants to help maintain his party's control over the executive branch will try to be seen as a hands-on executive who dispenses aid to voters. Hence, when something bad happens in an election year—the Mississippi River floods, wildfires burn in the West—the president is quicker to declare a federal disaster than he would in a nonelection year. In 1996, President Clinton was so eager to feel his constituents' pain that he declared 75 disasters, more than twice the number he declared in 1995 and 70 percent more than he did in 1997. But Michel-Kerjan says it's too simplistic to blame the president. This is a bipartisan, bibranch development. Once the president declares a disaster, Congress decides how much cash to allocate for disaster relief. And in years when members of Congress are eager to be seen delivering goodies and care to their constituents, they're likely to be more welcoming of opportunities to demonstrate their compassion.

The case isn't bulletproof. Every year, of course, there are some disasters that require assistance. And natural disasters are not evenly distributed through time. Nor are the spikes evident in every election year: 2000, 1988, and 1968 were comparatively benign on the disaster front. "If you don't have many major things happening, it's hard for the president to invent them," says Michel-Kerjan. In 1988, there were only 11 officially declared disasters. (Perhaps Reagan and Bush I felt the Dukakis challenge was so feeble that there was no need to act precipitously.) And while the general rise in the incidence of disasters may also seem politically suspect—a typical year in the 2000s saw about twice as many disasters as a typical year in the 1980s—there has been a distinct rise in development in marginal areas and in the value and concentration of property along coasts and rivers, in forested areas, and on fault lines. Twenty-five years ago, when trees burned in forests or when hurricanes smacked the coast of Florida, they were far less likely to cause massive property damage and threaten lives.

But Michel-Kerjan suggests that even controlling for such factors, there does seem to be an important political effect. It's hard not to conclude that during a typical election year, the threshold for a disaster seems to be set somewhat lower—he notes that either there has been a systematic peaking of presidentially declared disasters during election years or "God is against presidents." Still not convinced that politics plays a role? Michel-Kerjan says similar effects have been seen in studies on disaster declarations and elections outside the United States. And go to the FEMA Web site's state-by-state data. Since 1953, the top four states in number of disasters are Texas, California, Florida, and New York. Sure, these are very large states that are in some ways uniquely susceptible to natural disasters. Florida, Texas, and California have very long coastlines, while California is prone to earthquakes, mudslides, and devastating wildfires. But New York? Did you notice that the four states with the highest incidence of federal disaster declarations are also the four states with the most electoral votes?

Thus far, 2008 has gotten off to a relatively rough start, with 10 federally declared disasters in the first two months. The final tally for the year will likely depend on a number of factors: the severity of the Western wildfire season, the number of hurricanes that develop in the Gulf of Mexico, and how much President Bush wants to help John McCain succeed him.

Daniel Gross is a longtime Slate contributor. His most recent book is Better, Stronger, Faster. Follow him on Twitter.