Could Climate Change Turn Natural Disasters into a Direct Threat on Democracy?
Tom Johansmeyer uses data from the United States to explore an unanticipated side effect of climate change.
Natural catastrophe events could threaten global climate change progress simply by directly influencing the outcomes of major elections. It’s the sort of threat that seems obvious when mentioned to you, but which would never occur to you otherwise. In fact, PCS®, the group I lead at Verisk™, only explored the issue because of a close call in the United States in 2020. The good news is that few nations face massive natural catastrophe risk, and the number narrows further when you look at the countries exposed to repeated significant natural catastrophe risk and narrower still when only mature representative democracies with large economies are considered.
The United States is the most frequent target of large tropical cyclones, and despite that exposure, the number of such events (or other natural catastrophes) occurring shortly before presidential elections has remained low. Only a handful of major natural disaster events occurred close enough to U.S. presidential elections to possibly influence their outcomes, and in each of those cases, underlying circumstances weren’t sufficient to make a difference. However, several close calls – including Hurricane Zeta a little over a year ago – should serve as a warning to other nations exposed to tropical cyclone events about the potential threats to the normal operation of their most important institutions.
While the United States may be the specific focus of this article, given the availability of relevant data, the impact of severe weather – and climate change – requires broader global consideration. It’s possible for a natural disaster resulting from climate change to sway an election in a way that would reinforce climate-detrimental policy, ultimately exacerbating the climate conditions that swayed the election in the first place.
More catastrophes, broader threats
To understand the impact that natural catastrophes could have on election outcomes, you first need to understand the mechanics of how it could change the outcome of an election. Many would begin with concerns about storms affecting voter positions on issues or policy, but for the purpose of this analysis, PCS has decided to consider only catastrophes so large and close to an election that they could impede the physical movement of voters regarding their ability to reach polling places. Put simply: We’re looking at catastrophes so big that they keep voters from the ballot box, resulting in voter turnout volatility that could result in a win for the otherwise minority party candidate.
Further, only events occurring in the October or early November before a U.S. presidential election are considered, and historically, they haven’t been common. It’s roughly a 30-day window (since presidential elections are the Tuesday after the first Monday in November) that opens only once every four years. Within that time constraint, a “perfect storm” of conditions would have to be met, and so far, that hasn’t occurred.
To provide a sense of the cadence and magnitude of U.S. catastrophe events, it’s worth reviewing the 20 years ending December 31, 2020. PCS designated an average of 36 catastrophe events per year in the United States during that period, with an average annual insured loss of more than US$29 billion. Of course, many of those events are small; the threshold for declaring a catastrophe is an industrywide insured loss of US$25 million, with a significant number of insurers and insureds affected. That said, the annual average industrywide insured loss shows that there are plenty of devastating events in the mix.
Additionally, frequency is on the rise, as evidenced by the data provided (keeping in mind that the spikes in the 1970s and 1990s result in part from the fact that PCS increased the minimum reporting thresholds for catastrophe events at the ends of each of those periods). Yet, that doesn’t mean there’s a straight line from the frequency of PCS-designated catastrophes to worsening climate change.
PCS reports on events that cause sufficient insured losses to meet our threshold. A major PCS event could come during a below-average year for storm frequency and intensity, given that we report by industrywide insured loss totals. Hurricane Sandy wasn’t a particularly powerful storm, but it hit heavily populated and economically robust parts of the United States. Conversely, a hypothetical major hurricane that hits an empty stretch of land may not have a similar impact on insurers.
In general, PCS data is an imperfect proxy for climate change, but that doesn’t mean there aren’t cases where PCS data can be quite useful in understanding some climate change impacts. Overall, the average number of PCS-designated events per decade surged from only 10 in the 1950s to 33 in the 1980s to 40 in 2010s. Since 2013, the number of PCS-designated catastrophe events has gone up every year, reaching 71 in 2020 and already more than 60 in 2021, as of this writing, an indicator that climate change conditions are worth further, more specific exploration.
The potential implications of climate change are, of course, far-reaching, and even an increased rate of tropical storm landfalls on population centers could be profound. However, as you move from individual, community, and municipal impacts to how natural disasters shape political institutions, the effects tend to be second or third order. Aid and government relief, for example, may be allocated to communities hit by major storms – and in a way that could ultimately influence elections as candidates seek favor with voters, but that is much different from seeing a catastrophe event directly change voter behavior, for example, by restricting their mobility.
Close calls could get closer
PCS analyzed historical major catastrophe events occurring in October and early November of election years to ascertain the potential extent of the threat that natural catastrophe events could pose to presidential elections, and thus the policy and diplomatic positions that might come from the elected administration. The close call in Georgia in 2020 prompted PCS to conduct this research. Hurricane Zeta passed through some of the most tightly contested states only days before last year’s presidential election. Two states with close results that were seen as having the potential to shape the nation’s final result – Georgia and North Carolina – were in the hurricane’s path.
The notion that a major catastrophe, like a hurricane, could directly influence the outcome of a presidential election takes us into uncharted territory. Although there have been some major storms that have shaped world events, to Christopher Columbus’s own near miss in 1502, the United States has yet to see a natural catastrophe change an election –going back at least 70 years. It doesn’t seem that the close calls have even been all that close, as there haven’t been enough cases for detailed analysis.
A natural disaster shortly before an election is only part of the collection of factors that could change its outcome and consequently influence climate policy. During presidential elections, some states can be more important than others. A hurricane causing extensive damage in a state that tends to vote heavily for one party is unlikely to alter the future, while a storm hitting one with a tightly divided electorate could have profound implications for the presidency, and by extension, the world’s climate agenda. In addition to causing physical damage in narrowly divided states, a storm would also have to impact enough of these divided states to possibly swing the Electoral College tally from one candidate to the other.
Since the inception of PCS, the United States has had 18 presidential elections. The only years with natural disaster events that could be called significant under even the loosest of definitions were: 1964, 1968, 1996, 2012, 2016, and 2020. Election Day may fall within hurricane season, but because it’s toward the end of the season, it isn’t threatened often.
A review of key states in presidential elections going back to 2012 shows how rare the necessary conditions are. Interestingly, all three elections during that period (2012, 2016, and 2020) featured a impactful natural disaster shortly before the election, although 2012 and 2020 were most important, with the former Hurricane Sandy and the later Hurricane Zeta. The table below shows the number of insurance claims per vote cast in the relevant catastrophe event, as well as the difference between the candidates votes. To imply that a disaster event might have helped shape the election, the results would need to show a high rate of claims per vote cast, as that could signal the extent of human impact, and the delta between candidates’ votes would have to be narrow, indicating that the election was tight and could be swayed.
Historically, Hurricane Sandy in New Jersey showed the greatest human impact with claims per vote cast reaching a rate of nearly 15 percent, with the same storm in New York next at just shy of 8.5 percent. However, both states voted solidly in one direction, suggesting an extremely remote risk of the storm changing the outcome of the election. Further, the Electoral College results would have required more than two states turning. Eight years later, Hurricane Zeta pushed a rate of claims per vote cast above 7 percent – but with a difference between candidate vote tallies pushing 20 percent. As with New Jersey and New York in 2012, the state’s election was decidedly one-sided, meaning that the storm wouldn’t affect the outcome.
Top States by Claims/Total Votes Cast
Georgia was a bit different from those cases. The rate of claims per vote cast reached only 1.29 percent, and the margin between the two presidential candidates was a mere 0.24 percent. Given the narrow spread, it’s possible that, had the storm been more devastating, it might have changed the state’s result. Last year, storm risk may have been mitigated by pandemic-related early voting, which would have decreased the importance of mobility given that votes had already been cast.
When reviewing the situation in Georgia in 2020, there’s a lot that could have gone differently. Had the storm been more powerful or slower, had early voting not become a priority because of COVID-19, had the path of the storm shifted – small shifts could have made a profound difference. A different outcome Georgia would not have upended the 2020 election, but North Carolina was affected by Hurricane Zeta, as well, although the election-related risks were lower than they were for Georgia. However, it’s yet another “What if?” to add to the list of catastrophe-related scenarios that could change how the United States (and the rest of the world) affect international relations and thus address climate change.
Missing from the analysis are Florida and Texas, two of the most catastrophe-prone states in the United States. Both states have had close elections and major catastrophe damage (particularly from hurricanes), but they haven’t had both in the same year. Florida’s 2000 election was even narrower than Georgia’s in 2020, and the state has seen many large hurricanes. The same could be said for Texas. Although it hasn’t had a 2000-style close call, the state does tend to have tighter voting deltas than some realize. Both states have had more than 20 catastrophes per year for the past twenty years.
How the climate could affect the climate change debate
The question remains: What if it actually happened? The risk is real, and the effects of a major hurricane, for example, could determine four years of climate policy for the United States, which would in turn influence the objectives and negotiating positions of other world powers. The climate debate is already delicate. A change in administration in a country with a high-profile role in the ongoing climate debate could shift the dynamic for everyone else at the table.
Election catastrophe planning and resilience could provide near-term relief, and the work done recently for pandemic-era voting would provide a useful starting point. Consistently available early and mail-in voting procedures could mitigate the risks associated with casting the bulk of votes on a single day. If votes are cast in advance or secured by mail, the risks of voters having to stay home because of weather conditions declines precipitously. It may seem like a small measure, but during a difficult hurricane year, it could make a profound difference for progress on climate change.
Tom Johansmeyer is head of PCS, a Verisk business, which estimates the industry-wide insured losses from disaster events around the world. He writes and speaks regularly on natural catastrophes, cyber attacks, and political violence events.