Connecting state and local government leaders
It’s smashing, in the bad way.
America’s 18,000-odd car dealerships face many threats. They are unpopular. Younger generations are driving less. Electric cars require less maintenance. Also, every so often chunks of ice fall from the sky and smash all the cars.
For car dealerships—and for the web of automakers, banks, and insurance houses that they sustain—hail is a menace. Because intense hailstorms form suddenly, as tight, dense wallops within larger thunderstorms, dangerous hail is challenging to predict. A squall can blow up with little warning, dump a few minutes of walnut-size hail, and damage hundreds of new cars as they sit on a lot. And a storm that’s hard to detect in advance is harder still to authenticate afterward. So when a car dealership files a hail claim, the insurance company usually doesn’t have much choice but to take its word, check a few photos, and cut a check.
Hail is, you might say, a hard problem. It is also becoming a bigger problem, a new study argues. The study, released today in draft form to The Weekly Planet, says that hail is expanding its footprint across the country at the same time that it is becoming more frequent in the central United States’ so-called Hail Alley. “Hail isn’t moving from Denver to Ohio—it’s falling in Denver and Ohio,” says Alex Kubicek, an author of the study.
I spoke with Kubicek and the study’s other authors last month. Their work is worth your attention for two reasons. First, they argue persuasively—though not conclusively—that hail should be taken seriously as a major climate impact, as much a breaching edge of global warming as wildfires. When scientists list climate change’s biggest threats, they’re more likely to rattle off floods, droughts, and heat waves than severe hail. Yet if hail is increasing in spread and severity, it should make that list: Its ability to cause rapid, sudden, and unpredictable physical damage is outmatched only by tornadoes and wildfires. When baseball-size hail smacked the Denver area in 2017, it caused the state’s costliest natural disaster ever. Hail in Alberta, Canada, has caused similar damage.
This runaway hail risk is discombobulating the insurance market: In 2018, the insurance giant Zurich North America announced that it would no longer insure hundreds of midwestern car dealerships because of “catastrophic” hail damage.
Second, I’m intrigued by who the study’s authors are. Kubicek and his colleagues have scientific credentials, but they’re not academics. They work at an insurance start-up called Understory, which protects car dealerships against hailstorms. Kubicek is, in fact, Understory’s chief executive.
Insurance! Let’s talk briefly about why it is interesting. Insurance is where the economy of spreadsheets and Zoom invites meets the economy of flesh and steel and falling chunks of ice. To build a financial economy, you need to make its constituent parts act in a mathematically dependable and predictable way. But human bodies and weather systems over central North America (to give two examples) are messy and prone to surprises. Yet if you bundle enough of them together over time, you smooth out their unruliness and produce something that, on the whole, has a predictable level of messiness.
In a typical insurance policy, an insurance provider “indemnifies” against risk, which means that it pays out for the full cost of an event once that event has been shown to have happened. Understory sells a different kind of insurance, called “parametric insurance.” Here, an insurance company pays out a fixed fee automatically whenever a trigger event occurs.
Parametric insurance was invented in the developing world, but lately it’s being used more in the United States. It can be used for any climate-related insurance, says Upmanu Lall, an engineering professor at Columbia University—food, agriculture, drought, hail, tornadoes, wind, or wildfire. Parametric insurance is both more concrete than indemnity insurance and also a little more precarious, a fitting instrument for a lower-trust country. (It’s precarious because parametric insurance won’t ever affirmatively cover the full cost of disaster in the same way that an indemnity policy would.)
In Understory’s case, the trigger event is hail. Every dealership with an Understory-associated insurance policy puts a joystick-like sensor on its roof. It can detect the size and speed of falling hail, and when it senses a hailstorm, it triggers an automatic payout.
And that ends the section of the newsletter where I talk about insurance. By positioning itself in this way, Understory happens to have collected a new kind of hail data. It has a ground-truth record of where hail has landed, a surprisingly hard type of data to come across in the U.S.
“The observational data for hail is … dangerous. It’s less quality-controlled,” John T. Allen, a meteorology professor at Central Michigan University who wasn’t involved in this study, told me. “If you hit a city, you get lots of reports, but if you don’t hit a city, you don’t get many reports at all.”
Understory’s data are the key to the new study. Kubicek and his co-authors used machine learning to associate the spots where they know hail has actually fallen with the radar data observed at the time. Then they found similar radar data going back to 1979, extrapolating from their ground-truth data to create a historical record of hail fall.
More frequent hail is a nuisance by itself, but the damage depends on where it falls. In Hail Alley, roofers and car garages are used to dealing with hail. “If you have a hailstorm hit Dallas, there’s a whole market around that,” Kubicek said. “If you have a hailstorm hit Columbus, Ohio, or Nashville, Tennessee, then it will be costly in ways it wouldn’t be elsewhere.”
“The most shocking thing for us is that [hail’s] moving to new territory at essentially a rate of growth of 1.1 percent year over year,” he said. “It works out to a new Delaware getting hail every year.”
Now for the caveat: The paper hasn’t been peer-reviewed. I was able to circulate it to a few hail scientists, and would describe them as impressed with Understory’s data but skeptical of its methodology. “I think they’re on the right track when it comes to ground-truthing, but there’s a lot more that could be done,” Allen said. Often, the paper skips over statistical work or data cleaning that a more polished academic study might require, he added.
Yet these scientists did not dispute its broadest findings: Climate change seems to be making hail worse.
“A consensus does seem to be emerging” that hailstorms will become more severe “as the planet warms,” Julian Brimelow, a hail researcher in the Canadian government, wrote in an email. “We are on a path for some devastating and incredibly expensive hail damage years in the future. This is why insurance and re-insurance companies are so concerned.”
And that’s what I find truly noteworthy about the Understory study. It shows that some climate science is now happening entirely in the private sector: Climate-change data are detected, standardized, and converted into insurance payouts before they ever reach the hands of an academic.
In this newsletter, we talk a lot about how climate change is a function of the economy, how it emerges from the physical infrastructure of the fossil-fuel-based energy system. What Understory shows is that this is now true in a much more mundane way too. Climate science is now part of the ordinary course of business for insurance companies.
Robinson Meyer is a staff writer at The Atlantic.