A small engineering team at a cavernous lab in South Carolina spends its days setting buildings, fences and bushes on fire — and studying what happens next.
The research is for the U.S. insurance industry, which has suffered historic losses over the last decade in blazes that have leveled entire communities.
The industry in response has tapped the scientists to investigate how wildfires spread through urban areas, a field of study that has taken on greater importance as climate change fuels bigger and more destructive wildfires.
“The impact that it’s having on our communities is the likes of which we have never seen in past decades,” said Anne Cope, who leads the team of engineers at the Insurance Institute for Business and Home Safety, an industry-backed nonprofit. “We have to put a stop to it.”
A range of factors have complicated that work. Among them: Wildfires only recently have become a major issue for the industry, which means both the science and data surrounding their impact on communities are less advanced when compared with other disasters, such as hurricanes.
Major wildfires also can be more complex to model in part because blazes, unlike hurricanes or earthquakes, are often started by humans and can be suppressed by them, too.
The industry in response has ratcheted up investment in catastrophe modeling and building science research to keep pace with the mounting threat.
The ultimate goal: to more confidently account for the risk; help homeowners minimize it; and stabilize insurance markets that are reeling amid inflation, a shifting regulatory environment, rising reinsurance costs and climate-juiced disasters.
“It’s just a very volatile time right now for companies as they are trying to understand, and model, and price risk related to a wildfire,” said Karen Collins, vice president of property and environment at the American Property and Casualty Insurance Association (APCIA), an industry trade group.
Responding to that reality “is this really huge effort that’s underway both at the state level as well as at the federal level,” Collins added.
A ‘dramatic shift in losses’
Wildfires have torched U.S. forests and grasslands throughout history. But it wasn’t until the last decade or so that fires more regularly started destroying hundreds of homes at a time, generating billions of dollars in insured losses in states such as California and Hawaii.
“Hurricanes are so destructive in that anybody on the Eastern Seaboard had that threat of entire communities being wiped out. But with wildfire, that didn’t start happening right away,” said Carole Walker, executive director of the Rocky Mountain Insurance Information Association, an industry group.
Wildfire-related insured losses have skyrocketed over the last decade. Nine of the 10 costliest wildfires in U.S. history through 2022 took place in or after 2017, according to the Insurance Information Institute. And data from SwissReshows that global insured losses between 2011 and 2020 for wildfires alone were more than five times higher than losses in the previous three decades.
Those losses are due in large part to blazes such as Colorado’s 2021 Marshall Fire, which destroyed more than 1,000 structures, amounting to at least $2 billion in insured losses.
Before those types of events began occurring more frequently, insurers didn’t have to pay very close attention to the issue “because the number of fires that occurred and the likelihood of them” destroying neighborhoods were relatively small, said Michael Young, vice president of model product management at Moody’s RMS.
The last decade’s “dramatic shift in losses,” Young added, has forced the industry to consider wildfire a “catastrophe peril” for the first time and to approach it accordingly.
Doing so has entailed tapping catastrophic risk modeling firms such as Moody’s RMS to run simulations of future hypothetical wildfires that provide insight into how a given wildfire would impact a particular type of neighborhood — and what financial losses would result.
That’s different from traditional “actuarial accounting,” which uses past disasters and the claims they generate to predict future risk. Young said catastrophe modeling, which incorporates a sprawling set of data about wildfires, buildings and current economic conditions, can help insurers better account for wildfire when making underwriting and pricing decisions in risky areas.
California for years has prohibited insurers from using those models out of transparency concerns. The state plans soon to reverse that rule given the role “cat models” can play in boosting insurers’ confidence around pricing wildfire risk.
“The reliance on those models is taking on a new life, or a different life, than it has in the past,” Michael Conway, who serves as Colorado’s insurance commissioner, said in an interview.
While the models are important, Conway added, regulators lack transparency into how they work, what types of information they account for and how often they’re updated. He said that makes it difficult to know how exactly insurers are making underwriting and pricing decisions — a problem Conway said could necessitate legislation down the line. Oregon, for its part, is already moving in that direction.
The industry also has invested in research to get a better grasp of how fire jumps from wild lands to urban neighborhoods, as well as how mega-fires can be prevented.
IBHS, the group based in South Carolina, is the industry’s main source of that information. The organization has a team of three researchers, led by Cope, who carry out wildfire-related projects.
Just one example: research focused on vented attics and crawl spaces, which are meant to ensure airflow throughout a home — but can also catch embers during a fire and “burn the house from the inside out,” said Cope. That research has entailed testing different types of vents to see which ones prevent that outcome.
The project fed into the 2022 release of a standard that provides homeowners with specific steps they can take to mitigate their risk — and, at least in theory, boost their insurability. IBHS’s standard focused on hail and wind, meanwhile, has been around for more than a decade.
“We’re pushing that standard out into more states and trying to get momentum and investment to take action to actually achieve this,” said Collins of APCIA, who said the standard was first rolled out in California.
“But it’s going to take some sweat equity and some dollar investment,” she added. “That’s the hurdle we’re trying to work through.”
Humans are a big X factor
Catastrophe models and scientific research have evolved rapidly, and they already are helping insurers better account for both wildfire risk and on-the-ground efforts to mitigate it in their underwriting and pricing decisions, industry experts say.
But they don’t give the industry all of the information it needs to understand how these fires happen, how they can be prevented and what both mean for a particular home’s insurability.
“I think it’s certainly on the right track. But I wouldn’t say that the industry has this completely figured out,” said Roger Grenier, a senior vice president at Verisk, another catastrophe modeling firm.
That’s the case for multiple reasons. While there’s a long history of research into how fires ignite and move through wildlands, less time and attention has been spent on modeling fires as they transition from forests into neighborhoods, Grenier said.
The same goes for the science around what happens when different types of buildings are exposed to embers or radiant heat. IBHS and other groups’ experiments have moved the needle in that regard by demonstrating that wildfire mitigation, when done well, can dramatically reduce homeowners’ risk — sometimes by as much as 50 percent.
But there’s still some uncertainty among insurers around how to quantify the specific risk reduction that results from certain mitigation steps — and translating that information into decisions about where to offer insurance and how to price it.
The industry claims that’s in part because insurers lack sufficient historical data from real-world fires that would provide a better sense of how effective certain mitigation steps really are.
“We know from scientific experimentation and post-event [analysis] that when things are done super well, the house doesn’t burn,” Cope said. “How much does one thing contribute versus the other? Difficult to tell, both because of the lack of experience, and things either burned completely to the ground or they didn’t.”
Insufficient data isn’t the only factor that makes pricing wildfire risk — and risk mitigation — complicated. For starters, humans have a major influence over wildfire behavior given that they can start fires, put them out and make buildings more or less resilient. That makes it more challenging to model fire because it requires modeling human behavior.
Also important are the nuances of wildfire mitigation itself. While removing vegetation or wooden fences around a home are steps in the right direction, homeowners must make a whole suite of changes to meaningfully reduce their risk, some of which have to be maintained over time.
And then there’s the reality that meaningfully reducing wildfire risk doesn’t just hinge on what one homeowner does — it depends on the actions of every homeowner in a given area.
“You have to do it to your home, your neighbors need to do it. And oh, by the way, vegetation grows back, so it also has to be recurring,” said Collins of APCIA. “There’s uniqueness there that is different from any other peril at this moment in time.”
Collins was among those who said the industry is working to get its arms around the issue so companies can more confidently do business in fire-prone areas — and incentivize homeowners to do what they can to draw down the risk.
That could entail offering customers discounts on their premiums for reducing their risk on a property- and communitywide level. California in 2022 became the first state to mandate those discounts. Insurance companies already have submitted those discounts for review, but the majority have not yet been approved by the state.
“Companies do support risk-based pricing. And so if you have a higher risk, there’s generally a corresponding higher cost. But if you lower your risk, you will generally see your rates go down,” Collins said.
But that won’t happen overnight.
“It may take time as we have more data and more confidence in what the value of what that discount should be,” she added.
