AI, Insurtechs and the Coming Wave of US Climate Migration
Plus, the Bitcoin ETF passes a milestone.
Number of the Week: $125 million (explanation below)
AI, Insurtechs and the Coming Wave of US Climate Migration
By Holly Sraeel
When the World Meteorological Organization projected in May that global temperatures would likely hit record highs in the next five years, most people—climate scientists aside—were not prepared for that probability to hit immediately.
Yet that’s exactly what happened: July is on pace to become the single hottest month on Earth on record. Extreme climate change, coupled with El Niño conditions, means that 2023 now has an 81% chance of becoming the warmest year on record. The heat is so bad that protection from unprecedented record temperatures must become a “human right,” said climate scientist Brenda Ekwurzel in an MSNBC interview.
That sentiment applies to safe shelter. Sweltering heat is causing more frequent wildfires in western states, particularly in California where residents face limited home insurance options and steep premiums. Already in crisis, insurers could soon see losses climb in Vermont, New Hampshire, Maine, and parts of Minnesota, Wisconsin, Michigan, Alaska and the Pacific Northwest because of elevated wildfire risks.
Climate change is strengthening hurricanes along the East and Gulf Coasts. Florida has been battered, prompting 13 insurers to abandon the state. These exits could spread to northeastern states if losses from severe storms and catastrophic flooding continue to pile up. A recent storm produced up to two months of rain in two days in Vermont while also drubbing parts of New York, Connecticut and Pennsylvania.
Outside of high-risk states, homeowners rarely consider the impact of climate change on insurance premiums, even in areas emerging as climate hotspots. Between 2022 and 2023, U.S. climate and weather disasters have caused $209.6 billion in damages, including severe storms ($48.5 billion), drought ($22.6 billion), flooding ($6.1 billion) and wildfires ($3.2 billion).
The mounting losses for insurers are unsustainable. Insurtechs like CAPE Analytics, ZestyAI and Kettle, all based in California, are using AI- and machine learning-driven analytics to give insurance companies the ability to more precisely model climate risks, predict the likelihood of losses, and accurately price and underwrite property and casualty (P&C) policies.
CAPE’s property intelligence analytics combine computer vision, machine learning and geospatial imagery to give P&C insurers predictive insights about property condition and characteristics, which are tested against a historical claims database with millions of claims, to help quantify exposure, vulnerability and risk mitigation at the individual property level.
Its risk analytics targeting wildfire, wind and hail make it possible for insurers to predict the impact that individual property characteristics have on structural vulnerability to weather events. “We can make statistically significant statements such as ‘a certain amount of vegetation within 5 feet of a home increases that home's vulnerability to wildfire by X%’,” said CAPE CEO Ryan Kottenstette. The company can then combine relevant property characteristics into scores to help insurers better assess risk and segment their pricing to reduce their risk exposure. Another benefit: The granularity of the analytics “trickles up” to the reinsurance industry, Kottenstette said, by reducing the uncertainty that reinsurers have in the carrier's portfolio risk they are underwriting.
The company has raised more than $75 million and has 80 enterprise customers, including some of the world’s largest insurance companies and banks. Its analytics are being used in the contiguous 48 states, Canada and Australia and are licensed through subscription-based plans.
Similarly, ZestyAI’s climate and property risk analytics combine aerial imagery, permit, transaction, weather and IoT data with AI, turning “more than 200 billion data points into comprehensive digital records and property-specific risk scores that are 10 times more accurate than incumbent models,” said ZestyAI chief product officer Kumar Dhuvur.
The company has raised $56 million and works with P&C insurers across North America. Its scoring gives insurers a more accurate picture of risk based on specific characteristics of a property and its likely resilience against climate and weather risks. For each score, insurers can view the top three reasons behind that score and if homeowners have a question about the level of risk they’ve been assigned, insurers are able to outline the factors driving that particular risk score. Homeowners can then take actions based on the scoring information provided.
Its wildfire model looks at the vegetation around a residential property, what the roof is made of, whether the parcel is sloped and other property factors. The company also has models that assess hail and wind, as well as factors such as roof condition and complexity, property debris, the presence of pools or decks, etc., so that insurers can make better underwriting decisions. “A rainy year followed by several years of drier-than-normal summers can lead to excess vegetation. That may slightly change the risk profile of a property,” Dhuvur said. “Likewise, heat and sun also affect roofs. By understanding the true status of a roof, insurers get a clearer picture of how well it will stand up to other severe weather conditions like wind or hail.”
Unlike CAPE and ZestyAI, Kettle launched with a primary focus on the $400 billion reinsurance industry and wildfire risks in California. With a 3x increase in $1 billion catastrophes over the past 15 years and a 60% drop in the average ROI in the reinsurance industry, it was clear the models were no longer working, according to Kettle COO Nathaniel Manning.
Kettle’s deep learning and AI-driven models are trained to look for patterns in more than 3 petabytes of data around catastrophic wildfire events. Since the company began running its model in 2020, ~20,500 properties have burned in CA, 19,934 of which—97.24%—were in the top 25% riskiest areas in the state, according to Kettle’s model. The company now covers wildfire risk in the lower 48 states, with the Pacific Northwest and Colorado showing the most demand outside of CA. As Manning sees it, Kettle’s job is to accurately capture the risk so that insurers’ pricing is fair for policyholders; those who live in areas that are prone to higher risks will pay more, while those in areas with lesser risk will see that reflected in their policies.
Beyond reinsurance, Kettle now also underwrites insurance, focusing on areas most exposed to climate risk where P&C insurers have pulled out. “The big folks leaving is going to drive a ton of properties from the admitted [market] into the [excess and surplus] market,” Manning said, the latter of which is where Kettle is active.
Kettle has raised $29.7 million in funding. Last month, in partnership with Reask, it launched a parametric hurricane (tropical cyclone) reinsurance product in Florida.
Given the intense interest in AI and signs of catastrophic climate change everywhere, is more venture capital for insurtechs on the way? Global insurtech funding was up 37.6%, quarter on quarter, from $1 billion in Q4 2022 to $1.39 billion in Q1 2023, according to Gallagher Re’s Global InsurTech Report. Early-stage funding inched up 3.8% quarter on quarter (despite an 18.3% drop in early-stage deals), and 37.7% of insurtech deals—by individual deal counts, not funding amount—were in the “early-stage incubation” category. Meanwhile, mega-round funding comprised 12.9% of total Q1 2023 insurtech funding, the lowest contribution rate since Q1 2020.
But perhaps the biggest takeaway from the report is this: If 2021 was an “anomalous year” driven by mega rounds, then it’s possible “that 2022 was in fact a return to the trajectory that pre-2021 trends were otherwise heading.” Gallagher Re’s Fibonacci sequence projection on 2021 from the prior years’ data equals approximately 450 deals totaling $7.5 billion, 2022 totals 515 deals and $8.1 billion, and 590 deals totaling $8.5 billion for 2023. If Gallagher Re’s projections for 2023 hold true, then insurtech funding is set to pick up (see chart):
The only certainty of climate change is that things are going to get much worse if commitments to drastically reduce carbon emissions by 2030 and hit net zero by 2050 are not met. It’s not looking good. The increasing severity of climate change in the U.S. and its impacts on housing markets, the built environment, business, population (out-migration and in-migration), and state/subnational economies are sobering.
None of this should come as a surprise. Still, many Americans have been slow to accept the stark reality of the risks posed by climate change, including whether or when to leave areas likely to face more climate disasters and losses. “Americans have been conditioned not to respond to geographical climate threats as people in the rest of the world do….By comparison, Americans are richer, often much richer, and more insulated from the shocks of climate change. They are distanced from the food and water sources they depend on, and they are part of a culture that sees every problem as capable of being solved by money,” ProPublica’s Abrahm Lustgarten, investigative climate reporter, observed in his gripping three-part series on global climate migration produced in partnership with The New York Times Magazine.
Grist climate journalist Jake Bittle, author of The Great Displacement: Climate Change and the Next American Migration, makes a distinction between migration and climate displacement, noting that, in the past, migration was typically motivated by an “economic reality” and marked by a voluntary movement with a “specific destination in mind.” While acknowledging that climate change is causing migration, Bittle sees the aggregate movement in the US today as somewhat “chaotic” in that people didn’t want to leave and, in some cases, climate displacement deemed to be temporary occurred multiple times.
Climate migration and displacement will soon become a bigger part of American life, with millions of people affected and consequential social, financial and economic impacts. People who have the means and desire will migrate from high-risk areas, while those who lack the resources to permanently relocate or are incentivized to stay in high-risk areas will find themselves periodically displaced. “Migration is an adaptation strategy that is costly—both financially and, in some cases, socially, culturally and emotionally. Not all individuals and households will have the ability to pursue migration as an adaptation strategy, limiting their adaptation choices to technologies and behaviors available in their home origin,” said Tamma Carleton, assistant professor of economics at the Bren School of Environmental Science and Management, University of California, Santa Barbara. “For example, increased expenditures on cooling, fire mitigation, flood resilience, and other adaptation costs are likely to be faced by households remaining in the most climate-risky regions of the country.”
Could AI modeling accelerate climate migration in some regions of the country? Maybe—and that might not be a bad thing. “If AI results in increased precision of risk calculation results, and such risk signals are conveyed through market mechanisms to insurance policyholders, then this improved information could help people make their own best decisions about migration by accurately weighing costs and benefits of moving,” Professor Carleton pointed out. “In some cases, this may lead to higher outmigration, but it's not clear to me that this is problematic if outmigration is the best option for individuals facing rising climate risks. Distorting the insurance market by not allowing the insurance premiums to reflect true risk would lead to too many people located in high-risk areas, imposing private- and public-sector costs for recovery when disaster strikes.”
FINvestments
🦈Number of the Week: Barry Silbert doesn’t have much to brag about these days, but it’s safe to say the Digital Currency Group CEO will take what he can get. CoinDesk, a once-obscure, inside baseball media startup covering cryptocurrencies and blockchain, which he bought for $500,000 in 2016, is now reportedly being sold to an investor group for $125 million, according to The Wall Street Journal. If only the math on this were that simple: In an ironic twist of fate in early November 2022, CoinDesk broke the story about balance sheet irregularities at crypto trading firm Alameda Research and its outsized reliance on FTT tokens, which were issued by sister company FTX, a crypto exchange, both of which were founded by Sam Bankman-Fried. The unusually cozy relationship between the two SBF entities unleashed panic in the crypto markets, causing the collapse of FTX, the indictment of Sam Bankman-Fried, and CoinDesk parent DCG’s crypto empire—once valued at $10 billion—to tank, including the Chapter 11 bankruptcy filing of Genesis Global Trading. At one point, Binance was said to be in talks to acquire CoinDesk, according to Blockworks, but the talks stalled. If the $125 million sale of CoinDesk closes, it falls short of an earlier high estimate of $300 million, but soundly clears the $75 million the company was reportedly more likely to fetch.—Holly Sraeel
🦈The US hit a regulatory milestone this week, when the Securities and Exchange Commission (SEC) approved for comments a proposal for a spot Bitcoin exchange-traded fund (ETF) from Valkyrie. It’s not clear why this proposal slipped through when the SEC has blocked similar proposals for years, but presumably a year from now the market will be flooded with Bitcoin ETFs. Intriguingly, the SEC made the point that some investors have been buying the stocks of Bitcoin-heavy companies (such as Tesla), and that a Bitcoin ETF was probably safer and more logical. —James Ledbetter
Holly Sraeel is an award-winning business journalist, editor, media executive. She is a former SVP, president and group editorial director of SourceMedia (now Arizent), a content strategist focused on global financial services, fintech and sustainable business, and a contributing writer at FIN. Email her at holly@finnewsletter.com.