The Uninsurable Future
There is a phrase that is starting to terrify actuaries and delight litigators: the "uninsurable risk." When I started in this business, the term was reserved for nuclear war or asteroid impacts—events so catastrophic that no rational pool of capital could cover them. Now, it’s being applied to coastal property in Florida, or even entire zip codes in California’s fire zones. In 2023, State Farm and Allstate announced they would stop writing new homeowner policies in California, citing wildfire risk and construction costs. That wasn’t a market exit; it was a surrender to physics. The problem isn’t just that disasters are more frequent. It’s that the *tail* is fatter. We used to model a "100-year flood" as a statistical outlier. But we’ve now had three "100-year floods" in Houston in the last seven years. When the tail starts eating into the body of the distribution, the entire concept of pooled risk breaks down. Insurers are left with two choices: raise premiums to astronomical levels that make coverage pointless, or withdraw from the market entirely. Both options are terrible, but they reveal a deeper truth: the private insurance market is not designed to absorb the systemic, correlated shocks that climate change is now delivering. I remember a conversation with a risk manager from a mid-sized utility company earlier this year. He told me, only half-jokingly, that his company’s liability insurance now costs more than their actual maintenance budget. That inversion—where the price of protection exceeds the cost of prevention—is a market signal that regulators should be screaming about. Instead, we see states like Florida creating "Citizens" insurance, a state-backed insurer of last resort that is, frankly, insolvent by any honest accounting. The public sector is being dragged into subsidizing a risk that private capital has correctly identified as unpriced. That’s not a sustainable equilibrium; it’s a ticking fiscal bomb. What bothers me most, though, is how we talk about this. We say "uninsurable" as if it’s a property of the risk itself. It’s not. It’s a property of our *models*. We have the data to know that Miami Beach will be underwater—literally—by 2100. But because our mandates are quarterly and our models are backward-looking, we kick the can down the road. The consequence is that the people who need insurance most—those in high-risk zones—are the ones who get priced out first, often lower-income families who can’t relocate. Climate risk is not just an environmental problem; it’s a profound equity problem, and the insurance industry is the lens through which we can see it, all too clearly. ##Actuarial Blind Spots
Let’s talk about the models. I spend a lot of time looking at Monte Carlo simulations and catastrophe (CAT) models, and I can tell you with confidence: they are all lying to us, just not intentionally. The core issue is that traditional actuarial science relies on historical frequency and severity data. But climate change is a non-stationary process. The past is not a prologue; it’s a distraction. When you build a model that thinks a Category 5 hurricane is a 1-in-50-year event, but the ocean temperatures that fuel those storms are now at a baseline that makes them a 1-in-10-year event, your model is not just wrong—it’s dangerously wrong. I recall a specific project at JOYFUL CAPITAL where we were stress-testing a portfolio of catastrophe bonds. The bond prices were based on modelled expected losses, and the models were using historical wind speed records from the 1980s. When we re-ran the analysis using sea surface temperature projections from the IPCC’s RCP 8.5 scenario, the expected losses doubled. Doubled. Yet the market was still pricing those bonds as if nothing had changed. That gap between modelled risk and physical reality is the actuarial blind spot, and it’s a multi-billion dollar hole waiting to swallow someone. Part of the problem is regulatory. Rating agencies and state insurance commissioners require that rates be "actuarially justified," which often means they must be based on credible, historical data. But that regulatory framework was built for a stable climate. There’s a profound misalignment between what the law demands (evidence of past losses) and what the future demands (projections of new conditions). I remember telling a colleague, "We’re driving a car looking only in the rearview mirror, and the road ahead is on fire." He laughed, but he didn’t disagree. The industry is slowly starting to wake up, though. Some of the more sophisticated reinsurers, like Swiss Re and Munich Re, have begun to incorporate forward-looking climate scenarios into their models. But the adoption is uneven, and the methods are still nascent. For every forward-thinking actuary, there are ten who are still using the same spreadsheet they were using in 2005, just with more rows. The challenge is not just analytical; it’s cultural. Actuaries are trained to be conservative, to trust the data. But the data is no longer trustworthy. We need a new paradigm—one that embraces uncertainty rather than trying to eliminate it. ##Transition Risk and Stranded Assets
While physical risk gets the headlines—fires, floods, hurricanes—the quieter, creepier risk is the transition risk. This is the risk that as the world moves toward a low-carbon economy, entire asset classes will become worthless. Think about it: if we actually achieve net-zero by 2050, which I hope we do, then the value of every fossil fuel reserve on any company’s balance sheet will be stranded. And who insures those assets? You guessed it: the insurance industry. I was at an industry conference in London last year, and a panelist from a major oil and gas insurer made a comment that stuck with me. He said, "We don’t insure carbon; we insure *temporary* carbon." It was a half-joke, but it was also a confession. The underwriting community knows that the long-term viability of their fossil fuel book is questionable, but short-term profits are just too tempting. This is the classic tragedy of the horizon, where the time horizon of a CEO (3-5 years) is vastly shorter than the time horizon of the climate crisis (30-50 years). Transition risk hits insurers from two sides. On the investment side, insurance companies hold trillions in assets, and a significant chunk of that is in energy companies, utilities, and industrial firms that are carbon-heavy. If those valuations collapse due to policy changes, carbon taxes, or technological disruption, the solvency of insurers is directly threatened. On the underwriting side, they provide coverage for the construction, operation, and liability of fossil fuel infrastructure. As litigation increases—we see more and more municipalities suing oil companies for climate damages—insurers are on the hook for defense costs and settlements. There’s also a more subtle, behavioral effect. Transition risk creates political risk. When governments impose carbon pricing, or mandate energy efficiency standards, that changes the risk profile of insured properties. A building that was compliant yesterday might be obsolete tomorrow. This creates a patchwork of regulatory uncertainty that makes underwriting incredibly difficult. How do you price a commercial property policy when you don’t know if the building will be legal to operate in ten years? You either price in massive uncertainty, which kills the market, or you ignore it, which kills you. There is no comfortable middle ground. ##The Data Deluge
Here is where I get to my own swim lane: data strategy. The irony of climate risk is that we have too much data, not too little. Satellite imagery, IoT sensors, hyper-local weather models, flood plain maps, property-level granularity—the raw material is there. But the insurance industry, for all its talk of "big data," is still woefully behind in actually using it. Most carriers are running legacy systems that are older than the actuaries using them. The data is siloed, messy, and often unstructured. At JOYFUL CAPITAL, we’ve built what we call a "climate data fabric"—a layer that ingests disparate data sources, cleans them, and unifies them into a single analytical platform. The goal was to give our underwriting partners a real-time view of climate exposure at the property level, not just at the zip code level. The difference is stark. Zip code averages hide a lot of variation. Two houses on the same street can have vastly different flood risk due to elevation, drainage, and soil composition. When you move to parcel-level data, the pricing becomes more accurate, and more importantly, it becomes fair. You’re not penalizing a homeowner on a hill because of a neighbor in the valley. But the challenge isn’t just technical; it’s methodological. Standard regression models fall apart when dealing with extreme climate events. You need machine learning algorithms that can handle non-linear interactions, that can learn from sparse event data, and that can extrapolate beyond the historical record. This is where AI comes in. But here’s the dirty secret: AI models are only as good as their training data. If you train an AI on thirty years of weather data, it will learn that the world is the way it was, not the way it will be. Garbage in, gospel out, as we sometimes say. The real frontier is in *scenario generation* using Generative AI. Instead of asking "what happens if a flood hits this area," you can now ask "what happens if a flood hits this area *given* a 2°C warming scenario, coupled with property-level vulnerabilities, and a specific failure of a local levee?" That kind of what-if analysis is the future, and it’s just starting to emerge. But it requires a level of computational power and cross-disciplinary expertise (meteorology, hydrology, finance, computer science) that most insurance companies simply don’t have in-house. That’s why partnerships with fintech and data analytics firms are going to be the competitive battleground of the next decade. ##AI-Driven Pricing
So how do we actually use AI to solve the pricing problem? I want to be clear: I’m not a techno-optimist who thinks AI will save us. But I do think AI can soften the blow. The key insight is that AI allows for *continuous underwriting* rather than snapshots at policy renewal. Currently, a home insurance policy is priced once a year based on when the policyholder signs. But climate risk changes month by month, even week by week. A wildfire risk model that is accurate in January is stale by August. We worked with a specialty agricultural insurer to implement a dynamic pricing model for crop insurance. Instead of a fixed premium for the growing season, we created a model that adjusts the premium every two weeks based on soil moisture, temperature forecasts, and commodity prices. The result? The insurer could offer lower premiums to farmers in low-risk areas while turbocharging premiums for those in high-risk zones, all in real time. The adoption was slow at first—farmers hate unpredictability—but the aggregate claims ratio improved by 14% within two growing seasons. That’s the power of granularity. However, there’s a dark side to AI-driven pricing. It can easily become discriminatory. If we’re not careful, AI will price out entire communities based on variables they can’t control. Correlation is not causation, and an AI model might find that "postal code 90210" is correlated with low risk, but that’s just a proxy for income. We need to build fairness constraints into the models from the start, not as an afterthought. This is why at JOYFUL CAPITAL, we insist on an "ethics review" for every pricing model we deploy. It’s not just about regulatory compliance; it’s about maintaining social license to operate. And let’s not pretend the models are perfect. I once saw a model that was pricing flood risk for a coastal city in the Southeast. It was incredibly sophisticated, using LIDAR elevation data and neural networks. But it had a bug: it didn’t account for sea-level rise. The model assumed the coastline was static. When we flagged it to the developer, he said, "Well, the client didn’t ask for sea-level rise." That quote sums up the state of the industry—we build models that answer the questions we think to ask, but the climate crisis is fundamentally about the questions we haven’t thought to ask yet. AI can help, but only if we program the right questions into the system. ##Green Insurance Products
Amidst all the doom and gloom, there’s a growth area that genuinely excites me: green insurance products. These are policies that actively incentivize climate resilience and sustainable behavior. For example, some insurers now offer premium discounts for homeowners who install solar panels, green roofs, or flood-proofing measures. Others have introduced "pay-as-you-drive" auto insurance, which reduces premiums for people who drive less, cutting both emissions and risk. These are not just marketing gimmicks; they are fundamentally changing the relationship between insurer and policyholder. I think the most promising area is parametric insurance. Unlike traditional indemnity insurance, which pays out based on actual losses, parametric pays out when a specific trigger is met, like wind speed exceeding 125 mph or rainfall exceeding 3 inches in 24 hours. This structure is ideal for climate risk because it allows for instant payouts, no claims adjuster visits, and no lengthy litigation. After Hurricane Maria devastated Puerto Rico in 2017, the government paid out a parametric policy within two weeks, while traditional claims took years. That speed is critical when people have lost everything. But parametric insurance has its challenges, too. Basis risk is the big one—the risk that the trigger is met but you don’t actually suffer a loss, or vice versa. If a farmer buys a drought policy triggered by rainfall, but a local irrigation system saves their crops while their neighbor loses everything, the farmer gets paid despite no loss, while the neighbor gets nothing despite a loss. That’s messy. Yet we’re seeing innovation in this space, using AI to create more precise triggers based on satellite vegetation indices or soil moisture levels rather than a single weather station. It’s not perfect, but it’s a start. What I’m most excited about is the concept of "resilience bonds." These are insurance-linked securities where the coupon is tied to the insured’s investment in proactive resilience measures. If a city invests in flood walls and the flood walls work, the bond pays a higher return. If they don’t invest, the return is lower. This flips the financial incentive structure so that we are paying for prevention, not just for losses. It’s a small market today, but I believe it’s a preview of where the industry is headed. We need to move from being *reactive financiers of disaster* to *proactive investors in resilience*. That’s the only game-changing move available. ##Investment Strategy Shifts
Insurance companies are not just underwriters; they are also among the world’s largest institutional investors. The global insurance industry manages around $35 trillion in assets. That’s a war chest. Historically, the investment strategy was conservative—bonds, real estate, infrastructure. But climate risk is forcing a fundamental rethink. A bond in a fossil fuel company is now viewed as a risky asset, not because of credit rating (which might be fine on paper), but because of the transition risk we discussed earlier. If the carbon bubble bursts, that bond is worthless. I’ve been involved in several asset-liability matching exercises for insurers, and the tension is always the same: long-duration liabilities (like future pension payments or policyholder claims) need to be matched with long-duration assets. But long-duration assets are typically in real estate and infrastructure—exactly the sectors most exposed to climate physical risk. A commercial real estate portfolio with a strong footprint in coastal Florida is a time bomb. The yield might look great today, but the storm surge in 2040 will wash it away. This is pushing insurers toward a new asset class: climate-aligned investments. That includes green bonds, sustainable infrastructure, and even investments in climate adaptation technology. The returns are often slightly lower than conventional assets, but they offer a form of "double dividend"—financial return plus reduced systemic risk. I recall a sovereign fund we advised who was initially hesitant to invest in mangrove restoration along their coast. The financial model showed a negative return on a cash-flow basis. But when we ran it with a co-benefit analysis—storm surge reduction (which lowers insurance costs), carbon sequestration (which has market value), and tourism enhancement—the net present value turned positive. We need more of this kind of holistic thinking. However, I caution against "greenwashing" in the investment space. Just because a bond is labeled "green" doesn’t mean it’s climate-resilient. I’ve seen green bonds issued for infrastructure projects that are themselves located in flood zones. That’s not environmentalism; that’s taking a risk and painting it green. The industry needs standardized, rigorous frameworks for assessing the climate alignment of investments. The Task Force on Climate-related Financial Disclosures (TCFD) is a step in the right direction, but its implementation is uneven. Until we have mandatory, comparable, and audited climate disclosures across all asset classes, the investment leg of the insurance industry will remain stumbling in the dark. ##Regulatory Pressures
You can’t talk about climate risk and insurance without talking about regulators. And here, there’s a strange split personality. On one hand, regulators are demanding that insurers do more to address climate risk. On the other hand, they’re actively preventing insurers from pricing it correctly because they don’t want to be accused of allowing "unaffordable insurance." This is a classic catch-22. Regulators want insurers to be solvent, but they also want premiums to be politically acceptable. Those two goals are increasingly incompatible. I was recently in a meeting with a state insurance commissioner who confided to me that they were under political pressure to keep premiums low in wildfire-prone areas. "If I approve the rate increases the actuarial table demands, I’ll lose my job," he said. So instead, they privately pressure the insurers to "find other savings." That doesn’t work. The laws of physics don’t care about political survival. The result is a facade of affordability that masks systemic insolvency. It’s a dishonest game, and the players know it. The good news is that the regulatory tide is starting to turn. The European Union’s Solvency II directive has been amended to include climate risk in the Own Risk and Solvency Assessment (ORSA). In the US, the National Association of Insurance Commissioners (NAIC) has released a framework for climate risk disclosure. But the devil is in the details. Disclosure is not the same as action. An insurer can disclose all their climate risk and still do nothing about it. We need regulators to mandate capital charges for climate-exposed assets and require evidence of scenario analysis. This isn’t just about protecting policyholders; it’s about protecting the financial system from cascading failures. Insurance is the backbone of commerce. If the backbone breaks, everything breaks. I also foresee a significant role for central banks. The Network for Greening the Financial System (NGFS) has been vocal about how climate risk poses a threat to financial stability. I expect we’ll see capital requirements that specifically target climate risk within the next decade. This will hurt in the short term—insurers may report lower profits as they set aside more capital. But in the long run, it’s the only way to ensure the industry survives the century. A insurance sector that is honest about climate risk is a insurance sector that will have the reserves to pay claims when the big one hits. That’s worth more than a quarterly earnings beat, don’t you think? ##Conclusion and the Road Ahead
Let’s be clear: the impact of climate risk on insurance is not a future problem. It is a present, active, and accelerating reality. We are seeing withdrawal of private insurance from high-risk areas, we are seeing actuarial models fail, we are seeing a radical reshuffle of investment portfolios, and we are seeing the emergence of new, data-driven tools that might make it all manageable. But the clock is ticking. The insurance industry has a unique position in the climate crisis. It is both the diagnostician and the patient. As a third-party risk manager, it holds the keys to incentivizing or disincentivizing certain behaviors across the entire economy. If insurance gets climate risk right, it can accelerate the transition to a low-carbon economy. If it gets it wrong, it will face a wave of insolvencies that dwarfs the 2008 financial crisis. I don’t have easy answers. But I know that the way forward requires a tight collaboration between climate scientists, data scientists, regulators, and financial professionals. It requires us to break down the silos that have historically separated weather forecasting from actuarial science, and to invest in the computational infrastructure to make sense of it all. It requires an honest conversation about what risks are insurable, what risks should be borne by the public sector, and what risks we simply cannot bear at all. And above all, it requires us to be humble. I’ve been on projects where I thought I had the perfect model, only to be humbled by a "once-in-a-century" event that happened twice the same year. We are not in control. But we can be prepared. We can build systems that are robust to uncertainty, that adapt as the climate adapts, and that put resilience at the core of every premium we charge and every investment we make. That, I believe, is the only way forward., ##JOYFUL CAPITAL’s Perspective
At JOYFUL CAPITAL, we view climate risk and insurance not as a compliance exercise but as a data strategy opportunity. Our role is to build the analytical foundations that allow insurers to move from raw data to actionable insight, and we are committed to democratizing access to these tools. We believe that the integration of AI, satellite observation, and localized climate models will fundamentally transform the underwriting process over the next five years. We are also vocal about the need for industry-wide data standards. In our work with dozens of insurance and financial clients, we see the same problem repeatedly: data is collected in different formats, stored in different silos, and queried with different semantics. This makes efficient risk aggregation nearly impossible. Our solution is to advocate for and co-develop open-source data platforms that allow for seamless data sharing between stakeholders, always with privacy and security at the core.Finally, we are investing heavily in what we call “climate scenario intelligence”—not just the static stress tests that regulators ask for, but dynamic, constantly updating scenarios that reflect the latest reality. We plan to release a public index next year that ranks state-level insurance markets by their climate risk-adjusted resilience. We want to provide a public good that helps policymakers, insurers, and citizens make better decisions. The climate crisis is the single biggest challenge to the insurance industry in its history. But we firmly believe that with the right tools, the right mindset, and the right courage, the industry can not only survive but emerge as a leading force in building a resilient, sustainable future.
.png)
