By Ron Dembo
Most institutions still evaluate climate adaptation the way they'd evaluate a compliance expense: what's the average annual loss, what does the fix cost, does the arithmetic pencil out. CCOT's argument is that this framing misses the entire point of why the investment matters.
Three structural habits drive the miscalculation.
First, expected-loss analysis averages across every simulated future, which by construction buries the catastrophic years inside a smooth number — and it's precisely the catastrophic years where adaptation earns its keep.
Second, most models price hazards one at a time, when the real damage usually comes from chains: a drought hardens soil, which worsens flood runoff, which overwhelms drainage, which knocks out power. Treat those as separate, independent events and the combined loss estimate comes in too low.
Third, conventional scenario analysis typically picks two or three pathways to represent the future. Against a validated set of 2,000-plus simulated pathways, that's not a simplification — it's discarding 99.9% of the distribution, concentrated exactly in the extremes that matter most.
The consequence compounds: undercounting the tail doesn't just produce a slightly-too-low number. It systematically understates adaptation's value precisely where that value is highest — the point at which a loss threatens solvency, breaches a covenant, or triggers a capital call.
The central move in CCOT is an analogy, treated formally rather than loosely: paying for adaptation today is financially identical to buying a protective put option on your own balance sheet.
You pay a known amount now (the capex); in exchange, your downside beyond a certain point is capped. Once you accept that structure, the entire toolkit of options pricing — volatility analysis, optimal timing, staged exercise — becomes directly applicable to a capital-expenditure decision that most institutions currently run through a simple payback calculation.
| **Component** | **Financial Protective Put** | **Climate Adaptation Investment** |
| What's at risk | Asset the investor already holds | Physical infrastructure / balance sheet |
| Source of the downside | Market price decline | Climate-driven capital losses |
| The premium | Put premium paid upfront | Upfront adaptation spend |
| The payoff | Avoided market loss | Avoided climate loss |
| Effect of more uncertainty | Higher volatility → higher option value | Higher climate uncertainty → higher option value |
The framework's headline formula compresses this into one line:
Net option value = expected-loss savings + tail-risk capital benefit − upfront cost.
The first term is what conventional analysis already captures: the average annual loss you avoid, multiplied out over the planning horizon.
The second term is what conventional analysis misses entirely: when adaptation shrinks your worst-case exposure, you can hold less regulatory or economic capital in reserve against it, and that freed-up capital has real value — typically priced at 8–12% a year for a financial institution.
The third term is simply what you spend. When the total is positive, adaptation clears its hurdle on a risk-adjusted basis. When it's negative, the paper is explicit that adaptation may still be required by a regulator or a covenant — but at least the true cost is now visible rather than assumed.
Standard real-options theory carries a well-worn piece of advice: when the future is uncertain and a decision is hard to reverse, waiting has value, because you might learn something that changes the answer. Applied naively to climate adaptation, that logic suggests uncertainty is a reason to delay.
The paper's central result inverts this. Climate uncertainty doesn't spread evenly across outcomes the way ordinary market volatility does — it widens the tail specifically, without necessarily moving the average. Because adaptation's payoff is concentrated in that same tail, a wider tail increases the value of protection faster than it increases the value of waiting.
Run the numbers, and the threshold at which the model says “invest now” actually falls as uncertainty rises. In the paper's worked example, a 20% increase in the variance of the loss distribution — a moderate widening, not an extreme one — lowered the investment trigger by roughly 11% and pulled the recommended timing forward by an estimated year to eighteen months.
There is further refinement: alongside ordinary day-to-day climate variability, it separately models sudden, discrete shocks — the kind associated with tipping points like an ice-sheet collapse or a major ocean-circulation slowdown, rather than gradual drift. Building those jumps into the model doesn't soften the result. It reinforces it: accounting for tipping-point risk makes the case for earlier investment stronger, not weaker.
The framework also offers a non-obvious explanation for a pattern regulators and researchers have already observed: banks and insurers tend to adapt ahead of the corporate borrowers and counterparties who sit on their balance sheets. The usual explanations reach for culture — better governance, stronger risk appetite, more sophisticated boards.
CCOT's result doesn't need any of that. Two institutions holding identical physical assets, facing an identical loss distribution, will still reach their “invest now” threshold at different times if their cost of capital differs — because the value of releasing regulatory capital against a smaller tail scales directly with how expensive that capital is to hold in the first place. A large, systemically important bank facing a wide funding spread gets more value from the same flood barrier than an unrated corporate borrower does, purely as a function of balance-sheet structure. The paper's own illustrative ranges put globally significant banks and insurers at the high end of this capital-cost spectrum and unrated corporates at the low end — which, if the model holds, predicts exactly the adaptation-timing gap already visible in practice.
For real estate specifically, this has a direct and slightly uncomfortable implication: a REIT and its lender may have systematically different adaptation timelines built into their financing structure alone, independent of who actually cares more about the risk. Worth knowing before that gap turns into a covenant dispute.
**For insurers and reinsurers: **the framework's tail-risk measure translates directly into claims-tail reduction, which is the same language reserve release and premium relief are already priced in.
**For asset managers and REITs: **net option value is built to sit next to NPV in a standard capital-budgeting stack, and staged investment is priced as a sequence of options rather than an all-or-nothing commitment — useful for boards reluctant to approve a full multi-year programme in one vote.
**For banks and lenders: **the model draws an explicit line from adaptation to probability of default (via reduced balance-sheet volatility) and loss given default (via reduced expected loss on collateral) — the two numbers credit committees actually move on.
**For regulators: **the paper proposes a minimum bar for adequate disclosure: an institution that can only produce a hazard score, and can't compute a comparable net option value, hasn't yet demonstrated it understands the financial consequences of its own exposure.
The Numerical Illustration — Read With One Caveat
Both papers use the same worked example: a multi-hazard manufacturing facility in the Philippines. Without adaptation, expected annual loss is $2.8M, but the worst 5% of simulated outcomes reach $42M — a fifteen-fold gap between the average year and the bad one. With a $2.2M adaptation investment, expected loss falls to $0.9M and the worst-case figure falls to $8M. Expected-loss analysis alone would value this at roughly a 4.3x return. Folding in the tail-risk capital benefit lifts that to a return of roughly 7x, with just over half of the total value coming from the tail component that conventional analysis assigns zero.
The caveat: this exact example, with identical figures, appears in both papers, and neither one states whether it's a real audited facility or a stylised illustration. Treat it as illustrative until RiskThinking.AI confirms otherwise — the mechanism it demonstrates is sound either way, but the specific multiple shouldn't be quoted as an observed real-world result.
Old Approach vs. CCOT, Side by Side
| **Dimension** | **Conventional Appraisal** | **CCOT** |
| Loss basis | Two or three scenarios | 2,000+ simulated pathways |
| Core metric | Expected annual loss / simple payback | Net option value (expected loss + tail capital benefit − cost) |
| Hazard treatment | Single-peril, independent | Correlated multi-hazard chains |
| Timing | Static now-or-never decision | Formal trigger rule, updates as data arrives |
| Effect of uncertainty | Ambiguous / usually ignored | Explicitly lowers the investment threshold |
| Link to regulatory capital | None | Direct mapping to Basel III/IV, Solvency II, IFRS S2 items |
The reframing at the centre of both papers is simple to state: adaptation is not a cost, it's the premium on an option, and the option's value is driven overwhelmingly by a tail that conventional analysis can't see. Whether or not every formal result survives peer review in its current form, that reframing is doing real work — it gives boards, lenders and regulators a common unit (a risk-adjusted return, not a hazard score) for a decision that's currently made on instinct or compliance minimums as often as on numbers.
To request access to the full paper, please complete this form: Access Paper
A financial framework that values climate adaptation spending as a protective put option rather than a sunk cost, using a fully stochastic loss model instead of a handful of scenarios to price both the average benefit and the tail benefit of protection.
You pay a known amount today (the adaptation capex) in exchange for capping your downside if a catastrophic climate event hits — the same trade as buying insurance on a financial position, just applied to physical infrastructure and balance-sheet risk.
Because greater uncertainty in this context mainly widens the worst-case tail rather than shifting the average outcome, and adaptation's payoff is concentrated in that same tail. The benefit of protection grows faster than the benefit of waiting, so the model's investment trigger falls as uncertainty rises — the opposite of standard real-options intuition.
η is the model's regulatory-capital cost coefficient — how expensive it is for a given institution to hold capital against climate risk. Institutions with a higher cost of capital get more financial value from the same physical adaptation, which is why the model predicts they'll reach their investment threshold earlier.