Two things can be true on the same afternoon: the wind gauge read 62 knots — two under the hurricane line — and the town was under four feet of water. Only one of those facts is in the contract.
That gap has a name. Underwriters call it basis risk, and it is the central fact of parametric insurance — not a defect in the paperwork, not a calibration bug someone patches next season. It is the price of speed, agreed to in advance by everyone who signs.
Here is the plain-language version: what parametric insurance is, why the money can arrive and still miss the place it was written for, and what an underwriter, an allocator, or a finance ministry can do about the part no trigger will ever cover.
what parametric insurance actually is
Parametric insurance is a contract that pays a pre-agreed amount when a measured index crosses a stated threshold — wind speed at a location, rainfall over a window, ground-shaking intensity, a river gauge reading, sea surface temperature, snow water equivalent. Payment is triggered by the measurement, not by the damage.
The moving parts are few, which is the point.
- The index. One agreed number, produced by an agreed reporting agency, on an agreed schedule.
- The trigger. The threshold at which the contract turns on — often a ladder, so a category 3 pays more than a category 1.
- The limit. The most the contract will ever pay, regardless of what happened.
- Settlement. The reporting agency publishes; the payment goes out. No adjuster walks the site. No one argues about depreciation on a roof.
That is the whole appeal. Traditional claims adjustment is the slow part of insurance, and slow money is the wrong shape for the first month after a disaster, when the useful spending is debris clearing, pumps, payroll, and getting crews in the water before the damage compounds.
Speed is the trade. To pay in two weeks instead of two years, the contract stops asking what you lost and starts asking what the instrument read.
Everything that follows is a consequence of that one sentence.
basis risk in plain language
Basis risk is the distance between the index and the loss at the place. The index is a proxy. The place is the thing you care about. They are correlated, never identical, and the contract settles on the proxy.
It cuts both ways, and both directions are real.
- The index misses low. The place is wrecked and the payment is small or zero, because the number stayed under the threshold. This is the one that ends up in the newspaper.
- The index misses high. The number fires, the money lands, and the damage on the ground turns out to be modest. Awkward for a private buyer, genuinely difficult for a public one, who now has to explain a payout in a place that looks fine.
Where the gap comes from is not mysterious. It is a short and well-known list.
| where the gap comes from | what it looks like on the ground |
|---|---|
| geography | The trigger geometry says the storm center passed outside the covered area. The eyewall — where the worst wind lives — crossed it anyway. |
| peril mismatch | The index measured wind. The loss was water: surge, runoff, a river that came over the bank two days later. |
| scale | The index is regional or gridded. The loss is a parcel, a reef, a neighborhood, a single treatment plant. |
| timing | The measurement window opens and closes on a clock. Compound events — a burn scar, then rain three weeks later — do not respect the clock. |
| data revision | Storm intensity gets reanalyzed after the season. The contract settles on the agreed source at the agreed time, not on the final scientific record. |
The reef pilots are the most-discussed nature version of this problem, and the anatomy of one real trigger — cat-in-circle, a modest limit, what the money can and cannot buy — is laid out in a payout is not a reef.
indemnity vs parametric vs modeled-loss
Three ways to decide what a contract owes you. Each one buys speed with certainty, or certainty with speed.
| indemnity | parametric | modeled-loss | |
|---|---|---|---|
| what triggers payment | Your verified actual loss | A measured index crossing a threshold | Losses run through an agreed catastrophe model using the event's real parameters |
| who determines it | An adjuster, sometimes a court | A calculation agent, applying the reporting agency's number | The model, run on your exposure file |
| typical speed | Months to years | Days to weeks | Weeks to months |
| basis risk | Lowest — you are paid for what you lost | Highest — the index is a proxy | Middle — model error rather than index error |
| what gets argued about | Causation, valuation, exclusions | Whether the number was measured correctly | Model version, exposure data, assumptions |
| best at | Making a specific balance sheet whole | Getting cash moving before anyone knows the total | Bridging the two for a large, well-mapped book |
| cannot do | Move fast | Match your actual loss | Escape the model's own blind spots |
None of these is the smart choice in general. They are answers to different questions, and sophisticated programs stack them: parametric for the first 30 days, indemnity underneath for the real number.
you might be thinking: better data closes the gap
Fair. And partly true. Denser sensor networks, satellite-derived footprints, dual triggers, hybrid structures with a payout ladder tied to two indices instead of one — all of it narrows basis risk, and the last decade of trigger design has narrowed it a lot.
But narrowing is not closing, and the reason is structural rather than technical. An index is a simplification chosen because it is cheap, fast, and hard to dispute. Past cheap, public, hard-to-dispute data, each step you take to make it track your actual loss more tightly makes it more specific, more expensive to verify, and slower to settle. Push that all the way and you have rebuilt indemnity, handed back the speed, and paid for the privilege twice.
Basis risk is not the failure mode of parametric insurance. It is the fee.
Which is why the honest version of the sales conversation is not "our trigger is tight." It is: here is the residual, here is roughly how big it is, here is who is holding it.
And to be clear, none of this is an argument against the product. Parametric insurance is one of the best available answers to the working-capital hole in the first month after an event, it has moved real money to places conventional cover would not touch, and the people building sovereign and conservation triggers are doing careful, useful work. The critique here is of the job, not the craft.
who is actually holding the gap
Basis risk does not disappear when a contract is signed. It moves to whoever did not get paid for it.
If you underwrite or cede. A parametric hedge is not indemnity relief. The recovery you booked and the claims you owe are different events, and the difference sits on your balance sheet in the quarter it happens. That residual belongs on the page next to the parametric line, not in a footnote.
If you are a government. The ministry that buys a sovereign parametric still owns the distance between the payout and the repair bill. Pre-arranged money is a genuine improvement over an emergency supplemental six months late — and it is still a line item about who writes the check, not about how big the check has to be. That second question is the one that compounds across seasons. Pre-arranged reactionary finance walks the whole sovereign toolkit.
If you hold the other side. On the capital-markets end of the same structure, basis risk is why a note can pay its coupon while a coast is wrecked, and why it can take a loss while claims turn out modest. Your outcome and the place's outcome are correlated. They are not the same event, and the paperwork never claimed they were.
the part no trigger covers
Read the comparison table again and notice what all three columns have in common. Indemnity, parametric, modeled-loss — every one of them answers the same question: who pays, and how fast, after.
Not one of them changes whether 62 knots becomes four feet of water in the town.
That outcome is decided somewhere else entirely: by the width of the mangrove belt, by whether the floodplain upstream is still allowed to flood, by soil that holds rain instead of shedding it, by forest structure on the slope above the intake. Those systems are the reason two identical wind speeds produce two very different loss numbers in two different places. They are also, almost everywhere, funded at a fraction of the loss they avoid — and never by the contract that pays when they fail.
This is not a pitch for a better index. We do not run a trigger, and there is no cleverer circle here drawn around the same reef. It is the other column on the page: the money that goes to the living system before the season, so the number the gauge eventually reads has less to do.
hold the place, not only the index
Ensurance is the cousin of the products above that runs on the other clock — funding for a named natural asset, now, rather than a payment conditioned on a future event. A certificate is a holdable claim tied to one named place, where the money funds protection now rather than compensation later. A coin is the indirect version: trading activity routes proceeds toward protection without naming a single parcel.
What that is not, said plainly, because the words in this post make it easy to mishear:
- It is not a parametric policy. Nothing pays you when a threshold is crossed.
- It is not a catastrophe bond, a sidecar, or a rated note with a documented coupon.
- It is not a substitute for cover. If you need cash in the first 30 days after a hurricane, buy the parametric.
And it has its own honest failure mode, which is worth naming before someone else does: the place can still get hit. What you bought was work done in advance — protection, restoration, stewardship on a specific piece of ground — not a payment afterward. There is no trigger to argue about because there is nothing to trigger. The instruments are live and the volumes are small. That is the accurate description of the stage we are at.
The useful framing is not either/or. Basis risk is the two-sided gap between the index and the loss at the place. Shrinking the loss is one of the few things that reaches the uncovered part. A mangrove does not close basis risk — it changes the loss the index is trying to stand in for.
what to do with this
If you underwrite, cede, or advise on placement — put the residual on the page. Then look at the balance-sheet case for investing in the systems that lower the number your triggers are chasing: solutions for insurers.
If you want the anatomy of one real trigger — read a payout is not a reef. It is the whole mechanism of a live parametric written for nature, including the limit and what the limit cannot do.
If you want a position in the other column — browse certificates and see what a named place looks like when the funding is present-tense. Small, live, and specific about which ground it is attached to.
This is educational content on instrument structure — not investment advice, not legal advice, and not an offer of any instrument.
the series
Catastrophe bonds pay after. Ensurance pays now. Six posts:
- what a catastrophe bond actually is
- insurance-linked securities are still insurance
- the payout that misses the place — you are here
- you can take this position without a 144a
- a sidecar funds the book. a certificate funds the source
- what you actually hold