The industry loss warranty, known as an ILW, is the most disconcerting instrument of the family, because it breaks with the principle that governs all insurance: it does not look at the loss of the party buying it. What triggers it is the level of insured loss suffered by the whole market following an event, measured by a recognized index. If that industry loss exceeds an agreed threshold, the seller pays the buyer a pre-agreed amount, whether the buyer lost a lot, a little, or nothing.
The wording is of a simplicity that explains the instrument's success. An ILW is described by three data points: the event covered, the industry loss threshold, and the amount paid. There is no notification, no survey, no argument about what the underlying policy covers, and no run-off period. Settlement follows publication of the index estimate, so weeks rather than years. For a buyer that needs capacity fast, mid-season, it is sometimes the only instrument available.
Its most widespread form is binary: the threshold is crossed and the whole amount is due, it is not and nothing is due. That discontinuity is brutal and entirely deliberate, it is not a drafting flaw. There are variants known as layered, where payment follows the industry loss between two thresholds rather than falling at once, and variants that make payment conditional both on the index being crossed and on the buyer having suffered a loss of its own, which brings the instrument closer to ordinary reinsurance.
That last variant is worth pausing on, because it says what the instrument is as a matter of law. A purely index-based ILW, with no condition of own loss, resembles a wager on an event whose consequences the buyer need not suffer, and that resemblance has accounting and regulatory consequences depending on the jurisdiction: a contract that transfers no insurance risk is not treated as reinsurance. The own-loss condition exists largely for that reason, and it changes the classification of the deal as much as its economics.
Who sells, and who buys. Sellers are most often funds specializing in insurance risk, drawn to a liquid, standardized instrument whose exposure is described in two numbers and compares from one contract to the next. Buyers are insurers and reinsurers seeking to cover a catastrophe exposure quickly, to complete a program where one layer found no taker, or to protect against a late-season peak. It is an over-the-counter market operating at the speed of a financial one.
What the buyer gives up in exchange for that speed is named and not negotiable: basis risk. The gap between its own loss and the industry loss that triggers payment can run either way, and it runs the wider the more its portfolio is concentrated on one area or one type of risk. A regional insurer whose market share is tiny within the national index buys an instrument whose trigger depends almost entirely on what happens to others.
Hence a practical rule practitioners apply before looking at the price. An ILW is chosen by comparing the geography and composition of one's own portfolio with those of the chosen index, not by comparing two prices. Two ILWs at the same threshold and the same amount are not worth the same to two different buyers, and the same ILW is not worth the same to one buyer from year to year if its portfolio has moved. That is the exact reverse of an indemnity layer, which follows the insured wherever it goes.
An insurer present in a single coastal region buys in June 2026 a binary ILW of 25 million euros, triggered if the national market's insured loss exceeds 8 billion euros on a windstorm event. In October, a storm hits the country: the index publishes an estimate of 9.1 billion, and the insurer's own loss comes to 6 million euros. In December, a second storm devastates its region: its own loss reaches 31 million, and the national index publishes 5.4 billion. What is the outcome of the two events?
The two events illustrate both directions of basis risk on the same contract in the same year, which is rare but perfectly possible. On the first, the 8 billion threshold is crossed since the index publishes 9.1, so the 25 million is due in full, whatever the insurer's own loss: it collects 25 million against a 6 million loss, a positive basis of 19 million. That is not a windfall without consequence, and it is the point accounting and sometimes the regulator will look at: an instrument paying far beyond the loss suffered raises the question of how much insurance risk is really transferred, and that is precisely what the own-loss variant exists to correct. On the second, the threshold is not crossed since the index publishes 5.4 billion, so nothing is due, even though the insurer suffers its heaviest loss of the year at 31 million. That is negative basis risk, and it is the instrument's structural problem: a portfolio concentrated on one region can be devastated by an event that stays modest on a national scale. The lesson for the renewal is not that the ILW malfunctioned, it did exactly what was written. It is that the index was badly matched to the portfolio: a regional index, or a lower national threshold, would have tracked this house's loss experience more closely, and that comparison between one's own geography and the index's geography is made before the price is discussed.
- 01An industry loss warranty triggers on the whole market's insured loss, measured by an index, and not on the loss of the party buying it.
- 02It is described by three data points: the event, the industry threshold, the amount paid. No notification, no survey, no run-off period.
- 03Its most widespread form is binary, and that discontinuity is deliberate: the threshold is crossed and everything is due, or it is not and nothing is.
- 04The variant conditioning payment on the buyer's own loss exists largely for classification reasons: a contract transferring no insurance risk is not treated as reinsurance.
- 05What the buyer gives up in exchange for speed is basis risk, and it widens the more its portfolio is concentrated.
- 06An ILW is chosen by comparing your portfolio's geography with the index's, not by comparing two prices.