The insurer's soundness, or the last link nobody discusses

This whole certification has examined risks carried by a third party: a buyer, a bank, a state. One remains that is never discussed and is nevertheless the last link in every chain of reasoning, that of the insurer itself. A cover is only a promise to pay later, bought from a company that can deteriorate between signature and loss, and the exporter that spent three months choosing its buyers often chose its insurer in half a day, on price. That imbalance is the stranger in that an insurer's failure is the only one striking the whole book at once.

One must first see that this risk is not symmetrical with ordinary credit risk, because it materializes at the worst moment. A credit insurer deteriorates when general loss experience rises, that is, during a crisis, that is, exactly when its insureds are notifying. The correlation is structural and works against the insured, like the currency correlation seen elsewhere: one discovers one's carrier's fragility when one needs it, not three years earlier when one could have moved.

What can be looked at is more accessible than one thinks and requires no analyst. Insurance undertakings publish indicators of how well they cover their regulatory requirements, and a carrier whose margin approaches its threshold will see its supervisor intervene long before it defaults. Financial ratings exist and say something, provided they are read as a dated opinion rather than as a guarantee. And the carrier's concentration matters as much as its size: an insurer heavily exposed to a sector or region one shares with it is a poor insurer for oneself, however excellent it is in general.

Reinsurance then changes the reasoning and one must know in which direction. It is what allows a mid-sized carrier to withstand a major loss, so it is protective; but it is not the insured's contract, which has no rights against reinsurers and generally does not know their identity. An insurer paying badly because its reinsurance panel is arguing is no less bound to its insured, but it is slower, and that slowness appears in no public indicator. That is why a settlement reputation, hard to measure and easy to ask about, is worth more than a figure.

Then the practical question: what does an insured do whose insurer deteriorates mid-contract. Three routes exist and none is comfortable. Cancelling means finding cover elsewhere, and a new insurer will not take on buyers already in arrears, which leaves the deteriorated part of the book unprotected. Waiting for expiry is reasonable if expiry is near and the deterioration slow. And spreading, that is, placing part of the programme with a second carrier, is the only route that builds something, but it is put in place before the crisis and not during: at the moment one needs it, nobody takes half of a book in difficulty.

One must add what protects nonetheless, and what distinguishes insurance from most receivables. Insurance undertakings are supervised, technical provisions are controlled, and an insurer's failure almost always proceeds through an organized portfolio transfer rather than an abrupt liquidation: contracts are taken over by another carrier, often on worse terms but with continuity. An insured therefore rarely loses everything; it loses time, terms, and sometimes access to a country the successor does not write.

The discipline comes to three steps costing an hour a year. Look, once, at the regulatory coverage indicator and the rating of one's carrier, not to make a science of it but to spot a deterioration from one year to the next. Ask one's broker what settlement times are observed at that carrier, information it holds and nobody publishes. And raise the question of spreading once the programme becomes significant, because the only moment when part of it can be placed elsewhere is the moment one does not need to.