Commercial insurance is a bet that the buyer usually loses and is happy to lose. You pay a premium, most years nothing happens, and the money buys the certainty that the year something does happen will not end the company. Almost every business accepts that trade without examining it.
It gets examined when the premium stops looking like a price for risk and starts looking like a price for being in a category. Companies with clean loss histories have spent several renewal cycles absorbing increases driven by what happened to somebody else in their industry, and at a certain point the finance director asks the obvious question: what if we kept it?
Keeping the risk means keeping the profit and the loss
A captive is an insurance company the parent owns, into which it pays premiums for risks it chooses to retain rather than transfer. If the losses come in below the premium, the difference stays inside the group instead of leaving it. If they come in above, the group funds the shortfall out of its own balance sheet, which is the entire point and the entire risk.
The structures that work share a shape. The captive takes the predictable, high-frequency, low-severity layer — the fender benders, the small property claims, the routine liability — where the company's own experience genuinely is the best predictor, and buys commercial reinsurance above an attachment point for the catastrophic tail that would otherwise be existential. Retaining the boring layer is a financing decision. Retaining the tail is gambling.
What is drawing mid-sized companies in now is not the tax treatment, which is where captives got their reputation and most of their trouble. It is the pricing environment: the same repricing that has made insurance behave like a second inflation for households has run through commercial lines, and a company with genuinely better-than-average losses has been paying for the average.
The discipline required is the part most groups underestimate. A captive needs capital that cannot be swept back to the parent when cash is tight, actuarial reserving that survives an auditor, claims handling that is not simply the operating business marking its own homework, and a board that will decline to write coverage the parent wants at a price the parent likes. Every failure mode is a governance failure rather than an underwriting one.
Timing is the other trap. The moment a captive looks most attractive is after a long stretch of low losses, which is also the moment a company's own history is least informative about what it will owe. Groups forming one while equipment finance tightens and every other credit line is harder should be clear that they are adding a call on liquidity precisely when liquidity is scarce.
Cyber is the line where this gets genuinely difficult. It is the coverage companies most want to retain, because the premiums have risen fastest, and the one where their own loss history says least about the next event — which is why cyber insurers have become the de facto regulators of corporate security in the first place. Retain that layer and the company loses the underwriter who was, in practice, enforcing its controls.
