Insurance Float, Explained

Key takeaways

• Insurers collect premiums now and pay claims later; the money held in between is float, capital they can invest.
• A disciplined insurer can run near break-even on underwriting and still profit from investing the float.
• The danger is underpricing risk to grow float: it works until a bad year of claims arrives.
• Any business that collects cash before delivering, whether through deposits, prepayments or gift cards, enjoys a small version of the same advantage.

The timing gap that becomes an asset

Insurance looks like a simple bet: customers pay premiums, the insurer pays claims, and profit is whatever premiums exceed claims and expenses. That picture misses the engine. Premiums arrive today; claims are paid later, sometimes months later (auto damage), sometimes decades later (liability injuries that unfold over years). In between, the insurer holds an enormous pool of money that belongs, in an actuarial sense, to future claimants, but sits, in a practical sense, in the insurer's investment account.

That pool is float. It behaves like a loan from policyholders with no fixed repayment date, and, crucially, if the insurer prices its policies well, the "interest rate" on this borrowed money can be zero or even negative: policyholders effectively pay the insurer to hold their money.

The scale is what makes this a business model rather than a curiosity. An insurer writing steady business holds float continuously, because as old claims are paid out new premiums flow in, so the pool is permanent even though every dollar in it is temporary. Warren Buffett's shareholder letters have explained for decades that this is why Berkshire Hathaway is, at its core, an insurance operation: float is the low-cost investment capital that funds everything else.

The two engines: underwriting and investing

A property-and-casualty insurer runs two linked engines. Underwriting is the insurance itself: pricing risk, selecting whom to cover, paying claims. Its report card is the combined ratio: claims plus expenses, divided by premiums. Under 100% means the insurer made money on the insurance alone (an underwriting profit); over 100% means it paid out more than it took in.

Investing is the second engine: the return earned on the float while it waits to be paid out as claims. Here is the model's elegance. An insurer with a combined ratio of exactly 100% is running the insurance at break-even, which means it is holding billions of other people's dollars at zero cost and keeping every dollar of investment return on them. Even a combined ratio slightly over 100% can be rational: paying a 2% effective rate to hold float beats many alternatives for raising capital, as long as the investments earn more.

This dual-engine structure explains insurer behavior that otherwise looks strange: ferocious attention to claim reserves (misjudge future claims and your float was never yours), sensitivity to interest rates (the investment engine's output moves with them), and the industry's pricing cycles as competitors fight for premium volume.

How the model blows up

The float model has a specific, recurring failure mode: underpricing risk to grow the pool. Premium volume today is certain and visible; claims are later and estimated. A management team hungry for growth can cut prices, win volume, watch float swell, and report success for years, because the bad claims from underpriced policies arrive on a delay. When they arrive, the insurer discovers its reserves were fantasy, the float it invested was never really its own, and the hole compounds: it must pay yesterday's underpriced claims out of tomorrow's weaker position.

Insurance history is a graveyard of exactly this story: aggressive growth, reserve releases praised as discipline, then a reserving crisis. Long-tail lines (liability, workers' compensation) are the most dangerous because the feedback loop is slowest: you can misprice for a decade before the truth surfaces.

The disciplined posture, preached in Berkshire's letters and practiced by the durable underwriters, is the opposite: price risk properly even when it means shrinking during soft markets, reserve conservatively, and treat float growth as the byproduct of good underwriting rather than the goal. In other words: the float is only cheap capital if the insurance underneath it is sound.

Float thinking for ordinary businesses

You do not need to be an insurer to use the principle. Any business that collects cash before delivering value holds a small float, and the cumulative effect on cash flow is real. Deposits and retainers: contractors, agencies, and custom manufacturers who collect up front are financing the work with the customer's money instead of their own. Prepayments and annual plans: a SaaS collecting a year in advance holds eleven months of unearned cash, which is why annual-plan discounts are usually worth it to the seller. Gift cards: retailers hold the cash now; some cards are redeemed much later and a fraction never at all. Memberships and subscriptions billed ahead of usage run the same physics, as do marketplaces that hold seller balances for days.

The design question for an operator is simply: where in my model could cash legitimately arrive earlier than the cost it corresponds to? Each yes shortens your cash conversion cycle, the same mechanism covered in the working-capital briefing, and at the extreme it turns growth from a cash consumer into a cash generator.

The honesty caveat mirrors the insurer's: money collected before delivery is an obligation, not income. Businesses that spend their deposits as profit are running a tiny reserving crisis of their own: the gym that sells lifetime memberships to fund this month's rent, the contractor whose next deposit pays for the last job. Float is leverage, and leverage points both directions.

Put it to work

Look for float-like dynamics you can add honestly: deposits on custom work, annual prepay with a discount, retainers, memberships billed ahead. Track the obligation side as seriously as the cash side: money collected before delivery is borrowed, not earned. The cash-conversion-cycle tool shows what earlier collection does to your numbers.

Sources & references

Linked entries open the named source directly. Entries without a link say exactly what kind of reference they are — and how to check them yourself.

• Berkshire Hathaway — shareholder letters (float explained by its most famous user)
• Insurer annual reports via SEC EDGAR