Loss reserves and the uncertainty that hides in the balance sheet

• Case reserves: an estimate for a specific claim that has been reported. An adjuster looks at the mesothelioma lawsuit and pencils in a number.
• IBNR (incurred but not reported): money for claims that have happened but nobody has told the insurer about yet. In 2020, thousands of future mesothelioma victims were already sick and did not know it. The insurer still owes those claims. IBNR is the reserve for the unknown.

Walk one claim from cradle to grave

Follow the mesothelioma claim through the accounting:

2005, occurrence. The worker is exposed. The insurer collects premium and books a profit for the year. It also sets up IBNR for exposures like this, but the number is small and highly uncertain.

2026, reported. The diagnosis comes. The claim moves from IBNR into a case reserve. The adjuster estimates, say, a mid six-figure settlement.

2027 to 2030, developed. Litigation drags. Medical costs climb. Co-defendants go bankrupt, so this insurer's share rises. The case reserve gets revised upward twice.

2031, paid. The claim settles for far more than the original 2026 estimate.

Here is the punchline. The final cost was not booked in 2031. It kept flowing back to hit earnings in every year the estimate was revised. When reserves for old years turn out to be too low, the shortfall shows up as adverse development, a fresh charge against current profit for a policy sold long ago.

The loss triangle: how actuaries see the future

Actuaries do not estimate claims one by one. They watch how whole groups of claims behave over time using a loss development triangle.

Group all claims from the same accident year (the year the loss occurred). Then track how much you have paid on that group as time passes. The result looks like a triangle because older years have more history.

Here is a simplified paid-loss triangle (figures in millions, illustrative only):

Read a row left to right: accident year 2021 had paid 100 after 12 months, then 180, growing to 275 after 60 months. The claims "develop" as they settle.

The empty cells are the future. To fill them, actuaries calculate development factors, the ratio from one age to the next. From 12 to 24 months, 2021 grew 100 to 180 (a factor of 1.80), 2022 grew 105 to 190 (about 1.81). Average those and you get a factor you apply to younger years that have not aged yet.

Multiply the factors out and you project each accident year to its ultimate loss, the total it will eventually cost. Subtract what you have already paid, and you get the reserve. This is the chain ladder method, the workhorse of reserving.

For a clear, free walkthrough of the mechanics, the Casualty Actuarial Society keeps public educational material at casact.org.

Where the estimate goes wrong

The triangle assumes the past predicts the future. When that assumption breaks, reserves break with it.

Trend shifts. Suppose medical inflation or jury awards ("social inflation," the tendency for legal settlements to rise faster than general inflation) accelerate. Yesterday's development factors understate tomorrow's costs. Reserves set on old patterns come up short.

Slow-emerging losses. Asbestos is the classic case. Insurers in the 1970s had no factor in their triangle for a mass tort that would keep developing for fifty years. The U.S. insurance industry has paid out tens of billions on asbestos, far beyond original estimates (widely cited, though totals are still debated).

Changing mix. If you start writing riskier policies, your old triangle no longer describes your new book.

The danger is that all of these push in the same direction: reserves too low. And low reserves make current profit look better, which is exactly why regulators scrutinize them.

🎬 [VIDEO: "Loss Reserving and the Chain Ladder Method" — youtube.com — a concise visual explainer of how actuaries build a loss triangle and project ultimate losses]

How this hits the income statement

Reserves are not a footnote. They drive the headline profitability metric.

The combined ratio measures underwriting performance: losses plus expenses, divided by premiums. Below 100% means underwriting profit. Above 100% means the insurer paid out more than it took in.

Losses in that ratio include the change in reserves. So if an actuary decides last year's reserves were 200 million too low and strengthens them, that 200 million becomes a current-period loss. The combined ratio jumps. A year that looked profitable can flip.

This is why analysts obsess over reserve development disclosures. Every P&C insurer publishes a schedule (in the U.S., Schedule P of the statutory annual statement) showing how prior years' reserves have developed. Consistent adverse development is a red flag: it suggests the company has been under-reserving, meaning past profits were partly an illusion.

The reverse also happens. Favorable development (reserves that turn out too high) releases money back into current earnings. Some critics watch for insurers that lean on reserve releases to smooth earnings in weak years.

Judgment, incentives, and honesty

Reserving is technical, but it is also human. Management picks a number within a range the actuaries provide. That number affects reported profit, executive bonuses, and the stock price.

The pressure is obvious: a lower reserve today means higher profit today. That is why there are guardrails.

• The appointed actuary must sign a formal opinion each year that reserves are adequate. Their professional reputation is on the line.
• Regulators require reserves to be booked at a best estimate, not the low end of a range.
• Rating agencies and auditors independently pressure-test the triangles.

None of this removes the uncertainty. It just forces it into the open. A well-run insurer treats its reserve range honestly and books toward a prudent point within it. A poorly run one anchors low, harvests short-term profit, and hopes the tail behaves.

It usually does not.

Reading an insurer through its reserves

If you want to judge an insurer's real quality, ignore a single year's profit and look at the pattern:

• Does the reserve development schedule show steady favorable development, or repeated adverse surprises?
• How much of the book is long-tail, where estimation risk is highest?
• Is current profit driven by underwriting, or by releasing old reserves?

An insurer that consistently over-earns by releasing reserves may simply be running out of cushion.

Key Takeaways

• The largest liability on a P&C insurer's balance sheet, the loss reserve, is an estimate, not a settled bill. For long-tail lines it can be wrong for years.
• Reserve errors flow into current earnings as adverse (or favorable) development, so a change in an old estimate can flip today's reported profit into a loss.
• The loss triangle and chain ladder method project ultimate losses from historical development patterns, but they fail when trends like social inflation or slow-emerging mass torts break the pattern.
• The combined ratio includes reserve changes, so watch reserve development disclosures (Schedule P in the U.S.) to see whether reported profits are real or borrowed from the future.
• Reserving involves human judgment and incentives. Honest insurers book toward a prudent best estimate; the ones that anchor low eventually pay for it when the tail catches up.