Claims reserve
What is a claims reserve: claims reserves explained
A claims reserve is the money an insurer sets aside to pay the remaining cost of a claim that has been reported but not yet fully settled. It is carried on the balance sheet as a liability, it is an estimate rather than a payment, and it is revised up or down every time better information about the claim arrives.
For insurers
See the damage before you send anyone
Venta Capture, a product of VentaVid, sends the policyholder a link. They film the damage on their own phone, guided step by step, and the evidence lands with the claim.
Every open file on your desk carries one. Added up across the book, those estimates are the largest number on a property and casualty insurer's balance sheet, and the one that is hardest to get right.
What does a claims reserve mean on the balance sheet?
Loss reserves sit in two layers, and claims teams only control one of them.
- Case reserves: the handler's estimate of the remaining cost of a specific reported claim. This is the claims department's number.
- IBNR: incurred but not reported. The actuarial provision for claims that have happened but have not reached the insurer yet, plus the expected development on files that are already open. This is the actuary's number.
- LAE reserves: loss adjustment expense, split into allocated LAE tied to a specific claim (a loss adjuster, an engineer, legal costs) and unallocated LAE covering the running of the claims function itself.
Incurred loss is the sum of what has been paid plus what is still reserved. Move a reserve and you move incurred loss, which moves the loss ratio, which moves the underwriting result. That is why reserving discipline is a financial control, not paperwork.
How claims reserves are set
Three methods do almost all the work, and most insurers use all three at different points in a claim's life.
- Formula or average reserve: a standard opening amount by claim type, applied automatically at first notification. Cheap, fast, and wrong on any individual file by design.
- Case by case estimate: the handler prices the actual exposure from the evidence in the file, usually once the damage or injury is understood.
- Roundtable or technical reserve: for large or complex files, a group sets the number against a documented rationale, often with legal or medical input.
Good practice ties every reserve movement to a documented reason. If the number changed, something in the file changed: a new estimate, a liability decision, a medical report, a set of images that showed the damage was worse than described.
Claims reserve example: how one motor file moves
A collision is notified on Monday. The system sets a formula reserve of 4,500 dollars, which reflects the average for that claim type and nothing about this vehicle. The customer's photos arrive on Thursday and show structural damage, so the handler raises the case reserve to 14,000. The engineer's report two weeks later confirms a total loss and adds a hire car exposure, taking the reserve to 21,500. Settlement is 20,300, and the file closes with a small release.
Three weeks of that movement was not new damage. It was the same damage becoming visible to the person holding the pen.
Reserve strengthening, releases and adverse development
When reserves set in earlier accident years turn out to be too low, the shortfall lands in the current year's result. That is adverse development, and it is where reserving errors become an earnings problem.
It is a live issue in liability lines. AM Best's 2025 P/C Snapshot, published July 2026, reported that US commercial auto liability took another 2.0 billion dollars in reserve deficiencies in 2025, mostly on the recent accident years 2023 and 2024. Favourable development works the same way in reverse: reserves that prove redundant are released and flatter the current year.
Neither is free. Both distort the loss ratio of the year they land in rather than the year the loss belongs to, which is why accident year and calendar year loss ratios can tell very different stories about the same book.
What distorts a claims reserve
- Stale files: a reserve that has not moved in nine months is rarely accurate, it is usually unattended.
- Thin evidence at notification: a reserve set on a verbal description and three unclear photos is a guess with a decimal point on it.
- Anchoring: handlers adjust reluctantly away from the opening formula number, so early estimates persist longer than the facts support.
- Reserving to authority limits: a reserve capped just under a referral threshold is a governance failure, not an estimate.
- Inflation and legal trend: repair costs, medical costs and settlement values move after the reserve is set, particularly on long tail claims.
- Missing recoveries: subrogation, salvage and contribution reduce net cost, and reserves that ignore them overstate the liability.
What claims reserves are confused with
A reserve is not a payment, and it is not a provision for unexpired risk. The unearned premium reserve covers cover you still owe on policies in force. The claims reserve covers losses that have already happened. Nor is a reserve a settlement authority: what a file is reserved at and what a handler may agree without referral are two separate controls, and treating them as one produces both under reserving and slow settlements.
The fastest route to more accurate reserves is better information earlier. Most of the movement in the example above came from evidence that could have existed on day one. That is the same lever behind claims cycle time, and it is why the quality of insurance claim documentation at first notification of loss shows up in the reserving numbers weeks later. Accurate remote triage at intake is what separates a formula reserve from an informed one.
