How a claim affects your premium — the record, the discount and the renewal

Nothing happens to your premium on the day of the crash. The price moves at renewal, by which time the accident is a long way behind you, and that gap is why the letter reads as though it arrived from somebody else’s file.

The question people actually ask comes before the claim rather than after it: “will making this claim cost me more than it pays?” It can be answered, though not with a number, and the answer has the shape of a sequence — the points at which an event becomes a record, a record becomes a price, and the price stops moving. Most of the surprises in this area come from not knowing which point you are standing at.

The call, which is already an event

A reported incident and a paid claim are not the same record, and in some markets both can sit on your file. An enquiry — you describe what happened, ask whether it would be covered, decide the repair is cheaper than the excess and leave it there — can be logged as a notification, and a notification can be something a future application asks you to declare. This is not universal, and it is not a reason to hide an accident from an insurer you are contractually obliged to inform. It is a reason to ask, before you describe anything, how the conversation is being recorded. An insurer that treats enquiries as claims will normally say so.

The file opens, and carries a number nobody has paid

Once a claim exists, the insurer attaches an estimate to it — a provisional figure for what the file is expected to cost, set before the repair is agreed and long before anyone has paid anything. It is an internal accounting device, not a valuation of your car, and it moves as the file develops.

It matters here because a file that is open is a file with an unknown cost, and a renewal that falls while the claim is still open may be priced against that unknown. Whether a claim is treated as open or closed, and what happens if it settles for far less than the estimate, is the sort of thing insurers handle by their own practice rather than by a published rule. Asking whether your renewal can be revisited once the file closes is a reasonable request.

The file closes, with a figure and a verdict attached

When the claim settles, it acquires the two properties that follow you: an amount paid and an attribution of fault. Fault here means the insurer’s own conclusion about responsibility, including a split, and whether it recovered its outlay from the other side. That conclusion, and not the police report or your own account, is what the record carries.

The amount paid, curiously, does less work than people assume. A total loss and a modest repair can sit on a claims history as entries of the same kind, distinguished by their figure but not by their weight in the way a driver would expect. That is one root of the feeling that the price movement is arbitrary: the thing that changed your life and the thing that scuffed a wing are, on the file, both a claim.

The renewal after it, where two separate things move

At renewal the claim is priced, and it is priced twice by two different mechanisms that people run together.

The first is contractual and you can read it. A no-claims discount, or a bonus-malus scale, is a schedule of discount steps earned by claim-free years and stepped back by a claim. It is written into your policy, the step is defined in advance, and the effect is arithmetic. In markets where the discount reaches a substantial share of the price, losing it can hurt more than anything the underwriter does — the discount you had built is often what is most at stake, and it is the part you can find out about before you claim, by reading the document or asking what a claim would do to your step.

The second is the reprice, and you cannot read it. Your claims history is an input to a rating model along with everything else the insurer knows about you, and the model returns a price. There is no line item. How much an insurer has to explain about that price, and to whom, is not something a general page can settle: the rules below name the supervisor for your jurisdiction, and the question belongs with it.

Fault matters to both, but not in the same way. A claim fully recovered from the other driver’s insurer often leaves the contractual discount intact, because the insurer ended up out of pocket for nothing. The history entry remains regardless. This is the case people find hardest to accept, and they are right that it is strange: the two mechanisms are answering different questions, and only one of them is asking who was to blame.

The renewals after that, and the insurer you move to

The record follows the driver rather than the policy. Switching insurer does not reset it — a new insurer asks about the past directly and, in many markets, can check what it is told against shared claims records. How far back the question reaches, and how long a shared record is kept, is set where you are rather than generally, and it is not something we are going to state as a general fact. It is worth asking before you answer an application rather than after.

What does not vary is the consequence of answering the question wrongly. A declared claim produces a higher price. An undeclared one produces a policy that an insurer may treat as misrepresented when you next need it, which is a different order of problem, and it arrives at the worst moment — after a loss, not before. A price you dislike is a price; a policy that does not respond is not insurance.

What is actually in dispute

Insurers will tell you that a claims history is among the better predictors they have of future cost, that this is a statistical statement rather than a moral one, and that pricing it is what keeps premiums attached to risk at all. Set against that is a driver whose parked car was hit, who did nothing, whose insurer lost nothing in the end, and who is nonetheless being charged more — a product that penalises its own use, which is a strange thing to sell.

We think the prediction is probably sound and the objection still lands, because they are about different things. Nothing about a model’s accuracy answers the question of whether a person who was hit while stationary should carry the cost of that accuracy, and that is a question about what the product is for. It would move us to learn that insurers priced the not-at-fault, fully recovered case as a separate category and could show why it still costs more — some say their own experience supports exactly that, and none of them show the working.

The number nobody can give you

Not what your claim will cost you. That figure exists only inside a calculation you cannot see, produced by a model whose inputs are not published, and it depends on an insurer’s appetite in a particular month as much as on anything you did. The part you can work out in advance is narrower and more useful: what the repair would cost you outright, what your excess takes off the claim anyway, what step your discount would fall by, and how many renewals the entry will be asked about. Those four answers are in documents you already have or can ask for in a phone call, and they decide most cases before the unknowable part is reached.