A number that has to be guessed and is treated as known
Almost every quotation asks how many miles or kilometres the vehicle will cover in the coming year. It’s one of the strangest questions in the whole form, because it asks for a prediction about the future and then records the answer as a fact on which the contract is based. Nobody can know it. A new job, a house move, a relative falling ill or a change of school run rewrites it entirely.
Insurers ask anyway, because exposure genuinely matters. A car that spends most of its life stationary cannot be involved in many collisions, and one that covers a very large distance is exposed far more often. That relationship is real, and it survives every attempt to argue it away. What is less obvious is how much of the price it actually moves.
Distance is a weaker predictor than people expect
Risk per mile is not constant. A driver covering a long distance mostly on fast, well-engineered roads with predictable traffic is exposed to fewer of the events that generate claims than a driver covering a short distance entirely in dense urban traffic full of junctions, cyclists, parked cars and reversing manoeuvres. Claims frequency concentrates where conflicts are frequent, which is not where speed is highest.
That is why very low mileage does not always produce the reduction people anticipate, and why the difference between a moderate figure and a fairly high one is often smaller than the difference between two postcodes. Mileage is one input among many, correlated with other things the model already knows, and its independent contribution is correspondingly modest.
What makes mileage unusual among rating factors is that it leaves a trace. The odometer is read at every service and at every periodic roadworthiness test, and in many countries those readings are recorded in a database that anybody can query with a registration number. A claim also produces a reading, because an assessor writes it down when valuing or repairing the vehicle.
So a declared figure that bears no relation to reality is not a secret. It sits alongside a documented history that can be compared against it in a few seconds. Insurers do not audit every policy, since the cost of doing so would exceed the benefit, but they look when a claim arrives, which is precisely the moment a policyholder least wants a discrepancy to appear.
Being wrong by accident is different from being wrong on purpose
A genuine forecast that turned out low because life changed is not the same thing as a figure chosen to reduce a price. Most insurance law and regulation distinguishes between the two, treating an honest and reasonable answer as satisfying the duty even when events prove it wrong, while a careless answer attracts a proportionate remedy and a deliberate one attracts a much harsher outcome.
The practical protection is to answer the question that was asked, using the best information available on the day, and to tell the insurer when circumstances change materially during the year. A mid-term adjustment for a longer commute usually costs a modest additional premium. A claim declined for a mileage declaration nobody can defend costs the value of the car. The remedies available differ by jurisdiction, so the local rules govern.
Commuting distance is a separate question with its own purpose
Many forms ask separately how far you travel to work, and that is not redundancy. Commuting mileage is concentrated at the hours when roads are busiest and drivers are most tired or most rushed, and it is repeated on the same route in all weathers whether or not the journey is worth making that day. Discretionary mileage can be postponed; a commute cannot.
The same question also establishes the use class, and use is a cover question rather than a pricing one. A policy issued for social use only doesn’t become a commuting policy because the price would have been slightly higher. Driving outside the declared use can leave a claim unpaid and, in some jurisdictions, leaves the driver without the insurance the law requires.
Producing a number you could defend later
The most reliable method is to work from evidence rather than instinct. Two odometer readings a year apart, or the readings on consecutive test certificates, give a measured figure for how the vehicle has actually been used. Adjust it for anything you already know is changing, round it sensibly, and keep a note of how you arrived at it.
Estimating upward slightly is usually cheap insurance against an awkward conversation, because the effect on premium is generally small while the effect of an indefensible under-declaration can be severe. And if the number turns out badly wrong halfway through the year, saying so is the cheapest thing you will do all year. Insurers are entirely used to it, and the adjustment machinery exists for exactly this.
