The cheapest car to buy is not the cheapest to keep
The economics of an old car look excellent right up until they do not. Depreciation has already happened to somebody else, the purchase price is small, and for a period the vehicle asks for nothing beyond consumables. Then a component fails that nobody stocks, and the cost of keeping the car stops being about labour rates and starts being about supply.
This is the part that rarely appears in any comparison of running costs, because it does not average smoothly. It arrives as a single event that can exceed the value of the vehicle several times over, and its likelihood rises with age in a way that is hard to predict for any individual car and quite predictable across a fleet.
Parts supply follows a lifecycle of its own
A manufacturer supports a model heavily while it is in production and for a period afterwards, and independent suppliers fill in behind for anything with enough demand to be worth tooling for. Common service items on a car that sold in large numbers stay available for a very long time, which is why a mass-market model from twenty years ago can still be cheap to service.
What disappears first is the awkward middle: model-specific trim, electronic modules, unique castings, anything that was fitted to one variant for two years. Those items are supported while the manufacturer holds stock and then not at all, and the market that replaces the manufacturer is dismantled vehicles, specialist remanufacturers and, increasingly, nobody.
A body panel is a shape, and a shape can be reproduced by anyone with the right equipment. A control module is a piece of hardware carrying software, often coded to a specific vehicle, sometimes requiring a manufacturer tool to marry it to the car. When that module fails on a car outside its support window, the repair is not a matter of finding a part but of finding somebody able to make a substitute work.
This has changed the shape of old-car ownership considerably. A vehicle of the previous generation could be kept running almost indefinitely by anyone competent with a spanner, because everything on it was mechanical and therefore repairable. Contemporary vehicles reaching the same age carry sealed assemblies, integrated modules and diagnostic systems, and their failure modes are often replace-only.
Waiting is a cost even when the part is cheap
A component sourced from a dismantler in another region, or backordered from a manufacturer, takes days or weeks. During that time the vehicle occupies a workshop bay or a driveway, and its owner is either paying for alternative transport or doing without. That indirect cost is invisible in any repair invoice and is frequently larger than the part.
It also interacts badly with insurance. Where a car is being repaired after a claim, delays in obtaining parts extend the repair, and any replacement vehicle benefit is usually tied to repair time within limits that a long backorder can exhaust. Where a car is simply old and broken, there is no benefit at all, and the household absorbs the whole of it.
The point at which repair stops being rational
Every ageing vehicle reaches a threshold where the cost of the next repair exceeds what the car is worth, and the temptation is to treat that as the decision point. It usually is not. The better question is what the repair buys: a car that will run for another two or three years without further large bills is worth spending on, even at more than its market value, because the alternative is buying somebody else’s unknown history.
Where the answer is that a large repair merely defers the next large repair, the arithmetic changes. Corrosion is the clearest example, because it does not stop, and structural rust is one of the few conditions that reliably ends a vehicle regardless of how well everything else works. A periodic roadworthiness test tends to be where that verdict is finally delivered.
What this means for how the car is insured
Two effects are worth knowing. First, the settlement after a total loss is based on what the vehicle was worth, not on what replacing it with something equivalent and reliable would cost, and on an older car those two figures diverge sharply. Second, parts availability influences whether an insurer repairs at all, because a repair requiring an unobtainable component becomes a total loss by default.
Neither is an argument for or against any level of cover, which is a judgement depending on the vehicle, the household and local law. It is an argument for holding a realistic figure in mind for what the car would actually be worth on the day it is written off, since that is the number the policy is built around and it is almost always lower than an owner expects.
