A market where good history can't be proved is a market for lemons.

In 1970 the economist George Akerlof explained why the used car market punishes honest sellers. The paper was called The Market for “Lemons”, and it won him a share of the 2001 Nobel Prize in economics.

His argument is uncomfortably simple. The seller of a used car knows whether it has been looked after. The buyer does not. Since the buyer cannot tell a well-kept car from a neglected one, they refuse to pay a well-kept price - they offer something closer to what an average car is worth, because that is the risk they are actually taking.

The consequence is the part people miss. That average price is a discount charged to the person who did everything right, and a subsidy paid to the person who did not. And it gets worse over time: if the honest owner won't accept the average price, they keep the car instead of selling it. The good cars leave the market first, which raises the share of bad ones, which pushes the price down again. Economists call this adverse selection.

Akerlof also named the cure

The paper doesn't stop at the problem. It points to what it calls counteracting institutions - guarantees, warranties, brand names, licensing - mechanisms that let a seller credibly signal quality instead of merely asserting it. The word doing the work there is credibly. Anyone can claim a full service history. The claim is worthless precisely because it costs nothing to make.

That is the gap PeachStamp is built to close. Not a nicer place to store your receipts - a way to make a true claim about your car cost something to fake.

Where the name comes from

The bad cars in this literature are lemons - a word borrowed from used-car slang, popularised by a Volkswagen advert in 1960, a decade before the paper.

The good cars are peaches. Same car, same forecourt, same asking price, and no way for a buyer to tell them apart. That indistinguishability is the entire problem - a lemon is only worth what it is worth, but a peach is worth less than it is worth, and the difference is taken from whoever looked after it.

Which is also why the mark is a tick rather than a score. We are not grading cars into good and bad - the record does that, or fails to. All this does is make the difference visible to somebody who could not otherwise see it.

More than fifty years after Akerlof, an enthusiast who has spent thousands maintaining a car properly still has no reliable way to prove it.

The folder in the glovebox

The traditional answer is a wallet of invoices and a stamped service book. It works, right up until it is lost, water-damaged, thrown out by a previous owner, or simply disbelieved. Paper proves nothing about when it was written.

The manufacturer's record has a hole in it

Most brands moved to digital service records years ago. But an independent specialist - which is exactly where an enthusiast takes a car worth caring about - generally cannot write to the manufacturer's system. Their work shows up as an invoice and nothing more. The perverse result: the more discerning the owner, the worse the official record can look.

History checks look for the bad, not the good

HPI, carVertical and the rest are genuinely useful, and they answer a different question. They tell a buyer whether a car has been written off, stolen, clocked or has finance outstanding. That is the floor, not the ceiling. Nothing in those reports can demonstrate that you replaced the timing chain early, at your own expense, because you cared.

And a photo of an invoice is not proof

This is where most apps in this space quietly stop. Letting an owner upload a picture of a receipt into their own account creates a filing cabinet, not evidence - the owner controls the file, the date and the contents. A record only carries weight with a buyer if it was hard to fabricate after the fact.

A credible signal has to be expensive to fake and cheap to check. Everything here is built backwards from that sentence.

Say exactly how strong each claim is

Trust products fail by overclaiming. Every fact about a car carries a visible tier, and we never let a weak one dress up as a strong one:

• StatedTyped in by the owner. No evidence behind it.
• SealedA document was uploaded, fingerprinted and independently timestamped. Proof it existed on that date and has not been altered since.
• CorroboratedSealed, and its dates and mileage fit the official MOT record without contradicting it.
• ConfirmedThe garage that carried out the work has signed it off.
• OfficialTaken directly from DVSA or DVLA. Not editable by anyone.

Anchor it to something nobody can edit

Every MOT test records an odometer reading, once a year, for the life of the car, and the owner cannot change it. That gives us a skeleton of truth. An invoice claiming a service at 40,000 miles that sits between official readings of 52,000 and 61,000 is not suspicious - it is impossible, and we say so.

The same check runs the other way. Mileage that goes backwards between tests is flagged. Years with no record at all are shown as holes in the story, on your own car as much as anyone else's. A score that can only go up is worth nothing.

Make the proof worth having before you sell

A vault nobody opens is a chore. So the same record that convinces a buyer is also the thing you show off while you own the car: your garage, past and present, with the documents kept private and only their weight made public.

Uncertainty in this market is not a rounding error. It is priced, it is large, and it is paid by the wrong people.

Where the money is

The premium varies by who did the work: a full main-dealer history is generally worth 8–12% on a car under eight years old, a full independent history 5–8%. The penalty for having nothing is far heavier than the reward for having everything - which is precisely Akerlof's point restated in pounds.

How we would charge for it

Keeping records is free. That matters: the vault only becomes valuable once it is full, and nobody fills a vault they are being metered on. The moment worth charging for is the handover - when a complete, sealed, checked history transfers to the next owner and demonstrably moves the price.

A subscription for passive storage is the obvious model and we think it is the wrong one. It charges people during the years they get least from it, and it competes on price with apps that are free.

Why the UK, deliberately

The MOT is a national annual inspection with a public data feed - a free, tamper-proof mileage spine that most countries simply do not have. Building around it means the verification is stronger here than we could make it anywhere else. That is a narrow moat, but a deep one, and 7.6 million transactions a year is not a small pond.

Figures from SMMT used car market data (2026), and published UK trade estimates for service-history premiums including HPI. The interactive model in the lemons problem is illustrative and not drawn from transaction data - it demonstrates a mechanism, not a forecast.