One in five flats now sells for less than it cost

Every first-time buyer gets the same advice, usually from someone who bought in 1998. Get a flat. Get on the ladder. Build some equity, then trade up to a house. The flat is the first rung, and the ladder only goes one way.

We wanted to know how often that still holds. So we took HM Land Registry's record of every registered sale in England and Wales, 31.3 million of them, matched every home we could to its next sale, and read the two receipts side by side.

Since the start of 2024, 19.8% of flat resales came in below what the owner had paid. One in five. For houses the figure is 4.2%, roughly one in twenty-four. Flats made up 18.9% of all resales in that time, and 52.1% of the ones that lost money.

Why average flat prices hide the losses

The usual way to tell this story is with averages. Outside London, the median flat sold for £137,995 in 2007 and £175,000 in 2025, up 26.8%. The median house went from £172,500 to £280,000, up 62.3%.

Those numbers point the right way, but they compare different homes. The flats sold in 2025 are not the flats sold in 2007, and comparing averages of different homes measures the mix, not the market, the trap we described in the postcode average is a lie. An average also can't answer the question a buyer actually cares about: how often does a flat sell for less than its owner paid?

The record can, if you follow each home from one sale to the next. We matched each sale to the same home's next sale, kept pairs where the home was the same type both times and at least six months passed in between, and ended up with 13.6 million pairs of receipts for the same front doors. Every figure from here on is one home, bought and then sold.

The flat that never caught up

Here is the share of resales that came in below the previous price, by the year of the resale.

Before the financial crisis, almost nobody sold at a loss. Then 2008 hit flats and houses alike, and through 2012 around one flat in four and one house in five went for less than it had cost.

Houses then recovered. By 2022, only 2.2% of house resales came in under the purchase price. Flats bottomed out at 13.4% in 2017, and since 2022 the line has climbed every year. In 2025, 20.2% of flat resales lost money against 4.2% of houses, a gap of 16 points, about twice as wide as at any point in the crash itself.

When a flat does sell at a loss, the typical shortfall is £20,000, or 9.6% of what the owner paid. That is in cash terms, before stamp duty, legal fees, two lots of moving costs and inflation, every one of which makes the real picture worse.

Bought at the bottom, still sold at a loss

The easy explanation is timing: these must be people who bought at the top and are still underwater. Some are. Three in ten flats bought in 2007 and resold since 2024 still went for less. But the pattern doesn't depend on the year you bought. Pick any purchase year from 2004 to 2021 and flats resold since 2024 lost money more often than houses bought the same year, every time. Flats bought at the very bottom of the market, in 2012, sold at a loss about as often as houses bought at the very top, in 2007. The 2016 buyers make the point most plainly: 29.4% of their flats sold at a loss, with a typical total gain of 8.5%, against 2.0% of their houses, with a typical gain of 38.7%.

Nor is it that people sell flats sooner. They do, a little: the typical flat here changed hands after 8.8 years, the typical house after 9.7. Hold the period constant and the gap stays. Among homes owned for five to ten years, 25.7% of flats sold at a loss, against 2.1% of houses.

Price doesn't explain it either. Flats are cheaper, and cheap homes are supposed to swing harder, but among homes bought between 2014 and 2019 the dearer the flat, the worse it did. Flats bought for £300,000 to £500,000 resold at a loss 33.5% of the time, and above £500,000, 38.2%. Houses in the same bands came in at 2.0% and 3.8%. Under £100,000 the gap narrows, to 14.6% against 3.6%, but it never closes.

New builds, with their own reputation for losing value, are part of it. Flats first bought new resold at a loss 34.1% of the time since 2024. One south London block, first sold in 2017, shows how completely it can go wrong: all 28 of its flats resold since 2024 went for less than the previous owner paid, 27 of them by the people who had bought them new. They paid a median of £587,500 and got back £475,000. Take new builds out altogether, though, and second-hand flats still sold at a loss 16.6% of the time, against 4.1% for second-hand houses.

And it isn't a London story. London flats did a little better than the rest, 18.1% sold at a loss against 20.5% elsewhere. We ran the same comparison in every local authority with at least 200 flat resales and 200 house resales since 2024. There are 258 of them. In all 258, flats sold at a loss more often than houses.

Liverpool loses, Bristol holds

How big the gap is depends a great deal on where you are. These are the five worst and five best of the 82 districts with at least 800 flat resales since 2024.

In Liverpool, more than one flat in three changed hands for less than its previous price. In Bristol it was fewer than one in twelve. We haven't tested why each place lands where it does: local supply, who the flats were built for and how new the stock is are the likely candidates. London's boroughs are spread right through the list, from Tower Hamlets (27.4%, seventh worst) to Waltham Forest (8.4%, second best).

Why flats lose value more often than houses

The record shows that flats lag, not why. One common explanation can be checked against it; the rest are context.

• Houses get extended. A house owner can add a loft room and sell a bigger house than they bought; a flat leaves the same size it arrived. The record can't see extensions, so some of the house advantage here is likely to be owners' own building work rather than the market. But the option to extend is part of what a house gives you and a flat doesn't.
• The lease. Almost every flat is leasehold, with service charges and ground rent that keep rising after completion. The lease matters, but it can't be the whole story. The rare flats sold freehold, about 2% of flat resales and an unusual group, lost money 9.4% of the time: less often than leasehold flats, but still more than twice the overall rate for houses. Leasehold houses (4.8%) barely differ from freehold ones (4.2%).
• Building-safety paperwork. Since the Grenfell fire in 2017, questions about cladding and external walls can stall a flat sale or stop it outright, as we covered in what actually kills a house sale. A house doesn't carry that risk.
• Fewer landlords bidding. Landlords have long favoured flats. Since April 2016 they have paid a stamp duty surcharge on additional homes, raised from three to five percentage points in October 2024, and from 2017 higher-rate landlords lost much of their tax relief on mortgage interest. Since then, a growing share of homes has been bought by companies.

None of these is proven flat by flat here.

What this means if you're buying one

A flat is not a bad home. Millions of people live in flats they love, in places where a house was never on the table, and most flats in this record sold for more than they cost. The trouble is the story attached to them: that the flat will quietly grow and hand you the deposit for the next step. Since 2024, one in five has done the opposite.

So buy the flat for the flat. Holding on is no cure: flats sold after three to five years lost money 19.6% of the time, and after five to ten years 25.7%. Only the ten-to-fifteen-year group did clearly better, at 11.6%, and most of those owners bought in the trough after the crash. When you buy matters more than how long you stay. If your plan depends on selling at a profit in a few years, treat it as a bet and look hard at what moves the odds. Check what flats in the same block and street have actually resold for, not the local average. Know that new-build flats resold at a loss a third of the time. Put the lease length, ground rent and service-charge history among your first questions, not your solicitor's last ones, and ask about the building's external walls before you offer.

That is the record we check for any address in England and Wales: what this flat and its neighbours resold for, the tenure, the planning history, the building. You can ask about a specific property or look it up in Explorer. What a home is actually worth explains why sold prices beat estimates, and the 25-year cost of a leasehold does the sums on the lease.

The ladder is still there. One rung in five now gives way, so check what the flats around yours have resold for before you put your weight on it.

Methodology and limitations

Source. HM Land Registry Price Paid Data for England and Wales: 31,299,627 registered sales dated 1 January 1995 to 31 March 2026, analysed on 26 September 2026.

Building the pairs. We kept sales of detached, semi-detached and terraced houses and flats priced between £10,000 and £10 million, excluding other property types such as commercial sales, and kept one entry where an address had more than one on the same day. Each sale was matched to the next sale of the same address (postcode, house number or name, flat number and street). A pair counts only if the property type is the same at both sales, the sales are at least six months apart, the second sale is not flagged as a new build, and the second price is between a fifth and ten times the first, to remove obvious keying errors. That gives 13,608,200 pairs. "Sold for less" means the second price was strictly below the first, in cash terms.

The cuts behind the figures (resales from 1 January 2024 to 31 March 2026 unless stated).

• Headline: flats 19.8% sold for less (230,393 pairs); houses 4.2% (990,369). Detached 3.9%, semi-detached 3.5%, terraced 5.2%. Flats were 18.9% of resales and 52.1% of loss-making resales (45,605 of 87,547). Median shortfall on a loss: flats £20,000 (9.6%), houses £15,000 (7.7%). Median time between sales: flats 8.8 years, houses 9.7.
• By resale year (flats / houses): 2007 4.9% / 2.1%; 2009 27.4% / 19.3%; 2012 28.0% / 20.3%; 2016 14.5% / 8.8%; 2017 13.4% / 7.6%; 2022 14.2% / 2.2%; 2023 16.6% / 3.2%; 2024 18.7% / 4.1%; 2025 20.2% / 4.2%. 13.4% is the lowest flat figure after 2012.
• By purchase year, 2004 to 2021: flats 7.6% (bought 2012) to 31.1% (bought 2017); houses 1.2% (2012 and 2013) to 8.4% (2007). Flats were higher in every purchase year. Bought 2007: flats 30.0% (n=10,563). Bought 2016: flats 29.4% (n=14,271, median total gain 8.5%), houses 2.0% (n=50,394, 38.7%).
• By purchase price, bought 2014 to 2019 (flats / houses): under £100k 14.6% / 3.6%; £100k to £150k 15.1% / 1.9%; £150k to £200k 18.6% / 1.4%; £200k to £300k 26.6% / 1.4%; £300k to £500k 33.5% / 2.0%; over £500k 38.2% / 3.8%.
• By holding period (flats / houses): 3 to 5 years 19.6% / 5.7%; 5 to 10 years 25.7% / 2.1%; 10 to 15 years 11.6% / 1.6%.
• Tenure and new build: leasehold flats 20.0%, freehold flats 9.4% (n=4,748), leasehold houses 4.8%, freehold houses 4.2%. Flats first bought new 34.1% (n=42,233); second-hand flats 16.6% (n=188,160), second-hand houses 4.1%.
• The south London block is the only street address in England and Wales with at least 25 flat resales since 2024 where every one was a loss. It is a single development, first sold in 2017: 28 of 28 resales were losses, 27 of them by the original buyers, at median prices of £587,500 bought and £475,000 sold. We have not named it.
• Places: London flats 18.1%, elsewhere 20.5%. All 258 local authorities with at least 200 flat and 200 house resales show a higher rate for flats. The district table ranks the 82 with at least 800 flat resales; London's 31 boroughs among them run from 7th (Tower Hamlets) to 81st (Waltham Forest).
• Averages (all sales, not pairs): shown outside London because London's far higher prices would swamp a combined median. Flats £137,995 (2007) to £175,000 (2025), houses £172,500 to £280,000.

Limitations, stated plainly.

• Figures are nominal. We have not adjusted for inflation, stamp duty, fees or moving costs, all of which would push more sales into real losses.
• Matching on address is imperfect. Renumbered flats, renamed streets and inconsistent entries will break some pairs, and a house extended between sales still counts as the same house.
• The record we analysed does not separate standard sales from the Land Registry's "additional price paid" entries, such as repossessions and some sales to companies. They are a small share of sales, but they will be in some pairs.
• Sales from early 2026 are still being registered. They are about 7% of the pairs since 2024; on their own (flats 24.8%, houses 5.0%) they are provisional, which is why the year table stops at 2025. Leaving them out entirely, the headline becomes flats 19.5% (215,582 pairs) and houses 4.2% (918,370): still one in five.
• England and Wales only. Scotland and Northern Ireland keep separate registers.
• This study measures outcomes, not causes. The reasons discussed above are established drivers of flat prices, not findings about individual homes.

Every figure can be reproduced from the public Price Paid dataset using the rules above.