Do new-builds lose value? 253,000 UK repeat sales, 2010–2025

The property pages have repeated the same claim for years: buy a new-build and you'll lose 10–15% of its value the moment the next owner sells. It's a tidy soundbite. But until now, no UK publication has tested it at scale against the underlying transaction record.

We pulled every address in HM Land Registry's Price Paid dataset where the same property was sold first as a new-build between 2010 and 2019, and again as a non-new-build resale between 2020 and 2025. That gave us 253,139 matched repeat-sale pairs with an average gap of 7.4 years between the two sales. We then compared each pair's actual price change to what the UK House Price Index (UK HPI) recorded for the same local authority and same months — isolating the part of the gain that's specific to new-build first-owner resales from the part driven by the wider market.

Headline finding: across the 253,139 pairs, the new-build first-owner resales gained an average of +28.8% nominally, while the local HPI rose +37.8% over the same months. That's an average underperformance of −9.0 percentage points. Around 24% of pairs beat their local HPI; 52% underperformed it by 10 percentage points or more.

So the depreciation claim has empirical backing — but the size of the effect, and where it concentrates, is very different from the headline version most buyers hear.

Method, in one paragraph

Repeat-sales analysis is the standard way of measuring like-for-like price change because it holds the property itself constant — same address, same building, same plot. We constrained the sample to pairs where the resale was at least 2 years after the new-build registration and both sale prices were £30,000 or more (filters out the small number of Land Registry placeholder entries and same-day sub-sales). For each pair, we looked up the local authority's UK HPI value in the month of the new-build sale and the month of the resale, and computed an "excess return" — the pair's actual percentage gain minus the HPI's percentage gain. The HPI itself includes both new-build and resale transactions, so the excess return measures any first-owner premium decay relative to the market average, not relative to a hypothetical resale-only index. Data fetched from HM Land Registry Price Paid (loaded 2026-05) and the ONS UK House Price Index series (latest release: February 2026 print) via Homecost's data warehouse.

Holding period: the decay is gradual, not a cliff

The often-repeated "lose 10–15% in the first 5 years" framing assumes a single, front-loaded drop. The data shows something different — a slow, year-by-year drift below the local HPI that compounds with time held.

Three observations from the table.

First, the first three years actually look fine. At 2 years held the excess is essentially zero, and at 3 years held the new-build resale beat the local HPI by 6.7 percentage points on average. That contradicts the standard "drive it off the forecourt" framing entirely. The premium buyers paid on the new-build is not immediately marked down on first resale.

Second, the gap opens in years 4 through 7 — at a rate of roughly 1.5 to 2 percentage points per year. By year 5 the underperformance is around 6 percentage points, not 10–15. The widely-quoted figure overstates the early decay by a factor of two.

Third, the longer the property is held by the first owner, the larger the cumulative gap. At 10 years it's −18 percentage points; at 15 years it's −26. That's consistent with a slow erosion of the new-build premium rather than a one-off mark-down, and it lines up with industry surveys that find buyers are willing to pay more for "new" than for "five years old" but not measurably more for "five years old" than for "fifteen years old".

A note on the smallest hold-period buckets: the 15-year bucket only contains 1,108 pairs because our resale window ends in 2025 and our new-build window starts in 2010 — there isn't much room for longer holds. Read the long-end numbers with that caveat in mind.

Property type: this is mostly a flats story

The other widely-held belief about new-builds is that they all lose value at roughly the same rate. They don't. Splitting the 253,139 pairs by property type shows the underperformance is concentrated in one category.

Flats underperform their local HPI by 17.6 percentage points over an average 7-year hold. Detached, semi-detached and terraced houses all sit in a tight 5 to 7 percentage point band — within statistical range of each other and a long way away from the flats figure.

A handful of plausible explanations have been floated in the academic literature on UK off-plan housing (LSE working papers from the early 2020s, RICS occasional research). New-build flats are typically sold in identical-unit blocks where the resale market has many close substitutes, which pulls prices toward unit averages rather than premium pricing; the first owner pays for marketing, branding and finishings that the second buyer values less. They're also concentrated in regenerating urban locations where the pre-2020 development boom pulled forward demand, and many were sold to leveraged investors and Help-to-Buy users whose 2020–2024 sale timing coincided with the post-stimulus rebalancing. Houses, by contrast, are usually sold into mixed-stock streets where the new-build sits alongside older homes; they get re-priced as "a house on this street" rather than "another unit in this block".

The practical takeaway: the much-quoted "new-build depreciation" warning applies most strongly to off-plan flats. For a new-build detached or semi-detached house, the data shows roughly a 5 percentage point gap to the local HPI over 7 years — the equivalent of giving up about 8 months of average UK price growth at the rates seen in the latest ONS UK HPI release. Material, but not catastrophic.

Country: England and Wales tell roughly the same story

The split between countries is narrower than the property-type split.

Scotland and Northern Ireland operate separate Land Registry systems — their transactions are not included in HM Land Registry's Price Paid file used here, so this analysis is England and Wales only.

What this means in pounds

The percentages are abstract; a worked example helps. Take a £300,000 new-build sold in mid-2018 and resold in mid-2025 — a typical 7-year hold. The local HPI gain over that window for a representative English authority was around 30%. That implies an HPI-matched resale of roughly £390,000. The data above says the average new-build first-owner resale fell short by ~9 percentage points, which works out at roughly £27,000 — the resale lands closer to £363,000 instead.

If the same property had been a flat instead of a house, the gap widens to roughly 18 percentage points, or about £54,000 against an HPI-matched expectation. If it had been a detached house, the gap narrows to around £15,000.

These are averages. Half the flats in the sample underperformed by less; half by more. Variance within each bucket is wide — the standard deviation of excess returns inside any property-type cell runs at roughly 20 percentage points, so a single property's experience can easily deviate from the cohort mean. The numbers above describe the central tendency across a quarter of a million transactions, not what any specific buyer will see.

Caveats — what this analysis does not show

A few honest limits, because repeat-sales analysis isn't a complete answer.

1. Selection bias on resold properties. We can only measure what was sold. Pairs where the first owner is still in the property are missing from the dataset by construction. If owners who chose to sell within 7 years are systematically different from those who held on — for example, more likely to be in less-liked blocks, more likely to be forced sellers — the dataset will overstate the negative excess.

2. Improvements aren't subtracted. Some resale prices include the value of a new kitchen, an extension, solar panels. Repeat-sales work treats the property as constant; in practice it isn't, and any improvements made by the first owner push the resale price up without lifting our HPI control. That makes the negative excess return more conservative — the true new-build premium decay is likely modestly larger than the −9pp headline.

3. The HPI is not new-build-stripped. The ONS UK HPI uses all transactions including new-builds, so it has the new-build premium baked in on both sides of the comparison. A resale-only index would arguably be a cleaner control. We don't have one for the UK; the ONS doesn't publish one. The closest equivalent would be a hedonic resale-only re-build of the same data, which we may do in a future piece.

4. The 2025 new-build count looks artificially low because Land Registry registration of new-build transactions runs months behind completion. That doesn't affect this analysis — we only use the new-build flag on the first leg of each pair (2010–2019), where the registration lag is long since cleared. The 2025 sales appear here as resales only, where registration timing is normal.

5. Help-to-Buy and Shared Ownership distortions. A portion of the 2013–2023 new-build sales in our 2010–2019 first-leg sample were Help-to-Buy purchases; the resales would be at market price without the HTB equity loan. The Land Registry Price Paid file records the purchase price including any HTB loan, not the buyer's net contribution, so the comparison remains apples-to-apples in price terms — but the buyer-side experience of those transactions was different.

How to check this against your own postcode

Homecost lets you see every Land Registry transaction on any UK postcode, with the new-build flag visible on each. If you're considering an off-plan flat or a new-build estate, search the surrounding postcodes — particularly ones with developments completed 5 to 10 years ago — and look at what those units sold for on first resale. The Land Registry record is the ground truth; what an agent or developer brochure says is not.

For a worked example, try Manchester M1 1AE (a city-centre flat-heavy postcode), Birmingham B1 1AA (regeneration-corridor flats and townhouses), or Bradford BD1 1AA (mixed older stock with new-build in-fill).

Two sibling pieces sit alongside this one for context: the cost comparison of new-build versus resale homes — which looks at headline price premium at point of sale rather than depreciation on resale — and the long-form true cost of buying a £300,000 home, which puts depreciation into context against the much larger lines (deposit, stamp duty, mortgage interest) in a buyer's total cost picture. For category browsing, see all market-analysis articles.

Bottom line for buyers

The "new-builds lose 10-15% in the first 5 years" claim, as usually stated, is wrong. At 5 years held, the average new-build first-owner resale on Land Registry data underperforms its local HPI by about 6 percentage points — material, but considerably less than the property-press figure. The decay continues at roughly 1.5–2 percentage points per year and compounds with hold length, reaching −18 percentage points at 10 years and −26 at 15.

The headline number hides a much sharper split by property type. Flats account for almost all of the headline depreciation: −17.6 percentage points over a 7-year hold. New-build houses — detached, semi, or terraced — sit much closer to their local HPI, with gaps of 5 to 7 percentage points that look more like noise than catastrophe.

For a buyer comparing a new-build to a resale, the cleanest read of the data is: if you're buying a new-build flat, price in some HPI underperformance on resale; if you're buying a new-build house, the gap to the local market is small enough that other factors (location, EPC rating, transport, school catchment, stamp duty treatment, mortgage rate at completion) will matter more.

This article is general information drawn from public records, not financial or property advice. Past price patterns do not predict future prices; nothing here is a forecast. Speak to a qualified surveyor, mortgage adviser or conveyancer before acting on any property purchase decision.