Key takeaways
- The average UK new build sold for £332,197 in December 2024 against £260,674 for an existing home. That gap is £71,523, or 27.4%.
- Not all of the gap is a premium. New builds are around 9% of transactions and a different mix of property, and the index hides the split for the latest two months.
- NHBC Buildmark runs 10 years, but the builder's duty to put right anything below NHBC requirements covers the first two years. After that a defined structural list applies.
- 35% of buyers surveyed 9 months after completion in 2021 to 2022 reported 16 or more separate problems, and 65% of that group were still waiting for repairs.
- 80% of new homes sold by the 11 largest builders in 2021 to 2022 carried an estate management charge, which averaged £358 a year in 2022.
The new-build premium was £71,523 in December 2024, and some of it is not the house
You're standing in a show home and the question you want answered is whether the brochure price is a fair one. Here's the arithmetic, from HM Land Registry's UK House Price Index.
In December 2024, the last December the index has fully revised, the average new build in the UK sold for £332,197. The average existing home sold for £260,674. The gap is £71,523, or 27.4%. That is the figure people mean by the new-build premium, and it is not new. The same series puts the new build average at £174,292 in January 2005 against £135,955 for an existing home, a gap of £38,337.
Two things follow from that, and they pull against each other. Part of the gap buys something real: a 10-year warranty, and nothing to fix on move-in weekend. Part of it is not a premium at all, because the two averages describe different houses. Working out which part is which is the whole job, and the rest of this piece does it in three places where the money actually moves after completion.
Why the most recent new-build figure is the least trustworthy one
Start with what the index can and can't do. The UK HPI uses "hedonic regression and mix-adjustment" to "account for the change in composition", feeding in attributes such as floor area, the number of habitable rooms and an Acorn demographic classification of the postcode. It does not follow that the published average price for new builds is comparable with the published average for existing homes. Those are averages of different populations of sales.
New builds are a small and awkward slice. HM Land Registry puts their weight at "around 9%" of the headline series, and their registrations arrive late. The effect is large enough that the property-status breakdown is withheld for the latest two months, "as first estimates for new build are based on a small sample which can cause volatility in initial estimates."
The size of that volatility is worth seeing. On 16 July 2025, UK new build annual inflation was published at 29.9% for March 2025. Under the improved imputation method the ONS introduced afterwards, the same month would have printed 12.1%. For scale, the historical record high for that series is 13.1% in the 12 months to May 2016. A number more than double any figure the series had ever produced went out as an official statistic, because too few new build sales had been registered in time. At the end of 2016, new builds were roughly 9% of transactions and drove "over 80% of the revision."
So the premium quoted from a recent month is usually too big, and anyone quoting this year's number is quoting an estimate that has historically been revised down. December 2024 is used here for that reason, not because it is the most flattering.
The warranty is 10 years long; the part that covers your kitchen is two
The second place the arithmetic moves is the warranty, and it is routinely read as longer than it is. An NHBC warranty is not one product. Under the Buildmark policy for homes registered from 1 April 2024, cover comes in sections, and those sections do different jobs.
For the first two years, the builder "must put right anything covered by Buildmark that does not meet the NHBC requirements." That is the broad cover, and it is the builder's obligation rather than an insurance claim. It is also where the cosmetic exclusion bites: the policy "does not cover cosmetic damage which is reported to the builder more than seven days after the completion date." A scuffed worktop noticed in week three is not a warranty matter.
After that, section 3 runs for "the eight years after the builder warranty period, up to 10 years from the completion date", and it covers damage to a listed set of parts. Foundations, external cladding and render, roofs and ceilings, load-bearing walls and structural floors, flues and chimneys, staircases and screeds, double or triple glazing panes, and underground drainage you're responsible for. Sound transmission is excluded. So is cracking or mortar erosion that doesn't weaken structural stability or let weather in. So is damage that only affects floor coverings.
There's also a threshold. Section 3 claims have to clear a minimum claim value set out in the policy schedule, and that value "increases by £50 on 1 April every year from the end of the builder warranty period." Two areas of damage from one defect attract one minimum claim value; two areas from two defects attract two. The headline financial limit is generous, at the original purchase price up to a maximum of £1,000,000, rising by 5% of the original limit on each anniversary. The binding constraint for a household is almost never that ceiling. It's the list, the threshold and the calendar.
Most snags get fixed, and the cost sits in the tail
The third place is new build snagging, and the honest summary is that the average experience is fine and the average is the wrong statistic.
The Competition and Markets Authority analysed the National New Homes Customer Satisfaction Survey for four survey years to 2021 to 2022. It found "a statistically significant increase over time in the proportion of homeowners reporting higher numbers of snags, with 35% of respondents to the 9-month CSS in 2021 to 2022 reporting 16 or more different problems."
What matters next is what happened to those problems. The chart plots the share still unrepaired at the 9-month survey, by how many were reported, across a base of 38,870 respondents. For buyers reporting 1 to 5 problems, 18% were still to be rectified. At 6 to 10 problems, 31%. At 11 to 15, 44%. At 16 or more, 65%.
That is not a straight line, and the shape is the finding. Reporting more faults does not get you proportionally more repairs. It correlates with getting a smaller share of them done, which is the opposite of what a queue should do. The CMA's composite measure of overall dissatisfaction with problems and their handling caught 12% of customers in 2021 to 2022, and it "had hardly changed over the previous 4 years."
Put a number on the exposure. New build completions ran to 122,012 in 2025 on NHBC registrations, which cover "70%+ of all new homes built in the UK". If a third of those buyers land in the 16-or-more group, that's tens of thousands of households a year carrying an open repair list into their second year of ownership. Those are the years when the builder's broad obligation is still running and about to stop.
80% of new homes from the biggest builders came with an estate management charge
The fourth place is the one buyers rarely price at all, because it doesn't appear in the mortgage calculation. The CMA found "a growing trend by developers to build estates with privately managed public amenities". On its figures, 80% of new homes sold by the 11 biggest builders in 2021 to 2022 were subject to estate management charges. Those 11 builders supplied around 40% of homes in Great Britain in the same year.
A new build service charge on a leasehold flat and an estate management charge on a freehold house are different instruments. The government's response to the CMA describes the second kind, as charges on "residential freeholders living under private management arrangements". An estate management charge pays a private company to maintain roads, drainage and open space that a council would otherwise adopt. The CMA measured what it costs: the amount "varies greatly, from just under £60 to just under £1,000 in 2022, with an average charge of £358." Households paid "at least £260 million" in these charges in 2022. The CMA projected "more than £4.4 billion" over a 25-year period, and that was without allowing for price rises or wider adoption of the model.
£358 a year against a £332,197 purchase is 0.11% of the price, which sounds trivial and is the wrong way to read it. It's a standing claim on the household budget, and the CMA found homeowners "unable to switch estate management providers". It also isn't the whole bill. The CMA reported that "unclear administration or management charges" can often make up 50% or more of the total. It also found one estate management company charging an additional "c.40%" in event, consent and permission fees on top of its annual charge, in its last financial year. Repair liabilities can be lumpy in a way a service charge on a flat rarely is: the report cites one case of "a potential share of £100,000 for upgrades to a pumping station."
This sits alongside the running costs any owner carries, which is why the true cost of owning a home is a separate calculation from the purchase price. On a new estate there is simply one more line in it, and it's one the buyer can't shop for.
The law written to fix this passed in 2024 and has not commenced
Part 5 of the Leasehold and Freehold Reform Act 2024 was drafted for exactly this problem. Section 74 would make an estate manager's costs recoverable "only to the extent that they are reasonably incurred", and only where works are "of a reasonable standard". Section 77 would let a homeowner apply to a tribunal for a determination of whether a charge is payable at all.
None of it is in force. legislation.gov.uk carries Part 5 as a prospective version, and records every section from 72 to 99 as "not in force at Royal Assent". Section 124 leaves commencement to regulations the Secretary of State may make, and as at 24 September 2026 the page shows no commencement note bringing Part 5 into effect. The UK government's October 2024 response to the CMA accepted the enhanced-protections recommendation in principle and marked mandatory adoption of amenities as "Further work required".
For a buyer, the practical position is that the charge is contractual and the statutory remedy is drafted but dormant. That's a different risk from one the law has priced.
The strongest case against all of this is the satisfaction survey
There's a serious objection, and it comes from the industry's own independent data rather than from a press release. On 24 March 2026 the Home Builders Federation published the latest National New Homes Customer Satisfaction Survey results. 93% of buyers would recommend their builder. 90% were satisfied with the quality of their home, and 88% were happy with the service after they moved in. The survey is run by warranty providers and the data is verified by Ipsos. That is not a marketing claim, and it deserves to be taken at face value.
Here is how to referee it against the CMA's numbers, because both are true. The HBF headline comes from the survey "sent to home buyers eight weeks after they receive their keys". The CMA's snagging figures come from the 9-month survey. The gap between them is not a contradiction; it's the point. Satisfaction measured at 8 weeks captures the purchase. Satisfaction measured at 9 months captures the repair.
The CMA saw the same split in the same dataset: the 8-week survey showed 89% would recommend their builder in 2021 to 2022, while the 9-month survey produced the 35% and 65% figures above. The CMA also noted the risk of "perverse incentives for housebuilders to focus on quality and customer service only until the point at which consumers submit their survey responses."
One more thing weakens the reassurance a little. Only 30% of the 2026 respondents had previously bought a new home, so most were rating an experience they had nothing to compare with. A buyer who is happy at 8 weeks and has never owned a new build before is not evidence that the 9-month tail is small.
What this evidence cannot tell you
The biggest limitation is the one the headline number invites. £71,523 is a difference between two averages, not a like-for-like premium. Nobody in these sources has published a matched estimate of what an identical house costs new versus second-hand, and this piece does not have one. Treat the gap as an upper bound on the premium and not as the premium.
The resale question has the same shape. Over the 20 years from January 2005 to December 2024, the new build index rose from 54.5 to 103.9. That's a gain of 90.6%. The existing-property index rose from 52.7 to 101.0, a gain of 91.7%. Those two numbers are close enough to say that new builds as a category have not systematically underperformed. They cannot tell you what happened to a specific house. The new build index measures homes that are new at each date, not a cohort of 2005 new builds followed through to resale. The thing a buyer wants to know, whether their own new build sold at the existing-home price five years later, is not in this data.
The snagging numbers come from people who chose to answer a survey, in the 2021 to 2022 survey year. The CMA read four survey years and found the share reporting large numbers of snags rising, but the levels quoted here are one year's. The estate charge figures come from information requests to 14 of 15 estate management companies and are dated 2022. The £358 average sits inside a range running from just under £60 to just under £1,000, so the average tells you very little about any individual estate. And every number here is British. The house as an investment comparison that holds in one national market doesn't travel.
What would change the conclusion
Three things would, and they're worth watching in this order.
If Part 5 of the 2024 Act commences, the estate charge stops being an unreviewable contractual obligation and becomes a reviewable one, with a reasonableness test and a tribunal behind it. The £358 wouldn't vanish. The tail, the 40% in consent fees and the £100,000 pumping station share, is what a reasonableness test is for.
If local authorities go back to adopting roads and drainage, the charge disappears entirely for new estates. That is the CMA's own first recommendation, and the one the government marked as needing further work. That would remove a permanent cost from a buyer's budget rather than capping it.
If the 9-month survey figures improve, the case in this piece gets weaker on its strongest point. HBF puts the 8-week recommend score over 90% for a sixth consecutive year, and it has told us almost nothing about the tail. The figure that matters is the share of buyers reporting 16 or more problems, and whether the 65% unrectified rate comes down.
The premium itself is the least interesting part of this. Price the warranty cliff at 2 years, the estate charge at £358 a year with no ceiling and no exit, and the chance of a repair list still open at month nine. A buyer who has done that has priced the thing properly, and may still decide the new build is worth it. A buyer who has priced only the cost of moving house and the mortgage has priced the purchase, and none of the three costs that follow it.