How is the mansion tax valued? | LonRes

What it costs, and where the four cliff edges fall

The shape of this is already familiar to most agents: a flat annual surcharge bolted onto council tax in England, announced in the Autumn Budget 2025 and consulted on between 19 May and 14 July 2026, in four bands. The table is here for reference.

Liability falls on the legal owner rather than the occupier, which captures freeholders, leaseholders whose lease was originally granted for more than 21 years, companies and trustees. First bills land in March 2028 for the April 2028 start, and the Treasury expects to take around £430m a year from 2028/29.

The design matters more than the headline. These are slab charges, not marginal ones, so there are four cliff edges rather than one. A property valued at £2,499,000 pays £2,500. A property valued at £2,500,000 pays £3,500. One thousand pounds of value costs a thousand pounds a year, every year, indexed.

How will the mansion tax be valued?

Mostly by automated valuation model. The Valuation Office Agency (VOA), which sits within HMRC, will run what the consultation calls a model-assisted valuation approach, combining AVMs with professional valuer judgement, drawing on property databases, stamp duty records, local authority planning records and geospatial data. Press reports in February 2026 put VOA recruitment at up to 1,000 officials, though the agency did not confirm a figure at the time and HMRC has since said most of those hires will not work on the surcharge.

Two dates decide everything. Properties will be assessed at their estimated market value in 2026 – this year – and the list will then be revalued on a five-yearly cycle, with the next revaluation in 2033. Note the gap. A band set on 2026 evidence governs the five charging years from April 2028 to March 2033, so 2026 evidence is still setting bills seven years after the valuation. Improvements made after the valuation date do not feed into the band until the revaluation.

AVMs are not the weak link they are sometimes assumed to be. Lenders rely on them for mortgage valuations, and for a national exercise covering millions of homes they are the only practical method. Their limits are narrower and more specific than that: thin comparable evidence, properties that are unique or heavily altered, and markets moving faster than the data behind them.

Prime London above £2m is where all three overlap. Average achieved prices across prime London fell 7.9% in the year to July 2026, and in Q2 2026 prime London recorded its largest annual fall in achieved prices since 2009, down 7.5%, with prime central London down 9.0%.

The consultation’s own design reflects that, pairing the model with professional valuer judgement and reserving inspections for cases where comparable evidence is thin. So does the OBR, which has budgeted for one in five owners to challenge their band and 40% of them to win. Neither the consultation nor the OBR assumes the model will be right every time at this end of the market.

Will a valuation officer inspect your client’s home?

Most will not. A Sunday Telegraph report at the end of August said valuation officers would carry out in-person interior inspections of properties near the £2m threshold, assessing room counts, storeys, bedrooms and bathrooms. HMRC narrowed it on 24 August, saying most valuations will be made without a home visit, using publicly available data, third-party data and satellite imagery, with internal inspections only where attributes can only be confirmed internally or a re-measurement is required.

Powers of entry are neither new nor unusual here. Section 25A of the Local Government Finance Act 1992 has allowed a valuation officer to enter a dwelling since 2015, subject to First-tier Tribunal approval and at least three clear days’ written notice.

Which is the whole point. An inspection establishes what the house is. It does not establish what houses like it sell for. That second question is the one that moves a band, and it is answered by achieved prices.

Why the appeals are an evidence problem, not a tax problem

Because the tax is fixed and the valuation is not. Owners will be able to challenge both their valuation and their liability, with an initial eight-month challenge window at launch and six months thereafter, and appeals heard by the Valuation Tribunal for England. Taxpayers keep paying while a dispute runs and are refunded if it is reversed.

The OBR’s own assumption tells you what the government expects. Applied to 156,000 properties, it implies in the order of 31,000 challenges and 12,000 reductions. Each one is decided on comparables.

The proposal that has had least attention is this one. The consultation proposes that estimates from estate agents or property listing websites be inadmissible as evidence in a challenge. Propertymark has criticised that exclusion, arguing that agents hold exactly the local knowledge the exercise depends on.

The distinction being drawn is between an opinion and a record. An agent’s estimate is an opinion. An achieved sale price on a comparable property is a record. The first would be ruled out; the second is the evidence the assessment is built on.

A formal challenge will usually be argued by a chartered valuer or surveyor. The comparable evidence it rests on is the agent’s to supply, and the agent who supplies it is in the room.

How the market has already moved around £2m

Not gently, and not evenly. Several research desks have published on the surcharge; the pricing-behaviour figures below are Hamptons’ own. David Fell found that 83% of offers on homes priced within 10% of £2m kept the price beneath the threshold, against 64% a year earlier, while behaviour around the £1m mark – outside the tax – remained unchanged. That second finding is the useful control: it points to the tax rather than the wider market. The three figures below are counts of homes coming to market, not price movements.

Both sides have an interest in a price that sits below the line. The buyer carries the charge once they own it, and the seller knows a sub-threshold asking price reaches a wider pool. The negotiation is not over whether to stay under. It is over who absorbs the difference, and that is a live conversation on every instruction within reach of a threshold. It is being had against a market where more than half of prime London sales in July 2026 completed after at least one price reduction, the average discount to asking price was 10.4%, and homes taking more than 12 months to sell gave up 19.3% against 3.9% for those selling inside three months.

Will the mansion tax go up in the Budget on 28 October?

Nobody outside the Treasury knows. But the direction of travel is not subtle. Andy Burnham took office as Prime Minister on 20 July 2026 with John Healey as Chancellor, and the Mail on Sunday reported in July that the threshold could drop from £2m to £1.5m. Nothing has been announced, and the consultation response has not been published.

Tax Policy Associates and Hamptons have both modelled the effect, on different baselines, so read them as direction rather than arithmetic. Tax Policy Associates has roughly 127,000 properties in scope becoming roughly 243,000, close to a doubling. Hamptons, analysing its own proprietary data, estimates the increase at 102% across England, 79% in London and between 150% and 180% across the Midlands and the North.

Knight Frank made the regional point from the other direction in analysis published before the Autumn Budget 2025, calculating the threshold that would capture an equivalent share of stock in each region: £829,000 across London as a whole, £408,000 in the West Midlands and £288,000 in the North East. Its analysis covered Great Britain; the surcharge applies in England only.

Read those percentage increases carefully, though. They look dramatic outside London because the base is so small, not because the tax is about to reach the suburbs.

Which is the point worth holding onto. Lowering the threshold widens the tax politically and adds a great many more homes in absolute terms. It does not stop this being a London and South East tax.

Tom Bill, head of UK residential research at Knight Frank, has said that if the Budget brings a broader package of asset and wealth taxes, the current bands “may prove to be merely introductory rates”. He warns that a third successive Budget targeting high-value property, after 2024 and 2025, could set back any recovery.

There is precedent for the drift, and the profession has put it on the record. Hannah Catt, senior associate in private property at Charles Russell Speechlys, notes that the annual tax on enveloped dwellings began as a charge on property above £2m in 2013 and now starts at £500,000, and warns of mission creep: “If the surcharge can be implemented with relative ease, there is a real risk of mission creep.” She also points out that council tax bands themselves have not been updated since 1991.

Stamp duty is the other lever. Burnham ruled out scrapping it on 27 July, though the trade is not treating that as settled.

The concentration is the political problem. Kensington and Chelsea council calculates that four London boroughs – itself, Westminster, Richmond upon Thames and Wandsworth – would contribute around £270m, more than half the national bill. The parliamentary constituency of Cities of London and Westminster alone accounts for an estimated 10.5% of the surcharge, on PolicyEngine’s modelling of a £400m yield. A tax this geographically concentrated is politically exposed in both directions.

Who pays the mansion tax on a rented property?

The owner, which in practice means the landlord. Propertymark has asked for landlords to be able to choose whether they or their tenants carry the charge, mirroring the flexibility that already exists in council tax, but as drafted the liability sits with the legal owner regardless of who lives there.

What agents should do before March 2028

Four things, none of which require the legislation to be finalised.

The Budget will decide how many people this tax reaches. Comparables will decide what each of them pays.

Where LonRes fits

Every argument above resolves to the same question: what did comparable property actually sell for. Asking prices do not answer it. The consultation would rule portal estimates out as evidence. A model will produce a number, and whether you accept that number or challenge it, you need the same thing to judge it.

LonRes holds achieved sale prices across prime London, alongside the Sales Archive that lets you build a comparable set by street, period, size and date rather than by postcode alone. Live Sales shows what is trading now, including off-market stock that does not appear on the portals. This is the gap LonRes was built to close.

The agents who spend the next eighteen months quietly building evidence files for their £2m-plus clients will own those relationships when the band notices arrive. The ones who wait for the legislation will be starting where everyone else starts.

Frequently asked questions