Using price paid data to sense-check a viability input

Price paid data will not value your scheme. It will tell you, in about an hour, whether the sales values someone has put into an appraisal sit inside the range of what has actually sold nearby, and that is usually the question you need answered before a viability negotiation starts. HM Land Registry publishes every reported residential transaction in England and Wales as an open data set under the Open Government Licence, refreshed monthly, and it is the same underlying source the UK House Price Index is built from.

What it cannot do is replace the comparable evidence a valuer assembles. A Viability Assessment submitted to a council is a professional document, and councils normally expect a suitably qualified author working to the RICS professional statement on financial viability in planning. Professional review of any draft is strongly advised. Use price paid data to interrogate the inputs, and to know when to push back.

What the dataset contains and what it omits

Each row carries the price, the date of transfer, the postcode, the property type, whether the property was newly built, the tenure, the address components and a category flag. The property type codes are the ones you will use most: D for detached, S for semi-detached, T for terraced, F for flats and maisonettes, and O for other.

Two fields decide whether a row is usable.

The new build flag separates newly built sales from established stock. The category flag distinguishes standard arm's length transactions from additional price paid entries, which include repossessions, bulk transfers, transfers to corporate bodies and other sales that do not represent open market value. Filter the second category out before you calculate anything.

The omissions matter more than the contents.

• No floor area. There is no square metre figure anywhere in the data set, so a rate per square metre cannot be derived from it alone.
• No unit specification. Number of bedrooms, parking, outdoor space and condition are all absent.
• No off-market or undisclosed transactions, no transfers not for value, and no sales that were never lodged in the form the data set draws on.
• Nothing about incentives. A new build sold with a deposit contribution, paid stamp duty or a part exchange appears at its headline price.

That last point is the one that quietly inflates a new build comparable set.

Matching transactions to house types

Work from your own schedule of accommodation outwards. For each unit type in the scheme, define the comparable set before you look at any prices, and write the definition down.

A workable definition names the property type code, the tenure, the new build flag, a postcode radius or a list of postcode sectors, and a date window. Then join the addresses to floor areas from the energy certificate register, because the register carries total floor area for most dwellings marketed or let since 2008. That join is what turns a price into a rate you can compare with your own gross internal areas.

Expect the join to fail on perhaps a fifth to a third of addresses in any set, mostly through address formatting rather than absence. Clean the worst cases by hand and accept the rest as a sample.

New build against second hand values

The new build premium is real, and it varies by location and by product. Testing it is one of the more useful things this data supports, because the appraisal input is usually a single figure and the evidence is a distribution.

Pull two sets for the same area over the same window, one with the new build flag set and one without, restricted to the same property type. Compare the medians rather than the means, because a handful of large detached sales will drag a mean upwards. Then look at the spread. A wide spread in the established set with a tight cluster of new build prices usually means a single development dominated the new build rows, which makes it one comparable rather than twenty.

Where a scheme's appraisal assumes a new build premium above what the local evidence supports, that is a point to raise early rather than at the section 106 stage. Our note on reading a council's housing land supply monitoring covers the other half of that conversation, which is what the authority already believes about local delivery.

Adjusting for date and market movement

Transactions are historic by definition, and the date of transfer is the completion date rather than the date the price was agreed. On a new build bought off plan the gap can be a year or more, so a completion in March may reflect a price agreed the previous spring.

Index older transactions forward to your valuation date using the UK House Price Index at local authority level rather than a national figure. Record the index series, the base month and the adjustment applied for every comparable, because an adjustment you cannot reproduce is an adjustment the district valuer will discount.

Keep the table. It is the appendix that makes the rest of the argument checkable.

Where the data misleads on flats

Flats are where this exercise goes wrong most often, for four reasons.

Leasehold prices reflect the unexpired term, the ground rent and the service charge, none of which appear in the data. A short lease transaction sitting in your comparable set will pull the average down by an amount you cannot see.

Bulk transfers of blocks to institutional investors appear as single rows at a price that is not a unit value, and the category flag is how you exclude them. Shared ownership transactions record the price paid for the initial share rather than the full value. And in buildings affected by remediation or cladding issues, sales in the recent past may reflect a discount specific to that building.

The practical response is to restrict a flat comparable set to standard category, freehold or long leasehold where the term can be checked, and to inspect every row above or below the interquartile range by hand.

Using it to test a consultant's assumption

The point of the exercise is a conversation, so frame the output as questions rather than as a valuation.

1. What is the source of the sales value for each unit type, and is it a valuer's opinion or a benchmark.
2. How many transactions sit behind each figure, and over what window.
3. Has the figure been indexed, and from what date.
4. Does the new build premium assumed match the local evidence.
5. If the appraisal shows a deficit, how sensitive is that deficit to a five per cent movement in sales values.

That fifth question is usually the decisive one. On a scheme of 30 to 80 units a modest change in the sales value assumption can move the affordable housing offer materially, which is why the input deserves more scrutiny than the build cost. What that means for the offer itself belongs in the Affordable Housing Statement rather than buried in the appraisal.

Citing it in a viability submission

Cite the data set by name, the download used and the month of publication, and state the Open Government Licence attribution. Then be explicit about what the data supports and what it does not.

Say that price paid data has been used to test the reasonableness of the sales value assumptions, that floor areas were taken from the energy certificate register, and that formal valuation evidence is provided separately. A district valuer reading that sentence knows exactly what weight to give the appendix. A submission that presents indexed price paid rates as a valuation, without a valuer's opinion behind it, tends to be discounted in full.

It also helps to check how the authority has treated similar evidence before, and searching planning decision data across authorities is the quickest route to the relevant committee reports.

If you are assembling the wider evidence base for a major application, browse all 66 PlanScope report types to see what a scheme of your scale usually triggers. For specialist technical reports, qualified professional input may be required. PlanScope provides AI-assisted draft reports for planning applications. Reports should be reviewed before submission. We do not guarantee planning approval.