How to Read Land Registry Price Paid Data at the Viewing Stage

Asking prices are opinions. Sold prices are facts. HM Land Registry’s Price Paid Data is the public record of what changed hands, for how much, and when — and it is the closest thing an English buyer has to an independent check on the number an agent has put on a property.

It is also widely misread. The dataset has specific inclusions, specific exclusions, a registration lag, and a handful of fields whose meaning is not obvious. Used carelessly it produces confident, wrong conclusions. Used properly it is the single most useful free dataset in a purchase.

What the dataset is

Price Paid Data covers all residential property sales in England and Wales that were sold for value and lodged with HM Land Registry, going back to January 1995. Each row is one transaction: price, date of transfer, postcode, property type, tenure, address components, and the local authority geography¹.

It is not a sample and it is not an index. It is the register.

What it leaves out

The exclusions matter more than the inclusions, because each one is a way for a street’s apparent history to be incomplete¹:

• Sales not lodged with HM Land Registry. If it was never registered, it is not there.
• Transfers that were not at arm’s length — gifts, transfers following a court order, compulsory purchases.
• Right to Buy sales at a discount.
• Leases of seven years or less.
• Transfers involving the division of a share in a property, or a mortgage rather than a sale.
• Vesting deeds covering multiple properties, where no single price is attributable.

The practical effect: a house that last changed hands within a family, or was bought under Right to Buy, will simply be absent. Absence of a record is not evidence that nothing happened.

Category A and category B

Every row carries a PPD category, and mixing them is the most common analytical error¹:

• Category A — the standard entry: a single residential property sold for value.
• Category B — additional entries including transfers under a power of sale (repossessions), buy-to-lets where identifiable by a mortgage, transfers to non-private individuals, and sales where the property type is classed as “other”.

Category B prices are not comparable with category A prices. They include forced sales and portfolio transfers. Any median or average you compute should say which categories it includes — and for “what would a normal buyer pay”, the answer is category A only.

The fields that get misread

old/new — Y means a newly built property, N an established building¹. New-build sales typically transact above second-hand stock of the same type and size, and on estates a run of new-build registrations can move a postcode’s apparent average sharply. Filter them out before comparing to a resale.

duration — F for freehold, L for leasehold¹. A leasehold flat and a freehold house on the same street are not comparables, and neither are two leasehold flats with very different lease lengths. The register records tenure; it does not record the lease term, service charge or ground rent, which are the things that actually differentiate leasehold values.

date of transfer — the completion date, not the registration date and not the date the price was agreed. The price was agreed some weeks or months before it, so even a fully registered sale is evidence about an earlier market than its date suggests.

property_type — D detached, S semi-detached, T terraced, F flat or maisonette, O other. “Other” is a genuine catch-all and is worth excluding from most comparisons.

The lag, and why the last two months look wrong

Price Paid Data is updated monthly, but the time between a sale completing and being registered typically ranges from two weeks to two months — so the most recent two months of data are incomplete and will keep filling in after publication¹.

This produces a specific, avoidable mistake: looking at the last quarter, seeing low transaction counts, and concluding the market has stalled. It has not. The registrations have not caught up. When you are comparing periods, compare like-for-like windows that are both fully settled, or state the snapshot date so the reader knows what was in the file when you ran it. We do that on every data story — see Surrey towns by price per square foot for an example of how the window gets declared.

Using it on an actual offer

1. Define the comparable set narrowly. Same property type, same tenure, same broad size, same side of the main road. Five genuine comparables beat fifty loose ones.

2. Check transaction volume before you trust an average. If three houses sold in the postcode last year, the “average” is one unusual sale away from meaningless. This is the thin-market problem, and it is why a headline median can move without any individual home changing value.

3. Separate composition from direction. A median that falls because the mix of what sold changed is telling you about the mix, not about value. A median that falls while like-for-like properties sell for less is telling you about value. They look identical in a spreadsheet.

4. Adjust for time explicitly. A sale from eighteen months ago is evidence about eighteen months ago. Say what you think has happened since, and why, rather than treating the number as current.

5. Strip the new-builds out of resale comparisons. And if you are buying the new-build, accept that your resale comparables in five years will be the second-hand ones.

6. Read the extremes, not just the middle. The top and bottom sales on a street usually have a reason — a plot, a condition problem, a probate sale. Finding the reason tells you where your target property sits.

7. Bring it to the negotiation as evidence, not as an argument. “Number 14, same type, same tenure, sold in March for X” is a conversation. “The data says you are overpriced” is not.

What our reports do with it

Home-Checker’s property market section reads the same HM Land Registry Price Paid Data: the average price for the postcode, prices broken down by property type, a year-by-year price series rather than a single trailing figure, and the recent sales behind those numbers — falling back to outcode level where a postcode has too few transactions to be reliable, and recording the gap where neither has data. For a Property Report, sales are matched to the specific address using the Land Registry address components.

For the county-level view of how these numbers move, see the Surrey property market overview, the most expensive postcodes in the county and the Surrey area data.

House price data is sourced from HM Land Registry Price Paid Data under the Open Government Licence v3.0 and is subject to a registration lag. This guide is general market commentary and does not constitute financial or investment advice.

Run a Home-Checker report for any English postcode to see the local market picture alongside schools, crime, flood risk and broadband.

Footnotes

1. HM Land Registry, About the Price Paid Data (gov.uk/guidance/about-the-price-paid-data) — coverage, exclusions, PPD category A and B definitions, old/new and duration field definitions, monthly update cycle and the two-week to two-month registration lag. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶