Category A and category B sales in Price Paid Data
Every row in Price Paid Data carries a category marker, and mixing the two together is the single easiest way to draw a wrong conclusion from sold-price history. This guide covers what each category holds, how to tell them apart, and what to do with a category B row.
What the two categories are
| Category | What it covers | Reads as market value? |
|---|---|---|
| A — standard price paid | A single residential property sold for value | Yes — this is the comparable you want |
| B — additional price paid | Transfers under a power of sale (repossessions), buy-to-lets where they can be identified by a mortgage, transfers to non-private individuals, and sales where the property type is recorded as "Other" | Not reliably — treat individually |
The split was introduced in October 2013. Before that date the dataset published standard transactions only, so records from the earlier period are all effectively category A — which matters when you compare a 2005 sale against a 2020 one.
Why category B prices drift from market value
Category B is not a quality flag and does not imply anything is wrong with a transaction. It groups sales that happened under conditions a normal buyer would not face:
- Repossessions — a lender selling under a power of sale is discharging a debt, often on a timetable, which reliably drags the price below open-market value.
- Transfers to companies and other non-private buyers — the price may reflect an internal restructure rather than a negotiation between two parties at arm's length.
- Buy-to-lets identifiable by a mortgage — an investment purchase is a different calculation from an owner-occupier's, and the two need not price alike.
- Property type recorded as "Other" — the residual type, used where the property does not fall into the standard detached, semi-detached, terraced or flat categories. Comparing one against an ordinary home is rarely like-for-like.
How a mixed average misleads — worked through
Suppose a street records four sales in a year: three ordinary sales at £310,000, £325,000 and £330,000, and one repossession at £215,000. The raw mean across all four is about £295,000 — below every one of the three genuine market sales. A buyer using that figure as "the street average" would under-read the street by roughly £25,000, and a seller using it would price into a loss.
Separating them gives a category A mean near £322,000, plus a single category B data point worth knowing about separately. Same rows, same dataset, two very different conclusions — and this is the specific failure mode behind the warning on our sold house prices page that a raw street average can mislead.
How to tell which you're looking at
- In the raw dataset, the category is an explicit column on every record in the monthly Price Paid Data downloads.
- In a tool, check whether categories are distinguished at all. Many consumer sites publish the merged list with no marker, which is why two sites can quote different "averages" for the same street from identical source data.
- By eye, a single row far below its neighbours in the same period is worth checking before you treat it as a comparable.
Frequently asked questions
What is a category B sale in Land Registry data?
An additional price paid entry: repossessions and other transfers under a power of sale, buy-to-lets where they can be identified by a mortgage, transfers to non-private individuals, and sales where the property type is recorded as "Other". Category A is the standard entry — a single residential property sold for value.
Should I ignore category B sales when comparing prices?
Do not average them in with category A, but do not discard them either. They are genuine transactions that carry information about a street; they are simply not like-for-like comparables for an ordinary open-market sale.
Why do two websites show different average prices for my street?
Most often because one merges category A and B and the other does not, or because they cover different date ranges. Both can be reading the same underlying HM Land Registry data.