Calculated, not counted

How brownfield cover is calculated

Brownfield cover is our own calculation from three government datasets. No council publishes it. The method is below in full, including what it leaves out and a version we got wrong first, and the working is free to download.

The calculation

1. Housing need. Each council's figure under the government's standard method: 0.8% of its existing homes a year, raised where homes are expensive relative to local earnings. We recompute it from the Office for National Statistics' affordability ratios and the government's dwelling stock figures, and our total for England, 371,117 homes a year, matches the published figure exactly.
2. Over 5 years. The period for which a council must show a supply of deliverable land.
3. Brownfield capacity. The homes the council's own brownfield register says its live sites can take, using the lower estimate each site gives. Sites marked as built, and sites that have left the register, are not counted.
4. Brownfield cover is capacity divided by 5 years' need. Under 100% means the council must find land beyond its brownfield register. The difference, in homes, is shown on each council's page as the homes still to find.

Where we do not publish a figure

Of 295 councils with a housing need figure, 146 get a cover figure. The rest are shown with their need and delivery, and a note saying why:

• Incomplete registers. Where fewer than 80% of a register's sites give a housing estimate, capacity cannot be totalled. Adding up only the sites that do would make those councils look short of brownfield when they are short of data.
• Registers missing from the national dataset. Absent is not zero.

How we read the registers

• A site recorded as 0 minimum and 0 maximum homes is treated as giving no estimate. Some councils use 0 to mean "not stated" for every site.
• A site with a minimum of 0 but a positive maximum is counted at the maximum. Some councils record only an upper figure; read literally, a permitted site for several hundred homes would count as none.
• Permission statuses are written by each council in free text. We group more than twenty variants into four: with permission, without, pending, and built. A site allocated in a local plan but without permission counts as without.

A version we got wrong

Our first version divided the homes still to find by the number of homes each council already has. It put Westminster, Harrow, Redbridge, Ealing and Bromley at the top. None of them is short of brownfield land: their registers list sites with no housing estimates, so their capacity added up to nothing. It was also mostly a ranking of house prices, because dividing by existing homes cancels most of the need figure and leaves its affordability adjustment. Brownfield cover does neither: it tracks how much brownfield a council has, not how expensive it is.

What it does not account for

• Permissions granted on land not on the register. Greenfield sites with planning permission also count towards supply, and no complete national dataset of them is published. Cover understates total supply for that reason.
• Phasing. Registers list total capacity without saying when it will be built, so a register can cover more than five years of need.
• Register quality. Each council maintains its own, and the publisher says the national dataset "may be incomplete".
• Any particular piece of land. A council's position says how hard it will have to look beyond brownfield. It does not say whether a given site would get permission.

England as a whole

Across the 146 councils with complete registers: need of 183,906 homes a year, brownfield registers covering 79% of 5 years of it, and land for 196,063 homes still to find. The council figures add up to the England figure exactly.