Which councils refuse the most extensions?

Draw the same single-storey rear extension on the same terraced house and post it to two different planning departments, and your odds of a refusal change by a factor of thirty. In Belfast, fewer than one extension application in a hundred is refused. In Brent, it is closer to one in four. Same national policy, same permitted-development rulebook, the same drawings — a thirtyfold difference in the answer.

That gap is the finding of this piece. We analysed published planning records for 46 UK authorities and ranked those with a meaningful number of decided extension applications by the share refused. The councils at the top and the councils at the bottom are not a little different. They are running what feel like two different systems.

The scope, stated up front

This is not a national league table, and we won't pretend it is. The study covers 46 authorities — a mix skewed toward the front of the alphabet (Aberdeen to Brent), plus a cluster of Kent districts and a sample across Scotland, Wales and Northern Ireland. It is not every council in the UK, and the missing ones could sit anywhere on the range.

Within those 46, we looked at every application whose description mentions an extension — 71,376 of them — and kept the 60,023 that reached an actual decision. Across all 46, the aggregate extension refusal rate is 9.0%. That is the number to hold in your head while you read the table, because the story is not the average. It is the distance from it.

The ranking

Refusal rate is refusals divided by decisions — approvals (including approvals with conditions) plus refusals. Withdrawn and still-undecided applications are excluded. We only rank councils with at least 200 decided extension applications, so a handful of small or lightly-covered authorities drop out (they are listed in the methodology).

It's the council, not the extension

The most useful thing in this data is a myth it quietly kills. People treat "extensions get refused" as a property of extensions — as if a rear extension were an inherently risky ask. It isn't. Across our 46 authorities, extensions are refused at 9.0%, and all application types together are refused at 9.4%. An extension is, on average, very slightly safer than the typical planning application. Nationally, it is a boringly ordinary request.

What isn't ordinary is the postcode you file it in. The refusal rate ranges from 0.7% to 23.5% across councils running the identical national framework. The single biggest predictor of whether your extension gets waved through is not the design, the materials, or the neighbour — it is which desk the application lands on. That is an uncomfortable fact for a system that is meant to apply one rulebook, and it is the strongest possible argument for reading your specific council's record before you pay anyone to draw anything.

Why the top of the table looks like inner London

The councils clustered at the top — Brent, Barking & Dagenham, Barnet, Bexley, Bradford, Bolton — are not random. Most are dense, heavily terraced, high-pressure urban authorities, and density is what turns an extension from a private decision into a contested one. On a detached plot with ten metres to the boundary, a rear extension harms nobody and sails through. On a Victorian terrace where your rear wall is your neighbour's daylight, the same extension triggers the objections planning exists to weigh: overlooking, loss of light, cramming, the erosion of a garden the local plan is trying to protect.

Several of these boroughs also make heavy use of Article 4 directions, which strip permitted development rights across whole streets — so work that would need no application in a leafier authority becomes a full application here, and a full application can be refused. The refusal rate isn't only measuring how strict the committee is. It is partly measuring how many marginal jobs the borough forces into the formal system in the first place. Both effects push the same way, and both are properties of the place, not your plans.

At the other end, the authorities refusing under 1% — Belfast, Angus, Ards & North Down, Allerdale — tend to be lower-density, with more space around each dwelling and fewer party-wall conflicts to adjudicate. When an extension can't hurt anyone, there is very little for a planner to refuse.

What to actually do with this

Three things follow, and none of them are "move house".

Find out whether you even need to ask. A large share of ordinary extensions fall under permitted development and need no application at all — which means no refusal is possible. Our answer to do I need planning permission to build an extension? walks through the size and siting limits that flip a job from "just build it" to "apply first". If your extension is inside those limits, this whole table is irrelevant to you.

Check whether your rights have been switched off. Permitted development is a default, and Article 4 directions, conservation area status and use class complications all revoke it in ways that don't show up in a photo of the house. A lawful development certificate is the instrument that puts "this is allowed" in writing before you start — and the document a future buyer's solicitor will ask for.

Read your council before you design, not after. If you are in a top-of-table authority, that is not a reason to give up — it is a reason to design for the refusal risk: keep the massing modest, respect the building line, pre-empt the overlooking objection, and consider a pre-application conversation. The refusal rate is high there precisely because marginal schemes get tested, so being unmarginal is most of the battle. Our guide to researching a neighbourhood before buying covers where these constraints cluster, and the wider planning explainer sets this refusal data in context alongside decision times and what actually gets refused.

If you want your own council's record — and the flood, conservation and permitted-development picture for a specific address — that is exactly what our property assessment is built to surface. You can also just ask about a house directly.

Methodology

Source and scope. Published local-authority planning records, analysed on 18 July 2026: 232,235 applications across 46 authorities.

Extension set. We selected the 71,376 applications whose published description mentions an extension. This is a text proxy, not a formal development class. We checked the obvious contamination — "extension of time" and "time extension" phrasing — and found only 22 records (0.03%), so the set reads as genuine building extensions.

Refusal rate. The denominator is decided applications: approved, approved with conditions or refused. The numerator is refusals. Withdrawn, undecided, unresolved, referred and blank outcomes are excluded. Authorities need at least 200 decided extensions to enter the ranking.

Limitations, stated plainly.

• 46 councils, not national. The set is alphabetically skewed (Aberdeen–Brent) plus Kent and a UK-nations sample. Do not read a national ranking into it; the councils not studied could sit anywhere on this range.
• Recency. The decisions are overwhelmingly recent — roughly 73% landed in 2023 or later — so this reflects the current regime, not a long historical average. (A small number of records carry obviously dirty decision dates; those don't affect a state-based count.)
• Six councils excluded for fewer than 200 decided extensions: Barrow (160), Babergh & Mid Suffolk (72), Blaenau Gwent (30), Blaenau (69), Bournemouth/Christchurch/Poole (65) and Alderney (0). Small samples make a refusal rate noisy, so we left them out of the ranking rather than mislead with a two-decimal figure built on a handful of cases.
• Applications, not households. A refused scheme is often revised and resubmitted until it passes. We are counting applications and their outcomes, not counting how many homeowners were ultimately told no forever.
• A refusal rate is not a verdict on a council. A high rate can mean a strict committee, or it can mean a dense borough that funnels far more marginal work into the formal system via Article 4 directions. The number is a real, useful signal for your odds — it is not a scorecard of officer reasonableness.

Council names are shown in readable form.