There are two ways a council in England can lose the ability to refuse housing it does not want.
The first is well known, and this publication has written about it several times. If a council cannot demonstrate a five year supply of deliverable housing land, the presumption in favour of sustainable development engages and its own protective policies carry reduced weight.
The second is the Housing Delivery Test. It measures what a council actually built against what it was supposed to build, over three years. Deliver less than 85% and a 20% buffer is added to the supply the council has to demonstrate, which makes the first test harder to pass. Deliver less than 75% and the presumption applies directly, by a second route, whatever the land supply position says.
It is a serious mechanism. It is the difference between a council that can say no and a council that mostly cannot.
Which raises a question that turns out to have an uncomfortable answer. What is your council's current score?
Every Housing Delivery Test result, and when it was published.
The Housing Delivery Test results are published by the Ministry of Housing, Communities and Local Government in a collection on GOV.UK. Every measurement since the test began sits there in one list. This is that list.
| Measurement | Published |
|---|---|
| 2023 | 12 December 2024 |
| 2022 | 19 December 2023 |
| 2021 | 14 January 2022 |
| 2020 | 19 January 2021 |
| 2019 | 13 February 2020 |
| 2018 | 19 February 2019 |
The most recent result is the 2023 measurement, published on 12 December 2024.
There is no 2024 measurement. There is no 2025 measurement. The collection page carries no notice explaining why, no holding statement, and no revised date. It stops.
What the Housing Delivery Test Measurement Rule Book promised.
The Housing Delivery Test Measurement Rule Book was published in July 2018, when the test was created. It set out how the calculation works and when the results would appear: annually, in November each year, covering the three preceding financial years.
Set that against the record above. The 2018 measurement arrived in February 2019. The 2019 measurement in February 2020. Then January, then January, then a gap of nearly two years to December 2023, then December 2024, and then nothing.
The promise of November publication has never once been kept. What has changed is that the slippage used to be measured in weeks and is now measured in years.
Why the current score describes housebuilding that ended in March 2023.
The naming is worth pausing on, because it hides the real age of the data. Results published in one year measure delivery over the three financial years before it. The 2023 measurement, published in December 2024, assessed what councils built in the three years ending 31 March 2023.
So the score currently attached to every local planning authority in England, the one that determines whether a 20% buffer applies and whether the presumption engages by the delivery route, describes housebuilding that finished more than three years ago.
Consider what has changed since March 2023. The standard method was rewritten in December 2024, which for many councils more than doubled the number they are now measured against, and did so without altering a single completion in the period being measured. In August 2026 housebuilding applications reached a five year high while new build completions fell to their lowest level in over a decade, which is a combination the score is incapable of describing.
None of that is in the score. The score stopped in March 2023.
How Cotswold District Council applied a 20% buffer to itself.
The clearest illustration of what this does sits in the evidence base Cotswold District Council published in July 2026 for its new Local Plan.
The council records its Housing Delivery Test measurement as 95%. On that figure it is required to apply a 5% buffer to its housing land supply, and it says so, citing the December 2024 framework.
Then it applies 20% instead.
Its reasoning is set out plainly. Given how far delivery has fallen short of the revised standard method over the past two monitoring years, the council anticipates that it will soon fail to deliver more than 75% of its requirement, and it expects that position to persist for several years. So it has taken what it calls a precautionary approach and applied the harsher buffer to the first three years of the plan period.
Read that again with the publication gap in mind. The 95% is the score from the measurement published in December 2024, describing delivery to March 2023. The council does not believe it any more. It cannot see the number that would replace it, because that number has not been published. So it has estimated its own likely failure and imposed the consequence on itself in advance.
A council that cannot see its own score cannot know which rules apply to it. Cotswold has responded by marking itself down, and it is hard to argue it is wrong to.— The Editor · Ground Level, August 2026
This is not a criticism of Cotswold. Volunteering a harsher test than the one you are formally subject to is a defensible response to an absence of information, and an inspector examining the plan in 2027 is more likely to accept a trajectory built on a pessimistic assumption than an optimistic one. It is a rational thing to do.
It is also a council marking its own paper, because the examiner has not marked it.
How many councils may be below the 75% threshold.
The honest answer is that nobody knows, and that is the point.
In the absence of official results, the consultancy Pegasus Group has published estimates. Its figures suggest that the number of authorities subject to the presumption in favour would rise from 59 on the 2023 measurement to 80 on a 2024 measurement and 92 on a 2025 measurement, and that the number facing a 20% buffer would rise from 87 to 97 and then to 126. It estimates that more than 150 authorities may have a 2024 result below 75%.
Those are estimates by a private firm, and they should be read as estimates. This publication cites them because they are the only figures available, not because they carry the authority of a published measurement. The entire difficulty is that no such measurement exists.
The department has acknowledged the delay elsewhere. It attributes it to the volume of planning reform and consultation, and has said it is considering ways to speed up future collections, including combining the 2024 and 2025 data collections into one exercise.
Combining them would produce a single result covering a longer period. It would also mean that the 2024 measurement, as a distinct assessment of a distinct three years, never appears at all.
Why an unpublished Housing Delivery Test matters.
There is a fair objection to make here, and it should be made before a reader makes it.
Very few planning decisions turn on the Housing Delivery Test alone. The five year land supply test does most of the work, it is calculated by councils themselves, and it is updated far more often. A council with a healthy supply and a poor delivery record is in a much better position than the reverse. So a delay to the delivery test results is not, on its own, why fields are being built on.
That is true, and it is not the argument.
The argument is about what a test is for. The Housing Delivery Test exists to hold councils to account for what they actually deliver, and the consequences it carries are automatic. A council does not get to argue about whether the presumption applies. It applies, on a number, published by the department, on a date.
Take away the publication and the accountability does not become gentler. It becomes unpredictable. Councils cannot plan against a measurement they cannot see. Developers cannot rely on it. Inspectors weigh a figure that describes a period ending three years before the hearing. And a district drafting a plan today, as Cotswold is, has to guess at its own future failure and write the guess into its evidence.
The government has spent two years telling councils that housing delivery is the test that matters, while not publishing the test.
What to watch for in the 2024 and 2025 results.
The 2024 results, if they appear as a separate measurement, will be the clearest signal. If the two collections are combined instead, the department will have quietly retired a year of accountability, and the combined figure should be read with that in mind.
The second thing to watch is the replacement National Planning Policy Framework, consulted on between December 2025 and March 2026 and promised for summer 2026. It has not yet been published. The buffer thresholds and the delivery consequences sit inside the framework, and a rewrite is the natural moment either to fix the publication problem or to make it permanent.
Your council's Housing Delivery Test result is a matter of public record. So is the fact that the record stops in December 2024.