Ask why a rural district with a National Landscape across most of it has been told to build at more than twice its previous rate, and one answer comes back more often than any other. The homes are for people arriving from elsewhere. Population is rising because of migration, so the government needs somewhere to put people, so the countryside is being built on.
It is a coherent theory. It is also, unusually for an argument about planning, entirely testable, because every input to it is published. The Home Office reports asylum support by local authority. Out of area homelessness placements are counted by the Ministry of Housing, and the Office for National Statistics publishes both the population projections and the affordability ratios that drive housing targets. The formula that produces a council's number is set out in guidance and can be worked through by hand.
So this piece does not argue with the theory. It looks up the numbers, one system at a time, and reports what they say. Cotswold District is used throughout as the worked example, because it is the district where the targets have risen most sharply and the constraint is greatest.
Where asylum seekers are actually housed, by local authority.
Since April 2022 every local authority in England has formally been a dispersal area. That is a real policy change and it is often the starting point for the argument above.
Nationally, at the end of March 2026, 97,519 people were on asylum support. That is 9% down on a year earlier and 21% below the September 2023 peak of 123,758. Of the 93,650 actually housed, 68,720 were in longer term dispersal accommodation and 23,950 were in initial or contingency accommodation. Hotels specifically accounted for 20,885, which is 63% below the September 2023 hotel peak of 56,018.
The distribution is heavily uneven. Across the UK the average is 14 supported asylum seekers per 10,000 residents. The North East is highest at 26 and the North West next at 25. The South West is the lowest region in the United Kingdom, at 6.2 per 10,000, or 3,639 people. Forty of the 361 lower tier authorities in the UK contain no housed asylum seekers at all. Glasgow has the most of any single authority, at 3,724, or 57.3 per 10,000 residents.
Cotswold District is one of the forty.
The Home Office table for 31 March 2026 records the district's supported asylum total as zero. Not a small number. Zero: no dispersal accommodation, no contingency accommodation, no hotels, and nobody on subsistence only support. Neighbouring Tewkesbury is also at zero. Gloucestershire's asylum population sits in Gloucester city, at 316, and Cheltenham, at 69.
The district's arrivals under the other resettlement routes are on a similar scale. 383 people have come to Cotswold under Homes for Ukraine since 2022, cumulatively, sponsored by local hosts, and 4 under the Afghan Resettlement Programme.
One figure that circulates locally is worth handling carefully. A Gloucestershire county estimate of roughly 3 arrivals per 1,000 residents combines all three routes: asylum support, Homes for Ukraine and Afghan resettlement. It is not a dispersal figure and should not be quoted as one. The county Director of Public Health recorded 2,454 refugees and asylum seekers across a population of 659,276 in 2024, which is 0.37%, below the UK average of 0.43%.
Out of area temporary accommodation placements, and where they go.
The second version of the argument is about temporary accommodation. London boroughs, unable to house homeless households locally, place them elsewhere, and the receiving areas absorb the demand.
The first half of that is true and the scale is significant. At 31 March 2024, London boroughs had 28,540 households in temporary accommodation placed outside their own district. Nationally, out of area placements rose 42% between the third quarter of 2018/19 and the same quarter of 2023/24, from 23,450 to 33,350. By the end of June 2025 the figure was 42,080 households.
The second half does not follow, and the reason is in the same tables. Of those 28,540 London placements, about 85%, or 24,300, remained within London. The largest destinations outside London were the South East, at 2,420 households, and the East of England, at 1,560. Nationally, MHCLG records that 86.3% of out of area placements stay within the originating authority's own region.
For the South West the regional total rounds to zero. The only receiving authorities in the South West that appear in the London table are Gloucester and Sedgemoor, both suppressed because they hold fewer than five households. Cotswold does not appear as a receiving authority at all.
The two systems do interact, and it would be wrong to leave that out. The National Audit Office recorded the Home Office accommodating around 100,000 asylum seekers in local authority areas as at December 2023, and the Public Accounts Committee warned in October 2023 that Home Office competition for housing was driving up prices and worsening the homelessness pressures councils already faced. That is a real effect on the housing market. It is not a placement of people into Cotswold District, because the tables show that is not happening.
What is driving UK population growth, on the ONS projections.
Here the folk theory is broadly right, and it should be said plainly.
The ONS 2022 based principal projections, published on 28 January 2025 and still the current set, project the United Kingdom growing from 67.6 million at mid-2022 to 72.5 million at mid-2032. That is a rise of 7.3%, or 4.9 million people. Almost all of it is net migration. Births and deaths over the period roughly offset one another, so natural change is approximately zero and the projection assumes long term net migration of 340,000 a year from mid-2028. England grows fastest of the four nations, at 7.8%.
So national population growth is a migration story. That is what the official projections say and there is no serious argument about it.
The question this piece is asking is narrower and it is the one that matters for a planning target. Does any of that reach the formula that told Cotswold District what to build?
The standard method housing formula, worked through step by step.
It does not, and the way to demonstrate it is to do the arithmetic.
The standard method was rewritten in December 2024. The change that matters is that household projections, which are the only part of the old calculation that counted people, were dropped altogether. What replaced them is dwelling stock.
The calculation now runs in two steps.
Step one. Take 0.8% of the existing housing stock. Cotswold District had 46,213 dwellings in 2023, on MHCLG's own live table. That produces a baseline of 370 homes a year.
Step two. Apply an affordability adjustment based on the average of the five most recent ONS median workplace based affordability ratios. For Cotswold, covering 2019 to 2023, those ratios are 13.38, 14.10, 15.70, 14.35 and 15.32, which average 14.57. The formula takes that average, subtracts 5, divides by 5, multiplies by 0.95 and adds 1. That gives an adjustment factor of 2.82.
Multiply the baseline by the adjustment and you get 1,042 homes a year. The official figure at the time was 1,036, and the small difference is ONS back revision of the ratio series rather than different arithmetic.
So of that number, roughly 370 comes from the stock baseline and roughly 666 comes from the affordability multiplier.
Nowhere in either step is there a term for people. Not households, not births, not deaths, not migration, not the population projections described in the section above. The inputs are a count of existing buildings and a ratio of house prices to earnings.
And the ratio deserves a closer look, because it is where 64% of the number comes from. It compares median house price, which in Cotswold was £452,500 for the year ending September 2023, against median workplace earnings, which were £29,528 in 2023. Workplace earnings measure what jobs located in the district pay. They do not measure what the people buying the houses earn.
In a district where a substantial share of buyers are retired, or commuting to Swindon, Cheltenham, Oxford, Bristol or London, or buying a second home, the two figures describe different populations. The gap between them is read by the formula as unmet local need, and converted into a housing target.
Put the four systems together and the result for this district is the inverse of the theory. The place ordered to build 1,036 homes a year, since revised upward to 1,054, houses no supported asylum seekers, receives effectively no out of area homelessness placements, and has a housing target that contains no population input of any kind.
The fields are being built on for a number, and the number does not contain people. It contains house prices and it contains wages, and it reads the distance between them as need.— The Editor · Ground Level, September 2026
What these figures do and do not show.
Three qualifications, because a piece like this is only worth reading if it states its own limits.
This is one district. Cotswold is the worked example because it is the case with the sharpest target rise and the greatest constraint. The national policy geography is real: full dispersal does exist, out of area placements are rising, and migration is driving population growth. A district with a large asylum population and heavy inward placement would produce a different local picture, and the tables would show it. The point is not that the systems do not exist. It is that in this district they can be checked, and they are not the mechanism.
The formula being a formula is not an argument against the homes. Nothing above establishes that Cotswold does not need housing. The waiting list is long, local earnings do not buy local houses, and young people leave. The affordability ratio is measuring something real. The objection is narrower: a ratio driven substantially by who is buying is being used as a measure of how many homes local people need, and those are different questions with different answers.
Some figures here move. The asylum data is a snapshot at 31 March 2026 and is published quarterly. The out of area placement tables are periodic. The standard method output is recalculated as the affordability ratios and stock estimates refresh, which is exactly why the district's figure has already moved from 1,036 to 1,054 without anything happening on the ground. What does not move is the structure: a stock baseline, multiplied by a price to wage ratio, with no person in it.
Why the arithmetic explanation is the stronger one.
There is a temptation, when a widely held explanation turns out not to survive the tables, to treat the correction as the whole story. It is not, and treating it that way would miss what the numbers actually show.
The countryside around Cotswold market towns is being allocated for development at a scale nobody local voted for. That is true, it is documented in the council's own plan, and this publication has followed several individual cases through to their conclusions. The pressure is real.
What the figures establish is what is producing it. Not a policy of moving people into the district, because the district houses almost none. An arithmetic operation, performed on a spreadsheet, that reads the distance between what houses cost in the Cotswolds and what jobs in the Cotswolds pay, and converts that distance into an obligation to build.
That is a harder thing to campaign against than a conspiracy, because there is nobody to blame and no decision to reverse. It is also, unlike the alternative, checkable line by line by anyone who wants to open the tables.
Every figure in this article comes from a published dataset. The Home Office local authority tables, MHCLG's live table on dwelling stock, and the ONS affordability ratio series are all free to download.