51,000 homes - where does that come from?

A single number is doing a lot of work in the argument over Sheffield's Local Plan. People are being told that if the Plan is not adopted, the city faces a housing requirement of around 51,000 homes and would have to find room for more than 13,000 on the Green Belt, against roughly 3,900 Green Belt homes in the current proposals. The figure has appeared in published political statements and is repeated in debate about the Plan.

It is being used to close the debate down rather than inform it. The message is that objecting carries a penalty, and that the penalty is more development on the Green Belt.

The underlying concern is a fair one. The Leader of the Council, Cllr Fran Belbin, has warned of a real risk that rejecting the plan leads to "a new, higher housing target being imposed", requiring still more Green Belt land to be included. That is a legitimate argument about risk, and this article does not dispute it.

What this article examines is the specific figure of 51,000 homes, and the further claim about how much Green Belt would follow.

How the number was built

The 51,000 figure was not invented. But it is the product of two assumptions, and both have been overtaken by national policy.

Under the "standard method" that applied before December 2024, Sheffield's local housing need was calculated as 2,236 dwellings per year, to which a 35% increase was then applied because Sheffield was one of twenty urban authorities that had to use it. That increase was called "the cities uplift".

That produced 3,018 dwellings per year. Sheffield City Council's December 2022 housing land supply report records the number of dwellings per year required on exactly that basis.

Then, multiply 3,018 by the current Plan's seventeen-year period, 2022 to 2039, and the result is 51,306 - the figure now being quoted back to councillors as the price of voting the Plan down.

That calculation was put to the Local Plan examination in 2023, in representations from housebuilders and landowners arguing that Sheffield's requirement should be higher than the Council proposed. It was never an adopted requirement, and the Inspectors did not accept it.

What the Inspectors did conclude, in their post-hearings letter of January 2026, was that 2,236 dwellings per year represented a justified and reasonable housing requirement for Sheffield, giving 38,012 homes across the 2022–2039 period. That is the figure the current Plan is built on.

Neither part carries over automatically

The 35% increase was removed when the Government replaced the standard method in December 2024. Sheffield City Council recalculated its own local housing need shortly afterwards and published the result in its April 2025 housing land supply update: a new figure of 2,389 dwellings per year, against a previous figure of 3,036. The national change reduced Sheffield's calculated need by about a fifth.

The method changed again on 17 August 2026, when the Government published a rewritten National Planning Policy Framework (NPPF). The calculation is now in Annex D. It has two steps.

• Take 0.8% of the authority's existing dwelling stock, using the Government's Table 125.
• Adjust that figure according to the five-year average of the ONS median workplace-based affordability ratio.

Affordability data: ONS Table 5c, Sheffield, 2021–2025. Annual ratios: 6.15, 6.33, 6.49, 5.77 and 5.88; mean 6.124, published as 6.12. Stock: MHCLG Table 125. 252,656 × 0.008 = 2,021.25; affordability factor 1.2128; result approximately 2,451 dpa.

On the rules currently in force, a replacement Sheffield plan would be prepared under this method.

The plan period has changed too. Policy PM2 of the August 2026 NPPF requires a local plan to make provision for a period of no less than ten years from the point of adoption, reduced from fifteen. Longer periods remain permitted where the circumstances justify them. What is not required is that a replacement Sheffield plan reruns 2022 to 2039.

Rotherham is not evidence for Sheffield

Rotherham is also offered as proof that restarting the plan increases the housing requirement. Rotherham's figure did rise sharply after the December 2024 reforms, from 544 to 1,111 homes a year, and the council began a new Local Plan as a result.

But the same reform moved Sheffield in the opposite direction, from 3,036 to 2,389 homes a year.

The standard method runs on each authority's own housing stock and affordability data, so one authority's outcome carries no information about another's. Rotherham shows that national policy changes can move a plan's numbers.

A range, not a rival number

We are not going to answer 51,000 with a competing figure of our own. Nobody can state the housing requirement of a replacement Sheffield plan today, and anyone who claims to is making a mistake.

The inputs update annually, the plan period would be a matter for the new plan, and SYMCA's South Yorkshire Spatial Development Strategy could set Sheffield's requirement before a local plan does.

What can be shown is the range that current and recent figures produce across plausible plan periods. Every cell below is arithmetic, not forecast.

Green row: the requirement the Plan is currently built on. Red row: the abolished method from which the 51,000 claim derives. Thirteen years is shown because a plan begun now would take roughly three years to prepare and must then run ten years from adoption.

The 13,000 Green Belt figure

The second half of the claim - that 13,000 homes would need to be built on the Green Belt - is weaker than the first. There is an obvious candidate for its origin: 51,306 minus 38,012 is 13,294 more homes required. We have not found a published calculation setting out how the Green Belt figure was reached, so we do not assert that this is the method. But if it is, the arithmetic does not do what is claimed for it.

13,294 is the difference between two housing totals. It is not a measure of Green Belt capacity or Green Belt need. Converting one into the other requires a fresh assessment of urban capacity, brownfield land, densities and site availability, and then a Green Belt assessment. None of that work exists for a plan that has not been started.

National policy also stands in the way of a direct conversion. Before an inability to meet need can support altering Green Belt boundaries, an authority must examine reasonable alternatives, including making as much use as possible of previously developed and under-used land and optimising development densities. Where release is necessary, priority goes to previously developed land in the Green Belt, then to grey belt which is not previously developed, then to other Green Belt locations.

That order of priority is absolute. The Government has been explicit that it sits within an overall judgement about sustainable patterns of development, including sustainable transport, so a better-connected site on higher-performing Green Belt can come forward without every previously developed and grey belt option being exhausted first. It remains a priority that has to be worked through and evidenced, site by site, in a new plan. It is not a guarantee that Green Belt would be spared.

The risk of starting again is real

The case for restarting should be made honestly, as it carries costs that councillors should consider.

The Inspectors' figure of 38,012 would no longer be secured. It is the product of the current examination, on this evidence base, under the Framework that applied when the Plan was submitted. A replacement plan would be prepared under the August 2026 NPPF and would have to establish its own requirement. That requirement could come out lower. It could also come out higher, and Policy HO2 allows a requirement above the standard method figure where an authority takes on unmet need from neighbours or plans for economic growth.

There is a further consequence that matters more in practice than the plan-making arithmetic. Without an up-to-date plan, Sheffield's position on housing land supply weakens, and under the August 2026 NPPF a shortfall in the five-year supply, or a Housing Delivery Test result below 75%, can count as evidence of unmet housing need in decisions on individual applications. That same evidence feeds the grey belt route. The practical effect is that decisions shift from the plan, where allocations are debated in public and tested by Inspectors, towards appeals brought by developers on sites of their choosing.

What a restart would and would not change

Two things can be said with reasonable confidence.

A replacement plan could use a shorter period. A plan covering ten years from adoption commits less land now than one covering seventeen. On the current annual figure, that is a total in the region of 24,500 rather than 38,012, with the later years addressed by a subsequent plan against evidence that exists at the time rather than assumptions made today.

Green Belt release would be reassessed under the current priority order. Many of the proposed Green Belt releases are open farmland, not previously developed land and not obviously grey belt. Under the current NPPF they sit behind previously developed Green Belt and grey belt, subject to a sustainability judgement. That is a materially different starting point from the one under which they were selected for the current draft Plan.

The Inspectors have found that the current evidence base does not identify enough suitable land outside the Green Belt to deliver 38,012 homes. A new plan might find more brownfield land, through different density assumptions or a fresh call for sites. It might not. What cannot be claimed is that restarting would save Sheffield's Green Belt or require more Green Belt. The work would have to be done again, and nobody can state the outcome in advance.

What we are asking councillors to do

This article does not ask any councillor to vote a particular way on the Local Plan. It asks for something narrower.

1. Do not repeat a figure you cannot source. Before using 51,000, ask which annual figure and which plan period produce it under the rules now in force. The old 51,306 calculation cannot be reproduced using current standard-method inputs over the current Plan's 17-year period, and no current rule requires that total.
2. Ask for the Green Belt working. A difference between two citywide housing totals is not a Green Belt capacity assessment. If more than 13,000 Green Belt homes are being predicted, the land supply calculation behind that prediction should be published.
3. Consider the real risks, not unsupported figures. There are genuine costs to starting again, and they are set out above. They are strong enough to argue on their own terms. They do not need a number from an abolished method to support them.

Decisions about Sheffield's Green Belt are irreversible in a way that housing figures are not. Councillors will be asked to take one of them. They are entitled to take it on figures that can be traced to the rules currently in force, and to have the evidence behind any figure put in front of them before they are asked to act on it.