State primary schools in the 594 English towns we can measure hold 2,707,010 places. 255,319 of them are empty.
That is not a scandal and mostly not a mystery. Births in England peaked around 2012, have fallen most years since (the ONS series is here), and the dip has now worked all the way through the primary years. Buildings sized for the peak hold the trough. Nationally that shows up as primaries running at 89.2% of capacity across measured towns; 307 of 594 towns sit below 90%, and 62 below 80%.
What the national number hides is where. This is the same register read town by town.
The emptiest primary schools in England
| # | Town | Places | Pupils | Filled |
|---|---|---|---|---|
| 1 | Ross on Wye | 826 | 541 | 65.5% |
| 2 | Bridport | 840 | 561 | 66.8% |
| 3 | Tidworth | 1,635 | 1,099 | 67.2% |
| 4 | Ilkley | 1,330 | 904 | 68.0% |
| 5 | Amesbury | 1,290 | 884 | 68.5% |
| 6 | Thetford | 2,754 | 1,923 | 69.8% |
| 7 | Teignmouth | 1,050 | 734 | 69.9% |
| 8 | Earl Shilton | 1,260 | 881 | 69.9% |
Three different stories share this table. Tidworth is a garrison town whose families arrive and leave with postings, so its rolls swing in ways no birth rate explains. Bridport and Ross-on-Wye are market towns ageing in place. Teignmouth is the coastal version of the same thing, and its age profile says so.
The counterexample is instructive: Bexhill-on-Sea is one of the oldest towns in England by median age, and its primaries still run at 98% of capacity. A retirement town keeps few schools, and the children it does have concentrate in them. Emptiness tracks the gap between what was built and who lives there now, which is why the fill figure and the age profile sit on the same page.
The other end
Tyldesley runs at 108.9% of stated capacity, which is possible because schools can admit above their published number in bulge years. Towns over 100% cluster where new housing outran school building. Bolton sits at 98.7% with nearly 20,000 places, which for a town its size means the system is full in the only sense that matters: try finding a place at the school you actually want.
How to read the number, and how not to
Fill is places against pupils from one register at one date, summed only over schools that report both. It says nothing about whether a school is good, and an emptying school is often a well-run school in a town with fewer children. What it does predict, run low for long enough, is closure: capacity follows rolls eventually, and the towns at the top of this table are where that conversation happens first.
One number we refuse to publish sits next to this one. Schools per child sounds simple, but the census gives town ages in five-year bands and primary age is four to ten, so the honest version on each town page counts children aged five to nine, a complete band, and says exactly that. A cleaner-sounding number would be a made-up one.
Source: the Department for Education's Get Information about Schools register (capacity and each school's latest census return), located to ONS built-up areas by postcode; towns with three or more state primaries and 10,000+ residents, where at least nine in ten schools report both figures. A school reporting pupils at 125%+ of its stated capacity is treated as a stale capacity record and excluded. England only. Each town's own page carries its figures.