Demographic change and service exposure index
A composite score from 0 to 100 built from five capped components: White British share change, asylum support rate, EAL demand growth, non-English main language, and foreign-born growth. It measures exposure to demographic change. It does not measure service cost, service capacity, or observed demand.
Top 30 areas by exposure score
#AreaScoreBandWhite British change
1 Reading 97 High exposure -18.6pp 2 Hillingdon 93 High exposure -21.4pp 3 Coventry 92 High exposure -21pp 4 Nottingham 92 High exposure -21.7pp 5 Wolverhampton 92 High exposure -21.2pp 6 Manchester 91 High exposure -23.2pp 7 Sandwell 91 High exposure -24.6pp 8 Hounslow 90 High exposure -13.5pp 9 Leicester 90 High exposure -18.2pp 10 Luton 90 High exposure -18.5pp 11 Croydon 87 High exposure -16.7pp 12 Derby 87 High exposure -20.6pp 13 Peterborough 87 High exposure -19.9pp 14 Blackburn with Darwen 86 High exposure -22.3pp 15 Barnet 85 High exposure -14.7pp 16 Birmingham 83 High exposure -20pp 17 Slough 83 High exposure -13.6pp 18 Stoke-on-Trent 82 High exposure -20.3pp 19 Middlesbrough 82 High exposure -16.6pp 20 Lambeth 82 High exposure -14pp 21 Hammersmith and Fulham 82 High exposure -15.6pp 22 Bolton 81 High exposure -27.5pp 23 Bradford 80 High exposure -17.5pp 24 Newcastle upon Tyne 80 High exposure -19.3pp 25 City of London 80 High exposure -17.2pp 26 Oldham 79 High exposure -23.2pp 27 Telford and Wrekin 79 High exposure -13.7pp 28 Barking and Dagenham 79 High exposure -21pp 29 Ealing 78 High exposure -9.9pp 30 Redbridge 78 High exposure -14.4pp How the score is built
Each component is the input value multiplied by a fixed weight, then capped at 20 points. The five capped values are summed, the total is capped at 100, and the result is rounded to a whole number. There is no normalisation and no weighting derived from data. The multipliers and the caps were chosen when the score was written.
| Component | Input | Source dataset | Multiplier | Cap | 20 points reached at | Areas scoring 0 |
|---|---|---|---|---|---|---|
| White British share change | Absolute change in White British share, 2021 to 2041 | Census 2021 custom ONS dataset (single year of age by ethnic group by sex), projected by the Hamilton-Perry method | 1.5 | 20 | A change of 13.3pp in either direction | 0 of 318 |
| Asylum support rate | Supported asylum population per 10,000 residents, quarter ending 31 December 2025 | Home Office Immigration Statistics, asylum support by local authority | 0.5 | 20 | A rate of 40 per 10,000 | 65 of 318 |
| EAL demand growth | Minority pupil share minus minority resident share, in percentage points | DfE School Census 2024/25 pupil ethnicity by local authority, against the Census 2021 resident share | 1.0 | 20 | A gap of 20pp | 192 of 318 |
| Non-English main language | Share of residents aged 3 and over whose main language is not English, as measured in 2021 | ONS Census 2021 TS029 Proficiency in English via NOMIS (NM_2048_1) | 1.0 | 20 | A share of 20% | 0 of 318 |
| Foreign-born growth | Change in foreign-born share, 2021 to 2051 | ONS Census 2021 TS004 and Census 2011 KS204 country of birth via NOMIS, projected by the Hamilton-Perry method | 0.5 | 20 | A change of 40pp | 13 of 318 |
Band cutoffs
| Band | Rule |
|---|---|
| High exposure | Score 70 or above |
| Moderate exposure | Score 45 to 69 |
| Low exposure | Score 25 to 44 |
| Stable | Score below 25 |
The band names shown here are display labels. The category values inside fiscal-resilience.json are unchanged, so anyone reading the data file directly sees the generator's own wording against the same cutoffs.
What the score does not do
- A missing component scores zero, not unknown. Where an input is absent the generator substitutes 0 and the total is still published as if the component had been measured. An area with no school data therefore scores lower than a comparable area with school data, rather than being marked incomplete. The counts in the table above are areas scoring zero on each component, which is a missing input and a genuine zero combined: the published file does not distinguish them.
- Two of the five components are levels, not changes. The asylum support rate and the non-English main language share are both stocks: what an area holds now, measured once. The other three are changes over time. An area with a large settled population whose main language is not English scores on that component whether or not anything there is changing, and the two kinds of measure are added together as though they were the same kind.
- The index has not been validated against observed service demand. No comparison has been made against NHS interpreter spend, school place pressure, waiting lists, adult social care caseloads, or any other measure of what services actually deliver. The score has never been tested against an outcome.
- It measures exposure, not cost or capacity. A high score means the area is projected to see fast demographic change on these five measures. It does not follow that services there cost more, are under more strain, or are failing. Demographic change is not shown here to cause service pressure, and this page does not claim that it does.
- The components use different time horizons. White British share change runs 2021 to 2041, foreign-born change runs 2021 to 2051, the non-English main language share is a Census 2021 observation, and the asylum and school inputs are later single snapshots. They are summed as though they were comparable.
- Direction is discarded. The White British share component uses the absolute change, so an area where the share rises scores the same as one where it falls by the same amount.
Scoring is computed in scripts/model/generate_national_dashboards.mjs and published to src/data/live/fiscal-resilience.json. Underlying datasets are listed on sources.