PROVENANCE
Sources &
method.
Stalled London contains no estimated, modelled or illustrative numbers. Every figure on the site is either published directly by a government body or is an arithmetic aggregation of published records, described below. Data retrieved 2026-08-01.
1. Stalled homes and the register
We queried the Datahub’s residential-unit and application indexes for every London borough decision between 2015 and 2021 with a residential unit record (160,000+ unit rows). A scheme is counted as stalled when it was permitted and has no actual commencement date recorded, and its status is either “Lapsed” (permission expired) or “Approved” (live consent, no start logged). That yields 7,818 schemes and 100,477 homes, of which 19,988 were consented in affordable tenures (social rent, affordable rent, London affordable rent, shared ownership and other intermediate products, as classified in the Datahub tenure field).
The register table lists the 164 largest of those schemes, with the application reference, site address, tenure mix and decision date taken verbatim from the Datahub record.
2. Demand and delivery layers
The demand layer uses each borough’s homes required for 2020–23 and the delivery layer uses its published Housing Delivery Test percentage, both read from the official measurement spreadsheet. Across the 33 London authorities the test records 111,299 homes delivered against 126,991 required. Consequences (presumption in favour of sustainable development, 20% buffer, action plan) are the government’s own designations.
3. Planning delays layer
Percentage of decisions on applications for major development determined within the 13-week statutory period, April 2024 – March 2026, per local planning authority. We deliberately show the raw statutory measure rather than the headline “within agreed time” figure, which counts extensions of time agreed with the applicant and therefore sits near 100% for almost every borough. Both numbers appear on the borough scorecard.
4. Stall risk score
The only composed number on the site. For each scheme: years since permission (capped at 12) weighted 40%, homes consented (capped at 1,000) weighted 30%, affordable share of the scheme weighted 20%, and the inverse of the borough’s Housing Delivery Test score weighted 10%. Expressed 0–100. It ranks published facts; it is not a prediction.
Known limitations
Datahub records are submitted by boroughs, and completeness varies. A missing commencement date is evidence of a reporting or delivery gap, not proof that a site is empty today — several large schemes recorded as lapsed have since been consented again or partly built under later permissions. Layers that could not be sourced to a verified published dataset were not included, and no borough figure on this site is interpolated.