Recorded shoplifting in each Business Improvement District, compared with the area around it, September 2023–August 2024 against September 2025–August 2026. Shoplifting fell at least 10 points faster than the surrounding area in — of — BID zones and rose at least 10 points faster in —. Of the verdicts that survive every robustness check, — fell and — rose. Falls and rises are close to evenly matched, so the comparison shows no consistent direction for BIDs as a whole. It does not show cause: stores, police operations and recording changes move these numbers too.
This is a first pass, published to be corrected. Most boundaries outside London are provisional (grade C). We are writing to councils for each BID's property list and boundary, and inviting every BID to send its map, term and services. Method, data problems and every column are explained in the methodology and data dictionary.
BID zones with a shoplifting verdict. "Robust" means the verdict survives every check: without supermarket points, without the single biggest location, on alternative boundaries, and after the re-pinning check. Northern Ireland zones sit in the national build.
| Group | Zones | Fell | Rose | In line | Fell, robust | Rose, robust |
|---|
Gap = BID % change minus comparator % change (not a change in a rate). Sep 2023–Aug 2024 to Sep 2025–Aug 2026, with the same months dropped from both years where a force's data has a break (Devon and Cornwall Dec–Aug; Hampshire Sep–Jun; Northumbria no Jan–Feb; Gloucestershire Sep–Jan; Dyfed-Powys no Jan–Apr). Faded bars failed the robustness checks (verdict changes without supermarket points or without the single biggest location); hatched bars, and figures marked *, rest on small numbers (fewer than 100 offences in year one). Use the buttons to show one group.
National build rows add no-suspect rate, staffing model, crime tech and paid police time; gaps in those columns for London and pilot rows are gaps in the register, not findings. Boundary grades: A official GIS, B official street list or map, C traced map or own definition (provisional). Small numbers = fewer than 100 shoplifting offences in year one. Charge and no-suspect rates are for shoplifting recorded Apr 2025–Mar 2026.
| Zone | Group | Force | Role | Comparator | Shop Y1 | Shop Y3 | Shop % | Comparator shop % | Gap in % change | Verdict | Robust | Small numbers | Caveat | All crime % | Comparator all % | ASB % | Comparator ASB % | Shop charge rate, zone v comparator | No suspect, zone v comparator | Staffing model | Crime tech | Paid police time | Boundary | Operating model | Window |
|---|