Data Limitations & Caveats
- Nationality, not ethnicity. All datasets use nationality (country of citizenship). Nationality is not a proxy for ethnicity or race.
- "Returns" includes many outcomes. Enforced removals, voluntary departures, port refusals. Not all are forced removals.
- No individual-level linking. Prison population and deportation data cannot be linked at the individual level — all cross-references are aggregate correlations.
- Correlation, not causation. A high prison population relative to deportations does not imply those individuals should or could be deported.
- Prison population is a snapshot. It measures who is in prison at a point in time (stock), not annual flow. Direct comparison with annual returns is approximate.
- Age-sex adjustment matters. When adjusted for the younger, more male demographic profile of foreign nationals, non-UK nationals are slightly underrepresented in prison overall (Migration Observatory, Sept 2025).
- Population denominators are uncertain. For small nationalities (Afghan, Eritrean), ONS estimates may understate the actual UK population by 2–3x, which inflates per-capita rates.
Sources:
Home Office Immigration System Statistics (Feb 2026) |
Home Office Immigration Enforcement Returns Cost (Mar 2026) |
MoJ Offender Management Statistics Quarterly (Jan 2026) |
MoJ Tribunals Statistics Quarterly (Dec 2025) |
MoJ PNC via Centre for Migration Control FOI (2025) |
MoJ CCSQ Court Interpreter Tables |
House of Lords 'Lost in Translation' (Mar 2025) |
Metropolitan Police CMC FOI (Jul 2025) |
ONS Long-term International Migration (May 2026) |
ONS NEET Bulletin (Feb 2026) |
ONS Births by Parents' Country of Birth (2024) |
Centre for Social Justice / HMRC payroll analysis (May 2026) |
Eurostat Returns of Irregular Migrants (2025) |
Migration Observatory, Oxford (2026) |
House of Commons Library (Mar 2026)