More or Less Behind the Statistics
More or Less Behind the Statistics

Why are so many ex-prisoners being sent back to prison?

Tim Harford investigates some of the numbers in the news. This week: (00:38) The number of people being recalled to prison after early release in England and Wales is rising sharply. Why? (07:37) Is the UK millionaire an endangered species? We ask Duncan Weldon - he’s like David Attenborough, but fo

Featured Speakers

BBC Host

Topics Discussed

Episode Summary

Executive Summary: This episode fact-checks viral claims about prisons, millionaires, wildfires, and seagull food theft. It finds early prison release is linked more to higher recall rates and probation strain than to more reoffending; millionaire counts are shaky estimates driven by inflation and asset prices; Europe’s wildfire picture depends heavily on whether Russia and agricultural burning are included; and the seagull food-loss headline is inflated by dubious survey methods.

Main Topics: Prison recalls and early release (Priority: 5/5): The programme tests whether early prison release caused a surge in recalls. Experts explain recalls are mostly for probation breaches, not new crimes, and that probation pressure and shorter preparation time after release likely drive the rise. Headline about prisoners 'returning to jail at rapid rate' (Priority: 4/5): The BBC headline is scrutinized for implying causation. The discussion shows recall numbers rose sharply, but that does not mean early release directly makes people reoffend more. UK millionaires and the '7% decline' (Priority: 5/5): The Telegraph claim about shrinking millionaire numbers is unpacked as a model-based estimate using wealth data, inflation adjustments, and a Pareto distribution, not a direct count of rich people leaving Britain. Wildfires in Europe and global trends (Priority: 5/5): A widely shared graph from Our World in Data is examined. Europe looks low only because Russia dominates the regional total and because agricultural burning is included; excluding that, the EU has a high year. Seagulls stealing food worth £171 million (Priority: 4/5): The seagull loss estimate is challenged as coming from a survey that conflates 'targeted' with 'stolen' food and extrapolates an implausibly high frequency of incidents to the whole adult population. How statistics shape news narratives (Priority: 5/5): Across all segments, the episode shows how definitions, sampling, model assumptions, and headline framing can radically alter public understanding of the same underlying data.

Key Arguments: A rise in prison recalls does not prove early release increases reoffending; most recalls are for non-compliance rather than new offences. Probation services are stretched and risk-averse, which likely pushes more recalled prisoners back into custody for minor breaches. The number of UK millionaires is not directly measured well; estimates depend heavily on wealth modelling, inflation, and asset-price movements. The millionaire headline is more about people crossing an arbitrary wealth threshold than about ultra-rich individuals emigrating. Europe’s wildfire totals are distorted by Russia’s inclusion in the regional aggregate; the EU-only picture is very different. Wildfire datasets can include agricultural burning, so 'wildfire' headlines may not reflect what listeners imagine as accidental forest fires. The seagull-loss figure is likely inflated because the survey counts being 'targeted' as if food was actually stolen. The average value used in the seagull estimate likely overstates actual losses because partial thefts or mere harassment were treated too similarly. Survey responses suggesting nearly annual seagull theft are implausible and may reflect poor survey quality or bot-like answers.

Data Points: Prison recalls (year ending March 2026): 51,419 - BBC cited this as evidence of more people being sent back to prison after early release. Year-on-year change in recalls: +28% - Recalls increased compared with the previous year. Recall rate among probation caseload: 74 recalls per 100 people - Cassia Rowland said the rate rose from 33 in 2018 to 74 in 2025. Recall rate in 2018: 33 per 100 people - Baseline used to show the increase in recalls over time. Recall composition: About 20%-25% for a further offence - Most recalls are for breaches/non-compliance rather than new crimes. Serious further offences charged in 2024-25: Just over 800 people - People charged with murder, attempted murder, rape, and similar serious offences while on probation or just after supervision ended. Increase in serious further offences charged: About 50% - Compared with 2022-23. UK millionaires estimate: Under 442,000 - Adam Smith Institute tracker estimate referenced in the Telegraph headline. Change in millionaire estimate: 7% fewer than last year - The headline claim about a shrinking millionaire population. Wealth measure used: Constant-price sterling millionaires - The tracker adjusts wealth to 2025 pounds rather than counting literal millionaires in each year. UK adults in seagull survey: 2,000 - Compare the Market’s Great Gull Index survey sample. Respondents saying food had ever been stolen or targeted: 71% - Used to extrapolate seagull losses to the UK adult population. Average reported theft frequency: 0.8 times per person per year - Applied to the estimated affected population in the seagull calculation. Average reported value of food: £5.50 - Used in the seagull-loss estimate, though the survey included 'targeted' incidents. Seagull food-loss estimate: £171,889,518 - Headline figure derived from the survey and extrapolation. Europe wildfire burn share from Russia: Around 83% - Explains why Russia dominates Europe-wide wildfire totals in the data used by Our World in Data. EU wildfire ranking this year: Around the fifth highest - If Russia is excluded and the EU is examined separately, the year is not a record low. Global burned area from Africa: Around 70% - Shows why Africa largely determines global wildfire totals. Global wildfire burn trend: Around 25% decline over the last few decades - Attributed mainly to land-use change, especially conversion of savannah/grassland to cropland.

Pivotal Quotes: "we don't have any evidence that early release makes you more likely to re-offend, but we do know that it makes you more likely to be recalled to prison" — Cassia Rowland: Explanation of why the prison-coverage headline is misleading about causation. "There is no little black book of millionaires." — Duncan Weldon: Point that UK millionaire counts are model-based estimates, not direct administrative counts. "whether it's a high or low welfare year in Europe, depends on whether it's a high or low year for Russia" — Hannah Ritchie: Clarifies why Europe-wide wildfire totals are dominated by Russia in the dataset.

Implications: Listeners should treat headline statistics cautiously: definitions, sampling, and geographic aggregation can radically change the story. Policy debates on prisons, wealth, climate, and consumer behavior need the underlying data context, not just viral numbers.

🔓 Sign Up for Unlimited Episode Search

About More or Less Behind the Statistics

Tim Harford and the More or Less team try to make sense of the statistics which surround us. From BBC Radio 4

View all episodes from More or Less Behind the Statistics