Government waste

Episode Summary

Executive Summary: This More or Less episode examines how numbers are used—and misused—in public debate. It debunks exaggerated claims about local-government waste and laptop savings, explains how riot arrest statistics and alcohol-hospital data can be misleading, interviews intelligence chief Alex Allen on statistics in spying, assesses whether the Olympics deliver economic or sporting benefits, and calculates the odds of an uncanny Scrabble coincidence.

Main Topics: Zombie statistics in government waste claims (Priority: 5/5): The programme challenges ministers’ repeated use of a consultancy-based claim that local authorities could save 20% of procurement spending, arguing the figure was extrapolated from a tiny, unrepresentative sample and should not be cited as robust evidence. Misleading procurement comparisons in Cabinet Office laptop claims (Priority: 4/5): A Cabinet Office press release claimed departments paid vastly different sums for 'the same laptop'; fact-checking revealed the laptops were not comparable and that the press release overstated its case, though one secondary example was accurate. Interpreting riot arrest and reoffending statistics (Priority: 4/5): The episode tests whether Ken Clarke’s claim that nearly three-quarters of those charged over riots had previous convictions proves rioters were mostly repeat offenders, concluding the figure may reflect police selection and the known prevalence of prior convictions among convicted offenders. Correlation vs causation in alcohol policy (Priority: 3/5): A report linking off-licence density with hospital admissions for underage drinking is presented as a classic correlation-not-causation problem, with no causal mechanism demonstrated. Statistics and intelligence analysis (Priority: 5/5): Alex Allen explains how statistical reasoning helps intelligence work by managing uncertainty, framing probabilities, testing hypotheses, and countering groupthink through red-teaming and alternative-hypothesis analysis. Olympics: economics and legacy claims (Priority: 5/5): Experts argue the London Olympics were unlikely to generate a direct financial profit or major tourism gains, and that much of the regeneration would have happened anyway; promised sports participation gains also fell far short of targets. Scrabble coincidence and probability (Priority: 3/5): A listener’s repeated draw of the word 'janitor' is analysed probabilistically, showing the event is rare but not extraordinary once multiple games over decades are considered.

Key Arguments: The 20% local-government savings claim is a zombie statistic: it was derived from a consultancy leaflet and extrapolated from a tiny mobile-phone sample to all procurement. Government press releases can be factually shaky even when they contain one accurate example; fact-checking requires comparable items and transparent methods. A high share of riot defendants having prior convictions does not prove most rioters were known criminals, because arrests were often made later from CCTV and prior convictions are common among convicted offenders generally. Off-licence density and underage alcohol admissions are statistically associated, but the report does not establish that off-licences cause the admissions. In intelligence analysis, statistical training helps interpret uncertainty, assign probabilities carefully, test competing hypotheses, and avoid groupthink. Olympic operating costs can be covered by ticket and broadcast revenue, but the big infrastructure and legacy costs usually mean the Games are not a clear financial win. Evidence for tourism boosts from the Olympics is weak, and much of East London’s regeneration was already planned for other infrastructure projects. Legacy claims that the Olympics would significantly increase sports participation were not met; the target was later quietly scrapped. A repeated Scrabble draw that feels miraculous may still be unsurprising when many games are considered over a long period.

Data Points: Potential local authority savings: 20% - Claim cited by ministers as achievable on local authority third-party spending Local authority third-party spending: £50 billion - Base spending amount used in the waste-savings claim Implied annual saving: £10 billion - Approximate saving suggested by applying 20% to £50 billion Old VAT rate revenue comparison: almost as much as VAT to 20% - The alleged savings were compared rhetorically with tax changes Procurement sample mentioned in critique: £600,000 - Ben Goldacre said the consultancy extrapolated from a tiny sample of mobile-phone spending Cabinet Office laptop claim: £2,000 vs £350 - Press release claimed different departments paid this much for the same laptop London riot arrests initially reported: around 685 - Metropolitan Police arrests reported the day after three days of rioting London riot arrests later reported: more than 2,300 - Arrests in the capital reported about a month later Previous convictions among adults convicted in Crown Court: around 80% - Used to contextualize Ken Clarke’s claim about riot defendants Previous convictions among Magistrates Court convictions: around 93% - Higher-level prevalence for less serious offences One-million-participants target: 1,000,000 - Olympic legacy sport-participation goal set by the previous government Required annual participation increase: 200,000 per year - Implied pace needed to meet the one-million target over five years Progress halfway to target: 109,000 - Active People Survey figure cited as halfway through the period Olympic London infrastructure cost: £9.3 billion - Commonly cited cost of building the Olympic site and hosting infrastructure Probability of drawing 'janitor': about 1 in 20,000 - Estimated chance of getting the seven-letter Scrabble word on a single draw Probability of drawing 'janitor' twice in a row: about 1 in 500 million - Simple back-to-back probability for two independent draws Probability of drawing 'janitor' twice in 1,000 games: about 1 in 400 - Adjusted for the listener having played many games over decades

Pivotal Quotes: "It's possibly one of the most ridiculous figures I've ever seen quoted in Parliament." — Ben Goldacre: Critiquing the government’s 20% local-authority savings claim "The 20% figure comes not from a department of competent statisticians in the civil service ... It comes from this incredibly flimsy advertising leaflet from a private management consultancy firm touting for business." — Ben Goldacre: Explaining why the procurement-savings claim is not credible "The press release said, and I quote, the move will end poor value for money contracts such as those where government departments and agencies paid between £350 and £2,000 for the same laptop." — Tim Harford: Highlighting the Cabinet Office’s overstated comparison

Implications: Listeners should treat headline statistics, especially in politics and media, as claims to verify rather than facts to accept. The episode underscores the value of methodology, context, and probability in judging policy, security, and economic promises.