Debate, Reform, tax evasion and ants

Episode Summary

Executive Summary: This episode fact-checks election and science claims: debate soundbites on taxes, NHS waiting lists and small boats; Reform UK's draft 'contract' claims on excess deaths and net zero; the realism of parties' £6bn anti-avoidance plans; poll reliability with 1,000+ samples; and a Springwatch ant myth. The show repeatedly finds that timing, assumptions and methodology matter more than headline numbers.

Main Topics: Election debate fact-checking (Priority: 5/5): The Tim Harford/Kate Lamble segment dissects Sunak and Starmer's claims on tax rises, NHS waiting lists and migration, showing how each side uses selective time periods and speculative assumptions to frame the debate. Reform UK manifesto claims (Priority: 5/5): The programme scrutinizes Reform's 'contract with the people,' challenging assertions about excess deaths, vaccine harms and net zero using expert testimony and current evidence. Tax avoidance and evasion funding claims (Priority: 5/5): Reporter Nathan Gower explains why both Labour and Conservatives' promised £6bn yearly gains from clamping down on tax avoidance/evasion are plausible but uncertain and not guaranteed by the NAO. Polling sample size and errors (Priority: 4/5): A listener question leads to a discussion of why around 1,000 respondents can be enough for a national poll, and why past polling failures were often due to sample quality and turnout/model assumptions rather than sample size alone. The ants vs humans pub fact (Priority: 3/5): The show revisits and updates the claim that ants weigh as much as humans, concluding humans still outweigh ants, though new research raises the estimated ant population and leaves some uncertainty. Evidence and interpretation in public claims (Priority: 4/5): Across the episode, the recurring theme is that raw numbers can be misleading unless their source, timeframe, assumptions and definitions are clearly understood.

Key Arguments: A £2,000 tax-rise claim can be reverse-engineered from a four-year estimate and divided across households, but annualized it is closer to £500 per household per year and depends on many speculative assumptions. Starmer's claims about Conservative plans to abolish national insurance and inheritance tax refer to ambitions or rhetoric, not announced policies. Sunak's NHS waiting-list claim is complicated: the list fell from its peak, but it remains higher than when he made the promise, and strikes only explain a small part of the increase. On small boats, Sunak is correct year-on-year (2023 lower than 2022), while Starmer is correct for 2024-to-date, which is the highest ever at that point in the year. Reform's excess-deaths claim is inaccurate: current population death rates are near best-ever levels overall, and evidence does not support a vaccine-harms link causing appreciable excess deaths. Reform's net-zero language misrepresents climate science; reaching net zero would stabilize warming, and future sea-level rise would be lower than under continued fossil-fuel emissions. Both parties' £6bn tax-enforcement promises are based on the tax gap estimate, but the NAO says recouping revenue requires prioritization, sustained focus and further investment—not a guaranteed windfall. A poll of about 1,100 people is enough for roughly a ±3% margin of error nationally; larger samples reduce sampling error but cost more, and real polling errors often come from non-random samples and weighting/model choices. The ants fact remains false in the simple form stated: even with a revised estimate of 20 quadrillion ants, humans still outweigh ants overall, though uncertainty remains due to incomplete data and ant mass assumptions.

Data Points: Claimed tax rises from Labour: £38 billion over four years - Conservative framing used to produce the repeated £2,000 per working family figure Implied annual tax rise per household: about £500 per household per year - Annualized interpretation of the Conservative £2,000 claim Households used in calculation: 18 million working households - Denominator used by Conservatives for the tax-rise claim Labour green prosperity plan: largest spending item - Conservatives' estimate relied heavily on this being tax-funded, though Labour said much would be borrowed Conservative national insurance abolition: £46 billion a year - Labour's rebuttal of a non-policy described by Starmer as a Conservative plan Inheritance tax abolition: £10 billion a year - Another Labour rebuttal of a non-announced Conservative policy NHS waiting list at Sunak pledge: 7.21 million - Treatment pathways in January 2023 when Sunak promised to cut waiting lists NHS waiting list peak: 7.77 million - September 2023 level after Sunak's pledge Current NHS waiting list: 7.54 million - Latest figure cited in the episode Waiting-list growth from strikes: just over 200,000 referrals - Health Foundation estimate of the impact of NHS strikes Share of current waiting list added by strikes: less than 3% - Contextualizing the strike impact Small boat crossings in 2022: almost 46,000 - Used to assess year-on-year change Small boat crossings in 2023: over 29,000 - About a third lower than 2022 Small boat crossings in 2024 to date: more than 10,000 - Highest ever at this point in the year Reform excess deaths claim: nearly as high as during the Covid pandemic - Claim challenged by actuary Stuart MacDonald Current mortality comparison: similar to the best years on record (2019 and 2014) - MacDonald's assessment of overall population death rates Working-age mortality: still much lower than during the pandemic - MacDonald's caveat about younger groups Ant estimate old paper: 10 quadrillion - Earlier academic estimate cited in the ant fact Ant estimate new paper: 20 quadrillion - Updated estimate from a newer global data set Ant-to-human weight ratio: humans weigh about 5 times as much - Latest estimate after revising ant numbers and weight assumptions Earlier ant-to-human weight ratio: humans outweigh ants by about 8 times - Previous correction to the pub fact Poll sample size for ±3% error: about 1,100 people - Statistical target discussed by Paula Surridge Poll sample size for 2.1% error: 2,000 people - Larger sample reduces sampling error, but at higher cost Mega-poll sample size: 10,000 people - Sampling error falls to about 1% in very large polls Conservative tax-gap reduction: something like £5 billion in today's terms - Historical precedent cited for tax enforcement gains Potential tax collected if all behaved perfectly: over £30 billion - HMRC-type tax gap discussion Failure to take reasonable care: £11 billion - Category larger than tax avoidance and evasion losses Tax avoidance and evasion gap: £6 billion a year - Estimate both parties cite when promising extra revenue

Pivotal Quotes: ""The waiting lists are coming down."" — Rishi Sunak: Debate claim about NHS performance, examined against longer-term waiting-list data ""In a word, no."" — Stuart MacDonald: Answer to whether excess deaths are nearly as high as during the Covid pandemic ""There's not just a nice anti-evasion, anti-avoidance button that you press and then billions of pounds starts flowing into the Treasury."" — Helen Miller: Explanation of why promised tax-enforcement revenue is uncertain

Implications: Listeners should treat campaign numbers as estimates shaped by assumptions, timeframes and definitions. The episode shows why independent verification matters and why headline claims on taxes, health and climate often oversimplify reality.