School re-opening, Germany’s Covid-19 success and statistically savvy parrots

Episode Summary

Executive Summary: The episode examines COVID-19 risk communication, arguing that children and teachers face very low direct health risks from reopening primary schools, while broader family risks are also relatively limited. It then explains why Germany’s death toll was far lower than the UK’s, critiques misleading UK testing statistics, outlines how pandemic borrowing is financed, and ends with research showing New Zealand kea can reason about probability and combine cues intelligently.

Main Topics: School reopening and COVID risk to children (Priority: 5/5): Professor David Spiegelhalter argues that the direct risk to primary-age children is extremely low, especially compared with everyday risks, though rare inflammatory complications remain under study. Risk to teachers and families (Priority: 4/5): The discussion expands to school staff and households, noting that teachers’ risk appears similar to other professionals and that healthy younger adults faced measurable but still relatively small mortality risk. Why Germany’s COVID death toll was lower than the UK’s (Priority: 5/5): Kavita Puri explains that Germany’s early testing, contact tracing, quarantine, and earlier effective lockdown timing likely explain most of the difference, rather than mysterious system advantages. Government testing statistics and the 100,000 target (Priority: 5/5): The programme dissects how the UK government inflated testing figures by counting posted kits on dispatch and including research tests not used for diagnosis, making the headline target misleading. How pandemic borrowing is financed (Priority: 4/5): Bloomberg economist Jamie Rush explains that large government borrowing is being supported by central-bank money creation and bond buying, delaying but not eliminating eventual repayment choices. Parrots and probabilistic reasoning (Priority: 3/5): Researchers show that New Zealand kea can reason about probability, integrate barriers and preferences, and use multiple information sources in a way suggesting domain-general intelligence.

Key Arguments: Children’s direct COVID mortality risk is extremely low; reported deaths among 5-14 year olds in England and Wales were exceptionally rare relative to the size of the population. Reopening schools is not presented as an extraordinary risk to children, though policymakers must still account for rare complications like Kawasaki-like inflammatory syndromes. Teachers and educational staff do not appear to have markedly higher occupational risk than other professionals, based on ONS occupational analysis. Germany’s lower death count is largely explained by earlier widespread testing, quarantine, and lockdown at an earlier stage of the epidemic curve. A one-week-earlier UK lockdown could have reduced deaths dramatically, potentially to around German levels, according to modelling cited on the programme. UK testing statistics were misleading because they mixed diagnostic tests with posted kits and research samples, inflating the count beyond actual completed diagnostic tests. Counting multiple tests on one person can be clinically justified in some cases, but it should not be used to claim more people have been tested than actually have been. Pandemic borrowing is being enabled by central banks buying government debt with newly created money while inflation remains low, but this is only a temporary financing mechanism. Kea parrots demonstrated not just pattern recognition but flexible integration of probability, physical constraints, and experimenter behavior, indicating broad cognitive abilities.

Data Points: Children aged 5-14 in England and Wales: 7 million - Population base used to assess COVID death risk among school-age children COVID-marked deaths among 5-14 year olds: 1 death - Reported on death certificates so far, cited as extremely low risk Annual deaths among 5-14 year olds: about 600 - General background mortality for perspective Road deaths among 5-14 year olds in England and Wales: about 50 per year - Used to compare everyday risks with school reopening risk COVID deaths in healthy 20-40 year olds in English hospitals: about 30 deaths out of 30,000 COVID deaths - Healthy young parents without pre-existing conditions Estimated population of healthy 20-40 year olds: 13-14 million - Denominator used for mortality-risk estimate Mortality risk for healthy 20-40 year olds: about 3 in a million - Approximate chance of death at peak epidemic Relative risk difference by age: 10,000-fold - Comparison between over-90s and under-15s over a five-week period Over-90 COVID mortality over five weeks: more than 1% - Illustrates very high risk for the oldest group Under-15 COVID mortality over the same period: about 1 in 7 million - Illustrates extremely low risk for school-age children Germany reported COVID deaths: around 8,000 - Official figure compared with the UK UK reported COVID deaths: around 35,000 - Official figure compared with Germany German testing capacity by late March: 50,000 tests a day - Germany was already testing at scale before UK lockdown UK testing capacity by late March: less than that each week - Contrast with Germany’s early testing effort Germany reported deaths at lockdown: 86 deaths - Germany locked down earlier in the epidemic curve UK reported deaths at lockdown: 359 deaths - UK locked down later in virus time Modelled effect of a UK lockdown one week earlier: about a quarter of reported COVID deaths - Cited from James Liyanage modelling Government claimed tests on 15 May: 136,486 - Headline figure including posted kits and research tests Diagnostic tests actually carried out on 15 May: 69,900 - Figure after removing posted kits and research tests People tested on 15 May: 43,298 - Shows tests exceed people because some individuals are tested more than once UK government testing target: 100,000 tests a day - Target claimed to have been reached multiple times Government’s later testing target: 200,000 tests a day by end of May - Announced as an expanded goal UK extra borrowing in 2020 estimate: £220 billion - Bloomberg in-house estimate due to pandemic Pre-pandemic expected borrowing: about £55 billion - Baseline before coronavirus struck Cost of announced policies within extra borrowing: about $140 billion - As stated in transcript; likely intended as about £140 billion in context Tax shortfall due to recession: about $80 billion - As stated in transcript; likely intended as about £80 billion in context Kea cognitive test performance: passed probability, barrier, and preference tasks - Evidence of domain-general reasoning abilities

Pivotal Quotes: "There have been, based on the data so far, extremely low risks to children." — Professor Sir David Spiegelhalter: Assessing the health risk of reopening primary schools "If the UK had gone into lockdown a week earlier, we would have seen only a quarter of the reported COVID deaths that we've had, about the same as in Germany." — Tim Harford citing modelling by James Liyanage: Explaining the UK-Germany death toll gap "A swab in the post is not the same as a swab in the nostril." — Kate Lambell: Critiquing the government’s inflated testing statistics

Implications: Listeners are urged to judge COVID policies by real risk and transparent metrics, not headline claims. The episode shows how timing, testing, and reporting shape outcomes and public trust, while also highlighting surprising animal cognition research.