Assessing NZ's Covid lockdowns
Did the scientific community overwhelmingly support lockdowns?
This is the third of a series reflecting on 3 years of the Covid pandemic. Please feel free to submit your own reflections to Asymmetric Information. You could draft a post (email the editor), or write a comment on this post.
I became aware of Covid-19 at some point in December 2019. The first significant piece of information I noticed concerned the Diamond Princess, a cruise ship, which was quarantined off Japan in early February 2020 after cases were discovered. Covid spread rapidly through the ship, which was carrying 2,700 passengers (mostly old) and 1,000 crew (mostly healthy young adults). All crew and passengers were tested: 700 tested positive (mostly asymptomatic), and 9 to 14 died (all but one aged 70+). On this basis, I thought Covid was going to be a moderate problem for the 70+ group (both in terms of the death rate and the low residual life expectancy of these people), and was unconcerned for myself — despite being close to the relevant age group.
Further information revealed that the Infection Fatality Rate (the probability of dying if infected) was strongly related to the presence of serious pre-existing medical conditions (as well as age). Thus, a healthy 70-year old had even less risk of dying of Covid than the Diamond Princess data suggested, even if they caught it (which was far from certain). I was therefore very surprised at the way Covid has been portrayed in the mainstream media. In particular, that media advanced the following principal ideas:
Covid was the greatest public-health crisis in living memory.
Scientific experts in NZ (and elsewhere) were overwhelmingly of the view that lockdowns were warranted, in order to avoid an extraordinarily high death toll.
Some economists took a different view, on the basis that saving lives was costly and the costs (in the form of GDP losses) exceeded the monetary value of the additional lives saved by lockdowns.
The view that one could put a monetary value on life was highly distasteful, and it was alien to governments and their scientific advisers.
The NZ government (and typically those elsewhere) accepted the views of scientists (“following the science”), and therefore locked down.
The only one of these points that is clearly true is 3. The other points range from highly disputable to clearly wrong.
Obesity is a bigger health crisis than Covid
Consider point 1. It is standard practice amongst public health experts to assess the health consequences of any event in terms of the Quality Adjusted Life Years (QALYs) that are lost. The government’s principal external advisers in March 2020 were Professors Baker and Wilson of Otago University, who contemporaneously estimated the death toll from Covid at 10,000 (if mitigating measures short of a general lockdown were adopted), and also estimated the residual life expectancy of the victims sans Covid at five years.1 The expected loss in Life Years was then 10,000*5 = 50,000. These potential victims were generally in poor health, and a conservative discount on their quality of life (of 20%) leads to a loss in Quality Adjusted Life Years (QALYs) of 50,000*0.8 = 40,000.
This is a large number but it is dwarfed by that associated with another public health crisis: obesity. About a third of adults in NZ are obese, which is about 1.2 million people.2 Obesity shortens life expectancy by about 8 years, and reduces the quality of life before (premature) death.3 So, for an average aged obese person (of say age 40), obesity reduces residual life expectancy from a further (say) 40 years to 32 years. Using a 20% reduction in the quality of life until (premature) death, the loss in QALYs per person would then be 8+32*0.2 = 14.4 QALYs. With 1.2 million such people, the aggregate QALY loss would be 17 million QALYs. This is over 400 times that for Covid. This bears repeating. Obesity is over 400 times more serious than Covid was expected to be.
With the benefit of hindsight, an upper bound on the deaths from a mitigation rather than a lockdown strategy would be Sweden, which had the highest Covid death rate amongst European countries that did not lockdown (2,200 per million).4 Applying this rate to NZ’s population of 5 million implies 11,000 deaths, which is close to Baker and Wilson’s forecast of 10,000. So, whether using contemporaneous forecasts or ex-post estimates, obesity is over 400 times more serious than Covid.
Obesity isn't the only public health crisis that dwarfs covid. So do recreational drugs, of which the worst seems to be alcohol. There are about 800,000 NZrs who are defined as "hazardous drinkers".5 This causes accidents leading to injuries, highly aggressive behaviour, relationship problems, financial problems, worsening of mental health problems, and poor decisions in various areas such as crime. It also reduces life expectancy by several years.6 For an average aged hazardous drinker (of say 40 years), suppose this reduces their residual life expectancy by 3 years (from 40 to 37) and reduces their quality of life in their remaining 37 years by 10%. The loss in QALYs for each such person would then be 37*0.1 + 3 = 6.7. With 800,000 such people, the aggregate loss in QALYs would be 800,000*6.7 = 5.3 million QALYs. Again, this dwarfs covid, by a factor of 130.
My Memorandum trumps your Declaration?
Turning to point 2, the pandemic induced the drafting (in October 2020) of the Great Barrington Declaration (which opposed general lockdowns in favour of protecting only vulnerable groups whilst allowing transmission in the rest of the population, leading to development of herd immunity). In turn, this prompted the drafting of the John Snow Memorandum (which favoured general lockdowns). Both declarations were drafted by scientists, and invited signatures. In respect of the signatories who were scientists, Prof John Ioannides of Stanford University assessed their quality (based upon citation figures for their publications), and concluded that both documents were signed by a “massive number of stellar scientists”.7 This rebuts the suggestion that scientists were overwhelmingly in favour of lockdowns.
Monetary valuation of lives is standard public-sector practice
Turning to point 4, it is standard practice in NZ and elsewhere for government bodies to value lives or QALYs for the purpose of determining which public health interventions are warranted. For example, the government-funded BODE³ project assesses the cost effectiveness of health interventions in NZ, and those managing the project (public health academics including some of those advising the government on Covid) value a QALY at $45,000 for the purpose of determining which interventions should be undertaken, i.e., which lives should be saved.8
The Treasury promotes a Cost-Benefit-Analysis model for use in the wider public sector (CBAx2022), which includes a central QALY value of $36,000.9 Again, this determines which projects should be undertaken, and therefore which lives should be saved. In addition, the Ministry of Transport values a life at $4.88m, and uses this figure to assist in its decisions on which road safety projects to undertake.10 In short, putting a monetary value on life for the purpose of deciding which lives to save is standard practice by NZ government entities. As it is for government entities elsewhere.
Lockdown action before evidence
Turning to point 5, the actual factors that induced the NZ government’s initial lockdown decision are not clear, but there was either no advice or no consistent advice to do so from its external public health advisers. These external advisers were primarily Professors Wilson and Baker of Otago University, up to late March 2020, and thereafter a group of science academics headed by Professor Hendy of Auckland University. In respect of Professors Wilson and Baker, they recommended lockdown in a paper dated 19 March 2020:
“New Zealand should consider a short pulse (a few weeks) of intense social distancing, including bringing forward the school holidays and temporary closures of most businesses, social meeting places and public transport.” 11
They reiterated this in a report one day later for the Ministry of Health:
“For Covid-19, we suspect that elimination can probably be achieved, based on the experience in China. An elimination goal (which we strongly recommend) would require use of more intensive and disruptive interventions early in the pandemic to interrupt transmission. … All of these interventions have some costs…”.12
No cost-benefit analysis was provided in support of either of these recommendations. However, in a paper dated three days later (23 March 2020) and cited in footnote 1 below, they estimated deaths at 6,500–13,000 with a mitigation strategy (involving isolation of the over 60s), and 500 from a successful eradication strategy. Upon converting these deaths to monetary values, they concluded that
“This implies that…society might be prepared to spend up to $650 million to $2.48 billion to ‘give eradication a go’. These are quick calculations. The government needs to instruct the Ministry of Health and Treasury to check and update our analysis.”
This was a clear recommendation to the government to estimate these costs and pursue an eradication strategy only if these costs were below their threshold figure. There is no evidence that the government did estimate costs before locking down, and Baker and Wilson’s advice to do so contradicts their earlier unconditional support for lockdown in pursuit of elimination.
Interestingly, three months later, a paper, whose authors included Professor Wilson, conducted a cost-benefit analysis on mitigation (as in Sweden) versus suppression (as in Australia) and elimination (as in NZ). They concluded that, from a societal perspective (which involves taking account of deaths, morbidity, health costs and GDP losses), mitigation was favoured for a value per QALY of up to US$240,000.13 So, using Baker and Wilson’s preferred value per QALY of NZ$45,000 (US$30,000 at the time), mitigation was strongly favoured in NZ. The best that can be said of this is that Baker and Wilson did not have a consistent view on lockdowns.
Turning to the team of scientists headed by Professor Hendy, and in a report dated 25 March 2020, Table 2 of the paper presented predictions of the death tolls in NZ from a range of possible control strategies.14 However, none of the control strategies examined involved lockdown (which included closing down all but essential businesses). The only places of work that are closed down in any of the control states in their Table 2 are schools and universities.
Thus, in choosing to lockdown in late March 2020, involving the closure of all non-essential businesses, the government went well beyond the contemporaneous advice from Professor Hendy’s team, and also beyond some of the contemporaneous advice offered by Professors Baker and Wilson. The lockdown decision was therefore presumably triggered by some mix of observing the behaviour of other governments, a belief that the majority of New Zealanders favoured doing so, and fear.
Bad news sells
How is it then that the narrative in the mainstream media seemed to be so much at variance with the facts? Most journalists might have been unaware of how much more serious the obesity problem was. They might also have concluded from reading the work of other journalists (who chose to focus on quoting scientific opinions in favour of lockdowns) that this reflected the mainstream view amongst scientists, leading them to also focus upon pro-lockdown scientists, and each acting in that way had a self-reinforcing effect on the group. They might also not have been aware of papers revealing that placing a monetary value on lives was standard practice in the health and transport sectors. They might also have failed to notice that none of the control strategies examined by Professor Hendy’s team included lockdown (closing down all but essential businesses), and been unaware of the papers by Professors Baker and Wilson in which lockdown was conditional on a cost-benefit analysis which, when done, concluded against it.
A more cynical explanation may also have contributed: journalists want to be read, bad news sells more than good news, the worst news is that the pandemic was so severe that only lockdowns could avert a huge death toll, and this was supported by portraying the scientific community as overwhelmingly in support of lockdowns.
By Martin Lally
See Blakely, T., Baker, M., & N. Wilson., (2020). The Maths and Ethics of Minimising Covid-19 Deaths in NZ, Public Health Expert 23/3/2020.
See Ioannidis, J.P. (2022) Citation impact and social media visibility of Great Barrington and John Snow signatories for COVID-19 strategy. BMJ Open.
For example, amongst the papers on the Presentations of the Burden of Disease Epidemiology, Equity & Cost-Effectiveness Programme (BODE³) website, the following paper invokes this figure for valuing a QALY: Wilson, N., Grout, L., Summers, J., Mizdrak, A., Nghiem, N., & Cleghorn, C., (2021). A League Table of Cancer Control Interventions that Spans Primary Prevention to Palliation.
See pages 4 and 7 of Social Cost of Road Crashes and Injuries June 2021 update (transport.govt.nz).
Baker, M., & Wilson, N., (2020), Why New Zealand needs to continue Decisive Action to Contain Coronavirus, 19 March, .
See page 12 of Baker, M., Kvalsig, A., Barnard, L., Gray, L., Verrall, A., & Wilson, N., Potential Preventative Interventions for the “Stamp it out” and “Manage it” Phase of the Covid-19 Pandemic in New Zealand, prepared for the Ministry of Health, 20 March 2020.
See Blakely, T., Bablani, L., Carvalho, N., Andersen, P., Abraham, P., Katar, A., Boujaoude, M., Akpan, E., & Wilson, N., (2020), Integrated Quantification of the Health and Economic Impacts of Differing Strategies to Control the COVID-19 Pandemic, 17 June. The authors used the acronym HALY (Health Adjusted Life Year) rather than QALY, but the terms are synonymous.
See James, A., Hendy, S., Plank, M., & Steyn, N., (2020). Suppression and Mitigation Strategies for Control of Covid-19 in New Zealand, working paper.
 | A guest post by
|
Loss of life is one cost. So is reduced quality of life while ill, and over longer periods for those drawing that unlucky straw. Looking more and more like Covid does long-term harm to the immune system, increasing the cost of other illnesses. And also worth noting that places that never undertook NZ's lockdowns had their own impediments to economic activity when illness rates were very high.
And there remains a categorical difference between protecting people from non-communicable illness, where no real externality argument applies (recall that fiscal externalities are largely pecuniary rather than technological), and infectious disease.