This is my longest ever post by about a factor of four. I have, until now, respected a self-imposed word limit of 2000. I promise that this post will be the only one I make which breaks that rule, but I could not shoehorn all I wanted to say into my normal size post. My apologies to any who find it indigestible. After getting this off my chest, I will avoid the subject for a while.
Towards the end of 2019 a respiratory virus epidemic began in Wuhan and started to spread to other countries. Early reports from China indicated that the virus was causing human deaths on a large scale. It appeared to spread rapidly from one person to another.
Western countries initially responded with measures recommended in their existing flu pandemic plans, including the promotion of hand washing and the use of PPE in hospitals dealing with Covid patients (although many countries immediately found themselves short of PPE). Otherwise, no substantial public policy responses were mounted, and there was, initially, no attempt to restrict, test, or quarantine arrivals from China, Northern Italy, or other infection hotspots. Western countries mostly took no substantial action until late March 2020. Far Eastern countries, having been through the SARS epidemic some years ago, responded much more rapidly (starting in January 2020) with incoming border restrictions and a test, trace and isolate system (TTI) focused on locating and quarantining those infected.
Western epidemiological modellers began to build scenarios in which they assumed that there was no pre-existing immunity in any human population, and that people could become infected, shedding new virus particles, and infecting others, without themselves having any symptoms at any time (ie. asymptomatic, not pre-symptomatic). It was further assumed that asymptomatic transmission was a major driver of the spread of infection. It was assumed the virus could be transmitted via a contaminated surface. Ferguson’s paper of March 2020 stated the assumption that Covid 19 had “comparable lethality to H1N1 influenza in 1918” – Spanish flu, and that 81% of the population would be infected. Some surprisingly crude arithmetic (81% of 65m people times a 1% infection fatality rate) led to an expected total of 510,000 deaths in the UK, resulting from an unmitigated epidemic. The models were (and still are) built on an elapsed time basis from the index case, disregarding the seasonality effects typical of respiratory infections. Models built on the basis of these, and other, behavioural assumptions, also suggested that social restrictions (lockdowns, face coverings etc) would be highly effective in limiting the spread until vaccines could be developed; Ferguson’s model suggested that the combined effect of all the NPI measures he considered, could reduce the total deaths from 510,000 to as low as 8,700.
Largely as a result of such models, a year ago, in late March 2020, the UK and much of the world began an experiment, using mandatory business and school closures and whole-population stay-at-home orders (“lockdowns”) in an attempt to control the Covid epidemic. These measures were taken alongside many other precautions and interventions, all of which overlapped in uncontrolled ways, forming a chaotic experiment at the end of which, we will have little definite proof of which specific interventions helped, which harmed, and which just did nothing. We have to acknowledge that, despite the natural desire for certainty, we will get no firm, cut-and-dried answers to these questions.
The use of lockdowns on such a scale had no precedent and could therefore only be assumed to be effective. The argument given at the time in favour of the policy was that the empirically untested epidemiological models (especially the Imperial College/Ferguson model) predicted an unacceptably high death rate from an unmitigated Covid epidemic, but also a very much lower death rate if whole-population social controls were mandated. No attempt was made, however crude, to carry out a cost-benefit analysis of this policy, partly because the lockdown was initially supposed to be brief – a few weeks, a period which would do limited economic and human damage. Johnson initially talked about “three weeks to flatten the curve” and “we can send the virus packing in 12 weeks”. The adoption of such policy was therefore driven by the twin assumptions of a very high unmitigated death rate, and of highly effective NPI policies of short duration.
This pro-lockdown argument was widely accepted in the UK and has been the official narrative ever since. Almost all public discussion starts from the assumption that lockdowns work (however that might be defined). It has reached a point at which those who question this assumption are treated as a danger to public health and their arguments are not engaged with. Dissent has been equated with sedition. The opposition has gone along with the official narrative and has provided no push-back to the policy responses.
However, it’s been a year, and there are now real-world outcomes available to assess whether the models were realistic, and what effects the lockdowns had. We can now revisit and test the original assumptions to see if they were correct.
How bad is Covid-19?
This question is important because it has long been acknowledged that the restriction of basic human rights is allowable on public health grounds if it is “necessary and proportionate”. However, the law is weak in that it does not state when and in what forum proposed restrictions should be held up to those tests, nor the legal standard which a case for restrictions should meet. The burden of proof is not imposed. By default, this gives governments an unrestrained power.
The coronavirus is indeed deadly, but the question is, how deadly, and deadly to which people? Early estimates of fatality rates were (as usual in most epidemics) far too high. Public ignorance of epidemiology was understandable, but professional ignorance inexcusable: the WHO issued a tweet on 3rd March 2020, stating “Globally, about 3.4% of reported #Covid19 cases have died. By comparison, seasonal flu generally kills far fewer than 1% of those infected. – Dr Tedros”. These two sentences are both (approximately) true, but the linking words (“By comparison”) equate the case fatality ratio (the higher number) with the infection fatality ratio for flu. It is therefore a highly misleading statement, relying on public ignorance of the huge difference between IFR and CFR to make an apples-to-oranges comparison. In line with this statement, the definitions of established medical terms such as “case” have been misused throughout the last year, and the long established principles of cause-of-death attribution overturned.
It became clear at an early stage that the severity of Covid-19 is strongly age-dependent. The US CDC currently estimates the IFR to be 0.002% for under 18s, and 0.05% for 18 – 50s (or, about 0.02% for all under-50s together). The small number of deaths represented by these figures are very largely those who had other, pre-existing, serious conditions. It is therefore questionable whether it is truthful to call Covid a deadly disease for the ¾ of humanity who are under 50 and in average condition for their age. This pattern of lethality is very different to that of the 1918 Spanish flu, which killed millions of under-50s, but Covid was painted simply, and in black-and-white terms, as deadly for everyone (Spanish flu is variously estimated to have caused between 100,000 and 200,000 excess deaths in the UK).
The IFR is much higher for older people: 0.6% for 50 – 64 year olds, rising as high as several percent for over 70s. The median age at death of those dying of Covid has been 82. So it is definitely a deadly disease for those age groups.
It turns out that Covid-19 presents a high risk of death to a few – chiefly the very old and those with other medical conditions – and a low risk of death to the large majority, those who are of school and working age.
To answer the question, how bad is Covid, we need some context. According to the ONS https://www.ons.gov.uk/peoplepopulationandcommunity/birthsdeathsandmarriages/deaths/bulletins/excesswintermortalityinenglandandwales/2017to2018provisionaland2016to2017final , there were 50,100 excess deaths during the 2017/18 winter flu season (not all necessarily attributable to flu). About 78,000 died in England and Wales in the 1951 flu season (out of a population of 44m) https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3294686/ . Such peak death rates have been seen eight times in the last 40 years, without much comment or any public policy response. We are used to it and don’t worry about it; there is a flu vaccination program, and that’s all we do. So, in the UK, Covid has been between twice and three times worse than a bad flu year. But, Britain has the world’s fourth worst Covid death rate. In most other Western countries, Covid deaths were roughly twice as high as a bad flu year. This is still a lot of deaths, but we have to see the level of the threat in that context.
We accept 78,000 deaths a year from smoking, something which is entirely avoidable. The public policy response is limited to measures to discourage smoking, and heavy taxation. We live with this level of death, and people remain free to decide for themselves whether to take this risk. And, that is 78,000 deaths every year – while the Covid epidemic is more of a one-off event. In 2019, there were 36,800 deaths from alcoholic liver disease, and again, the response is limited to advice, and discouragement by taxation. We have closed all the pubs and bars in the country to prevent Covid, but have never done so to prevent the deaths and other harms directly caused by alcohol. Similarly, air pollution is variously estimated to cause 40,000 premature deaths per year, and again, that’s every year, not just one epidemic year, yet we do disgracefully little about it, given what is in our power to do.
In a normal year, about 166,000 people die of cancer in the UK. Many of these deaths are unavoidable, but the ONS estimates that cancer accounts for about a third of deaths from avoidable causes (those which are treatable or preventable), which are typically over 130,000 per year https://www.ons.gov.uk/peoplepopulationandcommunity/healthandsocialcare/causesofdeath/datasets/avoidablemortalityintheuk . We respond to this with targeted measures such as screening programs (mammography, colonoscopy, smear tests etc) but otherwise no more than general lifestyle and diet advice.
At the peak of the second wave, the week of January 15th, a total of 18,676 people died in England and Wales (of all causes). This was less than the 20,566 who died in the week ending 7th January 2000, a bad flu year; or the 20,116 who died in the week ending 8th January 1999, also a bad flu year. In neither of those years was there any suggestion that we should close small businesses and all stay at home.
So, on the scale of things, Covid-19 has been a worse-than-average respiratory pathogen, but hardly the Black Death. The initial estimates of fatality ratios were an order of magnitude too high, leading to threat-inflation and panic responses. Its impact is severe on the old and the vulnerable; the vast resources poured out over the last year would, if focused on the protection of those at risk, have been more than enough to mitigate the death toll, while those not at any significant risk could have been allowed to make their own decisions, as usual, about their behaviour – as they were in some countries which have had similar or better public health outcomes.
We accept, through familiarity, various causes of similar numbers of deaths. We choose to take avoidable risks because of a common acceptance that life is not, and should not be, risk-free. But when Covid-19 appeared, and amid initial uncertainty as to how deadly this disease is, there was a period of inaction (during which relatively low-impact but effective measures could have been taken), followed by a sudden panic, at which point the bar was set far higher, and it remains higher today. Governments assumed the authority to take away individual choice and make one-size-fits-all decisions about human social behaviour. We have to ask, why is the Covid death toll unconscionable when the deaths from other causes in other years were not? Any why are the deaths caused by lockdown ignored? There has been an inexcusable indifference to the harm done by lockdowns, to the point of a failure even to acknowledge they exist. It has become a failure so great, like the Iraq war for the Blair administration, that the government cannot now ever acknowledge that it was a mistake.
Were the assumptions made correct?
Many have been infected with this virus, but few people turned out to be susceptible to severe disease. There does appear to have been a significant, non-zero level of pre-existing immunity; this is now widely accepted in the academic literature. The idea that 81% of the population would be infected has been quietly dropped down the memory hole.
The assumption about asymptomatic spread has also been falsified. The WHO (always a little late to the party) finally acknowledged in June 2020 that “asymptomatic spread of coronavirus is very rare”. How the assumption arose in the first place is not clear, given that previous medical experience showed that “in all the history of respiratory viruses of any type, asymptomatic transmission has never been the driver of outbreaks. The driver is always a symptomatic person” (A Fauci, 28 Jan 2020).
The virus has also proved to be seasonal, with infection and illness levels reducing during the warmer months, just like other respiratory viruses. The assumption about surface contamination was completely wrong; no cases have been recorded of this kind of transmission. Nor is the virus easily transmitted outdoors.
A number of effective treatments have been found which substantially reduce the severity of the disease and therefore, the case fatality ratio. Budesonide and Dexamethasone are effective in reducing the severity of illness, and vitamin D is an effective preventative. Asymptomatic people (ie. people who in any other context are not ill) are not major drivers of the spread of disease. Lockdowns have not worked and have caused great harm. In short, all the main assumptions made in the models used in February and March 2020 were incorrect by a wide margin. Yet, most Western countries are still continuing with lockdown policies, or similar restrictions, to some degree. The incorrect assumptions made a year ago, still have us restricted and impoverished. There has been no public process to monitor the improving state of knowledge and recalibrate our public policy responses accordingly.
The effects of Public Policy responses
In the UK, certain policy measures were taken which have proved to be either ineffective or counter-productive. In the case of stay-at-home lockdowns, with tight restrictions on going outdoors, and all outdoor sports and activities prohibited, this may have made the epidemic worse. It is now clear that it is extremely difficult to catch Covid outdoors; negligibly few such cases have been identified. On the other hand, mandating people to stay indoors puts them in an environment more conducive to infection. People who are active outdoors gain general physical and mental health benefits, including higher vitamin D levels, which is effective in preventing infection. Restrictions on outdoor activity have therefore been at best ineffective, and at worst, actively counter-productive.
In the first wave of the epidemic, some 40% of deaths were in care homes, and a similar proportion of hospital deaths were of patients who did not have Covid on admission – hospital-acquired infections. Policy responses, designed to “protect the NHS”, included discharging Covid patients from hospital directly into care homes, while they were still infectious – this was a major cause of the terrible death rate in the care home sector. This practice has stopped now, and we have learned other lessons about protective equipment and hygiene measures, which, coupled with the vaccination of residents and staff, mean that any future Covid death rate in the care homes will be far lower (there is also the “dry tinder” effect – many of the most vulnerable are now dead). However, the conclusion is that while taking ineffective measures such as prohibiting the young and healthy from going outdoors, the government was also failing to take effective measures in the places where it mattered most. The fact that these stable doors are now mostly closed does not change the fact that tens of thousands of deaths occurred last year as a result of public policy errors.
Do Lockdowns Work?
Asking this question, we immediately run into the problem of defining what is meant by “work”, and what a convincing answer might look like. Pro-lockdowners are still relying on their models and assumptions as their counterfactual case, rather than looking at the mounting evidence of real-world outcomes. My argument is entirely evidence- and outcome-based, therefore it may be that there is no common ground for a discussion. There is also disagreement over where the burden of proof should lie, and whether a precautionary approach could justify lockdowns, at least a year ago, when there was greater uncertainty.
Naturally, we can think of many uncontrolled factors which will have an effect on outcomes, and there are probably also some unknown factors yet to be thought of. However, the case for lockdowns requires that they have a very significant positive impact, big enough to show through any confounding factors, since they have such immense negative impacts – human, social and economic – including deaths. Hence Johnson’s recent statement (14th April) emphasising lockdowns over vaccination, for example. Lockdown was his decision, and if the lockdown wasn’t the overwhelming factor, it wasn’t justified, therefore he has to argue that lockdowns are the overwhelming factor driving down the epidemic numbers.
We might ask, what would a satisfactory evidence-based argument in favour of lockdowns look like? Firstly, evidence should show a positive correlation between the use of lockdown policies, and public health outcomes, especially death rates. This would be evidence against the null hypothesis (that lockdowns are ineffective). Then, further evidence should show that the positive correlation was causative, not coincidental. This would be the limit of scientific argument; we would then move on to the moral and political arguments around the balance of costs and benefits, and the rights and wrongs of imposing heavy costs (up to and including death) on one section of society, for the benefit of another.
For now, we are still at the beginning of that chain of argument; testing the null hypothesis. While there will always be uncontrolled factors which distort like-for-like comparisons between countries which did lockdown and those which didn’t, a comparison of US states is a useful starting point. Although there is a wide spread in the figures, the average death rates per capita are very similar between lockdown (1374 per million) and no-lockdown (1278/m) states. The outcome in Fig 1 is entirely compatible with the null hypothesis. In Europe, the discussion tends to centre around Sweden, the only large country which did not mandate a lockdown. Sweden’s death rate from Covid to March 22, 2021 is 1289 per million; for Germany, France, Italy, Spain and Poland (the most populous EU nations which locked down), it averages 1332/m. These figures are not significantly different, so the outcome is again, entirely compatible with the null hypothesis. Nor is there much difference if we compare Sweden to itself in other years; Fig 2 shows all-cause mortality in Sweden since 1900. The stand-out line is the year of Spanish flu. 2020 was worse than 2019, but better than any year before 2013. Another country which did not lock down, Belarus, has suffered just 227/million Covid deaths, while several Far Eastern nations such as South Korea and Taiwan have had far lower death rates.
Fig 1 US states Covid deaths per capita, lockdown vs no-lockdown
Fig 2 Sweden all-cause mortality vs year
In addition to comparing one country to another, we can look at trends within a single country which introduced lockdown measures, such as the UK on 23rd March 2020. The trend of new cases had been following a constrained growth model (such as the Gompertz function) with a peak on 1st April. The introduction of lockdown did not disturb that trend at all. The death rate peaked on 8th April; a peak which was already inevitable before the lockdown began, given the 18 day average interval between infection and death.
A study https://arxiv.org/pdf/2005.02090.pdf by Wood et al contains the graph Fig 3:
Fig 3 (Wood et al): grey dots are hospital deaths from Covid-19. Black line is inferred fatal infections given the disease duration. Red vertical lines mark the start of each of 3 lockdowns. Relative peak amplitudes are not directly comparable as the IFR has declined with the development of more effective treatments
It is clear from this graph, that lockdowns have consistently been imposed after a peak in fatal infections, and have made no noticeable impact on the shape of the curves. This does not prove the null hypothesis, but it does show that the lockdowns cannot have been responsible for turning the fatal infection rate down.
This is not to say that lockdowns have no impact at all on the epidemic’s trajectory. That cannot be proved. But this, and the previous examples, do prove conclusively that lockdowns are not necessary for the epidemic to decline, measured by fatal infections and deaths. Therefore, the argument (repeated on 14th April by Johnson) that “the lockdown … has been overwhelmingly important in delivering this improvement in the pandemic” is demonstrably false; and the legal requirement for emergency powers (that they should be both necessary and proportionate) has not been met. If Johnson doesn’t grasp this himself, his scientific advisers must; but none have contradicted him.
Britain has had among the most stringent Covid restrictions in the world and spent a total of 213 days under national stay-at-home orders from 23 March 2020 to 29 March 2021. This is among the highest totals in the world. And yet, Britain has also suffered the fourth highest Covid death rate in the world. As well as causing painful cognitive dissonance among those who went along with lockdowns, that single fact is leading a lot of people to ask, why didn’t it work? All that sacrifice and we still have the world’s fourth highest death rate. What went wrong, and why, a whole year later, are we still doing the things that didn’t work, and not doing what worked elsewhere?
The table below uses the stringency index – a measure of the combination of NPI policies, not just lockdown or no-lockdown – compared to 2020 Covid death rates:
Country Level of stringency Covid-attributed deaths per million in 2020
Taiwan 0.53 0.3
Japan 0.88 27
Estonia 0.90 173
Finland 1.01 101
Norway 1.10 80
Denmark 1.30 224
Singapore 1.32 5
Sweden 1.31 861
South Korea 1.41 18
Germany 1.54 404
United Kingdom 1.59 1066
France 1.59 985
Italy 1.65 1218
Australia 1.77 35
Canada 1.81 410
United States 1.91 1058
The level of stringency indicates the relative severity of a nation’s combined policies of school closures, workplace closures, restrictions on public events, restrictions on gatherings, closures of public transport, stay at home requirements, restrictions on internal movement, and restrictions on international travel.
Sources: The Oxford Coronavirus Government Response Tracker, Policy indictors (C1-C8), containment and closure index; Worldometer, Coronavirus, December 31
We can pick apart the “stringency index” if we like, but the simple fact is – there is no correlation between the use of lockdown policies and the death rate, in countries which are otherwise comparable and had outbreaks of Covid at around the same time.
So, why is grudging consent for lockdown policies still so widespread? Partly it’s because most people, for better or worse, accept what they are told by authority figures. Partly it’s because it is truly awful to confront the increasingly obvious fact that the immense sacrifices we have made through lockdowns, were wasted. Also, though, it’s because there was an appeal to common sense in the original lockdown pitch: that if you stop people meeting other people, they will have fewer opportunities to pass on infections. Many people simply accepted that argument, and still do, without further thought. What’s wrong with it?
The prevalence of Covid deaths in hospitals, care homes, and to a lesser extent other institutions, shows that the necessary concentration of elderly and vulnerable people in such indoor environments created the conditions in which the majority of fatal infections occurred. Lockdowns did not abolish such institutions; we cannot close the hospitals and care homes. To some extent, therefore, focusing public policy on closing business premises, factories, schools, shops and outdoor sports, was a huge distraction from the fact that the infected and the vulnerable were herded into indoor spaces, with mixing going on (especially the discharge of Covid patients from hospitals into care homes) in ways which actually made the transmission of infection more likely; while the healthy were forced to spend more time together indoors as well, doing nothing to reduce the likelihood of transmission among those not at risk of dying.
The fig leaf of a “common sense” argument for lockdown policies was a cover for the fact that no real threat analysis had been made. The Government did not understand who was at risk; who was not at risk; and which measures could have protected which group. While implementing ineffective policies restricting the young and healthy, they failed almost completely to implement policies which could have protected the elderly and vulnerable. This was a failure of rational thought on a vast and catastrophic scale.
What does science tell us?
A substantial amount of statistical work on the epidemic is being published, with a range of conclusions, and we may expect a great deal more over the coming years. It is normal, and to be expected, that honest analysts will differ, given the many variables and the data quality issues; and it is also to be expected that many people will seek to defend their established position, given how much is at stake. Looking for a consensus is therefore no guide to the right conclusions.
However, already, even a superficial look at the outcomes shows that there is no obvious and positive correlation between lockdowns and a lower death rate from Covid (or lower all-cause mortality). The null hypothesis, while it cannot be absolutely proved, is certainly in front on the evidence. More detailed analysis will seek to allow for the uncontrolled differences, including population density, GDP, ethnicity, climate, urbanisation level etc. Already it is clear that the age profile and the prevalence of obesity are strongly correlated with death rates, for example. But there is no such strong correlation for lockdown policies. Taiwan, with a similar population, urbanisation, and age profile to the UK, has had 10 Covid deaths. That’s not 10 per million; just 10 in total, despite its proximity to the origin of the virus. Yet, the pro-lockdown argument relies on it being the sledgehammer which controls the epidemic by brute force: if lockdowns are not overwhelmingly effective, there is no case for their use, as the harm they cause will be greater than any good they do.
You might argue that age-demographics and the prevalence of obesity are not open to government action. Yet this is a tacit acknowledgement that lockdowns were a case of “something must be done, this is something, therefore this must be done”. Also, it’s not entirely true that obesity is not open to government action. Britain has been getting fat for years, and last summer, the very first thing the government did when the initial lockdown was being lifted, was to offer a 50% subsidy for eating out.
What’s the alternative?
It’s worth pointing out that the alternative to mandatory business and school closures, and whole-population stay-at-home orders, is not “do absolutely nothing” and “let it rip”. That, of course, is the straw-man argument put forward by the pro-lockdowners, and a rather crude attempt to pass their own guilt for what has happened onto those of us advocating something more effective. When I refer to the guilt for Britain’s excessively high Covid death toll, plus the death toll of lockdowns, I should clarify that I’m willing to believe that the pro-lockdowners, at least in March 2020, were perfectly sincere in their hope that it would do good; they had no malicious intent back then. But the longer this situation goes on, the harder it is to maintain that outlook in the face of all the evidence. It is only human to be wrong, and that is forgiveable. What is culpable is staying wrong, long after the error is clear.
What should be very obvious is that Covid has “ripped” anyway; what Britain did, lockdown and all, was ineffective or counterproductive, especially so in hospitals and care homes. The alternative is not business as usual pre-2019. Many voluntary changes including widespread home working, hand hygiene etc. had already occurred before the first lockdown was imposed. The use of border restrictions at the right time – before the infection was widely seeded in the UK – and an effective TTI system were critical to the success of countries like Taiwan, along with targeted protection for the vulnerable. The use of these effective measures also required a certain degree of preparation – stocks of PPE, and sufficient NHS capacity – which had been deliberately run down over recent years for largely financial and political reasons. It is the policies and outcomes of countries which were better prepared and which reacted more effectively, which are the alternative to which Britain’s late and repeated lockdowns should be compared.
So, if the null hypothesis cannot be ruled out, how did so many governments manage to do an end-run around the question, and establish lockdowns as a received wisdom? After all, the next question, if you can show some evidence against the null hypothesis, is to demonstrate a causal relationship. So far, I’m not aware that anyone has produced any evidence of this type. In a clinical trial of a drug, the null hypothesis (that the drug has no significant effect) can be ruled out by showing a significant correlation between the drug and better outcomes, in a double-blind controlled trial. Showing a causal relationship for that correlation would involve studies of the biochemical mechanism for the drug’s action. The corresponding work in the epidemic case would involve studies of the transmission of the virus and experimental work to stand up the many assumptions about that, built into the epidemiological models. However there’s a problem with that.
The Imperial College model – by far the most influential – not only wildly overestimated the death rate without lockdown (as tested in places such as Sweden and Florida) but also underestimated the death rate if lockdown-type restrictions were imposed. In short, it heavily overestimated the death rate in the no-lockdown scenario, but also heavily overestimated the beneficial impact of social restrictions. Most of the assumptions built into the model had not been tested or quantified in real-world situations. Hence the WHO’s 2019 report “Non-pharmaceutical public health measures for mitigating the risk and impact of epidemic and pandemic influenza” warned of the flimsy empirical basis for epidemiology models such as the one developed by ICL. “Simulation models provide a weak level of evidence.”
In late October, Johnson was once again panicked into a lockdown by alarmist modelling. The graph in Fig 3 was distributed to MPs to get them to vote through the November lockdown:
Fig 4: Chart used in a government presentation (no-lockdown scenarios) – Halloween 2020
The government’s chief medical and scientific advisers, Chris Whitty and Patrick Vallance, stood in front of this graph, knowing at the time that the 4000-going-on-6000 deaths per day scenario had already been falsified by real outcomes at that point. They spent the rest of the week backtracking and expressing regret for doing so, having been called out for such alarmism, but too late; the lockdown vote had been passed. And yet, the pro-lockdown argument is still based on claiming credit for the huge difference between such alarmist models and the real-world outcome. Still has no explanation been offered by the modellers as to why their outputs are so wildly and consistently wrong on the high side; plenty of explanations are offered by others, but somehow only the modellers have the ear of government. Nobody in the government or SAGE seems to ask: these estimates are out by a factor of several times, so which of our assumptions are wrong? Because to do so would destroy the only case that has ever been made for lockdowns.
It is interesting to note that the curves in Fig 4 all predicted steep falls in the death rate, following peaks variously anticipated to be in late December or early January, in a no-lockdown scenario. Three of the four models appear to show the death rate reaching close to zero by the end of March 2021, without lockdown. That feature of the modelled trends, is the only one that is borne out by reality. In this respect, these no-lockdown models predicted the outcome we have had, with a lockdown. And yet, now that we are on this predicted downslope, and have reached very low death rates, the pro-lockdowners (led by Johnson) are attributing this trend to the January 5th lockdown, claiming credit for something they had predicted would happen anyway! The alarmist model was dishonest when it was first presented; to now claim that the current falling figures are attributable to lockdown is double dishonesty.
Johnson (14th April 2021) made a statement that it was “the lockdown that has been overwhelmingly important in delivering this improvement” rather than the vaccination program. However, this is based on the presumption that there is anything needing to be explained and attributed. Given that his own modellers had already predicted deaths falling to background noise levels by April without any measures being taken, there is nothing that requires to be explained or attributed, whether to lockdowns or vaccination. It is what the government told us would have happened in any case. Johnson might as well be claiming the credit for the sunrise this morning.
During the Black Death, there was a widely held belief that bubonic plague was God’s punishment for sinners, and that repentance and prayer would end it. We could call that the prayer hypothesis. The null hypothesis would be that praying made no difference to the course of the plague. Like today, no controlled trial was possible. The plague eventually went away, leaving many people convinced that their prayers had done the trick. They saw the epidemic die down, and correlated that with their prayers. Given their pre-existing, deeply held religious beliefs, this left them firmly convinced of the prayer hypothesis. We must avoid falling into the same logical trap.
The modern equivalent is that “confining the entire population to their homes for over a year makes such a significant positive impact that it is a necessary and proportionate response to an epidemic with an infection fatality ratio below 0.02% for people under 50”. That is the Lockdown Hypothesis. Look, the believers say, we locked down on 5th January, and behold, the plague is going away. This is no more valid than the prayer hypothesis of the 1300s. One can understand and forgive the error of the medieval people given their mindset; but today, we at least claim to be logical and scientific in our thinking, so there is no such excuse.
Having failed to give convincing answers to the questions addressable by science, governments moved straight on to swerve the next questions; what level of collateral damage caused by lockdowns is acceptable; and why should the heaviest burden of lockdown policies fall on those at least risk of serious illness or death from Covid? In short, is it right, or preferable, for one 40 year old to die of cancer than for two 80 year olds to die of pneumonia? Is it right to privilege one cause of death over all the others?
The biggest unresolved question created by the use of lockdowns is, what is the measure of their success, and when can we return to normal life? Some proponents of lockdowns argue for total eradication, while others at least imply that once the death rate from endemic Covid is down to a reasonable level, we can get on with normal life. This raises the question nobody dares to answer; how many deaths are acceptable? In the winter flu season of 2017/18, there were more than 50,000 excess deaths in England and Wales, and it went almost completely unnoticed in the press and public sphere. 78,000 people die prematurely each year from smoking, and we continue to defend their right to choose to smoke. These numbers should set an expectation of what is within the parameters of a normal, acceptable national death rate, the sort of level which is tolerated without drastic policy responses and without attempting to shift the burden of death from those on which it would naturally fall, onto others. Yet, public policy has been entirely focused on deaths from one, highly visible cause – Covid – and has neglected to give similar weight to other causes.
Without specifying the measure of success or defining the end point, the door is wide open to shifting the goalposts. The original rationale for lockdown one year ago was that it would be short – “three weeks to flatten the curve” – and prevent the NHS being overwhelmed. It was compared to August in France, when many businesses close for a month but basic services carry on and the economy survives unharmed. A year later, the original rationale, objective, and sales pitch are no longer mentioned. Despite the accumulating evidence that it’s not effective, the lockdown – or something like it – continues, supposedly to keep out possible variants.
The abandonment of rational thought
A number of the points made above relate to the lack of rationality in public policy. Examples of this include:
- The continuing failure to carry out any cost-benefit analysis reflects “this time is different” thinking and the back-covering defence of entrenched positions
- Scientific ideas can be defined as those capable of being objectively disproved, yet lockdown policies, while claiming to be driven by “the science”, have never acknowledged the possibility of being falsified, nor set criteria for disproof
- The over-reliance on models, despite the underlying assumptions being untested, and a continuing refusal to revisit those assumptions in the light of real-world outcomes
- Consistent failure to present Covid statistics in context
- Repeated confusion of correlation with causation (as in Johnson’s claim that the spring fall in the epidemic was attributable to his decision to lockdown)
- The emphasis on death counts rather than the long established use of life-years to assess public health policies
- The suppression of different views and refusal to engage in logical argument
It has to be recognised that all sides in the discussion share the problem that we cannot know what would have resulted from different choices. Pro-lockdowners rely on claims that outcomes would have been much worse without lockdowns, and seldom acknowledge that this cannot be proved. Nor do they easily engage with analysis of outcomes in countries which did not lockdown. On the other hand, our knowledge of the negative impacts of lockdowns is also limited. We have only statistical methods to estimate how many will die from undiagnosed and untreated non-Covid conditions, or from the economic impacts of lockdown; and statistical arguments are often unconvincing to the public.
All the same, justifying a novel policy option such as lockdown should require that the net benefit be very substantial. The imposition of great harm, up to and including death, on certain sections of the population, for the benefit of other sections, might possibly be justifiable if the net gains were overwhelming and indisputable – let’s say, a factor of ten. If this were the case, we would see it clearly in the data from all countries: those which did, and those which didn’t lockdown. We would also see it in those countries which did lockdown, in before-and-after data.
While correlation doesn’t prove causation, the lack of correlation is an absolute slam-dunk. The failure of lockdown policies to show an overwhelming impact is now beyond any doubt at all. That anyone is still advocating them is a profound failure of rational thought
Where are we now?
Because our government never had a coherent strategy to deal with Covid, one which contained answers to these questions, they are forever making it up as they go along. Having announced a timetable for lifting restrictions, and despite the epidemic disappearing almost completely in the UK, they are now dithering over their decisions.
We are in far better shape now (April 26th) than in January 2021. The number of positive tests per day is well under 2,500 (probably largely accounted for by the operational false positive rate); 33.6m have been vaccinated (including almost all in the at-risk groups); the average daily deaths-with-Covid figure is 17.6 (down from 1,283 on Jan 19th) and still trending down (although it cannot reach zero, as it is includes all-cause deaths among those who have had a positive test within 28 days, and among a numerous group, someone is bound to die of something). There are 1,961 Covid patients in hospital, down from 38,800 in January, and also trending down. These trends in the figures, along with the modelling from late last year (for what it’s worth), all suggest that we are on the normal downslope of an epidemic curve which peaked in late December, and the vaccination programme is reinforcing that and making sure that the numbers stay down. The schools re-opened a few weeks ago with absolutely no impact on the levels of infection, which surprises only SAGE. Two thirds of adults in Britain have received at least one dose of vaccine; if the vaccine is effective – as controlled trials showed it to be – we are now at, or possibly over, the herd immunity threshold, the level at which the disease no longer spreads. Under last year’s tier system, the whole country would now be in tier one. The NHS is tired, but not anywhere near being overwhelmed; indeed, it is all the more important to normalise things and deal with the 5m people now on waiting lists for other reasons. All-cause mortality is now below normal. By any standard, we are not in an emergency. Yet we are still under emergency restrictions.
The damage caused by lockdown is cumulative. Every day causes more economic and human harm, more deaths, more businesses which will never re-open, more livelihoods lost, and more widening of the poverty gap. Each additional day of lockdown is something which needs to be justified, but the thing which has been glaringly absent all year is a cost-benefit analysis of lockdowns. The government has made no serious attempt to evaluate the cost of lockdowns, and thereby is playing down their impact. This impact is heaviest on the poor and those with insecure livelihoods; that is true both nationally and globally. Among their many other side-effects, lockdowns transfer suffering from the rich to the poor.
On the basis of some of the biggest sunk costs of all time, nobody in either main party will now be able to acknowledge that lockdowns don’t work, any more than Brexiters can ever acknowledge the harm they have caused with that misguided policy. It is just too big a deal to acknowledge that it was a mistake, apologise, and move on. Our political system and its advisers are fully invested in safety-through-lockdown and will never be able to change their position. What if an eventual public inquiry concludes that the government’s response was disastrously wrong? The question only has to be asked to suggest the obvious answer: the terms of reference will be set, and the inquiry leader selected, to make sure this outcome cannot happen. The inquiry will be defined so as to vindicate the government’s basic lockdown approach, not call it into question. The inquiry will be allowed only to consider whether the government should have locked down sooner, or for longer; not whether it was barking up altogether the wrong tree.
The precedents which have been set
As well as the direct harm of lockdowns, there is huge constitutional harm which will become clear over time. The government took powers under the Coronavirus Act, which was passed in just five days, without a vote, and without any of the line-by-line scrutiny given to the introduction of the 5p charge for plastic shopping bags. It is striking how there is always time to write the most draconian laws, but not to read them. Many, possibly most, of the rights listed in the Universal Declaration of Human Rights (1948) and the ECHR (1959) are either removed altogether, or infringed, by the lockdown. After a whole year of this, it is becoming a legally established fact that these fundamental, supposedly inalienable rights are not actually rights at all, but privileges which ministers can turn on and off like a light. Johnson, always inclined to dodge mechanisms of accountability, was already keen, before all this, to restrict judicial review and re-write the Human Rights Act 1998; now he has seen how easily draconian measures can be passed. Neither Parliament nor the courts have stood in his way; there has been no opposition worth the name. Ministers have become used to controlling us by executive fiat; through measures which are not humbly recommended to Parliament for debate, amendment, and voting on, but announced from the new Press Room at Downing Street, by Ministers standing in front of starched flags and telling us how our lives will be. Ministers have taken the right to decide whose lives matter, whose illnesses get treated, whose livelihood is “essential”, who can practise their religion, who can earn a living, how long you must wait for your case to be heard in a court, whether you can receive your education, which lovers can meet, and in every respect, when and where you can go about your life.
Our rights and freedoms, established by so many struggles down the centuries, have been treated as a luxury to be dispensed with as soon as something moderately bad happens. The precedent is now firmly in place, that these freedoms are no longer our birthright, but are in the gift of our rulers. Already it seems that the freedom to assemble and protest is to be removed permanently; protest is henceforth to be something the authorities will permit only as long as it causes no “nuisance”, in the opinion of a police officer or the Home Secretary.
When rulers take extra powers, they seldom give them up again. Emergency powers are always sold as being temporary, but somehow fail to expire. The Patriot Act, passed in the US in just six weeks after 9/11, granted the government a long (and obviously pre-existing) wish-list of intrusive powers which made a mockery of the US constitution and bill of rights; many of them are still in force. So it will be in Britain after the epidemic. Some powers will expire, but it will have been established that we can all be put under prison-like restrictions at the stroke of a Prime Minister’s pen. An underlying principle of liberal democracy has been surrendered.
Governing by edict; disregarding basic rights and freedoms; and the vast over-reaching of government authority beyond its normal range and into the micromanagement of personal behaviour, has created the toolkit for tyranny. Tyrants do not all come carrying guns and building gulags; I’m perfectly willing to accept that the people in government sincerely believe they are acting for good reasons and with the intention to protect against a public health threat. That doesn’t matter; the principle is now established and the precedent set, that when it feels the need to do so, government can reach into all these areas of life and rule by edict. If they subsequently stop doing so, the precedent remains for the tyrants of the future, that they can take such power. The Conservatives used to be the party of limited government; well, they’ve made a very Big State now. They have shown that belief in limited government only applies during the good times; come even a small emergency and it goes right out the window. Covid happened to come at a time when new technologies, especially mobile phones, have created immense new types of data about every individual, our actions, movements and communications. Just as other trends, such as the decline of high street shops, have been accelerated by the epidemic, so has the toolkit of social control expanded ever faster. The use of cash has declined much more quickly in the last year; more and more of our transactions are now done through the banking system, leaving a trail; while we have at times, been made to sign in to pubs, and encouraged to install a tracing app which monitors who we meet. The idea being floated now, of Covid passports, raises the prospect of the introduction of a kind of ID card through which almost anyone will have access to some aspect of our medical history.
There are some on the political right who still do believe in limited government, and have until recently provided the only push back against the use of all these powers. The left mostly went along with the official narrative, although the burdens of lockdown are heaviest on the poor, the insecure, the disadvantaged, the young, and the global South. The Labour party have not only failed to oppose the government, they have secured nothing in return for their support; not even a promise of some kind of restitution after the epidemic. The government will continue to run down the NHS and there is no plan to strengthen the education system after all the damage done to it. The rich are still getting richer even with much of the economy shut down, and Labour has been the enabler of all this.
The role of government communications; the fearsome and threatening advertisements; the excessive punishments for breaking the rules; and the slavish support of almost all the press, together these things have manufactured popular consent for lockdowns. People have been told to leave all the thinking to the experts, or scientists, or ministers; people who are smarter than us. Democracy has been put on hold; not just through postponed elections and a sidelined, rubber-stamp Parliament, but stifled debate and withered opposition. But there are now signs that it’s no longer working – more people are disregarding the rules, and complaining about them. Elements of the press, beginning on the political right but increasingly also on the left, no longer buy the lockdown narrative. Scepticism is no longer the preserve of professional contrarians like Nigel Farage; the Mail and the Telegraph have started printing anti-lockdown opinion pieces, and they reach a lot of the Tory heartland. Johnson doesn’t care what the opposition thinks, but he does care what his own voter base thinks. Manufactured consent can disappear very rapidly once it starts to erode. There is a gathering sense that people are just waiting, increasingly impatiently, for all this nonsense to be officially called off. A huge protest march took place in London on 24th April; it was only reported in the context of some trouble afterwards in Hyde Park, but shows that plenty of people are now willing to turn out and protest against lockdowns.
Political failures
Government communications have been terrible throughout this episode. Often contradictory, seldom clear, and frequently changing, ministers have taken their turn to stand in front of those starched flags and tell us how our lives will be. Vacuous slogans are repeated everywhere from radio commercials to motorway overhead signs, such as “stay local”, but there is no legal or official definition of “local”. Even the police don’t know the difference between the law – which we are obliged to obey – and guidance, which is a request for us to behave in certain ways. The entire government approach was based on dividing all activities into essential and non-essential categories, but these have never been defined in law. The extreme government over-reach has encouraged Police over-reach; they were drawn into making roadside checks on drivers’ purposes for travel, or employing drones to survey dog walkers thought to have gone too far from home. Meanwhile the opposition failed to oppose any of this structure of badly-written law and ministerial edict. Ministers often claimed to be guided by “the science”, but many of the restrictions are arbitrary (eg. how many people can meet, how often people are allowed to go out for exercise) for which there has never been the slightest scientific evidence. The opposition failed to demand a full cost-benefit analysis of policy; and where policy was originally precautionary, they failed to demand proper monitoring of effectiveness; they gave the government their support without getting anything in return, not even legally binding promises of restitution for the NHS, the education system, and full support for those hardest hit by restrictions. Emergency powers have been in force for far too long, and still have no firm end in sight. The opposition have still failed to demand a full return to normal, rather than a new-normal in which some of the recently introduced powers and restrictions are indefinite. They have never engaged in the discussion of where we are going with all this. The government have flown by the seat of their pants, without a plan; the opposition have failed to hold them to account for this, or develop their own plan. Still, more than a year into this ongoing disaster, we have no date for a public inquiry; the opposition have failed even to establish the public understanding that it is a disaster. Never has the government been able to articulate its overall policy objective: do they intend to eliminate Covid completely, or reduce it to an acceptable level, and if so, what level is that? There has been no attempt to hold a grown-up conversation as to what an acceptable level of mortality might be, despite the fact that everyone knows such a thing exists and that it cannot be zero. We have submitted quietly while the government followed a “whatever it takes” approach in which no attempt was made to measure the costs or the harms of the responses, which have, as a result, been allowed to run vastly out of control. Above all, nobody, back in March last year, asked: what would make me wrong? What new information would prove me wrong about this approach? If the world of politics had been more humble and accepted that not everything is, or ever can be, under their control, and that their emergency policies could turn out not to work, we might by now have dropped the failed approaches. It was the job of opposition to point out what wasn’t working, and wiser heads would not have invested so much in an uncertain approach. Sadly, government and opposition are both fully bought in and cannot now row back. It will be a generation before the facts can be dispassionately acknowledged.
Covid-19 has been a genuine crisis which has killed a large number of people and played havoc with the lives of many others. But it too will pass and the worst of it has passed already. It must not be allowed to become an excuse for permanent, even higher levels of governmental intrusion and control.
In what we think of as an age of reason, it seems we were primed to expect an apocalypse. Out of fear of a deadly threat, we went along with an untested approach which was assumed, without evidence, to work. Now we have a year’s worth of evidence that it doesn’t work, but every day, the government continues to prevent normal social contact, to destroy businesses and jobs, and to ruin children’s education, their mental health and future prospects. Today’s youth have had to sacrifice their schooling, happiness, prosperity, career prospects, sense of purpose, sanity and freedom, all for no benefit to themselves. Those who work as actors, musicians, and waiters; bar, restaurant and hotel staff and owners; instructors of all kinds, and those in a wide variety of hospitality, tourism, sport and retail businesses, are watching in despair as everything they have ever worked for is steadily destroyed. A vast amount of normal economic output has been prevented, and £400bn of public money has been poured out, for a gain which is still only assumed to exist at all.
Our rulers have done what rulers often do: they said, “give up your rights and freedoms, trust me, and I will protect you”. History has proved this is never a bargain to accept. In the face of an overstated threat on one side of the scale, no rational or measured consideration on the other side can ever match up. Then, in the name of protecting society, harm is done. In this case, the harm is to those who will die from other illnesses which were deprioritised; those in the care homes who went unprotected; those whose livelihoods were sacrificed and who will suffer the deaths of despair, unemployment and poverty. They have been killed, just as dead as anyone who was ever put against a wall and shot.
In a tacit acknowledgement that the original Covid no longer has a death rate which justifies bringing the world to a standstill (if it ever did), we are invited to remain paralysed with fear that some possible future variant will emerge that will be worse. Everything that makes life worth living has been sacrificed to this fear. Churches and theatres have been closed. Childrens’ football games and rounds of golf became criminal offences, despite there being absolutely no evidence that outdoor sport was a risk to anybody. Getting your hair cut or having a pint with friends was off the menu for months. Music festivals still hang in the balance. Holiday travel may resume but with a significant cost barrier.
We now have a far better understanding and a good measure of the severity of the epidemic and the consequences of how we respond to it. Even if we didn’t know what we were doing a year ago, we do now. We are treating a case of head lice with a guillotine.
Notes from Interesting Times 