"By shutting down the NHS, driving people to suicide, and creating poverty and unemployment that worsens people's long-term health, lockdowns themselves carry enormous mortality costs that make them more costly than mere economic analysis can measure. As a result, lockdowns lead to many deaths - perhaps more than they prevent."
The degradation and closure of NHS services has been because of Covid, not lockdowns. Lockdowns involve restrictions on shopping, eating out, travel, and business - but not healthcare. Where NHS Trusts have tragically had to stop treating people for conditions like cancer and heart disease, and paused elective surgeries, it's been because Covid cases have been so high that they've had to divert resources (like beds and medical staff) to treating those patients, and because the risk of those patients catching Covid in hospitals was too high. Health services have been suspended not because of lockdown, but because of the Covid pandemic. And in the absence of lockdowns, there would have been more Covid cases, undermining non-Covid care even further.
The idea that "hospitals are empty" (sometimes illustrated with photographs of empty corridors or unsourced anecdotes) is not correct. The Health Service Journal reported on 18 January that Covid patients occupied over a third of acute inpatient beds across 58 hospital trusts. The number of beds occupied arguably understates the pressure on the system from infectious patients, who need to be isolated.
Suicide rates have not risen during lockdowns, either in the UK or worldwide. The evidence is very strong on this. In the UK, there was no statistically significant rise in suicides after lockdowns were imposed, and there has not been a rise in rates of self-harm. This is true across other developed countries as well: reports from Germany, Norway and Massachusetts, USA, show either no rise in suicides or a fall in suicides during initial lockdown measures. So far, Nepal is the only country where a study shows a rise in suicides associated with Covid. A recent study from Japan shows a fall in suicides during the first half of 2020, and a rise in the second half, but no increase overall on previous years (NB: Japan did not have a full lockdown during 2020 at any point). Even Victoria, in Australia, which experienced one of the world's longest lockdowns, saw no rise in suicides in 2020 compared to previous years.
Lockdowns have a mixed relationship with violence and other causes of death. The effect of lockdowns on crime is a complicated picture: studies have found that reports of domestic violence and cybercrime have increased during lockdowns, while reports of harassment, rape and violence have dropped. In parallel, studies from around the world have found that deaths from homicides and traffic accidents have all dropped during lockdown orders. Meanwhile, the decline in pollution has also likely saved lives. In total, this suggests that using the measure of excess deaths may cause an underestimate of the death toll of Covid-19 – as deaths from many other causes have been declining during the same period.
Recessions are very bad - but usually don't raise mortality rates. Counterintuitively, recessions are associated with short-term falls in mortality. As Nobel Prize-winner Angus Deaton said, summarising the evidence, "Many papers, many places, many times find that all-cause mortality falls in recessions". It's a mistake to assume that a lockdown-induced recession should necessarily involve extra deaths – the opposite is usually the case.
Lockdowns do save lives. This analysis from November looked across over 200 countries at effective non-pharmaceutical interventions at preventing the spread of the virus and saving lives. All of the measures associated with national lockdowns are found at the very top of the effectiveness table.
Another paper found that "closing all educational institutions, limiting gatherings to 10 people or less, and closing face-to-face businesses each reduced transmission considerably", while additional stay-at-home orders had a small effect beyond these.
In the UK, an Imperial College report concluded that if the March 2020 lockdown been introduced a week earlier, this “could have reduced the first wave death toll (up to 1st July 2020) from 36,700 to 15,700 … while delaying lockdown by a [further] week would have increased the deaths to 102,600”. The higher death rates in the counterfactual "later lockdown" scenario are shown in green on the graph below:
Source: This preprint.
A common criticism of lockdowns is that people would have made voluntary behaviour changes anyway which would have slowed viral transmission. The same report addresses this by showing that the timing of the sudden shift away from exponential growth was timed precisely with the UK government’s ‘stay at home’ order. This reduced the reproduction rate suddenly below 1 across all regions of the country, as shown by the different coloured lines. Even a short further delay in this national behaviour change would have increased the death rate substantially, as per the model above.