Toby Young has been given plenty of stick for an article which asked whether the UK government has overreacted to the coronavirus crisis by locking down the economy, and which concluded that it had. In my view, he was right to pose the question, but gave the wrong answer.
The thrust of his argument is that the economic costs of the measures being taken are too high to justify ‘extending the lives of a few hundred thousand mostly elderly people with underlying health problems by one or two years’. This is a clumsy way of expressing the problem and, as I shall explain later, his numbers are way off. Nonetheless, he makes a potentially valid point.
It’s common practice for health economists to put a monetary value on people’s lives according to the number of years they have left, and the quality of that life. This is the concept of a ‘QALY’, or quality-adjusted life year, where one QALY is one year in ‘perfect’ health.
The Treasury’s Green Book, which provides guidance on how to appraise and evaluate policies, projects and programmes, values one QALY at £60,000. This is an answer to the ‘normative’ question of what value society should place on health outcomes, based on ‘willingness to pay’.
The National Institute for Health and Care Excellence NICE uses a lower figure of £30,000 in assessing new treatments for the NHS. This figure is based on a more pragmatic assessment of the funds actually available (and is the one that Young uses). In both cases, a monetary value is being placed on a human life.
Crucially, this is not ‘eugenics’, nor is it about people’s ‘wealth-producing capacity’ or ‘economic productivity’ (as many of Young’s critics assume). If you believe this concept is morally wrong, imagine you were on the Titanic and had a straight choice between rescuing a healthy child or a sickly old man. Whom would you save, and why?
Similarly, we have to draw the line somewhere in allocating limited resources. For example, if we could save just one person’s life at a cost of £10,000, almost everyone would surely say yes. But what if it would cost £1 million (possibly)? Or £1 billion (presumably not)?
Indeed, policymakers are making these sorts of judgements all the time. The same techniques are used to assess the value of lives that might be saved by road safety improvements, and in determining awards of damages in court judgements. The BMA’s own guidance to medical professionals underlines that it is both legal and ethical to prioritise treatment among coronavirus patients.
But I still don’t agree with Young’s conclusion. Let’s begin by dissecting his cost-benefit analysis, assuming this approach is indeed the right one here. (I’ll come back to that assumption too.)
On the benefit side, he has attempted to value the additional lives that might otherwise be lost to COVID-19 if the government took a more relaxed approach.
To be fair, his conclusion does not actually rely on the assertion that the tougher measures would only extend the lives of those saved by ‘one or two years’, which is obviously false. (Young has taken an out-of-date number for the average age of those who have died, and compared this to the average UK life expectancy at birth rather than at the age already reached.)
In practice, he uses a reasonably high number for the potential number of fatalities (370,000) and a more realistic (but still low) figure for the years of life lost (an average of 11). Assuming each year of life is valued at £30,000, this implies a total cost of £330,000 per life, and £122 billion in total. But both these figures are still questionable, especially if the NHS becomes so overwhelmed with patients that it cannot effectively look after younger people with non-coronavirus related problems too. And even relatively young and health medical staff are vulnerable to a high ‘viral load’.
There is also a huge amount of uncertainty here. If it goes horribly wrong, the first figure could be as high as 500,000, and the second as high as 20. We’d then be looking at a total cost of £330 billion. And you could also justify using a much higher number for the cost of a life, such as the ‘Value of a Statistical Life’ saved by a road safety improvement, which is more than £1 million. At the very least, the Green Book’s £60,000 for a QALY seems closer to the concept that Young is trying to capture.
On the cost side, Young starts by focusing on the headline figure of £350 billion for the additional government support announced by Rishi Sunak on 17th March. He does then acknowledge that most of this (£330 billion) took the form of loan guarantees that may never actually be needed, but only cuts his number to £185 billion.
In reality, the direct costs to the Treasury of the measures announced so far are likely to be in the range of £50-100 billion. What’s more, this can be seen as a transfer from one group of people (mainly current and future taxpayers) to others (such as those whose incomes would otherwise have disappeared), rather than a net loss to the economy.
Young also makes the fair point that recessions themselves can cost lives. There is plenty of evidence that life expectancy is lower in more deprived areas of the UK, and that the pace of improvement generally has slowed since 2011. It’s not a huge leap to conclude that an economic slump could itself have a major impact on the health of the nation.
However, the reality this time may again be more complicated. For a start, there is a big difference between a temporary drop in GDP, even a very big one, and a prolonged recession. Provided the great majority of businesses, jobs and basic incomes are protected – which is the focus of the economic policy responses – normal life should resume relatively quickly once the emergency health measures are lifted.
Some would also argue that it was ‘austerity’ that led to the deterioration in health trends in the 2010s, rather than the recession itself. That’s still debatable (something for another day), but there is no doubt that the fiscal responses to coronavirus are very different from those that followed the global financial crisis.
A lot of the economic costs are also inevitable, whatever policymakers had done. For example, the government’s early analysis of coronavirus suggested that, in a reasonable worst-case scenario, up to a fifth of the labour force might be off work at any one time. That would have hit GDP hard, even if the government had done nothing (and probably by more without the measures that the government has taken).
Put another way, the economic measures being taken now have both costs and benefits, as they prevent bigger losses further ahead, including the accumulated loss of future income (potentially substantial) that would result from the premature deaths of a large number of younger people.
In the meantime, people stuck at home might be bored, and even depressed. But most seem to accept the need for the tougher measures, and community spirit seems high. Studies of previous recessions in the US have also shown a mixed impact on overall health, with an increase in the number of suicides in particular offset by fewer deaths due to other causes, such as traffic accidents.
On this basis, it would be relatively easy to redo Young’s analysis with different and better numbers and conclude that the government’s measures are actually good ‘value for money’.
Finally, though, it is worth asking whether the standard cost-benefit analysis, which you might apply to the approval of a new medicine or an individual railway project, is appropriate to a national emergency, like the coronavirus crisis. To use an extreme example, we wouldn’t have assessed the pros and cons of fighting World War II in this way.
Similarly, we may well be willing to put a much higher price on protecting the lives of people who face a ghastly end in a converted conference centre, especially if we could avoid this simply by chilling at home for a few weeks longer. By all means let’s keep everything under review, but, ‘yes’, I do think the lockdown is worth it.
PS (added on 3rd April).
I’ve since come across a few interesting US papers on the economics of pandemics. These studies (which have also been picked up elsewhere) support the view that the coronavirus lockdown is indeed ‘worth it’…
First, a cost-benefit analysis of social distancing for the US, which suggests that the benefits in terms of lives saved could be worth as much as US$8 trillion (40% of GDP)…
Greenstone, Michael and Nigam, Vishan, Does Social Distancing Matter? (March 30, 2020). University of Chicago, Becker Friedman Institute for Economics Working Paper No. 2020-26.
Second, a study of how different US cities responded to the flu pandemic of 1918: those that intervened earlier and more aggressively ‘do not perform worse and, if anything, grow faster after the pandemic is over‘.
Correia, Sergio and Luck, Stephan and Verner, Emil, Pandemics Depress the Economy, Public Health Interventions Do Not: Evidence from the 1918 Flu (March 30, 2020)
Finally, a study which finds that a *temporary* deterioration in the economy is actually associated with a small *improvement* in overall health outcomes (because an increase in the number of suicides is more than offset by other changes, e.g. fewer traffic accidents).
31 thoughts on “Is it worth it?”
You can always mess about with figures, that’s how Ferguson came up with 500,000, then 20,000 and then 5,160, all from messing around with figures.
Young was making a good point, and you say so too, just the figures are/maybe a little out. The point stands though I think
It’s a shame you and people like you didn’t question the change in tack the Gov took a couple of weeks back when Ferguson’s number of 500,000 deaths was bandied about. It would have made for an interesting discussion, particularly around modelling.
Then the question arises, how long would be too long? That question should be answered now because if the answer is say 10 weeks and it is clear that in 10 weeks the problem will still be there, we should cut our losses now. We need to be sure that the cost we are paying will not just be wasted.
The problem here is that we simply can’t say now how long the measures will have to last be effective, and therefore don’t have enough information to cut out losses now. Or put another way, I suggest the risks of easing restrictions too soon is probably greater than the risks of keeping them in place a little longer than necessary.
“measures will have to last be effective” What do you think effective is?
CV-19 isn’t going away. I there are around 1.5 million vulnerable (UK) people most of which will die from it without intensive care and probably about 1/5th will die regardless. Most of the rest of us are going to have to get infected and become immune. For most of us it will be like a dose of flu. Some of us will need intensive care to survive. A few of the rest of us will die regardless. The aim of the lockdown is to protect the vulnerable and limit the rate the most of us get infected so there is sufficient intensive care for those that need it to survive. Once we have got over any hump in the pipeline we should be aiming at a transmission rate of 1 so we have a constant number infected. As more become immune the transmission rate will drop and we should ease the lockdown to bring it back up towards 1 so we get though the pandemic asap.
Eventually there may be so little CV-19 kicking around that the vulnerable will dare leave home, if not they will have to wait for a vaccine.
How long? If we know what proportion of the most of us will need intensive care. If we know how many intensive care beds we have. If we know how many days it takes someone to go through infection and become immune. If we know what proportion of the population need to be immune to negate the need for any non-vunerable lockdown. If we know these things working out when the lockdown will end is easy, but, the lockdown will be easing throughout the process anyway.
The most important metric to monitor is the load on health services and the more intensive care places our health service can provide the faster we get through it. I can’t help thinking the most useful testing is going to be for immunity so people can know they don’t need to be locked down and get the economy moving again.
Agreed, testing seems to be key now.
Do you agree that only key workers should be allowed to leave their homes for work surely there are various safeguards that could be implemented to allow people to continue working where physical distancing can be achieved.
Is there a risk that the government scheme will incentivise employers to furlough staff unnecessarily.
If the deaths show no sign of levelling out then presumably the restrictions may have to be extended, though beyond the obvious categories it is hard to say who is a key worker and who isn’t.
The job retention scheme might create some perverse incentives, but I’m not sure why employers would want to furlough staff unnecessarily who would otherwise be fully occupied and making money for the firm.
It has been suggested that the scheme might encourage businesses to concentrate work among a small number of employees and furlough the rest, rather than sharing it more evenly. But that may be exactly what ‘social distancing’ requires. Firms would also struggle to continue to pay full wages to employees who are only occupied part time. It makes more sense to allow those who most need to stay at home to do so, with the government covering most of their wages until the crisis passes.
“At least, fewer people will be drinking themselves to an early death down the local pub, or losing their lives to traffic accidents.” – I, like many others, am drinking far more than normal, and my 77 year old mother, still a keen golfer, is enduring far more physical and emotional stress by being prevented from leaving her tiny flat and exercising in the fresh air – perhaps you could factor these numbers into the equation also?
Fair points, and we probably all know people in similar positions. (Hopefully your mum, like mine, is still going out for some exercise, even if not golf.) The problem is that it is almost impossible to factor these things into a simple equation. If we could, I’d also point to the distress that might be caused to the loved-ones of the hundreds and thousands of people who might die if restrictions were lifted too soon, and to those caring for them. There are lots of tragic choices here and no easy answers.
Also, drinking in a pub is drinking in a supervised, managed environment. Even before COVID-19, problem drinkers mostly chose the home (or park) environment to over-indulge rather than a public house. Nobody in Asda tells you that you’ve had enough. Now pubs are off-limits all the supervision is gone, and deaths will rise.
Fair point, and not the best example I could have chosen. I’ll tweak the text accordingly.
This is what concerns me most about the lookdown. I support measures that reduce the spread of covid-19 however where physical distancing can be achieved relatively easily e.g. a golf course these activities should be continued. Understandably restrictions on numbers may be required but clearly some people being able to play golf is better than none.
I guess the problem here (as we’ve seen with walkers in the Peak District) is not the golf (or the walking) itself, but the unnecessary journeys to get there, contact in car parks, changing rooms and club houses etc. Very hard to get the balance right.
“To be fair, his conclusion does not actually rely on the assertion that the tougher measures would only extend the lives of those saved by ‘one or two years’, which is obviously false. (Young has taken an out-of-date number for the average age of those who have died, and compared this to the average UK life expectancy at birth rather than at the age already reached.)”
Young’s rationale might be false, but the vast majority of deaths are from people with 1 or more comorbidities. I don’t think these people have the usual life expectancy for their age, and so Young’s estimate of 1-2 years sounds far closer to the truth than 11 years.
Perhaps (I’m not a health expert!) but the 11 years is a figure Young himself cites, from work by Noah Carl. Also worth bearing in mind the risk that the NHS gets so overwhelmed with coronavirus patients that it can’t deal properly with people with non-coronavirus related problems. No easy answers here!
“the vast majority of deaths are from people with 1 or more comorbidities”
Of course it kills the weak first because the not so weak are being saved by intensive care. The life expectancy of those it kills will be much higher if the health services become overwhelmed.