Its hard to know where to start with arguments like this. But its important to try, because they need combatting. So I'd like to propose a crowd-fisking effort. Please send me examples of mistakes of fact or logic in Young's argument. I'll start with a few which immediately occur
1. Young's assumption that the mortality from COVID-19 is a fixed figure which isn't influenced by the mitigation actions we take. This is obviously not correct. If NHS facilities are overwhelmed, then the mortality goes up.
If cases are spread out, more of the serious ones can access ICU, or ventilators, or regular doctor and nursing assessments. If cases are more bunched together, more patients won't and therefore more patients will die. This doesn't even occur to Toby who just assumes a fixed
death rate which he has copied and pasted from a single epidemiological study.
2. Young's death figure assumes that the death rate from COVID-19 won't change over time. Again, fairly obvious problems with this. Numerous drugs are being trialled now. As are dozens of vaccines. If serious cases arise later, there's a good chance our treatments are better.
3. Young's figures focus only on *direct mortality from COVID-19*. That ignores (a) the extra deaths which will occur from *all the other diseases* because *our healthcare system can't treat them if it is overwhelmed by COVID-19.
If you get *any disease at all* that requires intensive care at a time when *there are no ICU beds available* that's not a great outcome for you. That's a lot of illness, including a lot of illness among younger people.
(b) Young's figure also focuses only on mortality and its avoidance. He doesn't account for the much larger number of people who will need critical care, often for many weeks, but don't die. Failing to take mitigation dramatically increases their numbers.
5. Young assumes that the vast majority of the costs of COVID-19 is within govt's discretionary control. This is absurd. Much of this cost - probably most - can't be avoided regardless of the strategy we take. And some costs would be *higher* if we end lockdown early.
The costs of caring for COVID-19 patients evidently has to be paid regardless - and would be higher without mitigation. Industries which depend on travel between countries will be shut down regardless of what we do for as long as international travel remains heavily restricted.
Consumers and businesses won't return to normal behaviour once the formal lockdown ends, so there will be a large downturn in economic activity regardless. So the implicit framing of £200 Bn vs £0 is absurd.
Young makes no effort to break down the costs or calculate what share the govt actually relates to the lockdown. This way he has the largest possible number for making his polemical claim that the cost is illogical. Yet logic would demand a better demonstration of true cost.
While we're on the subject of cost, Young provides no source for his £350 Bn price tag. He calls it "spending" so I assume he means extra govt expenditure. Yet this most recent IFS estimate puts the price tag at £120 Bn, one third of Young's estimate: ifs.org.uk/publications/1…
That's my initial list - please keep adding further counter-arguments in replies! The more holes in this daft arg we can identify, the less credible it will be.
Ryan Bourne offers some more criticisms here:
It seems both the cost and benefit sides of Young's equation are massively wrong:
Cost - he says £350Bn or "generously" £185Bn. IFS says £50Bn (additional spending) or £120 Bn (additional borrowing)
Benefit - he says price of quality adjusted life year (QALY) is £30k. Its £60k
Cost - he says £350Bn or "generously" £185Bn. IFS says £50Bn (additional spending) or £120 Bn (additional borrowing)
Benefit - he says price of quality adjusted life year (QALY) is £30k. Its £60k
Benefit (2) He says saving the life of an 80 year old yields one extra year of life - but that's based on life expectancy *at birth*. Life expectancy of people who are *already 80* is 9-10 years, per several replies on this thread.
Combining these errors, Toby Young is out by a factor of at least 1.5-7 times on cost (as this assumes we keep his obviously silly assumption that all costs could be avoided) and out by a factor of *20 TIMES* in benefit (ten times as many life-years and twice the benefit per yr)
If the costs of mitigation are closer to £50Bn than £200Bn, and if the benefits of mitigation are £60k*10 years instead of £30k*1-2 years, then the cost-benefit analysis looks awfully different. And that's without questioning any of the *other* shonky assumptions.
A few more points on costs. Lets' imagine the (hopefully) fanciful counterfactual world where the govt listens to Toby Young and the lockdown is ended immediately after Easter. Are the costs of COVID-19 then fully prevented? Obviously not.
Firstly, there are the costs *already incurred* by the measures taken before Easter. Those aren't changed, and they will be substantial.
Secondly, there are the costs which will be incurred by combatting the epidemic under any particular framework. These will likely be higher.
Secondly, there are the costs which will be incurred by combatting the epidemic under any particular framework. These will likely be higher.
If the lockdown if lifted earlier, before the epidemic is fully stamped out, it will start spreading again. Which means more ICU beds, ventilators, doctors, nurses, not to mention huge numbers of man hours lost to sickness.
Thirdly, there are the costs which will be incurred by behaviour changes people implement regardless of what the government does. If pubs, clubs, cinemas, hotels etc opened immediately after Easter, I dare say numbers going would be quite a bit down on the no-COVID baseline.
People, rather obviously, are not going to behave as if a once in a century health crisis has entirely stopped simply because the lockdown ends. Plenty of avoidance behaviour will continue, and therefore the economic costs associated with it will continue.
Fourth, there are the costs associated with what other countries do. They won't life lockdowns early just because we do, and might indeed impose extra restrictions on us *because* we lift our lockdown early.
These are just the first fixed costs which occur to me casually. They're already more than sufficient to make it abundently obvious that the baseline for comparison on the lockdown is not "business as usual".
If Young was/is really serious about having a cost-benefit debate he should present a calculation of the *net added cost of the lockdown* compared to the *net added cost* in other scenarios. Not just wave around a big number which includes a huge element of sunk cost.
(note - on point 3 above "stamped out" is a poor choice of phrase. What I meant to say is until we reduce cases and increase resources to the point where proper testing and tracing is feasible meaning we can keep caseload manageable/avoid exponential growth)
OK, a couple more sets of objections have occurred to me. Firstly, lets think a bit more closely about QALYs - Quality Adjusted Life Years. Young assumes going rate for one QALY - one year of top quality livin' - is £30k. Fair enough. But he underestimates the QALYs COVID costs.
Young assumes the cost of COVID comes entirely through the QALYs lost by those who die. I.e. the years they would have otherwise been alive. But that's obviously incorrect, overlooking at least four sources of QALY cost:
1. The QALY cost of the illness/hospital time of those who get sick but recover. A mass epidemic means millions of days of illness, thousands upon thousands of days spent in ICUs, on intubation, unable to eat, unable to sleep, unable to breathe. Those are a QALY cost.
2. The QALY cost of illness/hospital time for those who die. COVID death is slow. Many weeks in isolation, tended by people in space suits, unable to see or speak to families. These grim months of suffering need inclusion in Young's calculus too .
3. The QALY cost of long term effects of COVID on those who recover. Both direct effects, which are by definition unknown at present but unlikely to be zero but also side effects of medical interventions - medicines, intubation surgeries, damage to organs etc.
4. The QALY cost of the mental health impacts of all this suffering and death on those who recover, relatives of those who die, and health care providers. Such things deserve an entry in the accounts also.
Again, those are just the ones I've thought of, I'm sure followers can think of more. On to the second new objection I have - Young is arguably committing a *category error*, applying a utilitarian cost/benefit caclulus in an area where it is not appropriate.
To illustrate what I mean by this, consider organ transplant. There is a chronic shortage of organs. We could alleviate this shortage by killing off a number of people society considers of low worth (long term prisoners, say) and harvesting their organs. Obviously, we don't.
Why don't we? Because, in general, society does not consider cost/benefit analysis to be the sole criterion for judging social rules. We don't max out the QALYs by killing off prisoners for their organs because organ harvesting is, by general consensus, *unacceptable*.
The NICE criteria Young deploys are, some repliers have pointed out, *NOT* used for judging life saving interventions. Most likely due to moral/ethical constraints such as the one I illustrated. COVID lockdowns are, clearly, a life saving intervention.
Ergo, it is not appropriate to judge them using utilitarian criteria im the way Young does. It just isn't acceptable for a government to let people die in large numbers who would otherwise not die, whatever the QALY calculus might say.
This is a different category of objection to my other ones. Rather than point out that every number he uses is probably wrong (and every one does seem likely to be wrong), this is highlighting that the calculation is simply not a legitimate one to make.
Some further well made economic criticisms from @jdportes here:
And some other related economic args in favour of shutdown from @Noahpinion here:
@Noahpinion A good example of how to think properly about cost-benefit analysis, which Young would have benefitted from reading and thinking about, is available here (h/t @jdportes )
cato.org/blog/time-move…
cato.org/blog/time-move…
@Noahpinion @jdportes In particular, Young's analysis fails on the following of the Cato paper criteria:
1. Doesn't compare realistic scenarios
3. Doesn't vary price put on lives saved properly (uses single price)
4. Doesn't produce comprehensive measures of costs or benefits
...
1. Doesn't compare realistic scenarios
3. Doesn't vary price put on lives saved properly (uses single price)
4. Doesn't produce comprehensive measures of costs or benefits
...
@Noahpinion @jdportes 5. Doesn't account for behaviour changes which will happen regardless of policy
6. Doesn't account post-crisis rebounds in GDP (i.e. most of the GDP lost is likely deferred, not lost *forever*)
7. Not comparing apples with apples
6. Doesn't account post-crisis rebounds in GDP (i.e. most of the GDP lost is likely deferred, not lost *forever*)
7. Not comparing apples with apples
@Noahpinion @jdportes Another point made in a few comments now is we also need to net out from lockdown cost benefits produced by the lockdown - deaths and injuries avoided (fewer accidents); greenhouse gas emissions reductions. These will likely be only a fraction of the costs, but not insignificant
Another carefully argued response to Young here:
Some useful detail here on the different thinking behind the NICE £30k per QALY and Green Book £60k per QALY from @PadraigDixon
@PadraigDixon TL;DR - the £60k Green Book figure is based on thinking closer to what Young is trying to do, but Young uses the £30k NICE figure (which he misinterprets)
Aside from all the problems with Young’s cost-benefit analysis, there is a whole other line of criticism to pursue - namely that his call to relax restrictions immediately after Easter shows reckless ignorance of COVID’s epidemiology and disease characteristics
Take for example this observation from today’s COVID press conference in Italy - average ICU stays for serious cases are *2-4 weeks*.
That means even once you have case loads under control there will be a lag of several weeks before burdens on the NHS begin to relax. Even the most optimistic of the analyses I’ve seen would suggest flattening out just underway by Easter.
Therefore at least 2-4 weeks further restrictions would be needed to clear out the accumulated serious cases. In other words, relaxing at Easter would risk unleashing a second wave of COVID on an NHS still dealing with the peak of the first wave
There may be a cost-benefit case for relaxation later (though such a case needs to address the objections up thread). But there quite obviously is not one for relaxing in just two weeks’ time.
