Streetwise Professor

March 19, 2020

Are We Destroying Society In Order to Save It?

Filed under: China,Economics,Politics — cpirrong @ 6:40 pm

In 1968, journalist Peter Arnett claimed that a U.S. major had told him that a particular village in Vietnam, Ben Tre, had to be extirpated: “It became necessary to destroy the town to save it” (from the Vietcong), sayeth the major (according to Arnett). This has entered American discourse as “we had to destroy the village to save it.”

That phrase came to mind when contemplating the havoc wreaked by the CCP Virus. Europe is shutting down, country by country. Parts of the US have shut down. Others are on the verge of shutting down. The economic carnage is immense. Governments talk of spending trillions of dollars in various forms of relief: the loss of output/income will probably be measured in trillions.

Contra Hayek, it is the curious task of an economist to ask whether it’s worth it. That is, economics is predicated on the concept of scarcity, which in turn implies that every choice involves a trade-off. You want more of a good–or in the present instance, less of a bad–you have to give up something.

What price are you willing to pay? How much is saving 1000 lives worth? 10,000?

Orders of magnitude. Let’s say that shutting down the US economy through radical social distancing, quarantines, etc., saves 1000 lives, and costs $1 trillion. That works out to $1 billion per life. Moreover, the lives saved are most likely aged, infirm, sick individuals with short life expectancies and poor life quality.

Is that a price you are willing to pay? There is no right answer: the answer is subjective. Your answer may differ from mine. But when making decisions, it is a question we have to answer.

Increase the death toll by 10, and you are still at $100 million/life. This is far beyond any value of life estimate used in other regulatory and policy decisions.

If the cost of an economic shutdown is $1 trillion, you would have to save on the order of 100,000 lives to approximate the value of a statistical life (around $10 million) the US government uses for other policy making purposes.

I know that most people recoil at such calculations. The idea of valuing lives in dollars violates most people’s moral intuitions.

So let’s focus on lives. A major recession–or depression, which is not inconceivable–costs lives. Suicide rates go up. Substance abuse goes up, which costs lives in the near term (overdoses, fatal vehicle accidents) and the long term (substance abuse shortens lives). Stress-related fatalities (heart attack, stroke) go up. Murder rates go up. Consumption of health care declines, leading to premature deaths.

And then we can start talking about quality of life.

Pretty soon it adds up. We are not just evaluating the trade-off of lives for money. We are evaluating the trade-off of lives for lives.

That is, always remember Bastiat: think of the unseen. There is an unseen public health cost associated with major economic dislocation. That unseen cost has to be weighed against the cost that is right in front of our faces at present, i.e., the death toll from CPCV-19/20.

It is of course difficult to estimate, or even approximate, the various costs. Our radical ignorance about the virus makes it difficult to assess what the death toll would be under various policies. Similarly, we are operating in completely unexplored territory in trying to estimate the economic cost, let alone the health cost, of more or less draconian restrictions on our lives and movement.

But we have to at least confront the trade-off. Acknowledge it. Grapple with it. My strong sense is that the monomaniacal focus on controlling spread of the virus, the costs be damned, is operating according to the logic of destroying society in order to save it. That logic was absurd in 1968. It is absurd in 2020.

Print Friendly, PDF & Email


  1. Did they influence us to hoard toilet paper so we would not hoard food and disrupt the supply chain?

    Comment by Summit NJ — March 19, 2020 @ 7:43 pm

  2. UK was probably the only major economy willing to listen to this argument, but they did an about face following a public outcry. It’s a democracy and people get scared and want the government to do something. Costs and benefits do not enter the equation.

    Comment by John Hall — March 19, 2020 @ 8:29 pm

  3. John Hall: the UK volte-face came after a public outcry, true – but also after their modelling team (UCL?) admitted they had gotten some of the parameters wrong, and the short term demand on health services would well exceed their surge capacity.
    Prof: Regulatory agencies like the EPA use $10M as the value of a human life because if you’re in the business of creating regulations to “protect human life” the higher that value, the more regulations you can justify – and that’s a nice measure of job security for an EPA employee. I can’t dig out the study, but some guys at Princeton (IIRC) reviewed how willing people were to undertake risky behaviour (high speeds on highways) versus the time-benefit of arriving sooner. They came out with the result that people implicitly valued their own lives at $2.5M to $3.0M – and I saw another one that gave a similar answer (but I can’t even recall the methodology, so take that for what it’s worth). Again, for what it’s worth, the average American carries less than half a million dollars in life insurance – which is at least one data point in establishing what a human life is worth when the holder of that life has to trade-off against it.

    Comment by dcardno — March 19, 2020 @ 9:25 pm

  4. Your argument is not on point. The exercise of flattening the curve is an attempt to preserve hospitals, not individuals. If all the people who will sicken, sicken at once, the medical system will be overwhelmed. You know, like a run on the banks. The effort is to stretch out the crisis over time to allow hospitals to still function, for these patients as well as for all other patients. The same numbers will die, but we hope they don’t take the medical staffs and facilities down with them!

    Comment by Sinyet — March 19, 2020 @ 10:04 pm

  5. Interesting idea to look the cost per death at different mortality rates for a trillion.
    Doesn’t seem to add up. But then I think about other big numbers like the Army. So, how much did we spent to get each individual ISIS member and what was their ‘potential damage’?
    This sort of exercise is fruitless. You can put a value on your family’s life, good for you, I cannot

    Comment by John Robert McFarlane — March 20, 2020 @ 1:12 am

  6. @ John Hall – the Netherlands too. Not sure if they too have changed tack.

    @ dcardno – Imperial College did the study. There has since been comments that it may have been unduly pessimistic.

    If Italy is anything to go by, you could be looking at 100s of thousands of deaths in the US & across Europe, more if there are multiple waves, reinfections, civil unrest* etc. Govts will probably have to throw everything they’ve got at it. If we do come out the other side with functioning economies then we can just add everything to our collective national debts and wait for inflation to do its work.

    Also, I’d love to be living in one of those technocratic autocracies at the moment. They’re the ones best placed to survive this mess.

    * Craig you’ll no doubt be pleased to hear I’ve been making inquiries about purchasing a firearm (talk about eating one’s words). Any recommendations? Something small and discreet or big and showy?

    Comment by David Mercer — March 20, 2020 @ 3:33 am

  7. Something that would make the decision making much easier is that random sample of the asymptomatic population that the prof. wrote about earlier. If the virus is scarce in that sample then the measures we are taking might very well make sense but if the virus is already present in large numbers then we are indeed destroying the economy for nothing.

    Comment by Andrew Stanton — March 20, 2020 @ 8:08 am

  8. Matt Ridley observed the other day that Western societies struggle with preparing for an event that has a low probability of occurring, but has a high risk of catastrophe if it does.

    Here, exhaustion of emergency health care resources is the main risk of catastrophe. Absent better information on the true risk of severe infection that would require hospitalization (a result of the delay in having sufficient testing), overreacting to preserve precious health care resources is the proper response.

    Nassim Taleb makes the point that insurance is cheaper than catastrophe, which is why under convexity it is better to act early, quickly, and cheaply. It’s pretty clear that we would be better off economically if the measures to stem the pandemic over the past week had been taken in late January.

    That said, I’m still cautiously optimistic. Using number of deaths per day, the US is well behind Italy.

    The following is Italy’s death count from 1 to 200:

    1 2 3 7 11 12 17 21 29 41 52 79 107 148 197 (15 days – Feb 21 to March 6)

    The following is the US death count from 1 to 200:

    1 6 9 11 12 15 19 22 26 30 38 41 49 57 68 86 109 125 154 206 (20 days – Feb 29-March 20)

    So, it took Italy 15 days to get to 200 deaths, while it took the US 20 days to get to 206.

    US deaths per active cases (with the exception of one spike) is lower than Italy’s and declining. Italy’s deaths per active cases continue to increase. Inasmuch as Italy is currently testing more than the US, that’s another encouraging signal.

    Comment by Tom Kirkendall — March 20, 2020 @ 12:35 pm

  9. Re. Craig: I think it would be more resonable to assume a range between 100,00 to 1M expected deaths and therefore a unit ‘cost’ of 1 to 10M dollar per person.
    Re. Tom: Go to:
    Scroll down until Section “La Traiettoria degli Altri Paesi”.
    Select “Italia” and “USA”.
    You may see that US contagion is growing at a faster pace than Italy’s.

    Comment by Andrew — March 20, 2020 @ 1:06 pm

  10. The unseen (h/t Bastiat) may be favourable. I’m in the vulnerable cohort so I don’t want to shout this from the rooftops, but…
    Influenza used to be called the “old man’s friend”. This is a form of influenza. Generally it is only fatal in the old and the infirm.
    Suppose all the over 80s die. The social care conundrum solves itself.
    Additionally the virus hits hardest on those too stupid to take elementary precautions e.g. washing hands. Award them a Darwin prize.
    Suppose the over 70s die. Here you save on pension provision and, crucially, inherited wealth goes to a younger generation earlier.
    If the mortality of 70s and 80s is high, property investment looks a mug’s game for demographic reasons.
    What to do with the inherited wealth? Start up businesses.
    Start ups generally fail but the usual law still applies, one third will succeed. So once the plague is over, you will experience an economic boom.

    Comment by philip — March 20, 2020 @ 2:13 pm

  11. “Your argument is not on point. The exercise of flattening the curve is an attempt to preserve hospitals, not individuals.”

    Flattening the curve is nice in theory, but when you think about it, it is really trivially true. Of *course* slowing the spread of a virus lowers the rate it spreads. I mean, it’s really a tautology.

    The only way that curve is meaningful is if scales are applied to each axis, and solid justification is given for them. But the y-axis is based on terrible data and models that are as atrocious as our weather modeling. So that’s not going to give an accurate number. What is the high peak, really? Don’t say “it’s the worst case”, because the worst case is everyone gets infected and reinfected by this until we all die.

    And the x-axis has never been defined by anyone. When does the low curve peak? Is it next week? Is it in six years? Because closing down everything for a couple weeks, even if drastically overkill, will still be pretty mild on the world economy. But six years? Try World War 3 and billions dead.

    And over what? Again, we don’t know the differences in the peaks, but either way, the conventional wisdom (we know how reliable that is) is that the area under both curves is the same. That means “flatten the curve” is really only a game of keeping medical demand under medical supply. But that has costs, too. My own governor, bless her heart, has said that people should avoid getting “non-critical” health care right now. There’s a lot of non-critical health care that prevents critical health care. When does THAT mortality curve get placed on this “flatten the curve” chart?

    All this to say that “flatten the curve” on its own is dumb.

    Comment by Joshua Postema — March 20, 2020 @ 4:17 pm

  12. @Sineyt. You indicate a potential benefit. You don’t address the cost. Thus, you miss the point.

    Comment by cpirrong — March 20, 2020 @ 5:18 pm

  13. @Joshua Postema. It’s become a mantra. Those are usually a (poor) substitute for thought.

    Comment by cpirrong — March 20, 2020 @ 5:18 pm

  14. @Andrew–The figures are all over the map, in part for reasons I’ve discussed in other posts. The numbers you cite are at the extreme, extreme end of the distribution.

    Given the variation in the estimates, it boils down to risk aversion. If you are averse to risk, overweight the bad outcomes.

    But . . . we have to consider risks in all dimensions, including the risks associated with shutting down the world economy. Those are not trivial.

    It’s not an easy call, and I’m glad I don’t have to make it. The point of the post was to emphasize that there are trade-offs, and heretofore the debate has largely ignored them.

    Comment by cpirrong — March 20, 2020 @ 5:23 pm

  15. @Tom–And when you consider that the US population is about 5.5x Italy’s . . .

    Comment by cpirrong — March 20, 2020 @ 5:24 pm

  16. “The same numbers will die, but we hope they don’t take the medical staffs and facilities down with them”

    This seems to be the underlying assumption in all those “explanations” in the media, except I have never seen anyone explain where it comes from.

    What if actually many more people will die because of the shutdown of the society, not just the economy? How many more people die because nobody is around to rescue / call 911 when something happens to them? Because they cannot take that flight to visit Baden-Baden? Etc, etc.

    You cannot train new medical staff very quickly, but like in any economic endeavor, you can always substitute capital for labour. What is clear is that much more elaborate modeling is needed than two curves, but the governments do their usual “we have to do SOMETHING” thing.

    Comment by Ivan — March 20, 2020 @ 11:59 pm

  17. One of the latest tidbits floating around the net is that about 7 million Chinese have recently either closed, or had closed for them, their mobile phone accounts.

    You could say that the fall in economic activity meant that 7 million people could no longer afford to pay their mobile bills.

    The more sinister inference is …

    And if the mortality rate is that high, then that may change the cost-benefit analysis to the extent that maybe shutting down a good portion of the economy for a few months, to allow authorities to get a handle on the virus, is a good thing.

    But no one in power has asked me for my opinion, so I’ll just stick to working from home, surfing the net, washing my hands for thirty seconds each time and gargling warm salty water.

    Comment by Ex-Global Super-Regulator on Lunch Break — March 21, 2020 @ 12:18 am

  18. @Andrew Here are deaths weighted by the total number of cases as of March 20th, 2020 from

    China 8.2%
    Italy 8.57%
    Spain 5.07%
    Germany 0.34%
    U.S. 1.34%
    Iran 7.29%
    France 3.57%
    S. Korea 1.16%

    Ranked by the total number of cases, our death rate is closer to South Korea’s than Italy’s.

    Comment by Tom Kirkendall — March 21, 2020 @ 5:33 am

  19. If it’s not pretty lethal why did the Chinese despots close their economy down? I assume that they thought it was lethal. Were they wrong?

    If it’s not pretty contagious how did so many people on the Diamond Princess catch it? Was the quarantining aboard slapdash? (Surely not as slapdash as the so-called quarantining that the evacuated American passengers were then subjected to? Dear God, federal officials seems to be just awful at their jobs.)

    Is it true, as claimed, that only 5% of the Wuhan patients sent to the ICU survived?

    If that were to prove representative, would there be any point in moving heaven and earth to avoid the ICUs being overwhelmed?

    OK, I accept that our data is biased and incomplete, that our knowledge is “radically uncertain”. But I’d still like to see some sort of coherent attempt at a comparison of predicted lives shortened by virus (weighted by quality of life, if that’s possible) versus ditto by draconian social and economic measures.

    At the moment we seem to be combining “let’s panic now” with “let’s hope something will turn up”. Maybe that’s the best that’s available as a matter of politics in a liberal democracy. Maybe it might even work out not too badly. Maybe.

    Comment by dearieme — March 21, 2020 @ 6:56 am

  20. Oh bugger! “… data are …”

    Comment by dearieme — March 21, 2020 @ 6:58 am

  21. @cpirrong: Thanks post, thought-provoking as always. I suspect the average man will be more receptive to such utilitarist calculation if his family (the relevant economic unit for most people) can cash in a significant part of the 10 million USD you mentioned as the arbitrary value of a statistical life. But then.. moral hazard.

    Comment by Tarik — March 21, 2020 @ 7:44 am

  22. Re: The Diamond Princess:
    The Diamond Princess was a virtual Petrie dish for this virus. Before they knew that the virus was a problem, the epidemic raged on the ship, with infected crew members cooking and cleaning for the guests, people all eating together, close living quarters, lots of social interaction, and a generally older population. Seems like a perfect situation for an overwhelming majority of the passengers to become infected.
    And despite that, some 83% (82.7% – 83.9%) of the passengers never got the disease at all…why?
    Slightly less than half the passengers (48.6% ± 2.0%) who got the disease showed NO symptoms. If this disease is so dangerous, how come half the people who got it showed no symptoms at all?
    Reference: “Estimating the infection and case fatality ratio for COVID-19 using age-adjusted data from the outbreak on the Diamond Princess cruise ship” which is a PDF you can find here:

    Comment by Richard Whitney — March 21, 2020 @ 3:31 pm

  23. Thank you all for this interesting discussion.

    Re. Craig: I share your view about the opportunity to consider risk trade-offs. Possibly, within a holistic approach, where various perspectives are jointly taken into account. It seems to me clear that most governments around the world lack knowledge, view, and focus.

    Re. Tom: I was more interested about rates of growth instead of absolute figures. These latter might not be so meaningful as we think. In Italy, both numerator and denominator in the death rate are biased. Numerator includes anyone died WITH (not necessarily OF) Covir-19. As long as the vast majority of dead have 3 or more standing diseases, Covir-19 is likely to be a triggering factor for death, not the actual main cause. Denominator comprises those who have undergone a test through a swab. Now, to have a swab done, you must be in critical conditions. As long as numerator is overestimated and denominator is underestimated, the ratio is incredibly high compared to the actual figure. Germany stays on the other side of the story: they claim to have 84(!) dead over about 22500 cases, lower than Singapore … hard to believe (I do not know why, but this reminds me Dieselgate, Deutsche bank case, and other amenities of the kind). However, rates of growth might still be comparable.

    Comment by Andrew — March 21, 2020 @ 4:20 pm

RSS feed for comments on this post. TrackBack URI

Leave a comment

Powered by WordPress