Should South Africa lock down?

The lockdown will lead to 29 times more lives lost than the harm it seeks to prevent from Covid-19 in SA, according to a conservative estimate contained in a new model developed by local actuaries.

The model, which will be made public today for debate, was developed by a consortium calling itself Panda (Pandemic ~ Data Analysis), which includes four actuaries, an economist and a doctor, while the work was checked by lawyers and mathematicians. The process was led by two fellows at the Actuarial Society of SA, Peter Castleden and Nick Hudson.

They have sent a letter, explaining its model, to President Cyril Ramaphosa. In the letter, headed “Lockdown is a humanitarian disaster to dwarf Covid-19”, they call for an end to the lockdown, a focus on isolating the elderly and allowing children to go back to school, while ensuring the economy restarts so that lives can be saved.

The paper also is at the link, and it is perhaps more of a rough and ready calculation than a formal model per se.  Nonetheless South Africa has a relatively young population and the core points are well taken:

In SA, they estimate that 5.4 years of life have been lost per Covid-19 death. They then multiply this by the range of deaths which they predict – 20,000 – as well as the actuarial society’s prediction of 88,000 fatalities. They factor in that the lockdown will have reduced some deaths, but not all. In the end, their model translated into a minimum of 26,800 “years of lives lost” due to Covid-19, and a maximum of 473,500 years. (This, critically, shouldn’t be confused with the actual number of fatalities expected from Covid-19.)

The actuaries then used the figures predicted by the National Treasury to model the impact on poverty. On Friday, the Treasury estimated that between 3-million and 7-million jobs will be lost due to the measures taken to combat the virus. The actuaries then work out that, conservatively, 10% of South Africans will become poorer, and as a result, will lose a few months of their lives.

It is a good question how many of the models used for the West have taken into account the “demonstration effect,” namely that poorer (and much younger) countries will be tempted to follow the same policies.  I’ve yet to see a good discussion of this.


It sort of is not a good question, because you know the answer is ZERO.
well maybe it is a good question among those categories of questions to which one knows the answer.
Miss Marple and Poirot were specialists at that .....

I'm sure the lawyers that reviewed this project can provide you with the answers you seek.

I've wondered the same thing: why poorer countries have leapt into adopting the Western lockdown model. India especially, where the loss to life and well being of sliding back to more poverty are potentially staggering. I alternate between thinking that (i) Covid-19 a few months ago seemed far worse than it is, (ii) it's another instance of the Indian elites mimicking Western trends, or (iii) it's a way for the authoritarian-tending ruling party to become more authoritarian.
Returning to the post: it would be good to see more examples of non-rich/Western perspectives.

Poorer countries have less to lose from a lockdown. People and businesses tend to be less leveraged. There’s less value to be destroyed overall. Also, if just .1% of India’s population dies, that’s ~1.3 million people. India’s healthcare infrastructure can’t handle even a modest outbreak. People will die who otherwise would have survived with better healthcare. Recent headlines downplaying the effect of ventilators are misleading - they still save an appreciable number of lives, and that’s just one resource that wouldn’t be available to many Indian victims of a COVID-19 epidemic. Also consider that, because of the population density and sanitary conditions, the average patient with COVID-19 in India will have been exposed to a higher more fatal viral load than the average American patient.

Poorer people in poorer countries have everything to lose in a lockdown. If you only have a little, then losing a little means losing everything. The safety net barely exists. If you don't work today you don't eat today.

Over here we can just sit at home for a few months and collect government bailout assistance. In India, the only choice for many was mass migration out of major cities and walking hundreds of miles on foot to rural villages.

It’s a rich world’s disease. If you’re poor in a poor country, you have bigger problems like making your $ 2 a day so you can eat.
And what social distancing are you gonna do in the townships, the favelas or the Makoko slum of Lagos?

Certainly, this list represents the cream of the GRE crop - four actuaries, an economist and a doctor, while the work was checked by lawyers and mathematicians.

An American law school professor who predicted America 5000 coronavirus deaths is currently off by more than an order of magnitude, but he was working alone.

The real question is why the Nigerian response to ebola did not lead to a demonstration effect in the west, where in many countries, the ability to trace and isolate is clearly below the level of a corrupt 3td world oil exporter.

Overdue analysis for every country. Living years lost is an improved measurement on the toll the virus is taking compared to total deaths. No surprise policy doesn’t follow though since number of deaths is more closely correlated to number of voters lost.

I think the number I’ve read for excess deaths caused by the Great Recession was 8%. We are truly stuck in “we don’t know everything” paralysis, and lockdown was safest political answer.

We have a statistic called Quality-Adjusted Life Years for measuring this kind of thing. For some reason, I haven't seen anybody yet try to even estimate the impact in QALY terms.

Even after such bungled handling, the arrogance is breathtaking. India has plenty of epidemiologist, actuaries, economists even to build their own models and take decisions accordingly. Even people in poorer countries love their older population.

No or they will never develop herd immunity.

Every time someone writes herd immunity without including endemic or vaccination, take a drink - of something appropriately non-alcoholic, since the befuddlement is already obvious.

South Africa
Average age: 27.6
Population over 60: 8%
Per capita GDP: approximately $6,400

"Over 60" has a different health meaning in poorer countries.

Eh. As I've long said, obvious even to a five-ten minute calculation from age distribution of deaths and fatality rates by age (whether well studied Chinese returnees to Wuhan or mass seralogical IFR from the West, adjusted by demography of recorded fatalities:cases).

'Premature imitation' to take Tabarrok's term, for sure.

The only caveat for SA is if their stable HIV pop really takes a beating from this. But for other young regions without that problem, yeah, come on.

It's one thing for the "economic harms mean no health harms" to propose that for rich countries. Sure they're probably being Clever Sillies who point to probably spurious recession mortality studies ("Mortality didn't go up during the Great Depression!"), but it's plausible that only extreme poverty matters for health. But it's another thing to propose that extreme poverty doesn't matter for health in poor countries, and that lockdown wouldn't create more extreme poverty for 'em.

To add onto this, I'd love to see the intersections between the "Recessions are totally fine and have no impact on health and life expectancy" crowd today, and the crowd on the left who screamed for about ten years that austerity deepens recessions and this leads to "excess deaths".

Either recessions in rich countries are harmful for health or they're not, but it can't be that they both are when governments are practicing very limited spending restraint (and so at worst mildly slowing recoveries), whole having no effects at all during an unprecedented act of government mandated self imposed reduction in economic activity.

South Africa's population: 59 million
It's per capita GDP is close to half that of China, a middle income nation that has successfully contained the virus and is eliminating it.

South Africa's new daily infection rate is about half Australia's at its peak. If level 3 restrictions, which are short of a lockdown, are as effective in South Africa as they have been in Australia then perhaps the virus could be eliminated from the general population in two months. However, despite geographical similarity, conditions are not the same, so that can't be counted on.

5 weeks of Level 5 restrictions in South Africa have not caused a decreased in cases.
If the lockdown was eliminating the virus then it would be worth it, but its not working.

Should South Africa lock down?

None of us know.

No country can justify lengthy lockdowns based on the known death rate alone. 1% IFR * (5 - 10 life years lost per death) = ~1 month lost per person on average, in the unlikely event that every person in the country gets the disease.

You can't do indefinite involuntary lockdown and gamble with economic collapse, just to save 1 month of life per person. It would be crazy and intensely unethical.

Therefore, almost the *entire* justification for lockdowns comes down to doomsday scenarios where the virus turns out to be far more damaging than we think, for instance if it turns out to be more lethal the second time you catch it. (There's a mechanism for this called antibody-dependent enhancement that occurs with some diseases, notably dengue.)

How likely are these doomsday scenarios? That's the hundred trillion dollar question. I'd guess it's hard to get the math to work out so the lockdowns are worth it. But truthfully I don't know, and I am a little frightened at some of the possibilities.

More discussion on this point would be welcomed.

SA has a very high unemployment rate (over 50% if those who are not looking for work are included), with blacks/coloureds making up the bulk of the unemployed. And the level of inequality as between whites and blacks/coloureds is staggeringly high. Not surprisingly, crime is high and so is brain drain (mostly whites). An economic lock-down will hit whites economically more than blacks/coloureds for the simple reason that whites have a lot more to lose; indeed, the majority of blacks/coloureds have little to lose. On the other hand, I understand the point of this study and the reasons for the recommendation to lift the lock-down: because blacks/coloureds have so little, they are at high risk of starvation if they lose the little that they have. That's the reality. It's a perverse way of looking at the consequences of an economic lock-down, but SA is perverse. By way of comparison, the rush to lift the lock-down in the U.S. is being led by a similar analysis: who suffers more or less from lifting the lock-down. That's the reality. Morality left the barn long ago.

Yes and EU and WHO are conditioning help packages on implementing a lockdown. The arrogance of EU and western elites has no limits.

If true, beyond even the immediate consequences, giving a bunch of mostly unstable regimes a huge financial incentive to enact legislation giving them the powers to place the population under house arrest is a fascinating example of handing them exactly the rope to hang themselves that they don't need.

This crisis will give us a better sense of who is actually against authoritarian rule under all circumstances, if anyone (real, consistent, any-and-all-circumstances "anti-authoritarians"). We will also see who opposes it only when it represents power for traditional institutions, males, natives, blue collar workers, small businessmen, and the army (Tyler's "brutes" and Clinton's "deplorables") and enthusiastically endorse authoritarian powers when placed in the hands of the university, scientists, engineers and doctors (backed of course, by the army and police), to do "heroic materialism".

Different anonymous here.

I wonder how skeptical I should be of "deaths due to lockdown?" I don't think anyone has shown me a data set. They are theoretical at this point, right? Which means that you have a modeled root cause (for instance staying away from doctors). Which means you can tackle the cause (with promotion of telemedicine as a first screening?)

In general I just trust the framework that we can't do something about COVID-19, because we can't do anything about the associated harms.

Let's just do things!

s/just trust/distrust/g

IMO this is a different argument from the one that "in a developing nation COVID-19 might never become a leading cause of death, so keep your eye on the ball."

Just treat it with the relative seriousness it deserves.

Certainly you should be skeptical, but you should be skeptical too of the deaths-without-lockdown models, which have shown themselves also to be "more of a rough and ready calculation than a formal model per se."

I don't need theory for bodies piled up in UHaul trailers.

Video widely available, but I don't think anyone should actually watch it.

I wonder how skeptical I should be of "deaths due to lockdown?" I don't think anyone has shown me a data set. They are theoretical at this point, right? Which means that you have a modeled root cause (for instance staying away from doctors). Which means you can tackle the cause (with promotion of telemedicine as a first screening?)

I had someone on Facebook telling me that a friend of hers just died from an appendicitis, the person was avoiding going to the doctor/hospital and telemedicine didn't catch it.

All well and good but not having a lockdown and filling bars, tattoo parlors and beauty salons wouldn't make me feel any safer going to the hospital if they were filled with Covid patients dying, out of sight.

"Certainly you should be skeptical, but you should be skeptical too of the deaths-without-lockdown models, which have shown themselves also to be "more of a rough and ready calculation than a formal model per se.""

Indeed but such models are not needed. The answer to "what would happen if we did nothing and let the virus just do its thing" is a real lot of people would die. If I'm going to push you out of an airplane and I give you the choice of a parachute or some time to punch numbers into a very formal model to see where and how you'll hit the should choose the parachute.

This is why we need to understand if hospitals are actually "overrun" under non lockdown distancing. Sweden (and even worst case in NY) suggests possibly not really.

If they're not overrun, then there would be hospital capacity.

Depending on definitions of overrun. A definition that's like "if non-zero people in hospital have covid, I won't even go there for cancer treatment or a measles jab" / "I won't go to a hospital unless the disease is eliminated" is probably too high a bar.

If Sweden has a higher in-hospital death rate then Denmark then perhaps they are stressed.

It's not like "overrun" is the only bad case.

Maybe, but overrun is not exactly a hidden variable we infer by death rates.

Just ask them how many beds, ICU, overtime, etc. If they say nothing too high, not overrun, not stressed and death rates are probably just low denominator or demographic diffs.

Remember the old joke about the economist looking for his keys under the street lamp. Did you lose them here? No. Well why are you looking here? Errr duh, the lights here.

I've read the portion of infected that needs hospitalization is roughly 10% with 6% just needing hospitalization and 4% needing a ventilator. Knowing we're not going to run out of ventilators is somewhat comforting. I'm not eager to see what's its like to go on one. Nor am I eager to discover 5, 10, 15 years from now having this virus offers surprise secondary problems.

I would avoid a hospital unless I felt I really had to go. I would also avoid crowds and businesses even if the economy 'reopened'. Maybe not 100% but drastically less than anything I did before.

I have no particular reason to be this cautious. I'm 46. I'm reasonably healthy with no known medical problems. I walk several miles a day. I could probably lose 10 pounds but I'm 30 pounds lighter than I used to be years ago. When I was caring for my wife, who passed from cancer, I would have been much more paranoid about infection control but I'm about 90% right now. I speak to other people who are following the same protocol that I am. Interestingly even people who are younger, very fit, who go even further than I do. No grocery store, everything delivered. Boxes/bags wiped down before being stocked in the house. These people, I note, are professional, high earners and pretty smart.

Smart money is not going to abide by a decision to open prematurely. People who aren't quite as smart but still bright follow what the smart money does as well.

Re; rates of hospitalization, I suspect those are "case" rates, and probably quite high ones, not "infection" rates.

Of course individuals should make decisions for themselves on whether they feel they need to go or not, but not without consultation with a doctor and with realistic medical advice on the prevalence, hospitalization and fatality rates, and without realistic advice on probability of infection from attending an appointment. And ideally with information about whether they already have some degree of immunity from a weak infection already.

Re; "elite emulation", to take the example further, in this thing, billionaires may ensconce themselves on private islands with private medical staff. But for the rest of society to blindly following their example as closely as possible within more limited resources, even if that seems to be following smart money (they're smart, they have money), is probably a worse idea than realistically trying to assess risk and your means yourself.

Assessing risk is pointless. Today a story came over that 5 kids who came down with Kawasaki disease in NJ. This disease has an incidence of about 19 per 100,000 in the US. Fluke? Perhaps but the kids did have Covid and supposedly the virus almost never harms children. Do you have any kids 5 or under? Shall we try giving them the virus to see if there's a link?

The lesson here isn't that we should assess risk for ourselves. If you think you're assessing risk you're being conned. We have no real experience with this virus and the idea that we should just let the bulk of the population get it for herd immunity is nothing more than experimenting without informed consent on hundreds of millions of people all because we can't figure out how to give the guy who works at Starbucks $2k per month for 3 months. This is the mentality of a society that is flushing itself down the drain.

Anyway my impression is not that the bottom is mindlessly emulating the elite. They do not have private islands to hang out at. They are, however, pulling back and distancing as much as they can and they will double down if a panic is set off with cases and deaths exploding. Or even more if people who thought they passed the virus without incident suddenly discover strange things are happening.

"Re; rates of hospitalization, I suspect those are "case" rates, and probably quite high ones, not "infection" rates." Possibly. I remember Stephen Jay Gould had a piece on being diagnosed with cancer that had a very short time until death. He noted, however, that figure was just an average of all patients with that type of cancer. He was younger than was typical, caught it at an earlier stage than was typical, etc. He expected to live longer than the average and he did. But then he died of it.

Years later I had the same experience with my wife. Well the average is 5 years but many women are diagnosed later in life....that average includes women who reject traditional therapies and try quack ones. etc. etc. She got a little over the 5 year mark. Be careful of injecting too much wishful thinking into the figures.

So good to see this. I'm sure there's plenty to quibble about in the analysis but at least they did the analysis and went about it honestly. In the US, this decision has been presented as a trolley problem where we throw the switch to minimize Covid deaths but pretend that the only thing on the other track is "the economy." We ignore the human suffering and death on the other track. The truth is that there is a lot of human suffering and death on the other track. Some of it is Covid related...the people that couldn't be saved even by the extreme measures we've taken. But there's also all the ways that GDP tracks with good health outcomes. And all the missed medical procedures and diagnostics. And all the programs that supported worldwide reductions in poverty and hunger. And all the unemployed who will no longer have health insurance. Many, many people will die because of lockdowns. I don't know how to quantify this but just acknowledging it is a start.

"The truth is that there is a lot of human suffering and death on the other track. Some of it is Covid related...the people that couldn't be saved even by the extreme measures we've taken. But there's also all the ways that GDP tracks with good health outcomes. And all the missed medical procedures and diagnostics."

Calling bullshit on this. Sorry. Why have we gone through umpteen posts here over epidemiologist models, death counts, etc. yet an economics blog just takes it as a given that "there's a huge cost". How huge a cost? Why is it so huge? Are there policies that are causing it to be bigger than it need be or are there policies that could make the cost less extreme. Nope, it's just a 'huge cost', gotta open now, we're bored working from home and miss hanging out at Starbucks so we now care about small business, or the unemployed, something something.

I've mentioned this before, but a 2016 Harvard School of Public Health study published in The Lancet estimated that the global financial crisis in 2008–2010 caused 260,000 excess cancer-related deaths in OECD countries.

There you go. There's an example of "death on the other track", caused by an economic crisis much milder than the current one. And it's not economists who said this, but public health experts.

"time-series analysis provided an estimate of more than 40 000 excess deaths due to a subset of treatable cancers from 2008 to 2010, on the basis of 2000–07 trends. Most of these deaths were in non-UHC countries." UHC = Universal Health Care

So if you have a spike in unemployment but avoid people losing their healthcare, you have dramatically fewer deaths.

What do we learn?

Set your economy up to kill 1M people if there's a downturn, you'll have to sit by and let a plague kill up to 1M people before its worth it to do anything to stop it.

Set your economy up to kill fewer people if there's a downturn, it's worth it to defend against diseases.

Sounds like we should revoke tenure for economists and let epidemiologists have a year or two running their profession to see if they do better.

I fear the demonstration effect. There are many, many reasons why policies chosen by rich countries are probably not appropriate for countries like South Africa:

1) Lockdowns are supposed to buy time to do things various things (get medical capacity in place; test; trace; etc), but in countries with very limited state capacity these things are not going to get done anyway, so the time is simply wasted. Or if only a little bit gets done, it isn't worth the economist cost. Point is that cost-benefit ratio is dramatically different for poor vs rich countries due to differing levels of state capacity.
2) Impact of the disease is likely to be much lower for poor countries with much younger populations.
3) Lockdowns don't work as well to control the disease in poor countries. People that live in crowded townships can't isolate themselves from each other. Forcing them to stay indoors probably makes things worse when "indoors" for most people means a tiny shack shared by multi-generational families.
4) Lockdowns don't work as well to control the disease in poor countries with lots of migrant labour. Workers return to their homes, potentially spreading the disease to other parts of the country that may not have been badly affected otherwise.
5) Lockdowns are difficult to maintain for long in large parts of lawless countries like South Africa, where there is very little trust in or deference to the authorities. Except, that is, in the formal sector, big business, etc, where most of the economic activity takes place.
6) Heavy-handed enforcement of lockdowns by corrupt or incompetent institutions of state can lead to further breakdown of trust in government, civil unrest, etc.
7) Emergency powers to deal with pandemic are more easily abused and extended indefinitely in poor countries with weaker democratic institutions and traditions.
8) Poor countries often don't any fiscal space to cushion the economic impact of a lockdown. South Africa certainly doesn't.
9) Poor countries may have monetary ammunition to cushion the economic impacts, but in the long term this could lead to a slippery monetary slope far more easily than in developed countries. Current discussion in South Africa exhibits worrying signs of this. (Up to now the central bank has been remarkably well run and independent.)
10) Poor countries typically have very limited (or no) social/economic safety nets, so economic effects of lockdowns are far worse for the typical person than in rich countries.
11) The further down the income curve you go, the more geared are health/mortality outcomes to economic outcomes. So the actual health/mortality effects of the "cure" can end up being orders of magnitude worse than the disease (i.e. the point of the paper Tyler linked to).
12) More generally, it is perfectly plausible that "saving some (mainly old, mainly already ill) peoples' lives" ranks far further down the societal priority list for a poor country than a rich one.

I worry that when all of this is over, South Africa will be seen as the country with the single most catastrophic policy response to the crisis.

Very interesting discussion.

May be very interesting but not a single person here has noted the most important fact about what is driving Ramaphosa'a policy, which is not pressure from the EU or the WHO. If anything, he is pressuring them to be tougher.

It is a reaction to the do-nothing policies of President Mbeki, Mandela's successor, and the predecessor of Ramaphosa'a predecessor, the seriously corrupt Zuma. Mbeki for a ridiculously long time refused to do anything serious about the HIV/AIDS crisis, denying that it involved sexual transmission and claiming it was due to an imperialist plot by US and UK figures. It was Mandela himself who finally pressured Mbeki to change his do-nothing policies after one of Mandela's sons died of the disease.

Anyway, it is this past history that has motivated Ramaphosa to introduce the policy he has that includes vigorous followups and house visits and so on, all this learned during the experience with th eHIV/AIDS epidemic.

"Generals always fight the previous war"

"Read the methodology before you comment" modern-day version of "think before you speak" TRy to steady your focus for longer than a youtube video. Read the entire paper as you are drinking the cool-aid of ill-founded conclusions.

The report attributes all costs (years of life lost) from economic collapse solely to the lockdown. How sensational and simplistic: Two artificial extremes to play off against each other (lockdown = Eco loss vs no lockdown = Eco profit). See Sweden's growth contract by 10% Q2 and Unemployment rise by 5%.

This counterfactual of SA with no lockdown, there would be significant economic collapse as the global economy collapses, demand falls due to loss of life and safety measures, and employees are too ill to work. There is also a knock-on economic cost (and thus cost of lives lost) when breadwinners die across the age spectrum which does not appear to be costed in the YLLO analysis. Loss of scientists that should have had a legacy to change SA see Dr. Gita Ramjee.

I think economic losses are directly related to the strength of your health care system’s readiness for a pandemic. Japan and SKorea smaller economic losses.

Pandemics Depress the Economy, Public Health Interventions Do Not!

Comments for this post are closed