Last March, right at the beginning of the pandemic lockdowns in the US, I did a short podcast with Jake Fisher about the likely economic effects of lock down. You can listen to the podcast here: Eric Hammer (Part 2): The Economic Impact of the Coronavirus Pandemic. (If you don’t think an hour and 48 minute podcast is short, the first podcast I did with Jake ran over three hours.)
The past month, a friend of mine from high school listened to the podcast and sent me a few questions about what I had said and whether I still thought it was a good idea. He graciously agreed to have the response serve as a post here, under the presumption that it probably took me three hours to write it. Closer to nine, probably, but still, here we are!
(His response to my response might be the next post, as my response to his response veers into political economy and governance and away from direct cost benefit analysis of pandemic response. And it isn’t like this needs another 3,000 words.)
Steve (from the recent past of the future) asks:
You made a statement about lockdowns...which was along the lines of "people should make the best decisions for themselves on whether to social distance, and the level of lockdown practices they should take. They know their situation best"
Now that you saw a year's worth of behavior...do you still hold that opinion?
I personally saw people who were very opposed to the most minor of inconveniences of wearing a mask. She was going to switch churches because the 2 hours a week she sat in a church and wore a mask was a bridge too far. She traveled around and saw her mother in law in Florida during July surges...her mother in law died of covid in October
Without public policy...most large gatherings would have continued...and in my opinion, with people choosing to attend.
So...do you think public health policies of lockdowns are needed, since people may not choose the best choices of gatherings or masks...and we do not live in a world where spreading of viruses can be contained only to those who make those choices?
Short answer: Yes. I think the people who needed to quarantine, the elderly and otherwise infirm, did well to isolate themselves independently of policy. People outside of that group have such a low rate of serious illness that quarantining has massively higher costs than any benefits. (See the CDC stats here as of 7/1) Arguably people self-quarantined more than was necessary before there were mandates. The lack of correlation between states' or countries' quarantine or lockdown procedures and COVID deaths per capita supports that, along with at least one study that found that government lockdowns didn't change people's behavior much. (I will try and dig that one up.)
Long answer: Good questions, and get to the core of the externality question of when it is valuable for policy to intervene in human affairs. I think there are a few angles to consider, grouped based on what we knew at the time and what we know in hindsight.
1: Who is affected? Contrary to what some commentators have recently stated (*ahem* Tyler Cowen *ahem*) we did know at the time that the vast majority of deaths were in the over 65 age group, with those younger being largely unaffected. We also knew that the large majority of fatalities had accompanying comorbidities. (Alas, I don't really know how to demonstrate this other than pointing to the podcast... if someone has a link to the old CDC site data that I got the information from at the time, or something similar, that would be great.)
This isn't to say that all working age and younger people are immune, or what the long term effects would be. Any coronavirus, whether COVID-19 or regular seasonal flues, will kill the young here and again. However, like regular flu, we did know that the vast majority of deaths would be among the old and infirm. If one takes a peak at the CDC's top ten sources of yearly fatalities, flu and chronic respiratory conditions are routinely right up there. Regular flu is probably underreported as official autopsies are generally not undertaken when an old person dies of natural causes.
Importantly, unlike regular flu, the very young are entirely spared from COVID, so thank god for that.
Why is who is affected so important? Well, it means that who needs to worry about being infected, and thus who needs to quarantine or lockdown changes drastically. If people are all at a moderate risk of death, then it makes sense to keep everyone away from everyone else. If those who are threatened are actually an easily identifiable subset, say the elderly and already infirm, imposing lockdowns on everyone makes rather little sense. Instead, those at risk can lockdown along with those who live with them as necessary, while everyone else can go about their business just like with other flu seasons. Locking everyone down imposes very large costs with very little benefit, as those at low risk aren't gaining much safety by staying inside most of the time.
2: What are the benefits of lockdown? This takes a bit of unpacking, as there were many different benefits cited, some of which make sense and some of which do not.
The first big benefit is not getting the virus in the first place. How possible it is to avoid getting the disease is, and was, uncertain, given that you have to acquire food and medicine eventually. We never really did figure out how many people caught it but never showed symptoms, as quick, cheap and easy testing never happened. Presumably more quarantine means less risk, although there are probably bigger gains from avoiding large gatherings and wearing a mask than there is from also never leaving your house or seeing your family. Definitely diminishing returns. These benefits are also a function of number 1 above: if you are not likely to die (ie. you are pretty healthy and under 65) the personal benefits from being less likely to get the virus are pretty small. If the virus just makes you feel super tired and sick for three to five days (my experience and the experience of those I know who have had it) that's pretty minor. If you are post retirement age or have comorbidities, then it makes a lot of sense personally to lockdown or quarantine yourself because getting the virus is more likely to be deadly than getting a common flu.
The second benefit of locking down is not spreading the virus to others. However, this benefit is directly tied to the above benefit of the person getting the virus from you avoiding the virus. If you spread the virus to me, someone who is young and healthy enough to be fine, it is low cost to me, so avoiding that spread is of low benefit. If you spread the virus to your grandmother that likely is a high cost to her, so avoiding that is a large benefit. If you spread the virus to me, and I don't realize it and then might spread it to my grandmother, your spread to me is indirectly costly, mitigated by how likely I am to spread it to someone whose cost is higher. If grandma is well protected or quarantined, you spreading to me without my knowledge is low cost because she isn't likely to get it, and if not it tends to be higher cost.
A third benefit of lockdowns was the "flattening the curve" concept. The fear was that if a bunch of people got the virus and got sick all at once, the hospitals would be overwhelmed and people who might live with proper care would not be able to get the care and die. We had seen such things happen in China, as well as Italy as I recall. How much of that is a function of virus action, insufficient medical capabilities, or a population with lungs weakened from pollution (Wuhan has really unhealthy air) is up in the air. It does not seem to have occurred many other places in Europe or the USA. At the time, however, overwhelming the medical facilities was a fair concern, and locking down densely populated areas or regions with a low proportion of hospital beds per capita for a bit might not have been a bad idea. (Yes, you can look up hospital beds per capita. Before you click, guess what the correlation between hospital beds per capita and deaths per capita from COVID is.)
I don't see much of a correlation either... possibly deaths slightly correlating with more beds is due to both correlating with a higher level of elderly residents.
Other people have recommended, at least implied, that locking down until there was a vaccine and it could be applied to everyone was the thing to do. Early on, that was not a good idea. Firstly, we had no idea if a vaccine could be made (although one was ready for testing very early, it wasn't clear it worked and the FDA surely wasn't going to let it through for a long time.) Secondly, it was not, and is still not, clear that a vaccine would actually make you permanently immune. Corona viruses are notoriously hard to vaccinate against, and indeed only one virus has been eradicated by vaccination, that being smallpox; polio still pops up now and again. Thirdly, as will be discussed below, can you shut down an economy for months without causing massive collapse? I doubt anyone who has ever had anything to do with business or personal finance would suggest a month of downtime wouldn't be a problem.
So, on whole, were there benefits to be had or even expected from lockdowns? Well, if one was worried about overwhelming the local hospitals, perhaps in those <2 beds per 1000 residents states, then sure, maybe lockdown for a bit to save some time. Many people seemed pretty happy to limit their social and economic activities quite a bit in the early days for exactly that reason, without being told. On the other hand, flattening the curve for more than a week through locking down was pretty infeasible, since who has that much food and medicine in their house, even if they don't need to go to work?
That leaves us mostly with the benefits of not spreading the virus, which depend largely upon which people the virus is being spread. Avoiding getting the elderly and otherwise infirm exposed probably was a good idea, but it seemed the young could be exposed and infected with very little risk. Since the at risk individuals are easy to identify, there is little risk of exposing people at high risk accidentally, so little danger in spread among the population at large.
**From the FUTURE we also know that the virus… really wasn’t that deadly. Look again at the above graph with deaths per 1,000 residents and beds by same. Deaths is the vertical axis. The highest value was just under 3 deaths per thousand people, or 0.3% of the population. Given how skewed the deaths were, with approximately 80% in the over 65 crowd, that isn’t a terribly frightening number in terms of years of life lost. The virus wasn’t additive, that is, it didn’t kill ~500,000 people on top of everything else, but was cause of death instead of something else that was likely to kill someone. See this awkward picture of the chart from the CDC link above. Note the row All Ages and All Sexes in the middle. Out of 4.9 million deaths from all causes, about 600,000 involved COVID, 530,000 involved pneumonia, and 293,000 involved both COVID and pneumonia. In other words, of the ~500K that died from pneumonia, ~300k also had COVID.
The point here is not that 600,000 deaths isn’t a lot of people in a year and a half. It is that of those 600,000 many were going to die anyway of something else that year or the year after.
Ivar Cummings had a really excellent video going into the UK’s health services’ data on deaths. I finally found it this morning, but alas, all that remained was the smoking corpse where YouTube had banned it for wrong think. To summarize, possibly badly from recollection, the UK has very detailed records on patients due to their universal government supplied medical care. They also have rules where if someone comes into the hospital with problem X and it looks like they will die from problems Y and Z in a bit anyway, they do not try to fix problem X. Again, because they have universal government supplied medical care. Plus side, they write that down! The short version was the most of the people who came to the hospital with COVID fell into the category of “Well… he isn’t long for this world anyway. Let’s direct resources elsewhere.” COVID certainly didn’t help, but it was more the straw that broke the camel’s back rather than the tree trunk that did the same.
** I added this part after replying to Steve, as I was trying to find the Ivar Cummings video before I wrote it, then forgot before I hit send. If anyone knows where that video is outside of YouTube, I would appreciate letting me know!
3: What are the costs of lockdowns? This one we knew a bit about at the time, but probably really underestimated compared to what the results really were. I know I underestimated the costs in the podcast, and even I wasn't optimistic about them. I suspect it is because the costs of lockdown increase exponentially with time. A one week lockdown is pretty cheap, basically a long holiday vacation people didn't plan for. A one month lockdown is shockingly expensive as supply chains start to break and people run out their savings. A six month lockdown is a huge problem, as businesses close, severe shortages develop and the complex network of interrelated and interdependent systems of trade and specialization start to come crashing down. Q2 2020 saw a 30% drop in GDP from Q1. That's... incredible. The Great Depression was less than 15%. Ouch.
(All that said, we won't have finalized numbers for GDP for a while now, so that might be revised a bit. Still, here's a graph from the Bureau of Economic Analysis.)
So we knew there would be costs, but we probably didn't know how crazy high they would be. We still don't know until we understand the fallout from the huge amounts of debt financed emergency spending the government has done, the inflation that seems to have kicked off, and how policy makers try to reign in that inflation while not making the debt unsustainable by raising interest rates. I am kind of curious to see how that pans out, although I would be a lot more inclined to grab some popcorn and watch the horror show unfold if I didn't have kids who will have to deal with the outcomes. I am pretty sure it is going to be very bad.
Anyway, we knew lockdowns were going to be very expensive in early 2020, though we probably underestimated just how expensive. So how long could we bear that cost? Well, for a week or two to "flatten the curve", that didn't seem too bad. Longer? We are going to have to wait and see.
More importantly, and closer to the original question, what costs do individuals and families face? These costs are knowable only at that low level. As the difficulty of defining what an "essential" business or employee demonstrated, this kind of low level detail is entirely illegible to policy makers. Individuals, families and businesses all have to assess the costs themselves. Some people and families might be able to afford to go a few weeks without work, or even just work from home, while others might not have that option or the savings to deal with being put out of work. Many small businesses went under in the past year, and I suspect the full tale of the forced shifts in the national economy won't be told for a while yet.
There are non-monetary costs as well. Increased rates of depression, suicide, drug and alcohol deaths, divorce, joblessness, all those are going to be hard to quantify but definitely take a toll on people. These costs will never be fully understood and taken into account. Human misery is often merely a footnote in the history of grand plans.
Also throw in the medical care NOT received by people due to hospitals and doctors not being allowed to treat patients with non-COVID related issues. Cancer screening? Nope. Mental health visits? Nope. The list goes on, and is partially catalogued in this article from late 2020.
Another important aspect, much forgotten, is that by locking down we kept people from getting infected. Which sounds insane, because obviously that is the point! But consider that we know that people under the age of 65 without comorbidities are very unlikely to get seriously ill, and these people are also those likely to spread the disease via going to work, the store, all the usual adult things. Well, if they all get the virus they generate antibodies and, hey presto, become vaccinated in effect. Once that happens herd immunity kicks in pretty quickly, and those at greatest risk to the virus don't have to quarantine as long. At worst, those at risk need to stay quarantined until there is a vaccine proper for them, but everyone else goes back to life as normal. In effect, locking down everyone regardless of risk means the virus stays in circulation longer instead of hitting everyone quickly and then running out of hosts. Sort of like when we were kids and one of us got chicken pox, our parents made sure all of us got it at the same time to avoid dealing with it later, especially because getting chicken pox for the first time as an adult is super miserable while getting it as a kid isn't fun, but passable with sufficient ice cream and cartoons.
A related problem is that by locking people down and keeping them from getting infected, we are effectively forcing rationing on the virus, ensuring it will have a supply of unprepared hosts around. Unless you can keep everyone entirely isolated until a fully effective vaccine can be administered, as soon as people come out of lockdown into the outside world where COVID exists, some will get it, resulting in secondary waves, tertiary waves, etc. as new groups are exposed. Now, if you can keep everyone entirely isolated and have a vaccine on hand such that you can reasonably eradicate the virus in the wild before letting the quarantined people out into the now clean world, that works fine. If you can't guarantee 100% quarantine and you don't have any prospect of a vaccine that can be administered to such a large percentage of the population that the virus dies, well then you are just wasting your time. And money. And all the quarantined people's lives, the portion of which you have to lock them up at least.
Whew... lotta costs there. I would argue that pretty much all of those costs were known, or easily surmised, in early 2020, although as a low ball estimate. I don't think we fully realized how much the economy was interconnected and thus how laughable the idea of closing everything but essential businesses was. Turned out a pretty large chunk of the economy was essential. I don't think we fully realized how exponentially expensive lockdowns were going to be. Did people arguing for extended lockdowns expect an unemployment rate of 13% in May of 2020? Possibly not. Should they have? Probably, but since that's the highest since the Great Depression according to the BLS, well, I suppose it is hard to blame them for not foreseeing that they could tank the economy THAT badly. People never do seem to foresee how badly a government can wreck an economy.
COST vs. BENEFIT TIME!
Ok, deep breath. Took a long time to get here. Do I stand by my assessment that people should be allowed to make their own decisions as isolate or lockdown as they saw fit? More generally, do I still think that the policies of lockdowns and other interventions to stop the spread of the virus were a good idea?
I do. I stand by that statement. I think the massive unemployment, contraction in production, and general dislocation of lives both economic and social was too much of a price to pay for the seemingly tiny benefits that the lockdowns offered. One might argue that the benefits were lessened because people didn't take them seriously. That might be true, although I would not say that a positive point of policy controls is that people ignore them. It is in fact a negative feature of policy controls that some people ignore them while others pay the price. I would also note that the Amish communities of central PA (where I was living for most of 2020 to present, moving from MN in June 2020) did not seem to follow any of the mandates, and so far as I can find have not been decimated by the virus.
Further, there is no evidence that lockdowns helped at all, which is actually really surprising even if you are, like me, skeptical of them. I'd have guessed that more lockdowns meant fewer deaths, although with sharply diminishing returns. Nature published at least one article on this here, although they note there are objections. What sort, and why those are not being published as non-confirmatory replications, is unclear.
Had I the ability to go back in time and advise world leaders on COVID policy, I would have suggested a quarantine of the elderly and infirm, with state support for their food and medical needs while they are stuck at home with their families. Along with mask recommendations and suggestions that people avoid riding in subways, gathering in poorly ventilated places or the like, I think that would about do it. Hopefully it would have helped without causing such massive costs. I would recommend that if I had a time machine, but even at the time it seemed like a pretty reasonable course of action, while lockdowns seemed pretty excessive.
I can not tell you how happy I am that there is a public record of my saying that, too. Compared to the nonsense coming from the CDC and other federal and state government officials at the time and since, I sound like a bloody prophet. I feel like Cassandra, but I suppose it is better to be right but ignored instead of ignored and wrong. Given the continuing employment and deference paid to the aforementioned officials, there is apparently zero consequences to being wrong yet paid attention to.
** Also update: after writing this, Scott over at Astral Codex Ten posted about lockdown effectiveness. I am generally on board with his analysis, although I think his methods understate the economic costs of lockdowns. Largely this is due to the interactions between states, and how this tends to pull states towards the average despite their policy differences. I will write up a little bit about this, but for those interested NOW, definitely check it out.