Lockdown Bait And Switch
The case for harsh COVID lockdowns is weak, but there is an ongoing attempt to formalize the pro-lockdown position as the thing all smart people know.
At the beginning of 2021, I was asked to do a review of a book on COVID-19 by Richard Horton, the editor of The Lancet. As I read through the book, the realization struck me that the purpose of this book was as a piece of self-conscious history. The reason it was being written was to put a flag in the ground, stake a position, and say “this is the official history of COVID”. The goal of the book was not so much to tell a clean history of the pandemic but to start the long work of establishing the narrative of the pandemic that will one day become the formal history taught in classes on public health and in schools around the world. This book was an opening salvo in a war over the narrative.
I was reminded of the rush to establish narrative dominance when I saw this tweet:
In order to read this without giving the LA Times a click, you can go here.
My first reaction to this was laughter. It is not at all clear that COVID lockdowns saved lives without harming economies. I find it ridiculous that someone could, with a straight face, assert that. But what truly bothers me about this is that everyone is treating the discussion of lockdowns like the same game of political football that is played with every other topic when this is the single most important practical policy issue that Americans will be dealing with in the next 12 months.
This Is Not A Game, Stop Treating It Like A Game
I am tired of people who are even now treating all things COVID as some kind of game or political football that needs to be moved 5 yards down the field. This whole “Actually, Lockdowns Not Only Save Lives, But They Have No Downsides” is transparent nonsense. It’s not that it is stupid, it is that it is a talking point from “smart” people that is meant to convince other “smart” people that all the “smart” people made the right decisions and are not only justified in their past decisions but fully vindicated in every possibly way by The Data, which is the great oracle that grants us the grace of life and the fullness of knowledge, praise be to The Data.
I watch as pundits still treat this like a game or, at best, a science experiment. The reality is that the efficacy of lockdowns is the single most important issue to human beings who operate and exist in the real world. People are making enormous life decisions about what to do next, where to live, how to educate their kids, and who to trust and much of it depends on how a state is going to approach this topic.
If “The Narrative” convinces powerful people that lockdowns work, they will institute more lockdowns. That will mean more school closures and more business closures. It will mean that the people under this regime will live in the ongoing shadow that their livelihood is in danger because the widely accepted approach to all future pandemics (and possibly even the tail end of this pandemic) is to close things down or mandate behaviors.
We are already seeing this in some states with the coming school year. California is keeping mask mandates for children in place regardless of vaccination status because the people in charge in California have been convinced that mask mandates save lives.
Given this reality, it is with an unbearable helplessness that I watch The Narrative barrel forward with its absurd and disconnected blame game. The pundits fiddle with Excel spreadsheets while real people are trying to figure out how to live their lives. The disconnect between the people who implement policy and the people who suffer the effects of that policy has never been wider.
OK, But Do Lockdowns Work?
When I read the article above, I discovered the author was deferring his evidence of lockdown efficacy to Noah Smith, who wrote about it here.
His article makes all the feints of a compelling argument but pulls a pretty duplicitous bait-and-switch. In order to “prove” that lockdowns saved lives, he points to several weak studies that compare different countries to each other. Then, in order to prove that lockdowns didn’t have any overarching economic effect, he points to adjoining states to show that they were hit with similar economic downturns.
The problems here are multitude:
1) Many studies on the efficacy of lockdowns suffer from a “post hoc ergo propter” fallacy in which COVID starts spiking in a country and then the country, having recognized that COVID was spiking, instituted a lockdown. These studies presume that the lockdown was effective, so they then calculate that it was effective due to the fact that cases did not continue rising forever.
These studies are teeth-gnashingly prevalent and were completely shredded by Phillipe Lemoine, who took the models in question, applied them to a random epidemic curve, and found that the model would *always* predict that lockdowns saved lives. It didn’t matter what the epidemic curve looked like, it didn’t matter when the lockdowns were put in place, it didn’t matter what the lockdown details looked like. The model was built for no other reason than to prove that lockdowns worked, no matter when they were applied and no matter how strict they were. As a result, the model proved that lockdowns worked.
This is the proof that Noah Smith was so excited to share, the proof that found its way into the LA Times. This model, under every circumstance and using any set of data, will always show that lockdowns work. It is mathematically designed to do so.
This is the evidence that is being shown to politicians who (let us be kind while maintaining some honesty) are not equipped to judge the honesty or veracity of the data even while they pull the levers of power that deeply impact the lives of every citizen in their care.
2) The country-by-country lockdown procedure was enormously variable. Some countries implemented almost no mandatory lockdown, relying instead on individuals to independently assess their own risk. Others locked things down to a degree we American can scarcely comprehend. Germany’s lockdown procedures were so stark that you needed a permit just to go for a drive in the country by yourself. No state in the US came anywhere close to this kind of lockdown, and it is in comparing Germany to the US that Noah demands that we should prefer Illinois over Iowa for lockdown policy.
3) Judging lockdowns on an “open-to-shut” one-dimensional line misses a lot of details, especially on the state level. Some states demanded that lockdown policy be instituted top-down from the state health department to every county in the state only to have their dictates largely ignored in some (usually rural) counties. Other states allowed their counties and municipalities to make their own rules, in which high-density urban centers ended up with more restrictions and tougher mandates while rural regions eschewed them. On an “open-to-shut” scale, the first state looks very closed while the second one seems largely open. In reality, however, citizens in both states were doing the same things.
4) Lockdowns are likely to loosely follow the preference of the population. A group of people who can effectively work from home, who can comfortably avoid social interaction, who would act in a severely cautious way of their own accord are very likely to approve of lockdowns that take their own personal preferences and voluntary behaviors and apply them as state-enforced mandates across the entire state.
Issues 1 and 2 are grappled with to one degree or another in Scott Alexander’s incredibly long post on lockdown effectiveness. But he completely misses the part where the metrics that we use to judge a lockdown are more a reflection of the political and cultural preferences of the state’s population than an indicator that the actual lockdown caused changes in behavior. And he (and everyone else I’ve read about) completely miss the question of culture.
The Question of Culture
In my many many hours looking at COVID data both across the country and across the world, I’ve come to the extremely uncommon position that a substantial component of COVID spread was “culture”. I was alerted to this potential variable by looking through data on excess deaths in Europe. Most data-focused analyses of COVID in Europe have been looking at this map and trying to ask how lockdown differences in different countries map to these patterns of infection.
But as I was looking through this data, I found that this regional pattern of excess death happens every year. Yes, COVID amplified the excess death rate, but the relative rates of excess death are pretty consistent.
It would be absurd to look at these maps and say “this is related to lockdowns” because none of these countries had lockdowns during these years. Using lockdowns as our single variable assessment is fraught when the data we’re looking at is entirely ignoring the base rate of non-COVID infection in non-lockdown years.
I think “culture” includes a lot of variables like multi-generational housing, the importance of community, and cultural patterns and customs of behavior. Did Finland have such low COVID rates because they had effective lockdowns? Or because this meme is funny?
Do you greet a friend with a kiss or a hug or a handshake or a polite bow? That’s a cultural pattern and they are layered together with a lot of other things that might impact COVID spread in the US. For example, a church local to my area in Washington was still holding virtual services 2 months ago, long after state policy allowed them to returned to in-person services. But these virtual services were only available for the English service. The Spanish service has been meeting in person since at least January. That’s a cultural difference and these differences are not captured anywhere in the data or in formal lockdown metrics.
The difficulty with assessing culture is that the best yard stick with which we can judge “culture” in the United States is race and ethnicity, but that doesn’t even begin to differentiate the vast differences in culture among non-Hispanic white populations in the US.
Culture is going to be an incredibly complex thing to try to understand in the context of lockdowns because culture influences how a group of people is likely to respond to government demands. In Germany (I’ve been told) the idea that you would buck against a government health mandate is unthinkable. In the United States, we have a much more… um… variable view of how we should respond to authority. We have some groups who are faithful followers of whatever the state health department has told them to do and some groups who think deference to authority is anti-American.
In the end, there are two kinds of people when it comes to COVID restrictions:
The people who are going to do what they think is appropriate regardless of government mandates
The people who wanted to be less careful than they were being, but were coerced by government rules into being more careful
I think group 2 is a really small group. Most people who wanted government mandates were always going to abide by those mandates regardless of whether they were implemented or not. In my own state, I’ve seen that the mandates that were implemented were often ignored and poorly enforced and I see people even today who go above and beyond what the government mandates require. People did what they thought was safe and they would have done it regardless of the position of the state health department.
Unfortunately, “what people would have done regardless of restrictions” is an enormously difficult thing to capture. During the pandemic, most people really just did what they thought was safe. The only state lockdown that I believe was effective over and above what people would normally have done was Hawaii’s extremely strict border control that relied on multiple tests and traveler quarantine.
But a question that is under-represented in all these studies is “Are lockdowns worth the restriction of freedom and the demand that we force everyone into this specific pattern of behavior?” Because we’re uncomfortable asking that question, we retreat to the pure numbers. That way, the argument can become “did lockdowns save lives?” and then we get to argue about this topic using a single metric while avoiding the question of when we should implement government restrictions on individual freedom.
Is There a Time For Lockdowns?
I am not inherently anti-lockdown. I think there are some pretty obvious scenarios where the state or federal government can restrict citizen freedoms in the interest of saving lives. But I think those scenarios need to be pretty stark and pretty short. I think the events of March 2020 met that criteria.
But I’ve long maintained that the reason Seattle did not become New York City in the early days of this crisis had nothing to do with government lockdowns and everything to do with the decision making of individuals. Large employers (especially tech companies) pulled employees out of the office the first week of March. My church closed services weeks before the state told us we had to. Over half of the students at our local elementary school were already staying home before the schools were closed down. My last pre-pandemic restaurant visit was March 3rd and there were only a few occupied tables in a popular restaurant. Pike’s Place was closed and the city was a ghost town.
Two weeks after that, the New York City looked like this:
What saved Seattle was not the government, but people looking at a risky and uncertain environment and making their own free choices.
I find it very likely that, if implemented earlier in March, lockdowns would have saved lives in New York City. I think lockdowns are an appropriate option in very limited circumstances. But the data needs to obviously and unambiguously point toward a substantial reduction in harm before the government forces unwilling citizens to certain actions or restrictions. When the knowledge of risk is fully disseminated and understood and citizens have the information and ability to make reasoned and balanced decisions, then the case for lockdowns is too weak to stand.
Looney Tunes: Crowing Pains
Henry Hawk must have been popular by the time this cartoon came out because the animators barely spend any time introducing Henry and just kick him right into the fray. We get an establishing scene introducing the characters (Foghorn Leghorn, Sylvester Cat, Barnyard Dawg, and Henry Hawk) and then jump straight into the “Henry doesn’t know what a chicken is” portion of the short. Foghorn dresses Henry up like an egg and sics him on Sylvester.
This isn’t my favorite Foghorn short, it’s a little paint-by-the-numbers for me without a lot of good or unique gags. The cutest part is when Henry struggles getting into his egg suit. This watchable, but a bit stale.
I could use a bit more of an explanation of how certain countries could have consistently higher excess deaths than others. Is the chart measuring against a European average, or do more and more people keep dying in Spain each year, or what?
You’re forgetting one possible reason that Seatlle or SF did substantially better than New York. Perhaps the virus was circulating freely among the population In winter 2019.