Do shutdowns reduce coronavirus deaths significantly? An update

In an August post, I argued that shutdowns reduce coronavirus deaths significantly. I based my argument on common sense — the disease spreads through human contact so, up to a point, reducing human contact reduces the number of infections — and on comparisons between the experiences of neighboring jurisdictions that have taken different approaches to shutting down.

I thought it might be worthwhile to update the comparisons I made in the August post. As in that post, unless otherwise indicated, the figures I cite come from Worldometer.

The main comparison I discussed was between the experiences of Norway, which locked its economy in the face of the virus, and Sweden, which did not. When I wrote my post in late August, there had been only 49 deaths per one million people in Norway. In Sweden, there had been 575.

Where do things stand now? Fortunately, there have been few coronavirus deaths in either country since late August. In Norway, the number of deaths from the virus per one million people is virtually unchanged at 51. In Sweden, the number is 582.

There have been changes in both countries when it comes to infections. Sweden may have hoped to achieve herd immunity by not locking down. However, new reported cases of infection began rising in September. By the end of the month, new cases were at around 600 per day — up from around 300 in July. Fortunately, as noted, there has been no spike so far in the number of deaths attributed to the virus.

Norway has also had a spike in new cases. That spike is more pronounced than in Sweden — from about 10 new cases per day in July, to about 60 in August, to about 180 by the end of September. But even so, Norway is still reporting fewer new cases per day per capita than Sweden is. And like Sweden, Norway has not yet seen a spike in deaths attributed to the virus.

Denmark also went on lockdown, but the Danes reopened the economy earlier than Norway did. As might have been expected, Denmark had less success in preventing coronavirus deaths than Norway, but considerably more success than Sweden, which didn’t shut down.

In August, the per capita number of deaths from the virus in Denmark was 107. Now, the number is 114.

Denmark, though, has had a larger spike in new cases than Sweden and Norway. In mid September, new reported cases in Denmark were around 600 per day, roughly the same as Sweden, but with a population only about half as large. Lately, the number of new cases in Denmark has fallen slightly to 400-500.

The total number of infections per capita in Denmark during the pandemic is roughly half of the number in Sweden. The total number of per capita deaths is about one-fifth of Sweden’s.

There are, of course, demographic differences between the three Scandinavian countries I am comparing. But I’d be shocked if they can come close to explaining the dramatically different results between the three.

It’s also worth noting that, before their policies diverged, the coronavirus health outcomes in Norway and Sweden were not very different.

My August post also compared coronavirus outcomes in several upper Midwest states. I compared South Dakota, which never shut down, with Wisconsin, which shut down until mid May when a court order reopened the economy, and Minnesota, whose shutdown wasn’t lifted.

Per capita deaths attributed to the virus in South Dakota were almost identical to the number in Wisconsin. This, despite the fact that population density is greater in Wisconsin. South Dakota has no city remotely like Milwaukee. Thus, the fact that the per capita death numbers in the two states are similar suggested to me that Wisconsin’s shutdown for about a month and a half prevented loss of life due from the virus.

Since my last post on this subject, deaths from the coronavirus have spiked in both Wisconsin and South Dakota. However, the spike is greater in South Dakota. It now attributes 280 total deaths to the virus per capita compared to 240 such deaths in Wisconsin. Before, as noted, the numbers were almost identical.

Since Wisconsin has been off of lockdown since the late Spring, I don’t think we can draw conclusions from this particular comparison with South Dakota. However, I think we can say that South Dakota hasn’t achieved anything like herd immunity.

Now let’s consider Minnesota. When I wrote my last post, it attributed 321 deaths per capita to the virus — a higher number than South Dakota, which never locked down, and Wisconsin, which locked down for a shorter period of time.

Today, Minnesota attributes 382 deaths to the virus. So it’s still doing worse, by this measure, than the two neighboring states.

However, in my August post, I took into account the fact that a very high percentage of all Minnesota deaths from the virus occurred in long term care facilities — a much high percentage than in South Dakota. This analysis seemed relevant to the question at hand because a lockdown isn’t designed to prevent the virus from spreading among those confined in these facilities.

After taking long term care facilities into account, it appeared that Minnesota had fewer deaths per capita than South Dakota among those not trapped in these places, even though demographics and population densities would tend to make South Dakota a less deadly place than Minnesota in a normal pandemic.

The percentage of Minnesota deaths from the virus that occur in long tern care facilities has been trending downwards for some time. The trend has continued since my last post on this subject. However, at 71 percent, it still far exceeds the percentage in South Dakota (48 percent). Thus, it’s still true that Minnesota has fewer deaths per capita among the population outside of long term care facilities than South Dakota does.

This also appears to be true when one compares Minnesota and Wisconsin. However, the data on deaths from the virus in long term care facilities as a percentage of all deaths from the virus in Wisconsin aren’t robust.

Wisconsin offers another way of getting at the question of the effectiveness of lockdowns in reducing coronavirus infections and deaths. Wisconsin lifted its lockdown due to a court order issued in mid May.

What happened after the lockdown was lifted? Nothing much for a while. However, by mid July the number of new reported cases per day had doubled.

Since then, Wisconsin cases have spiked significantly. The state is now reporting six times as many new cases per day as it did when the lockdown was lifted.

Minnesota did not experience this trend. Between mid May and mid August the number of new reported cases per day remained roughly constant. Lately, the number has risen significantly, but there has been nothing like the Wisconsin spike. New cases per day in Minnesota are less than double what they were in mid May. In Wisconsin, as noted, they have increased by a factor of six.

Reported cases are, in part, a function of the number of tests administered. However, Minnesota does more testing per capita than Wisconsin, so the large spike in Wisconsin cases, as compared to Minnesota, probably can’t be explained by the number of tests.

None of this analysis means that shutdowns are beneficial, on balance. There are costs associated with shutdowns that have to be considered. And even if shutdowns are beneficial on balance, that’s not necessarily an argument for shutdowns of a very high level of severity. However, I believe relevant data comparisons continue to show that shutdowns significantly reduce coronavirus infections and deaths.

UPDATE: My original post on this subject didn’t discuss North Dakota, so I didn’t mention it above. However, it’s worth pointing out that North Dakota didn’t issue a stay-at-home order. Per capita deaths from the virus in that state are roughly at the same level as in Minnesota, notwithstanding demographic differences that would cause one to expect a much lower number.