The Corona-Panic (or, the Corona Mass Hysteria Event of 2020) continues, more narrative-driven and evidence-detached than ever.
Now that we are as far into it as we are, it’s (past) time to question our assumptions, to re-evaluate. After all, we can now compare theory and assumption with observed reality.
The previous post (Part VII, Sweden’s Vindication) compared the influential Neil Ferguson assumptions (that is, his guesstimates) of millions of deaths from the virus and millions more “swamped hospitals”-related deaths under a Swedish-style No-Lockdown approach, with the observed reality in No-Lockdown Sweden through late April.
The Transmission Rate Curve (“R0”) Revisited
In Part III, section 9, I included a graph which this post will examine in more detail, along with a parallel data-set on mobility, with the goal of trying to find possible causal factors behind the beginning of the decline:
This graph is of “R-Naught” (R0), which is the transmission rate (or reproduction rate) of the virus, and was calculated by the Robert Koch Institute (RKI), Germany’s public health agency, in a report published April 23 (see Figure 5 [“Abb. 5”] in the report). (Graph also posted at Swiss Propaganda Research‘s “A Swiss Doctor on COVID19” series about this time.)
Before saying anything else, and if you don’t read any more of this post, notice that the transmission rate declined and fell below 1.0 before the Lockdown order.
The data from other countries look similar. Here is Switzerland’s:
This post will focus on a close analysis of the German case, to see if we can find any proof of whether social-distancing, either personal-choice-based or government-mandated, “worked” in line with our assumptions, which is to say the mantra heard throughout March and April 2020. What kinds of changes in behavior are associated with the decline of the transmission rate?
The higher the line at a given time, the faster a virus is spreading at that time. A number above 3.0 means rapid spread is underway. A rate that falls and stays consistently below 1.0 signals the approaching decline and end of an epidemic, at least the transmission phase. With the “1.0 barrier” itself breached several days before the Lockdown order(s), and the peak predating it by as much as much as two weeks (!) in Germany’s case, questions need answers. What caused the decline?
The Social Distancing Manta
No assumption of Corona-Land was as influential, as central to the victory of the pro-Panic side, than the mantra of “social distancing” and its related sub-mantras and slogans (“flatten the curve;” “stop the spread”), enforced from above, with skeptics publicly shamed and dissidents including expert ones, suppressed.
Namely, the pro-Panic side has long said: “We need(ed) social distancing because it is (was) the ONLY way to cause the rate of transmission of this coronavirus to decline.”
This sounds reasonable.
Is it true? What is the evidence for it? Can we reconstruct evidence from March to corroborate it?
But first a word on what the Social Distancing Mantra implied, a key to the puzzle:
The Assumptions behind the Social Distancing mantra
Implied in the advocacy of social distancing by the pro-Panic side are two things:
- (implied) “The virus is unusually dangerous and so requires special measures like this” (in fact, studies continue to come out concluding that the Wuhan Coronavirus is not unusually dangerous; e.g., the latest study coronavirus randomized population testing in Iran published in late April finds a ca. 0.1% fatality rate in a certain province in Iran, making it one of the latest to find a fatality rate somewhere in or near the 0.05% to 0.15% window, pointing again to Just the Flu‘s vindication);
- (implied) “To enforce social distancing and stop this Virus Apocalypse, we need extreme measures, mass shutdowns, mass unemployment if necessary, ruined businesses and bankruptcies if necessary, and ‘Lockdown’ policies.” In other words, the Lockdowns rested on the theory that only social distancing could reduce the spread and “flatten the curve,”and that therefore anything done in the interest of Social Distancing was, ipso facto, a good thing.
I have challenged the first assumption there in other posts, as have many others. All studies of which I am aware that have come out over the past six weeks have discredited the notion that the Wuhan Coronavirus and “COVID19” disease is unusually, alarmingly dangerous. The media-quoted rates of 3% or 5% deaths are consistently shown to be too high by at least 25x, even up to 75x too high, because of the statistical-fiasco (bad-data fiasco) involving the “denominator problem.”
In this post I want to look at the second assumption, but will go further and try to “strike the root” in the Thoreau sense:
Let’s dare to ask if the Social Distancing premise was right. What I mean is, Did the changes in behavior we have called “social distancing,” either voluntary or forced, cause the decline in the transmission rate? That is a basic assumption and ought to be asked, and all to few have asked it.
Next, if we find changes in behavior are associated with the transmission-rate decline, what was the role of government-mandated measures (including the extreme shutdowns, stay-at-home orders, and ‘lockdowns’), vs. voluntary measures?
We have some intriguing data on all of this and I think we have the outlines of an answer.
Extreme Social Distancing was not effective
The answer is that there is no evidence for the most extreme “lockdown”-type measures taken being in any way useful, as will be demonstrated in the rest of this post but as is readily visible from the reconstructed R0 transmission-rate graphs, such as the two posted above.
As for voluntary measures: There is, surprisingly, no evidence that even these are associated with the “turning of the ride” and the beginning of the decline in the transmission rate, at least in Germany.
We are left with the surprising conclusion that the epidemic appears to have been running through the regular cycle for a flu-virus, beginning to decline “on its own” (as it were) irrespective of government measures imposed (shutdowns) and perhaps even of the major changes in public behavior that predated the shutdown/lockdown orders.
If true, this undermines a very central premise of the pro-Panic side.
The rest of this post will be a close study of the decline in the transmission rate (“R0”) in Germany, specifically, looking at associated changes in behavior and finding the chronological relationship between them.
The multiple “curves” of any flu virus epidemic
At this point it should be specified/reminded that there is a lag-time between each of the various steps in an epidemic:
- Transmission (i.e., having bodily contact with the virus and it going from one person’s system to another’s);
- Onset of first symptoms for some number of those who come into contact with the virus (depending on the virus, this can be a very small %);
- Hospitalization of the minuscule number for whom symptoms are severe;
- Entering an intensive care unit (ICU) for a portion of those who seek hospital treatment;
- And, finally, Death for a portion of the weakest and most vulnerable, as in the very aged or sick, generally drawn from those who sought hospital treatment.
Each step is subject to a time lag; the entire process will be up to weeks long for the median case who dies. A very, very large majority will not advance much beyond the first step, either zero symptoms or those too mild to notice, often similar to the light “under the weather” feeling a person might get multiple times a year.
The fact that so few ever have symptoms and few of those with symptoms seek treatment can cause a major under-reporting at the beginning of a peak-flu-event like this. By the time one case is confirmed, there are probably going to be hundreds who have been exposed, and the curve continues its upward movement. That and other technical limitations and delays in testing increased the natural lag even more, and so each step of the way, people, the media, and governments were reacting (and continue to react) to the situation as it existed weeks earlier; even if they privately admit the mistake, other forces have taken over and the pro-Panic side is entrenched and jealously guards its gains.
Each of the lines — transmission, symptoms, hospital visits, hospitalizations, intensive care intakes, and deaths — can be visualized as a curve, similar to the two curves presented in Part VII (“Sweden Vindicated“) which show Sweden’s daily coronavirus-positive deaths and ICU intakes. Note that in the Swedish case, the ICU intake curve precedes the Death curve:
The transmission curve, though, will be the first to rise, but we aren’t able to know what the transmission curve looks like until weeks after the fact. It will show a rise, have a peak period, and then a fall. The exact mechanisms behind each stage, the movements in each curve in each specific place, are for specialists to look at (there are many, including weather).
That having been said, we should be interested in dating the peak and the start of the decline in the transmission curve (technical name: “R0”).
When did Germany’s transmission curve peak? Why?
Looking back at the Robert Koch Institute (RKI) graph, we see that the 95%-confidence-estimate has the transmission-rate peaking in Germany at some point within the 72-hour window of late-PM March 8 to late-PM March 11. (This is the date range within the confidence interval for where the peak was and therefore where the decline begins.)
An immediate observation is that the large event ban looks to maybe have been beneficial. At least chronologically, the announcement of the policy predates by the peak.
Health Minister Jens Spahn (b.1980; lifetime politician; and 2020s-, 2030s-, or 2040s-era Chancellor-wannabe) announced the ban on large events with over 1,000 people on the afternoon of March 8, effective the next day (March 9). While the order predates the peak, direct contributing causality is debatable; March 9 was a Monday, and it’s unlikely there were many events with over a thousand people scheduled for early that week anyway being a work-week and not a holiday week. The secondary effect of people taking personal precautions, now that the government was announcing such things, is more plausible.
The large-event ban is merely one single data-point, though, and society is highly complex. As we are really interested in what (could have) caused the decline, we need much more data. A reasonable suggestion, given where the media narrative has been, and a common one you’ll hear, is that it is entirely because people began changing their usual lives and “social distancing.”
On the idea that the peak came almost two wees before Germany’s general ‘lockdown’ orders were issued, anti-Panic commenter Kratoklastes suggested:
It’s almost as if human beings, who make decisions about how to live their lives, had made a bunch of risk-mitigation decisions all by their own self
This is a highly reasonable conclusion to draw; there should be some good evidence for it to confirm it. We may also thereby be able to find what changes were effective, on which more momentarily.
Needless to say, if we go with “people did it on their own,” it means at the least that the moderate measures of ca. March 10-15 period were an adequate response, and that the escalation-spiral of increasingly strict measures up to ‘Lockdown’ were not necessary at all.
On Herd Immunity
There are always other factors always at play that could affect this curve. The most fundamental mechanism that drives a decline of transmission rate in any flu-strain’s spread, causing it to “peak and decline on its own” in a given place, is herd immunity. When enough people start getting exposed to a new flu-virus strain, there are fewer people available who have not yet been exposed and the rate is expected to fall. This is not hard to understand.
This is to say that the natural flu-virus-cycle always looks something like that RKI graph, at some scale at some time in some place. Throw a dart at a calendar. Whatever date you hit, odds are that some virus in your region is doing something that looks like what is seen in the RKI’s calculated Wuhan Coronavirus transmission rate (R0, R-Naught) in Germany graph. Some virus of some kind will be following that arc: Emerging, rising, peaking, declining. All unbeknownst to us, all emerging and declining for one reason or another, a series of superspreading events, weather, whatever. Most are very minor; occasionally some are moderately more serious and trigger “peak flu events” which become visible in total-mortality data, usually in December, January, or February but spikes that begin in March are not necessarily surprising.
All flu viruses do this. This is what led Knut Wittkowski to recently say,
Most of us don’t bother paying attention to such things (why would we?) but specialists know the pattern at work.
The interesting thing in 2020 is how many caved in to the alarmists and disregarded this long-established knowledge…
Germany falls to pro-Panic forces
As mentioned above, the (announcement of the) ban on large events is chronologically associated with the peak and the start of the decline in the transmission curve. The problem with even this reasonable large-event-ban order, is that, in retrospect, and in this case, even that kind of order may have been a mistake in that it was probably a contributing factor to the Panic atmosphere (the Corona-Panic). To that extent, it may have been a net negative (and, in any case, it’s likely most/all such large events would have been postponed or cancelled anyway).
Why did Germany cave in to the Panic?
It may be lost in the flux of events since then, but Germany was actually one of the holdouts against the most extreme measures, and its government leaned towards a Swedish approach until it, too, caved in. This was after the UK and US folded to the pro-Panic wave (following the evil deed by Doctor Frankensson who released his “imaginary horror scenario” guesstimate of millions of deaths March 16).
As I see it, one must analyze this thing in terms of both calculations like R-Naught and social-political factors that contributed to the vortex of Panic that caused so much unnecessary damage. And after all, the one influences the other.
Measuring how/when people changed their behavior: The Google Mobility data
In response to a discussion on the decline of the coronavirus transmission rate in Germany, “J A A” wrote:
If you look at the google community mobility report for Germany here, it suggests a correlation between the decline in the rate of transmission and the decline in the mobility of people in Germany. While lockdown was ordered on March 23 [actually evening March 22], German authorities were already suggesting it before. For example, Merkel had an address to the nation on March 18 where she said Germans should stay at home.
I recognize this is just a correlation between two sets of data and might not be significant, but it might indicate the effectiveness of staying at home.
Before looking at the Google mobility dataset in detail, here is the transmission-rate graph again for reference:
Recall now that R0 in Germany peaked at some point within the 72-hour window of late-PM March 8 to late-PM March 11.
The transmission rate (“R0”) declined as follows:
- (The peak occurs between late-PM March 8 and late-PM March 11)
- R0 was below 3.0 already as of the morning of March 12;
- It hit 2.0 by about late in the day on March 14;
- It further declined below 1.5 as of the morning of March 17;
- It fell below 1.0 about the morning of March 20;
- (The severe Lockdown, banning all meetings between more than two people and closing all businesses, was ordered early evening March 22, effective March 23).
I would suggest the general period of interest is the twelve days between March 8 to 19, which is when we see the substantial decline; the more specific window to look for possible causal factors driving the decline in the transmission rate, I would suggest, is March 11 to March 16, when it dropped from ca. 3.0 to already the cusp of 1.5. Any contributing measures taken by either the government, or by people independently, or by some other (natural) mechanism, that reduced the rate of transmission will have to be found within that window.
Here is the Google Mobility data for Germany:
What do we see about changes of behavior in the Google mobility data?
Here is what I see:
- Use of mass transit begins to majorly decline beginning on the weekend of March 14-15, with the down-trend more likely beginning Sunday March 15 than Saturday March 14 (this is close chronologically to the large fall-off in British public transit ridership, which was at 80% the usual rate on the same weekend, but down to <30% by the next weekend, according to one study; Ferguson’s release of his doomsday-fantasy-scenario on Monday March 16 is associated with the British fall-off);
- Workplace visits start to steadily go down starting Monday March 16, and were down by 40% the usual rate by Friday March 20;
- It looks like visits to “retail and recreation” start to drastically fall also about March 16, reaching ca.70% below normal by Friday evening March 20 through Sunday March 22, with the latter date being when the general Lockdown was ordered, meaning therefore that changes in behavior all predated the Lockdown.
Note that these are not sudden fall offs, but trends, mainly driven in Germany’s case by the sum of people’s individual choices.
But here is the interesting part:
Each one of these declines in observed activity post-dates the 95%-confidence interval for when the R0 peak was hit and began its decline, which was, at latest, the late evening of March 11 (a Wednesday) or as early as late-evening March 8 (a Sunday). In other words, the transmission-rate decline was underway before the changes in behavior (social distancing), as measured in the Google Mobility data.
This may indeed be good evidence, in and of itself, for herd immunity theory in practice. Herd Immunity may at times be misunderstood to be an all-at-once phenomenon, but in practice follows an arc, much like so much else in life if graphed in aggregate, driving down the transmission-rate curve.
Other reports of people’s activities can be reconstructed from media reports. I will leave this to others, but will suggest one more and briefly look at it:
During an early part of the Corona Panic in the US, there was a focus on churches.
An early and high-profile report out of the US (March 9) of a coronavirus-positive pastor at an Episcopal church in Georgetown, Washington DC, who was thought to have “infected several parishioners,” led to the kneejerk decision by the Episcopal diocese of Washington to suspend all services at all its churches for three weeks (a decision effectively expanded to several months as the Panic escalation-spiral swirled ever upward, and bans on church services were imposed after pro-Panic forces seized control of governments).
A look at what churches, and congregants, were doing during the critical period in Germany may therefore be of interest. A few very specific dates are relevant: The evening of Wednesday February 26th (Ash Wednesday); the Sunday mornings of March 1st, 8th, and 15th. These are by far the busiest times for churches near the period of interest.
It appears there were no disruptions at all to services in Germany through Sunday March 1. Transmission was high by this time but it was not yet known. The Panic had yet to begin.
We might expect slightly lower attendance for Sunday March 8, given the early stages of the Panic cycle, promoted by the media there as everywhere else, but services were held as usual.
It appears that most church services in Germany were held as usual through Sunday March 15 (the Protestant churches reportedly held services usual on March 15 with increased safety measures, though some had already cancelled services). The largest churches/cathedrals had suspended services in accordance with the ban on large events and some local/regional government decisions. It appears that the churches that hadn’t suspended services by March 16 were heavily leaned on to do so by the government starting that day (the Catholics announced a suspension of services with the last diocese confirming the suspension March 18. Government “guidelines” recommending the suspension of services at all houses of worship were issued March 16), and very few, if any, services were held as usual on Sunday March 22, just ahead of the general Lockdown order (after which no congregation was allowed to meet at all; unprecedented in peacetime).
This is a lot of detail on churches in Germany, but the salient fact is: While the overall transmission-rate decline starts not later than late-PM March 11 (Wednesday), the church-related attendance falloff dates only in part to Sunday March 15, and is in full only from Sunday March 22. Leading up to Sunday morning services on Sunday March 15, the transmission decline was underway for at least 3.5 days and as many as 6.5 days.
The bulk of the changes in church attendance and therefore of close contact at churches, related to suspensions of services, post-date the beginning of the start of the transmission-rate decline.
The later Lockdown order that closed all churches was so late that R0 was already below 1.0 before Sunday March 22.
Church attendance may not be high in Germany, but this is a close look at yet another place where changes in behavior post-date the decline in the transmission rate.
Summary and Conclusions
We can also conclude a few surprising things about the decline of coronavirus transmissions in Germany in March 2020, which deserve to be better known:
- The peak of the transmission rate was between late-PM March 8 and late-PM March 11 (the RKI appears to give only a 2.5% chance the peak was after 12:01am March 12),
- After the peak, a decline began which turned into sustained decline, creating a familiar bell-like curve in the transmission rate;
- Given that the transmission rate had sunk below 1.0 some days earlier, the Lockdown order which was announced March 22 (effective March 23) was unnecessary and needlessly economically and socially damaging ;
- The major changes in the behavior of the public, surprisingly, all post-date the start of the transmission-rate decline;
- There should be no doubt that people’s changed behaviors contributed to the lowering of the transmission rate, visualizable within the graph by” keeping downward pressure on the curve” starting about March 16 (a Monday), and to a limited extent the immediately preceding weekend based on the clear changes in behavior seen in the Google data;
- While personal-choice social distancing may have helped with “downward pressure” on the transmission-rate curve, it did not itself trigger the start of the decline, which may have peaked on its own in accordance with the beginning stages of herd immunity in some parts of Germany (a well-underway process that was disrupted/delayed by the Lockdown);
- The same conclusions will apply anywhere else the epidemic curve had reached the stage it had in Germany by early March.
As for the question posed at the start of this post, on whether Social Distancing was necessary and important: Inconclusive.
It appears Social Distancing is yet another way within “Corona” that the data does not quite align with the narrative.