Coronavirus Fact-Check #18: Would an earlier lockdown REALLY have “saved 23,000 lives”?

The UK’s Covid Inquiry just published their Module 2 report, and though it’s over eight hundred words long, most people only really seem interested in this short quote from Baroness Hallett:

Had the lockdown been imposed one week earlier than the 23rd of March, the evidence suggests the number of deaths in England alone in the first wave, would have been reduced by 48%, that’s 23,000 fewer deaths”

It’s being repeated A LOT. But is it true?

No, it is nonsense.

The evidence on lockdowns is clear, they had little to know impact on the (alleged) disease, and were a “global catastrophe”, likely killing hundreds of thousands and impoverishing millions more.

So why is Hallett pretending otherwise? And what does she mean by “the evidence”?

Let’s find out.

Hunting down this mythical “evidence” is a longer and more drawn out process than you might imagine. I’ll go into detail.

First you have to go to the Covid Inquiry website and either open or download a copy of the report.

Then, you have to find the specific quote. Since none of the people talking about it on social media have actually read the report, none of them are citing where the quote specifically came from.

This means we have to do a Ctrl-F search of the document. I looked for “48%”, which returned two results.

First you get the same thought as the press conference, in more reporty and less clickbaity language, from the “Executive Summary” of Chapters 3+4:

the Inquiry accepts the consensus of the evidence before it that the mandatory lockdown should have been imposed one week earlier. Had a mandatory lockdown been imposed on or immediately after 16 March 2020, modelling has established that the number of deaths in England in the first wave up until 1 July 2020 would have been reduced by 48% – equating to approximately 23,000 fewer deaths.

I added that emphasis to highlight two things.

1. The language usage: “The consensus of the evidence before it”, makes it sound like there’s multiple evidences when only one is cited.

2. The nature of “evidence”: It’s a modelling study. A model should not be cited with language as strong as “would have been”, “may have been” is more correct.

Anyway, this is informative about the nature of “the evidence”, but still doesn’t actually say what it is.

For that we must move down to the second search result – Chapter 4, paragraph 129 (because I believe in proper sourcing, even if the British government doesn’t):

Professor Ferguson told the Inquiry that in later work which analysed the impact of restrictions in England:

“we explicitly modelled the counterfactual scenario of moving the lockdown of 23rd March back to 16th March, and estimated mortality … would have been reduced by 48%“

That’s from Professor Neil Ferguson, who you all doubtless remember. Also note the very important words “modelled” and “estimated”, which we’ll come back to later.

So, now we know what “the evidence” is — it’s an estimation based on a computer model produced by a man who is famously wrong about pretty much everything and who thought lockdown was so important he broke it to visit his married girlfriend.

But that doesn’t mean we should dismiss it without reading it. So how do we do that?

Well, first we must make a mental note of the footnote “484”, because this is only a governmental report years in the making and costing 190 million pounds, so they obviously haven’t got the resources necessary to hyperlink their footnotes.

Then you have to manually scroll down to the end of chapter 4, click the “show footnotes” button, then manually scroll down to footnote 484, where we finally get some links. Yay!

One, to this very long (and very dull) witness statement from good old Neil, and a second to this paper:

source

This paper is ALL of the “evidence” which “establishes” that an earlier lockdown “would have saved 23,000 lives”.

Literally, the entirety of it.

And it’s…rubbish. It’s only eight pages long (not including footnotes), and the word “estimate” appears 83 times. That’s ten times per page.

It’s also a computer model and, as I wrote last year:

Modelling studies involve feeding data into a computer programme, then asking it to form conclusions…they are only as reliable and useful as the data you use. In fact, you can very easily make them produce any result you want by feeding in the “right” (bad) data.

Or, as Seth Brundel more succinctly puts it in The Fly: “Computers are dumb. They only know what you tell them.”

In this instance, the crew at Imperial College told their computer:

1. “Covid is a dangerous, deadly disease” (which isn’t true)
2. “Lockdowns are an effective method of disease prevention” (equally untrue).

Then these clever little experts asked, “Computer, given these assumptions, what would happen if we had locked down sooner?”

At which point the computer spat out the only answer it could.

You see, by curating the available data, the authors of the study made it logically impossible for the machine to do anything but support their pre-ordained conclusion.

Now, let’s return to the original question: Would an earlier lockdown have saved 23,000 lives?

Absolutely, definitely, 100% no.

But, we can also ask, was Baroness Hallett lying when she said “the consensus of the evidence” before the inquiry “suggests” it?

Technically, no.

The “consensus” of the (so-called) “evidence” placed before the inquiry was crafted to “suggest” exactly that.

By curating the available data, the authors of the report have made it logically impossible for the machine to do anything but support their pre-ordained conclusion.

Because government inquiries are just like computers – they’re dumb, and they only know what you tell them.