Data isn’t (always) evidence
Inflation is embedded. Team transitory is discredited, and those who said it peaked a couple of months ago were clearly lost in a maze of wishful thinking.
Why did the monetary authorities get it so wrong? Why don’t they just follow the data? The reason they’re so often behind the curve is this… data isn’t evidence.
I’d better explain.
Theories don’t just emerge from data; the theory we accept determines the data we take notice of, and the data we ignore
In the mid-nineteenth century, there were two maternity wards at Vienna’s general hospital. Ward I was notorious. Pregnant women begged to be taken to Ward II. Indeed, it became commonplace for women to refuse to enter the hospital and give birth in the street rather than be admitted to the dreaded “ward of death”.
While Ward II had a maternal mortality rate of around four per cent – about average for the time – in Ward I, for every ten women admitted, only nine came out alive. The rest succumbed to the mysterious “childbed fever”.
In July 1846, the hospital appointed a young doctor named Ignaz Semmelweis. He was determined to get to the bottom of the mystery. However, most of the theories floating around could be rejected out of hand. Poor diet? No, the food was the same in both wards. Overcrowding? No, the wards had the same number of beds, and there were fewer patients in Ward I.
A more promising explanation was psychological. Whenever anyone was dying, a priest came to administer last rites. Because of the building’s layout, he had to walk through Ward I, regardless of his ultimate destination.
Perhaps the sight of a priest in his last rites garb filled the women with dread and in some cases, scared them to death?
What data was available?
The set is infinite. It includes the colour of the priest’s shoes; the name of his horse; the prior activities of the doctors on the ward; even the number of leaves in Vienna Woods. Of course, none of these is relevant and Semmelweis ignored them.
Data isn’t evidence; evidence is relevant data. OK, but what determines which data is relevant? That is the key question. The answer is… the theory we are investigating.
It may be counterintuitive, but theories don’t just emerge from data; the theory we accept determines the data we take notice of, and the data we ignore.
Back in Vienna, the priest stopped walking through Ward I but the dreadful mortality rate remained. The psychological theory was itself dead.
Semmelweis described himself at this time as being “like a drowning man searching for a straw.”
Then tragedy struck. Exhibiting all the symptoms of childbed fever, a friend and colleague of Semmelweis died an agonising death. A few days earlier, the man’s finger had been punctured by a scalpel one of his students had just used in an autopsy.
Was there a connection? Although to us this seems obvious, it’s essential to remember there was absolutely no concept of microbes or bacteria at that time, let alone the germ theory of disease. The hospital authorities thought this latest conjecture was outlandish. A dangerous invisible force? What nonsense!
Nevertheless, Semmelweis carried on. He discovered that while students often went straight from the morgue to Ward I, they rarely did so to Ward II. Could “cadaveric matter” somehow be the culprit?
From now on, he insisted, everyone entering the maternity wards must thoroughly wash their hands.
It worked. In 1848, the death rate in Ward I fell from 10% to 1.27%; in Ward II, it fell to 1.33%.
Semmelweis’ work in Vienna shows two things: First, finding the correct theory is no easy matter. Secondly, the dividing line between relevant and irrelevant data is set by the theory we’re using.
According to psychological theory, the presence of a priest on the ward was vital data; if a priest passed through the ward there would be increased maternal mortality. On the other hand, the movements of students between morgue and ward was irrelevant.
Now, monetary authorities make predictions according to their own economic models. Every model embodies specific economic theories. This means, of course, that some data is going to count as evidence, and the rest is irrelevant.
You may recall in early 2021 the Bank of England’s conviction that inflation wasn’t about to hit because “Inflationary expectations are well anchored.” According to its economic models, any forthcoming inflation had to show up in expectations data. That’s incorrect. They were waiting for their priest in the room.
Was there anything the models made irrelevant that was in fact important? Yes: the influence of changes in the money supply on the rate of inflation. Incredibly, according to the models, the huge increase in the money supply during the pandemic did not pose an inflation risk. One thing I think we can safely predict is an imminent change in the economic models used by the authorities. To be continued…
Peter Lawlor is the Principal Economic Advisor to 7Ridge Capital. He was formerly the Principal Economic Advisor to the German Stock Exchange (Deutsche Börse), and continues to act as an adviser to senior Wall St figures and political leaders. These are his own views and should not be imputed to any organisations with which he is, or has been, affiliated