Furthermore, the rest of the variation is highly questionable. The vertical scale on the graph is compressed. The actual difference in the alleged U is only about 0.1, or 3% of the absolute scale value of 3.3. That is a tiny difference. I'd frankly be surprised if it was even statistically significant, let alone indicative of an actual underlying systemic change.
So I'm calling shenanigans here.
UPDATE: I searched for "pandemic" rather than "COVID" so I missed this:
> Later, as he put it, he “thought that everything was COVID.” Now, Blanchflower says, he sees that was an error. “COVID extended trends that had existed before,” he says.
But the data presented do not support this conclusion at all. There are no "trends that had existed before" in the data since all pre-covid data is grouped together in a single data set.