I say this as someone who's been full-time WFH for over a decade; I'm not taking a position one way or another.
So the "steep dive" that you talk about is a single year -1.7% decrease vs a decade of constant increase, going as high as 4.6% in a single year?
I feel that chart is almost deceitful. Every year that this is above 0 is a percentage win no matter the delta. So if that chart showed
* 2030 -> 20%
* 2031 -> 1%
It'd feel like a steep dive, but 2031 would still be an _increase_ in productivity over 2030.
It's common to draw economic indicators this way.
Basically, if one year had +3 and the next -3, you would be back at the original value.
The largest ever 4.6 at the peak of WFH in 2020 and the fall to a normal 2.2 in 2021 both represent fast growth in productivity. The only "bad" point there is the -1.7 at the peak of RTO in 2022. So the GP's analysis is indeed bullshit.
(But anyway, there are some much more compelling explanations for 2022 than RTO, do not take it as evidence against mandating office work, it's not strong enough for that.)
People think you can measure this as a single number. At best, for a very limited number of workers, who don't WFH anyway, you can measure it. But you cannot compare those workers with the workers next to them doing similar but slightly different jobs even.
So it's a huge mess and the idea of a graph of it for the USA is like making a graph of the average number of testicles. Except people take it seriously!?
It goes down because you now have to pay him more for the same amount of work. But it also goes up because marginal utility of his work has increased.
Productivity shot up early COVID when people were scared and WFH started?
That's a pretty big leap.
Manufacturing and Durable Manufacturing had the biggest decreases and WFH isn't a thing in those sectors. https://www.bls.gov/news.release/prod2.nr0.htm