Ummm... household data for Japan isn't household data for China.

The quote said "East Asian" countries, hence Japan.

What I had wanted to emphasize from both the quote and the interview, though, was the degree to which we can be far more rational about economics simply by acknowledging the realities of basic accounting: surpluses somewhere must equal deficits elsewhere. Before this reality, huge quantities of implicitly normative "economics" dissolve.

It's not at all obvious to me that the quote has anything to do with that. The quote seems to be arguing "if you use current data, you would think X->Y. But if you used past data, you would think X->~Y, so we should be reluctant to use causal reasoning like this."

Because of an argument I came across earlier which failed spectacularly badly (which also had to do with China), I pattern-matched and said "wait, would we actually think X->~Y if we looked at past data?". According to the paper Lumifer found, at least, the answer is "no, we would think X->Y."