Super fair point, you are right. And FWIW, the source I linked to does not differentiate between PhD and Master's, IIRC. I do assume there's a positive benefit, but I'd have to see whether that's supported using some other source. And the GP comment is right as well, regardless of financial benefits; going back to school will come with some limitations and difficulties. I imagine that's true even in the post-career scenario.
Supposing a PhD earned 20% over a masters and it takes 3 years, then it could be a net negative return unless you have 15 years or more of career left after returning to work. That is a valid point and something to consider before going back. The Fed report I linked to is actually arguing that recent trends have changed the balance of opportunity costs when getting degrees. (Though I happen to think the report leaves out some important details and thus frames things in a misleading way.)
Even though I brought up financial side, what I dislike about boiling it down to financial opportunity cost of attending school is that tends to frame the entire outcome as a financial decision, and ignores the opportunity cost or future value of a life getting to do research. In my mind it's better to do the PhD because you want to rather than for the money, and I suspect that people who care more about the research than the money tend to fare better as a byproduct. (That's pure speculation though.)