Item 75: Close the loop: policy-dependent date costs

Every existence result in the affordability round is E0-E2 of its own hierarchy: the date costs L_t are exogenous, given to the scheduler by an environment that does not react. In a live market both inputs to the cost react to the policy. The market maker's slack and the opposing volume m_t respond to how hard the enforcement position has been pushing, and the exclusion depth D^r_t responds to whether the reason has been enforced — that is the point of enforcing it. So the scheduler's own spending changes the prices it will face.

Settle E3-E5. The tractable first question is monotone: if enforcing a reason weakly decreases its future depth, does the persistence criterion survive, and is the greedy tranche rule still causal-optimal? The adversarial one is whether a policy can be driven into a friction trap it created — cheap dates spent early, depth restored, and no affordable dates left.

Deliverable shape: a stated theorem with exact-rational fixtures, or a trap instance. Acceptance check: the round's runner is discovered by python3 tests/run.py and passes.

Context: projects/normativity/legitimacy/rounds/2026-08-31-normative-affordability/CLOSED_LOOP_EXISTENCE.md for the hierarchy and what E0-E2 already give, SHARP_PERSISTENCE.md for the exogenous criterion the closed loop has to be compared against, and SHARP_TIMELY_SERVICE.md §6, which names this as the largest gap in its hypotheses. A solution ships: the criterion, and a statement of which of the exogenous theorems survive the loop unchanged.