Written 2026-08-03, before pass-2 evidence collection. Criteria for the load-bearing hypotheses. Pass-1 provisional calls do not bind these adjudications; where pass-1 evidence already speaks, it is cited but the criteria govern.
H2.1 — Majority of under-5s are not in paid non-parental care
- Supported if: best available national estimate of children under 5 in paid arrangements (parent- or subsidy-paid, including copays) is <50%, with paid/unpaid separable at the arrangement level.
- Refuted if: ≥50% in paid arrangements.
- Indeterminate if: paid/unpaid cannot be separated within ±10pp of the 50% line.
H4.1 — Wage parity accounts for a majority of incremental cost
- Supported if: in the FTE model, (parity payroll − current payroll) ≥ 50% of total incremental steady-state cost under central assumptions.
- Refuted if: <35% under central assumptions.
- Indeterminate: 35–50%, or if facility/admin terms are too uncertain to fix the denominator.
H4.2 — Pipeline, not wage level, caps the early ramp
- Supported if: required net new recruits per year under the central ramp exceed plausible annual pipeline throughput (credential issuance + adjacent-sector recruitment capacity) even at parity wages, for ≥3 of the first 5 years.
- Refuted if: any US jurisdiction demonstrably grew its credentialed workforce faster than pipeline throughput via wages alone; or modeled recruit demand fits within throughput at parity wages.
- Refinement standing (pass 1): split channels — retention (wage-responsive, fast) vs. recruitment (pipeline-bound, slow). Adjudicate on the recruitment channel.
H9.1 — New spending is absorbed by prices rather than capacity
- Adjudicate per instrument (pass-1 finding): supported for demand-side instruments if fee growth ≥ subsidy growth in ≥2 national cases; refuted for supply-side operating grants if prices fell or held while funds flowed at ≥$10B scale (ARPA evidence already meets this).
H11.1 — Reconciliation strips universality-plus-quality
- Supported if: parliamentarian precedent shows behavioral mandates on private actors struck (≥2 cases) AND no case of a market-wide standards mandate surviving.
- Refuted if: any market-wide standard (not a grant condition) survived Byrd review.
- Refinement standing: "finance survives, mandates die"; grant conditions on new money are the contested middle.
H12.3 — Employer benefits crowd out the universal-provision constituency
- Supported if: (a) historical analog (ESI) shows constituency effects in ≥2 independent analyses AND (b) §45F-expansion uptake concentrates among large, high-wage employers.
- Refuted if: employer-benefit growth coincides with rising business-coalition support for public provision (testimony, coalition membership) over ≥5 years.
- Indeterminate otherwise — expected; this may not resolve from US data (register entry stands).
H14.2 — Segmented architectures dominate uniform ones
- Supported if: in the formal scorecard, the best per-band architecture outscores the best uniform architecture in ≥2 of 3 bands under ≥3 of 4 objective weightings.
- Refuted if: a uniform architecture ranks first in ≥2 bands under most weightings.
Deviations log opened
See deviations-log.md.