Track SPM-specific input gaps: housing, WIC, school meals, child support, workers comp, and expense deductions

[MaxGhenis]

Problem

SPM-specific poverty resource and deduction components need explicit Populace handling so the resource definition is complete and release diagnostics show unsupported families rather than silently omitting them.

Relevant families:

• HUD housing assistance / capped housing subsidy resources.
• USDA WIC resources.
• USDA school lunch/breakfast resources.
• HHS OCSE child support received and paid.
• Workers' compensation benefits.
• MOOP, work expenses, and childcare expense validation/imputation benchmarks.

Scope

For each family:

1. Identify the relevant PolicyEngine-US input/output variables and whether Populace should store an input, impute an upstream input, or let the engine compute a formula-owned variable.
2. Identify source status and whether an Arch source package exists.
3. If source data are missing, make release diagnostics report the family as unsupported/source-gap.
4. Once source-backed inputs exist, add Populace imputation/target mapping and tests.

Acceptance criteria

• A Populace US SPM component inventory exists with model variables, formula/store decision, source status, and release behavior.
• Missing source families are surfaced in release diagnostics/manifest as source gaps.
• Populace does not hard-target CPS SPM values to fill these gaps.
• Where source-backed inputs are available, tests confirm inputs or upstream formula inputs are populated.

[MaxGhenis]

PE-US-data backlog migration note from #46: this issue should absorb the relevant SPM/resource source-gap work from PolicyEngine/policyengine-us-data#1113, #736, #733, #735, #739, #740, #732, #182, #445, and #381. Keep these as source-gap diagnostics or source-backed inputs/targets after classification; do not hard-target CPS SPM itself.

[daphnehanse11]

SNAP parity note from the July 1 SNAP input audit:

This issue is directly SNAP-relevant, not only SPM-poverty relevant. The current Populace H5 has several raw CPS/SPM fields that old policyengine-us-data converted into PolicyEngine-facing SNAP income/deduction inputs, but the Populace output surface does not consistently expose those PE names.

Examples:

• child_support_received from CPS CSP_VAL
• child_support_expense from CPS CHSP_VAL and/or reviewed SPM child-support fields
• spm_unit_pre_subsidy_childcare_expenses from SPM_CHILDCAREXPNS
• spm_unit_capped_work_childcare_expenses from SPM_CAPWKCCXPNS
• workers_compensation from WC_VAL
• disability_benefits from DIS_VAL1 / DIS_VAL2, excluding workers-comp sources
• veterans_benefits from VET_VAL

These feed SNAP gross/net income, deductions, elderly/disabled deduction treatment, and work-requirement exemption logic. For an OBBBA SNAP work-requirement analysis, missing aliases make the model depend on defaults or formula outputs rather than the observed CPS/SPM inputs already present in the source table.

Related focused issues: #242 for hours, #243 for SNAP reported/take-up, and #244 for eligibility/exemption flags.