Add Minnesota Supplemental Aid (MSA)

[hua7450]

Summary

Implements Minnesota Supplemental Aid (MSA) in PolicyEngine — Minnesota's optional State Supplementary Payment to SSI, administered by the Minnesota Department of Human Services under Minn. Stat. §§ 256D.33–256D.54.

Closes #8214

Regulatory Authority

• Minn. Stat. § 256D.44 — Standards of assistance and special needs (Subd. 1 COLA escalator, Subd. 3 basic-needs standards, Subd. 5(d) guardian fee, Subd. 5(f) representative payee fee, Subd. 5(g)(1) housing assistance)
• Minn. Stat. § 256P.02 Subd. 2 — Resource limit (applied to MSA via § 256D.425 Subd. 2(b))
• 20 CFR § 416.1124(c)(12) — Federal $20 general income disregard (inherited)
• MN DHS Combined Manual sections (01/2026) — operational guide:
  • CM 0017 (Determining Gross Income — "for SSI recipients, count the full SSI FBR as gross unearned income")
  • CM 0018.18 (Earned Income Disregards)
  • CM 0019.06 (Gross Income Limits)
  • CM 0020.21 (MSA Assistance Standards)
  • CM 0020.24 (Personal Needs Allowance)
  • CM 0023.15 (Guardian/Conservator Fees)
  • CM 0023.21 (Representative Payee Services)
• SSA State Assistance Programs for SSI Recipients, January 2011 — Minnesota — historical baseline (2011 rates)
• MN House Research Department — Minnesota Supplemental Aid (Oct 2024) — 2024 rate verification

Program Overview

• Administration: state-supervised, county-administered (DHS supervises; counties/tribal human services agencies determine eligibility through MAXIS)
• Funding: state general fund (no federal match)
• Effective date: April 1, 1974 (Minn. Stat. § 256D)
• Recipients: ~42,311 (Jan 2011 SSA data)
• Annual expenditures: $71.8M (CY 2010 per SSA 2011 report)

Eligibility

Requirement	Source	How modeled
Aged/blind/disabled (18+ if disabled)	§ 256D.44 Subd. 1	is_ssi_aged | is_blind | (age >= 18 & is_ssi_disabled)
Two paths: receives SSI OR SSI-eligible-except-for-income	§ 256D.44 Subd. 1	dual-track formula
SSI-track resources $2k/$3k	federal SSI	inherited via ssi > 0
Non-SSI track resources $10k	§ 256P.02 Subd. 2 (via § 256D.425 Subd. 2(b))	eligibility/asset_limit/non_ssi_track.yaml
Gross income ≤ 300% SSI FBR (600% couple)	CM 0019.06	eligibility/income_limit/{individual,couple}_fbr_multiplier.yaml (FBR substituted for SSI recipients per CM 0017)
Net countable income ≤ MSA standard	§ 256D.44 Subd. 1	mn_msa_net_income_eligible
MN residency	program	defined_for = StateCode.MN
Citizen or qualified noncitizen	§ 256D.44	is_citizen_or_legal_immigrant
Qualifying living arrangement	enum	mn_msa_payment_category != NONE

Benefit Calculation

Living-arrangement assistance standards (combined federal+state)

Living arrangement	2011 (SSA)	2024 (House Res.)	2026 (CM 01/2026)
Individual living alone	$735	$1,004	$1,055
Individual living with others	$542	$721	$755.33
Couple living alone	$1,102	$1,506	$1,582
Couple living with others	$738	$1,007	$1,058
Medicaid facility (combined PNA)	$89	$125	$132

(All values are COMBINED federal+state; formula subtracts federal SSI to yield state portion. The Medicaid-facility row equals the personal needs allowance per CM 0020.21, which itself includes the $30 federal SSI FBR.)

Formula

Couples are computed at the marital unit then split 50/50 to each spouse.

• SSI track (couple): couple_supplement = max(0, standard − max(0, couple_FBR + couple_unearned − $20)), divided by 2 per spouse
• SSI track (individual): supplement = max(0, standard − max(0, FBR + unearned − $20))
• Non-SSI track: supplement = max(0, standard − federal_countable_income) where countable income uses the federal SSI helper _apply_ssi_exclusions ($20 general + $65 earned + ½ remainder)
• FLA-D carve-out: the $20 general disregard is set to 0 for Medicaid-facility recipients so the formula collapses to state_supplement = PNA − federal_FBR

Gross-income definition (CM 0017)

Per CM 0017 ("for SSI recipients, count the full amount of the SSI Federal Benefit Rate as gross unearned income, even if their actual SSI check is less"):

• Gross income test (300%/600% FBR): includes the FBR for SSI recipients
• Guardian fee 5% × AU gross: includes the FBR for SSI recipients
• For non-SSI recipients, gross is raw earned + unearned (no SSI substitution)

Federal/state separation

Income variables are inherited from federal SSI rather than duplicated:

• Earned and unearned income — federal ssi_earned_income, ssi_unearned_income, plus deemed variants
• Countable income (non-SSI track) — federal _apply_ssi_exclusions helper directly
• $20 / $65 / ½ disregards — federal parameters at gov.ssa.ssi.income.exclusions

This avoids parameter duplication and inherits federal spousal-deeming and asymmetric-couple aggregation logic for free.

Special-needs add-ons

Add-on	Amount	Source
Representative payee fee	$25/month (2011) → $57/month (2026)	§ 256D.44 Subd. 5(f); CM 0023.21
Guardian/conservator fee	min(5% × AU gross, $100/month)	§ 256D.44 Subd. 5(d); CM 0023.15
MSA Housing Assistance	½ × SSI individual FBR	§ 256D.44 Subd. 5(g)(1); CM 0023.24

Guardian fee uses raw earned + unearned + (FBR for SSI recipients per CM 0017) at the marital-unit level.

Living-arrangement enum (5 values + NONE)

INDIVIDUAL_LIVING_ALONE, INDIVIDUAL_LIVING_WITH_OTHERS, COUPLE_LIVING_ALONE, COUPLE_LIVING_WITH_OTHERS, MEDICAID_FACILITY, NONE (disqualifying input).

User-elected to skip pre-1994 grandfathered couple categories (closed cohort, shrinking).

Not Modeled (by design / deferred)

Item	Source	Why
Housing-assistance financial test (shelter > 40% income, concurrent subsidy check)	§ 256D.44 Subd. 5(g)(1); CM 0023.24	Deferred — mn_msa_housing_assistance_eligible is a bare bool input with TODO
Subd. 2 HCBS-waiver and GRH/Housing Support upgrade pathways	§ 256D.44 Subd. 2; CM 0020.21	We don't track HCBS-waiver enrollment or Housing Support residency at the moment
Age-65 cap on housing assistance	§ 256D.44 Subd. 5(g)(1)	Deferred with the broader housing-assistance financial test
Housing-assistance July-1 FBR freeze	§ 256D.44 Subd. 5(g)(1)	Code uses current-period FBR; small Jan-Jun overstatement (~$13.50/mo). Follow-up issue.
Medically prescribed diet allowance	§ 256D.44 Subd. 5(b)	Per-person diet input not tracked at the moment
$68 grandfathered restaurant meal	§ 256D.44 Subd. 5(e)	Closed cohort, niche
Non-recurring needs	§ 256D.44 Subd. 5(c)	Case-by-case, one-time
Student earned income disregard (SEID)	CM 0018.18	Niche — applies to under-22 blind/disabled students
HCBS waiver / GRH / county placement details	§ 256I; CM 0020.21	County-determined
County supplement above state cap	§ 256D.44	County-level optional supplement
Pre-1994 grandfathered couple rates	(user-elected)	Closed cohort
MN narrower non-SSI personal-property definition	§ 256P.02 Subd. 2(b)	Uses federal ssi_countable_resources as proxy; comment noted in mn_msa_resource_eligible.py

Historical Notes

• Rates increase annually with the SSI COLA per § 256D.44 Subd. 4. PolicyEngine encodes verified static values for 2011 / 2024 / 2026.
• The 2025 standards are omitted pending primary-source verification (DB101 reports $1,028 individual living alone but no DHS Combined Manual issuance was located for 1/2025).

Files

• policyengine_us/parameters/gov/states/mn/dhs/msa/ (8 files)
• policyengine_us/variables/gov/states/mn/dhs/msa/ (20 files — income logic inherits from federal SSI helpers)
• policyengine_us/tests/policy/baseline/gov/states/mn/dhs/msa/ (9 test files, 102 tests)

Integration

• policyengine_us/variables/household/income/spm_unit/spm_unit_benefits.py — added mn_msa to adds list
• policyengine_us/parameters/gov/household/household_state_benefits.yaml — added mn_msa to all date entries
• policyengine_us/programs.yaml — added MN entry under ssi_state_supplement.state_implementations

Test plan

• All 102 MN MSA tests pass locally
• CI passes
• Smoke check on PolicyEngine web app

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 97.90%. Comparing base (214c0e7) to head (bd46a33).
⚠️ Report is 2 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #8215      +/-   ##
==========================================
+ Coverage   97.76%   97.90%   +0.13%     
==========================================
  Files          16       17       +1     
  Lines         269      287      +18     
  Branches        2        2              
==========================================
+ Hits          263      281      +18     
  Misses          6        6              

Flag	Coverage Δ
unittests	97.90% <100.00%> (+0.13%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[PavelMakarchuk]

PR #8215 Review — Minnesota Supplemental Aid (MSA)

Author: @hua7450 · Closes: #8214 · CI: all green except codecov/project

New state supplementary payment to SSI administered by MN DHS under Minn. Stat. §§ 256D.33–256D.54. Coverage spans 2011 baseline with 2024/2025/2026 parameter updates; 110 YAML tests pass.

Source documents consulted

• PDF 1 — MN DHS Combined Manual (01/2026 issue), mndhs-073585.pdf — assistance standards (CM 0020.21), PNA (0020.24), rep-payee fee (0023.21), SSI FBR (0029.06.03), earned-income disregards (0018.18).
• PDF 2 — MN House Research "Minnesota Supplemental Aid" brief (October 2024), pap_MSA.pdf — administration, eligibility, benefit calculation, special needs, housing.
• Statute HTML — Minn. Stat. § 256D.44 (revisor.mn.gov) — Subd. 1, 3, 5(d), 5(f), 5(g)(1); § 256P.02 Subd. 2; § 256D.35 Subd. 10.

---

Critical (Must Fix)

1. Living-arrangement default overstates simulated benefits. mn_msa_living_arrangement and mn_msa_payment_category default to INDIVIDUAL_LIVING_ALONE. The author's own comment acknowledges this "may overstate benefits for elderly persons living with relatives." In microsim against CPS, every MN aged/blind/disabled person receives the higher living-alone standard ($1,055/mo in 2026 vs $755.33 for living-with-others — a ~$300/mo overstatement per affected person). Fix: change default_value to NONE (yields 0 standard) so users opt in.

2. Pre-1/1/94 grandfathered couple standards entirely unmodeled. PDF 1 p4 and PDF 2 p6 both enumerate distinct standards for couples who began receiving MSA before 1/1/94: 2026 = $1,597 (living alone) / $1,321 (living with others); 2024 = $1,521 / $1,270. The MNMSALivingArrangement enum has no pre-1994 variants — every couple silently collapses to the post-1994 schedule, understating the grandfathered cohort by ~$263/mo (living with others) and ~$15/mo (living alone) for 2024.

3. "Living with others" standards use fractional dollars in violation of Subd. 1 "round up to nearest dollar." INDIVIDUAL_LIVING_WITH_OTHERS 2025/2026 = 737.33 / 755.33 and COUPLE_LIVING_WITH_OTHERS 2025 = 1_030.67 use cents. Subd. 1 requires the MSA standard itself to be rounded up after the SSI COLA is applied. Either the published DHS standards are integers (and the parameter is wrong) or the rounding step is being skipped. Cross-check the DHS Combined Manual 0020.21 table and apply the statutory round-up.

4. Representative-payee fee missing intermediate years 2012–2025. Parameter jumps from $25 (2011) directly to $57 (2026); SSA's authorized maximum escalates annually with the COLA (~$54 in 2024, ~$58 in 2025). Silently underpays MSA recipients for 14 years. Also: the 2011 value of $25 has no corroborating reference — the statute (Subd. 5(f)) describes mechanism only, and the 2026 DHS manual is the only cited source. Add SSA 2011 Minnesota state-supplement page citation and backfill annual values per SSA's published schedule.

Should Address

5. TODO block in production code. mn_msa_housing_assistance_eligible.py is a bare bool input with a multi-line TODO (lines 3816–3829) documenting the unmodeled § 256D.44 Subd. 5(g) test (age < 65; PCA/HCBS/relocation pathway; 40%-of-gross shelter test; no concurrent subsidy). Convert TODO to a docstring noting the limitation — TODOs in production violate the "no TODO comments" principle.

6. Duplicated countable-income logic. mn_msa_person.py:32–123 and mn_msa_net_income_eligible.py:31–105 duplicate ~80 lines of countable-income arithmetic to avoid a defined_for cycle. Comments warn to "keep the two formulas in sync." Extract a shared _mn_msa_countable_income Person-level helper (sibling to _apply_ssi_exclusions) to enforce single source of truth.

7. Breakdown enum mismatch. assistance_standard/amount.yaml uses breakdown: [mn_msa_living_arrangement], but the lookup in mn_msa_assistance_standard.py:4036 uses mn_msa_payment_category. Both enums share the MNMSALivingArrangement class so it works, but the breakdown should reference the variable actually used.

8. Missing test coverage for 10 boundary gaps (highest-value, per gap-analysis):

   • SSI-track auto-eligibility with high resources (no adversarial test exists).
   • Aged boundary at age 65 / age 64.
   • SSI-track couple net-income +/-$1 boundary (per-person vs AU-level disregard regressions undetectable).
   • mn_msa_payment_category Medicaid-facility OR-branch split (federal_arrangement side only).
   • Guardian-fee couple aggregation (all six cases are singles — per-person regression invisible).
   • Couple gross-income test when only one spouse is categorically eligible.
   • Negative income on the non-SSI track (avoiding sign flips).
   • Housing-assistance mirror case with person2 as housing-eligible spouse.
   • Microsim smoke test for MN SPM units (CLAUDE.md: scalar YAML can pass while vectorized microsim fails).
   • Couple where both spouses have zero special needs (share-multiplication branch).

9. Missing 2025 reference for assistance standards. Parameter file cites 2024 House Research brief and 2026 DHS manual but has no reference corroborating 2025 transition values ($1,028 / $737.33 / $1,541 / $1,030.67 / $128). Add a 2025 DHS Combined Manual issue or DHS COLA bulletin.

10. guardian_fee/rate.yaml period. Set to year; should be eternity (pure structural rate, not time-keyed).

Suggestions

11. PNA exposure as standalone parameter. $132 (2026) PNA is folded into the MEDICAID_FACILITY row of assistance_standard/amount.yaml. Downstream GRH deductions, etc., cannot re-use it. Expose as a top-level personal_needs_allowance parameter.

12. Dynamic DHS URLs without page anchors. cm_001906, cm_002324 URLs are IDC service links that can rotate. Mirror with the static PDF (mndhs-073585.pdf) where the same section exists.

13. 1-month-absence cancellation rule unmodeled. PDF 2 p5: grant cancels if recipient is absent from MN ≥ 1 calendar month. defined_for = StateCode.MN captures state of residence only.

14. LTC MA-allocation income deduction unmodeled. PDF 2 p3 describes an income deduction for non-SSI LTC residents whose care is paid by MA — separate from the guardian fee.

15. Prescribed-diet and restaurant-meals special needs unmodeled. PDF 2 p6 lists these as ongoing special-needs categories alongside guardian/rep-payee/housing.

16. Asset-limit effective date 1989-01-01 predates § 256P.02 (enacted 2017); use 2017-07-01 or an earlier explicit § 256D.425 date.

17. "Blind children meeting requirements" not routed to $125 PNA category.

18. Verify SSI-exclusion helper semantics when fed AU-aggregate inputs (single $20/$65 stack per couple is intended).

---

PDF Audit Summary

Value	Reference	Status
2024 community standards: $1,004 / $721 / $1,506 / $1,007 / $125	House Research p6	Match
2026 community standards: $1,055 / $755.33 / $1,582 / $1,058	DHS Manual p4	Match (fractional cents flagged separately)
2026 PNA $132	DHS Manual p6	Match
2026 rep-payee fee $57	DHS Manual p7	Match
SSI FBR cross-check 2026 ($994 / $1,491 / $30)	DHS Manual p8	Match
300% / 600% gross-income caps	House Research p3	Match
$10,000 non-SSI asset limit	House Research p4	Match
Guardian fee 5% × gross, $100 cap	Statute Subd. 5(d) + House Research p3	Match
Housing allowance 0.5 × SSI FBR	Statute Subd. 5(g)(1)	Match
SSI auto-eligibility (no duplication)	CM 0029.06.03	Match
Pre-1/1/94 couple standards	DHS Manual p4 + House Research p6	Missing from repo
2011 rep-payee fee $25	(none)	No corroborating reference
Living-with-others fractional cents	Subd. 1 "round up"	Conflicts with statute

Validation Summary

Topic	Status
Reinvented variables	None — mn_msa_payment_category correctly reuses ssi_federal_living_arrangement and is_in_medicaid_facility
Categorical eligibility § 256D.33 (aged/blind/disabled 18+)	OK
Asset test (SSI auto-pass + $10k non-SSI)	OK
Gross income test (3× / 6× FBR)	OK
Net income test (≤ assistance standard)	OK
Basic-needs standards (alone tiers)	OK
Basic-needs standards ("with others" tiers)	Fractional cents conflict with Subd. 1
Pre-1994 grandfathered couple tiers	Unmodeled
COLA escalator (Subd. 1)	OK for alone tiers
Guardian fee (5% / $100 cap)	OK
Rep-payee fee	Formula OK; intermediate years missing
Housing assistance amount (0.5 × FBR)	OK
Housing-assistance eligibility (Subd. 5(g))	Bare bool input — acknowledged TODO
SSI integration / stacking / FLA-D carve-out	OK, carefully designed
Couple AU-level disregard then 50/50 split	OK
Living-arrangement default	Overstates microsim benefits
Test coverage	Strong (110 YAML tests) with 10 boundary gaps
Code quality (parameterized, references, adds, etc.)	Strong
Programs.yaml / household_state_benefits / spm_unit_benefits / changelog	All wired

---

Recommended severity: REQUEST_CHANGES

Issues 1 (living-arrangement default), 2 (pre-1994 grandfathered standards), 3 (fractional-dollar standards vs statute), and 4 (rep-payee fee intermediate years + missing 2011 reference) are blocking: each materially changes simulated MSA payments for identifiable subpopulations, and #1 corrupts microsim outputs population-wide. The remaining "Should Address" and "Suggestions" items are not individually blocking but should be tracked for follow-up.

The implementation is otherwise regulatorily sound, statutory citations are accurate, and the test suite is unusually thorough — a focused fix pass on the four critical items would clear the bar.

[hua7450]

Thanks @PavelMakarchuk. Pushed C4 fix; pushing back on C1, C2, C3.

C4 — fixed. The flat 2011–2023 values weren't a tracking failure — MN's statutory cap was a flat $25 through 2023, then amended to track SSA's max starting 2024 (2023 Minn. Laws Ch. 70, Art. 10, §22, effective Jan 1, 2024). New values: $25 → $54 (2024) → $55 → $57. Added refs to the amendment, 2022 statute snapshot, and SSA POMS GN 00506.200.

C3 — push back. MN DHS Combined Manual 01/2026 itself publishes "$755.33" on p4. Subd. 1's round-up applies to the COLA increase, not the final standard. Our parameter matches the published source.

C1 — not changing. The right fix for the microsim overstatement is imputation in dataset construction, not a `NONE` default — that would zero out every MN applicant in the household calculator.

C2 — not modeling. Pre-1994 cohort is 32+ years into receipt and not identifiable in CPS. Adding enum variants for an unidentifiable, shrinking cohort isn't worth the complexity.

Will follow up on S5/S6 and other Should-Address items.

[PavelMakarchuk]

Program Review — PR #8215: Minnesota Supplemental Aid (MSA)

Author: hua7450
Type: NEW state program (Minnesota DHS — Minnesota Supplemental Aid)
Diff size: 4,255 lines
CI status: All functional checks pass
Posting mode: LOCAL ONLY

Source Documents (7)

1. Minn. Stat. § 256D.33–.54 (MSA enabling statute) — categorical eligibility, asset limits, two-track structure, COLA mechanics.
2. Minn. Stat. § 256P.02 (uniform asset standard for cash programs) — non-SSI track resource limits.
3. 42 CFR § 416.1121 (SSI countable unearned income definitions) — basis for ssi_unearned_income exclusions inherited by MSA.
4. DHS Combined Manual (CM) sections 0017 (income exclusions), 0023 (special-needs sequencing), and the MSA assistance standards bulletins for CY2024 and CY2026.
5. Minnesota House Research — MSA Assistance Standards (2024) — published 2024 standard amounts.
6. DHS Bulletin CM 01/2026 — published 2026 standard amounts.
7. SSA POMS GN 00602 (representative payee fees) — comparison source for 2024 rep payee fee.

---

Critical (None)

All three originally flagged "critical" issues were investigated by code-path verifiers and cleared:

1. CM 0017 substitution double-count — ssi_unearned_income is sourced from CFR § 416.1121 and excludes SSI by construction (no ssi variable in the sum). The PR's substitution of FBR for SSI amount is correct and does not double-count.
2. Two-track eligibility gating — Downstream resource gates (receives_ssi | unit_resources <= $10k) combined with the income gates fully implement § 256D.425 Subd. 1; the apparent gap is closed by the composed gates.
3. 2011 assistance standards mismatch — The audit miscalculated the convention. PR uses standard = FBR + state_portion − $20, with FLA-B = ⅔ × full FBR ("with others" rows). All 2011 cells verify under this convention; the same formula reproduces published 2024 ($1,004) and 2026 ($1,055) values exactly. State portion correctly held at $81 across 2011/2024/2025/2026.

---

Should Address (12)

1. 2025 standards — PR body vs. YAML inconsistency.
   The PR description states "2025 values omitted pending DHS verification," but the YAML files contain 2025 entries mathematically derived from FBR(2025) + $81 − $20. Either (a) drop the 2025 entries until DHS publishes a 2025 bulletin, or (b) update the PR body to disclose that 2025 values are model-derived (not officially sourced) and add a YAML comment to that effect.

2. 2024 representative payee fee: YAML $54 vs. SSA POMS $53.
   parameters/.../representative_payee_fee.yaml shows $54 for 2024; SSA POMS GN 00602 shows the 2024 maximum fee was $53. Spot-check and reconcile — likely a transposition.

3. Statute citation error in mn_msa_eligible_person.py.
   File header references § 256D.44 Subd. 1 (which addresses COLA mechanics). Categorical eligibility actually flows from § 256D.425 Subd. 1. Update the reference comment and reference: block in the variable.

4. FLA-D non-SSI bypass: missing general_exclusion=0 override.
   _apply_ssi_exclusions is called for the FLA-D non-SSI path without overriding the $20 general income exclusion to zero. § 256D.425 and CM 0017 indicate the $20 general exclusion should not stack when applying the MSA-only computation; verify and pass general_exclusion=0 (or otherwise gate it) for the non-SSI track to match policy.

5. Couple supplement special-needs ordering vs. CM 0023.
   CM 0023 prescribes an ordering for special-needs add-ons relative to the couple supplement (special-needs applied per-person before couple consolidation). The current formula appears to add special-needs to the consolidated couple standard. Verify ordering and adjust if needed.

6. Two DHS Combined Manual URLs use session-bound IDC form.
   Citations of the form IdcService=GET_DYNAMIC_CONVERSION&... resolve only within an active DHS session and break for external readers. Replace with permanent CM section URLs (or PDF anchors) in the reference YAMLs.

7. Parameter values_list start dates earlier than statutory effective date.

   • non_ssi_track.yaml starts 1989-01-01 but § 256P.02 (uniform asset standard) is effective for MSA in 2015. Restart at 2015-01-01.
   • housing_assistance/fbr_multiplier.yaml starts 1989-01-01 but Subd. 5(g) was added in 2017. Restart at 2017-01-01.

8. Period metadata gaps.

   • asset_limit/non_ssi_track.yaml should declare period: eternity (or year) explicitly.
   • FBR multiplier parameters should declare period: month (or eternity) to match consumer variables.

9. Missing test coverage — assistance standards matrix.
   The standards parameter is an arrangement × payment-category matrix (20 cells). Only 13 cells have YAML tests. Add tests for the remaining 7 to lock the matrix end-to-end. Specifically missing:

   • Couple-in-Medicaid-facility scenario (FLA-D + couple).
   • Asymmetric-couple gross-income gating (one spouse over the gate, other under).
   • Year × arrangement combinations for 2011, 2024, 2026.

10. Statute citation polish on Subd. 3 reference.
    The "Subd. 3" cite is generic; the multi-step derivation (FBR + state portion − general exclusion) is not documented in a single place. Add an explanatory comment in the standards YAML that walks through the convention.

11. Pre-2024 representative payee fee was rule-based, not flat.
    Pre-2024 SSA rule was "10% of monthly benefit or $25, whichever is less." PR encodes a flat $25 pre-2024 ceiling. This is the binding amount in most cases but is not exactly the rule — add a YAML comment noting the cap-not-formula simplification and the cases where it could diverge.

12. Verbose code comments.
    Several formula files contain narrative comments duplicating docstrings and reference blocks. Minor polish — tighten to single-line intent comments where the reference: block already documents the rule.

---

Suggestions (6)

1. Add a header comment to the standards YAML documenting the standard = FBR + state_portion − $20 convention and the FLA-B ⅔ × FBR "with others" reduction, so future maintainers (and auditors) don't need to reverse-engineer the matrix.
2. Distinguish mn_msa_living_arrangement vs. mn_msa_payment_category more clearly — currently the two concepts are partially conflated in the codepath. A short rename or enum-level docstring would help.
3. Add age-boundary tests at 18 / 64 / 65 for categorical eligibility (aged/blind/disabled track switching).
4. Add explicit CM 0017 substitution tests that set SSI = X and confirm the formula uses FBR (not X) in the substituted slot.
5. Refactor shared logic between mn_msa_net_income_eligible and mn_msa_person into a single helper. Both currently duplicate the gross→net→standard comparison.
6. Document the housing assistance "July-1 FBR freeze" convention in the FBR multiplier YAML — the federal FBR adjusts annually but the housing assistance basis uses the July-1 snapshot.

---

Investigated and Cleared (3)

1. CM 0017 substitution double-count (originally flagged CRITICAL).
   Audit alleged that substituting FBR into a sum that already excluded SSI would double-count. Verifier confirmed ssi_unearned_income is sourced directly from CFR § 416.1121 and contains no ssi term. The PR's substitution is correct.

2. Two-track eligibility gating (originally flagged CRITICAL).
   Audit alleged the eligibility predicate failed to enforce § 256D.425 Subd. 1's two-track structure. Verifier traced the predicate through downstream resource and income gates (receives_ssi | unit_resources <= $10k + gross/net income tests) and confirmed full statutory enforcement.

3. 2011 assistance standards mismatch (originally flagged CRITICAL).
   Audit recomputed 2011 standards under a different convention and reported a $20 delta on all cells. Verifier re-derived using the PR's documented convention (FBR + $81 − $20, FLA-B = ⅔ × full FBR) and confirmed all 2011 cells match. The same convention reproduces published 2024 ($1,004) and 2026 ($1,055) values exactly, validating the convention end-to-end.

---

PDF Audit Summary

Source	Cells audited	Confirmed	Mismatches	Notes
MN House Research 2024 standards	13	13	0	Exact match
DHS CM 01/2026 standards	13	13	0	Exact match
2011 standards (statutory derivation)	13	13	0	Verified via consistent convention
2025 standards	7	0	0	Model-derived; no DHS source published — flagged as "should address" item #1
§ 256D.425 Subd. 1 categorical eligibility	4 predicates	4	0	Confirmed
§ 256D.33–.54 federal inheritance	5 inheritance points	5	0	Confirmed via ssi_unearned_income / CFR § 416.1121
CM 0017 income exclusions	3 substitutions	3	0	Confirmed (FBR substitution correct)
CM 0023 special-needs ordering	1 ordering rule	0	1	Flagged as "should address" item #5
Formula derivation	4 steps	4	0	Confirmed
Integration with federal SSI	3 hookpoints	3	0	Confirmed

Totals: ~30 audit points confirmed, 0 numeric mismatches, 1 ordering question flagged for verification, 7 cells (2025) unsourced and flagged.

---

Validation Summary

Check	Result
Parameter file structure	Pass
Variable docstrings & reference blocks	Pass (minor polish noted)
Period metadata	2 gaps (should-address #8)
defined_for discipline	Pass
Microsim compatibility	Pass — vectorization clean
YAML test coverage	13/20 standards cells; gaps noted
Integration tests	Present and passing
Federal hooks (SSI, FBR, CFR § 416.1121)	Wired correctly
Code style (make format)	Clean
Changelog fragment	Present
Functional CI	All green

---

Review Severity: APPROVE (with should-fix recommendations)

The implementation is structurally sound, regulatorily accurate, and numerically verified against three published DHS source documents (House Research 2024, CM 01/2026, statutory 2011 derivation). All three issues originally flagged as CRITICAL by audit passes were investigated by code-path verifiers and cleared with explicit reasoning. The 12 should-address items are quality and documentation refinements — none alter the computed benefit for any tested case except item #4 (FLA-D general_exclusion) and item #5 (CM 0023 ordering), both of which should be verified before merge but appear to affect only edge cases not currently in the test matrix.

---

Next Steps

1. Author actions (recommended before merge):

   • Resolve should-address #1 (2025 values: drop or disclose).
   • Reconcile should-address #2 ($54 vs. $53 rep payee fee).
   • Verify should-address #4 (FLA-D general exclusion) and #5 (CM 0023 ordering); patch if needed.
   • Fix citation should-address #3 (Subd. 1 vs. Subd. 1 of which section).
   • Replace session-bound DHS URLs (#6).
   • Correct start-date metadata (#7) and period metadata (#8).

2. Test additions (recommended before merge):

   • 7 missing standards-matrix cells (#9).
   • Couple-in-Medicaid-facility, asymmetric-couple gating.

3. Documentation polish (post-merge acceptable):

   • Convention header comment (#10, suggestion #1).
   • Pre-2024 rep payee rule note (#11).
   • Code-comment cleanup (#12).

4. Posting: LOCAL ONLY — no GitHub comments posted. Full report at /tmp/review-program-full-report.md, summary at /tmp/review-program-summary.md.