[PavelMakarchuk]

Summary

gov.hhs.medicaid.eligibility.categories.parent.income_limit still held the 2021 values for Texas, Florida, Mississippi and Wyoming (0.19, 0.32, 0.26, 0.56) in 2024 and 2025: #9475 updated these four states from 2026-01-01 only. Connecticut stayed at 1.60 from 2021, so it also missed the October 1, 2024 cut to 138% of FPL (Public Act 24-81). Without a change, 1.60 also applied in 2026 and later.

This PR adds 2024-01-01 and 2025-01-01 entries for the four frozen-dollar states. Each one uses the method of the 2026 entries: the state's published standard for a family of three, divided by that year's guideline for three ($25,820 in 2024, $26,650 in 2025), with the 5-point MAGI disregard. It also adds CT: 2024-10-01: 1.38. South Carolina (67%) is already correct on main (#9475), so this PR does not change it.

State	Standard for a family of three	Source	2024	2025	Before
TX	$230/month, C-131.2 Revision 15-4, effective October 1, 2015	Texas Works Handbook C-131.2	0.157	0.154	0.19
FL	$303 adults standard (dated April 1992) + $183 standard disregard (their $486 sum is the MNIL, dated January 2014)	DCF ESS Appendix A-7	0.276	0.269	0.32
MS	COE-075: $492 (March 1, 2024) and $495 (March 1, 2025), 5% disregard included	MS DOM Eligibility Manual Appendix A-3, pages 2-3	0.229	0.223	0.26
WY	$873/month Family MAGI. Family Care rests on the 1996 AFDC standard	WDH income requirements, 2026 Healthcare Coverage Options in Wyoming, footnote 3	0.456	0.443	0.56
CT	160% before October 1, 2024, then 138% (CGS 17b-261(a): 155%, then 133%, plus the 5-point disregard)	Public Act 24-81, section 38, SPA CT-24-0025, CT DSS, 2024 HUSKY A parents and caretaker relatives transition	1.60 (January 1 value)	1.38	1.60

Cross-checks:

• KFF, Medicaid income eligibility limits for parents also uses a family of three and includes the disregard. It truncates to two decimals. May 2024 and January 2025: TX 0.15 and 0.15, FL 0.27 and 0.26, WY 0.45 and 0.44, CT 1.60 and 1.38, MS 0.22 in January 2025. The one exception is KFF's May 2024 value for MS, 0.27, which is 0.229 + 0.05. It adds the disregard to a figure that already includes it, and KFF's January 2025 value does not do this.
• MACPAC Exhibit 36 (July 2024, 2025) gives the highest share across household sizes, without the disregard: TX 12 and 11, FL 24 and 22, MS 19 and 17, WY 45 and 42, CT 155 and 133. Each new value is at or below the MACPAC figure plus 5 points, except MS in 2025 (0.223 against 0.22), which is a whole-percent rounding difference.

The 2018-2023 values for these four states are rounded shares that sit above each state's standard in most years. They are out of the scope of #10151 and are tracked in #10491.

Tests

• parent/medicaid_parent_income_limit.yaml, Cases 24-34: TX, FL, MS and WY for 2024 and 2025 (the arithmetic is in a comment on each case), CT 2024 (1.60, the January 1 value) and CT 2025 (1.38). Control: TX 2023 stays at 0.19.
• is_parent_for_medicaid.yaml: the issue's Connecticut parent at 1.414 FPL qualifies in 2024 (control) and does not qualify in 2025. A Florida parent at 0.30 FPL does not qualify in 2025 under 0.269 (0.32 admitted them).
• With the main parameter file, 11 of the new cases fail (9 limit cases and 2 eligibility cases), and the 3 controls pass. With this change, all 64 cases in the two files pass.
• Also run with this change, all passing: gov/hhs/medicaid + gov/hhs/chip (842), gov/aca + household + gov/fcc (1,046), partners (623), gov/states/{ct,tx,fl,ms,wy} (2,033). The code-health tests for reference URLs, parameter YAML keys and known dead URLs, and test_parameter_files.py, also pass (113).

Fixes #10151

axiom: TheAxiomFoundation/rulespec-us#1951 queued

🤖 Generated with Claude Code

[DTrim99]

PR Review: Set 2024-2025 parent Medicaid limits for TX, FL, MS and WY, and Connecticut's 138% limit from October 2024

Summary

The PR changes one parameter file, gov.hhs.medicaid.eligibility.categories.parent.income_limit (#10151):

• TX, FL, MS and WY. It adds 2024-01-01 and 2025-01-01 entries. Each one is the state's published family-of-three standard divided by that year's guideline for three ($25,820 in 2024, $26,650 in 2025), with the 5-point MAGI disregard. This is the same method as the 2026 entries from Update 2026 parent Medicaid limits in six frozen-standard states #9475. Before this PR, both years still read the 2021 values (0.19, 0.32, 0.26, 0.56).
• CT. It adds 2024-10-01: 1.38 for Public Act 24-81.

It also adds 11 limit cases, 3 eligibility cases and a fixed changelog fragment.

• Law and values. Every new value matches the state source for its year:
  • TX: $230 a month (C-131.2, Revision 15-4, effective October 1, 2015).
  • FL: $303 standard plus the $183 standard disregard, which make the $486 MNIL (Appendix A-7; the $303 is dated April 1992 and the MNIL January 2014).
  • MS: $492 and $495, disregard included (Appendix A-3, COE-075, effective March 1, 2024 and March 1, 2025).
  • WY: $873 (WDH Family MAGI column; the program rests on the 1996 AFDC standard).
  • CT: 160% until October 1, 2024, then 138% (CGS 17b-261(a) as amended by PA 24-81: 155% to 133% before the disregard).
• The PR body's Mississippi caveat is resolved. The MS DOM site still times out. I read the COE-075 rows from an archived copy of the same appendix (February 2026 upload). They confirm $492 (2024) and $495 (2025) for a household of three, on a State Set Income Limit of exactly $384.
• Cross-checks. KFF's parent series is exactly the PR's method, truncated to two decimals, in every year from 2021 to 2026 for TX, FL and WY. CT matches too (1.60 in May 2024, 1.38 in January 2025). MACPAC Exhibit 36 and the CMS December 2023 table agree.
• Code and tests. The change is parameter-only. The variable reads parameters(period) at January 1 and indexes by state code, so it stays vectorized, and 2024 correctly keeps CT's 1.60. I hand-checked all 14 new expectations, and all are right.
• Partners. The PR edits no partner file. No partner case's outputs change on the merged tree (see "Verified correct").
• CI and branch. All 27 checks pass, and the merge with main is clean.

Two things need fixing:

1. The axiom: line is missing. rulespec-us doesn't encode these values.
2. The Mississippi Appendix A-3 reference has no #page= anchor, although it is a 13-page PDF.

Critical (must fix)

None.

Should address

1. Add the axiom: line to the PR body.

The body at head 97b8670071 has no line that mentions axiom in any format. Pavel is staff and this is a policy fix, so the line is required, and axiom: needed isn't available.

encoded-correct isn't available either. rulespec-us encodes these limits only from the CMS eligibility-level table as of December 1, 2023, each with a single effective_from: '2023-12-01' version:

Module (us-xx/policies/cms/<state>-medicaid-chip-bhp-eligibility-levels.yaml)	Rule	Encoded	This PR (2024 / 2025)
us-ct	connecticut_parent_caretaker_adults_medicaid_fpl_limit	1.55	1.60 / 1.38 (the PA 24-81 cut is missing)
us-tx	texas_parent_caretaker_adults_medicaid_fpl_limit	0.12	0.157 / 0.154
us-fl	florida_parent_caretaker_adults_medicaid_fpl_limit	0.24	0.276 / 0.269
us-ms	mississippi_parent_caretaker_adults_medicaid_fpl_limit	0.19	0.229 / 0.223
us-wy	wyoming_parent_caretaker_adults_medicaid_fpl_limit	0.45	0.456 / 0.443

The CMS figures are pre-disregard, use the highest share across household sizes, and are converted at 2023 guidelines. They have no 2024 or 2025 versions, and CT has no October 2024 change. No rulespec-us issue (open or closed, pe-parity or not) mentions #10151, #10252, HUSKY or the parent/caretaker group. See "Axiom line assessment" for the recommended queued line.

2. The Mississippi Appendix A-3 reference needs a page anchor.

income_limit.yaml:42-43 cites .../2026/03/Appendix-A-3-MAGI-Income-Limits.-2014-to-present.pdf with no #page=. The appendix has 13 pages, one table per year from 2026 (file page 1) back to 2014 (file page 13). In the archived February 2026 upload:

• the "2025 MAGI INCOME LIMITS EFFECTIVE MARCH 1, 2025" table (Appendix Page A-3-l), with COE-075 for a household of three "$384 $495", is file page 2;
• the "2024 MAGI INCOME LIMITS EFFECTIVE MARCH 1, 2024" table (A-3-k), with "$384 $492", is file page 3.

As it stands, the link opens on the 2026 table, which shows neither value. CLAUDE.md's reference rules call for the href to open the first cited page and for the title to end with the page list.

Fix: change the href to ...-2014-to-present.pdf#page=2 and end the quoted title with #page 2-3. Confirm the page order against the March 2026 upload you cite. The search index shows its first page as A-3-m (2026), the same as the archived file.

Suggestions

1. Pre-2024 years are still stale, and Case 24 pins one of them. The PR leaves 2021-2023 alone, as the issue scopes it. But the same standards were already in force then (TX since October 2015, FL since January 2014, MS's $384 since at least 2020, WY's 1996 AFDC base). The same method gives:

   State	Main (2021-2023)	2021	2022	2023	KFF Jan 2021 / 2022 / 2023
   TX	0.19	0.176	0.170	0.161	0.17 / 0.16 / 0.16
   FL	0.32	0.316	0.303	0.285	0.31 / 0.30 / 0.28
   MS	0.26	0.260 ($476)	0.250 ($480)	0.236 ($488)	0.25 / 0.25 / 0.28*
   WY	0.56	0.527	0.505	0.471	0.52 / 0.50 / 0.47

   *KFF's January 2023 MS figure double-counts the disregard, as the body explains for May 2024 ($488 × 12 / $24,860 = 0.236, plus 0.05 = 0.286).

   Case 24 ("Texas parent limit stays at the 2021 value of 0.19 FPL in 2023") therefore asserts a 2023 limit about 3 points above Texas's $230 standard. For example, a Texas parent at 0.17 FPL in 2023 is a Medicaid parent in PolicyEngine but not under C-131.2. Either:

   • add 2022-01-01 and 2023-01-01 entries for all four states (MS 2022/2023 from Appendix A-3 file pages 5 and 4), plus TX and WY 2021, and update Case 24; or
   • keep the scope, open a follow-up issue, and rename Case 24 to something like "Texas 2023 limit is unchanged by this PR (stale 2021 value, see #...)".

2. Connecticut: cite the statute, and describe the transition fully.

   • Add CGS 17b-261(a) as amended by PA 24-81 next to the DSS page. OLR Report 2024-R-0174 (November 26, 2024): "Laws passed in 2024 ... lower[ed] the income limit for HUSKY A parents and caretaker relatives from 155% of the federal poverty level (FPL) to 133% of the FPL". CMS SPA CT-24-0025 (approved 03/14/2025, effective October 1, 2024) is the plan amendment. It reduces "the income standard for the eligibility group HUSKY A Parents and Caretakers from 155% to 133% of the Federal Poverty Level (FPL) effective 10/1/2024. Including the 5% income disregard, this effectively changes the income limit from 160% to 138% of the FPL." Citing it, and saying that 1.38 = 133% + 5 points, would tie the value to the statute and the state plan.
   • The comment at income_limit.yaml:86-89 says the cut applies "for applicants from October 1, 2024". The DSS page also applies it to current enrollees: those above 138% "will no longer qualify", coverage "may change or end on October 31, 2024", and earners get "up to an additional year of Transitional Medical Assistance". The January 1 reading (1.60 for 2024, 1.38 from 2025) is still the right annual choice. Only the comment is narrower than the source.

3. Mississippi wording. The comment at income_limit.yaml:170-171 says "about $384". The appendix's "State Set Income Limit" column is exactly $384 for a household of three in every table from 2020 to 2026. You could say so, note that each table takes effect March 1 (January-February 2024 used $488, 0.227), and drop the outage caveat from the body.

Verified correct

• Sources, quoted:

  • Texas Works Handbook C-131.2, "Medically Needy and Parents and Caretaker Relatives Medicaid, Revision 15-4; Effective October 1, 2015", TP 08 "One Parent" column, family size 3: $230 (two parents: $251). Unchanged through 2026.
  • Florida DCF ESS Appendix A-7 (one page, updated 02/02/2026), family size 3: Adults (Parents, Caretakers) $303, standard disregard $183, MNIL $486. Effective dates: "April 1992" for the adults standard and "January 2014" for the MNIL. Note: "MNIL ... formula is Adults (Parents/Caretakers) plus Standard Disregard". So $486 held in 2024 and 2025, and the 2026 entry's $600 is $486 plus the $114 MAGI disregard.
  • Mississippi DOM Appendix A-3, COE-075 "Low-Income Parents & Caretaker Relatives", household size 3, "State Set Income Limit" / "+ 5% Disregard":
    • $384 / $492, effective March 1, 2024 (file page 3);
    • $384 / $495, effective March 1, 2025 (file page 2);
    • $384 / $498, effective March 1, 2026 (file page 1, the existing 2026 entry).
  • Wyoming. The WDH Medicaid Income Requirements page gives Family MAGI: size 1 $529, size 2 $737, size 3 $873, size 4 $999. The 2026 Healthcare Coverage Options PDF (one page), footnote 3: "Wyoming's Family Care population is based on the 1996 AFDC standard and FPL estimates are illustrative only." The reference title quotes it exactly.
  • Connecticut DSS: "A new state law (Public Act 24-81) has lowered the income limit for parents and caretaker relatives to qualify for HUSKY A health coverage from 160% to 138% of the federal poverty level (FPL). This change starts October 1, 2024." CMS approved SPA CT-24-0025 on 03/14/2025, effective 10/1/2024: "from 155% to 133% ... Including the 5% income disregard, this effectively changes the income limit from 160% to 138% of the FPL." So both 1.60 and 1.38 include the disregard, as the file's other values do.

• Cross-checks:

  • KFF (family of three, disregard included, truncated):

    State	May 2024	Jan 2025	Jan 2026	PR / existing values
    CT	1.60	1.38	1.38	1.60 / 1.38 / 1.38
    TX	0.15	0.15	0.15	0.157 / 0.154 / 0.151
    FL	0.27	0.26	0.26	0.276 / 0.269 / 0.264
    WY	0.45	0.44	0.43	0.456 / 0.443 / 0.433
    MS	0.27*	0.22	0.21	0.229 / 0.223 / 0.219

    *0.229 + 0.05, double-counted, as the body says.

  • MACPAC Exhibit 36 (pre-disregard, highest across sizes): July 2024: CT 155, FL 24, MS 19, TX 12, WY 45. 2025: CT 133, FL 22, MS 17, TX 11, WY 42. Every PR value is at or below the MACPAC figure plus 5. The one exception is MS 2025 (0.223 against 0.22), and the unrounded figure is 17.3 + 5 = 0.223.

  • The CMS Medicaid.gov eligibility-levels table (December 1, 2023) gives the same 2024 figures: CT 155%, FL 24%($), MS 19%($), TX 12%($), WY 45%($).

• Years and dates.

  • The guideline for three is 15,060 + 2 × 5,380 = 25,820 (2024) and 15,650 + 2 × 5,500 = 26,650 (2025), matching gov/hhs/fpg.yaml.
  • medicaid_parent_income_limit is a YEAR variable that reads the January 1 instant. So:
    • CT gives 1.60 for 2024 and 1.38 from 2025, and 2026+ now gets 1.38 (it was 1.60).
    • The MS March 1 tables are applied from January 1.
  • 2026 is untouched.

• Encoding form. All four frozen-dollar states are stored as a family-of-three share of the guideline, including the disregard. This is the convention Update 2026 parent Medicaid limits in six frozen-standard states #9475 set and the file header documents, including the size-two error for TX and WY. MS's figure already includes the disregard, so no 0.05 is added, which matches its 2026 entry.

• Code and vectorization.

  • The formula is unchanged: p.income_limit[state], then where for VA and MO.
  • The only state-keyed lookup is by state_code_str, so the NH Add New Hampshire state supplementation for 2026 #10156 pattern doesn't apply.
  • medicaid_income_eligible rounds the monthly limit up to a whole dollar and compares inclusively.

• Hand-checked expectations (14), all correct:

  • medicaid_parent_income_limit.yaml:
    • Case 24, TX 2023: reads the 2021 entry, 0.19 (see Suggestion 1).
    • Case 25, TX 2024: 2,760 / 25,820 = 0.1069 + 0.05 = 0.1569, so 0.157.
    • Case 26, TX 2025: 2,760 / 26,650 = 0.1036 + 0.05 = 0.1536, so 0.154.
    • Case 27, FL 2024: 5,832 / 25,820 = 0.2259 + 0.05 = 0.2759, so 0.276.
    • Case 28, FL 2025: 5,832 / 26,650 = 0.2188 + 0.05 = 0.2688, so 0.269.
    • Case 29, MS 2024: 5,904 / 25,820 = 0.2287, so 0.229.
    • Case 30, MS 2025: 5,940 / 26,650 = 0.2229, so 0.223.
    • Case 31, WY 2024: 10,476 / 25,820 = 0.4057 + 0.05 = 0.4557, so 0.456.
    • Case 32, WY 2025: 10,476 / 26,650 = 0.3931 + 0.05 = 0.4431, so 0.443.
    • Case 33, CT 2024: the January 1 instant gives the 2021 entry, 1.60.
    • Case 34, CT 2025: the 2024-10-01 entry, 1.38.
  • is_parent_for_medicaid.yaml (Medicaid household of 2):
    • CT 2024 at 1.414: ceil(20,440 × 1.60 / 12 = 2,725.33) = 2,726, so the effective limit is 1.6004 and the parent is true. The child is false (not head or spouse).
    • CT 2025 at 1.414: ceil(21,150 × 1.38 / 12 = 2,432.25) = 2,433, so the effective limit is 1.3804 and the result is [false, false].
    • FL 2025 at 0.30: ceil(21,150 × 0.269 / 12 = 474.11) = 475, so the effective limit is 0.2695 and the result is false. Under the old 0.32 the effective limit is 0.3200, so the parent qualified and the case tells the two apart.
  • The case numbers (24-34) continue main's 23 without collision.

• Partner contract tests. The PR doesn't touch policyengine_us/tests/policy/baseline/partners/**, and main hasn't changed that folder since the merge base. TX is the only one of the five states in any partner fixture (250 TX cases; none for CT, FL, MS or WY). Of the 17 TX cases with a 2024 or 2025 period (all 2025):

  • the ones that output medicaid, is_medicaid_eligible or aca_ptc are childless single adults or a married couple with a child at 2.09 FPL ($55,600 for 3), so the parent limit can't bind;
  • the zero- and low-income parent cases (TANF two-parent, Head Start, Early Head Start, CTC at $2,600) are under both 0.19 and 0.154 and don't output Medicaid variables.

  No partner output changes on the merged tree, and Household API Partners passed at this head. No Ziming Hua sign-off is needed.

• Source-conflict ratchet (Tie source-conflict markers to the register shared with Axiom #10287). None of the added comments matches CONFLICT_COMMENT_RE or ATTEMPT_RE. The merged income_limit.yaml has 0 matching lines, the same as the merge base, and isn't in KNOWN_UNMARKED. The test YAML is outside the scanned roots. CI ran on 2026-10-23 after Tie source-conflict markers to the register shared with Axiom #10287 and Count the FL and SD LIHEAP comments in the source-conflict ratchet baseline #10334.

• Changelog. changelog.d/fix-10151-parent-medicaid-limits-2024-2025.fixed.md is a top-level fragment with the fixed type.

• References. The WY PDF and the FL appendix are single-page PDFs, and the CT, WDH and KFF references are HTML pages, so none needs an anchor. Each title corroborates its value. The one missing anchor is Should 2.

• Issue closure. Fixes #10151 is right. The issue asks for exactly the 2024-2025 TX, FL, MS and WY values and CT's October 2024 cut.

Axiom line assessment

The line is missing (Should 1). rulespec-us doesn't encode the provision correctly (see the table in Should 1), and no parity issue exists.

Recommended line. File a dispatch-ready pe-parity issue and use axiom: TheAxiomFoundation/rulespec-us#<n> queued. The issue should have:

• Modules.
  • Add a dated version to us-ct/policies/cms/connecticut-medicaid-chip-bhp-eligibility-levels.yaml, or better, a CT statute module for CGS 17b-261(a) as amended by PA 24-81 (133% FPL from October 1, 2024).
  • Add state-standard modules for the four frozen-dollar states. Suggested paths: us-tx/policies/hhsc/texas-works-handbook/c-131-2, us-fl/policies/dcf/ess-appendix-a-7, us-ms/policies/dom/eligibility-manual/appendix-a-3, us-wy/policies/wdh/medicaid-income-requirements. Each should hold the monthly dollar standard by household size.
• Corpus. axiom-corpus already holds the MS state eligibility manual lock (which lists appendix-a-3.pdf) and the WY health guidance lock (income-requirements.html). Check that TX C-131.2, FL Appendix A-7 and the CT text are ingested before dispatch.
• Law. Quote verbatim the C-131.2 table row (TP 08, size 3 $230), the Appendix A-7 adults, standard disregard and MNIL row ($303 / $183 / $486), the Appendix A-3 COE-075 rows ($384 / $492 for 2024 and $384 / $495 for 2025), the WDH Family MAGI row ($873), and the CT DSS / PA 24-81 text.
• Required outputs. The parent/caretaker income limit, as a monthly dollar standard by household size or as a share of the guideline for a family of three, for 2024 and 2025. For CT, the limit before and after October 1, 2024.
• Companion cases. Compute them by hand from the state tables, not from PolicyEngine output:
  • a Texas one-parent family of 3 with $230/month of MAGI income after the disregard qualifies, and $231 doesn't;
  • a Mississippi household of 3 at $495/month qualifies in 2025, and $496 doesn't;
  • a Connecticut parent of one at 141% FPL qualifies in September 2024 but not in November 2024.

CI status

All 27 checks pass at 97b8670071 (gh pr checks 10252). The workflow run on that SHA was created 2026-10-23 18:42 UTC. That includes:

• the 11 Baseline shards (incl. states shards 1-4);
• the 9 Contrib shards;
• Household API Partners;
• PR validation (lock, guard tests, lint, changelog);
• Microsimulation;
• the two Full Suite - Rest jobs;
• Package and compatibility;
• Quick Feedback.

None is pending or failing. I didn't run tests locally: CI ran the PR's two test files green at this head.

Branch status

The branch is 2 commits ahead of main (the change on 2026-10-16 and a merge of main on 2026-10-23) and 48 behind. git merge-tree --write-tree PolicyEngine/main 97b8670071 merges cleanly.

• Since the merge base, main has changed medicaid_person_is_required_to_file.py. It feeds Medicaid household income, but the PR's tests set medicaid_income_level directly, so they don't depend on it.
• Main has also changed test_source_conflicts.py, in rows unrelated to this file.
• Main hasn't touched the parent income-limit parameter, its variables or its tests.
• No other open PR touches medicaid/eligibility/categories/parent.

Verdict: REQUEST_CHANGES

[DTrim99]

All values verify: TX $230, FL $486, MS $492/$495, WY $873, and CT 160% -> 138% from 10/1/2024 (PA 24-81; SPA CT-24-0025). They match KFF exactly, and all 14 expectations hand-check; no partner output changes. Two should-address items: (1) there's no axiom line, and rulespec-us has only 2023 values and no CT cut, so please file a pe-parity issue and mark it queued; (2) the 13-page MS Appendix A-3 reference needs '#page=2' with a '#page 2-3' title. Three suggestions, including the still-stale 2021-2023 values. Full report in the comment above.

[PavelMakarchuk]

Thanks @DTrim99. Each item from the 2026-10-26 review, at head 4d96379814:

Should address 1, axiom: line: fixed. I filed TheAxiomFoundation/rulespec-us#1951 (pe-parity) and added axiom: TheAxiomFoundation/rulespec-us#1951 queued to the body. The issue has:

• the module paths: TX C-131.2, FL Appendix A-7, MS Appendix A-3, WY WDH Family MAGI, CGS 17b-261 and SPA CT-24-0025;
• the verbatim law and the required outputs;
• 15 companion cases worked by hand from the state tables and the HHS guidelines, including your TX $230/$231, MS $495/$496 and CT September/November 2024 cases.

TX C-130, FL A-7, MS A-3 (the same March 2026 URL), the WDH page and the SPA CT-24-0025 approval package are already in axiom-corpus manifests. The recorded blockers: PA 24-81 / CGS 17b-261 and the 2026 Wyoming coverage-options PDF are not ingested, the signed release pin was not checked, and no billed encoder run is approved.

Should address 2, Mississippi Appendix A-3 anchor: fixed in 4d96379. The href ends in #page=2 and the quoted title ends in #page 2-3. I checked the page order in the archived February 2026 upload: page 2 is "2025 MAGI INCOME LIMITS EFFECTIVE MARCH 1, 2025" (A-3-l, $384 / $495) and page 3 is the 2024 table (A-3-k, $384 / $492). medicaid.ms.gov still times out, so I could not open the March 2026 upload itself. axiom-corpus lists that URL with source_as_of 2026-09-10, so the file exists.

Suggestion 1, stale pre-2024 values: kept the scope and filed #10491. Case 24 is now "Texas 2023 parent limit still reads the 2021 entry of 0.19 FPL". A comment above it gives the $230-based 0.161 and points to #10491. I did not backfill 2021-2023 here because the gap starts in 2018. The 2018 entries are also above the standards by the same method: TX 0.20 against 0.183, FL 0.34 against 0.331, MS 0.28 against 0.272 ($471) and WY 0.59 against 0.554. A 2021-2023 backfill would leave the same error in 2018-2020, so #10491 covers 2018-2023 in one change. It has the full table and the MS file pages 4-9 ($471 to $488). No partner case for these four states has a 2018-2023 period.

Suggestion 2, Connecticut statute and transition: fixed in 4d96379. I added two references:

• Public Act 24-81, section 38, file pages 68-69. It amends CGS 17b-261(a) "(Effective October 1, 2024)" to "one hundred [fifty-five] thirty-three per cent".
• SPA CT-24-0025, pages 2, 5 and 11.

The comment now says that 1.38 is 133% plus 5 points. It also says that the cut covered current enrollees: coverage changed or ended on October 31, 2024, and earners could keep up to one more year of Transitional Medical Assistance. I did not cite the OLR report, because the act and the SPA are the primary sources.

Suggestion 3, Mississippi wording: fixed in 4d96379. The comment now says three things:

• the State Set Income Limit is exactly $384 in every table from 2020 to 2026 (checked on pages 1-7);
• each table takes effect March 1;
• January and February 2024 still used the 2023 table's $488 (0.227).

The outage caveat is gone from the body.

Tests. No expected value changed; Case 24 changed only its name. Local runs, all passing:

• the two changed test files (64 cases);
• gov/hhs/medicaid/eligibility (30 files, test_batched.py);
• the reference-URL, parameter-key, dead-URL, concatenated-URL and source-conflict code-health tests (257);
• test_parameter_files.py (8).

[DTrim99]

PR Review (round 2): Set 2024-2025 parent Medicaid limits for TX, FL, MS and WY, and Connecticut's 138% limit from October 2024

Summary

This round covers one author commit since our 2026-10-26 review: 4d96379814 (94 lines), which changes comments, references and one test name. No parameter value or expected output changed, so the values I verified in round 1 still stand:

• TX 0.157 / 0.154;
• FL 0.276 / 0.269;
• MS 0.229 / 0.223;
• WY 0.456 / 0.443;
• CT 2024-10-01: 1.38.

I re-read each state source for this round. Both Should items from round 1 are fixed:

1. Axiom line. The body now has axiom: TheAxiomFoundation/rulespec-us#1951 queued. The issue is OPEN, labelled pe-parity, dispatch-ready and written against head 4d96379814.
2. Mississippi anchor. The Appendix A-3 reference now opens #page=2, and its quoted title ends #page 2-3. The author could only check the February 2026 upload. I checked the March 2026 upload that the reference cites, using axiom-corpus's copy (sha256 70f1e370..., which matches the corpus lock). File page 1 is the 2026 table (A-3-m). Page 2 is "2025 MAGI INCOME LIMITS EFFECTIVE MARCH 1, 2025" (A-3-l, household of 3: $384 / $495). Page 3 is the 2024 table (A-3-k: $384 / $492). The anchor is right.

The three suggestions:

• Connecticut statute and transition, and Mississippi wording: both done. The CT references now cite the primary sources, Public Act 24-81 section 38 and SPA CT-24-0025, with page anchors I checked.
• Pre-2024 values: declined with a good reason and tracked in Parent/caretaker Medicaid limits: 2018-2023 values for TX, FL, MS and WY are rounded shares above each state's dollar standard #10491. Case 24 was renamed to say it reads a stale entry.

No new findings. The source-conflict ratchet holds, and the merge with main is clean. Main has not touched the parent income-limit file, its variables or its tests.

CI at this head has 28 pass, 2 fail and 3 pending. Both failures are the known main-caused Montana AGI invariant (Rest Python shard 3, fix #10496) and its aggregator.

Prior findings

Finding (round 1)	Status	Evidence
Should 1: add the axiom: line to the PR body	Fixed	The body ends axiom: TheAxiomFoundation/rulespec-us#1951 queued. rulespec-us#1951 is OPEN and labelled pe-parity (created 2026-10-27 15:18 UTC). It describes head 4d96379814 and has everything the queued form needs: module paths, corpus citations, verbatim law, required outputs, 15 companion cases worked by hand from the state tables, and the blockers. See "Axiom line assessment".
Should 2: MS Appendix A-3 reference needs a page anchor	Fixed	income_limit.yaml:42-43: the href is ...-2014-to-present.pdf#page=2 and the title is quoted, ending #page 2-3, which follows CLAUDE.md's reference rules. I confirmed the page order against the March 2026 upload, as described in the Summary.
Suggestion 1: pre-2024 values are stale and Case 24 pins one	Declined with reason (accepted)	The author kept #10151's 2024-2025 scope and filed #10491 (OPEN). It tables the same method for 2018-2023 (TX 0.183-0.161, FL 0.331-0.285, MS $471-$488, WY 0.554-0.471) with sources and MS file pages 4-9. The stated reason holds: the gap starts in 2018, so a 2021-2023 backfill would leave 2018-2020 wrong, and one change should cover all six years. Case 24 is now "Texas 2023 parent limit still reads the 2021 entry of 0.19 FPL", with a comment giving the $230-based figure, 2,760 / 24,860 + 0.05 = 0.161, and pointing to #10491. The test now says what it pins.
Suggestion 2: cite the CT statute and describe the transition fully	Fixed	Two references added. (a) Public Act 24-81, section 38, #page=68, title #page 68-69. File page 68 has "Sec. 38. Subsection (a) of section 17b-261 ... is repealed and the following is substituted in lieu thereof (Effective October 1, 2024)". File page 69 has "their parents and needy caretaker relatives, who qualify for coverage under Section 1931 ... with household income not exceeding one hundred [fifty-five] thirty-three per cent of the federal poverty level". (b) SPA CT-24-0025, #page=2, title #page 2,5,11. Page 2: "We approve Connecticut State Plan Amendment (SPA) CT-24-0025 with an effective date(s) of October 01, 2024." Page 5: "from 155% to 133% of the Federal Poverty Level (FPL) effective 10/1/2024. Including the 5% income disregard, this effectively changes the income limit from 160% to 138% of the FPL." Page 11: "FPL 133.00%". Approval Date 03/14/2025 is on every form page. The CT comment (income_limit.yaml:90-98) now says 1.38 = 133% + 5 points and that the cut also reached current enrollees (October 31, 2024, with up to a year of Transitional Medical Assistance). Leaving out the OLR report is fine, because the act and the SPA are the primary sources.
Suggestion 3: MS wording ("about $384", March 1 dates, outage caveat)	Fixed	The MS comment (income_limit.yaml:176-184) now says the State Set Income Limit is exactly $384 in every table from 2020 to 2026. The March 2026 copy confirms it, household of 3: $384 / $498 (2026, p1), $495 (2025, p2), $492 (2024, p3), $488 (2023, p4), $480 (2022, p5), $476 (2021, p6) and $475 (2020, p7). The comment also notes the March 1 effective dates and that January-February 2024 used $488: 5,856 / 25,820 = 0.227. The outage caveat is gone from the body.

New findings

None (0 critical, 0 should, 0 suggestions).

Verified

• Sources re-read this round (unchanged values, family of three):
  • Texas Works Handbook C-131.2, "Revision 15-4; Effective October 1, 2015": TP 08 One Parent $230 (Two Parents $251).
  • Florida DCF ESS Appendix A-7 (Updated 02/02/2026): size 3 adults standard $303 and standard disregard $183. MNIL $486, dated January; the adults standard is dated April.
  • Wyoming WDH Medicaid Income Requirements: Family MAGI size 3 $873.00 (size 1 $529, size 2 $737).
  • Mississippi and Connecticut: as quoted in the table above.
• Expectations. The commit changes only Case 24's name and adds a comment above it. Its expected value is still 0.19, read from the 2021 entry (2023 < 2024-01-01). No other expectation changed. I re-checked all 14 again:
  • TX: 2,760 / 25,820 + 0.05 = 0.1569 → 0.157; 2,760 / 26,650 + 0.05 = 0.1536 → 0.154.
  • FL: 5,832 / 25,820 + 0.05 = 0.2759 → 0.276; 5,832 / 26,650 + 0.05 = 0.2688 → 0.269.
  • MS: 5,904 / 25,820 = 0.2287 → 0.229; 5,940 / 26,650 = 0.2229 → 0.223.
  • WY: 10,476 / 25,820 + 0.05 = 0.4557 → 0.456; 10,476 / 26,650 + 0.05 = 0.4431 → 0.443.
  • CT limit cases: 1.60 (2024, January 1 value) and 1.38 (2025).
  • Eligibility cases:
    • CT 2024 at 1.414: ceil(20,440 × 1.60 / 12) = 2,726, so [true, false].
    • CT 2025: ceil(21,150 × 1.38 / 12) = 2,433, under 1.414 × 21,150 / 12 = 2,492.18, so [false, false].
    • FL 2025 at 0.30: ceil(21,150 × 0.269 / 12) = 475, under 528.75, so [false, false].
• New comment arithmetic. TX 2023 is 0.161 and MS January-February 2024 is $488 → 0.227. Both are correct.
• Reference format. All three new or changed multi-page references put the first page in the href and end a quoted title with #page 2-3, #page 68-69 or #page 2,5,11. That matches CLAUDE.md, with no page lists in an href. Each title corroborates its value.
• Source conflicts (CLAUDE.md "Conflicting sources"). No conflict choice is made:
  • CGS 17b-261(a) (133%), the SPA (133.00%, or 138% with the disregard) and the DSS page (138%) agree once the 5-point disregard is added.
  • TX, FL, MS and WY: the manual tables match MACPAC Exhibit 36 and the CMS levels table, which are pre-disregard and use the highest share across sizes.
  • So no register entry or # Source conflict marker is needed.
• Ratchet (Tie source-conflict markers to the register shared with Axiom #10287). On the merged tree, income_limit.yaml has 0 lines that match CONFLICT_COMMENT_RE or ATTEMPT_RE, and it isn't in KNOWN_UNMARKED. The new words ("formerly", "although ... still applied") are in a reference title or don't match any pattern. The test YAML is outside the scanned roots. CI ran after Tie source-conflict markers to the register shared with Axiom #10287 and Count the FL and SD LIHEAP comments in the source-conflict ratchet baseline #10334.
• Callers and vectorization. medicaid_parent_income_limit (YEAR, p.income_limit[state_code_str], then where for VA and MO) is the only reader. is_parent_for_medicaid_fc is its only consumer. Neither changed on main or in this PR, and the NH Add New Hampshire state supplementation for 2026 #10156 pattern doesn't apply. programs.yaml already has Medicaid at verified_start_year: 2021, so no registry change is needed.
• Partners. The PR doesn't touch policyengine_us/tests/policy/baseline/partners/**, and main hasn't changed that folder since the merge base. Household API Partners passed at this head. No Ziming Hua sign-off is needed.
• Changelog. changelog.d/fix-10151-parent-medicaid-limits-2024-2025.fixed.md is a top-level fragment with the fixed type.
• PR body. The test-count claim holds. On main, 9 limit cases (25-32 and 34) and 2 eligibility cases (CT 2025, FL 2025) fail. The 3 controls pass: Case 24, Case 33 and the CT 2024 eligibility case.

Axiom line assessment

axiom: TheAxiomFoundation/rulespec-us#1951 queued is valid:

• State. OPEN, labelled pe-parity, written against PR head 4d96379814, and it lists all three PR commits.
• Dispatch-ready, on the rulespec-us#1467 model:
  • Module paths. us-tx/policies/hhs/texas-works-handbook/c-130-medical-programs/c-131-2, us-fl/policies/dcf/ess-program-policy-manual/appendix-a-7-.../page-1, us-ms/policies/dom/medicaid/appendix-a-3, us-wy/policies/wdh/medicaid-income-requirements, us-ct/statutes/17b-261 and us-ct/policies/cms/medicaid-state-plan/spa/ct-24-0025. It also maps them to the existing us/regulations/42-cfr/435/110 and 435/603/d.
  • Corpus citations. Each module has a manifest name and an ingest status.
  • Verbatim law from all five states, including PA 24-81 section 38 and SPA pages 2, 5 and 11.
  • Required outputs with the policyengine-us variables, entities and periods.
  • 15 companion cases worked from the state tables and HHS guidelines, not from PolicyEngine output. They include our TX $230/$231, MS $495/$496 and CT September/November 2024 cases. I spot-checked these: F (FL size 2, $387 + $88.13 = $475.13 < $528.75), M (CT, 1.414 × 20,440 / 12 = $2,408.51 against $2,725.33), N ($2,350.60) and O ($2,492.17 against $2,432.25). All are right.
• Blocker (needed for queued). PA 24-81 / CGS 17b-261 and the 2026 Wyoming coverage-options PDF aren't in any axiom-corpus manifest. The signed release pin wasn't checked, and no billed encoder run has been approved.
• Gaps. The issue records the PE approximations (family-of-three share, January 1 reads of the MS March and CT October dates, the WY disregard assumption) as expected oracle mismatches. It also states "Source conflict (none)", which matches my reading.

CI status

Workflow run on 4d96379814, created 2026-10-27 15:17 UTC. Re-checked with gh pr checks 10252 at 16:16 UTC: 28 pass, 2 fail, 3 pending.

• Python tests / Full Suite - Rest (Python shard 3): fail. Main-caused, so a CI note, not a finding.
  • The log shows FAILED policyengine_us/tests/core/test_mt_state_agi_invariants.py::test_drawn_batches, with "1 failed, 1945 passed".
  • The failing draw is MT, year=2021, with a pension=12000 person. Actual 7,600 against desired 12,000 is the $4,400 Montana pension exemption.
  • This is the known failure on main since Add Montana's 2021–2023 partial pension, annuity, and IRA income exemption #10392. The fix is Check Montana's pension exemption in the Montana AGI pooling test #10496 (open), and the PR doesn't touch MT AGI. Merge main and re-run once Check Montana's pension exemption in the Montana AGI pooling test #10496 lands.
• Python tests / Check Rest Python shards: fail. It's the aggregator ("Rest jobs did not all succeed: plan=success, shards=failure") and follows from shard 3. Same note.
• Pending (queued, not started): Contrib states-shard-1, states-shard-2 and other-shard-1. All three passed at 97b8670071, and the only change since is comments, references and a test name.
• Passing, among others:
  • all 11 Baseline shards. contrib-hhs runs gov/hhs, including both PR test files, and passed all 5 batches;
  • Household API Partners;
  • Quick Feedback;
  • PR validation (lock, guard tests, lint, changelog);
  • Microsimulation;
  • Rest Python shards 1, 2, 4, 5 and 6, which run the code-health reference, URL and source-conflict tests;
  • the other Contrib shards.

I didn't run tests locally: CI ran the PR's test files green at this head, and no expectation changed.

Branch status

The branch is 3 commits ahead of main (8cbbbf5614, the merge 97b8670071 and 4d96379814) and 70 behind.

• Merge. git merge-tree --write-tree PolicyEngine/main 4d96379814 merges cleanly (tree c9ddfce184). The merged income_limit.yaml is byte-identical to the head's.

• Main's Medicaid changes since the merge base:

  • pregnant/postpartum_coverage.yaml;
  • the is_pregnant_for_medicaid.yaml tests;
  • medicaid_person_is_required_to_file.py and its test.

  None is in the parent folder or feeds the PR's cases, which set medicaid_income_level directly.

• Main-side guards:

  • Integrate modeled state unemployment insurance into income flows #9558 (test_unemployment_compensation_readers.py) and Count military SBP annuities in state benefit income lists #10317 (test_military_retirement_readers.py) are on main;
  • Count each Social Security benefit once in Pennsylvania CCW countable income #10408 (test_social_security_source_lists.py) is still open.

  The PR touches no income-source list or reader, so none of them applies.

• Sibling PRs. None of the open Medicaid PRs (Extend standard pregnancy CHIP through the postpartum period #10363, Update Medicaid aged, blind and disabled pathway disregards and asset limits (CA, CT, MO, AR, OK, SC) #10364, Correct CHIP and Medicaid pregnancy income limits and effective dates #10365, Count blind people and use the individual or couple in the Medicaid medically needy test #10370, Update Medicaid medically needy limits and switch on AR, IA, IL, NJ and ND #10374, Limit Arkansas's Medicaid aged, blind and disabled pathway to aged people, and apply Maine's savings exclusion and couple disregard #10386) touches the parent folder or is_parent_for_medicaid.yaml.

Verdict: APPROVE

[DTrim99]

Approving round 2: 0 critical, 0 should. Full report in the comment above. CI note: the Rest shard 3 failure (test_mt_state_agi_invariants) comes from main and is fixed by #10496; re-run after it lands.