Skip to content

Correct 2018–2020 Medicaid parent limits in UT, TN and VA and adult-group starts in UT and VA - #10874

Merged
MaxGhenis merged 11 commits into
mainfrom
medicaid-parent-limits-ut-tn-va-2018-2020
Nov 15, 2026
Merged

MaxGhenis merged 11 commits into
mainfrom
medicaid-parent-limits-ut-tn-va-2018-2020

Conversation

@MaxGhenis

@MaxGhenis MaxGhenis commented Nov 13, 2026 •

Copy link
Copy Markdown
Contributor

Correct Medicaid parent income limits for Utah, Tennessee and Virginia in 2018–2020, and start the Utah and Virginia adult groups on their coverage dates. The model denied some Utah parents in 2018–2020 and some Virginia parents in 2019–2020, and admitted Tennessee parents in 2018–2020 and Virginia parents in 2018 above their state standards.

Virginia began covering the new adult group, with income under 133% of the poverty guideline, on January 1, 2019 (CMS approval, page 8). Utah covered adults up to 95% from April 1, 2019 and up to 133% from January 1, 2020 (CMS approval, pages 7 and 19). The model started both groups on January 1, 2021. Start them on these dates, so Virginia parents in 2019–2020 and Utah parents in 2020 above the parent standard, but within the adult limit, qualify through the adult group.

Set Utah's raw parent limit to 0.60 in 2018–2019, using its approved 55% standard (UT-16-0025, page 7; DAR 41429) and the model's five-point convention. Use its restored dollar schedule (DAR 43706, made permanent by DAR 43796) in 2020.

Tennessee uses its fixed household-size dollar schedule (TN-14-0005, page 19; the same table in TN-21-0010, section H, effective October 1, 2021). Virginia uses all three locality groups from its dated 2017, 2018, 2019 and 2020 manuals. The model adds 0.05 to these raw parent ratios under its existing category convention.

Closes #10793

Invariants and tests

  • For calendar-year calculations, only the Utah, Tennessee and Virginia parent limits in 2018–2020, the Virginia adult limit in 2019–2020 and the Utah adult limit in 2020 change. Virginia's adult limit starts at 1.38 on January 1, 2019. Utah's starts at 1.00 on April 1, 2019 and rises to 1.38 on January 1, 2020. Other states and years keep their existing routes, including Missouri's override, Florida's and Texas's routes and Virginia's later schedule.
  • Limits stay positive for valid household sizes. Dollar allowances do not fall as household size grows. For a fixed household size, a higher Medicaid income level cannot restore parent financial eligibility.
  • medicaid_parent_income_limit.yaml adds eleven household cases, nine three-person boundary cases and fifteen schedule vectors. The vectors cover sizes 1–22 and both sides of each effective financial threshold.
  • The Virginia 2019, Virginia 2020 and Utah 2020 households expect is_adult_for_medicaid and is_medicaid_eligible true for the parent. Two more Utah 2020 households earn $9,500, which only the adult group admits, and $25,000, which only the 1.38 adult limit admits.
  • test_historical_limits_bound_income_and_preserve_size_schedules checks source amounts, income monotonicity and household-size monotonicity.
  • test_parent_limits_outside_historical_correction_equal_existing_routes checks other states, compatibility amounts and year boundaries. Florida's expected value comes from fl_medicaid_parent_income_limit in every year Florida's route applies, as Missouri's comes from its own variable.

Three-person households have annual guidelines of $20,780, $21,330 and $21,720 in 2018, 2019 and 2020. Dollar schedules convert to a raw limit with monthly_amount * 12 / guideline + 0.05.

State 2018 raw limit 2019 raw limit 2020 raw limit Parent eligibility in the issue's household The parent's Medicaid eligibility
Utah 0.600000 0.600000 0.424586 true, true, false true, true, true (2020 through the adult group)
Tennessee 0.980318 0.956329 0.940055 false, false, false false, false, false
Virginia Group II 0.383205 0.381364 0.384807 false, false, false false, true, true (2019–2020 through the adult group)

The Virginia household's adult limit reaches $29,436 a year in 2019 ($21,330 × 1.38 / 12 = $2,452.95, rounded up to $2,453 a month) and $29,976 in 2020; the parent earns $9,351. Utah's 2020 adult limit also reaches $29,976; the parent earns $11,429.

Validation

Pull request run 43676490127 at 41230025fe passed all 37 jobs on GitHub's merge with main a29c73d587.

  • All 76 cases in medicaid_parent_income_limit.yaml, both property tests in test_medicaid_parent_historical.py and the 103 source-conflict guard tests pass on this head. Each of the 35 cases this PR adds asserts a parent limit that main doesn't return, so all 35 fail on main.
  • CI runs test_medicaid_parent_historical.py in the selective job and in the test-other-python-rest-core group. In the selective job, the file passes in 14 seconds.

Methodology choices

  • Utah's adult limit on January 1, 2020 has two readings. Under CMS's approval, Utah covered adults up to 133% from that date. In S.B. 96, Utah's legislature set full expansion to begin once CMS approved a qualifying waiver, or by July 1, 2020 (Utah Code 26-18-3.9(5)(b)). Utah's rule, R414-312-6(3), still read 95%. Utah proposed 133% in DAR 52479, effective as early as March 9, 2020. The model follows CMS and the statute with 1.38; the rule's reading gives 1.00. Record the conflict in source_conflicts/candidates.yaml as us-ut:medicaid-adult-expansion-limit-january-2020, and pin the 1.38 reading with the $25,000 Utah case.
  • Keep the five-point disregard on the Tennessee, Utah and Virginia parent limits in 2018–2020, following the model's parent-category convention. Under 42 CFR 435.603(d)(4), a state subtracts the five points only for the eligibility group with the highest MAGI income standard. On that reading, the adult limit becomes the highest standard in Virginia 2019–2020 and Utah 2020, and those parent limits would fall by 0.05. A parent in that band whom the adult group also covers would stay eligible, as an adult instead of a parent. A caretaker outside the adult group, such as one aged 65 or older, would lose eligibility in that band.
  • Tennessee's and Virginia's 2018–2020 routes and Utah's 2020 route read the shared gov.hhs.medicaid.eligibility.categories.parent.fpl_disregard. Utah's 0.60 for 2018–2019 holds the five points in the national table. Florida's and Texas's routes read their own fpl_disregard entries, each cited to its state manual. Missouri reads gov.states.mo.dss.mhf.income_limit.fpl_disregard, which ends once Missouri's adult group becomes the highest standard, so it cannot share an entry that never ends.
  • Annual calculations use the schedule effective on January 1 and that year's federal poverty guideline. Utah keeps 0.60 for 2019; its May 2019 dollar table first affects 2020 annual calculations. Utah's April 1, 2019 adult row affects no annual result, since January 1, 2020 replaces it. Virginia's July 2020 amounts reach no annual result either, since the January 1, 2021 entries replace them.
  • Exact dollar schedules replace national summary percentages for Tennessee and Virginia in 2018–2020, and Utah in 2020. One statewide percentage would drop household-size and locality differences. Utah's approved 55% standard applies across household sizes in 2018–2019.
  • Tennessee's table supplies no additional increment, so the model keeps the size-20 amount above size 20. Utah adds $62 a month for each person above 16; Virginia adds each group's published increment for each person above 8.
  • The existing financial helper stabilizes monthly amounts at cents, rounds up to whole dollars, and compares inclusively at float32 precision. Boundary tests keep that convention; raw limit assertions stay separate.
  • For Utah, the model follows the approved plan and amended rule where CMS's 2018 summary reports 44%. For Virginia, it follows DMAS's dated manuals where 12VAC30-40-220 lists lower amounts, as specified in The model denies Medicaid in 2018 to Utah parents between 49% and 60% of the poverty guideline, whom Utah covered, and covers Virginia and Tennessee parents above their states' limits #10793. Record both conflicts in source_conflicts/candidates.yaml, with source captures and deciding comments.
  • Compatibility entries keep Utah's pre-2018 national value and Virginia's direct helper outside 2018–2020. These entries keep prior behavior; they do not verify those other years.
  • Virginia tests set Group II explicitly for the Norfolk example. Existing locality classification stays in use.
  • Medicaid's program-wide verified start year stays 2021. This change verifies historical parent standards and two adult-group start dates, not every Medicaid provision for those years.
  • The first property test disables Hypothesis's input-generation timing health check, as existing core tests do.

Microsimulation impact

We estimate no change in 2026 on the default dataset, where none of the seven measured outputs moves. This PR changes Medicaid limits and start dates for 2018–2020. We compared main at 61c743cb61 with this branch at 41230025fe.

Output Main total Change Records changed (weighted)
is_medicaid_eligible 75,766,036 people 0 0 (0)
medicaid_parent_income_limit $245.1m $0 0 (0)
va_medicaid_lifc_income_limit $3.1m $0 0 (0)
state_income_tax $576.5bn $0 0 (0)
income_tax $2,594.9bn $0 0 (0)
household_benefits $2,060.2bn $0 0 (0)
household_net_income $17,246.6bn $0 0 (0)

Axiom parity

axiom: TheAxiomFoundation/rulespec-us#2248 queued (42 CFR 435.110 and 435.119 with the Utah, Tennessee and Virginia parent standards for 2018–2020; the corpus lacks the states' sources and the poverty guidelines for those years, so the signed encoder can't run yet)

🤖 Generated with Claude Code

MaxGhenis and others added 11 commits November 13, 2026 08:14
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…inia

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Keep main's Virginia, Florida and Texas parent routes, add the Tennessee
and Utah routes after them, and keep Missouri's override last. Keep
main's parent YAML cases and candidates, then this branch's appended
cases and the two candidates.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Virginia began covering the adult group, with income under 133% of the
poverty guideline, on January 1, 2019 (CMS approval of July 9, 2020,
page 8). Utah covered it to 95% from April 1, 2019 and to 133% as of
January 1, 2020 (CMS approval of December 16, 2020, pages 7 and 19).
These rows match those in #10826.

The Virginia 2019-2020 and Utah 2020 parents that fail the lower parent
standard now qualify through the adult group, so the three household
cases expect eligibility. Two new Utah 2020 cases cover a parent at
$9,500 and one at $25,000; the second qualifies only under 1.38.

Utah's rule read 95% until DAR 52479, so the January 2020 value goes
into the candidates file as an open item, with captures of the four
sources.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Main's Florida route replaces the national parent ratio from 2017, while
its indicator is on. The property test now reads that indicator at each
year's January 1 and takes Florida's expected value from
fl_medicaid_parent_income_limit, as it does for Missouri.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Add DAR 43796 and its effective-date notice beside the Utah emergency
rule, and date the Utah amounts May 7, 2019, when DAR 43706 took effect.
The in_effect indicator still turns the Utah route on in 2020 only.

Add TN-21-0010, effective October 1, 2021, beside TN-14-0005; both show
the same section H table. Describe what each state uses in the in_effect
descriptions and what the model does in the shared disregard. Note that
the 2021-01-01 entries replace Virginia's July 2020 amounts before any
annual result reads them.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@MaxGhenis MaxGhenis changed the title Correct 2018–2020 Medicaid parent limits in UT, TN and VA Correct 2018–2020 Medicaid parent limits in UT, TN and VA and adult-group starts in UT and VA Nov 14, 2026
@MaxGhenis
MaxGhenis marked this pull request as ready for review November 14, 2026 21:39
@MaxGhenis MaxGhenis closed this Nov 15, 2026
@MaxGhenis MaxGhenis reopened this Nov 15, 2026
@MaxGhenis MaxGhenis closed this Nov 15, 2026
@MaxGhenis MaxGhenis reopened this Nov 15, 2026
@MaxGhenis
MaxGhenis merged commit 7cffc3f into main Nov 15, 2026
85 of 89 checks passed
@MaxGhenis

Copy link
Copy Markdown
Contributor Author

US + core hub merge audit, us#10874 at 41230025feebab7af53877243e4293535ed02780 (squash): Correct 2018–2020 Medicaid parent limits in UT, TN and VA and adult-group starts in UT and VA.

  • Reviews:
    • The PR has no GitHub reviews. Two independent review rounds ran outside GitHub, with a fix round between them.
    • The first round, at c37557162b, read 42 CFR 435.603(d)(4), DMAS's Appendix 3 tables for July 2017 to July 2020, 12VAC30-40-220, Utah DAR 41429, 43706, 43708, 43796 and 44129, Tennessee's TN-21-0010 section H and CMS's eligibility-levels table. It recomputed 689 historical rows in the YAML file against the source tables and found no mismatch. It ran 11 households on the merge base and on the head.
    • It found that the corrected parent limits denied Medicaid to Virginia parents in 2019–2020 and Utah parents in 2020 whom each state's adult group covered, and that three cases asserted those denials. For example, a Virginia Group II parent earning $9,351 went from eligible to ineligible. The branch also conflicted with main, main's Florida route would have failed the second property test, and the property file had never run against the full country system. Six minor findings asked for Utah's permanent rule as a source and a May 7, 2019 date on Utah's amounts. They also asked for one account of the disregard's copies, parameter descriptions that match the code and a comment on Virginia's July 2020 rows.
    • The fix round, through 41230025fe, merged main at 61c743cb61 and kept both sides of three conflicts. It added the Utah and Virginia adult rows from CMS's approvals, changed the three cases to expect eligibility through the adult group and added Utah 2020 cases at $9,500 and $25,000. It recorded the Utah January 1, 2020 candidate with four captures, took Florida's expected value from Florida's own variable in the property test and answered the six minor findings.
    • The fix round reported four runs. A mutant without the Utah and Virginia adult rows failed the five adult-route cases and no other. A mutant with Utah's 2020 adult row at 1.00 failed the $25,000 case alone. A mutant without Virginia's 2021 formula failed the second property test at 2021. On main at 61c743cb61, all 35 added cases failed and main's 41 passed.
    • The second round approved 41230025fe with no blocking finding. It fetched CMS's Utah and Virginia approvals, S.B. 96, the Utah Bulletin of February 1, 2020, TN-14-0005 and UT-16-0025, and confirmed each adult date and limit. It parsed the merge and found main's 41 cases verbatim beside the PR's additions, 458 candidate entries with no repeated id and no weakened assertion. It recomputed each changed or new expectation by hand. CI run 43643942759 passed all 38 jobs at the head on main at 61c743cb61, and its selective job ran the 76 YAML cases, both property tests and the 103 guard tests.
    • It listed three minor findings, which stay open under Next. The shared disregard parameter's description claims every year of the three routes. No case separates Virginia's 2019 adult limit of 1.38 from a lower value. The Utah April 1, 2019 adult row cites page 6 of CMS's approval, where the 95% start sits on page 19.
    • The reviews listed corrections to the description, and I made each before merge. The description now gives test counts and the CI jobs that run the property file in place of local timings. It limits the disregard statement to 2018–2020, says that Utah's 0.60 for 2018–2019 sits in the national table, names the adult limits that change by state and year, and uses an accepted axiom: form.
  • Methodology choices for maintainers:
    • Utah's adult limit on January 1, 2020 follows CMS's approval and Utah Code 26-18-3.9(5)(b), which give 133%, so the model stores 1.38. R414-312-6(3) still read 95% on that date, and DAR 52479 could raise it no earlier than March 9, 2020. Candidate us-ut:medicaid-adult-expansion-limit-january-2020 stays open and needs a ruling. Under the rule's reading the row would hold 1.00, and the $25,000 Utah case would expect no adult eligibility.
    • The model keeps the five-point disregard on the Tennessee, Utah and Virginia parent limits in 2018–2020. Under 42 CFR 435.603(d)(4), a state subtracts the five points only for the group with the highest standard, which the adult group holds in Virginia in 2019–2020 and Utah in 2020. Dropping the disregard there would cut those parent limits by 0.05. A parent whom the adult group also covers would move from the parent category to the adult category, and a caretaker outside the adult group would lose eligibility in that band.
    • For Virginia the model follows DMAS's dated manual tables over 12VAC30-40-220, which lists lower amounts and last changed on July 27, 2017. For Utah it follows the approved plan and the amended rule over the 44% in CMS's 2018 summary. Both candidates say ruled, and each rests on the specification in The model denies Medicaid in 2018 to Utah parents between 49% and 60% of the poverty guideline, whom Utah covered, and covers Virginia and Tennessee parents above their states' limits #10793.
    • Annual results read each schedule on January 1. Utah keeps 0.60 and no adult group for all of 2019, though the dollar table took effect on May 7 and adult coverage began on April 1. Virginia's 2018, 2019 and 2020 results use the July 2017, 2018 and 2019 tables, and the July 2020 amounts reach no annual result.
    • The existing income helper rounds each monthly limit up to a whole dollar, so it admits up to $0.99 a month above the state standard plus the disregard. The boundary cases and schedule vectors test that convention.
    • Above the last listed size, the model holds Tennessee's size-20 amount, adds Utah's $62 for each person above 16 and adds each Virginia group's increment for each person above 8. Utah's rule says "each family member", which the model reads as each member above 16.
    • The model treats Utah's Adult Expansion Population, which CMS approved as demonstration expenditure authority, as the adult group. It applies no enrollment limit, employer-coverage requirement or community engagement requirement. Medicaid's verified start year stays 2021, and compatibility entries keep prior values outside 2018–2020 without verifying them.
  • Gates checked live at merge:
    • gh pr checks exits 0;
    • the latest Pull request run at the head, 43676490127, passed and tested this head merged into main at a29c73d587;
    • MERGEABLE;
    • not a draft;
    • no reviewer's latest review requests changes;
    • the semantic-overlap guard: since a29c73d587, main has changed one file this PR touches, source_conflicts/candidates.yaml, where other landings added or edited entries. In the file merged with main at e6824771f8, each of the 475 entries matches the side that changed it, and the 195 tests in the files that read it pass on the merged tree. Between the PR's merge base 61c743cb61 and main at a29c73d587, main changed one file this PR touches, source_conflicts/candidates.yaml. There ten PRs add eight entries and edit nine others, and this PR adds two Utah entries and one Virginia entry, so the merged file holds 467 unique ids. Main changed nothing under gov/hhs/medicaid, gov/states/tn or gov/states/ut, and changed none of the three test files this PR touches. Under gov/states/va, Add Virginia EAP cooling assistance electric bill payment #10760 added ten cooling-assistance files under va/dss/eap and Count gambling winnings in federal gross income and fix the wagering losses deduction #10070 changed four itemized-deduction files, and none sits under va/dmas. Count gambling winnings in federal gross income and fix the wagering losses deduction #10070 adds gambling winnings to federal gross income, which Medicaid's income measure reads; this PR's households have employment income alone, so none of their results move. Date Medicaid adult-group limits from state coverage starts #10826, a draft that adds the same Utah and Virginia adult rows, hasn't landed. The branch merges into main at a29c73d587 and at e6824771f8 without conflict;
    • the merge pins the head.
  • Impact: we estimate no change in 2026 on the default dataset, where none of the seven measured outputs moves. This PR changes Medicaid limits and start dates for 2018–2020. Main 61c743cb61 against 41230025fe.
  • Next:
    • Date Medicaid adult-group limits from state coverage starts #10826 adds the same Utah and Virginia adult rows, byte for byte, beside rows for Maine, Idaho and Nebraska. After this squash it should merge main and keep main's Utah and Virginia rows unchanged, so each state and date has one row. Its axiom: line should name Encode the 2018-2020 Medicaid parent and caretaker relative income standards in Utah, Tennessee and Virginia, and the Utah and Virginia adult group start dates (42 CFR 435.110 and 435.119) (PE-US #10874) TheAxiomFoundation/rulespec-us#2248 for Utah and Virginia.
    • The Utah April 1, 2019 adult row cites page 6 of CMS's approval, and the 95% start sits on page 19. Correct the reference with Date Medicaid adult-group limits from state coverage starts #10826, which carries the same row.
    • fpl_disregard.yaml says the model applies the disregard in every year of the three routes, but Virginia's route runs again from July 1, 2025 without it. The description should name 2018 through 2020.
    • No case separates Virginia's 2019 adult limit of 1.38 from a lower value. A three-person Virginia 2019 case at $29,436 (eligible) and $29,437 (not eligible) would pin it.
    • Tennessee's TN-21-0010, effective October 1, 2021, keeps the same table, yet the route stops on January 1, 2021 and the model returns to 1.00, where the table gives about 0.93 for three in 2021. Utah returns to national values from 2021, and Virginia for 2021–2025. An issue can add each state's dated standards.
    • Four parameters hold the five-point disregard: Missouri's, Florida's, Texas's and the shared one. An issue can replace them with one dated parameter, or a rule that drops the disregard below the highest standard, once the 435.603(d)(4) reading has a ruling.
    • After this squash, the three candidates go into the shared register in rulespec-us, and the Utah January 1, 2020 one still needs its ruling. A follow-up PR then marks the deciding lines and deletes the seven KNOWN_UNMARKED lines this PR adds.
    • Eight of this PR's captures render their tables as CSV, and the register checks excerpts verbatim. The Utah 2018 and Virginia entries may need new verbatim captures beside them.
  • Axiom: Encode the 2018-2020 Medicaid parent and caretaker relative income standards in Utah, Tennessee and Virginia, and the Utah and Virginia adult group start dates (42 CFR 435.110 and 435.119) (PE-US #10874) TheAxiomFoundation/rulespec-us#2248 queued, a new pe-parity issue for 42 CFR 435.110 and 435.119 in these states, with companion rows from this PR's YAML. It covers Tennessee's TN-14-0005 table, Utah's UT-16-0025 and R414-303-4, Virginia's DMAS Appendix 3 tables, and the Utah and Virginia adult group start dates. The signed encoder can't run yet. The corpus holds later editions (TN-21-0010 and the current Utah and Virginia manuals) and none of the 2018–2020 texts. It lacks the two CMS approvals and the 2018–2020 poverty guidelines, and the release that rulespec-us pins carries no Medicaid source for these states. Encode the 2024-2025 Medicaid parent and caretaker relative income standards in TX, FL, MS and WY, and Connecticut's 133% FPL from October 1, 2024 (42 CFR 435.110; CGS 17b-261(a)) (PE-US PR 10252) TheAxiomFoundation/rulespec-us#1951, which the earlier description named, covers other states and needs no comment.

🤖 Generated with Claude Code

@MaxGhenis
MaxGhenis deleted the medicaid-parent-limits-ut-tn-va-2018-2020 branch November 15, 2026 14:42
MaxGhenis added a commit that referenced this pull request Nov 15, 2026
Resolve one conflict in source_conflicts/candidates.yaml, where main added
candidates from #10874 after the last entry this branch adds. Keep every
candidate from both sides.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
MaxGhenis added a commit that referenced this pull request Nov 15, 2026
On main, #10874 dated Utah's and Virginia's adult-group limits and cited
CMS approvals. This branch dated the same limits and cited state sources.
The values agree: Utah 1.00 from April 1, 2019 and 1.38 from January 1,
2020, and Virginia 1.38 from January 1, 2019. Each date keeps one value
and the references of both sides.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
MaxGhenis added a commit that referenced this pull request Nov 15, 2026
#10874 reported Virginia's and Utah's starts in the 2.135.4 changelog.
Against main, this branch still changes three starts: Maine on July 2,
2018, Idaho on January 1, 2020 and Nebraska on October 1, 2020.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
MaxGhenis added a commit that referenced this pull request Nov 15, 2026
Date Medicaid adult-group limits in Maine, Idaho and Nebraska from their coverage starts.

Under 42 CFR 435.119(b), the agency must provide Medicaid to adults aged 19 to 64 who meet the section's other conditions and have household income at or below 133% of the poverty guideline. CMS approved each state's election of the group: Maine's SPA ME-18-0006 effective July 2, 2018, Idaho's ID-19-0007 effective January 1, 2020, and Nebraska's NE-19-0002 effective October 1, 2020. Under 435.603(d)(4), the state must also subtract five percentage points of the guideline from income for the group with the highest standard. The model held no adult group in the three states until January 1, 2021.

Start each state's limit of 1.38, the 133% standard plus the disregard, on its effective date, and cite a state source beside each CMS approval. Add state sources for Virginia's and Utah's starts, which #10874 dated. The model reads the limit on January 1, so the adult group now opens to Maine adults in 2019 and 2020 and to Idaho adults in 2020. Nebraska's start changes no annual result.

Add 31 cases to `medicaid_category.yaml`: 30 for Kentucky, Virginia, Maine, Idaho, Utah and Nebraska in 2019, 2020, 2021 and 2026, and one for Texas in 2026. Six 2026 cases at $20,000 check that Medicaid eligibility bars PTCs under 26 USC 36B(c)(2)(B), and a seventh, in Texas, expects PTC eligibility. Add Python tests that read each of the six start dates and the day before, and two Hypothesis properties: the limits never fall from 2018 to 2026, and they stay at 1.38 from 2021 to 2035.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

1 participant