Skip to content

Decide Maryland Premium Assistance's 400% ceiling on the exact FPL ratio - #10320

Merged
DTrim99 merged 3 commits into
mainfrom
md-premium-assistance-exact-fpl-ratio
Nov 19, 2026
Merged

DTrim99 merged 3 commits into
mainfrom
md-premium-assistance-exact-fpl-ratio

Conversation

@MaxGhenis

@MaxGhenis MaxGhenis commented Oct 20, 2026 •

Copy link
Copy Markdown
Contributor

Still pending: full impact run, Axiom parity line, GPT-6.1 Sol review. Done: the fix and YAML cases, on current main.

Update onto current main (2026-11-18)

  • Merged upstream/main with no conflicts. Max's commit stays; nothing was rebased or force-pushed.
  • The ceiling now divides by aca_fpg, the poverty line aca_magi_fraction and the federal credit have used since Keep a filer who can be claimed as a dependent out of the premium tax credit family, and apply each state premium program's rule #10041: the prior-year guideline for the section 36B tax family. The division runs in float64.
  • The 12 fixtures that pinned tax_unit_fpg for the prior year now pin aca_fpg for the coverage year. No expected value changed.
  • New eligibility Cases 19 and 20 cover a joint return with one claimable spouse. The ceiling follows the one-person tax family: $62,600 is eligible, and $70,000 is not (447% of $15,650, though 331% of the whole unit's $21,150).
  • The sources settle the guideline year but not the family counted. COMAR 14.35.21.02B(3) defines the federal poverty level as the guidelines "most recently published" as of the first day of open enrollment, and MHBE's income thresholds use the 2025 guidelines for PY2026 and the 2026 guidelines for PY2027. That matches aca_fpg's guideline year. Those tables list "Household Size" without defining it, and COMAR 14.35.21.03A(1) refers to the "tax filer's tax household" without defining it. Insurance Article §31-125, which authorizes this program, and §31-122, which authorizes the Young Adult Subsidy, define neither the poverty level nor family size. Counting the section 36B tax family is a modeling choice: aca_fpg is the line of the federal credit this gate already requires. Cases 19 and 20 pin that choice, not a Maryland rule.
  • Modal run pa-10320-a: all 149 cases pass across the five Maryland YAML files (108), healthcare_benefit_value.yaml (22) and household_health_benefits.yaml (19), and all 14 tests in test_claimable_filer_ptc_invariants.py pass.
  • Mutation checks on Modal: restoring the old denominator (pa-10320-b) fails eligibility Cases 2 and 20. Restoring main's truncated gate (pa-10320-c) fails eligibility Cases 16, 17 and 18.
  • Maryland's 2027 rates are left for a later change. That change must revisit Case 19 of md_premium_assistance.yaml: it expects $18.22 at the 6.22% target, and at the adopted 6.33% the result falls to $0 ($2,500 less $2,525.67 on $39,900).
  • Not rerun: the differential sweep below, and the $62,756.50 row of the table below. The table's other three rows are covered by eligibility Cases 15, 16 and 17.

Bug

md_premium_assistance_eligible compared the truncated aca_magi_fraction (np.floor(100 * magi / fpg) / 100) with <= 4.0. Truncation maps every income from 400.00% to 400.99% of the poverty line to 4.00, so all of them passed. MHBE's adopted PY2026 parameters say "No state subsidy above 400% FPL", and the PY2027 parameters say "Eligibility: 400% FPL or below, eligible for APTC, and ineligible for Medicaid". No source I found prescribes truncating the percentage.

Under current law the gate also requires is_aca_ptc_eligible, whose income test ends at 400% from 2026, so incomes above 400% already failed end to end. The Maryland ceiling decides alone when a reform keeps the federal credit above 400%.

Executed on main (645fcd3; the file is unchanged at 740e1eb), single filer in Maryland, 2026, 2025 one-person guideline $15,650:

Income Exact % of FPL aca_magi_fraction Gate alone on main (is_aca_ptc_eligible supplied) Here
$62,600 400.000% 4.00 yes yes
$62,600.47 400.003% 4.00 yes no
$62,755.56 400.994% 4.00 yes no
$62,756.50 401.000% 4.01 no no

Fix

Decide the ceiling on the exact ratio max(aca_magi, 0) / aca_fpg, divided in float64, the pattern of #10305 as it stands on main. The target-contribution schedule still interpolates on aca_magi_fraction; only the eligibility edge moves.

Invariants

For every input, with the program in effect and an APTC-eligible member:

  1. Band = design. Eligible if and only if r <= 4, r = max(MAGI, 0) / aca_fpg.
  2. Only the sliver changes. Relative to main, eligibility can only be lost, and only for r in (4.00, 4.01). Under current law that set is already federally ineligible, so baseline results should not move; the impact run is the check.

No Hypothesis test is added: the gate is one comparison on a ratio, and the YAML cases hit the limit, 0.003 and 0.994 points above it, plus a reform case. The repo's test-design rules keep such cases in YAML.

Tests

  • md_premium_assistance_eligible.yaml: cases that supplied aca_magi_fraction now supply aca_magi against an aca_fpg of $10,000. New income-derived cases: $62,599 (399.994%) and $62,600 (400%) eligible; $62,600.47 (400.003%) and $62,755.56 (400.994%) ineligible. Case 18 sets gov.aca.ptc_income_eligibility[2].amount: true (federal credit above 400%): at $62,600.47 the household is federally eligible and Maryland-ineligible.
  • md_premium_assistance.yaml: seven cases supplied a fraction that disagreed with their own aca_magi, checked by re-deriving each fraction from its MAGI. With the gate reading MAGI, each would have judged eligibility and the target percentage on different incomes. None crossed 400%, so none changed result here; the inputs now agree, with expected values recomputed by hand:
    • Case 16, "225% FPL", used $35,000 (223.6% of $15,650): now $35,212.50; $1,061.50 → $1,052.77.
    • Case 17, a two-person unit "at 300% FPL", used $46,950 (222% of $21,150): now $63,450 with a self-consistent APTC of $2,680.38; $929.39 → $1,256.31.
    • Case 19, 2027 "250% FPL", used $39,125 (245% of the 2026 guideline $15,960): now $39,900; $66.425 → $18.22.
    • Case 21, "175% FPL", used $27,000 (172.5%): now $27,387.50 with APTC $4,630.625; $1,350 → $1,369.375.
    • Case 9 (2025), Case 10 (2028) and Case 12 (zero MAGI) are outside the program or have no income; their inputs are made consistent and their $0 / $2,500 results are unchanged.
  • Mutation check: main's gate fails exactly the three new above-limit cases (16, 17, 18); the recomputed amount cases pass on main too.
  • Local differential sweep, run before the move to aca_fpg (a script, not committed, per the repo's test-design rules): 256 tax units per year, 60% with income in cents within $2 of an edge and the rest uniform, guidelines $10,000-$80,000 including $20 multiples, comparing the eligibility gate with an exact-rational reading of the source on the float32 values the model receives. 0 mismatches here; on main the same draws give 93 mismatches, all in the truncation sliver.
  • Local run before the move to aca_fpg: the Maryland Premium Assistance YAML folder, 102 passed. Current results are in the update section above.

Impact

Pending: full-file microsimulation via the US hub impact queue, 2026 and 2027.

Coordination

#10041, merged 11/13, moved aca_magi_fraction's denominator to aca_fpg, and this gate's exact ratio now uses the same denominator. Part of the aca_magi_fraction audit (table: #10329) begun with #10305; federal edge in #10312.

axiom: pending

🤖 Generated with Claude Code

MHBE's adopted PY2026 design provides "No state subsidy above 400% FPL"
and the PY2027 design "Eligibility: 400% FPL or below". The gate compared
the truncated aca_magi_fraction with 4.0, so every income from 400.00% to
400.99% passed. The federal PTC gate ends at 400% under current law, so
the error showed only when a reform extends the federal credit; test the
exact ratio of nonnegative ACA MAGI to the prior-year guideline anyway.

Seven amount cases supplied a fraction that disagreed with their own
aca_magi (for example "225% FPL" with $35,000, which is 223.6% of
$15,650). With the gate now reading MAGI, each case would have judged
eligibility and the target percentage on different incomes; their
incomes and expected amounts are recomputed so both agree.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
DTrim99 and others added 2 commits November 18, 2026 14:08
Divide ACA MAGI by aca_fpg, the prior-year guideline for the section 36B
tax family that the federal credit has used since #10041, instead of the
whole tax unit's guideline. Divide in float64 so a ratio just above the
limit can't round down onto it. COMAR 14.35.21.02B(3) and MHBE's income
thresholds support the guideline year; Maryland's rules don't define the
family counted, so the comment calls that part a modeling choice. Pin the
test fixtures to aca_fpg and add a joint return with a claimable spouse,
whose ceiling follows the one-person tax family.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@DTrim99
DTrim99 marked this pull request as ready for review November 18, 2026 19:28
@DTrim99
DTrim99 requested a review from hua7450 November 18, 2026 19:28

@hua7450 hua7450 left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks, David. Maryland's ceiling checks out: the MHBE Board's PY2026 motion ("No state subsidy above 400% FPL") and PY2027 motion ("400% FPL or below") set an inclusive 400% limit, and MHBE's tables put it at exactly four times the guideline ($62,600 for one person in PY2026), which truncation would contradict. COMAR 14.35.21.02B(3) uses the same guideline year as aca_fpg. Approving, with one reference fix before merge.

Before merge

Each MHBE deck is listed twice. In md_premium_assistance_eligible.py lines 20-25, the PY2026 deck appears at #page=30 and #page=16, and the PY2027 deck at #page=25 and #page=12. Please keep one href per deck on the first cited page with a pages comment: # PDF pages 16,30 above the PY2026 href (#page=16) and # PDF pages 12,25 above the PY2027 href (#page=12). The pages themselves are right.

Optional

  • Comments. The comments on Cases 12 and 13 still say the gate reads aca_magi_fraction against the "2026 FPL".
  • Family size. The comment calling the federal tax family "a modeling choice" could cite COMAR 14.35.07.08. Cases 19 and 20 rest on a joint filer who can be claimed as a dependent, which 26 U.S.C. 152(b)(2) rarely allows.
  • Zero guard (see below).
  • 2027 rates (separate PR). The 2027 contribution schedule still uses the 2026 rates.

Series-wide (follow-up, not for this PR). The exact ratio max_(aca_magi, 0) / aca_fpg is now computed inline in each PR of this series, plus two California copies on main, and only 10312 guards a zero poverty line (territories, where aca_fpg is 0; the formula runs before the state mask, so a mixed simulation emits NumPy warnings, though outputs are unaffected). One guarded tax-unit variable for the untruncated ratio, read by every gate, would be a good follow-up PR once these merge.

@DTrim99
DTrim99 merged commit 3b54cc1 into main Nov 19, 2026
82 of 83 checks passed
@DTrim99
DTrim99 deleted the md-premium-assistance-exact-fpl-ratio branch November 19, 2026 03:58
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants