Pooled ASEC 2022/2023 vintages lack the NOW_* coverage recodes: reported Medicaid at interview thins to 24.6M vs ~58M survey under 65 on the certified artifact

[MaxGhenis]

Summary

On the certified default artifact (hf://policyengine/populace-us/populace_us_2024.h5@populace-us-2024-buildp-sparse-rmloss100-cae8640-20260728T011454Z), the reported-coverage person input has_medicaid_health_coverage_at_interview sums to 24.6M weighted under-65 persons versus ~57.5M survey-reported (CPS ASEC 2023 NOW_MCAID, MARSUPWT). Root cause: the pooled support spine stacks three ASEC vintages keyed by income year (source_year 2022/2023/2024 = the ASEC 2023/2024/2025 survey files), and the income-year-2022 and -2023 raw H5 inputs carry only 2 of the 18 NOW_* at-interview coverage recodes (NOW_GRP and NOW_MRK). The CPS-carried derivation maps missing/NaN source values to False, so roughly two-thirds of the weighted population — every 2022- and 2023-vintage person, 173.2M weighted under 65 — is structurally unable to report Medicaid (or TRICARE, VA, CHAMPVA, IHS, other-means-tested, or non-marketplace direct-purchase) coverage.

has_esi and has_marketplace_health_coverage_at_interview escape only because NOW_GRP and NOW_MRK happen to be the two NOW_* columns the archived raw-lane converter carried for all vintages.

Downstream consequence (policyengine-scorecard PR #69, annotation kff-medicaid-reported-denominator-sanity): the "no reported coverage" construct identifies 80.9M under-65 persons versus 25.9M survey uninsured-at-interview (NOW_COV==2, ASEC 2023) and ~27M ACS nonelderly uninsured, so uninsured-population constructs on the certified artifact are not survey-faithful. Modeled enrollment is separately CMS-calibrated and does not backfill the reported flag.

No artifacts were published or promoted as part of this investigation.

Reproduction

Weighted person counts on the certified artifact (person-table columns × household weight; the columns are physically present in the h5 and the engine surface reads them as-is), against full-file CPS ASEC 2023 (MARSUPWT/100):

input (under 65, millions)	artifact	ASEC 2023 survey	survey column
has_esi	146.9	164.7	NOW_GRP
has_marketplace_health_coverage_at_interview	21.4	11.4	NOW_MRK
has_non_marketplace_direct_purchase_…_at_interview	4.0	8.8	NOW_NONM
has_medicaid_health_coverage_at_interview	24.6	57.5	NOW_MCAID
has_champva_health_coverage_at_interview	0.3	0.4	NOW_CHAMPVA
has_tricare_health_coverage_at_interview	3.5	6.0	NOW_MIL
has_va_health_coverage_at_interview	0.7	1.6	NOW_VACARE
has_other_means_tested_…_at_interview	0.5	0.7	NOW_OTHMT
has_indian_health_service_coverage_at_interview	0.3	0.8	NOW_IHSFLG
none of the nine ("no reported coverage")	80.9	25.9	NOW_COV==2

The marketplace flag tracks NOW_MRK one-to-one on all vintages, so its level above the ASEC 2023 survey reflects vintage composition and calibration-era weighting, not the missing-column mechanism (out of scope here).

Artifact totals: 340.1M persons, 278.9M under 65 (57,240 households, 166,321 person rows; pe-us pin 1.764.6 per the build manifest).

Compact repro (certified snapshot cached locally)

import os, h5py, numpy as np
p = "~/.cache/huggingface/hub/datasets--policyengine--populace-us/snapshots/26dcad66867687f15735dc4926523e3741920836/populace_us_2024.h5"
f = h5py.File(os.path.expanduser(p), "r")
per, hh = f["person/table"][:], f["household/table"][:]
hw = dict(zip(hh["household_id"], hh["household_weight"]))
w = np.array([hw[h] for h in per["person_household_id"]])
u65 = per["A_AGE"] < 65
flag = per["has_medicaid_health_coverage_at_interview"].astype(bool)
print(w[flag & u65].sum() / 1e6)          # 24.6
for yr in (2022, 2023, 2024):             # vintage split
    m = per["source_year"] == yr
    print(yr, w[m & flag & u65].sum() / 1e6, np.isnan(per["NOW_MCAID"][m]).mean())
# 2022  0.0  1.00 / 2023  0.0  1.00 / 2024  24.6  0.00

Root cause

1. The pooled spine stacks three ASEC vintages keyed by income year. source_year on the certified person table: 2022 = ASEC 2023 survey (54,464 rows, 106.8M weighted), 2023 = ASEC 2024 (54,654 rows, 105.7M), 2024 = ASEC 2025 (57,203 rows, 127.6M). Pooling unions per-year columns via pd.concat (asec_pool.py#L141), leaving NaN where a vintage lacks a column.

2. The 2022/2023 raw H5 inputs carry only 2 of the 18 NOW_* fields. On the certified frame, NOW_MCAID, NOW_CAID, NOW_COV, NOW_MRKS, NOW_MRKUN, NOW_PCHIP, NOW_CHAMPVA, NOW_VACARE, NOW_MIL, NOW_IHSFLG, NOW_NONM, NOW_OTHMT, NOW_PUB, NOW_PRIV, NOW_DIR are populated on exactly the 57,203 2024-vintage rows and NaN on all 109,118 rows from 2022/2023. Only NOW_GRP and NOW_MRK are populated for all vintages. This is a property of the SHA-locked inputs themselves: policyengine_us_data/storage/census_cps_2022.h5 (146,133 persons = pppub23, 100% PERIDNUM match; 137 person columns; NOW_* = {NOW_GRP, NOW_MRK}), census_cps_2023.h5 (144,265 persons = ASEC 2024; 138 columns; the same two), census_cps_2024.h5 (142,125 persons = ASEC 2025; 162 columns; all 18 NOW_* fields). The 2022/2023 inputs match the archived policyengine-us-data extractor's column list before 410f63a7 ("Add CPS current health coverage anchors", 2026-03-20, which added the broader block); the 2024 input was extracted with it. The input-coverage manifest already documents this vintage asymmetry for other fields ("The 2022 and 2023 inputs also omit PECOHAB, A_EXPRRP, and A_FAMREL; only 2024 carries those alternatives" — is_unmarried_partner_of_household_head exclusion; same pattern in the employer_sponsored_insurance_premiums exclusion, Track SPM-specific input gaps: housing, WIC, school meals, child support, workers comp, and expense deductions #32).

3. The CPS-carried derivation silently maps missing to False. _fill_health_coverage_inputs maps has_medicaid_health_coverage_at_interview ← NOW_MCAID (L272) through _yes_code, where _source returns zeros for an absent column and fillna(0.0) for NaN. The certified flag equals NOW_MCAID == 1 exactly (weighted crosstab: 24.6M both-true, 0.0M in either disagreement cell) — no later stage rewrites it, and the export carries it (this is not the Sparse-57k export drops 94 no-formula input columns from the dense parent (tips, overtime, disability flags, ESI premiums, rent, QBI basis) #361 export-drop class; the column is present).

4. The 2024-vintage slice is survey-plausible; selection and reweighting push the other way. Within the 2024 vintage the flag carries 24.6M of 105.7M under-65 = 23.3%, versus 21.1% in full ASEC 2023 (57.5M of 272.7M). Joining the 2022-vintage rows back to pppub23.csv by PERIDNUM (100% match, zero age mismatches), the under-65 Medicaid-at-interview rate moves from 21.1% in the full survey to 23.8% among the selected source persons, 24.5% with support clones, and 25.8% under final artifact weights — selection and calibration mildly enrich reporters, the opposite direction from the thinning. Counterfactual scale: restoring the raw codes on the artifact's existing rows and weights gives ≈ 70M under 65 (22.7M on the 2022 vintage, verified directly against pppub23; the 2023 slice extrapolated at the same enrichment; 24.6M already on 2024); at survey rates ≈ 0.211 × 278.9M ≈ 59M. A rebuild reselects and reweights, so neither is a guaranteed post-fix value — the point is the scale, ~2.5–3× today's 24.6M.

5. The Medicaid take-up stage consumes the already-thinned anchor. medicaid_take_up.py anchors on the flag (eligible reporters always take up) and hard-fails only if the column is absent (L173-L177) — there is no guard on per-vintage coverage. The CMS state-count fill compensates in enrollment terms, so enrollment-side calibration passes while the reported-coverage denominator stays thinned; the thinned anchor also shifts enrollment composition from reported toward filled and overstates saturation pressure (Add Medicaid and CHIP eligibility-to-enrollment diagnostics #170's diagnostic direction).

Why certification passed

• release_input_coverage requires the column to be present with non-default signal ("Column mass parity is not column coverage" — the gate closes the absence class, Build J: full eCPS-exported-column coverage as a HARD release gate + asset-column restoration (SSI reforms score $0) #368); 24.6M true satisfies it.
• degenerate_input_signal (175 columns) passes any column that is neither all-default nor constant.
• health_input_signal checks 2 columns (selected_marketplace_plan_benchmark_ratio, takes_up_aca_if_eligible) — no reported-coverage flags.
• ecps_parity is a populated-layer presence check (158 layers, 0 gaps).

Nothing measures per-vintage signal or survey-mass fidelity for the reported-coverage family, so a flag populated for one of three vintages certifies.

Relationship to existing issues

• Pool default-value holes: engine-consumed input leaves with checkpoint sources the pool never populates (is_tanf_enrolled, strike_benefits, receives_medicaid/snap/wic) #591 (pool default-value holes): same family — engine inputs the pool under-populates. Its table notes the sibling receives_medicaid is never emitted while has_medicaid_health_coverage_at_interview "is emitted"; this issue sharpens that: it is emitted for the 2024 vintage only. Partial-vintage holes are invisible to presence-style checks.
• Sparse-57k export drops 94 no-formula input columns from the dense parent (tips, overtime, disability flags, ESI premiums, rent, QBI basis) #361 / Sparse-57k certified release zeroes untargeted input bases (IRA/HSA/SE-pension/childcare); CDCC and ALD reforms silently score ~$0 on bundle 4.18.8 #278 / Sparse-57k certified release drops SSI asset input columns; SSI resource-test reforms score $0 on bundle 4.18.8 #356 (export-side column drops): distinct mechanism — here the column survives export; the values were born thinned at source pooling.
• Ingest ASEC 2024 and 2025 into the US support pool (spine is two vintages behind) #296 (ingest ASEC 2024/2025 into the support pool): the natural vehicle for the data-side fix — any re-extraction should carry the full NOW_* coverage block for every vintage.
• Add Medicaid and CHIP eligibility-to-enrollment diagnostics #170 / Materialize Medicaid and CHIP take-up via anchored count-calibration (ACA stage pattern) #331 / Health take-up wiring + certified-stack computability gate (real residuals from #604) #605 (Medicaid take-up and enrollment diagnostics): enrollment-side consumers of the thinned anchor.
• Add coverage-universe targets for uninsured and private coverage #230 (coverage-universe targets for uninsured and private coverage): targeting the uninsured universe would have surfaced this earlier and remains the calibration-side complement.

Proposed fix

1. Data side (the actual fix): re-extract the income-year 2022 and 2023 inputs (ASEC 2023/2024) with the full Census person column set — not a NOW_*-only patch, so the next missing field (Track SPM-specific input gaps: housing, WIC, school meals, child support, workers comp, and expense deductions #32's NOW_OWNGRP/NOW_HIPAID/NOW_GRPFTYP, Populate remaining US tax-input and reported-observation layers #38's PERRP/A_EXPRRP/PECOHAB) does not repeat this cycle, while any new derivations stay in separate PRs — and rebuild the pooled source base. Do it as one schema migration with Ingest ASEC 2024 and 2025 into the US support pool (spine is two vintages behind) #296 (whose body predates the ASEC 2025 ingest: Build P already consumes income years 2022–2024) after reconciling survey-year vs income-year naming. The raw CSVs carry the data — verified for ASEC 2023 (pppub23.csv has NOW_MCAID, column 547). This runs in the gated data lane outside PR CI; no artifact publication is part of this issue.
2. Regression guard (PR Add per-vintage reported-coverage signal gate (#720) #744): a per-vintage reported-coverage signal gate — for each pooled source_year, every reported-coverage input must carry non-default signal (the crisp invariant violated today: 2022/2023 vintages have exactly zero true mass for seven of the nine flags). Ships red against current bases by design, like the Build J: full eCPS-exported-column coverage as a HARD release gate + asset-column restoration (SSI reforms score $0) #368 asset gates, and turns green with the re-extraction.
3. Scorecard follow-up: once a fixed artifact certifies, the kff-medicaid-reported-denominator-sanity annotation and the reported-uninsured rows in policyengine-scorecard PR Target latest eligible Ledger fiscal facts #69 can be revisited.

Numbers computed 2026-08-18 on the locally cached certified snapshot (26dcad668676…, sha256 48b9d479fb4f…) and asecpub23csv with policyengine-us 1.764.6 (the build pin).

[MaxGhenis]

Regression-guard PR (fix layer 2, the per-vintage reported-coverage signal gate): #744. The data-side re-extraction of the 2022/2023 ASEC inputs (fix layer 1) remains open here.

[MaxGhenis]

Corrections after a cross-family adversarial recompute (body edited accordingly):

1. source_year is the income year: 2022/2023/2024 = ASEC 2023/2024/2025 survey files. The SHA-locked inputs are policyengine_us_data/storage/census_cps_{2022,2023,2024}.h5 with 137/138/162 person columns and 2/2/18 NOW_* fields (census_cps_2022.h5 = pppub23, 100% PERIDNUM match).
2. The earlier "selection is coverage-neutral" paragraph was computed on a cross-year join (2023-vintage rows ↔ pppub23, a 32% CPS rotation-group overlap) and is retracted. On the correctly matched 2022 vintage, the under-65 Medicaid rate goes 21.1% (full survey) → 23.8% (selected persons) → 24.5% (with clones) → 25.8% (final weights): selection and calibration enrich reporters. The conclusion is unchanged and stronger — the thinning is entirely the missing columns; the post-fix counterfactual is ~59M at survey rates or ~70M on current rows/weights.
3. NOW_MCAID does appear in the archived extractor's history (410f63a7, 2026-03-20, on non-main history); the 2022/2023 inputs match the pre-410f63a7 column list. The "never appears" sentence was wrong and is removed.
4. Wording: the 2022/2023 inputs carry 2 of the 18 NOW_* fields, not "none of the block".

The root cause, numbers in the reproduction table, and the NaN→False mechanism are unchanged. Review findings on PR #744 (source-family × vintage key, export-frame rerun, provenance must be present, failure receipt) are being folded into the PR.

[MaxGhenis]

Data-side fix built and proven at the source stage (no pool/release build run yet).

Corrected inputs (_buildo-runtime/inputs/asec-720/census_cps_2022.h5, …2023.h5): every existing table/column byte-identical to the SHA-locked files (asserted), with the columns the 2024 input carries appended from the raw Census CSVs by PERIDNUM (100% match): the 16 NOW_* at-interview recodes plus PECOHAB/A_EXPRRP/A_FAMREL (the #38 prerequisites) and A_ENRLW/A_FTPT/A_FAMTYP/LKWEEKS/PTOTVAL; CENSUS_TAX_ID (extractor-derived) stays absent. Receipt with SHA-256s: inputs/asec-720/receipt_720.json; script: scripts/augment_asec_inputs_720.py.

build_us_puf_support_base.py --stage source_construction (seed 0, same flags as inc2_asec_producer.sh) before/after, gate = PR #744's us_reported_coverage_vintage_signal_gate on derive_us_cps_carried_inputs of the raw-stage frame (432,523 persons):

inputs	NOW_MCAID null share (2022 / 2023 / 2024)	Medicaid reporters	gate
SHA-locked	1.00 / 1.00 / 0.00	0 / 0 / 24,844	FAIL — 14 failures (7 flags × vintages 2022, 2023)
corrected	0.00 / 0.00 / 0.00	28,184 / 26,367 / 24,844	PASS

The corrected raw-stage checkpoint (out/asec-720-source/checkpoints/asec_raw_stage.checkpoint.h5, sha256 67e459381f173ae89e05012f26c3004e4d5668218b97568a18f3d37463362911) is a drop-in for the 25% candidate chain via tools/build_us_multispine_pool.py --asec-raw-stage-h5 … --asec-raw-stage-h5-sha256 …; hand-off note at out/asec-720-source/HANDOFF.md. The calibrated pool/release build waits for the host queue's explicit go (STOPPED-BY-MAX-2026-08-21 rule). Follow-up under #296: one extractor for every retained vintage with the full Census person set; this augmentation is the minimal verified bridge.