MicroZAMOD findings ledger — ZM_2025 full-surface validation (8 entries)

[MaxGhenis]

This is the outward-facing ledger of divergences we observed between MicroZAMOD (SOUTHMOD Zambia, bundle A4.0, systems ZM_2024/ZM_2025, dataset zm_2022_a2) and the governing Zambian instruments, found while building statutorily-grounded encodings of the ZM_2025 policy surface and validating them case-by-case against MicroZAMOD as the household-level oracle. Every encoded amount comes verbatim from captured official prints; every divergence below was observed live and is dispositioned in the oracle suite that covers it.

Model findings (8):

#	Instrument	Finding (probed live)
1	PAYE base	The model taxes a base net of employee NAPSA + NHIMA (ttb_s = yem − tsceepi_s − tsceehl_s); the Income Tax (Amendment) Act, 2024 s.4 makes the approved-fund deduction employer-only from 1 Jan 2025 and no NHIMA deductibility provision exists — PAYE understated 5,328/yr at 240,000 gross, 9,101/yr at 409,968
2	Turnover 0% band	Ninth Schedule Part II prescribes 0% at or below K12,000; the model charges 5% from the first kwacha (K500 at ytn 10,000; K600 at exactly 12,000)
3	Turnover ceiling bound	s.64A(2) reads "five million kwacha or less"; the model's strict < pays 0 at exactly 5,000,000 where the statute prescribes 250,000
4	Cigarette excise	Model K0.400/piece = the CY2024 K400/mille; the Customs and Excise (Amendment) Act, 2024 sets K452/mille from 1 Jan 2025 (probed 480.00 vs statutory 542.40 on 1,200 sticks)
5	Fuel excise	Model applies K2.07/l (the CY2021–24 combined petrol figure) to all fuel litres; CY2025 rows are petrol K2.34/l and diesel K0.75/l — petrol understated ~12%, diesel overstated ~2.8×
6	Clear beer	Model's flat 20% coincides only with the sorghum suspended rate; the statutory rate is 60% layered with the S.I. 66/71 of 2023 suspension (malt→20%, cassava→5%, SME/excess sorghum excluded, self-revoking 31 Dec 2026)
7	Opaque beer base	The statutory duty is specific (K0.25/litre); the model taxes opaque-beer expenditure at the 60% ad-valorem group rate — likely a survey-data proxy (expenditure, not litres)
8	SCT amounts	ZM_2025 pays 400/600 per month — annualizing the drought measures that end June 2025 (or anticipating the 2026 budget); the standing scheme is 200/400 (MCDSS Ministerial Statement; ZM_2024 matches it exactly), and severe disability is double-the-base, not base+200

Scope notes (not findings): the Minimum Alternative Tax (Act 10 of 2025) postdates the 30 June snapshot and is correctly absent; the CY2025 NAPSA ceiling (34,164 = 4 × NAE 8,541) matches the Cap. 256 s.19(3) formula but has no dedicated SI; the SWL/KGS/HGSM/e-FISP/FSP amounts are, per the MicroZAMOD country report's own citations, calibration assumptions with no official kwacha instrument.

Canonical artifacts

• Program tracker with the closing ledger: MicroZAMOD full-surface parity program #1
• Machine-readable entries with probed values: axiom_oracles/data/microzamod_issues.json (on the oracle review chain)
• Per-suite dispositions: the comparisons/zm-*.yaml configs in axiom-oracles

Scope and discipline. The SOUTHMOD bundle is licensed and non-redistributable: only comparison statistics and MicroZAMOD-produced values are recorded. Findings are stated against the captured official prints (National Assembly of Zambia / Government Printer), cited in the encodings. Corrections and context from the MicroZAMOD/SOUTHMOD team are very welcome in this thread.