UGAMOD validation result — UG_2025 full surface (zero findings)

[MaxGhenis]

This is the outward-facing result of validating UGAMOD (SOUTHMOD Uganda, bundle A4.0, system UG_2025, dataset ug_2024_a1) against statutorily-grounded encodings of the full UG_2025 policy surface — PAYE (including the additional 10% above 120m/yr), rental income tax, presumptive small-business tax (both regimes), NSSF employee and employer contributions, Local Service Tax, the Senior Citizens Grant, VAT, fuel excise, and the composed disposable-income identity.

Findings: zero. Every UGAMOD output we probed matches the governing instrument to within monthly sub-shilling rounding (worst divergence 0.4 UGX per year across ~50 live cases spanning every band boundary, both presumptive regimes, all ten LST bands, the SCG age gates, and the full-grid disposable-income identity). UGAMOD implements the complete PAYE schedule including the additional rate — in contrast to some sister models — and its statutory bases matched the captured prints everywhere we tested.

Canonical artifacts

• Program tracker with the closing ledger: UGAMOD full-surface parity program #1
• Per-suite comparisons: the comparisons/ug-*.yaml configs in axiom-oracles (live-exact runs recorded per suite)

Open items (documentation, not findings): the FY2025/26 assented prints (ITA/VAT/Excise/Appropriation 2025) were not yet capturable from public sources at validation time, and the Cap 338-vs-Cap 340 renumbering question is dodged via act-number citation paths.

Scope and discipline. The SOUTHMOD bundle is licensed and non-redistributable: only comparison statistics and UGAMOD-produced values are recorded. Corrections and context from the UGAMOD/SOUTHMOD team are very welcome in this thread.