How we score

We publish this page because public money needs a clear, auditable benchmark — the same transparency norm used in open-budget portals worldwide. It explains outcomes citizens see on Performance and the explorer; it does not document proprietary weights, signal-lake wiring, or enterprise adoption tools. FedhAI shows official budget lines next to an independent reference built from public fiscal record. The reference does not copy the office's published book. Scores are rerunnable rules — not a corruption finding, audit opinion, or Controller of Budget verdict. See Terms.

Efficiency (official vs reference)

We compare the sector mix on the enacted total: how much goes to health, education, infrastructure, and other service lines versus what the independent reference suggests for that level (county, national, ward). The headline percentage and Kenyan letter grade summarize that gap. Offices with an assembly- or Treasury-sourced programme book get a league rank on Performance.

Letter bands follow a Kenyan-style scale (for example, A from 80% upward on the efficiency percentage). An A means strong overall mix alignment with the reference on the whole envelope — not an endorsement of every sector choice, and not a finding of waste or corruption. Read the percentage and sector table when one line matters more than the letter.

FY 2024/25 county explorer

Published FY 2024/25 county rows are historical enacted-share shells with template sector mix for comparison — they are not programme league scores. The public explorer labels them Historical · not league-scored (no A–E grade). Use FY 2026/2027 for the current programme league after assembly ingest.

FY 2024/25 national (DoRA division)

The FY 2024/25 national row is the Division of Revenue Act revenue split (national share, county equitable share, Equalisation Fund) — not the Treasury ministerial programme book. It is not league-scored on Performance. Cash received on that row compares county exchequer disbursements to the enacted county share envelope, not the full KES 2.95T division total. Use FY 2026/27 national for ministerial programme vs reference league scoring.

Interim counties

Where we only have the county's enacted equitable share but not yet the full county assembly programme breakdown, we publish a transparent placeholder labelled (interim). Those rows show totals and cash received but are not league-ranked, because placeholder sector lines would falsely match the reference. When the real assembly estimates are loaded, the interim label drops and the county joins the league automatically after the next reference refresh.

Cash received

Cash received is cumulative Treasury exchequer from Gazette statements compared to the enacted envelope for that office (county equitable share, or for national ministerial books the aggregate transfers frame we document on the Kenya page). Figures can be partial-year until the fiscal year closes.

Where official programme data comes from

National: Treasury programme / Mwananchi and related published books. Counties: county assembly itemized revenue and expenditure estimates (PDFs on assembly or county sites, and public repositories). FedhAI ingests from those public sources — citizens do not need to upload PDFs for the public explorer to work. Counties or partners can supply verified files to speed ingest; see Pricing for office workflows.

Each county book is mapped to FedhAI sectors from the published document, and reference scores refresh on schedule. Scores and league rows update whenever the official lines change, without a separate "scoring release." Each county page links to the publisher's own copy of the approved budget, so you can check any figure against the source.

Projects and citizen checks

Project status, spending and reasons are reported by the owning office and shown as reported. FedhAI does not guess a status, and says so when an office has not updated a project for 90 days. Citizens can ask a question on a project and record whether the status matches what they see; the page shows totals only, and a project is marked contested when at least three citizens dispute it and they are at least 30% of those who checked. None of this changes a score: a contested label is a prompt for the office to answer, not a finding.

On each budget page

Open any book in the Public budget explorer for sector-by-sector official vs reference tables. We do not publish the full reference wiring or source-level weights on those pages.