Non-authoritative. Every figure here derives from mirrored public records; verify against the authoritative sources linked below before relying on any of it.

oregon-stories · from the oregon-budget corpus

An Oregon agency's 20 largest payees absorb a median 88% of its recorded spending — and that is a floor

For each of 80 Oregon agencies in FY2025, the corpus asked one question of the recorded expenditure mirror: what share of the agency's total recorded spending went to its 20 largest payee strings? The median answer is 88% — and every one of these figures is a floor, because payee strings are not de-duplicated upstream. The spread runs from 42% (HUMAN SERVICES, DEPT OF, 7,906 payees) to the 8 small agencies whose 20-or-fewer payees make 100% an arithmetic certainty rather than a finding.

Top-20 payee share of recorded spending — one dot per agency, FY2025

0%20%40%60%80%100%median 88%lowest: Human Services (42%)Dept of Education (67%)Health Authority (77%)
agency with more than 20 payees20 or fewer payees (8 agencies) — the top-20 share is necessarily 100%, so the dot is greyed and the agency is excluded from the table below

Every share on this chart is a floor. Payee strings are not de-duplicated upstream — the same vendor can appear under several spellings — so real concentration is higher, never lower.

Concentration vs scale — share against total recorded expense (log scale), FY2025

0%25%50%75%100%$10K$100K$1M$10M$100M$1B$10B$100B

Most concentrated at scale — the eight highest top-20 shares among agencies with at least 100 payees, FY2025

agency (as recorded)top-20 payee sharepayeesFY2025 recorded expense
PUBLIC EMPS RETIREMENT SYS99.5%174$916,080,198
LEGISLATIVE ADMIN CMTE98.5%227$178,913,223
GOVERNOR, OFFICE OF THE93.0%137$3,600,852
JUDICIAL DEPT92.8%1,492$275,265,718
HI-ED COORD CMSN90.5%602$1,940,344,652
HEALTH RELATED LICENSING BRDs90.1%116$1,541,115
FOREST RESOURCES INST, OR89.0%134$2,458,561
LABOR & IND, BUREAU OF88.9%172$10,276,447

Read this before quoting the chart

The export's own note, carried verbatim: GENERATED by src/build_story_exports.py — do not hand-edit. Per agency x fiscal year over the committed Parquet mirror (the same bytes the expenditure documents hash). VENDOR STRINGS ARE NOT DE-DUPLICATED upstream, so every top20_share here is a FLOOR — real concentration is higher, never lower. agency_registry_slug is null where no join doc carries one; null means unmapped, not unaffiliated.

How to read the numbers. 'Payees' are distinct payee strings in the state's expenditure records, not de-duplicated vendors — one vendor under three spellings counts three times, which can only push a top-20 share down; that is why every share is a floor. An agency with 20 or fewer payee strings shows 100% by construction; the most-concentrated table therefore admits only agencies with at least 100 payees, a threshold chosen here and stated here. Recorded expense is the agency's total spending from every source and implies nothing about appropriations. Coverage: 544 agency-year rows, FY2019–FY2025; the charts show FY2025 only. Concentration is a shape, not a verdict — a pass-through or grant-making agency is concentrated by design.