How we calculate the governance index for state agencies
The governance index compares state agencies with agencies of the same type on what can be measured in open data. It is shown as a quartile within the group, never as a single ranking.
Sources
- The Central Coordinating Register: every unit in the state administration (sector 6100) and every state enterprise. The type is derived from the organisation form, industry code and parent unit.
- DFØ, state accounts: monthly accounting files with organisation numbers. The budget is per chapter and post, so the deviation is computed for chapters where the agency posts at least 95 % of the operating amount.
- The Office of the Auditor General, through the Storting's recommendation on Document 1: which accounts may contain material errors and omissions. We read the sentence, check that the count matches, and match the names to agencies.
- Doffin: notices and awards where the agency is the buyer.
The parts of the index
| Part | Measure | Weight |
|---|---|---|
| Financial control | Deviation between accounts and budget on own operating chapters, both over- and underspending. A smaller deviation gives a higher place in the group. | 30 % |
| Audit | 100 when the agency is not named among the accounts with material errors. 50 is deducted for the latest year and 25 for each of the two years before. | 30 % |
| Employer | Sick leave and full-time equivalents per agency. Not included: the figures are not open per agency. | 15 % |
| Procurement | Share of awards with more than one tender (at least five awards with a known number of tenders). | 15 % |
| The Storting | Cases and questions in the Storting are context, not points: much attention is not bad, and little is not good. | 0 % |
Agencies are compared within their type: ministries, directorates and agencies, universities and colleges, justice, county governors and the Storting's bodies. A group with fewer than 5 agencies is compared with the whole state.
The index is computed only for years with both whole-year accounts and the annual audit. The audit of a year comes the winter after, so the latest accounting year has no index until then. Development is the latest year against three years before.
What we do not measure
We do not assess policy, goal attainment, the quality of services or individual cases. We look only at what can be measured in open data and what the agency itself controls: financial control, audit and procurement. Sick leave per agency is not openly available and is therefore left out.
Limits
State agencies have different missions and no profit, so the figures say little about how well they fulfil their mission. Budget deviations can have good reasons, such as postponed purchases. Mergers and new names can break time series. The audit is matched by name: a name we do not recognise is added as an alias by admin.
Audit years we have read
- 2024: 10 of 230 accounts may contain material errors and omissions. The Storting's recommendation
- 2023: 7 of 235 accounts may contain material errors and omissions. The Storting's recommendation
- 2022: 5 of 236 accounts may contain material errors and omissions. The Storting's recommendation
- 2021: 4 of 235 accounts may contain material errors and omissions. The Storting's recommendation
Corrections and replies
Agencies can report errors or comment on the figures. The comment is shown on the agency's page with date and sender.
Sources: Brønnøysund Register Centre (NLOD), the Norwegian Agency for Public and Financial Management (DFØ), state accounts (NLOD), the Storting (NLOD) and Doffin.