How the Government Accountability Score Is Computed
The 2026 edition
How the score works
The Government Accountability Score rates a country’s law from 0 to 100. It is a weighted average of the scores of five sections of law, each also 0 to 100, so a country that scores 100 in every section scores 100.
The score measures what the law requires, not whether government complies. A high score means the law on paper contains strong accountability provisions. A low score means the law is silent, permissive, or weak in ways that reduce accountability. Compliance, enforcement and outcomes are separate questions.
Three rules decide what goes into a score:
- A law that does not exist scores 0. Where we investigated and found no qualifying law (shown as None), the section counts as 0.
- A section that is not assessed is left out. The other sections’ weights are scaled up to fill its place, so it neither raises nor lowers the score.
- A score needs three quarters of the points. A jurisdiction is scored only when the sections assessed for it carry at least three quarters of the points in the table below, which in practice means open records and any three others.
A country’s page also shows its civic-space rating from the CIVICUS Monitor, beside the score. CIVICUS rates the freedoms of association, peaceful assembly and expression in law and in practice, so its rating answers a question the score leaves aside. It is shown for context and is never part of the score.
The five sections and their weights
| Section | Points | Intended share | Weight applied | Share it carries |
|---|---|---|---|---|
| Open records | 8 | 25.8% | 24.9% | 25.8% |
| Open meetings | 4 | 12.9% | 16.1% | 12.9% |
| Oversight | 7 | 22.6% | 26.0% | 22.6% |
| Fiscal transparency | 7 | 22.6% | 25.3% | 22.6% |
| Accessibility | 5 | 16.1% | 7.8% | 16.1% |
| Total | 31 | 100.0% | 100.0% | 100.0% |
Points say how much each section should count, and come from the rubric below. The intended share is a section’s points out of 31. The share it carries is how much of the differences between countries’ scores comes from that section.
Using the points directly as weights would not give each section its intended share, because a section’s influence on a ranking also depends on how far its scores spread (Paruolo, Saisana and Saltelli, Journal of the Royal Statistical Society A, 2013). So the weights applied are calibrated until each section carries its intended share, the method of Becker, Saisana, Paruolo and Vandecasteele (Ecological Indicators, 2017).
Accessibility’s weight looks low for that reason. Its scores are close to all or nothing (most countries have no qualifying law, and those that do score well), so a small weight already gives it its full intended share of the ranking.
The weights were calibrated on November 4, 2026 over the 95 countries with all five sections assessed. They are fixed for the 2026 edition and recalibrated once a year, so a rank moves between editions only when some country’s law changes.
How the points are set
Each section is marked 1 to 3 on three criteria, and its points are the sum of the three marks. The marks rate each section as a measure of government accountability, not the importance of the right behind it.
| Criterion | 3 | 2 | 1 |
|---|---|---|---|
| Recognition in international law | Binding treaty text on point | UN General Assembly resolutions, or a global standard-setter’s standard | No global instrument |
| Evidence of effect on accountability, in peer-reviewed studies | Several studies, consistent | Some studies, or contested | Little or none |
| Reach: how much of government the duty opens to the public | Every public body, all of its information or work | A central but bounded function | One channel or a subset |
Open records: 8 points
| Criterion | Mark | Basis |
|---|---|---|
| Recognition | 3 | Binding treaty text: the International Covenant on Civil and Political Rights, article 19, as the UN Human Rights Committee reads it in General Comment 34, paragraph 18; the Council of Europe Convention on Access to Official Documents (CETS 205); the UN Convention against Corruption, articles 10 and 13; SDG indicator 16.10.2. |
| Evidence | 2 | Much studied, with mixed findings: Escaleras, Lin and Register (Public Choice, 2010); Costa (Journal of Law, Economics and Organization, 2013); Cordis and Warren (Journal of Public Economics, 2014). |
| Reach | 3 | Every public body, and all of the information it holds. |
Open meetings: 4 points
| Criterion | Mark | Basis |
|---|---|---|
| Recognition | 1 | No global instrument requires public bodies to meet in public. The UN Convention against Corruption, article 13, asks for public participation in decision-making generally, and the Aarhus Convention covers environmental decisions only. |
| Evidence | 1 | No study isolating the effect of an open-meetings law was found. |
| Reach | 2 | The meetings where public bodies decide, a central but bounded part of their work. |
Oversight: 7 points
| Criterion | Mark | Basis |
|---|---|---|
| Recognition | 2 | UN General Assembly resolutions 66/209 (2011) and 69/228 (2014) on strengthening supreme audit institutions, which take note of INTOSAI's Lima (1977) and Mexico (2007) Declarations on audit independence; the Venice Commission's Principles on the Ombudsman Institution (2019), acknowledged by General Assembly resolution 75/186 (2020); the Paris Principles (General Assembly resolution 48/134, 1993). The UN Convention against Corruption, articles 6 and 36, binds states to independent anti-corruption bodies, which are one part of this section. |
| Evidence | 2 | Contested: Blume and Voigt (European Journal of Political Economy, 2011) find no clear-cut effect of how an audit office is designed, while Olken (Journal of Political Economy, 2007) and Avis, Ferraz and Finan (Journal of Political Economy, 2018) find that audits reduce corruption. |
| Reach | 3 | The conduct and accounts of every public body. |
Fiscal transparency: 7 points
| Criterion | Mark | Basis |
|---|---|---|
| Recognition | 3 | Binding treaty text: the UN Convention against Corruption, article 9(2), requires each state party to take appropriate measures to promote transparency and accountability in the management of public finances. Also General Assembly resolution 67/218 (2012) on the GIFT High-Level Principles on Fiscal Transparency, the IMF Fiscal Transparency Code (2019), the PEFA framework (2016), SDG indicator 16.6.1, and the OECD Best Practices for Budget Transparency (2002) as a reference. |
| Evidence | 2 | Some studies, consistent: Alt and Lassen (American Journal of Political Science, 2006; European Economic Review, 2006) find political budget cycles and higher debt where fiscal transparency is low, in one sample of OECD countries. |
| Reach | 2 | The budget and public money, a central but bounded function. |
Accessibility: 5 points
| Criterion | Mark | Basis |
|---|---|---|
| Recognition | 3 | Binding treaty text: the Convention on the Rights of Persons with Disabilities, articles 9 and 21; in Europe, the Web Accessibility Directive (EU) 2016/2102. |
| Evidence | 1 | One cross-sectional study (Kuzma, Yen and Oestreicher, 2009) finds more accessible government websites in countries with stronger disability laws; we found none that isolates the effect of a law. |
| Reach | 1 | One channel: public bodies' websites and digital services. |
Most published indices weight their parts equally. Equal weights are a judgment too, and not a neutral one: the same spread effect above means equal weights would let accessibility decide about half of the differences between countries. The RTI Rating, which scores records law feature by feature, sets points per feature by stated judgment; this rubric applies the same approach one level up.
How certain a rank is
A rank depends on the weights, and the weights rest on the judgment in the rubric. To show how much, the ranking is recomputed 600 times, each time with every rubric mark moved up or down by one at random (within 1 to 3) and the weights recalibrated. A country’s range runs from the 5th to the 95th percentile of its rank across those runs.
Across the 144 countries ranked, the median range is 9 places wide, and nine in ten are 18 places wide or narrower.
Two countries whose ranges overlap should not be read as clearly apart.
Fiscal transparency: the 10 dimensions
The fiscal transparency score is built from ten dimensions of budget and financial accountability law.
| Dimension | Max Points | What It Measures |
|---|---|---|
| Budget Publication | 15 | Requirement to publish annual budget before the fiscal year begins; machine-readable formats |
| Expenditure Reporting | 15 | In-year spending reports; transaction-level disclosure; online portal |
| Audit Independence | 15 | Supreme audit institution authority, independence from executive, public reporting |
| Procurement Disclosure | 10 | Competitive bidding requirements; contract publication; award justification |
| Beneficial Ownership | 10 | Disclosure of ultimate beneficial owners of government contractors |
| Debt & Liability Reporting | 10 | Public reporting of total debt, contingent liabilities, pension obligations |
| Revenue Transparency | 10 | Disclosure of tax expenditures, intergovernmental transfers, fee schedules |
| Asset Disclosure | 5 | Public official financial disclosure; public registry of government assets |
| Citizen Budget | 5 | Plain-language budget summary accessible to the general public |
| Enforcement | 5 | Penalties for non-disclosure; legislative review authority |
| Total | 100 |
Open records: how a records law is scored
The open records score says how strongly a records law protects the right to request government records, out of 100. It comes from one of two places, and each records page names which.
The RTI Rating. Where the Right to Information Rating (Access Info Europe and the Centre for Law and Democracy) rates the law, the score is that rating put on a 100-point scale: the rating is out of 150, so 127 becomes 85.
UnGovr’s own rubric. Where it does not, which is most state and province laws, we read six features of the statute:
| Feature | Max Points | What It Measures |
|---|---|---|
| Response deadline | 25 | How soon the agency must answer a request |
| Fees | 20 | What a requester can be charged, and any free allowance or waiver |
| Appeal route | 20 | Whether a refusal can be appealed, and to whom |
| Ways to file | 15 | How many ways the law allows a request to be made |
| Right to sue | 10 | Whether a requester can take a refusal to court |
| Penalties and legal fees | 10 | Penalties for unlawful withholding, and whether a winning requester recovers legal fees |
A law that is not in force (enacted but never commenced, or suspended) shows Not in force on its records page, with its score as written beside it.
The other sections
Open meetings and oversight each have their own methodology page with their dimensions and scoring rules.
Reading the Accessibility column. A number is the section score (0 to 100). None means we investigated the jurisdiction and found no qualifying digital-accessibility statute: only physical-access disability law, a general anti-discrimination law, treaty ratification, or non-binding guidance, none of which meet the standard of a separately citable statute with explicit web and ICT accessibility provisions. It counts as 0 in the score. A dash means the section is not assessed for that jurisdiction. Digital-accessibility law is a young, largely European and developed-economy legal category, so “None” is the correct and common result for much of the world.
- Open meetings score methodology: advance notice, public comment, closed session limits, documentation, enforcement
- Oversight score methodology: appointment independence, funding, removal protections, investigative authority
- Accessibility laws: WCAG standard required, scope, enforcement, monitoring requirements
What the score measures
The score evaluates the quality of law, not government performance. A jurisdiction with strong laws may have poor compliance; a jurisdiction with weak laws may perform better than expected through culture or practice. The score is a starting point for inquiry.
Data comes from UnGovr’s own research into statutory and regulatory text. Where law is ambiguous, we apply conservative interpretations: uncertain fields score zero rather than receiving partial credit.
Known limitations
- The weights rest on judgment. The rubric marks are stated and cited above, and the rank ranges show how far a different judgment would move each country.
- Subnational coverage is uneven. A state or province is scored only when three quarters of the points are assessed for it, and it is not ranked.
- Law changes. Scores reflect the law as of the date the data was last updated. Legislative changes may lag by months.
Edition: 2026