Walka Advisory

Data sources & coverage

Five public feeds underpin the dashboard. Coverage is near-complete for high-income jurisdictions and thins for lower-income tiers — the inclusion criteria at the bottom of the page explain which administrations are excluded from the headline analysis.

ISORA

International Survey on Revenue Administration. A multi-organisation survey of tax administrations run jointly by the IMF, OECD, IOTA and CIAT.

  • Dataflow: ISORA_LATEST_DATA_PUB v4.0.0 (published 2025-07-10).
  • Coverage: FY2018–FY2023. FY2024 not yet released (~18-month survey lag).
  • Portal: data.imf.org (SDMX API at api.imf.org).
  • Accessed: 2026-04-17.
ITTI

Inventory of Tax Technology Initiatives. OECD-led snapshot of digital, automation and analytics adoption across tax administrations. Used to derive the technology-adoption binaries (cloud, RPA, AI/ML, network analysis, data engineering).

  • Snapshot: FY2022, refreshed January 2025.
  • Coverage: 107 jurisdictions.
  • Portal: data-explorer.oecd.org (dataflow OECD.CTP.TAV:DSD_QDD_ITTI@DF_QDD_ITTI).
  • Access: public CSV.
OECD PIT & Social Security rates

Headline personal income tax and social-security contribution rates across OECD member states under different earnings, family and civil-status scenarios. Used in the OECD PIT & SS Rates section of the dashboard.

  • Dataflow: OECD.CTP.TPS:DSD_TAX_WAGES_COMP@DF_TW_COMP v2.1.
  • Coverage: 38 OECD member states.
  • Portal: data-explorer.oecd.org (SDMX API at sdmx.oecd.org).
World Bank

GDP, GDP per capita, population and GNI per capita tier classifications — used as macroeconomic context and to anchor the income-vs-maturity regression. Income-group thresholds are annual Gross National Income (GNI) per capita in USD under the World Bank Atlas method, a measure of average income per resident (distinct from GDP, which captures total economic output). Thresholds reflect the FY2025 classification and are updated annually.

  • Indicators: GDP (current USD) NY.GDP.MKTP.CD; GNI per capita tiers (Atlas method, FY2025 classification); population.
  • Coverage: all World Bank member states.
  • Portal: data.worldbank.org (public CSV / WDI API).
Currency conversion

Revenue figures reported in local currency are converted to USD using World Bank annual average exchange rates for the relevant year. GDP is taken directly from the World Bank in current USD.

Inclusion criteria

Three thresholds restrict the headline analysis to administrations that genuinely run a PIT regime at a meaningful scale — leaving 102 jurisdictions in-sample. Each threshold exists to keep the misfit rankings tracking substantive differences in PIT administration rather than structural or data-quality artefacts.

ISORA coverage ≥ 25%

Drops administrations that didn’t actually submit an ISORA return for FY2023. Their feature vectors would otherwise be almost entirely missingness flags, which under StandardScaler turn into outlier spikes that distort the analysis without representing real capability differences.

Excluded under this rule: Afghanistan, Comoros, Ethiopia, Iraq, Jordan, Lebanon, Libya, Myanmar, Philippines, Russian Federation, Sudan, South Sudan.

PIT ≥ 5% of total tax revenue

Drops jurisdictions where PIT isn’t materially part of the revenue mix — tax havens, oil monarchies, very small states — or where PIT revenue isn’t reported at all. Without this filter these jurisdictions would dominate the misfit rankings simply by looking unlike a typical PIT administrator, rather than by being meaningfully above or below their peers.

Excluded under this rule: Afghanistan, Antigua and Barbuda, Burundi, Bosnia and Herzegovina, Belarus, Belize, Bermuda, Bolivia, Brunei Darussalam, Central African Republic, Congo, Dem. Rep., Comoros, Cuba, Ethiopia, Guinea, Equatorial Guinea, Guatemala, Iraq, Jordan, Lao PDR, Lebanon, Libya, Maldives, Mali, Myanmar, Niger, Nigeria, Pakistan, Philippines, Paraguay, Russian Federation, Saudi Arabia, Sudan, El Salvador, South Sudan, Turks and Caicos Islands, Chad, Tonga, British Virgin Islands, Vanuatu.

Population ≥ 1.5M

Drops small states. PIT-administration profile is structurally different at this scale — whole ISORA survey blocks are skipped, simplified systems are common, and there’s often no functional separation between tax types — so these jurisdictions would dominate the misfit list through scale rather than substantive differences. 1.5M is the World Bank Small States Forum / IMF threshold used in tax-policy literature; 1M (the academic ‘microstate’ threshold) is a defensible alternative that would keep Trinidad, Mauritius, Cyprus, Bahrain and Estonia in the analysis.

Excluded under this rule: Aruba, Antigua and Barbuda, Belize, Bermuda, Barbados, Brunei Darussalam, Bhutan, Comoros, Cabo Verde, Cyprus, Dominica, Estonia, Fiji, Micronesia, Fed. Sts., Grenada, Guyana, Iceland, Kiribati, St. Kitts and Nevis, St. Lucia, Luxembourg, Maldives, Marshall Islands, Malta, Montenegro, Mauritius, Nauru, Palau, Solomon Islands, Sao Tome and Principe, Suriname, Eswatini, Seychelles, Turks and Caicos Islands, Timor-Leste, Tonga, Trinidad and Tobago, Tuvalu, St. Vincent and the Grenadines, British Virgin Islands, Vanuatu, Samoa.