Source agency directory

2,154 World Bank series placed on the What hierarchy, and the agencies FAND draws on.

FAND deliberately cherry-picks a small number of key indicators from each source agency. The point is to let users leave the site to verify — and come back. Every published value records the agency, table, year, and a direct URL to the original data.

How good are these sources? The agencies below are graded by others: the IMF’s Data Quality Assessment Framework grades the producer, Open Data Watch’s Open Data Inventory grades the release. FAND, which compiles from them, is graded by neither, so it grades each number instead. See Assessing National Statistics.

The World Bank’s series, placed on the What hierarchy

FAND draws a few numbers from many sources. Before choosing them, we took inventory: every series in ten World Bank Databank databases, placed one by one on the node of the What hierarchy it measures. The hierarchy is the 298-node map of what a balance sheet and income statement for a place could record. The inventory shows where the Bank’s statistical effort lands and, as plainly, where it does not.

Open any node below. The number beside it counts the series at or below it; inside, “series placed on this node” lists each one with its code and the databases that carry it. Greyed nodes have none: those gaps are what a balance sheet for a place must fill from other agencies, or estimate.

DatabaseEntriesPedigreeDatabank
World Development Indicators (WDI)1,5164source
Health Nutrition and Population Statistics (HNP)4744source
Sustainable Development Goals (SDG)4074source
Wealth Accounts (WA)1463source
Millennium Development Goals (MDG)1324source
Enterprise Surveys (ES)1153source
Environment, Social and Governance (ESG)713source
Identification for Development (ID4D)713source
International Comparison Program 2021 (ICP)454source
Worldwide Governance Indicators (WGI)363source

Pedigree grades the database’s own method (4 established or benchmark, 3 survey or composite), on the scale used in NUSAP-lite. The placements are ours and grade lower, 2: each series went to exactly one node by rule (its code prefix, keywords in its name, and the Bank’s topic), not by expert review. They were made in March 2026 on that month’s tree and carried to the current one, Rev7 of 20 October 2026, by a node-by-node crosswalk. A few series were then placed on nodes of their own: ICP headings on their expenditure heading, reserves on cross-border claim-rights, the debt components the Bank’s Debtor Reporting System sums on sovereign debt, other external debt stocks on cross-border duties, and agriculture and manufacturing value added on their industry sections.

The hierarchy, populated

2,154 distinct series (3,013 database entries) on 89 of 298 nodes. Direct series reach 58 of 228 leaves. The number beside each node counts the series at or below it; greyed nodes have none. FAND reads marks the 102 series FAND's pipeline reads, as inputs or comparators.

Series lists: March 2026 Databank extracts. Hierarchy: TL3What_Hierarchy Rev7, 2026-10-20.

The agencies FAND draws on

The inventory above covers one institution’s ten databases. FAND’s tables also come from the agencies below. Open a group to see them.

International sources (country-level data): 8 agencies
AgencyScopeURL
WDI — World Bank World Development IndicatorsGNI Atlas, GDP, population, education indicatorsdata.worldbank.org
IMF — International Monetary FundBalance of Payments (BOP), International Investment Position (IIP), Government Finance Statistics (GFS), World Economic Outlook (WEO)imf.org/data
FAO — Food and Agriculture Organization (FAOSTAT)Agricultural land, forestry, fisheriesfao.org/faostat
USGS — US Geological SurveyMineral Commodity Summaries (global reserves and production for 31 commodities)usgs.gov/centers/national-minerals-information-center
EIA — US Energy Information AdministrationGlobal energy reserves and production (fossil fuels)eia.gov/international
OECD — Organisation for Economic Co-operation and DevelopmentProductivity, fixed capital stock, depreciationstats.oecd.org
UN WPP — UN Population Division, World Population ProspectsPopulation, working-age share, life expectancypopulation.un.org/wpp
BIS — Bank for International SettlementsLocational banking statistics, cross-border flowsbis.org/statistics
US sources (state and county data): 8 agencies
AgencyScopeURL
BEA — Bureau of Economic AnalysisNational accounts, regional accounts, FDI positionbea.gov
BLS — Bureau of Labor StatisticsEmployment, wages by industry (QCEW), productivitybls.gov
US Census — Census BureauPopulation, ACS demographics, housing, County Business Patterns, SAIPE (county income)census.gov
EIA — Energy Information AdministrationUS energy production by stateeia.gov
USGS — National Map and minerals dataUS mineral deposits by state and countyusgs.gov
NOAA — National Oceanic and Atmospheric AdministrationClimate, fisheries, coastal economynoaa.gov
USDA NASS — USDA National Agricultural Statistics ServiceUS cropland, rental rates, agricultural productionnass.usda.gov
IHME / CHR — Institute for Health Metrics and Evaluation / County Health RankingsLife expectancy and health metrics at state and county level (human-capital component)healthdata.org · countyhealthrankings.org

The UX pattern

Every data value on a country / state / county page carries a small source marker. Click it → see the agency, table name, year, and a direct URL to the original data. Not a separate bibliography page — inline, contextual, and immediate.

The four-axis architecture treats Source as a first-class navigation axis: every number on the site has a verifiable provenance chain back to a public agency.

Directions for further work

  1. Refresh from the API, series by series. Re-pull the ten databases’ series lists through the World Bank API with each series’ last-updated date, so the inventory carries a vintage per series rather than one March 2026 snapshot.
  2. Inventory the other agencies the same way. Place the IMF’s (IFS, BOP, IIP, GFS), FAO’s, OECD’s and UN’s series lists on the same hierarchy. That would show whether the nodes the Bank leaves grey (liabilities in detail, ecosystem services beyond forests, cross-border positions) are filled elsewhere, and by whom.
  3. Review the placements where they concentrate. An expert pass on the five nodes holding 64% of the series, starting with Human Resources, would split them toward the leaves they measure and raise the placements’ pedigree from 2 toward 3.
  4. Finish placing cross-border series. Reserves, external debt stocks and the Debtor Reporting System components now sit on the cross-border nodes. Debt service, remittances and the Wealth Accounts’ net foreign assets still sit where the March rules put them, and should be placed by the same test: gross position, net position, or flow.
  5. Build the same accordion for US counties. Place the county series FAND reads (BEA’s county income and product accounts, including the residence adjustment that 1.1.2.3 and 1.2.1.2 capitalize; BLS QCEW; Census ACS and County Business Patterns; IRS county income; USDA and USFS land and timber) on the same hierarchy. This needs the county pages’ row codes aligned with the current hierarchy first: a few still carry April codes (produced assets at 1.1.1.1, hydropower at 1.1.2.2.4.4).