Who are the particularly vulnerable that adaptation finance is for? And did they become less vulnerable because of it?
The Green Climate Fund's record cannot answer either question. It can say which countries the money went to, and on that, the news is good. It cannot say which people it reached. Below: first the good news, then every single-country adaptation project the Fund has approved, set against the population it claims to serve.
Above the 50 percent floor the Board set.
Against 14.1 for regular proposals (−5.4; 90 percent interval −7.3 to −2.9).
13.3 against 12.7; indistinguishable from zero (−2.0 to 2.8). Delay does not roll downhill.
Each circle is one single-country adaptation or cross-cutting project, placed by the Fund's own adaptation indicator: the total of direct and indirect beneficiaries relative to total population (Board decision B.07/04, 2014). Right of the dashed line, direct and indirect beneficiaries add up to more than the country has residents. Gray: they exceed it only when the two counts are added, usually because the whole nation is counted as indirect beneficiaries. Black: the larger of the two still exceeds the most generous population ceiling. The implementers filled in the form the Fund gave them; the form cannot tell a resident from a beneficiary. Click a circle.
Reported counts against the population of the area served. It sorts the circles above from a desk: gray where the count is a convention, black where it exceeds the population under any reading.
A published tabulation script runs inside the implementer's systems and writes only aggregates by sex, age, tail group and intensity of support, attested by checksum.
With consent, in the respondent's language: does the person exist, was the support targeted, how much arrived, where does the household sit in the tail.
A linkage trustee compares encoded identifiers across projects and returns only the number of overlaps. No central register of the poor.
Data. Green Climate Fund open data, projects endpoint, 372 approved projects, frozen 12 October 2026; 216 single-country adaptation and cross-cutting projects with population data (of 218) (Cook Islands and Niue lack World Bank population). Beneficiaries are lifetime projections reported by accredited entities; population is World Bank SP.POP.TOTL, latest year 2020 to 2025. Generous reading (added after first results, logged as deviation D11): the larger of direct and indirect, against the larger of the highest population 2020 to 2025 and the latest population grown to the project's end year. A count above the population is a statement about the Fund's definition, not proof that any person was not reached; no implementer is accused of misconduct. Vulnerability: INFORM Risk Index Mid 2026 (primary) and ND-GAIN 2023 (sensitivity). Intervals: bootstrap, 90 percent. Registered analysis; deviations logged before results.