Who Funds the Pacific?

Two tools from the measurement layer series, pointed at the Indo-Pacific. The first exposed an error in a piece published the day before. The second shows that the smallest states in the world are almost fully covered in the global SDG database — and that this says less than it seems.

Part 1 · The $91 billion donor to Tonga

Ask d-portal, the standard IATI query interface, for every activity tagged to Tonga and sum the spend: $97.2 billion, 94% of it from the U.S. Department of State. Tonga’s GDP is about $0.5 billion. The cause is one activity, the U.S. Foreign Military Financing program: $90.2 billion, tagged to 131 countries, 53% Israel, 14% Egypt, 0.005% Tonga. IATI records the percentage. The naive query ignores it and credits the full amount to every country on the list.

Across fifteen countries, $1,797 billion of naive spend becomes $260 billion weighted. The share removed is highest for the smallest states, whose own aid is tiny beside the global programs they are tagged into.

Who actually funds the Pacific

Weighted, the picture is the one a desk officer would recognise. Australia is the largest reported funder in eight of nine Pacific countries; New Zealand is second or third in six; the Asian Development Bank and the World Bank fill out the top four. In Timor-Leste the largest “funder” is AidData, a research organisation that publishes third-party data on Chinese finance into IATI, with Australia second. The U.S. State Department, first in eleven of fifteen countries under the naive query, is not in the top six of any Pacific country.

The Pacific has few reporting organisations (37 in Tuvalu to 83 in Papua New Guinea) and moderate concentration: one to three funders reach half of spend, four to six reach 80%. Gini coefficients of 0.85–0.92 are below East Africa’s 0.93–0.95, and the long tail of tiny actors is much shorter.

Part 2 · What the SDG database knows about Tuvalu

The second tool queries the UN SDG Global Database for 33 headline indicators across all 17 goals and records, for each country, whether data exists and who produced it. Twenty-five countries: thirteen Pacific (including Timor-Leste), eight in Southeast and South Asia, and four high-income comparators.

Coverage is high even for the smallest states in the world. Tuvalu has data for 26 of 33 indicators; Nauru and Palau 27; Fiji 32, the same as Australia. Nineteen indicators have data for every one of the 25 countries. Pacific states average 29.2 of 33 (89%), Asia 31.6 (96%), the high-income comparators 30.0 (91%) — New Zealand and Singapore do not report the poverty indicators.

The gaps are specific, and they are the hard ones. Goal 10 (inequality) is measured in three of thirteen Pacific states: labour share of GDP has 15% coverage, bottom-40% income growth 23%. Undernourishment (54%), safely managed drinking water (62%), learning proficiency (69%) and unsentenced detainees (69%) follow. These need household surveys large enough to estimate the bottom 40%, learning assessments, and administrative justice data. No global model can estimate the income growth of Tuvalu’s poorest 40%.

Pacific states produce as much of their own data, proportionally, as Australia. Country-produced or adjusted data is 60% of the Pacific’s available indicators, 59% of Asia’s and 58% of the comparators’. Micronesia (67%) and Palau (63%) exceed Australia and Japan (53% each). This matches the global finding of the earlier SDG audit: about half of everyone’s SDG data comes from the global statistical apparatus, whatever their income. The coverage rate measures the completeness of the database, not the capacity of the statistical office.

What the two parts say together

Method

Aid: d-portal.org, activity and country tables joined on activity identifier, all activities with positive spend, ~78,000 activities across 15 countries. Weighted spend = lifetime spend × declared country_percent / 100. Lifetime totals, not annual. Tool: donor_concentration_weighted.py. SDG: UN SDG Global Database API, 33 indicators × 25 countries, presence of any data and the Nature attribute of the most recent records. Tool: sdg_coverage_indopacific.py. Both are samples, not censuses. Code and results in the public repository.