One education-funding pledge, two yardsticks: 46 of 186 countries clear both, 67 clear neither — and the two floors split by wealth

At the World Education Forum in Incheon in May 2015, governments adopted the Education 2030 Framework for Action, the roadmap for implementing Sustainable Development Goal 4. Its financing section asks countries to "allocate at least 4% to 6% of gross domestic product (GDP) to education and/or allocate at least 15% to 20% of public expenditure to education". The two thresholds are joined by "and/or": a country is treated as meeting the benchmark if it clears either the GDP floor or the budget-share floor. We pulled both series from the World Bank's compilation of UNESCO Institute for Statistics data — government expenditure on education as a share of GDP (SE.XPD.TOTL.GD.ZS) and as a share of total government expenditure (SE.XPD.TOTL.GB.ZS), database vintage last updated 2026-07-13 — and took each economy's most recent available year (mostly 2022–2024). Against the lower bound of each range: 86 of 190 reporting economies spend below 4% of GDP on education, and 125 of 189 spend below 15% of their budget. Of the 186 economies that report both figures, 46 clear both floors, 67 clear neither, and 73 clear exactly one. Fifty-five clear only the GDP floor and 18 clear only the budget floor. The median reporting economy spends 4.17% of GDP (above that floor) and 12.93% of its budget (below the other). The United States reports 5.42% of GDP and 11.29% of budget (2021); Viet Nam reports 2.89% of GDP and 15.45% of budget (2022); South Africa clears both, at 6.02% and 18.52% (2024).

The following is analysis, not fact. The "and/or" in the 2015 text is an escape hatch, and a deliberate one: a rich country with a large GDP and a small education share can still qualify, and so can a poor country that devotes a large slice of a small budget to schools. Two doors, either one counts. What the joined series show is that the two doors are not equivalent — they sort countries by wealth ^mmodeled. Of the 55 economies that clear only the GDP floor, 49 are upper-middle- or high-income, and 33 are high-income. This is the group that spends 4% or more of a large national output on education but for whom education is a modest line in a big government: the United States (5.42% of GDP, 11.29% of budget), the United Kingdom (5.91% / 11.82%), France (5.32% / 9.12%), Germany (5.24% / 10.68%), Canada (4.84% / 10.65%) and Brazil (5.62% / 12.93%) all pass on GDP and fail on budget. The GDP yardstick flatters them because their governments are large relative to schooling, not because schooling is a priority within the budget ^mmodeled. The budget-share floor runs the other way. Of the 18 that clear only that door, half are lower-middle- or low-income, six of them low-income. Viet Nam (2.89% of GDP but 15.45% of budget), the Philippines (3.93% / 16.55%), Guatemala (3.08% / 22.99%) and Sierra Leone (2.62% / 19.96%) devote a large share of a small budget to education, clearing the 15% door while missing the 4% one. For these states the budget-share floor is often the only benchmark they can reach, because 4% of a low GDP is a sum their treasuries do not command ^mmodeled. So the same pledge issues opposite verdicts on the same country depending on which of its two numbers is read. The United States "meets Education 2030 financing" by GDP and misses it by budget share; Viet Nam is the mirror image. For 73 of 186 economies — 39% — compliance is a matter of which agreed yardstick you pick up.

The strongest counter is that the "and/or" makes most of this a non-story: the framework asks countries to hit either floor, so a country clearing one has met the benchmark, full stop, and pairing the two measures to manufacture "split verdicts" reads a contradiction into a rule that was written to avoid one. On that reading the honest headline is that 119 of 186 economies clear at least one floor and only 67 clear neither — a benchmark two-thirds met, not a pledge in disarray. That is a fair point, and this piece reports the 119/67 split alongside the rest; the finding is not mass failure but that the two supposedly interchangeable doors are sorted by income, so "which door a country walks through" is itself information the single "and/or" verdict hides. A second, sharper caveat is about the data, not the rule. UNESCO's figures are meant to cover general government — central, regional and local — but coverage is uneven, and federal states are where it frays: Nigeria's reported 0.32% of GDP on education almost certainly reflects central-government-only reporting that omits state spending, not the true national total. Cross-country comparisons that lean on such figures carry a coverage caveat, which is why this piece anchors its claims on the distribution and on well-reported economies rather than on any single outlier. The benchmark language ("should aim to") is a recommendation, not a binding target; we report shortfalls against a stated floor, not a broken law.

From inside a finance ministry, the choice of yardstick is not academic — it is which number goes in the speech. A wealthy government defending its record cites the GDP share, where it clears the bar; a poorer government defending the same record cites the budget share, where a fifth of every dollar it spends goes to schools. Both are quoting the identical 2015 commitment, and both are, by its own terms, correct. The framework's "and/or" was written to let unlike economies sign the same page; its side effect is that no single number tells a citizen whether their government kept the promise, because the promise shipped with two rulers and no instruction on which to trust.

A correspondent checks one country against the pledge and finds it passing: 4% of GDP, box ticked. The gap opens only when all 186 economies that report both numbers are laid against both floors at once. Do that and the benchmark's two "interchangeable" doors turn out to open for different countries — 49 of the 55 that clear only the GDP floor are upper-middle- or high-income, while the budget-share floor is the door poorer states walk through. The 2015 framework let 160 governments adopt one commitment with two yardsticks, and a decade on those yardsticks disagree about who is keeping it for 73 of them. The United States passes by one measure and fails by the other; Viet Nam fails and passes on the reverse. No single-country desk sees this, because from inside one country there is only ever one number worth quoting — the one that clears.