More money, fewer readers? The funding-vs-learning paradox

What happened

Over the past decade governments and multilateral bodies have made some of the most explicit financing commitments in the history of mass education. In 2015, the World Education Forum adopted the Incheon Declaration / Education 2030, urging states to commit 4-6% of GDP or 15-20% of public expenditure to education. In 2020, Brazil made its basic-education fund (FUNDEB) permanent, constitutionally locking in financing and committing to raise the federal supplement from 10% to 23%. Yet over the same window the headline learning numbers moved the wrong way. According to the World Bank, the Learning Poverty Rate — the share of 10-year-olds who cannot read and understand a simple text — rose from 57% before the pandemic to an estimated 70% in 2022. On the OECD's PISA mathematics scale, run by the OECD, the OECD average fell from 489 in 2018 to 472 in 2022, the largest single-cycle drop on record.

Why it matters

The Factrail causal graph treats Public Education Funding as a structural enabler: it underwrites teacher salaries, class sizes, materials and digital access. The graph links the funding driver to lower learning poverty and to a lower out-of-school rate, but with multi-year lags (the model assigns roughly a five-year lag to the learning-poverty channel). That lag is central to the paradox. Financing reforms signed in 2015 and 2020 cannot, by the graph's own logic, have shown their full effect in the 2022 readings — and those 2022 readings were dominated by a pandemic shock the funding driver did not control.

How the fact(s) move the driver

Factrail's fact-to-driver links code both Incheon and FUNDEB as strengthening public education funding, with FUNDEB the stronger and higher-confidence of the two. The driver's own intensity series tells a more tempered story: it climbed from the early-2000s baseline to a 2015 peak around the Incheon commitments, dipped under post-COVID fiscal pressure in 2020, and only partially recovered by 2023. In other words, commitment rose, but realized, sustained financing flattened — and much global spending still falls short of the 4-6%-of-GDP benchmark Incheon set.

What changed relative to the normal line

Both indicators sit far from their norm lines. The learning-poverty norm is 0% (SDG 4's aspiration that no child reaches age 10 unable to read); a 70% reading is catastrophically above it. PISA's reference line is approximately 500 points, set near its early-2000s calibration; at 472 the OECD average has slipped well below its own historical baseline. The signal is not that funding failed mechanically, but that other forces overwhelmed it in this window.

Alternative explanations

Several readings compete, and Factrail does not assert a single cause. First, the lag: the strongest financing reforms are recent and their learning dividends may still be pending. Second, Teacher Quality and Supply — the most proximate in-school driver in the graph — barely improved over the period and was strained by pandemic disruption. Third, the COVID-19 school closures of 2020 inflicted a one-off learning shock that is plausibly the dominant explanation for the 2022 numbers. Fourth, money spent is not money spent well: governance and targeting determine whether budgets reach classrooms. These are interpretive links flagged at medium-to-low confidence, not proven attributions.

What may happen next

If the recent financing reforms hold and translate into teacher supply, the graph's lagged logic implies the learning-poverty rate should bend back down over the second half of the decade — but only partially, and unevenly across countries. Factrail's accompanying education forecast treats a partial recovery as the baseline, not a return to the norm line.

Sources

The figures cited here come from the World Bank's learning-poverty reporting, the OECD's PISA 2022 results, and the Incheon and FUNDEB policy records. Each indicator reading marked real in the dossier is anchored to an official figure; interpolated intermediate years are flagged and excluded from strong claims.