BLOG POST

“PISA-Shocks” in Africa

Des “chocs PISA” en Afrique

When the OECD published its first international student assessment (PISA) in 2000, several countries were taken aback by just how poorly students performed. This “PISA-Shock” prompted comprehensive reforms in Germany, resulting in measurable gains in learning by 2012. Cross-national assessments can influence policy through multiple channels: they can trigger new reforms, give reformers political ammunition to push through difficult changes, and stimulate cross-country policy borrowing.

Less well documented are similar shocks in Francophone Africa that followed the regional assessment, Programme d'Analyse des Systèmes Éducatifs de la CONFEMEN (PASEC). This blog post summarizes four cases: Benin, Côte d'Ivoire, Niger, and Senegal.

Côte d'Ivoire is the most dramatic. When the PASEC 2019 results were released, they made national news: only 17 percent of children had satisfactory mathematics levels by the end of primary school. The government responded with an unprecedented national consultation on education—the États Généraux de l’Éducation Nationale—that produced 42 reform measures backed by a planned $1.4 billion budget over 10 years. The centrepiece is the National Programme for Improving Early School Learning or PNAPAS, a cognitive science-based early reading programme now rolling out nationwide. Ivorian officials explicitly cited PASEC as a “compass” for the reform process. This case is a textbook assessment shock—poor results widely reported, political mobilisation, and a large-scale policy response that would not have happened without comparable cross-national data.

Niger experienced a similar shock, but earlier. PASEC 2014 revealed that 9 in 10 students had inadequate reading and mathematics skills, placing Niger last among all participating countries. The government responded with politically difficult reforms—evaluating and dismissing underperforming contract teachers, introducing national languages into early-grade curricula, and making teacher training a core focus of the education sector plan. The PASEC 2019 report described Niger as the most striking example of reform triggered by assessment results. And the results showed: Niger recorded the largest gains of any sampled country at the end of primary school. Average scores rose by 68 points in reading and 56 points in mathematics, equivalent to roughly 0.6 standard deviations—a huge improvement by any standard. Niger went from the lowest-performing system in 2014 to a level approaching the sample average in five years.

Niger responded to poor 2014 PASEC results with major reforms, followed by rapid improvement

Other countries show a softer version of the same pattern. Benin and Senegal both reported using PASEC results to inform their education sector plans, and both recorded substantial gains between PASEC 2014 and 2019. The causal link is less direct; these are cases where PASEC data shaped ongoing reform efforts rather than triggering a dramatic new response, but the direction is consistent.

PASEC has existed in its current format–designed for comparability over time and across countries–since 2014. In just two cycles—PASEC 2014 in 10 countries and PASEC 2019 in 14 countries—there have been at least two well-documented cases in which PASEC results directly triggered major policy reform. That is roughly one shock per cycle.

Not every cycle produces a shock in every country. But the historical pattern of at least one major reform generated per cycle means the expected value is high.

PASEC funding is at risk

In January 2026, Agence Française de Développement, France's development agency, announced it would no longer fund the PASEC secretariat—the small team in Dakar, Senegal, that coordinates the only comparable learning assessment across Francophone Africa. AFD will continue supporting individual country evaluations, but the permanent technical capacity that makes PASEC work as an international programme is now at risk. This is in the context of major cuts to total aid budgets from France and other donors.

Given the increasing strain, how much should donors continue to prioritise assessment over trying to improve outcomes more directly? This comes down to a return on investment decision, and our analysis suggests the returns for PASEC are high.

A simple cost-benefit analysis of PASEC

First, how much does implementing PASEC cost? The most recent cycle cost about $9 million per year across 21 countries, covering around 100 million children and combined government spending on education of roughly $14.5 billion per year. That works out to about 9 cents per child per year, or one dollar in every 1,600 spent on education.

The next cycle will be cheaper. PASEC has refocused its assessment on primary education and reduced its international costs, cutting per-country costs by around 40 percent. The cost per child should fall to roughly 5 cents.

How about the benefits? If PASEC costs less than a tenth of one percent of all education spending in these countries, that is how much it needs to improve the effectiveness of that spending to pay for itself. If having diagnostic data on learning outcomes leads to even marginally better decisions than having no data at all, PASEC clears the bar.

1. The value of PASEC-shocks

First, as we discussed above, the PASEC-shock reforms have likely led to substantial improvements in learning. Several studies connect improvements in early grade learning directly to higher lifetime earnings. Roughly, a one standard deviation improvement in test scores is associated with 11 percent higher lifetime earnings. So even a modest 0.05 standard deviation improvement—well below what a structured pedagogy typically achieves—translates to a 0.55 percent earnings increase for affected individuals.

Applied to Côte d'Ivoire, the arithmetic works as follows. The country has 5 million school-age children. A national reform like PNAPAS, triggered by PASEC, targets the entire school-age population rather than just a pilot group. Assume it improves learning by a conservative 0.05 SD. With a present value of lifetime earnings around $49,000 per person (at GDP per capita of $2,700, a 40-year working life, and a 5 percent discount rate), the gain per person is about $270. That gives total lifetime earnings gains of roughly $1.35 billion—for one country, from one assessment cycle. PASEC's total cost for all 21 countries over five years is $43 million.

Even under these deliberately conservative assumptions, this single channel produces a benefit-cost ratio of around 31:1. That is, counting only one reform, in one country, in one cycle.

2. Donors need PASEC too

International donors, including the World Bank, AFD, Global Partnership for Education (GPE), the Gates Foundation, Swiss Agency for Development and Cooperation, and UNICEF, use PASEC data extensively to design, monitor, and evaluate their education programmes. PASEC data are cited in at least 8 to 10 World Bank project appraisal documents for Francophone Africa, in multiple GPE country programme documents, AFD project frameworks, and UNESCO reports.

Education-related aid to sub-Saharan Africa amounts to roughly $4.5 billion per year, of which an estimated $1.8 billion flows to PASEC countries. If PASEC data improve the targeting of even 0.5 percent of this aid, the value is approximately $9 million, roughly equal to PASEC's annual cost.

This is not speculative. Donors design results frameworks around PASEC benchmarks, set disbursement triggers linked to PASEC indicators, and evaluate programme success using PASEC data. When a World Bank education project in Senegal cites PASEC 2019, which shows 65 percent proficiency in grade 6 mathematics, that indicator shapes the allocation of hundreds of millions of dollars.

3. PASEC drives research

Francophone Africa is one of the most under-researched regions in the world in terms of education. The REAL Centre at the University of Cambridge found that 95 percent of publications indexed in international databases on foundational literacy and numeracy are in English. A prior review found only around 150 to 200 peer-reviewed articles from 2000 to 2018 on education in Francophone Africa, compared to over 2,000 studies on Anglophone countries. Francophone Africa is around 30 percent less populous than Anglophone Africa, but produces roughly 10 to 15 times less research on education.

In this context, PASEC is, for many research questions, the only dataset available. It is the sole source of comparable learning outcome data across Francophone Africa. Without it, the region would be largely invisible in global education research.

PASEC scores are a core ingredient in the Harmonized Learning Outcomes (HLO) database, the most comprehensive global dataset on education quality, covering 164 countries. Through HLO, PASEC data feeds into the World Bank's Human Capital Index, the Learning Poverty indicator, the SDG 4.1.1 monitoring framework, and Our World in Data. PASEC feeds the global infrastructure that tracks whether the world is making progress on education. Without PASEC, these global metrics and databases would be missing learning data for roughly 20 countries.

National assessments aren’t enough

One natural objection is couldn't countries just run their own national assessments? Some already do. But there are at least three reasons why the international dimension is essential, and why the benefits described above cannot be achieved through national evaluations alone.

First, comparability drives accountability. When Benin's education minister sees that Senegal's students are outperforming Benin's, it creates political pressure that no national assessment can replicate. This is precisely what makes cross-national tests politically potent. Exposing performance relative to peers is more costly for governments than domestic assessments. It is also what gives reform-minded leaders a credible tool to mobilise public support for change. The fact that 21 countries voluntarily pay to participate, often with scarce domestic resources, is itself evidence that they value the comparison.

Second, quality assurance requires independence. National assessments controlled entirely by the ministry they are supposed to evaluate face obvious incentive problems. PASEC's international governance and standardised methodology provide a credibility that purely national exercises cannot match.

Third, the international programme generates economies of scale in technical development (test design, psychometric analysis, sampling methodology) that no individual country could afford on its own. The marginal cost of adding a country to PASEC is far lower than the cost of that country building equivalent capacity from scratch. These supply-side constraints are particularly binding in Africa, where small economies, linguistic diversity, and limited fiscal space raise the relative costs of cross-national testing. International donor support is therefore not a luxury but a precondition for sustaining regional assessment.

So what now?

Even under conservative assumptions, PASEC's benefits exceed its costs by a comfortable margin. Counting just one successful reform in one country per decade with a very conservative learning improvement, yields a benefit-cost ratio of around 31:1. Adding the planning, research infrastructure, and donor coordination channels only widens the margin.

At a few cents per child per year, PASEC is not expensive. The question is not whether Francophone Africa can afford it. The question is whether it can afford not to have it, and whether the international community will let a proven, cost-effective piece of education infrastructure disappear because of a coordination failure among donors.

In an ideal world, we’d have a regular comparable assessment covering every country and including all children of a specific primary-school age, such as 9 or 10. In its absence, donors should support existing platforms such as PASEC. Given the fiscal stressors that France, like most bilateral donors, is experiencing, leaving PASEC's funding to one country is risky and perhaps unfair. The role that PASEC plays in global education monitoring and research makes a strong case for multilaterals such as the World Bank or GPE to lead the fundraising challenge.

DISCLAIMER & PERMISSIONS

CGD's publications reflect the views of the authors, drawing on prior research and experience in their areas of expertise. CGD is a nonpartisan, independent organization and does not take institutional positions. You may use and disseminate CGD's publications under these conditions.


Thumbnail image by: Chalabala/ Adobe Stock