Theme 38 of 43
Basic Education, Literacy and International Comparison
Brazil's richest quartile of students scores 425 in PISA mathematics — still below the OECD average of 465.
Brazil has taken part in PISA since 2000, has run a national literacy assessment since 2023, and records in its school census the picture of special-education students enrolled in mainstream schools. Read separately, these registries produce predictable headlines about poor performance. Read together, they reveal something more uncomfortable: Brazil's shortfall is not merely a matter of internal inequality that could be fixed by lifting the bottom, because not even the top of the national socioeconomic distribution reaches the average of wealthy countries. The gap between Brazil's richest student and the average OECD student is wider than the gap between Brazil's richest and poorest.
Not even Brazil's top reaches the OECD average
In the most recent PISA round, Brazilian students average 378 points in mathematics and 412 in reading, against 465 and 480 across OECD countries — a shortfall of 87 points in mathematics, roughly three years of schooling by the OECD's own yardstick. Splitting Brazilian students into quartiles by PISA's socioeconomic index, the poorest quartile scores 347 and the richest 425. Brazil's internal spread (78 points) is narrower than what still separates its top from the OECD mean: even the Brazilian student from the most resourced family, with more books at home and better-educated parents, performs below the median student of a wealthy country.
| Group | Mathematics | Gap to OECD |
|---|---|---|
| Brazil — poorest quartile | 347 | -118 |
| Brazil — 2nd quartile | 364 | -101 |
| Brazil — 3rd quartile | 380 | -85 |
| Brazil — richest quartile | 425 | -40 |
| OECD average | 465 | — |
Brazil's socioeconomic elite scores 425 in mathematics; the median OECD student scores 465. The national ceiling sits below the international average.
Literacy rose 10 points in two years — and still leaves 1 in 3 behind
The national literacy assessment, applied at the end of second grade, shows the public network moving from 55.9% of children literate at the appropriate age in 2023 to 66.0% in 2025 — a 10.1 percentage-point rise across two cycles, a rare pace for educational indicators. The figure is simultaneously the theme's best and worst news: best because it shows that age-appropriate literacy policy produces measurable, fast effects; worst because even at that pace, a third of Brazilian children still finish second grade unable to read adequately.
| Year | Literate at appropriate age | Change |
|---|---|---|
| 2023 | 55.9% | — |
| 2024 | 59.2% | +3.3 p.p. |
| 2025 | 66.0% | +6.8 p.p. |
Ten percentage points in two years is real, accelerating progress — but 34% of public-network children still leave second grade without being literate.
In special education, failure concentrates where support should be strongest
Among special-education students enrolled in mainstream schools, the failure rate is 10.7% in the early years of primary school — nearly three times that of secondary school (3.8%) and triple the later primary years (3.4%). The pattern inverts the expected logic: failure normally rises with content complexity, but here it peaks precisely at the literacy stage, when specialised pedagogical support should be most intense. Dropout, by contrast, rises with age and reaches 4.5% in secondary school.
| Stage | Failure | Dropout |
|---|---|---|
| Primary — early years | 10.7% | 1.4% |
| Primary — later years | 3.4% | 2.5% |
| Secondary — regular | 3.8% | 4.5% |
Failure in special education is three times higher during literacy than in secondary school — the bottleneck is at the entrance, not in content complexity.
Fewer than half of municipal teachers teach what they studied
The hypothesis raised above — that Brazil's shortfall has a component of average system quality, not only inequality — can be tested directly. INEP's teacher training adequacy indicator measures how many teachers hold a degree in the subject they teach. In the later primary years, when teaching becomes subject-based, only 56.7% of teachers have adequate training, and in municipal networks that figure falls to 46.7%: fewer than half. In adult primary education it collapses to 31.6%. The data supports the hypothesis: a system where half the teachers teach material they did not study produces a low ceiling across every socioeconomic stratum, which explains why not even the richest quartile reaches the OECD average. The most counterintuitive finding, however, is that private schools have worse teacher training than public ones at several stages — 57.9% against 73.5% in early primary, and 46.9% against 64.0% in early childhood education.
| Stage | Public network | Municipal network | Private network |
|---|---|---|---|
| Primary — early years | 73.5% | 71.2% | 57.9% |
| Primary — later years | 55.7% | 46.7% | 61.4% |
| Early childhood | 64.0% | 64.0% | 46.9% |
| Adult education — primary | 32.2% | 27.0% | 22.3% |
| Adult education — secondary | 58.2% | 47.3% | 44.5% |
In municipal networks in the later primary years, 53% of teachers teach a subject they hold no degree in — and in adult education the share exceeds two thirds.
Nearly half of state teachers are on temporary contracts
The teacher training adequacy seen above gains an explanation when employment status is examined. In state networks, 48.5% of teachers work under temporary contracts — practically tied with the 49.8% who are tenured. Municipal networks have 34.4% temporary staff, and federal ones just 12.9%. Temporary hiring is the instrument networks use to plug vacancies at the start of the school year, and it selects for immediate availability rather than subject qualification: it is the mechanism by which a teacher without a mathematics degree ends up teaching mathematics. The two series fit together — the network with the most temporary staff is also the one where fewer than half of teachers in the later primary years hold a degree in what they teach.
| Network | Tenured | Temporary | Private-law / Outsourced |
|---|---|---|---|
| Federal | 86.4% | 12.9% | 0.8% |
| Municipal | 63.2% | 34.4% | 2.4% |
| State | 49.8% | 48.5% | 1.6% |
State networks have almost as many temporary teachers as tenured ones — and that is where subject-adequate training is scarcest.
Powerful cross-references
- PISA × Class: Brazil's richest quartile (425) falls below the OECD average (465).
- Brazil × OECD: an 87-point mathematics gap, roughly three years of schooling.
- Literacy × Pace: +10.1 p.p. in two years in the public network, from 55.9% to 66.0%.
- Special education × Stage: 10.7% failure during literacy against 3.8% in secondary school.
- Dropout × Age: rises from 1.4% in early primary to 4.5% in secondary.
- Internal × external inequality: the Brazil-OECD gap exceeds the rich-poor gap within Brazil.
- Teacher training × Network: only 46.7% of municipal later-primary teachers hold a degree in their subject.
- Contract × Training: 48.5% of state teachers are temporary, the network with the worst training adequacy.
Explanatory hypotheses
That Brazil's socioeconomic top fails to reach the OECD average suggests the problem is not merely distributive but systemic: were it only inequality, lifting the bottom would suffice; since the ceiling is also low, a component of average system quality — teacher training, effective instructional time, curriculum — affects every stratum. The acceleration in literacy, conversely, indicates that focused, well-measured interventions produce fast results, reinforcing the hypothesis that the shortfall is one of public policy, not student capacity. The concentration of failure in special education during literacy points to a scarcity of specialised educational support precisely at the stage where it is decisive.
Public policy implications
If not even the richest quartile reaches the international average, policies focused exclusively on equity will not close the gap — floor and ceiling must rise together, which runs through teacher training and career structure. The recent literacy pace suggests that public annual targets with census-level measurement work and should extend to other stages. And failure in special education during early primary indicates a need to expand specialised educational support at the literacy stage, where provision is currently thinnest and the cost of failure most enduring.