Education systems: models, performance and inequalities

The United States spends USD 14,603 of public money on each primary and secondary student every year, well above the rich-country average of USD 12,438. Its 15-year-olds scored 465 points in mathematics in the latest international PISA assessment. Estonia, which spends USD 10,303 per student, almost 30% less, scored 510.

That is a 45-point gap, the equivalent of more than two years of schooling on the scale the OECD uses. And money does not explain it. This is the starting point for everything that follows: an education system is not a machine that converts budgets into skills. It is a set of choices, often invisible, about who pays, who decides, who sorts, and at what age. Those choices produce radically different outcomes at comparable cost.

Why spending explains almost nothing

The intuition is powerful: more money, better education. International data contradict it, though not in the way most people assume.

A relationship between spending and results does exist, but it flattens sharply. Below a certain threshold of cumulative spending per student, every additional dollar clearly improves learning: classrooms get built, teachers get paid, textbooks get printed. Above that threshold, the correlation collapses. Wealthy countries all sit on the flat part of the curve, where what separates them is no longer how much they spend but how they spend it.

OECD figures show the spread. In primary and secondary education, average public spending is around USD 12,400 per student per year. It exceeds USD 21,000 in South Korea, Switzerland and Luxembourg. It stays below USD 4,000 in Mexico and Türkiye. In tertiary education the gap is even starker: over USD 25,000 per student in Luxembourg, Norway and Switzerland, under USD 5,000 in Chile and Mexico.