When Education Expenditure Matters: An Empirical Analysis of Recent International Data

We analyze the diminishing correlations between education expenditure and learning outcomes to address two fundamental questions: Do education systems with different levels of education spending have different student achievement levels? If so, at what amount of education spending does the relationship between increased expenditure and student achievement differ? Using data from a large group of countries around the world, we find that the association between education spending and student performance in mathematics is statistically significant among systems that spend below a threshold of US$8,000 per student annually (in purchasing power parity). Controlling for average income (GDP) per capita and income inequality, our estimates suggest that education spending is associated with increased student performance only among systems that spend below this threshold, with mean student achievement approximately 14 points higher on the PISA scale for every additional US$1,000 spent.