This paper investigates the determinants of grades achieved in maths by first-year students in economics. We use individual administrative data from 1993 to 2005 to fit an educational production function. Our main findings suggest that good secondary school achievements and the type of school attended are significantly associated with maths grades. "Ceteris paribus", females typically do better than males. Since students can postpone the exam or repeat it when they fail, we also analyze the determinants of the elapsed time to pass the exam using a survival analysis. Modeling simultaneously maths grades and the hazard of passing the exam, we find that the overall hazard rate of passing the exam is higher for those students who get the higher grades. The longer the students wait to take the exam, the less likely they are to obtain high grades. (Contains 16 notes, 7 tables and 1 figure.)