This paper uses data from a single School of a UK university to estimate the impact of a switch from end-of-semester to end-of-year final exams on student performance. The identification strategy exploits the fact that while the timing of final exams changed, the timing of mid-term exams remained the same. Estimates are based on a difference-in-differences methodology that compares final and mid-term exam scores after the switch, with final and mid-term exam scores before the switch. The empirical findings suggest that the shift of final exams to the end of the academic year had a negative effect on student achievement. Many changes in higher education are often done on the basis of financial, administrative or ideological considerations, underestimating the magnitude of their effect on student learning and student performance.