In this paper we evaluate the impact of need-based grants on university drop-out rates in the first year of enrollment, using student-level administrative data from all Italian universities in the period 2003-2013. We exploit the fact that not all eligible students receive financial aid due to limited resources to generate a treatment and a control group. Using this partition, we estimate the average treatment effect, i.e. the average effect on low income students, controlling for a set of observable characteristics by running regressions on blocks defined on the propensity score. Results point towards a sizeable effect of grants in reducing dropping out from higher education: around one third of these students would have left university in the first year in absence of the grants. This evidence is robust to a variety of specifications and sample selection criteria.