Recent economic expansion in Peru has apparently encouraged higher education demand. Averages, however, mask strong differences between expenditure groups and a potential anti-poor bias. We build a simple theoretical model that predicts a positive relation between education demand, households’ endowments and wage premia. Access to credit reduces endowments’ impact and also encourages education demand if its return is sufficiently large. These predictions are validated after estimating enrolment probabilities using a panelof households for the period 2004-2006. Results help explain the antipoor bias if we combine the relatively small expenditure growth experienced by the first quintiles and the reduction exhibited by the wage premium. Our results also highlight the importance of short run credit constraints and basic education quality as determinants of postsecondary enrolment. This s...