This paper models the university-to-work transition in a stochastic dynamic environment, where students may study and work simultaneously. The structural model is estimated using a unique panel data set with exogenous variation from changing threshold levels for maximum student grants. Estimates reveal that uniformly increasing student aid increases enrollment time. Policy simulations show that because of the non-linear effect of student working hours on academic achievement, however, tilting student aid towards those who work fewer hours increases graduation rates by 5 percentage points, but is ineffective in shortening time-to-graduation. A combination of tilting student aid and improving student abilities earlier in the education production process both increases graduation rates and lowers time-to-graduation. Including incentives into the student aid package with merit aid or timely graduation bonuses also tend to be effective policy devises to amend academic outcomes.