Dynamic effort choice in high school: costs and benefits of an academic track

I investigate high school tracking policies using a dynamic discrete choice model of study programs and unobserved effort. I estimate the model using data from Flanders (Belgium) and perform an ex ante evaluation of a policy that encourages underperforming students to switch to less academically oriented programs. This reduces grade retention by a third and dropout by 11%. Although it decreases college enrollment, the decrease in college graduation is small and insignificant. I also show that modeling effort is important; otherwise, smaller decreases in grade retention and dropout and larger decreases in college enrollment and graduation would be predicted.