There is mounting research evidence on both the importance of teacher quality in student learning and on the difficulty of identifying who actually is (can be) a good teacher. The participation of current teachers in the selection of new teachers may ease this informational problem if the former have superior information about teacher quality. However, a simple principal-agent model can show that asymmetric information between administrators and teachers may lead to an agency problem with rent extraction if teachers have an objective function with different arguments from teacher quality. In this study, I use a recent policy reform in Mexico to evaluate the effect on student outcomes of receiving a teacher hired through a standardized test versus one hired in a discretionary process with strong involvement from the teachers' union. My difference-in-differences results indicate that the allocation of test-hired teachers reduces exam cheating in junior-secondary schools and -in environments where cheating is rare- increases observed student achievement. I also find that joint committees of state officials and union representatives allocate the discretionary-hired teachers to schools in more "desirable" localities. These results suggest the existence of an agency problem with potential rent-extraction that the use of a hiring rule can mitigate.

ACCESS PAPER

ACCESS LATEST VERSION OF THIS PAPER