There is a cautionary lesson from a new working paper examining Houston's bold talent initiative. In many ways, it yielded the results intended by retaining a higher number of great teachers and exiting weak ones. It also led to a few unintended and unwelcome outcomes.

Houston, like other districts, recognizes the importance of having a workforce of great teachers – its new evaluation system, readopted in 2010, aimed at hiring great teachers, helping current teachers become great, and exiting ineffective teachers.

Julie Berry Cullen, Corry Koedel, and Eric Parsons look at the effects of one side of this multi-faceted policy – whether Houston is effectively exiting ineffective teachers.

The big finding is that ineffective teachers have left the district at a higher rate since this evaluation system was implemented. The even bigger finding is that they were especially likely to leave if they were working in lower-performing schools.

But another group also left those lowest-performing schools: effective teachers. They mostly left to take jobs at other less challenging Houston schools. The researchers suspect that this exodus occurred because, prior to implementation of its talent initiative, teacher evaluations did little to distinguish the strongest teachers. Because these evaluations clearly identify which teachers are in the top tier, talented teachers had more leverage with which to seek jobs in more desirable schools.

So what should policymakers learn from this? A good evaluation system is a good thing and can help nudge teachers out the door who aren't effective and don't show signs of improvement. But the district needs to give serious attention to how to keep great teachers in high-needs schools where they are needed the most.