Student loan defaults and performance management

Determining which colleges, public or private, have default rates that are "too high" is, as this NYT article suggests, a tricky problem. The University of Michigan has a very low default rate on student loans but Michigan is taking in high ability, high motivation people who will have little trouble finding a job even if they do not finish their degrees. These are very different than the sort of people who are attending the upper end of the proprietary school market (e.g. DeVry and U of Phoenix) and even more different from the lower end of that market (e.g. truck driving schools that advertise at 3 AM on ESPN).

What would be useful here is a regulation that embodied adjustments for the characteristics of the students admitted. Then colleges could be rewarded or punished based not on their absolute performance, which reflects both the population they serve and how they serve it, but instead on their performance relative to what one would expect given the characteristics of their students. Such an adjustment mechanism has been used for federal job training programs in the past.

My friends Carolyn Heinrich and Burt Barnow have written a whole paper on performance standards adjustment.

Also, it is important to note that, particularly in a program that is designed to help those who are otherwise credit constrained make investments in their human capital, the optimal default rate is not zero. An interesting paper idea would be to try and sort out the optimal default rate given evidence from the literature about the relevant elasticities for different types of students, the value of the human capital that might be obtained and so on.

Hat tip: Brian McCall's blog