Chang-Tai Hsieh and Miguel Urguiola find evidence that in Chile school vouchers caused schools to compete for the best students rather than compete to deliver better education.

Although statistically insignificant, the point estimates suggest that, if anything, test scores experienced a relative decline in communities where the private sector made greater inroads.

The results of this study would appear to me to suggest that vouchers need to be “progressive,” in that higher vouchers should be given to needier students. Progressive vouchers would change the incentives so that schools would compete for needier students, which in turn might create more incentive to compete on quality.

It could be that the majority of early experiments with vouchers are likely to prove disappointing. However, if we learn from successful and unsuccessful experiments and adopt best practices, my guess is that ultimately education systems that adopt vouchers will be superior.

UPDATE: More from Tyler Cowen and from Brad DeLong.

For Discussion. In the Chilean system, poorly-performing public schools were not shut down. How could this explain the authors’ results that areas where private schools made larger inroads were not areas in which overall education outcomes improved?