This dissertation develops and structurally estimates an equilibrium model of the Chilean school system and uses the model to assess the effect of teacher wage and accreditation policies on student achievement. In the model, potential teachers choose between teaching in a public school, teaching in a private school, working in the non-teaching sector and not working; parents choose whether to enroll their children in public or private schools; and private schools set tuition and teacher wages. The equilibrium of the model determines the distribution of student cognitive achievement. The estimation is based on a unique dataset that combines three rich data sources from Chile: student test score data, household survey data and teacher survey data. Chile is an ideal environment for studying school choice, because it has thirty years of experience with a school voucher system and a large private school sector. I use a two-step estimation approach that addresses the issue of potentially multiple equilibria. The estimated model fits the data well. Policy experiments show that an increase in the wages of public school teachers accompanied by minimum teacher competency requirements would increase average student test scores without a large increase in government costs. A larger increase in test scores could be achieved by making the wages of public school teachers more directly tied to their skills than in the current system. Under the existing voucher plan, both policies could be partially financed by the increased voucher revenues that result from an increased demand for public education.
Adviser: Kenneth I. Wolpin. Thesis (Ph.D. in Economics) -- University of Pennsylvania, 2012. Includes bibliographical references.