Association for Public Policy Analysis & Management

An extensive literature exists on the achievement effects of private school voucher programs. The vital question of who applies for and participates in voucher programs, however, is woefully understudied. The unresolved questions of what types of students are attracted to private school choice programs and what types tend to participate and persist in them when given the chance are central to a complete evaluation of the desirability of such initiatives. We leverage unusually rich data on the economic, demographic, and social characteristics of disadvantaged families that participate in a means-tested private school choice program located in a 20-square kilometer, highly urbanized slum area known as Shahdara, in East Delhi, India. Shahdara is an excellent context in which to study the potential segregating or equalizing effects of private school vouchers because the powerful forces of social caste and family income tend rigidly to stratify educational access in that community. Relative to the general population of school-age children in Shahdara, we describe the characteristics of the 1,618 applicants to the Ensure Access to Better Learning Experiences (ENABLE) program. Our data include key measures of relative student advantage such as parental literacy and employment status, average monthly household income, social caste, possession of a gas stove, family toilet access, whether students have a vision, hearing, or movement disability, and baseline performance on tests of Math, Hindi, and English proficiency. We use probit models to determine the student and family background factors that predict which applicants actually use an ENABLE voucher when offered one by lottery. Furthermore, we use Cox Hazard regression models to analyze the flow of students who attrit from the sample altogether or fail to comply with their treatment assignment over the course of our five-year panel data study. We document characteristics of the 37.6 percent of voucher lottery winners who leave the program during our study, as well as the 37.4 percent of lottery losers who ultimately enrolled in a private school, despite not winning a voucher to fund their attendance. By highlighting patterns in the economic and social characteristics of students in this context, our findings will shed light on implementation challenges and stratification concerns of private school choice, especially in the vital context of developing countries.