Peer Effects and Returns to Scale in the Education Production Function

Marc Luppino, Federal Trade Commission (FTC)

This paper is one of the first studies to investigate the potential peer effects induced by grade retention in elementary school. Specifically, I examine whether the sign and magnitude of grade repeater induced peer effects vary by gender, socio-economic background, and school location. I also test whether these peer effects vary non-linearly as one increases the number of grade repeaters assigned to a student's classroom. The findings of this study suggest that grade repeater induced peer effects are both heterogeneous and non-linear. In the majority of specifications that I estimate, I reject that such peer effects exhibit constant or decreasing returns to scale. Instead, I find that clustering of grade repeaters into one classroom generates large negative effects on higher performing students (e.g. female and non-poor students). Specifications that assume linear peer effects fail to pick up such effects.