Although research indicates that individual students benefit when they enroll in early algebra classes, evaluations of broad-based algebra acceleration often report negative effects. Using a regression discontinuity design, we replicate the positive effects of eighth-grade algebra placement on student achievement found in prior studies. We then demonstrate that eighth-grade algebra placement positively affects the achievement level of students' classmates, as well as the experience and value added of students' math teachers. Peer composition effects may help to reconcile the apparent contradiction between the positive effects of algebra on individual students and negative effects of algebra acceleration since these estimated effects are large enough to plausibly explain the majority of the effects of eighth-grade algebra on student test scores.