In 2002/03, the Charlotte-Mecklenburg Schools in North Carolina initiated a broad program of accelerating entry into algebra coursework. The proportion of moderately-performing students taking algebra in 8th grade increased from half to 85%, then reverted to baseline levels, in the span of just five years. We use this policy-induced variation to infer the impact of accelerated entry into algebra on student performance in math courses as students progress through high school. Students affected by the acceleration initiative scored significantly lower on end-of-course tests in Algebra I, and were either significantly less likely or no more likely to pass standard follow-up courses, Geometry and Algebra II, on a college-preparatory timetable. Although we also find that the district assigned teachers with weaker qualifications to Algebra I classes in the first year of the acceleration, this reduction in teacher quality accounts for only a small portion of the overall effect.