We estimate the effects of accountability-driven school closure in New York City on students who attended middle schools that were closed at the time of closure and students who would have likely attended a closed middle school had it remained open. We find that students who would have entered the closed school had it not closed attended schools that perform better on standardized exams and have higher value-added measures than did the closed schools. While we find that closure did not have any measurable effect on the average student in this group, we do find that high-performing students in this group attended higher-performing schools and experienced economically-meaningful and statistically-significant improvements in their 6th-, 7th-, and 8th-grade math test scores. We find that these benefits persisted for several cohorts after closure. We also find that closure adversely affected students, low-performing students in particular, who were attending schools that closed. For policymakers, our results highlight a key trade-off of closing a low-performing school: future cohorts of relatively high-performing students may benefit from closure while low-performing students in schools designated for closure are adversely affected.