Portfolio schools have become increasingly prevalent in urban school districts across the United States. These schools exemplify a model where school autonomy is integrated with district oversight, promoting a controlled form of autonomy. This study examines the impact of portfolio schools in Indianapolis on student achievement. To ensure balanced covariates between treatment and control groups, the study utilizes machine learning-generated inverse probability weighting and analyzes the ten years of data using student-fixed effect and event study approaches. The findings indicate that controlled school autonomy in the form of portfolio schools does not significantly affect student achievement overall. However, there is a negative effect for turnaround schools—a subset of portfolio schools—which serve disproportionately disadvantaged students. Additionally, portfolio charter schools, where disadvantaged students are overrepresented, perform worse than other charter schools. These results suggest that portfolio schools worsen student achievement in high-inequality settings. Potential mechanisms are discussed.