In 2011, the U.S. Department of Education granted states the opportunity to apply for waivers from the core requirements of the No Child Left Behind Act. In exchange, many states implemented systems of differentiated accountability that included a focus on schools with the largest achievement gaps between subgroups of students. We use administrative data from Michigan in a series of regression-discontinuity analyses to study the effects of these school reforms on schools and students. We find some evidence that targeting schools for such reforms led to small, short-run reductions in the within-school math achievement gap. However, these reductions were driven by stagnant performance of lower-achieving students alongside declines in the performance of their higher-achieving peers. These findings serve as a cautionary tale for the capacity of the accountability provisions embedded in the recent reauthorization of No Child Left Behind, the Every Student Succeeds Act, to meaningfully improve student and school outcomes.