Since the No Child Left Behind Act was enacted, grading and labeling of schools as low performing have been increasingly used as means to incentivize failing schools to raise student achievement. Using statewide high school data from Michigan, our regression discontinuity analyses show that the bottom 5% of schools identified as persistently lowest achieving (PLA), which is publicly announced and has consequential accountability, increased their student performance in writing and to a lesser extent in mathematics and social studies. The PLA effect in writing is quite robust, based on various sensitivity analyses. We find no improvement in student achievement for the bottom 6% to 20% of schools labeled as "watch list" that received no actual penalties and little public attention. Our findings suggest that schools respond differently to varying forms of labeling as low performing, depending on the levels of accountability pressure and social stigmatization.