A long-established and widely used approach to management in government has been the transmission of information up a hierarchy, and centralized decision-making and oversight; colloquially known as ‘command and control’. This paper examines accountability in such a system implemented at scale in Punjab, Pakistan. Using random variation in the intensity of accountability of the scheme, the paper shows that the corresponding de facto punishments had a negligible impact on school or student outcomes. It uses detailed data on the education production function to show that this fundamental component of command-and-control approaches does not induce bureaucratic action towards improvements in government performance.