This paper looks at the research and policy literatures on system level leadership and management in low and middle-income countries at the central and subnational level. It describes two paradigms – one drawn from principal-agent theory and the other from sociological approaches to organizational learning and system improvement – which have led to contrasting prescriptions for effective system leadership. Highlighting recent research and experience in three case countries (Brazil, Ghana, Rwanda) the paper reviews the limitations of some widely adopted policy interventions and diagnostic frameworks developed to improve education system leadership. It also explores new efforts to measure and support management and leadership capacity, especially at the middle tier of educational systems. The paper argues that much of what is understood about improving educational systems in LMIC contexts still focuses on getting policies, plans, technical systems, and procedures right – neglecting the potential impact of strengthening leadership capacity. This contrasts with the three decades of research on school and system improvement in OECD countries, which describes expansive roles for leaders and leadership at every level to improve the implementation of system-wide reforms.
