The Bank supported decentralization efforts in most study countries, often with good results (Honduras, India), but in some cases there was ambiguity in what the different levels covered, non-alignment of administrative and financial features of decentralization, and under-training of local government staff for their new tasks. The extent to which some forms of decentralization might be contributing to increased school system inequities has not been adequately assessed.

School-level management activities were relatively more effective and so were efforts to empower communities in school improvement efforts—at least for physical improvements but rarely with respect to improved teaching and learning. Various approaches to more equitable teacher distribution have been tried with mixed results, the most promising being the recruitment of local (often untrained) youth, as long as provisions can be made for their professional development, career paths and job security.

Project monitoring and evaluation has typically tracked outputs, as opposed to outcomes or impacts, but this appears to be changing. Bank support has helped governments to establish management information and student assessment systems and research capacity, but their quality and the degree to which they have been used for improving policy and practice have been limited.

Key lessons:

• Sector management and governance might have been better dealt with had there been better institutional and political assessments at the outset.
• Weak management incentives at all levels can be a constraint, especially to the improvement of educational quality. There are often more rewards for increasing the number of schools than for hard tasks of redistributing teachers, implementing a new curriculum, or doing effective monitoring and evaluation.
• Few Bank-supported country programs directly addressed teacher recruitment and performance incentives; particularly lacking are performance incentives related to student learning outcomes.