Centralization and Experimentation in the Implementation of a National Monitoring and Evaluation System

The general principles and technical procedures of monitoring and evaluation are well documented, but far less attention has been directed at the equally important process of field implementation in development settings. To identify general principles of implementation that may prove useful in building any monitoring and evaluation system, the authors focus on Malawi's experience over more than 2 decades in creating a national system for agriculture. Drawing on intensive interviews with managers and evaluators in Malawi's agricultural agency and on a range of evaluation documents, this study suggests that monitoring and evaluations systems generate higher quality information if they follow two principles of implementation. Information quality refers to the extent to which assembled data are reliable, valid, and focused on dimensions of interest to program managers and policymakers. The first implementation principle is to combine both centralized and decentralized elements in the organization of the monitoring and evaluation system. Central oversight helps to ensure uniform standards, and localized data collection ensures more sensitive solutions to local field problems. The second implementation principle is to build on an iterative process of field testing, refinement, and retesting of major procedures and indicators. A prolonged trial-and-error approach allows a monitoring and evaluation system to incorporate direct field evidence into procedural definitions, thereby incrementally sharpening data reliability and policy relevance.