This paper outlines a methodological approach for use by FAO to collect, analyze and monitor agricultural policy indicators (API) for developing countries. The aim is to establish a consistent and comparable set of policy indicators, allowing analysts to examine whether agriculture in particular countries is being stimulated or retarded by the set of policies employed. The API should also be useful in the context of quantitative models of policy impacts and market projections. Ideally, the indicators should function as building blocks for the more comprehensive policy measurement. This paper is presented in four parts. Part 1 gives an overview of the rationale for monitoring, the relationship between policy monitoring, quantitative modelling and evaluation, and the basic menu of choices of indicators. Part 2 provides a brief review of the studies that have calculated the extent of agricultural support in developed countries since 1973 and of the comparable exercises that have focused on developing countries, mainly published since 1990. In light of the apparent "success" or "failure" of the studies mentioned above, and of the authors' experience with many of them, Part 3 identifies several areas that in our opinion deserve special attention in the planning of the FAO-sponsored studies. We attempt to provide insights into methodological, computational and logistical problems that might be faced by FAO. Part 4 suggests a methodological approach for FAO, elaborating on the suggested coverage, definitions, classification criteria and methods of calculation.