A procedure is developed and applied for predicting ex ante impacts of agricultural research on aggregate poverty, using as an example the poverty-reducing impact of peanut research in Uganda. Market-level information on economic surplus changes is combined with a procedure for allocating income...

This paper suggests practical methods for assessing policy research programs, both ex post and ex ante. Measuring the benefits of policy research is difficult: the path of causation between research and policy change is nearly always uncertain; multiple factors influence any particular policy...