Allocative decisions concerning public sector agricultural research appear to be driven by both supply and, politically mediated, demand forces. In-sample Granger and Modified Sim's tests, along with post-sample predictive tests, suggest that simultaneity issues should not be ignored when modelling the research expenditure-output relationship. The results also provide strong evidence that the impact of research expenditures on agricultural output persists for at least 30 years. These lags are substantially longer than those commonly used for agricultural research to date. The lagged effect of output on research appears much shorter, at something less than 10 years.