Distributional Effects of Public Investments in Mozambique

In 2013, Mozambique’s government proposed a five-year agricultural investment plan that raised spending and targeted new investment areas, including fertilizer subsidies and agricultural research and extension. Evaluating sector-wide strategies is difficult since they comprise multiple interventions with spillovers. Numerous ex post studies evaluate specific investments, but the time-series or spatial data needed to estimate sector-wide returns is often lacking. Moreover, when new interventions are planned, there is no historical evidence on which to base analysis. To overcome these limitations, we develop a mixed-methods approach to evaluating the distributional effects of Mozambique’s investment plan – one that combines ex post analysis of specific investments with ex ante analysis of investment portfolios. We econometrically estimate investment impacts on farmer productivity, and then use these results to calibrate investment functions in a spatially-disaggregated CGE model. This permits experimentation with different levels and compositions of investments to evaluate how outcomes are improved. Econometric results (using propensity score matching) indicate farmers who use irrigation, receive extension advice, or use chemical fertilizers, have higher productivity. CGE analysis finds that the plan’s benefit-cost ratio and poverty impacts justify implementation. However, returns are much larger if resources are reallocated to research and extension and away from the current emphasis on irrigation. Greater spending on fertilizer subsidies also improves outcomes, but to a lesser extent. These findings are robust to a range of assumptions about investment costs and efficiency. We conclude that research and extension should be afforded a greater role in Mozambican investment plans. Our mixed-methods approach also greatly enhances the usefulness of ex post evaluation studies for sector-wide planning.