In HarvestChoice our primary goals are to understand the contemporary spatial patterns and time trends of smallholder productivity and, against that baseline, to evaluate the potential physical and economic consequences of a range of productivity-focused investments and policies. While we are concerned with increasing productivity at the plot and farm scales, we typically undertake our assessments of change at a landscape, country and even regional scale by estimating likely farm-scale productivity change for each 10km cell of our underlying databases. 

Conceptually, productivity growth can arise from interventions that increase output holding inputs constant, or that reduce inputs holding outputs constant. In practice, most interventions affect both outputs and inputs. For example, improved seeds can deliver higher yields but are typically more expensive than landrace seeds and require more input of nutrients, water and labor to realize their potential. Such improved seeds will increase productivity for a given farming enterprise in a given location if, after accounting is complete, the farmgate value of increased output has exceeded the increased farmgate costs of inputs.

In our economic analysis we typically express productivity in terms of a metric of cost per unit of output (e.g., US$/kg); by comparing this metric with and without a specific intervention we can express productivity change as a change in absolute cost per unit of output or, more often, as a percentage relative to the without-intervention unit cost. For many aspects of our work we assess only changes in partial productivity metrics such as crop yield (kg/ha) or livestock product yield (kg/animal). Additional levels of complexity present themselves in thinking about productivity changes over time, e.g. as weather and prices vary, but also in the context of “sustainable productivity growth” where the quantity and quality of unmeasured and uncosted inputs, typically natural resource inputs such as water, soil nutrients and pollinators, also need to be accounted for. One advantage of using both biophysical and economic productivity evaluation tools in tandem is that biophysical models are better able to provide a physical accounting of multiple inputs and outputs (e.g. crop yield models account for total biomass production, including both grain and crop residue, water use (from both rainfall and irrigation), soil nutrient uptake and changes in soil fertility status and, by implication, labor and energy input requirements).

Topic Contacts