Justin Kernoghan / TROCAIRE

Agricultural research and poverty impact assessments

At the ISPC-10 meeting on Wednesday, Doug Gollin (SPIA Chair) gave a presentation on the challenges of measuring poverty impacts of agricultural research. Some of his key points (I am paraphrasing)

ISPC 10 Copenhagen D.Gollin

  • Poverty reduction from agricultural research can occur through many different pathways. Not all poverty impacts of ag. research will transmit through price and quantity (productivity) effects. These effects may be important over the long term, but are not the only one. Hence, critical to disentangle these pathways. For instance, in case of the SPIA beans study (see SPIA project social impacts: poverty and hunger) , an ISPC Council Member asked how much of the income effect in question is from nitrogen fixation - it might be small and even harder to pick this up in a convincing way.
  • Because the contribution of ag. research to poverty reduction may be small (overall), IA methods need to be nuanced enough to pick up on this - one needs to recognize that it is hard to do this, and hard to do this in a convincing manner.
  • In the context of the above, the conclusion here is not that measuring poverty impacts of ag. research is unimportant or impossible. His argument is that there are specific things we can (and should) do to make this feasible. A diversity of methods is required. Donors need to understand the complexity, limited scope, and the highly context specific nature of this challenge. Not every ag. research innovation is going to generate poverty impacts - research systems often tend to have a few successes that are worth donors' entire spending.
  • With respect to methods, he emphasized that a diversity of research designs and methods, ranging all the way from qualitative to experimental, are required. For instance, there are cases where a small scale, time bound, extremely well-designed experiment may not, by itself, pick up poverty impacts. Where poverty impacts are possible through migration related pathways, descriptive studies could well be useful. The idea is to combine a set of studies at different scales to make a convincing case.
  • Relatedly, it is important that the poverty reduction agenda is on the minds of ag. research scientists. Measurement of poverty impacts begins at the stage of research design - when the program of research is being designed, in general. It is too important to leave to just economists. There are opportunities to do some really interesting research. Regular data collection on some key variables (e.g. adoption of improved varieties or NRM practices, real wages of unskilled labourers, food prices) would enable this.
  • Not every milestone or causal link in the pathway needs to be monitored from the beginning to the end. The idea being that IA studies may examine only a specific aspect of the pathway and generate evidence, but can (through the appropriate choice of research design/method) do this convincingly.
  • While it is not necessary to collect data from all beneficiaries/non-beneficiaries who are a part of the IA, and researchers can target a sub-set of questions at a sub-set of sample thereby lowering costs, one need to be aware that good poverty impact assessments require extensive resources, are costly and involve trade-offs.

The presentation (PDF) draws on a draft of a forthcoming paper that will briefly touch upon the theoretical evidence on links between ag research and poverty reduction (pathways, channels, and disaggregated effects); methodological approaches (incld. suitability to the IA context) drawing on a literature review of completed/on-going studies; and, include thoughts on appropriate research design and methods. Keep an eye on the SPIA site and ISPC's new page at http://ispc.cgiar.org/.