Monitoring for impact: Evaluating 20 years of soil and water conservation in southern Mali

A soil and water conservation (SWC) project has been going on in southern Mali since 1986. Donor support was gradually withdrawn between 1998 and 2002, but no final evaluation was undertaken to learn lessons from this long-term and large-scale experience. The objective of this present research was to find out how to evaluate impact, what the impact in Mali has been, and which recommendations could be made for monitoring and evaluation in SWC projects. A reconstructed logical framework made it possible to find out what was needed for the impact evaluation, what was available from project monitoring and external monitoring, and what additional data and analyses were required. Missing baseline data were substituted by reconstructed baselines and virtual time series. Between 1988 and 2002, agriculture has expanded and intensified, but crop yields have declined and nutrient balances are still negative. Further intensification is needed to halt and reverse the yield decline. The cause-effect chain between project activities and impact showed that the SWC extension approach was effectively increasing farmer adoption of SWC measures. Farmer adoption steadily increased, spread to neighbouring villages and continued after project withdrawal. Erosion control measures (live fences, stone rows, grass strips and check dams) reduced erosion with 50-70% and improved crop yields with 5-12%. Current annual farmer benefits of increased cotton production largely outweigh the annual SWC extension costs during the project. SWC projects are recommended to complete the logical framework and monitor accordingly, and to collaborate with external monitoring for a more efficient evaluation of impact. Achieving impact may take longer than the project life span. Therefore, project activities should be embedded in a long-term national programme. It also implies that to assess impact after a short project period requires proxy impact indicators that reflect a continuing change, rather than an end-status.