Why the Regreening Africa Panel cannot support Welfare Effects of Farmer Managed Natural Regeneration (FMNR) in Sub Saharan West Africa

Details

Supervisors

Van den Broeck, Goedele ; Wüpper, David

Faculty

Faculté des bioingénieurs

Degree label

Master [120] en sciences agronomiques et industries du vivant, à finalité spécialisée: Agricultural, Food and Environmental Policy Analysis

Abstract

enThis thesis set out to evaluate the socioeconomic impact of Farmer Managed Natural Regeneration (FMNR) and tree-based regreening activities during the first phase of the Regreening Africa initiative (2017–2023). The evaluation was based on a two-wave household panel comprising 4,505 households in Ghana, Mali, Niger and Senegal. The evaluation’s central finding is that the data cannot provide evidence of a positive relationship between FMNR and the study’s welfare measures, using country-specific tetrachoric-PCA wealth indices, the International Wealth Index (IWI), the Food Insecurity Experience Scale (FIES) and the Minimum Dietary Diversity for Women (MDD-W). The phase-in design does not deliver the experimental contrast it was built on. Lead-farmer learning reached control and treatment villages at similar rates, and treatment assignment no longer predicts realized regreening in any of the four countries, which collapses the first stage and renders the intention-to-treat (ITT) estimates null across all outcomes and countries. In response, the analysis turns to a descriptive dose-response specification relating household welfare to the initiative’s Regreening Action Index (RAI) through two-way fixed effects. The descriptive results are modest to zero and primarily concentrated on the baseline slope rather than on any change during the program period. A direct test of the program’s lead farmer mechanism associates benefits predominantly with the participating leadership households themselves, despite visible peer uptake of regreening. The primary contribution of this study is methodological. It transparently documents how a phase-in design erodes in the field and what a ladder of identification strategies can and cannot recover. It concludes that the welfare returns to regreening at scale remain unidentified in these data and calls for evaluation designs that protect the counterfactual.