Selection into agroforestry adoption and crop production value among smallholder farmers in Tanzania
| Authors: | Sonia Bruck, Olli-Pekka Kuusela |
| Year: | 2026 |
| Type: | Scientific Journal |
| Station: | Southern Research Station |
| DOI: | https://doi.org/10.1007/s10457-026-01689-z |
| Source: | Agroforestry Systems |
Abstract
Agroforestry is widely promoted in smallholder farming systems as a means of improving soil conditions, strengthening resilience to rainfall variability, and supporting rural livelihoods. However, households that adopt agroforestry may differ systematically from those that do not. We examine selection into agroforestry adoption and its relationship with crop production value per adult household member among smallholder farmers in Tabora Region, Tanzania. The practices promoted in the study area include planting the nitrogen-fixing tree Gliricidia sepium, pigeonpea–cassava intercropping, or a combination of the two. To account for potentially endogenous selection into adoption, we estimate a Poisson control-function model using household contact with extension agents as the exclusion restriction, complemented by instrumental-variable estimates and a sensitivity analysis. The control-function estimates indicate negative selection into agroforestry adoption: unobserved factors associated with a higher propensity to adopt are negatively related to crop production value in the outcome equation. Under the maintained identifying assumptions, the estimated average treatment effect on adopters is approximately USD 152 per adult household member, while the corresponding instrumental-variable estimate for households whose adoption decisions are affected by extension contact is approximately USD 412. Estimators that account only for observable differences produce estimates that are statistically indistinguishable from zero. The results suggest that agroforestry may be particularly beneficial for smallholders with relatively low counterfactual crop production value.