Deforestation and human development in the Brazilian tropical dry forest

The relationship between tropical deforestation and human development is unclear and contested. We evaluated the boom-bust hypothesis across agricultural frontiers in the Brazilian Caatinga dry forest, one of the largest dry forests in the world and home to 28 million people. We used panel data (1991, 2000 and 2010), and cross-sectional data (2010) from 1207 municipalities to assess how development indicators are linked to deforestation through a quasi-experimental approach. Our main finding is that deforestation in the Caatinga is associated with a boom-bust development pattern or at least to a stagnation in development in highly deforested municipalities. Municipalities at the advanced stage of deforestation (<33 % of forest cover remaining) in 1991 generally had higher development indicators than the initial stage (>66 % remaining), but differences between these groups disappeared by 2010. Intermediate stage municipalities (33–66 % remaining) consistently outperformed initial and/or advanced stage municipalities in four out of six development indicators (longevity, monetary income, extreme poverty prevalence, and child mortality), indicating a temporary ‘boom’ during frontier advance, followed by a stagnation. Evidence of a boom-bust was supported by cross-sectional analysis of 2010 data using propensity score weighting and a spatial autoregressive model. Overall, our findings contribute to on-going debate and strengthen the boom-bust hypothesis. By implication, the mere consumption of natural resources is inadequate to ensure sustained development progress. Achieving sustainability in Brazil's agricultural frontiers necessitates more than apolitical technical solutions; it requires active engagement by the state, non-state institutions, and society as a whole to address the country's deep-seated inequalities and imbalanced power dynamics. • Novel test of the boom-bust hypothesis in an old agricultural frontier. • Deforestation~development relationship assessed using panel and cross-sectional data. • Panel analysis showed weaker poverty reduction in highly deforested municipalities. • Propensity score weighting showed a quadratic deforestation~development relationship. • Spatial models confirmed better development outcomes at intermediate deforestation.