Publication: Anti-Environmental Signals and Deforestation: Evidence from Brazil
Magalhães de Oliveira, G., E. Cisneros, J. Sellare, and J. Börner (2026): Anti-Environmental Signals and Deforestation: Evidence from Brazil. The Journal of Politics, forthcoming, https://doi.org/10.1086/744065.
Our paper with Gustavo Magalhães de Oliveira, Jorge Sellare, and Jan Börner is now published online in The Journal of Politics. An earlier working paper version circulated under the title “Political signals and deforestation” and remains available on SSRN.
The paper asks whether political speech alone, absent any change in formal policy, can move environmental outcomes. We use Brazil’s 2019 transition to the Bolsonaro government and combine monthly anti-environmental signals from government Twitter accounts with pre-2019 municipal Twitter attention to forest issues in a shift-share design across 578 Legal Amazon municipalities. Municipalities more exposed to these signals cleared more forest, and the effect is concentrated where agricultural returns are high and where enforcement had historically been strong. Exposure does not predict local fines or embargoes, which points to changed enforcement expectations rather than changed enforcement itself.
Abstract
Political signals can influence behavior through information that alters enforcement expectations. We provide causal evidence from Brazil’s 2019 political transition, asking whether anti-environmental signals from the Bolsonaro government increased Amazon deforestation without formal regulatory change. We combine monthly government tweets with pre-2019 municipal Twitter attention to forest issues in a shift-share design across 578 Legal Amazon municipalities. A one-standard-deviation increase in exposure raises monthly deforestation by 6.8%. Effects are largest where agricultural returns are high and conservation enforcement was historically strong. Exposure does not predict municipal environmental fines or embargoes and effects do not vary with political allegiance, highlighting the role of perceived enforcement risk rather than immediate local enforcement or partisan interpretation. Political speech can weaken the effectiveness of environmental institutions even without changes in formal policy.
