Econometrics for environmental policy evaluation: Analysis of the ”Priority Municipalities” policy in the Brazilian Amazon.
Environmental Public Policy, Econometrics, Staggered Difference-in-Difference, Propensity Score Matching.
The present work intends to understand if the Brazilian policy of blacklisting municipalities, in the Amazon region, was efficient when created in 2008, and how it evolved over the years, until 2019. To do so, we applied econometrics methods. For the first question, difference-indifferences with propensity score matching was our choice, with deforestation increase divided by municipal area as our dependent variable. We found that when controlling for covariates and using fixed effects, the policy was efficient to decrease deforestation. Precisely, When listed, municipalities decreased deforestation per km2 by around 0.003. For the following question, we utilized difference-in-differences again, but with staggered treatment. We employed
two contemporary staggered diff-in-diff variations. In general, we found again that
the policy was efficient in 2008 and 2009, but its effect decreased over time. For robustness, we also answered both questions with different dependent variables, precisely: normalized and log-transformed deforestation increase. Next, we tried to explore the reasons behind the policy change. Our findings suggest that political alliances and the diminishing of resources focused on environmental conservation impacted the policy. Finally, we state some flaws and limitations of the paper, pointing opportunities for future studies.