Supplement to ‘ Deforestation in the Amazon : A Unified Framework for Estimation and Policy Analysis

Eduardo Souza-Rodrigues
Abstract:In this supplemental document, I provide a robustness analysis for the estimated regressions presented in the main paper (Section 2), and a detailed description of how the data set is constructed (Section 3). Section 2 is divided into five subsections. In Subsection 2.1, I discuss the estimated coeffi cients on the covariates in the land use quantile regressions. In Subsection 2.2, I present the estimated results when I exclude the proxies for monitoring efforts: the number of fines issued and the distance to the nearest IBAMA offi ce (the Brazilian Environmental Protection Agency). Consistent with the discussion in the main paper (that historically there had not been much enforcement in the Amazon rainforest), the inclusion and the exclusion of these variables do not affect substantially the coeffi cients on costs to ports, as well as on other regressors. Policy implications also are robust to the different specifications. In Subsection 2.3, I investigate how spatial dependence can affect the results. Coeffi cients on spatially lagged covariates are statistically significant, but results are not sensitive to the choice of weighting matrix nor to the cutoff distance. Subsection 2.4 drops the logit assumption in the paper and estimates a semiparametric model. The results establish the suitability of the logit model once there are no significant differences between the logit and the semiparametric models. Finally, in Subsection 2.5, I discuss the use of different productivity indices to calculate the demand for deforestation. I describe briefly the underlying economic model associated with the different indices, I present the corresponding estimated demands for deforestation, and I provide evidence that local yields do not respond to changes in transportation costs. ∗University of Toronto, Department of Economics. E-mail: e.souzarodrigues@utoronto.ca. Financial support from Charles V. Hickox Fellowship, Arthur K. Watson Fund, Yale University and Kernan Brothers Environmental Fellowship is gratefully acknowledged. All errors are mine.
Economics,Environmental Science
What problem does this paper attempt to address?