Essays on environmental and development Economics

• Autor(a) principal: Barros, Pedro Henrique Batista de
• Data de Publicação: 2024
• Tipo de documento: Tese
• Idioma: eng
• Título da fonte: Biblioteca Digital de Teses e Dissertações da USP
• Texto Completo: https://www.teses.usp.br/teses/disponiveis/12/12140/tde-14052024-170133/
• topic: Amazon; Amazônia; Brasil; Brazil; Causal Random Forest; Crescimento econômico; Deforestation; Desmatamento; Economic growth; Instituições; Institutions; Oil palm; Óleo de palma; Random Forest Causal; Remote sensing; Sensoriamento remoto

Paper 1: The causal impacts of local institutions on tropical deforestation are still little explored in the literature because they involve endogenous mechanisms that act, to a large extent, through socioeconomic and political channels that hinder identification. To fill such a gap, this paper contributes to the literature by exploring exogenous variations in local institutions to identify their effects on forest cover in Brazil, an ecologically and economically important country. To achieve this, we exploit exogenous geographical and historical variations to construct instruments for current institutions, assuming that initial conditions found by the country\'s settlers led to institutional designs that conditioned its subsequent development, explaining current institutions\' differences. Our main results show that the local institutional quality change has positive statistically significant heterogeneous causal effects on deforestation in Brazil, even after several robustness checks. To further explore this evidence, we used a Causal Random Forest algorithm to estimate individual treatment effects without ad hoc hypotheses, which further supported significant heterogeneous positive causal effects. This empirical evidence is important because demonstrates that public policies that aim to improve local institutional quality must adequately consider the potential side effects of deforestation. Paper 2: The existence of a trade-off in the relationship between economic growth and environmental quality is still an open debate, especially when considering the deforestation of tropical forests. Part of the literature states that the negative environmental impacts are focused on the early stages of development, when institutional quality is low, up to a turning point in which the economy moves towards sustainable development. However, many critics have supported that this is only a snapshot of a complex process, requiring additional empirical assessments to shed light on this controversy. In this context, this paper aims to contribute to the debate with a new approach to model the relationship between economic growth, capture by income per capita, and deforestation in the Amazon by controlling for institutional changes, market conditions, dynamic interactions, leakages, and spatial spillovers. After several robustness checks, our results supported the hypothesis that higher economic well-being is associated with lower deforestation rates in the Amazon and this relationship seems to be mediated by structural transformations and market access. Therefore, empirical evidence suggests that higher economic well-being could be reconciled with forest preservation in the Amazon. Paper 3: This paper maps palm oil plantations in the Eastern Amazon, the largest producer in Brazil, in 2014 and 2020, using machine learning algorithms, to estimate its causal effects on the trade-off between deforestation and economic activity. To achieve this goal, we combined optical spectral bands from Landsat-8, radar backscatter values from Sentinel-1, vegetation, and texture indices, and a linear spectral mixing model. The Random Forest algorithm presented the best classification with an overall accuracy of 94.53% and 95.53% for 2014 and 2020, respectively. Then, from a land use and land cover transition analysis, we identified an expansion of oil palm from 1,074 km² to 1,849 km²; around 156.88 km² (20.24%) occurred directly over vegetation cover. To overcome potentially complex endogenous mechanisms that hinder a causal interpretation for prior estimates, we propose to instrumentalize palm oil expansion using the maximum agro-climatically attainable palm oil yield from the Global Agro-Ecological Zoning (GAEZ). Our main results support that palm oil expansion in the Eastern Amazon has a statistically significant and positive causal effect on deforestation and a negative impact on economic activity in the non-agricultural sectors. In other words, palm oil expansion is increasing the environmental impacts of the region while creating centripetal forces that reduce performance in the industrial and service sectors, raising concerns about the social and environmental sustainability of this crop.