The environmental and agricultural effects of electing candidates connected to farming interests

Data
2020-05-15
Orientador(res)
Ferman, Bruno
Novaes, Lucas Martins
Título da Revista
ISSN da Revista
Título de Volume
Resumo

Politically connected individuals obtain favors in the form of public jobs and public credit, but do they also receive preferential treatment when it comes to environmental regulation? Also, what are the effects of electing rural producers? This paper employs a close-election RD design using data from Brazil, a country where environmental protection has global relevance, to estimate the environmental and agricultural effects of electing municipal candidates connected to farming interests. Via previously unused data on official property registries, we locate and georeference farms whose owners finance local electoral campaigns, and evaluate the variation of deforestation and agriculture within spatial buffers centered on those farms, after municipal elections. We also evaluate effects on municipal-level outcomes. Overall, we don’t find significant effects in electing farmers as either mayors or councillors on municipal-level planted and deforested areas. However, within small samples consisting only of large farmers, we do find significant and sizable effects, on both outcomes. These municipal-level results seem consistent with large farmers prioritizing - or being more able to prioritize - agriculture over environment, whilst in office. Our farm-level results suggest that electing councillors financed by farmers leads to significant increases (decreases) in farmed (deforested) areas within campaign donors’ farms, with effects on farmed areas being far more sizable. Electing mayors financed by farmers leads to significant increases in deforested areas within campaign donors’ farms, with insignificant effects on farmed areas. Although all farm-level variations plausibly come at the expense of within-farm vegetation, they seem particularly consistent with systematic manipulation of local environmental regulation only in the case of mayors financed by rural producers. This is coherent with the fact that the margin of environmental regulation that mayors control, enforcement, is far more flexible than the margin that councillors control, environmental legislation. Our farm-level results are considerably robust to buffer radius choice, and both our farm-level and municipal-level results come from samples for which our identification assumptions seem valid. Additionally, the vast majority of our results are robust to variations in the specification of our RD’s, and pass innovative tests that assess inference quality in finite samples. The totality of our results benchmark innovative recommendations on how to reduce deforestation rates in Brazil, and raise interest on whether environmental regulation should occur at local levels.


Descrição
Área do Conhecimento