Determinants of Deforestation and the Economics of Protection: An Application to Mexico

We estimate the probability of deforestation at the plot level for Chiapas and Oaxaca, two poor Mexican states with high levels of biodiversity. Results highlight the importance of physio-geographic endowments as well as policy variables and allow to test explicitly for aggregation bias. They also suggest that, if combined with information on the biodiversity value of specific plots, such models could be of great relevance for policy by allowing to identify the ex ante risk of deforestation. Copyright 2002, Oxford University Press.