Population, multi-scale processes, and land use transitions in the Ecuadorian Amazon

William Pan, Johns Hopkins Bloomberg School of Public Health
David L. Carr, University of California, Santa Barbara

A growing literature has identified agricultural expansion by small farmers in the tropics as an important cause of global forest clearing, yet institutional factors affecting small farmers' decisions have rarely been taken into account in empirical analyses. We examine the influence of context on household land use/cover change (LUCC) decisions and estimate a multilevel model of LUCC for forest, pasture, perennials, annuals and fallow in Ecuador’s northern Amazon from 1990 to 1999. We draw upon four strands of theoretical research to test relationships and interpret the findings: the peasant pioneer cycle, the hollow frontier, forest transition theory, and household lifecycle theory. We also examine differences within the study region, viz., sub-regional differences, and use descriptive and empirical model results to show that LUCC is jointly influenced by natural resource endowments, population changes, and community context, which may be moderated by location via road access. Off-farm employment and finca subdivision, which are affected by changes in population, are also strong predictors of LUCC. Community context plays a statistically significant role in land use change, accounting for 10% of total variation in our models. From a theoretical perspective, therefore, models that fail to take into account the effects of context, particularly in rapidly changing frontier environments, are missing a key component in explaining LUCC, which may result in incorrect statistical inferences. From a policy perspective, the results of this research suggest that policies in the Amazon must be targeted not just toward farmers, but toward structural factors at the community level.