SPATIAL EFFECTS OF LOCAL GOVERNANCE INSTITUTIONS ON ILLEGAL DEFORESTATION IN DEVELOPING COUNTRIES

Deforestation worldwide is a major concern. In developing countries, it is a merciless and devastating reality. My thesis addresses how local governance institutions' strength influences this phenomenon, focusing on the Colombian Andes. The theoretical analysis examines spatial patterns of illegal deforestation when enforcement is costly, and costs rise with distance from governmental centers. Those spatial patterns depend on the interaction between transportation costs incurred by farmers growing crops on deforested land and enforcement costs incurred by government officials conducting on-the-ground monitoring of deforestation. Areas closer to governmental centers can be monitored effectively and are thus less subject to illegal deforestation. Illegal deforestation is, therefore, more likely in areas where monitoring costs are high, but farmers' transportation costs are not. The calibration exercise then shows, that in this context, patches of deforestation might arise within the forest, causing unwanted forest fragmentation. Based on these results I study empirically, first, if the effect of access difficulty on deforestation may be non-monotonic in accessibility, causing forest fragmentation; and second, if this fragmentation is more likely to occur when enforcement is more costly. I approach this question in two manners: (1) using a cubic function of access difficulty interacted with measures of enforcement capacity and (2) non-parametrically using indicators for discrete ranges of access difficulty, again interacted with measures of enforcement capacity. I construct for this purpose a panel data set for the Colombian Andes from a variety of sources. Data on deforestation comes from satellite imaging at a 30mx30m resolution in two periods (2000-2005) and (2005-2010), this data was matched with biophysical variables such as, altitude, slope, precipitation, soil type, and roads using geographical information systems (GIS), as well as with socioeconomic variables which vary by municipality and time. The regressions show a significant non-monotonic effect of access difficulty on deforestation. The evidence shows that deforestation probability first decreases with access difficulty, and it then increases in remoter places. This evidences forest fragmentation as one moves away from roads. Moreover, this pattern is affected by the fiscal performance index (a proxy for enforcement capacity) of the municipalities, showing that municipalities with lower enforcement capacity have a higher probability to present illegal deforestation at remote places. This research adds to the deforestation literature, by studying the spatial reach of governance capacity and how it affects deforestation patterns. The findings highlight the importance of taking enforcement and monitoring costs as well as their spatial variation into account when designing land-use policies and defining the institutional arrangements, funding and monitoring processes to implement them