Population and deforestation in Costa Rica.

This paper addresses a central debate in research and policy on population and environment, namely the extent to which rapid population growth is associated with the massive deforestation currently underway in the tropics. Although temporal and spatial associations strongly suggest a connection between population growth and deforestation (Preston, 1994), some research indicates that the problem is more complex as it involves non-demographic mechanisms resulting from credit and capital market failures, lack of suitable mediating institutions securing property rights, wretched poverty, uneven land distribution, consumption patterns in developed countries, greedy multinational companies, ignorance and bad management by colonists of frontier land, and so forth (Gillis and Repetto, 1988; Bilsborrow and Ogendo, 1992; Myers, 1984; Palloni, 1994). This paper is an exploratory analysis of highly disaggregated data from Costa Rica-atropical country that in the 1960s and 1970s experienced one of the highest rates of deforestation and population growth in the world. It addresses the methodological problem of linking people and population pressure to land cover, a problem that arises from the fact that people usually do not live in the forests that will be cleared. To establish the population-land linkage the paper relies on a multidisciplinary geographic information system (GIS) platform, which was developed for this study with georeferenced data from two population censuses and a series of land cover maps. The key analyses in the paper use multivariate logistic regression to model the net impact of population growth on the 1973-83 probability of deforestation in about 31,000 parcels of 750 meters per side, which were covered with forest at the beginning of the period. Since conventional logistic models fail to account for sources of unmeasured covariates that could cause autocorrelation, we present in an appendix estimates which attenuate the impact of spatially relevant unmeasured covariates. (from Author's introduction)