Refugee camps and deforestation in Sub-Saharan Africa

Colette Salemi

Journal of Development Economics, Volume 152 (2021)


This article examines the effect of refugee camp openings on deforestation in sub-Saharan Africa. Previous research has suggested that refugee camps lead to increased deforestation in surrounding areas due to land clearing for cultivation and demand for fuelwood.

The author employs a spatially disaggregated research design using geographic ‘tiles’ of 1.1 square kilometers. About half of the analyzed tiles were not within 20 km of any refugee camps at any point in the study period, and these tiles serve as a comparison group in the econometric analysis.

The analysis draws on a number of data sources including: (1) the African Refugee Dataset (ARD) of georeferenced locations and years of operations of planned refugee camps between 1999 and 2016; (b) Global Forest Change (GFC) data on extensive margin deforestation for the years 2000 to 2012, which captures transitions from non-zero forest cover to zero forest cover in a given year for 30 meter grid cells; and (c) NASA’s Global Forest Cover Change (GFCC) data on intensive margin deforestation, which captures gradual reductions in an areas percent forest cover at five year intervals from 2000 to 2015. The analysis is disaggregated by the type of biome, i.e., grasslands or rainforests.

Key results:

  • There are statistically significant but very small reductions to extensive margin forest losses in response to camp openings, suggesting an increase in vegetation density. The analysis detects modest reductions in extensive margin forest loss in rainforest biomes but not grassland biomes. Each year of camp exposure is associated with a small fraction of one 30-meter grid cell within the tile not transitioning to zero forest cover. This avoided extensive margin forest loss occurs between 1–20 km from a camp. The author suggests two possible explanations for this result: (a) camps may reduce the returns to harvesting forest products, either because they impact the agricultural wage (and by extension, the opportunity costs of harvesting from forests); or (b) the security presence around camps discourages illegal logging activities.
  • There are very small declines in forest cover along the intensive margin in response to camp exposure. On average, camp exposure leads to annual forest cover losses amounting to less than one percentage point, with slightly larger effects in rainforest tiles. In grassland tiles 1–5 km from a camp, one year of camp exposure results in a reduction of tree canopy cover by 0.13–0.15 percentage points, while in rainforest tiles 1–5 km from a camp, one year of camp exposure is associated with a reduction in tile forest cover of 0.21–0.23 percentage points. While it is possible that the activities of refugees lead to some of these modest intensive margin losses 1-5 km from a camp, losses of forest cover 15 km or further from a camp suggest that the activities of other groups may also change in response to camp openings in a manner that influences forest cover.


Overall, the results suggest that refugee camps have a very small negative impact on intensive margin forest cover within distances of 1–20 km. These results challenge the preconception that refugees cause extensive deforestation in camp areas.