To date, there have been very few efforts to evaluate the impact of refugee camps on host landscapes. Yet many stakeholders believe refugee camp openings lead to forest loss in nearby areas. I use data on camp locations and years of operation and secondary geospatial data to produce a high-resolution panel dataset and estimate the impact of camps on the landscape while accounting for endogenous camp site selection. I define a difference-in-difference specification with tile fixed effects that exploits variation in camp openings and tile proximity to camps. I find that within the rainforest biome, camps are associated with a small reduction in extensive margin forest loss (land clearing) and a small increase in intensive margin forest loss. In the grasslands biome, camps lead to small increases in forest loss at the intensive margin but have little impact on the extensive margin. I find that camps in both biomes stimulate agricultural expansion, and I theorize that the camps may lead to less extensive margin forest loss in the rainforest biome due to indirect camp impacts on the opportunity cost of harvesting forest products for sale.