This paper examines the long-run urbanizing effect of refugee camps in Tanzania. Between October 1993 and April 1994, Tanzania experienced a mass influx of more than 800,000 refugees from Burundi and Rwanda, who were settled in camps located in the remote regions of Kagera and Kigoma. The population of these regions increased by more than a third and this was followed by a proliferation of humanitarian aid agencies and expatriate workers.
Exploiting the sudden and unanticipated nature of the refugee influx and their settlement locations, the authors employ a spatial difference-in-differences strategy that compares settlement patterns and spatial economic data both during and after the period of the camps’ operation. The authors compare localities in proximity of the camps with localities further away.
The analysis draws on several sources of data for the period from 1985 to 2015: (1) information on the geographic location of refugee camps and the dates of their operation from Maystadt and Verwimp (2014), Zhou (2014) and UNHCR; (2) high-resolution satellite data from the World Settlement Footprint Evolution (WSF) dataset showing the evolution of built-up areas each year at a high degree of spatial resolution; (3) spatial economic data on output from the Global Gridded Geographically Based Economic Data (G-Econ) dataset; and (4) spatial data on employment and consumption from the 2008, 2012 and 2015 Tanzania Living Standards Measurement Survey (LSMS).
Main findings:
- Refugee camps in Tanzania had a modest but persistent urbanizing effect on built- up areas within a 100 km distance of the camps, although the effect reduces with distance from camp. The built-up area within 100 km of a refugee camp increases on average by about 6 percent from the average pre-camp level of built-up area during the lifetime of the camp, and this increase largely persists after the camp has closed. The impact increases as the radius from the refugee camp is reduced. The effect is also greater the longer the camp is in existence and persists for years well beyond the closure of the camp.
- The urbanizing effect of refugee camps is higher in rural localities. Camps do not appear to have a significant urbanizing effect on localities that were urban before the camps’ existence, although most camps were situated well away from existing urban areas.
- Local economic activity grew faster in the region where camps were located than elsewhere. Areas with camps had significantly faster growth in regional output during the time camps were operational.
- Camp closure appears to shift employment away from non-agricultural activities in the long term. The likelihood of being in wage or salaried employment is lower around camps that closed long ago while the likelihood of engaging in unpaid or self- employment in agriculture is higher. However, there is no statistically significant difference in consumption around areas with camps that closed earlier compared to those that closed later.
The results demonstrate that refugee camps in Tanzania had a small yet enduring urbanizing effect on rural areas. Although regional output in areas with camps experienced faster growth during the operation of the camps, the increased likelihood of engaging in unpaid work and self-employment in agriculture long after camps closed suggests that the activity induced by camp was primarily in non-tradeable goods and services. The authors conclude that refugee camps in Tanzania resulted in “urbanization without growth” and did not bring about any structural transformation in the local economy.