Welfare Impact of Hosting Refugees in Ethiopia

Ashenafi Belayneh Ayenew

World Bank Policy Research Working Paper Series, No. 9613 (2021)



This paper examines the impact of refugee inflows on the welfare of host households in Ethiopia. The author examines the impact on consumption expenditure per capita and wealth of host households, and investigates three potential mechanisms for these effects, namely: (a) the labor market; (b) societal cooperation; and (c) prices. The author focuses on the period from the end of 2009 until the end of 2014, when the number of refugees hosted in Ethiopia increased from 125,910 to 660,987 people.

The author uses a difference-in-difference methodology that exploits large spatial differences in the intensity of the refugee inflows in villages over time, and an instrumental variable approach to address the possibility that refugees choose their settlement location based on the relative conditions in each location. The analysis is based on refugee data from UNHCR as well as household data from the Ethiopian Socioeconomic Survey (ESS) conducted by the Central Statistical Agency of Ethiopia (CSA) and the World Bank.

Main findings:

  • Refugee inflows negatively affected host consumption expenditure per capita, increasing the probability that host households fall into consumption poverty. A one percent increase in refugee intensity increases the probability of falling into consumption poverty by about 18 percentage points. The consumption effect occurs in rural areas with no effect in urban areas.
  • Refugee inflows negatively affected food consumption expenditure but not non-food consumption expenditure. Decomposing household consumption expenditure per capita into food, education, and other non-food components, reveals that refugee inflows negatively affect food consumption expenditure, while other components of household expenditure remained unchanged.
  • Refugee inflows had no statistically significant effect on household wealth. There are no differences in the effects on wealth and wealth poverty results between urban and rural areas. The finding that refugee inflows affect consumption expenditure and consumption poverty but don’t appear to affect wealth or wealth poverty might be because the wealth metrics are less sensitive to short-term shocks.
  • Consumption effects were driven by the displacement of individual hosts from salaried employment and increases in the prices of agricultural inputs (seeds and fertilizer). No evidence is found for other potential mechanisms investigated by the author including changes in self-employment in non-farm businesses, societal cooperation as measured by participation in customary labor-sharing arrangements, and prices of food items.

In his conclusion, the author identifies several development interventions that could potentially offset the welfare loss of hosting refugees such as: (a) cash transfer programs that include participation in temporary (casual) labor as one of the targeting parameters; (b) investments in skills and entrepreneurship training to assist rural hosts to engage more in self-employment in non-farm businesses or take up salaried permanent employment; and (c) provision of subsidized agricultural inputs (seeds and fertilizer) to refugee-hosting farm households.