Refugee mobilities in East Africa: understanding secondary movements

Alexander Betts, Naohiko Omata, Jade Siu and Olivier Sterck

Journal of Ethnic and Migration Studies, Volume 49 (2023), Issue 11 


This article examines the mobility aspirations of refugees in Kenya, Uganda, and Ethiopia, and includes an in-depth analysis of the mobility patterns of refugees in Kenya. The research challenges common assumptions about refugee mobility, that: (1) most refugee secondary movements (the movement of refugees from the first country in which they arrive) are South-North; (2) refugee movements are predominantly irregular; (3) aspirations to move translate into actual movements; and (4) refugees who remain in regions of origin are largely immobile. 

The authors study a range of refugee movement, including intra-urban/camp, inter-urban/camp, intra-regional migration, and inter-regional (i.e., international) migration. They consider three phases of mobility-related decision making: hope, expectation, and actual migration, and whether these are conditioned by indicators of capacity such as income levels.  

The analysis on hope and expectation draws on quantitative data collected between 2016 and 2017 from refugees and host communities living in Kenya (Kakuma refugee camp and Nairobi), Uganda (Nakivale refugee settlement and Kampala), and Ethiopia (Dollo Ado refugee camps and Addis Ababa). The sample includes 8,970 refugees and 7,638 members of the host populations, with data collected in each context being representative of the main refugee populations and host communities. Additional panel data was collected two to three years later in Kenya to track the movements of all individuals who were interviewed in the baseline surveys. The quantitative analysis was complemented with qualitative research, including focus group discussions and semi-structured interviews. 

Main findings: 

  • A large majority (more than 60 percent) of refugees expected to leave their host countries for a third country (i.e., not their host or home country). The proportion of refugees expecting to move to a third country is particularly large for Somali refugees in Kakuma (93 percent), Nakivale (65 percent), and Kampala (73 percent) due to shrinking business opportunities, police harassment, decreasing food rations, worsening security, and poor social services such as education and medical facilities. 17 percent of refugees expected to remain where they were living due to limited resources or networks necessary for onward migration. 
  • While financial capacity is not a strong determinant of expectations to migrate internationally, having networks abroad do predict expectations to relocate. Living standards do not predict expectations to migrate internationally, suggesting that financial constraints are not an important determinant of aspirations. Network variables (such as receiving remittances and having family networks in high-income countries) predict expectations to move, especially internationally. Refugees in camp-like contexts who speak English are also more likely to expect to migrate, especially internationally. 
  • Refugees are highly mobile. In Kenya, 23 percent of camp refugees and 37 percent of urban refugees change their primary residence each year. These rates of residency change are much higher than in host communities. This suggests that refugees are not sedentary in their first countries of asylum, despite the common assumption of immobility within research and policy circles. There isn’t any evidence that having a network abroad or receiving remittances affects refugees’ decision to move. 
  • Most refugee movements are internal, including camp-to-urban movement, inter-urban movement, and intra-urban movement. In Kenya, 6 percent of camp refugees and 24 percent of urban refugees moved internally, the majority within the camp or to a few cities. The most common type of internal movement is intra-urban movement, i.e., local movements within Nairobi. The main reasons for internal movements include the search for new work opportunities and real estate dynamics such as renovation, demolition, and rent increases. The relationship between internal migration and income levels appears to be U-shaped—most refugees in the middle of the wealth distribution appear to be less likely to move internally compared to the very poor or rich.  
  • While a large majority hope to migrate internationally, and a smaller majority expect to migrate internationally, actual international migration by refugees in first countries of asylum is rare. For example, 62 percent of refugees in Kakuma expect to migrate internationally, but only 14 percent actually do. The relationship between international migration and living standards appears to have an inverted U-shaped relationship, i.e., a positive relationship for most refugees, except for the very poor or rich. Additionally, there is some evidence that human capital partly predicts international migration, as individuals with vocational training are more likely to migrate internationally, and individuals with mental health issues are less likely to migrate internationally. 
  • Most international migration is to refugees’ home countries or to other countries in the Global South. Somali refugees in Kakuma took advantage of UNHCR’s repatriation program during the study period, citing several reasons for their decision to return including: (a) little hope of resettlement from Kakuma; (b) insecurity in the camp, (c) lack of access to higher education; or (d) repatriation cash benefits. Some South Sudanese refugees returned spontaneously without UNHCR assistance, but back-and-forth movements and family-splitting strategies are common among South Sudanese refugees in Uganda.  
  • Few refugees migrate to the Global North. Refugees have two main routes to migrate to the Global North: migration to Canada, the US, and Australia (through UNHCR’s resettlement program) and migration to Europe (through irregular channels). Only 1.2 percent of camp refugees and 4.9 percent of urban refugees move, regularly or irregularly, to rich countries each year. Most of these movements are regular, facilitated through UNHCR resettlement programs; 1.1 percent of camp refugees and 3.7 percent of urban refugees benefit from resettlement opportunities each year.  

These findings challenge the dominant belief that refugee mobility is reducible to irregular secondary movements from poor to rich countries, specifically: (1) secondary movements are predominantly South-South movements; (2) of the tiny minority of refugees who engage in South-North movement, most move via ‘regular’ channels to third countries (UNHCR resettlement programs or education visas); and (3) while many refugees may aspire to move to rich countries, few actually do; (4) while a large majority hope to migrate internationally, and a smaller majority expect to migrate internationally, actual international migration by refugees in first countries of asylum is rare; (4) internal migration within the country of first asylum is common.