This paper explores the connection between mass refugee return and the emergence of violence in post-conflict societies, by investigating the impact of mass refugee return to Burundi after the country’s 1993–2005 civil war. The author also considers how the experience of return migration affected individuals’ future behaviour, in the context of the 2015 electoral crisis in Burundi.
Burundi has experienced three major waves of forced migration. The first occurred in 1972, when a Tutsi-led government orchestrated a genocide against Hutu civilians, forcing an estimated 217,000 Burundians to flee to Tanzania. The second wave began in 1993 with the assassination of President Melchior Ndadaye and escalated during the civil war, sending hundreds of thousands of Burundians, primarily Hutu, to Tanzanian refugee camps. The third wave, in 2015, saw over 413,000 Burundians flee to neighbouring countries, including Tanzania, Rwanda, the Democratic Republic of Congo, and Uganda, amid a renewed political crisis. Starting around 2002, with the civil war ending, hundreds of thousands of Burundians living abroad returned to Burundi.
The author undertook in-depth ethnographic research in Burundi and Tanzania over 9 months from 2014 to 2016, including 258 semi-structured interviews with Burundian civilians, international humanitarian organization staff, Tanzanian and Burundian government officials, and Tanzanian villagers. The research was conducted in three primary areas: villages in Makamba Province in Burundi, the Nyarugusu Refugee camp in Tanzania, and Ilagala village, a small farming town in Tanzania’s Kigoma region.
Main findings:
- Refugee return to Burundi after the 1993-2005 civil war created new divisions between returnees and those who stayed. These labels were further subdivided by the era of initial flight (1972 or the 1990s) and, in some cases, by country of asylum. Burundians attributed certain characteristics to each group, such as language spoken, style of dress, and even how women carried their babies. These perceived differences fueled narratives about which group had better claims of national legitimacy. Social and political tension in Burundi after the civil war manifested in segregation between returnees and non-migrants, with accusations of discrimination and exclusion from both sides. These divisions were often intertwined with, but distinct from, existing ethnic cleavages.
- Tensions between returnees and those who stayed were most evident in conflicts over land. Because land in the interwar years had been occupied, expropriated, bought, or sold, both returnees and those who stayed often claimed the same land as rightfully theirs. The threat of losing land bred distrust, conflict, and violence between returning populations and non-migrants.
- Informal and formal governance practices exacerbated conflicts between returnees and those who stayed over land. Informal land inheritance practices in Burundi, based on patrilineal inheritance, created a fertile ground for conflict between male family members with different migration histories. This was further exacerbated by formal land governance practices, specifically the Commission Nationale des Terres et Autres Biens (CNTB), which initially promoted land sharing but later shifted to a policy of full restitution for returnees, particularly those from 1972. This policy, implemented retroactively, worsened relations between returnees and non-migrants, fueled violence against the CNTB itself, and ultimately reinforced the separation between the two groups.
- The 2015 electoral crisis in Burundi saw a mass exodus of refugees, but prior experiences of return migration significantly influenced who fled and when. Those who had previously returned and faced land conflicts, particularly the 1993 caseload returnees, were more likely to flee early, often citing fear of violence from those who stayed behind. They saw the national crisis as an opportunity to escape the local tensions they had already experienced. Later arrivals, often first-time refugees or those who had successfully reclaimed land, were more likely to stay until the national crisis reached a breaking point.
The author concludes that processes of out-migration and return can aggravate old rivalries and create new divisions between populations who were displaced across borders and those who remained in-country. New migration-related group identities are more likely to harden and become violent when post-conflict institutions intentionally or unintentionally favor individuals based on where they were physically located during wartime.