Twenty years later the refugees are back

Published 3 June 2015 / By Carlos Vargas-Silva

Back to Articles


In two recent papers we explored the long-term labour market consequences for Tanzanians of hosting refugees from Burundi and Rwanda. These two countries experienced violent conflicts in the mid-1990s and hundreds of thousands of its residents sought refuge in neighbouring Tanzania. Our analysis uses data from the Kagera Health and Development Survey (KHDS), collected in Kagera, the northwestern corner of Tanzania. The last refugee camp in Kagera closed in 2008. The longitudinal data was collected in 1991 and 2010 (i.e. before and after the “end” of the refugee shock).

In a shorter paper published in the American Economic Review we showed that Tanzanians who were employees and were more affected by the refugee shock had a higher probability of subsequently being in professional occupations and being part of a pensions program. In a longer paper forthcoming in the Journal of Economic Geography we find that greater exposure to the refugee shock resulted in Tanzanians having a higher likelihood of working in household plantations or caring for household livestock and a lower likelihood of working outside the household as employees. The latter effect was particularly strong for Tanzanians doing casual work before the shock. This coincides with anecdotal evidence of refugees concentrating in casual waged work and competing directly with Tanzanians for those jobs.

In both studies we find that the implications of refugee camps outlast the existence of these camps and that these impacts persist even after the refugees have returned home. One frequent question/challenge which we encountered while presenting preliminary versions of these papers in academic seminars was about the reasons for the permanent nature of these impacts. The logic of the argument was as follows: refugees had left the region and the natives should go back to their previous occupations. In economics jargon, there should be a return to the initial labour market equilibrium.

Of course, we were ready for this question. We would argue that the presence of the refugees transformed the host communities. Individuals develop skills over time that relate to their specific economic activities and it could be costly to change activities. A second explanation we often offered was that the “end” of the refugee shock was not really thought of as a real “end” and that many Tanzanians might have expected the refugees to be back again (i.e. renewed violence in Burundi and/or Rwanda). Therefore, their choice of economic activities could have in fact reflected those expectations. This explanation is grounded in examples from qualitative research in the area during the 1990s which highlighted that locals were not convinced that the peace agreements in Burundi and Rwanda would last forever. One paper for instance found that “local entrepreneurs started large tree farms with the expectation that refugees would again be around in twenty years.”

Burundi is now in the midst of a controversial election in which the President is running for a third term in office. Thousands have left the country given the likelihood of a new large scale violent conflict. Tanzania has been, once again, one of the main destinations of these refugees. As such, refugees have in fact returned in recent months to Tanzania. Our results suggest that this time many Tanzanians are likely to be ready to take full advantage of the potential economic opportunities related to the refugee inflow.

The research discussed in this blog is part of The Labor Market Impacts of Forced Migration project, funded by the UK Department for International Development (DFID) and the Institute for the Study of Labor (IZA) for the benefit of developing countries. The views expressed are not necessarily those of DFID or IZA.

Author affiliations: Isabel Ruiz, Official Fellow and Tutor in Economics at Harris Manchester College