An empirical study to determine which variables are most important in driving migration decisions/an extension of the Harris-Todaro model

Within the Economics of Developing Countries paper, I particularly enjoyed analysing the incentives that underpin micro-level decisions, such as the decision to migrate, or enrol in secondary school. My essay examined these incentives, arguing that risk considerations increase the incentive for rural-urban migration in developing countries. I have two potential thesis ideas related to this topic: one is to conduct an empirical study to determine which variables are most important in driving migration decisions. I believe the Young Lives study conducted by the Oxford Department of International Development would provide an excellent dataset for this purpose, as the cohort of children born in 1994-95 will have now reached an age where migration decisions typically take place. The second idea relates to a phenomena I observed while working with urban refugees in Kampala. I was curious as to why many refugees were leaving refugee camps in the north - where they were at least (virtually) guaranteed schooling, education and housing - to migrate to Kampala, where there was a high likelihood of them being unemployed, out of school and living in a slum. The traditional
Harris-Todaro model did not seem appropriate in this situation, since refugees faced an extremely slim probability of securing a formal job, due to language barriers, xenophobia and bureaucratic constraints. I theorised that the refugees were engaging in a similar act of risk-balancing, but were confronting a gamble with riskier odds but potentially much higher, non-income rewards. For instance, by migrating to the cities, refugees increased their chances of being resettled in a developed country fivefold, as their closer proximity to UNHCR headquarters better enabled them to 'play the system' and make their case heard. They were also more likely to acquire a sponsorship deal from a refugee-focused NGO, which are disproportionately located in the cities. These 'outcomes' cannot be captured by the HT framework, yet I believe they comprise the main drivers of rural-urban migration amongst refugees in developing countries. Thus, as a potential thesis, I propose an extension of the Harris-Todaro model: one which focuses solely on refugees and which expands the framework to include outcomes not related to expected income alone. This thesis would offer two insights: a useful broadening of the HT model to include non-income 'payouts' as well as a greater understanding of the unique challenges faced by refugees in developing countries, a topic which, I believe, has been overly neglected by the development literature. I have supplemented my academic interest in development economics with practical experience within the field. For the last two months, I have conducted an impact assessment for an education NGO in Kampala, assessing whether their adult education hubs have had a positive impact on local refugee communities. This experience has highlighted the difficulties faced when doing applied work in development economics, such as the problem of multiple causality, data limitations and language barriers. However, it has also introduced me to creative ways to avoid these difficulties, such as the use of qualitative data and randomised control trials. Before this, I also worked for the Oxford Microfinance Initiative on two projects, focusing on measuring the social impact of microloans in Assam (India) and Takeo (Cambodia). I was fascinated by some of the qualitative techniques we used to gather data, such as the use of Participatory Poverty Assessments to implicitly reveal the utility functions of our various interviewees. Working in the field and witnessing first-hand the local impact of microloans also sparked my interest in micro-level solutions to global development issues. A recent initiative I find particularly exciting is the use of mobile phone data to credit-score individuals who have been isolated from traditional credit markets.