s part of their effort to pool individual risk, households consider spreading their members over a plurality of locations, both inside andoutside their country of origin. At the same time, the world is full of ‘Chinatowns’ and ‘Little Italies’: people, whenever they move, tend to bunch in the same location. Bunching would thus appear to be fundamentally at odds with the desire to diversify risk. In Discussion Paper No. 1540, Francesco Daveri and Riccardo Faini provide a framework for reconciling spatial bunching with the spread of migrants.

Much previous work on migration determinants focused on the role of wage and unemployment differentials, under the Harris-Todaro (1970) hypothesis of risk neutrality of an individual migrant. Daveri and Faini retain the Harris and Todaro assumption of rationality, but instead study the decision to migrate when it is taken at the household level by risk-averse agents seeking a shelter against uncertain income prospects. The authors’ model builds on the portfolio approach to the determination of the optimal demand for children developed by Appelbaum and Katz (1991). Among the authors’ innovations. however, is an allowance for concave household-level mobility costs. If households are not too risk-averse, this concavity drives all members of the same family to migrate to the same place or not to migrate at all, thus explaining the ‘bunching’ phenomenon, while the diversification of destinations is the result of heterogeneity of tastes. Evidence from Southern Italy is consistent with the main predictions of the model.

Where do Migrants go? Risk-Aversion, Mobility Costs and the Locational Choice of Migrants
Francesco Daveri and Riccardo Faini