MIGDCM - Modeling Migration Intentions Using Advanced Discrete Choice Models - FDEF

The project at a glance

Start date:

01 Mar 2022
Duration in months:

36
Funding:

480 105
Principal Investigator(s):

Michel BEINE

About

This research programme proposes extensions of the use of the Cross-Nested Logit (CNL) approach to the modelling of location choices. The CNL allows to better capture the stochastic structure of these choices through the use of overlapping nests of alternative locations and to account for deviations from the property of independence of irrelevant alternatives. The research program extends the recent work of Beine, Bierlaire and Docquier (2021). This work might be seen as pioneer work that allows important issues like the substitutions across alternative destinations to be addressed. Nevertheless, it leaves many issues unaddressed. In this research programme, we tackle the most pressing issues of the use of the CNL. It will first evaluate the transferability of the CNL to other contexts than the particular case of India. It will develop the technique of alternative sampling to include a comprehensive set of destinations within the CNL. It will improve the specifications, in particular by modelling perceptions of individuals through latent variables and it will evaluate the relevance of using preparation plans for migration rather than pure intentions in capturing self-selection factors of mobility. The programme will run for a period of 36 months and will feature close collaboration between two teams, one located at the University of Luxembourg and one located at the Ecole Polytechnique Fédérale in Lausanne.