Estimating net migrations at regional and age-specific group levels

Following the Mediterranean Region trend, the Portuguese peripheral NUTS III regions are simultaneously affected by a human desertification and population ageing. In order to analyse the requirements to revert this process, DEMOSPIN project - Economically Sustainable DEMOgraphy - ReverSing Decline in Peripheral RegIoNs - is developing a joint demographic-economic model, where migrations are the link between the economic and demographic dynamics. The developed model estimates net migration for each age group, considering three kinds of migrants: i) the active age groups migrants (20-65 years old), responsive to job opportunities; ii) the older age groups (more than 65 years old), where migrations reverse the movements occurred in the past, and iii) the youngest age groups (less than 20 years old) which follow the parents migrations or are driven by opportunities to study elsewhere. This model was based on the census data for 1991, 2001 and 2011 and was developed to estimate migration based i) on socioeconomic factors: creation or destruction of employment (split in primary and nor primary sector), GDP per capita and demographic potential and ii) on the ancient migration, occurred 10, 20, 30, 40 or 50 years ago (for example, the migration of the 65 to 70 years old population related to migration of 25 to 30 years old 40 years ago). A satellite model was developed for migrants with ages below 20 years old where movement is related to parent's migration patterns. The coefficients of the regression model are specific to each age group and sex. The results show a very high adjustment for younger groups, between migration and socioeconomic drivers. For people older than 40 years, such adjustment is lower but it is partially compensated by the effect of migrations occurred in the past. Globally, the model provides an accurate description of NUTS III migration patterns. This paper will present, at first, the migration patterns of the Portuguese peripheral regions and the theoretical basis of migration estimation models and its determinants. Afterwards, it will describe the methodology underlying the model construction and the obtained results, providing a deep understanding of the migration dynamics of the regions under study.