Towards harmonisation of migration statistics: a probabilistic perspective
Beata Nowok, Netherlands Interdisciplinary Demographic Institute (NIDI)
Frans Willekens, Netherlands Interdisciplinary Demographic Institute (NIDI)
Inadequate and inconsistent data are a common and persistent problem in the field of migration. Deficiencies in migration statistics may be tackled using modelling techniques, which has recently been recognized by the European Union policymakers. The new Regulation on Community statistics on international migration (EC 862/2007), which obliges countries to supply harmonised statistics, provides for possibility of using estimation methods to adapt statistics based on national definitions to comply with the required one-year duration of stay definition. The main objective of this paper is to look at migration from a probabilistic perspective and provide an insight into a possible increase in harmonisation of the available flow statistics. Since migration is a random event, a probabilistic approach is natural and has been widely used in modelling migration. The novelty of this study consists in applying probability theory to harmonisation issues. In general, different migration statistics, which are always put in a discrete time framework, represent the same continuous process generating the data. They differ depending on how the data happened to be collected and how the statistics happened to be produced. We assume that migration events are generated by a homogenous Poisson process (it may be generalized by allowing migration rate to be variable). The application of a simply model amenable to probabilistic calculations, allows expressing different migration measures in terms of the underlying process and to find relations between them. The main focus is put on the time criterion used in migration definition. It refers to duration of stay following relocation, which is specified very differently among countries and constitutes the main source of discrepancies in operationalization of migration concept in the EU states. The model shows that this inevitably leads to different numbers of recorded migration and the level of discrepancies depends on migration rate.
Presented in Session 80: Data Sources, Measurement and Models