Monte Carlo methods in spatio-temporal regression modelling of migration in the EU

Spatio-temporal regression models are well developed in disciplines such as, for example, climate and geostatistics, but have had little application in the modelling of economic phe- nomena. In this study we have modelled migrations of workers and firms across the European Union during the period 1988-2005. The data set has been extracted from Eurostats Labour Force Survey (LFS) and contains information stratified by European region. It is interesting to determine whether the spatial component in the migration patterns is based either on neighbourhood, on distance or on some other metric (such as travel time). A preliminary test for spatial autocorrelation in residuals from an estimated linear model (Morans I test) suggests the existence of a significant spatial effect. The complete spatio-temporal model has been implemented using conditional auto- regressive (CAR) and simultaneously autoregres- sive (SAR) random effects in the Bayesian framework. In recent years, Bayesian methods have been widely applied to spatio-temporal modelling since they enable the use of Markov chain Monte Carlo (MCMC) samplers to estimate parameters of the model. In this talk, we consider various Monte Carlo approaches (including advanced MCMC methods) to estimate model parameters, and we compare different computing schemes.