Soraia Pereira
Centro de Estatística e Aplicações, FCUL

Official estimates of the labor market that are published quarterly by the INE (National Statistical Institute of Portugal) are based on a direct method using the Portuguese Labor Force Survey sample. These estimates are available both at national level and for NUTS II regions of Portugal.  Currently, knowledge of the labor market requires reliable estimates at a more disaggregated level, particularly at NUTS III level. However, due to the small size of these areas, there is insufficient information to obtain estimates with acceptable accuracy using the direct method. 

Recently, there have been considerable methodological developments to solve small area estimation problems. The majority of these methods are based on generalized linear models applied to areal data. 

In this study, we propose the application of a spatial point process approach, thereby taking into account the spatial location and information specific to the households, unlike the aggregate data models that do not take into account this information. We model the intensity of unemployment through a log-Gaussian Cox process model, using the SPDE and INLA methodologies.