Ana Viñuela (p), Alberto Díaz Dapena

Diana Gutiérrez Posada

For most workers, access to employment is severely restricted to the offers available in their municipalities or surrounding areas. Thus, the probability of being employed will depend not only on your personal characteristics -such as educational level, age, gender, civil status, number and age of children, etc.-, but more importantly on the socioeconomic characteristics of your environment or locality.
As part of the IMAJINE EU H2020 project, the objective of this paper is to study to what degree the chances of being employed are depending on the personal characteristics, the characteristics of the locality where people reside, or the characteristics of the region. Applying a multilevel logit model we will be able to identify the weight of the personal characteristics of the individuals in comparison to the exogenous factors, i.e, the socio economic context of the locality and region of residence.
The 2011 Microcensus databases for UK, France and Spain include the relevant variables to perform this type of analysis at local level (LAU2). Using microcensus individual data and collecting socio-economic information at LAU2 level (10,664 wards in UK, 36,683 communes in France and 8,111 municipalities in Spain) and for NUTs III regions, at local level we should observe the rural-urban dichotomy and/ or the relevance of location within the region. Ceteris paribus individual characteristics, results should show higher employability in metropolises than in rural areas, and a decrease with size. Also peripheral localities should pose lower chances to be employed than central areas. However, given the size and heterogeneity of the NUTS III regions in Europe, regional characteristics might not have as much influence over individual´s employability as the features of the locality of residence. In other words, patterns observed at local level must not necessarily hold at regional level.

Alberto Díaz Dapena, Esteban Fernández Vázquez, Fernando Rubiera Morollón (p), Ana Viñuela

Alfredo Cartone

See the extended abstract

Paolo Postiglione (p), Alfredo Cartone (p)

Alberto Diaz Dapena

see extended abstract

Esteban Fernandez-vazquez, Alberto Diaz Dapena (p), Fernando Rubiera Morollon, Ana Viñuela

Paolo Postiglione

Recent literature on economic geography suggests that significant disparities in social and economic variables are expected to happen at a local (sub-regional) scale. Most of the official indicators taken as reference when studying inequalities in the European Union (EU), however, are only observable at the scale of NUTS II regions. This prevents the analysis of disparities across local entities within the regions, which could be a substantial portion of the territorial inequality. Similarly, well-being indicators as head-count ratios of poverty can be studied for the EU at the same spatial scale from the figures contained in surveys like the Social and Living Conditions (SILC). Comparing poverty indicators across sub-regional entities, consequently, is not either possible from these figures.
This paper proposes a methodology to disaggregate the regional figures that are contained in household surveys at an aggregate regional scale into indicators produced at a smaller spatial scale. The methodology proposed here grounds on previous literature as Elbers et al. (2003) and Tazzoni and Deaton (2009). The point of departure of these works was the combination of household surveys that contain detailed economic information that is only reliable at an aggregate spatial scale with census data that is reliable at a small scale but not containing economic indicators. Applications of this idea have been applied by the World Bank to map poverty and inequality in countries like Cambodia, Mexico, Morocco South Africa or Uganda. This methodology, however, does not guarantee consistency between the disaggregated (sub-regional) estimates and the aggregate (i.e., regional or national) figures contained in the household surveys. The novelty of our proposal is that applies a second step to the initial estimates to adjust them and make them consistent with the regional and national aggregates. In particular, we propose the use of a Generalized Cross Entropy (GCE) estimator to the problem of spatial disaggregation similar to the approach in Bernardini-Papalia and Fernandez-Vazquez (2018) for small area estimation that exploits auxiliary information relating to observable variablesand adjust for consistency.
As an illustration we use the case of Spain where micro-data at the household level from the SILC and the most recent Census are combined in such a way. The resulting estimates of provide information of average household income and head-count poverty ratios for a set of more than five hundreds local areas that are consistent with the average indicators published in the SILC for the seventeen Spanish NUTS-II regions.

Diana Gutiérrez Posada (p), Maria Plotnikova

Ana Vinuela

See extended abstract