Small area estimation of divergence indexes under compositional multivariate Fay-Herriot models

E. Cabello García, D. Morales, A. Pérez Martín

This paper develops model-based predictors of proportions of employed men and women and divergence indexes between sexes by occupation sectors and areas. Since the direct estimators of the proportions add up to one in the occupational sections, they are compositions that can be imprecise if the sample sizes are small. We fit a multivariate Fay-Herriot model to log-ratio transformations of the direct estimators of the proportions. Small area estimators of the proportions and divergence indexes are derived from the fitted model and the corresponding mean squared errors are estimated by parametric bootstrap. Several simulation experiments designed to analyze the behaviour of the introduced estimators are carried out. We give an application Spanish Labour Force Survey data from 2022. The target is to investigate the state of sex occupational divergence by province in Spain.

Keywords: Small area estimation, multivariate Fay-Herriot model, compositional data, bootstrap, divergence index, occupation sectors, Labour Force Survey.

Scheduled

Mixed Models II
November 10, 2023  9:30 AM
HC3: Canónigos Room 3

Estimación de indicadores de pobreza en áreas pequeñas bajo sesgo de selección de la muestra

I. Molina Peralta, E. Berg, Y. Cho, M. Guadarrama, A. Eideh

Model-based prediction of small area labour force indicators under unit-level multinomial mixed models.

M. Bugallo Porto, M. D. Esteban Lefler, T. Hobza, D. Morales, A. Pérez Martín