Small Area Estimation of Poverty Indicators for Brazil Using Bayesian Beta Models
Seminario de Estadística y Actuaría
PhD. Denise Britz do Nascimento Silva
- Principal Researcher at the Society for the Development of Scientific Research (SCIENCE)
24 ABR 2025
11:00 h.
Aula Magna Leonila Vázquez
Transmisión: @matefcienciasunam
Goal 1 of the 2030 Agenda for Sustainable Development is to eradicate poverty for all people, everywhere. The assessment of progress towards this goal relies on statistical data for different geographical areas and population groups. Although reliable state-level poverty measures can be obtained from the Brazilian National Household Survey, the production of more disaggregated statistics requires small-area estimation methods. I will present the development of Bayesian small area models, with a Beta distribution and logistic link function, to estimate poverty incidence (headcount ratio), gap and severity for Brazilian municipality strata from 2012 to 2022. The Beta regression models incorporate area level socioeconomic auxiliary variables, as well as area and time random effects. The model-based estimates provide substantial improvement in precision when compared to direct estimates without evidence of bias. This is joint work with Maria Eduarda Gallo (the results are in her MSc dissertation from the Federal University of Rio de Janeiro - UFRJ) and Kelly Cristina Gonçalves (UFRJ – principal supervisor).
Informes: lizbethna@ciencias.unam.mx