A Hierarchical Bayesian Model to Standardize and Impute Residence Permit Data

• Activity Number: 355 - Advanced Bayesian Topics (Part 4)
• Type: Contributed
• Date/Time: Thursday, August 12, 2021 : 10:00 AM to 11:50 AM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Andrea Lisette Aparicio-Castro* and Arkadiusz Wi?niowski and Francisco Rowe and Mark Brown
• Companies: Department of Social Statistics, School of Social Sciences, University of Manchester and Department of Social Statistics, School of Social Sciences, University of Manchester and Geographic Data Science Lab, University of Liverpool and Department of Social Statistics, School of Social Sciences, University of Manchester
• Keywords: Bayesian inference; Residence permit data; International migration flows; Data model; Additive mixed gravity model; South America

Abstract

Under the current rapid changes in migration, residence permit data is one of the most common sources due to their availability, access and production frequency. However, using these data implies dealing with (i)country-specific legislations that define differently who constitutes a migrant, and thus, who is being counted; and (ii)dissimilarities of national data collection systems. We propose a Bayesian hierarchical model that integrates a data model and an additive mixed gravity model(BAMG-M) to overcome these two problems. The data model standardises the data by accounting for (i)minimum timing requirements to acquire a residence permit; (ii)data quality; (iii)migrants who do not need a residence permit because of the implementation of free movement agreements; and (iv)unauthorised migrants who are part of regularisation processes. The BAMG-M imputes missing migration flows, enabling modelling non-linear and non-uniformly-spaced trends. We illustrate the use of the model with South American residence permit data reported by SICREMI from 2000 to 2019. The output is a set of synthetic estimates of bilateral migration flows with measures of uncertainty.