The IIASA Population and Just Societies (POPJUS) Program is presenting new demographic research at the Population Association of America (PAA) Annual Meeting 2025.
The Population Association of America’s (PAA) annual meeting is a premier conference for demographers, social scientists, and health researchers from the U.S. and beyond. Since 1932, it has provided a platform for scholars at all career stages to present research, share findings, and network. The conference covers a wide range of topics, including migration, reproductive health, and issues of race and gender.
Gregor Zens from the Migration and Sustainable Development (MIG) Research Group will present a Bayesian factor model for analyzing age-specific mortality, fertility, and migration patterns across small demographic subpopulations. His approach captures common patterns while addressing data sparsity and heterogeneity through hierarchical priors. A case study on immigration flows to Austria demonstrats the model’s effectiveness, showing it outperforms traditional methods in both estimation and prediction.
Saturday, April 12 | 8:00 AM - 9:15 AM
Session: Statistical Demography: Bayesian Methods
Flexible Bayesian Modeling of Age-Specific Counts in Many Demographic Subpopulations
Analyzing age-specific mortality, fertility, and migration patterns is a crucial task in statistical demography. In practice, such analysis is challenging when studying a large number of subpopulations, due to small observation counts within groups and increasing heterogeneity between groups. To address these challenges, we develop a Bayesian factor model for the joint analysis of age-specific counts in many, potentially small, subpopulations. The model uses smooth latent components to capture common age-specific patterns across subpopulations and encourages additional information sharing through a hierarchical prior. It provides smoothed estimates of the demographic pattern in each subpopulation, allows testing for heterogeneity, and can be used to assess the impact of observed covariates on the demographic process. A case study of age-specific immigration flows to Austria, disaggregated by sex and 155 origin countries, is discussed. Comparative analysis shows that the model outperforms commonly used benchmark frameworks in both in-sample and out-of-sample predictive exercises.