Pauline Pohl - A microsimulation model for official population projections

  • Presenting author: Pauline Pohl (Statistics Austria)

  • Authors: Pauline Pohl (Statistics Austria), Philip Slepecki (Statistics Austria), Martin Spielauer (WIFO)

  • Session: B01A - Population Projections - Tuesday 9:00-10:30 - Ceremonial Hall

  • Slides: PDF

Population projections in official statistics are generally produced using the cohort component method. It is computationally simple, does not require a broad range of input data, and is well-established among demographers. However, it cannot account for complex and dynamic demographic processes, model interactions, or produce detailed results for individual-level characteristics. To overcome these limitations, Statistics Austria has implemented a microsimulation model for its official population projection that builds on the characteristics of individuals instead of entire cohorts and allows for the simulation of realistic life-courses. While most dynamic microsimulations in the social sciences include some modelling of demographic processes and several national statistical offices have microsimulation models in their “toolbox”, Statistics Austria is the first European statistical office using microsimulation to produce its official population projection. To mitigate the effect of this methodological break on the comparability of projection results over time, we start by replicating the results of the cohort component method in a microsimulation and gradually introduce new model features. As a first step, we incorporate a model of international migration which explicitly accounts for the relationship between the emigration risk and the duration of residence. In addition, the place of birth is included as an individual-level attribute in the form of detailed country clusters. As an ex-post validation, we compare the results of the microsimulation with the cohort component method and the observed data for Austria for the years 2012 to 2021. We show that the microsimulation projection matches observed emigration patterns more closely, especially following the increase in immigration to Austria in 2015 and 2016. The model will be gradually refined and it can be extended with additional modules for education, employment, health and other socioeconomic characteristics.