Project Population Growth in Switzerland Using the Cohort Component Method

Project the development of the population at different spatial levels (e.g., cantons, districts, municipalities) using the cohort component method, broken down by 101 one-year age classes, gender, and nationality (see e.g. Preston et al. (2000, ISBN: 978-1-557-86451-2)). Includes helper functions to download required population data and projection parameters as open data for Switzerland, plus tools for model evaluation against benchmarks such as population records or a base model.

Welcome to the 'propop' package

Overview

The goal of 'propop' is to create population projections using the cohort component method (see e.g., Preston, Guillot, and Heuveline, 2000; Statistik Aargau, 2025a; Statistik Aargau, 2025b). For a breakdown of the components, see this vignette.

The package was developed for use with projection scenarios from the Swiss Federal Statistical Office (FSO). The current scenarios are available for the years 2025-2055. As starting population, you can either use population data from the FSO or your own population data.

'propop' uses the same calculation method as the FSO (2020; only available in French) and therefore reveals almost identical results (see vignette Evaluate projections). The package can be used to project the development of different demographic groups for different scenarios and at different spatial levels (e.g., single canton or municipalities within a canton).

While the package was primarily designed for use with FSO input data (e.g., mortality rates), it should in principle also work for other contexts (e.g., custom input data). However, it is important to provide the required input data in the specified form.

Installation

To install the current github version of the package, make sure you have devtools installed and type:

devtools::install_github("statistik-aargau/propop")

Vignettes

The package includes four vignettes.

• Vignette 1 shows how to prepare the FSO population data and parameters. (Users can of course use their own data and parameters).
• Vignette 2 illustrates how to run population projections for a single region.
• Vignette 3 illustrates how to pepare and run population projections for subregions.
• Vignette 4 shows how to evaluate projections against benchmarks.

Features

• propop::propop() works with either two nationalities (usually Swiss vs. non-Swiss nationals) or without distinguishing between nationalities.
• The FSO projections are only published in five-year intervals, the most recent ones being based on population records from 2023. 'propop' enables you to run projections with more recent population records.

Limitations

• propop::propop() currently requires 1-year age classes (0-100+) and two genders.
• This package was developed for use with FSO parameters (e.g., mortality rate, or emigration rate). Most parameters are only available for cantons and the whole of Switzerland. If you wish to run projections at smaller scales (e.g., districts), you need to prepare the parameters for each spatial entity before running the projection. This vignette includes tips of how to prepare your input data.
• Similarly, if you wish to adjust parameters (e.g., mortality rates that vary between subregions), you need to prepare the parameters accordingly. To adjust birth rates, you can use the package propopbirth.
• propop::propop()offers the possibility to account for varying subregional migration patterns and migration between subregions. However, before using this feature, users must adjust or calculate the required parameters (see this vignette).
• Custom parameters (e.g., consideration of sub-cantonal spatial entities, adjustments to reflect regional differences, or addition of subregional migration parameter) can be passed through the parameters object to propop::propop().
• The FSO uses additional ex-post adjustments to ensure that all the sums involving different cantons add up, which leads to some differences between the projections from the FSO and 'propop'.

Future plans

• There are also possible future features that could increase the accuracy of the projections. For example, considering the attractiveness of municipalities or planned housing developments could help to better explain why population growth varies between municipalities.

If you are interested in contributing to these or other developments, please get in touch with the package maintainer.

Quick example

To run propop::propop() with the example data included in the package (canton of Aargau), use the following code:

library(propop)
projection_canton_2030 <- propop(
  parameters = fso_parameters,
  year_first = 2026,
  year_last = 2030,
  population = fso_population,
  subregional = FALSE,
  binational = TRUE)