Skip to main content
Log in

Evaluating Alternative Implementations of the Hamilton-Perry Model for Small Area Population Forecasts: the Case of Australia

  • Published:
Spatial Demography Aims and scope Submit manuscript

Abstract

Small area population forecasts are widely used across the public and private sectors, with many users requiring forecasts broken down by sex and age group. The preparation of small area age-sex population forecasts across a whole country or State with a multiregional cohort-component model is usually a time-consuming and expensive task. It involves the purchase of large datasets, considerable amounts of complex data preparation and assumption-setting, and substantial amounts of staff time. A quicker and lower-cost alternative is to use a reduced form cohort projection model, such as the Hamilton-Perry model. This paper presents an evaluation of various implementations of the Hamilton-Perry model, including an alternative version employing a combination of Cohort Change Ratios and Cohort Change Differences. It also evaluates the effects on forecast accuracy of smoothing the age profiles of Cohort Change Ratios and Differences, and constraining to independent population forecasts. Population ‘forecasts’ were created for small areas of Australia over the horizon 2006–16 and compared against population estimates. The most accurate implementation is found to be the Hamilton-Perry model using a combination of Cohort Change Ratios and Cohort Change Differences, smoothed age profiles, and with constraining to independent forecasts.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5

Similar content being viewed by others

Availability of data and material

The input data needed to replicate the population forecasts reported in this paper are available at https://doi.org/10.6084/m9.figshare.16822495.v1.

Notes

  1. A period-cohort refers to the Lexis space defined by a cohort over a specified period (Fig. 1). We use notation such as 0–4 – 5–9, which refers to the cohort aged 0–4 at the start of the projection interval and aged 5–9 at the end.

References

Download references

Acknowledgements

We thank the anonymous reviewers for their helpful comments on the initial version of this paper.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Tom Wilson.

Ethics declarations

Conflict of Interest

We have no competing interests to declare.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Appendix

Appendix

See Tables 8, 9, 10, 11 and 12.

Table 8 Mean error and bias of SA2 total population forecasts
Table 9 Mean error and bias of SA2 age-sex population forecasts
Table 10 Errors in forecasting the population age-sex distribution of SA2 areas as measured by the mean Index of Dissimilarity
Table 11 Errors in forecasting the population age-sex distribution of SA2 areas by population size and growth rate categories as measured by the mean Index of Dissimilarity, Model 2d
Table 12 Regression results for the MedAPE over age and sex, Model 2d, 2016

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Wilson, T., Grossman, I. Evaluating Alternative Implementations of the Hamilton-Perry Model for Small Area Population Forecasts: the Case of Australia. Spat Demogr 10, 1–31 (2022). https://doi.org/10.1007/s40980-021-00103-9

Download citation

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s40980-021-00103-9

Keywords

Navigation