Abstract
The Hamilton-Perry method is a variant of the Cohort-Component population forecasting method that has minimal data input requirements. It only requires the age distributions for a population at two points in time, which generally are two successive census enumerations. Although the method has gained acceptance, tests of its accuracy are limited. In this chapter we evaluate the forecast accuracy of the Hamilton-Perry method both in terms of age and total population (obtained by summing up the forecasted age groups). This evaluation is based on a sample of four states (one from each of the four census regions) and decennial census data from 1900 to 2010, which yield 10 census test points (1920, 1930, 1940,…, 2010) that provide a wide range of characteristics in regard to population size, growth, and age-composition that affect forecast accuracy. We conclude that the results are encouraging and suggest that the Hamilton-Perry method be considered when either a 10-year forecast of state populations by age or a total population are desired and components of change are not required.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
Although the name “Hamilton-Perry method” is virtually universal today, the first published instance of cohort change ratios for purposes of population forecasting is found in Hardy and Wyatt (1911), who built cohort change ratios from the 1901 and 1906 census counts of England and applied them to the 1906 census to generate a forecast for 1911. Hamilton and Perry acknowledge that they learned about this method from a description of it found in Wolfenden (1954), but they were unable to secure a copy of the 1911 article and were, therefore, not exactly certain what was done by Hardy and Wyatt. In any event, Hamilton and Perry deserve credit for providing a clear and detailed description of this approach to population forecasting in a journal was read by many demographers in the United States and elsewhere prior to the founding of demographic journals such as Canadian Studies in Population (first published in 1973) Demography (first published in 1966) and Population Research and Policy Review (first published in 1982).
- 2.
Other measures of precision can be used when the distribution of APEs is highly asymmetrical (Swanson et al. 2011, 2012). Two such measures are the “MEDAPE” (Median Absolute Percent Error); and “MAPE-R” (MAPE–Rescaled). The MAPE can be supplemented with MEDAPE, which is the median value of the APEs, or MAPE-R, a measure that minimizes the effect of outliers while preserving more information about the original distribution than MEDAPE (Swanson et al. 2011). One drawback of MAPE-R is that it is much more complicated to calculate than MEDAPE. Since MAPE is so widely known and forecast errors are generally stable across a variety of error measures (Rayer 2007), we forego using MEDAPE and MAPE-R in this study.
- 3.
In rapidly changing areas it is advisable to control Hamilton-Perry (HP) forecasts by age to an independent total population forecast (Smith et al. 2013:180–181). The application of constant CCRs can often lead to forecasts that are too high (low) in rapidly increasing (decreasing) areas. Smith and Tayman (2003) found that while uncontrolled HP forecasts generally had larger errors than the controlled forecasts, the patterns of errors by age groups very generally similar for both sets.
- 4.
While we caution against using the Hamilton-Perry method for forecast horizons longer than 10-years, we note that Smith and Tayman (2003) found that the Hamilton-Perry method produces levels of accuracy comparable to the cohort-component method over a twenty year horizon and Swanson (2016) found it useful over a long-term “reverse” projection used to reconstruct the population of Native and Part-Hawaiians in Hawaii (see also Chapter 25 in this book). Thus, we find that the method can be used over longer horizons, given the context and the population in question.
References
Baker, J., Ruan, X., & Alcantara, A. (2011). Incorporating spatial effects into Hamilton-Perry based projections improves accuracy and precision at the census tract level. Paper presented at the annual meeting of the Population Association of America, Washington, DC.
Fonseca, L., & Tayman, J. (1989). Postcensal estimates of household income distributions. Demography, 26(1), 149–159.
George, M. V., Swanson, D. A., & Tayman, J. (2004). Population projections. In J. Siegel & D. Swanson (Eds.), The methods and materials of demography (2nd ed., pp. 561–601). New York: Elsevier Academic Press.
Goldscheider, C. (1978). Modernization, migration, and urbanization. Paris: International Union for the Scientific Study of Population.
Hamilton, C. H., & Perry, J. (1962). A short method for projecting population by age from one decennial census to another. Social Forces, 41, 163–170.
Hardy, G. F., & Wyatt, F. B. (1911). Report of the actuaries in relation to the scheme of insurance against sickness, disablement, etc. Journal of the Institute of Actuaries, XLV, 406–443.
Land, K. (1986). Methods for national population forecasts: A review. Journal of the American Statistical Association, 81, 888–901.
Murdock, S., & Ellis, D. (1991). Applied demography: An introduction to basic concepts, methods, and data. Denver: Westview Press.
Pittenger, D. (1976). Projecting state and local populations. Cambridge, MA: Ballinger.
Rayer, S. (2007). Population forecast accuracy: Does the choice of summary measure of error matter? Population Research and Policy Review, 26, 163–184.
Smith, S. K., & Tayman, J. (2003). An evaluation of population projections by age. Demography, 40(4), 741–757.
Smith, S. K., Tayman, J., & Swanson, D. A. (2013). A practitioner’s guide to state and local population projections. Dordrecht: Springer.
Swanson, D. A. (2015). On the relationship among values of the same summary measure of error when it is used across multiple characteristics at the same point in time: An examination of MALPE and MAPE. Review of Economics and Finance, 5(1), 1–14.
Swanson, D. A. (2016, June 1–3). The number of native Hawaiians and part-Hawaiians in Hawaiʻi, 1778 to 1900: Demographic estimates by age, with discussion. Paper presented at the annual conference of the Canadian Population Association, Calgary, Alberta.
Swanson, D. A., & Beck, D. (1994). A new short-term county population projection method. Journal of Economic and Social Measurement, 20, 1–26.
Swanson, D. A. & Tayman, J. (2014). Measuring uncertainty in population forecasts: A new approach. In M. Marsili, & G. Capacci (Eds.), Proceedings of the 6th EUROSTAT/UNECE work session on demographic projections (pp. 203–215). Rome: National Institute of Statistics.
Swanson, D. A., & Tedrow, L. (2012). Using cohort change ratios to estimate life expectancy in populations with negligible migration: A new approach. Canadian Studies in Population, 39, 83–90.
Swanson, D. A., Schlottmann, A., & Schmidt, R. (2010). Forecasting the population of census tracts by age and sex: An example of the Hamilton–Perry method in action. Population Research and Policy Review, 29, 47–63.
Swanson, D. A., Tayman, J., & Bryan, T. (2011). MAPE-R: A rescaled measure of accuracy for cross-sectional, sub-national forecasts. Journal of Population Research, 28, 225–243.
Swanson, D. A., Tayman, J., McKibben, J. & Cropper, M. (2012). A “blind” ex post facto evaluation of total population and total household forecast for small areas made by five vendors for 2010: Results by geography and error criteria. Presented at the 2012 conference of the Canadian Population Society, Waterloo, Ontario, Canada.
Vaupel, J., & Yashin, A. (1985, August). Heterogeneity’s ruses: Some surprising effects of selection on population dynamics. The American Statistician, 39, 176–185.
Wolfenden, H. H. (1954). Population statistics and their compilation (Rev. ed.). Chicago: University of Chicago Press.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Appendices
Appendices
1.1 Appendix A.1: Hamilton Perry Method and the Fundamental Demographic Equation
In this Appendix we demonstrate that the Hamilton-Perry method satisfies the fundamental demographic equation. We begin by restating the fundamental demographic equation as follows:
where,
-
P i,t = Population of area i at time t (the launch year),
-
P i , t+k = Population of area i at time t + k (the forecast year),
-
B i = Births in area i between time t and t + k,
-
D i = Deaths in area i between time t and t + k,
-
I i = In-migrants in area i between time t and t + k, and
-
O i = Out-migrants in area i between time t and t+
then,
Expressing Eq. (A.1.2) in terms of Eq. (A.1.1):
where x + k > = 10, then,
since Ni = Ii – Oi, where x + k ≥ 10, we have
Equations (A.1.2), (A1.3) and (A.1.4) show that the Hamilton-Perry method is not only consistent with the fundamental demographic equation, but also closely related to the cohort-component method. The Hamilton-Perry method simply expresses the individual components of change—births, deaths, and migration—in terms of CCRs (Cohort Change Ratios). An important reason for a demographic forecasting method to be consistent with the fundamental demographic equation is to minimize the potential errors associated with hidden heterogeneity (Vaupel and Yashin 1985).
1.2 Appendix A.2: Population by Age, 1900 to 2010 by Decade (Tables A.2.1, A.2.2, A.2.3, and A.2.4)
Rights and permissions
Copyright information
© 2017 Springer International Publishing Switzerland
About this chapter
Cite this chapter
Swanson, D.A., Tayman, J. (2017). A Long Term Test of the Accuracy of the Hamilton-Perry Method for Forecasting State Populations by Age. In: Swanson, D. (eds) The Frontiers of Applied Demography. Applied Demography Series, vol 9. Springer, Cham. https://doi.org/10.1007/978-3-319-43329-5_23
Download citation
DOI: https://doi.org/10.1007/978-3-319-43329-5_23
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-43327-1
Online ISBN: 978-3-319-43329-5
eBook Packages: Social SciencesSocial Sciences (R0)