Demographic Research: Volume 35, Article 29
http://www.demographic-research.org 869
A population pyramid can be created in common spreadsheet or statistical software
without much difficulty, and is easy to understand. It presents immediately digestible
information on the age-sex structure of a population that would be less obvious in
tabular form. This distribution is important for understanding the demand for the wide
range of goods and services which vary by age and sex, such as baby products,
childcare, education, housing for first home buyers, household appliances, recreational
activities, aged care, and funeral services (Siegel 2002). Population age structure also
affects government spending and taxation receipts (Australian Government 2015). A
population pyramid additionally offers clues about a population’s fertility history,
mortality and migration, position in the demographic transition, and the likely influence
of its age-sex structure on future demographic change. The shape of a population
pyramid can hint at the economic or demographic role of the population in question, as
might be apparent for a university town, a popular seaside retirement region, or an area
with a communal establishment like a boarding school or prison. It can also be used to
check for potential data problems, such as age heaping. Finally, population pyramids
can also be quite entertaining, especially in dynamic form, as presented on some
national statistical offices’ and researchers’ websites. Examples can be found
• for Germany,
• for Australia, and
• for Moldova.
Several variations on the standard pyramid have been created. Many involve
disaggregating the bars into different categories of population, such as by marital status
(e.g., Statistics Canada 2013), household living arrangement (e.g., van Imhoff and
Keilman 1991), birthplace (e.g., Statistics Sweden 2009), educational attainment (e.g.,
Lutz, Cuaresma, and Sanderson 2008), labour force status (Statistics New Zealand
2015), ethnicity (e.g., Coleman 2010), and migrant status (e.g., Price 1998). Some
display shadows on the bars of the pyramid where there is an excess number of one sex
over another (e.g., Heenan 1965); others show age-sex structure with persons by age in
one histogram and age-specific sex ratios in another (Haak 1942). Commonly, two or
more population pyramids are overlaid on one another, often to compare age structure
over time (e.g., The Economist 2011), to show the outcome of alternative future
scenarios (e.g., Statistics Canada 2015), or to display probabilistic prediction intervals
(e.g., Keilman, Pham, and Hetland 2002).
However, population pyramids are not without their limitations. Because they
display only the size of each age-sex group in the population, it is usually not possible
to distinguish the relative contributions of different demographic processes to
population age structure. For example, if there is an indentation at the young adult ages,