You are here

Estimates of Australia's oldest old population


Official population estimates at the highest ages are often less accurate than those for younger ages in many countries which base their estimates on census counts. Official Estimated Resident Populations in Australia in the nonagenarian and centenarian ages suffer from some accuracy issues. Alternative methods based on deaths and survival at high ages can be applied to generate more reliable advanced age population estimates.

Download the data set

The data set is available to download free of charge. You agree to give appropriate credit by citing the below journal article.

Journal Article

Wilson, T., Temple, J. (2020): The rapid growth of Australia’s advanced age population. Journal of Population Research.

Australia’s population at the highest ages is growing rapidly, driving increasing demand for a whole range of services often required at advanced ages. Unfortunately, official population statistics tend to become less accurate as age increases, especially in the centenarian ages. The aims of the paper are to (i) describe the growth of Australia’s population at the highest ages using methods which provide greater accuracy and more detail than is available from official statistical sources; (ii) understand the demographic drivers of population growth at these ages; and (iii) urge caution with the use of official population data at the highest ages, and warn of some of the implications of doing so. State-of-the-art Extinct Generation and Survivor Ratio methods were employed to prepare robust estimates of Australia’s population by sex and single years of age for each year of the period 1971 to 2018. Results show that Australia’s population at the highest ages is growing very quickly. Over the 20 years to 2018 nonagenarians (those aged 90-99) grew in number from 63,000 to 181,000 (an increase of 185%), but centenarians (aged 100+) increased at an even faster rate, growing by 215% (from about 1,300 to 4,200). Official population estimates can be relied upon up to about age 85, but at higher ages the quality of single year of age population numbers may be poor. We encourage the ABS to adopt the methods described in this paper to produce official population estimates at the highest ages.


Associate Professor Jeromey Temple, The University of Melbourne

Dr Tom Wilson, The University of Melbourne