Australia's aged population is increasing faster than the government realises, population forecasting by Monash University researchers has shown.
Forecasting expert Professor Rob Hyndman, from the Department of Econometrics and Business Statistics, said the number of elderly people was increasing faster than official estimates and warned that the Australian Bureau of Statistics (ABS) was underestimating the future number of people aged over 85.
"The ABS forecast of that sector of the population for 2031 is 660,000, while ours is 845,000 -- a difference of 180,000, or 28 per cent," he said.
"That discrepancy could mean that we will be supporting more older citizens in 20 to 40 years than the government has provided for with health and community services.
"The projected growth in the proportion of elderly Australians means some hard decisions need to be made about the provision and funding of services for the aged over the next 20 years."
Professor Hyndman, director of the university's Business and Economic Forecasting Unit, said that while Monash and the ABS both used mortality and fertility rates to make their predictions, the university's predictions were more accurate because they allowed for complex dynamic changes in those elements.
"Mortality and fertility rates don't fall or rise smoothly. For instance, in the late 1980s and early 1990s, the mortality rate for people aged between 18 and 25 actually increased due to AIDS, suicides and road deaths. It has since started dropping again," he said.
The ABS was also inherently conservative in its predictions, Professor Hyndman explained, and worked on the assumption that the overall mortality rate would not continue to decline at the same rate that it had for the past 100 years.
"However, we take that trend over the past 100 years and extrapolate it into the future," he said.
"Our approach is unique because, for the first time, we can put probability limits around the population for different ages, and we can also attach a probability rating to the forecast itself -- about how certain we are that it will be correct.
"So instead of just predicting that in 50 years time there will be about one million people aged between 50 and 60, we can predict with 95 per cent probability that there will be, say, between 800,000 and 1.2 million in that age group.