The cohort-component method is the standard way to work out future population by age and sex. It tracks each age group (cohort) forward through time, applying assumptions about births, deaths, and migration to work out how the population changes. The ABS, state agencies, the United Nations, and Place Forecast all use this method.
The key idea is diagonal cohort shifting. The method follows cohorts along a diagonal of an age-by-time grid. A person aged 30 in 2021 joins the age-31 group in 2022. They join the age-32 group in 2023. And so on. The number of people aged a in year t comes from the number aged a−1 in year t−1. Then deaths are taken away and migrants are added or removed:
Population(age=a, year=t) = Population(age=a−1, year=t−1) − Deaths + In-Migration − Out-Migration
This makes sure no one shows up at two ages in the same year. A cohort can only shrink through death or grow through migration.
Births add newborns at age 0. Deaths remove people at each age. In-migration adds people moving into an area and out-migration removes people moving out of an area. Each part is broken down by age and sex. In Place Forecast, migration balancing is always on, keeping all parts consistent.
The method follows set steps. First, the jump-off year population gives the starting age-sex split. This is the last year of estimated resident population data. Second, the system works out rates from past data. It uses diagonal cohort shifting so rates use the right base. Third, assumptions are set for how rates will change. Fourth, the method moves forward year by year. It applies the assumed rates to calculate births, deaths, and migration for each age group.
One key rule: do not add rates across ages, years, or areas. Always add the raw counts first, then calculate rates from those totals. Adding rates gives wrong results.
Using this method for small areas brings extra challenges. Migration data is less reliable at smaller scales, and age-sex profiles show more variation. At this level, population change is often driven by housing supply rather than broader trends. Place Forecast handles this with dwelling-based target scaling. Where dwelling projections show that more dwellings will be built, the system adjusts in-migration to match that growth. This links dwelling projections to population through the cohort-component method via average household size.
Place Forecast uses this method at its core. Every projected figure comes from it. Small areas are forecast individually, and LGA figures are aggregated from their small areas. The method moves year by year from the jump-off population. It uses age-sex rates to build each future year's numbers.
Users can see each part on its own page. The fertility page shows birth rates. The death page shows mortality rates. The migration pages show people moving in and people moving out. The population page shows the combined result. This lets users trace any projected figure back to its source.