A Probabilistic Age-and-Gender Specific Malaysian Population Projection: Integrating Mortality, Fertility and Net Migration

Abstract

Population forecasting has become increasingly crucial for national planning, particularly in addressing demographic shifts including rising life expectancy, declining fertility rates and changing in migration flows. This study integrates three population components, including mortality, fertility and net-migration rates, into the probabilistic cohort component model in order to generate age-and-gender specific population forecasts. This study extends the existing projection methods by incorporating the Generalised Log Gamma model for fertility and the Heligman-Pollard model for mortality into the Lee-Carter model, along with the residual method for migration. Malaysian historical data, including populations, fertility rates and mortality rates by age group and gender spanning over 44 years from 1980 to 2023, were collected from the Department of Statistics Malaysia (DOSM), whereas net migration data were estimated from the demographic balance equation. These data were fitted into the proposed probabilistic cohort component method. Subsequently, the population projections were performed over the ten-year horizon. The results revealed that the Generalised Log-Gamma and the Heligman-Pollard models provide the best fit for Malaysia’s fertility and mortality data, respectively, as indicated by a low sum of squared errors. In terms of forecast accuracy, the GLG-Lee-Carter model produced highly accurate fertility projections, while the HP-Lee-Carter model performed best for male mortality based on mean absolute percentage errors. The projected population pyramids for Malaysia indicate that substantial demographic shifts resulted from a growing proportion of older adults and a shrinking young population. These trends suggest critical implications for national policy, particularly concerning labour market supply and the sustainability of social protection and retirement systems.