Kazi Zubair Hossain (khossain@isrt.ac.bd)
Institute of Statistical Research and Training, University of Dhaka Dhaka, Bangladesh
Abstract
The future trajectory of population size and composition is crucial for informed decision-making in public and private domains. Countries with continuous population growth, like Bangladesh, need accurate and prompt population forecasts at national and local levels to efficiently use limited resources and properly utilize the rising workforce. This study used ARIMA models to forecast age-sex-specific populations of Bangladesh until 2031 and then aggregated them to produce national forecasts. Aggregated ARIMA showed one of the lowest errors among the existing methods compared to the 2022 population census count. Additionally, combined forecasts yielded the highest accuracy compared to individual methods. The results present detailed insights into the future population of Bangladesh disaggregated by age and sex.
Keywords: Population forecasts, ARIMA, Box-Jenkins, combined forecasts, age-sex dynamics, Bangladesh
JEL Classification: C510, C520, C530, J110
