Effect of Age Group Length, Growth rate, & Degree of Development on the Aggregate & Disaggregate population forecasting Methods

Abstract

Abstract This research investigates the performance of selected aggregate and disaggregate methods for population forecasts. The data used is that given by the United Nations Population Division- Department of Economic and Social Affairs and covers the period 1950-2010. Also, we used simulated data that were generated by Mathematical packages. The objective, with respect to the disaggregate methods, is to determine their performance under various age group length, degree of development, gender & growth rate. It is found that for all methods the precision of the forecast decreases with increase in age group length & growth rate and increase with the degree of development. Forecast of female population is in general more precise than of male population. The cohort component method proved to be the best among the disaggregate methods in forecasting population for individual age groups. The exponential smoothing method, which as far as the author knows, is used for the first time as a disaggregate methods, provided the best forecast for the population total, followed by the cohort component method. As to the aggregate methods the cubic method showed the best performance. The sensitivity of the aggregate methods to the various factors is such that the precision of forecast decreases with increase in growth rate &increases with the degree of development. Looking at both aggregate & disaggregate methods the results confirm that the exponential smoothing method, suggested by the author, provided the best performance for forecast is the population total.