Population Forecasting

Abstract

Population forecasters have paid too little attention to forecast accuracy, uncertainty, and approaches other than the cohort-component method. They should track forecast errors and use them to adjust forecasts. They have chosen measures of forecast accuracy arbitrarily, with the result that flawed error measures are widely used in population forecasting. An ex- amination of past forecasts would help establish what approaches are most accurate in particular applications and under what circumstances. Researchers have found that alternative approaches to population fore- casting, including econometric models and extrapolation, provide more accurate forecasts than the cohort-component method in at least some situations. If they can determine the conditions under which these ap- proaches are best, they can use them instead of the established method or in combination with it. Methodological advances have made it possible to produce population forecasts with a greater degree of disaggregation and decomposition than before. If this decomposition allows a better understanding of the causal forces underlying population change, then decomposition may improve forecast accuracy. Even if disaggregation and decomposition do not im- prove overall forecast accuracy, they may lead to improved understanding or accurate forecasts of important components of the population, such as the elderly widowed population. Uncertainly has not been well-integrated into population forecasts. Researchers are pushing ahead in three main ar- eas: population forecasts that include probability distributions; combining expert judgment and statistical methods; and the specification of situations that provide an internally consistent forecast of the population under par- ticular circumstances. Evidence suggests that relying on experts to choose the fertility and mortality assumptions of the forecast has done little to im- prove forecast accuracy, but this is probably because expert opinion has been obtained in an unstructured way. Experience in other areas of fore- casting has shown how to use experts to improve forecast accuracy.