Woods Method: Demographic Forecasting Made Easy
Webb Sprague, State of Washington
One shortcoming of using the Cohort Change Ratio approach is that it confounds demographic dynamics due to both migration and survival into a single change ratio. In the Leslie matrix projection framework, this ratio is stored in the subdiagonal elements of the matrix, the section of the matrix most appropriate for survival ratios. We improve upon simple cohort change ratios in the Leslie matrix framework by separately assigning survival to the subdiagonal component and migration to the diagonal component of the Leslie matrix. We are able to estimate these change ratios separately by using a quadratic-optimization fitting procedure which also incorporates prior information on fertility and mortality via constraints. Cross-validation shows that our method has comparable accuracy for age and sex specific forecasting with other, more complex and time intensive, methods. Our method shares the simplicity of CCR approaches in that it only requires the population at time zero and time one, with the addition of fertility and mortality rates, rates which can be found easily for large geographies in the developed world or borrowed for regions where they are not available.