The overall objective of the proposed research is to develop and test a new methodology for projecting future trends in adult mortality by age and sex. This new procedure corrects shortcomings in existing forecasting methods, including the widely used Lee-Carter model. More accurate projections of adult mortality are essential to improving projections of the population size of the oldest age groups, and of the future cost of public pension and health care systems.
Specific aims are as follows: 1) Complete testing the goodness-of-fit of the logistic model for the age pattern of the force of mortality. Preliminary tests used country-specific data from the Human Mortality Data Bank for females and males ages 25-109 in 14 populations and demonstrated that 2 variants of the logistic models fit well for the period 1950-2000. These tests will be extended to the pre-1950 period in 8 countries for which such data are available. 2) Test the goodness-of-fit of the general shifting model. This model uses only 2 parameters to describe adult mortality change over time: 1 for the amount of the age shift in senescent mortality and the other for the level of background mortality. This approach introduces an alternative way of thinking about senescent mortality change. The conventional view that mortality change implies increases or decreases in age-specific mortality rates is replaced with the view that considers mortality change to be the result of changes in the timing of death. The model is expected to provide a concise mathematical description of past trends. 3) Develop a new methodology for projecting age-sex specific adult mortality rates over time based on the general shifting mortality model. 4) Test the new projection method by comparing its results with projections produced by the Lee-Carter procedure. ? ?
