An Evaluation of State Population Projections by Age

Stanley K. Smith, University of Florida
Jeff Tayman, San Diego Association of Governments (SANDAG)

Many studies have evaluated the forecast accuracy of projections of total population size, but few have considered the accuracy of projections by age group. For many policy purposes, however, the relevant variable is a specific age group rather than the total population. Examples include children for projecting school enrollment, women of childbearing age for projecting births, and the elderly for projecting Social Security expenditures. In this paper we evaluate the accuracy of several sets of age-group projections for states produced by the U.S. Census Bureau during the 1970s, 1980s, and 1990s. We compare the accuracy of these projections with the accuracy of a simpler version of the cohort-component method (Hamilton-Perry). We close with several conclusions regarding the accuracy and bias of age-group projections and the potential costs and benefits of using simpler projection methods.

Presented in Session 61: Estimates and Projections for Small Areas