Population Forecast Errors

Abstract

Projections of future populations are integral to many planning applications, yet are often poorly understood. This analysis focuses on the implications of the choices planners make when they construct projections. Specifically, it examines the impact of length of base period, analyzes the error structure of projection techniques for counties in the aggregate and by size and growth rates, investigates the role of averaging, and compares the performance of trend extrapolation and cohort—component methods. The article concludes by discussing forecast complexity, data quality, the role of assumptions, and other considerations of forecasting in a planning context.