This article deals with the forecast accuracy and bias of population projections for 2,971 counties in the United States. It uses three different population projection techniques and data from 1950, 1960,1970, and 1980 to make two sets of 10-year projections and one set of 20-year projections. These projections are compared with census counts to determine forecast errors. The size, direction, and distribution of forecast errors are analyzed by size of place, rate of growth, and length of projection horizon.
The housing unit (HU) method is used by public and private agencies throughout the United States to make local population estimates. This article describes many of the different types of data and techniques that can be used in applying the HU method, and it discusses the strengths and weaknesses of each. Empirical evidence from four different states is provided, comparing the accuracy of HU population estimates with the accuracy of other commonly used estimation techniques. Several conclusions are drawn regarding the usefulness of the HU method for local population estimation.
The housing unit (HU) method is often characterized as inferior to other methods for estimating the population of states and local areas. We believe this characterization must be challenged. In this article we evaluate population estimates produced by the housing unit method and by three other commonly used methods: component 11, ratio correlation, and administrative records.
State populations in the United States are characterized by large differences in current growth rates and historical growth trends. What demographic factors account for these differences? Population growth has only three components: births, deaths, and migration. In this study, we estimated the contributions of births, deaths, and migration to changes in population size between 1950 and 1980 for the 48 contiguous states in the United States.
Many different techniques can be used for making population projections. Most fall into four general categories: trend extrapolation, ratio extrapolation, cohort-component and structural. Techniques within these categories differ considerably in terms of their complexity and sophistication. A common perception among producers (and users) of population projections is that complex and/or sophisticated techniques produce more accurate forecasts than simple and/or naive techniques.
This book chapter starts by discussing the major producers of international, national, and subnational projections. Next, it provide a description of the methods and materials used in preparing three basic types of population projections: (1) trend extrapolation; (2) the cohort-component method; and (3) structural modeling. The authors briefly discuss methods for preparing related projections on such topics as school enrollment, employment, and households.
Developments in economic theory over the last 20 years have placed decisions regarding female labor force participation (FLFP) and fertility within a model of household decision making (HDM). In this one period, static model utility is a function of child services (including both number and quality of children), market goods and services, and leisure. At the outset of their marriage a husband and wife adopt a utility-maximizing lifetime plan of fertility, market work, nonmarket activities, and consumption of goods and services, subject to income and time constraints.
Many studies have found that population forecast errors generally increase with the length of the forecast horizon, but none have examined this relationship in detail. Do errors grow linearly, exponentially, or in some other manner as the forecast horizon becomes longer? Does the error-horizon relationship differ by forecasting technique, launch year, size of place, or rate of growth? Do alternative measures of error make a difference? In this article we address these questions using two simple forecasting techniques and population data from 1900 to 1980 for states in the United States.
A number of studies in recent years have investigated elnpirical approaches to the production of confidence intervals for population projections. The critical assumption underlying these approaches is that the distribution of forecast errors remains stable over time. In this article, we evaluate this assumption by making population projections for states for a number of tilne periods during the 20th century, comparing these projections with census enumerations to deter~nine forecast errors, and analyzing the stability of the resulting error distributions over time.
This article develops three different models of migration for cohort-component projections, each using a different base (i.e., denominator) for migration rates. The differences in the resulting projections are analyzed, and a number of conclusions are drawn regarding the construction of migration rates for use in cohort-component population projections.