Alternative Projections of the U.S. population

The U.S. Bureau of the Census recently released a set of population projections that include middle and high projections that we argue are too conservative. The projections discount the possibility of future baby booms and assume slow rates of mortality decline and low levels of immigration. In this article we explore the impact on the size and age composition of the U.S. population of alternative scenarios of plausible fertility, mortality, and immigration assumptions. We conclude that (1) the Census Bureau's highest projection might be interpreted as a reasonable middle projection, (2) a reasonable high projection would yield a U.S. population in 2080 some 300 million persons larger than the Bureau's highest projection, with the population 85 and older more than twice the Bureau's greatest estimate, and (3) uncertainty about the pace of population growth is substantially greater than the Bureau's projections suggest.     

Some aspects of the use of birth expectations data from sample surveys for population projections

The value of national sample survey data relating to birth expectations for projecting births is reassessed in the light of data limitations pointed out by Ryder and Westoff among others and of the methods of projection used by the United States Bureau of the Census. The annual level of fertility under the cohort-fertility projection method depends on the assumptions regarding completed fertility, about which the available survey data are fairly informative, and on the assumptions regarding the timing of births, about which the survey data tell us very little. Test calculations suggest that Ryder and Westoff have overstated the significance of timing relative to completed family size for the level of future births. We believe that the fall in the annual total fertility rate in the first half of the sixties is to be explained only in part by a general delay in childbearing; a moderate to substantial decrease in completed family size has also occurred. Analysis of the latest set of fertility projections of the Census Bureau also suggests that the assumptions about completed fertility are a much more important determinant of the level of future births than timing, both in the short and long term. Although the available expectations data cannot help in predicting short-term annual changes in fertility, they appear useful for making long-term projections of annual fertility. Expansion of the size, frequency, and content of the sample surveys and incorporation of parity and birth interval into the projection method may improve projections.     

Ethnic population projections for the UK, 2001–2051

This paper reports on projections of the United Kingdom’s ethnic group populations for 2001–2051. For the years 2001–2007 we estimate fertility rates, survival probabilities, internal migration probabilities and international migration flows for 16 ethnic groups and 355 UK areas. We make assumptions about future component rates, probabilities and flows and feed these into our projection model. This model is a cohort-component model specified for single years of age to 100+. To handle this large state space, we employed a bi-regional model. We implement four projections: (1) a benchmark projection that uses the component inputs for 2001; (2) a trend projection where assumptions beyond 2007 are adjusted to those in the UK 2008-based National Population Projections (NPP); (3) a projection that modifies the NPP assumptions and (4) a projection that uses a different emigration assumption. The projected UK population ranges between a low of 63 millions in 2051 under the first projection to a high of 79 million in the third projection. Under all projections ethnic composition continues to change: the White British, White Irish and Black Caribbean groups experience the slowest growth and lose population share; the Other White and Mixed groups to experience relative increases in share; South Asian groups grow strongly as do the Chinese and Other Ethnic groups. The ethnic minority share of the population increases from 13% (2001) to 25% in the trend projection but to only 20% under our modified emigration projection. However, what is certain is that the UK can look forward to be becoming a more diverse nation by mid-century.     

Population Scenarios Based on Probabilistic Projections: An Application for the Millennium Ecosystem Assessment

Probabilistic population forecasts offer a number of advantages to users. However, in some cases population is one component of a larger analysis that may take a different approach to uncertainty. For example, integrated assessments of environmental issues such as climate change or ecosystem degradation have typically used a small number of alternative scenarios to explore uncertainty in future environmental outcomes. In such cases, population projections that are provided only as probability distributions are difficult to use. I present a method of employing probabilistic population projections to derive individual, deterministic projections that can be used within scenarios for integrated assessments. The principal advantages of this approach are that (1) it provides a less ad hoc way of defining deterministic projections intended to be consistent with more comprehensive scenarios that describe, among other things, future socio-economic developments; (2) it provides more flexibility in specifying input assumptions for deterministic projections as compared to choosing off-the-shelf projections, allowing population assumptions to be tailored to the scenario; and (3) it provides a quantitative assessment of the uncertainty associated with any given deterministic projection. I describe the application of the method to the development of population projections used in integrated scenarios for the Millennium Ecosystem Assessment, an international scientific effort to assess the current conditions of and future outlook for global ecosystem goods and services. Results show that the MA scenarios are each consistent with a relatively wide range of demographic outcomes. For some scenarios, ranges of plausible outcomes in some regions overlap substantially, indicating that particular population projections could be consistent with more than one scenario. In other cases, uncertainty ranges for different scenarios are distinct, indicating that a projection consistent with one scenario is unlikely to be also consistent with another. Comparing variances of the conditional projections also provides insight into how much different storylines constrain future demographic developments. The development of the MA projections points to important areas of future research on correlations among demographic rates and on uncertainty across scales. It also serves as an illustration of how probabilistic and alternative scenario-based approaches to uncertainty can be combined within a single integrated analysis.This revised version was published online in April 2005 with corrections to figures 1-3.     

Population projections: states, 1995 - 2025

All states will have more people in the future, especially in the south and west, while population aging occurs as the baby boomers age. This report identifies population changes projected to affect the US's 50 states and District of Columbia during 1995-2015. Basic assumptions for state population projections are presented with regard to population, births, deaths, net international migration, and net internal migration. The methodology used to produce the report is also described. Total population and net change is presented in tabular format for each state over the period. These data are used as the basic input to many federal, state, and local projection models which produce detailed statistics on education, economic factors, labor force, health care, voting, and other subjects. State differentials in fertility and mortality are also projected to widen, reflecting the concentration of race and ethnic groups with high fertility in some states and differential migration patterns.     

Developing strategies for deriving small population fertility rates

Many agencies require population estimates and projections by ethnic group. These projections need ethnic-specific, age-specific fertility rates (ASFRs) but their inclusion is challenging since ethnicity is not recorded at birth registration. In this paper maternity data are used in a case study of electoral wards in Bradford, West Yorkshire, to develop fertility rates for small populations for a 1991 based projection. The challenge is to capture local variations in fertility by ethnic group when data are sparse. Small areas were grouped together using cluster analysis to define combinations with similar sociodemographic and fertility experiences so that sparse data could be aggregated to estimate reliable ethnic-specific fertility rates. For comparison, the data were aggregated into the 1991 Office for National Statistics area type classification. Fertility rates by single year of age for all area types were smoothed using the Hadwiger function. For the White ethnic group there were sufficient births to create ethnic-specific, ward-level ASFRs. For other ethnicities grouping of areas was necessary. The accuracy of the ASFRs in predicting births was assessed using mean absolute percentage error. Results show that for some minority groups district-level ethnic-specific fertility rates produced the most accurate birth estimates even though they were based on a larger area. This implies that rates created may be informative about the local area for White ethnic type but not in the same way for smaller ethnic groups. In terms of grouping strategies we recommend that existing classifications are assessed to determine how well variations in rates are stratified before embarking on a custom scheme. Where population sub-groups are small in some areas, it may be more reliable to use rates derived for larger areas and apply these to local populations. Inevitably, the rates used in a projection are a compromise but hopefully will still capture important dimensions of population change.     

Data quality and accuracy of United Nations population projections, 1950-95

Between 1951 and 1998, the United Nations (UN) published 16 sets of population projections for the world, its major regions, and countries. This paper reports the accuracy of the projection results. I analyse the quality of the historical data used for the base populations of the projections, and for extrapolating fertility and mortality. I study also the impact this quality has had on the accuracy of the projection results. Results and assumptions for the sets of projections are compared with corresponding estimates from the UN 1998 Revision for total fertility and life expectancy at birth, total population, and the projected age structures. The report covers seven major regions (Africa, Asia, the former USSR, Europe, Northern America, Latin America, and Oceania) and the largest ten countries of the world as of 1998 (China, India, the USA, Indonesia, Brazil, Pakistan, Russia, Japan, Bangladesh, and Nigeria).     

Population,projections, and policy: A cautionary perspective

Population projections depend on censuses, vital statistics and sample surveys, all of which have deficiencies that are most marked in the less developed countries (LDCs). Long-range projections by international agencies have recently undergone major revisions, while forecasts of the U.S. population have changed drastically over the past four years. The United Nations typically prepares “high,” “medium” and “low” projections. Even the high projection contains optimistic assumptions about fertility decline, while assumptions of constant or increasing fertility receive no serious attention. The paper suggests that high and constant fertility projections should receive more attention from policy makers, with medium estimates treated more as targets achievable only through considerable programmatic effort. At the same time, economic and social plans should be laid for dealing with the population sizes implied by the “high” variants.     

Comparison of microcomputer programs for making population projections: an update

"There are many programs for making population projections now available for use with microcomputers. This article reviews six of approximately 15 microcomputer population projection programs. Each program is compared to a standard set of criteria relating to such items as hardware and software requirements, input data requirements and specification of assumptions, methodology and documentation, and summary output indicators. Numerical results from projections of six test data sets reflecting different assumptions about mortality, fertility, and migration are compared. Qualitative comments are included for describing special features and for making an overall assessment of each program." (SUMMARY IN FRE)     

The United Nations on World Population in 2300

The Population Division of the United Nations biennially issues detailed population estimates and projections covering the period 1950–2050. The most recent revision of these estimates and projections, the 2002 assessment, was released in February 2003. At irregular intervals, the Population Division also publishes long‐range projections. The most recent of these, covering the period up to 2150, was issued in 2000, based on the 1998 assessment. On 9 December 2003, the Population Division released the preliminary report on a new set of long‐range projections, dovetailing with the 2002 assessment, that extend over a much longer time span: up to 2300 ( http://www.un.org/esa/population/publications/longrange2/longrange2.htm ). Unlike previous long‐range projections, which, apart from China and In‐dia, were prepared for large regional groupings only, the new projections are elaborated separately for 192 countries. Given the enormous uncertainties of the character of demographic trends over such an extended period, the information content of these projections is somewhat elusive. However, they are expected to be used to provide the demographic input for long‐range models of global climate change. Long‐range population projections also serve to demonstrate the unsustainability of certain seemingly plausible assumptions as to the future course of particular demographic parameters. In the present case, for example, the high‐fertility projection, reflecting a sustained total fertility rate at the relatively modest level of 2.35, by 2300 would yield a population of some 32 billion in the countries now classified as less developed. Or, in a yet more extreme exercise 0/reductio ad absurdum, maintaining constant fertility at present rates would result in a population size of some 120 trillion in the countries now classified as least developed. Apart from the “high fertility” and “constant fertility” models just cited, the projections are calculated for three additional instructive variants: “low fertility,”“medium fertility,” and “zero growth.” Underlying each of the five variants is a single assumption on mortality change: expectation of life at birth creeping up, country‐by‐country, to a 2300 level ranging between 88 and 106 years. International migration is set at zero throughout the period 2050‐2300 in each variant. Thus the projections are unabashedly stylized and surprise‐free, providing a simple demonstration of the consequences, in terms of population size and age structure, of clearly stated assumptions on the future course of demographic variables. Reproduced below is the Executive Summary of the preliminary report on the UN long‐range projections presented to a UN technical working group on long‐range projections at its December 2003 meeting in New York and slightly revised afterward. A full final report on this topic by the Population Division of the Department of Economic and Social Affairs of the United Nations Secretariat will be published later in 2004.     

Long‐Range Population Projections Made Simple

Recent developments in mathematical demography offer a new, simple means of producing long‐range population projections. The well‐known extant such projections, produced by the United Nations, the World Bank, and the International Institute for Applied Systems Analysis, rely on elaborate cohort‐component projection methods that require a large number of detailed assumptions and are difficult to replicate. Building upon recent results in the formal demography of nonstable populations, the authors show that analytic methods produce estimates of future population size very similar to those obtained through traditional methods. Simplicity is a virtue in making projections, allowing sensitivity tests of assumptions and avoiding the misleading impression of precision associated with more complicated methods. Cohort‐component methods should still be used for short‐ and medium‐term forecasts and projections. For the long term, however, analytic methods should supplement or even replace traditional projections.     

Comparison of microcomputer programs for making population projections: An update

There are many programs for making population projections now available for use with microcomputers. This article reviews six of approximately 15 microcomputer population projection programs. Each program is compared to a standard set of criteria relating to such items as hardware and software requirements, input data requirements and specification of assumptions, methodology and documentation, and summary output indicators. Numerical results from projections of six test data sets reflecting different assumptions about mortality, fertility, and migration are compared. Qualitative comments are included for describing special features and for making an overall assessment of each program.     

The future of Australia’s Indigenous Population, 2011–61

Existing projections of Australia’s Indigenous Population suffer from a number of limitations: problematic input data, unsatisfactory projection model design, and poor forecast performance. The aim of this study was to create a new model for projecting that population that better represents the demographic processes at work, and that makes use of a newly available data source on identification change. A new projection model is presented that explicitly incorporates ethnic-identification change, and mixed (Indigenous/Non-Indigenous) partnering and childbearing. It is a composite static–dynamic model which takes a multi-state form where data allow. The model was used to produce projections for the 2011–61 period. Rapid growth of the Indigenous Population is expected, with population momentum, identification change, and mixed partnering and childbearing shown to contribute more to growth than above-replacement fertility and increasing life expectancy. The future growth of Australia’s Indigenous Population is thus intimately connected to its interaction with the Non-Indigenous Population.     

Probabilistic Projections of the Total Fertility Rate for All Countries

We describe a Bayesian projection model to produce country-specific projections of the total fertility rate (TFR) for all countries. The model decomposes the evolution of TFR into three phases: pre-transition high fertility, the fertility transition, and post-transition low fertility. The model for the fertility decline builds on the United Nations Population Division’s current deterministic projection methodology, which assumes that fertility will eventually fall below replacement level. It models the decline in TFR as the sum of two logistic functions that depend on the current TFR level, and a random term. A Bayesian hierarchical model is used to project future TFR based on both the country’s TFR history and the pattern of all countries. It is estimated from United Nations estimates of past TFR in all countries using a Markov chain Monte Carlo algorithm. The post-transition low fertility phase is modeled using an autoregressive model, in which long-term TFR projections converge toward and oscillate around replacement level. The method is evaluated using out-of-sample projections for the period since 1980 and the period since 1995, and is found to be well calibrated.     

Probabilistic population projections for countries with generalized HIV/AIDS epidemics

In 2015, the United Nations (UN) issued probabilistic population projections for all countries up to 2100, by simulating future levels of total fertility and life expectancy and combining the results using a standard cohort component projection method. For the 40 countries with generalized HIV/AIDS epidemics, the mortality projections used the Spectrum/Estimation and Projection Package (EPP) model, a complex, multistate model designed for short-term projections of policy-relevant quantities for the epidemic. We propose a simpler approach that is more compatible with existing UN projection methods for other countries. Changes in life expectancy are projected probabilistically using a simple time series regression and then converted to age- and sex-specific mortality rates using model life tables designed for countries with HIV/AIDS epidemics. These are then input to the cohort component method, as for other countries. The method performed well in an out-of-sample cross-validation experiment. It gives similar short-run projections to Spectrum/EPP, while being simpler and avoiding multistate modelling.     

Demographic scenarios by age,sex and education corresponding to the SSP narratives

In this paper, we translate the five narratives as defined by the Shared Socioeconomic Pathways (SSPs) research community into five alternative demographic scenarios using projections by age, sex and level of education for 171 countries up to 2100. The scenarios represent a significant step beyond past population scenarios used in the Intergovernmental Panel for Climate Change context, which considered only population size. The definitions of the medium assumptions about future fertility, mortality, migration and education trends are taken from a major new projections effort by the Wittgenstein Centre for Demography and Global Human Capital, while the assumptions for all the other scenarios were defined in interactions with other groups in the SSP community. Since a full data base with all country-specific results is available online, this paper can only highlight selected results.     

Estimation of local demographic variation in a flexible framework for population projections

A cohort component projection of local populations based on sex and single year of age offers great value for planning local services, but demands data beyond the detail available. Local fertility, mortality and migration schedules by age and sex must be estimated sensitively to local variation if the results are to be of greater value than simpler methods of projection. Two approaches are compared, using data for the recent past: (a) direct estimation of local area age-specific schedules of fertility, mortality and migration based on data available to the national statistical agency; (b) graduation of national schedules using only local area population estimates by age, total numbers of births, and total numbers of deaths; age-specific migration is indirectly estimated from successive population estimates. These two approaches are compared with a projection using the same rates for each area. The three projections have been implemented for electoral wards in the Fife local government area of Scotland, using the flexible framework provided by POPGROUP software. Persuasive local population projections based on standard data for standard areas are feasible without the regular publication of migration flows.     

Comparative empirical evaluations of internal migration models in subnational population projections

While population forecasters place considerable emphasis on the selection of appropriate migration assumptions, surprisingly little attention has been given to the effects on projection outcomes of the way internal migration is handled within population projection models. This paper compares population projections for Australia's states and territories prepared using ten different internal migration models but with identical assumptions for fertility, mortality and international migration and with the internal migration model parameters held constant. It is shown that the choice of migration model generates large differences in total population, geographical distribution and age-sex composition. It is argued that model choice should be guided by balancing model reality with practical utility and model performance is examined against these criteria. Of the ten models evaluated the authors argue that the migration pool, biregional, and biregional with net constraints models offer a good compromise between conceptual rigour and practicality. If the projected origin-destination flows are required then one of the versions of the standard multiregional model with reduced data inputs is preferred. The large variation in projection outputs points to the need for a better understanding of the spatio-temporal structure of migration in Australia.     

The Census Bureau on Prospects for US Population Growth in the Twenty‐First Century

The U.S. Census Bureau periodically releases projections of the US resident population, detailed by age, sex, race, and Hispanic origin. The most recent of these, issued 13 January 2000, for the first time extend to the year 2100 and also include information on the foreign‐bom population. (Earlier projections were carried up to 2080.) The extensive tabulations presenting the new set, and detailed explanation of the methodology and the assumptions underlying the projections, are accessible at the Census Bureau's web site: http://www.census.gov . A brief summary of some of the main results of these projections is reproduced below from United States Department of Commerce News, Washington, DC 20230. (The Census Bureau is an agency of the Department of Commerce.) Uncertainties as to future trends in fertility, mortality, and net migration over a period of some 100 years are very great, as is illustrated by the massive difference in the projected size of the population for 2100 in the three variants produced. The “middle” projected population figure of 571 million (which represents a growth of some 109 percent over its current level) is bracketed by “lowest” and “highest” alternative projections of 283 million and 1.18 billion, respectively. With somewhat lesser force, the point also applies to the 50‐year time span considered in the well‐known country‐by‐country projections of the United Nations. These projections are also detailed in three variants: low, middle, and high. The UN projections (last revised in 1998) envisage less rapid growth in the United States during the first part of the twenty‐first century than do the Census Bureau's. The projected population figures for 2050 in the three variants (low, middle, and high) are as follows (in millions):

Prospects for fertility reduction and projections for future population growth in Kenya

In 1979 Kenya's annual rate of natural population growth was 3.8%. Data from the1989 Kenya Demographic and Health Survey indicate that significant decreases in fertility levels were experienced during the 1980s. Factors associated with conditions supportive of high fertility in Kenya are discussed, and progress toward attaining significant fertility reduction thresholds during the 1980s is assessed. Findings from recent fertility surveys are presented, and 1969–1989 national level family planning data are evaluated. Four population projections for 1985–2025 are presented and analyzed. One projection is based on official government growth targets; two are based on estimates provided by the United Nations and the Population Reference Bureau, and a fourth projection is based on the assumption that Kenya will attain an annual natural population growth rate of less than 1% by the year 2025. Each projection assumes that fertility declines will be experienced. Kenya's prospects for reducing the annual population growth rate to 1% within the next sixty years and a cost-sharing development policy are addressed briefly in the concluding section. Recent data suggest that Kenya will probably not complete the demographic transition before the year 2050, but Kenya should continue to move through the transition stage.