March 2007

Table of Contents

LIST OF FIGURES.............................................................................................................VI

LIST OF TABLES ...............................................................................................................VI

INTRODUCTION .............................................................................................................. 1

A. Description of the Spectrum System ............................................................. 1

1. Components ............................................................................................... 1
2. Software Description ................................................................................. 2

B. Uses of Spectrum Policy Models .................................................................... 2

C. Organization of the Model Manuals ............................................................. 4

D. Information about the POLICY Project ....................................................... 5

E. What Is DemProj?............................................................................................. 6

F. Why Make Population Projections? ............................................................... 6

STEPS IN MAKING A POPULATION PROJECTION.......................................................... 8

DEMOGRAPHIC INPUTS AND OUTPUTS ....................................................................... 10

A. Base Year Population ................................................................................... 11

B. Fertility............................................................................................................. 12
1. The Total Fertility Rate .............................................................................. 12
2. The Age Distribution of Fertility ............................................................... 15

C. Mortality......................................................................................................... 26
1. Life Expectancy at Birth .......................................................................... 26
2. Life Expectancy and AIDS ...................................................................... 28
3. Age-Specific Mortality............................................................................. 29

D. Migration........................................................................................................ 35

E. Current Population: Urban and Rural .......................................................... 36

iii
 F. Projection Outputs ......................................................................................... 37

PROGRAM TUTORIAL.................................................................................................... 41

A. Before You Get Started................................................................................. 41

B. Installing the Spectrum Program.................................................................. 42

C. Creating a New Projection .......................................................................... 42

1. Starting the Spectrum Program ............................................................ 42
2. Specifying the Projection Parameters in Projection Manager ........ 42

D. Entering the Base Year Data and Projection Assumptions ....................... 44

1. Specifying New Demographic Projection Parameters...................... 44
2. Entering the Base Year Data and Projection Assumptions................ 47
3. About the Editors...................................................................................... 48
4. Leaving the Demographic Data Editors .............................................. 60
5. Saving the Input Data ............................................................................. 60

E. Making the Projection................................................................................... 60

F. Examining the Output.................................................................................... 61

1. Graphs and Bar Charts............................................................................ 65
2. Tables ......................................................................................................... 65
3. Displaying All Age Groups ...................................................................... 65
4. Demographic Summary Table............................................................... 66

G. Saving the Projection ................................................................................... 66

H. Opening an Existing Projection ................................................................... 66

I. Closing a Projection ....................................................................................... 67

L. Aggregating projections........................................................................... 67

METHODOLOGY........................................................................................................... 70

A. Projecting the population by single ages ............................................... 70

B. Urban and Rural Projections ......................................................................... 75

REFERENCES.................................................................................................................. 77

GLOSSARY OF TERMS................................................................................................... 79

iv
ACRONYMS AND ABBREVIATIONS ............................................................................. 85

APPENDIX A ................................................................................................................... 1

THE COALE-TRUSSELL RELATIONAL FERTILITY MODEL ................................................... 1

APPENDIX A: THE COALE-TRUSSELL RELATIONAL FERTILITY MODEL............................ 1

A.1 Description of the Model.............................................................................. 1

A.2 Employment of the Model in Projections.................................................... 3

A.3 An Application to Bangladesh .................................................................... 5

v
List of Figures
Figure 1: Shift in Fertility Distributions - Bangladesh: 1975 to 1993-94 ....................................................... 17
Figure 2: Shift in Fertility Distributions - Taiwan: 1956 to 1983 ..................................................................... 17
Figure 3: Fertility Distribution by TFR: UN Sub-Saharan Africa Pattern ...................................................... 20
Figure 4: Comparison of Model Fertility Tables at a TFR of 4.0................................................................... 21
Figure 5: Shift in Fertility Distributions - Bangladesh and UN Asia Pattern................................................. 22
Figure 6: Shift in Fertility Distributions - Bangladesh and UN Sub-Saharan Africa Pattern ..................... 22
Figure 7: ASFRs Generated by Coale-Trussell Model Compared with Survey ASFRs:
Bangladesh, BFS 1975 ...................................................................................................................... 25
Figure 8: ASFRs Generated by Coale-Trussell Model Compared with Survey ASFRs:
Bangladesh, BFS 1989 ...................................................................................................................... 25
Figure 9: Model Life Table Mortality Probabilities - Coale-Demeny West Female Models Table:
Compared by Level of Life Expectancy ...................................................................................... 31
Figure 10: Model Life Table Mortality Probabilities - Coale-Demeny Female Model Table
Families: Comparison by Family at e(0) = 45 .............................................................................. 32
Figure 11: Model Life Table Mortality Probabilities - United Nations Female Model Table
Families: Comparison by Family at e(0) = 45 .............................................................................. 33
Figure 12: Typical Pattern of Urbanization ..................................................................................................... 37
Figure 13: Fertility Distribution by TFR: UN Sub-Saharan Africa Pattern ..................................................... 73
Figure 14: Fertility Distribution by TFR: UN Arab Pattern............................................................................... 73
Figure 15: Fertility Distribution by TFR: UN Asia Pattern ................................................................................ 74
Figure A.1: ASFRs Generated by Coale-Trussell Model Compared with Survey ASFRs:
Bangladesh, BFS 1975................................................................................................................. A-6
Figure A.2: ASFRs Generated by Coale-Trussell Model Compared with Survey ASFRs:
Bangladesh, BFS 1989................................................................................................................. A-6

List of Tables
Table 1: Declines in TFR from 1975 to 1990 by Level of Program Effort During 1982-1989 and
Socioeconomic Setting in 1985 ....................................................................................................... 14
Table 2. Calculation of the Distribution of Fertility by Age Group from Age-Specific
Fertility Rates ....................................................................................................................................... 16
Table 3: UN Model Tables of the Age Distribution of Fertility ...................................................................... 19
Table 4: UN Working Model of Life Expectancy Improvement During a Five-Year Period .................... 26
Table A.1: Coale-Trussell Fertility Control Schedule (vx)............................................................................. A-4
Table A.2: Bangladesh: ASFRs from Various Surveys, and Projected Rates ........................................... A-7

vi
 Introduction

I. A. Description of the Spectrum System

1. Components
USAID|Health Policy Initiative (HPI) and its
predecessor projects have developed computer models
that analyze existing information to determine the future
consequences of today’s development programs and
policies. The Spectrum Policy Modeling System consolidates
previous models into an integrated package containing the
following components:

• Demography (DemProj) – A program to make

population projections based on (1) the current
population, and (2) fertility, mortality, and migration
rates for a country or region.
Spectrum consolidates • Family Planning (FamPlan) – A program to project family
DemProj, FamPlan, Benefit- planning requirements in order to achieve national goals
Cost, AIM, and RAPID for meeting couples’ fertility intentions.
models into an integrated
package. • Benefit-Cost – A program for comparing the costs of
implementing family planning programs, to the benefits
generated by those programs.
• AIDS (AIDS Impact Model – AIM) – A program to project
the consequences of the AIDS epidemic including the
number of people infected with HIV, AIDS deaths, the
number of people needing treatment and the number of
orphans.
• Socioeconomic Impacts of High Fertility and Population
Growth (RAPID) – A program to project the social and
economic consequences of high fertility and rapid
population growth for sectors such as labor force,
education, health, urbanization and agriculture.
• Adolescent reproductive health (NewGen) – A program
to estimate the consequences of adolscent
reproductive heatlh behavior, including sexual debut,
marriage, pregnancy, abortion, and HIV/STI infection.
• Prvention of mother-to-child transmission (PMTCT) – A
program to examine the costs and benefits of
interventions to prevent the transmission of HIV from

1
 mother to child, including anti-retroviral treatment,
infant feeding options and delivery options.

2. Software Description
Spectrum is a Windows-based system of integrated policy
models. The integration is based on DemProj, which is used
to create the population projections that support many of
the calculations in the other components, such as FamPlan,
Benefit-Cost, AIM, and RAPID.

Each component has a similarly functioning interface that is

easy to learn and to use. With little guidance, anyone who
has a basic familiarity with Windows software will be able to
navigate the models to create population projections and
to estimate resource and infrastructure requirements. The
accompanying manuals contain both the instructions for
users, and equations for those who want to know exactly
how the underlying calculations are computed.

B. Uses of Spectrum Policy Models

Policy models are designed Policy models are designed to answer a number of “what if”
to answer a number of questions relevant to entities as small as local providers of
“what if” questions. The
primary health care services and as large as international
“what if” refers to factors
that can be changed or development assistance agencies. The “what if” refers to
influenced by public policy. factors that can be changed or influenced by public policy.
Models are commonly computerized when analysts need to
see the likely result of two or more forces that might be
brought to bear on an outcome, such as a population’s
illness level or its degree of urbanization. Whenever at least
three variables are involved (such as two forces and one
outcome), a computerized model can both reduce the
burden of manipulating those variables and present the
results in an accessible way.
Some of the policy issues commonly addressed by the
Spectrum set of models include:
• The utility of taking actions earlier rather than later.
Modeling shows that little in a country stands still while
policy decisions are stalled and that many negative
outcomes can accumulate during a period of policy
stasis.

2
• The evaluation of the costs vs. the benefits of a course
of action. Modeling can show the economic efficiency
of a set of actions (i.e., whether certain outcomes are
achieved more effectively than under a different set of
actions), or simply whether the cost of a single set of
actions is acceptable for the benefits gained.
• The recognition of inter-relatedness. Modeling can
show how making a change in one area of population
dynamics (such as migration rates) may necessitate
changes in a number of other areas (such as marriage
rates, timing of childbearing, etc.).

3
 • The need to discard monolithic explanations and policy
initiatives. Modeling can demonstrate that simplistic
explanations may bear little relationship to how the
“real world” operates.
A set of policies under • The utility of “door openers.” A set of policies under
consideration may not be consideration may not be acceptable to all stakeholders.
acceptable to all
Modeling can concentrate on favored goals and
stakeholders.
objectives and demonstrate how they are assisted by the
proposed policies.
• That few things in life operate in a linear fashion. A
straight line rarely describes social or physical behavior.
Most particularly, population growth, being
exponential, is so far from linear that its results are
startling. Modeling shows that all social sectors based
on the size of population groups are heavily influenced
by the exponential nature of growth over time.
• That a population’s composition greatly influences its
needs and its well being. How a population is
composed—in terms of its age and sex distribution—has
broad-ranging consequences for social welfare, crime
rates, disease transmission, political stability, etc.
Modeling demonstrates the degree to which a change
in age and sex distribution can affect a range of social
indicators.
• The effort required to “swim against the current.” A
number of factors can make the success of a particular
program harder to achieve; for example, the waning of
breastfeeding in a population increases the need for
contraceptive coverage. Modeling can illustrate the
need for extra effort—even if simply to keep running in
place.

C. Organization of the Model Manuals

Each manual begins with a discussion of what the model
does and why someone would want to use it. The manual
also explains the data decisions and assumptions needed
before the model can be run, and possible sources for the
data. It defines the data inputs and outputs. The manual
contains a tutorial, information on the methodology behind
the model, a glossary, and a bibliography.

4
 D. Information about the POLICY Project
The POLICY Project is a USAID-funded activity designed to
create a supportive environment for family planning and
reproductive health programs through the promotion of a
participatory process and population policies that respond
to client needs. To achieve its purpose, the project
addresses the full range of policies that support the
expansion of family planning and other reproductive health
services, including:

• national policies as expressed in laws and in official

statements and documents;
• operational policies that govern the provision of
services;
• policies affecting gender roles and the status of
women; and
• policies in related sectors, such as health, education
and the environment that affect populations.

USAID|Health Policy More information about the Spectrum System of Policy Models
Initiative Task Order 1 is and HPI is available from:
implemented byConstella
Futures Group in Director, HPI
collaboration with Constella Futures
Futures Institute, the
One Thomas Circle, , NW Suite 200
Centre for Development
and Population Activities Washington, DC 20005 U.S.A.
(CEDPA), Cultural Telephone: (202) 775-9680
Practices LLC, Social Fax: (202) 775-9694
Sectors Develoment [Link] [Link]
Strategies, Inc. White
Ribbon Alliance and
World Conference of
Religion for Peace.
or
USAID|Health Policy Initiative
U.S. Agency for International Development
Center for Population, Health, and Nutrition
1300 Pennsylvania Ave.
Washington, DC 20523 U.S.A.
Telephone: (202) 712-5787 or -5839

5
 E. What Is DemProj?
The demographic model in Spectrum, known as DemProj, is
a computer program for making population projections for
countries or regions. The program requires information on
the number of people by age and sex in the base year, as
well as current year data and future assumptions about the
total fertility rate (TFR), the age distribution of fertility, life
expectancy at birth by sex, the most appropriate model life
table, and the magnitude and pattern of international
migration (all of these inputs are discussed in Chapter III).
This information is used to project the size of the future
population by age and sex for as many as 150 years into the
future. If desired, the projection can also estimate the size of
the urban and rural populations. Linking DemProj with other
modules in Spectrum makes it possible to examine the
demographic impact of AIDS (AIM), the family planning
service requirements to achieve demographic and health
goals (FamPlan), the costs and benefits of family planning
programs and the socioeconomic impacts of high fertility
and rapid population growth (RAPID).
DemProj was first produced in 1980. Since then, it has been
used by a large number of planners and researchers around
the world. It has been updated from time to time in
response to comments and suggestions from users. This
current release incorporates a number of new features in
response to these comments.
DemProj is designed to DemProj (and the entire Spectrum system) is designed to
produce information useful produce information useful for policy formulation and dialogue
for policy formulation and
within a framework of easy-to-use computer programs. The
dialogue within a
framework of easy-to-use focus is on the generation of information useful for policy and
computer programs. planning purposes rather than detailed research into the
underlying processes involved. For this reason, the programs are
designed to be used by program planners and policy analysts.
DemProj uses data that are readily available and requires little
demographic experience beyond the information available in
this manual.

F. Why Make Population Projections?

Population projections are useful for a variety of purposes,
most commonly as a basis for planning. For example,
assessing a nation’s or region’s needs for new jobs, teachers,
schools, doctors, nurses, urban housing, or food requires
knowledge of the number of people who will be requiring

6
 services. Thus, population projections serve as the starting
point for most projections of future need.
Population projections are also important for policy
dialogue. A key aspect of the policy process is recognizing
that a problem exists and placing that problem on the
policy agenda. Since rapid population growth contributes
to many of the major development problems, population
projections are required to illustrate the future magnitude of
the problems. For example, a population projection is
required in order to illustrate the future threat to forests from
excessive use of fuel wood.
A discussion of the need for Population projections are required also in areas of the policy
and impact of an expanded process that involve proposals to solve problems. In this case,
program of immunization
population projections are required to show the magnitude of
requires a population
projection to show the the services that will be required for the proposed solution and
number of children who the improvements in population-based indicators that will
will need vaccination in the occur. For example, a discussion of the need for and impact of
future and the reduction in an expanded program of immunization requires a population
morbidity and mortality projection to show the number of children who will need
rates that could be
vaccination in the future and the reduction in morbidity and
achieved by the proposed
program. mortality rates that could be achieved by the proposed
program.
Looking at future population composition and size usually
requires producing several projections, for two reasons. First,
projections are based on assumptions about future levels of
fertility, mortality and migration. Since these are simply
assumptions, it is often wise to consider low, medium and
high variants of each assumption so that the range of
plausible projections can be determined. Second, when
population projections are used for policy dialogue, it is
usually important to show how various rates of population
growth would affect the projections. For example, an
analysis of the impacts of population growth on education
typically would include a high-growth projection to show the
magnitude of the problem under current conditions, and a
low-growth projection to show how a slower-growing
population would put less pressure on the education system
to meet national education goals.

7
II. Steps in Making a
Population Projection

There are seven key steps in making most population

projections. The amount of time spent on each step may
vary, depending on the application, but most projection
activities will include at least these seven steps.
1. Select geographic area. Population projections are
normally made at the national level. However,
projections may also be made for other geographic
areas, such as urban areas, capital cities, provinces,
districts, and catchment areas. For example, the trend
toward decentralization of public programs in many
countries has greatly increased the need for
projections at the district and provincial levels. The first
step in making a population projection is to decide the
most appropriate geographic area for the application.
2. Determine the period of the projection. Population
projections start at some base year and continue for a
certain number of years into the future. The base year
is often selected on the basis of data availability and is
usually the year of the most recent census or large-
scale survey. The number of years to project is
determined by the use of the projection. Planning
activities generally focus on short-term projections (five
years), while projections used for policy dialogue often
use a longer time horizon (10-30 years).
3. Collect data. At a minimum, base year data need to
be collected for the number of people by age and sex,
the TFR, and life expectancy at birth. Since the
population projection will only be as good as the data
on which it is based, it is worth the effort to ensure that
appropriate and high-quality data are collected and
prepared before starting the projection. (A special
feature of DemProj called EasyProj allows you to make
a projection quickly using data from the United Nations
World Population Prospects.)

8
4. Make assumptions. Population projections require
assumptions about the future levels of the total fertility
rate (TFR), life expectancy at birth, and international
migration. Assumptions regarding the most
appropriate model tables (see Sections III B and III C)
for fertility and mortality are also required. These
assumptions should be carefully considered and based
on reasonable selection guidelines.
5. Enter data. Once the base year data are collected
and decisions are made about projection assumptions,
DemProj can be used to enter the data and make a
population projection.
6. Examine projections. Once you make a projection, it
should be examined carefully. This scrutiny includes
consideration of the various demographic indicators
produced, as well as the age and sex distribution of the
projection. Careful examination of these indicators
can act as a check to ensure that the base data and
assumptions were understood and were entered
correctly into the computer program. This careful
examination is also required to ensure that the
consequences of the assumptions are fully understood.
7. Make alternative projections. Many applications
require alternative population projections. Once the
base projection has been made, the program can be
used to generate alternative projections quickly by
varying one or more of the projection assumptions.

9
III. Demographic Inputs and
Outputs

This section covers the DemProj inputs in the order that the
model needs them:

• Base year population by age and sex

• Fertility
• Mortality
• International migration

Overall, these inputs fit into one of the three population

Model schedules are often processes: fertility, mortality, and migration. A population
used to model patterns of projection accommodates these processes by using
mortality and fertility. information on the overall level of each process, and its shape
⎯or age pattern. For the sake of efficiency⎯and quite often
for accuracy⎯the shapes (age patterns) of the three
components are supplied by model schedules.
Demographic research has long observed and collected
patterns of data. Model schedules attempt to capture
these patterns with as few parameters as warranted. Such
model schedules are then used to evaluate the quality of
other demographic data; to smooth schedules of fertility,
nuptiality, migration, or mortality; to provide building blocks
in estimating levels and trends of demographic processes;
and to forecast the behaviors of these processes in a
population. In this last application, DemProj uses model
patterns for two processes, fertility and mortality:
Fertility Mortality
1. TFR 1. Life expectancy at birth
2. Age distribution of fertility 2. Age-specific mortality

These approaches are covered in Sections III B and III C

below.

10
 A. Base Year Population
All population projections must start somewhere. The
starting point is the number of people in the population by
age and sex in the base year. For both males and females,
the population is divided into five-year age groups from 0-4
to 75-79. There is also a final age group for those people
aged 80 and older.
Base year population figures are available from a number of
sources. Usually, the best source will be a national census.
Census reports always include tables showing the size of the
population by age and sex. Often these tables are
available for the national level and for provincial or district
levels.
Several types of Several types of problems can exist in raw census figures,
problems can exist in including underreporting, age misreporting, and undefined
raw census figures,
ages. Census tables should be adjusted to minimize these
including
underreporting, age problems. Underreporting can occur for a variety of reasons.
misreporting, and For example, occasionally certain geographic areas or social
undefined ages. status groups are entirely or partially missed in the census.
Sometimes respondents intentionally misreport in order to avoid
identifying military-age males, marriageable-age females, or
family members that may affect the family’s taxes. Age
misreporting can occur when ages are purposely or
unintentionally reported incorrectly, such as when ages are
rounded to the nearest five-year age group or when ages are
unknown. In most censuses, there will also be a certain
proportion of the population with age undefined. These
These problems should problems should be corrected through standard demographic
be corrected through procedures and reported in later census reports, but in many
standard demographic developing countries these reports may not be done or may
procedures and reported only appear many years after the census is completed.
in later census reports. Therefore, it is important to be sure that corrected census figures
are used whenever possible, rather than raw census tables.
There are other sources of population data if recent census
reports are not available. The Population Division of the United
Nations publishes a considerable amount of population
data. The most useful sources for population projections are
the Demographic Yearbook, which contains the most
recent census data for most countries; and the World
Population Prospects, published every two years and
containing population estimates and projections for most
countries of the world. The latest edition of World Population
Prospects may be a particularly useful source if country data
are unavailable, since it will contain estimates of base year
populations as well as assumptions about future levels of

11
 fertility, mortality and migration. World Population Prospects
contains estimates and projections of population by five-
year age groups that have been adjusted for misreporting.
These data may be used when reliable census data are not
available. Data files prepared using EasyProj (described
below in Section IV.C.2) use data from World Population
Prospects.
The U.S. Census Bureau also publishes a set of population
projections for the countries of the world called World
Population Profile. Other sources, such as the Population
Reference Bureau’s World Population Data Sheet or the
World Bank’s World Development Indicators, list the total
population of most countries of the world but do not provide
age- and sex-specific numbers.
Population estimates may also be accessed from various
Internet sites, including:
Internet sites may provide • the United Nations population projections at:
the latest population [Link]
estimates from several
sources. • the U.S. Census Bureau site at:
[Link]
• the Population Reference Bureau site at:
[Link]

B. Fertility
A population projection requires information about the level
of fertility (obtained through the TFR) and about its shape
(obtained through the age distribution).

1. The Total Fertility Rate

Base Year Estimates

The TFR is the number of live births a woman would have if
she survived to age 50 and had children according to the
prevailing pattern of childbearing at each age group. It is
not an average of the number of live births for currently
living women. Rather, it is a synthetic measure that
expresses the current level of fertility in terms of the average
number of live births that would occur per woman if the
current age-specific fertility rates remained constant and all
women survived to age 50.
Estimates of the TFR are available from a number of sources.
The best sources will be national fertility surveys, which have
been conducted by most countries. A large number have

12
 been conducted under a series of international projects,
including the Demographic and Health Surveys (DHS), the
Centers for Disease Control and Prevention (CDC) Fertility
Surveys, the Contraceptive Prevalence Surveys (CPS) and
the World Fertility Surveys (WFS). Information from these and
other national surveys is collected and reported in a variety
of sources, including summary reports from the DHS, the
Population Reference Bureau’s World Population Data
Sheet, and the World Bank’s World Development Indicators.

Future Assumptions
An assumption about the future TFR is required for most
population projections.1
There are several options for setting the TFR projection.
1. National projections. Many countries have official
population projections that include assumptions about
the future course of their TFRs, often with several
variants. If population projections are being made for
planning purposes, it is often recommended that the
official assumptions and projections be used.
2. National goals. Many countries have national
population goals that often include the TFR. It may be
useful to use these goals as a starting point for
projections. One projection may assume that the TFR
goal is achieved, while others may examine the effects
of a delay in achieving the goal. Sometimes goals are
expressed in terms of crude birth rates, population
growth rates, or contraceptive prevalence rates
instead of TFR. In these cases, different TFR assumptions
may be used to discover a TFR projection that is
consistent with the national goals for these other
indicators.
3. Published projections. The population projections
prepared by the United Nations Population Division and
reported in World Population Prospects include three
assumptions (low, medium and high) about future
fertility for each country included in the report. These
fertility assumptions may be used. The disadvantage to
using these projections is that it is not clear how the
“low,” “medium” and “high” assumptions have been
determined, but the UN generally considers the

1However, if the population projection is combined with a family planning

projection, then the future value of the TFR may be determined by the impact of
the family planning program as determined by the FamPlan module. See
FamPlan’s user’s manual for instructions.

13
 “medium” variant to be the most likely. The U.S.
Census Bureau projections also contain a set of TFR
assumptions.
4. Recent trends and international experience. If
information is available on the TFR for several years, it
may be useful to analyze the trends in TFR and to
develop a future assumption based on continuing past
trends. It should be noted, however, that past trends
cannot be expected to continue for very long into the
future. TFR rarely declines at a constant pace
throughout an entire demographic transition. Rates of
decline are often slow at first, increase during the
middle of the transition, and slow again as they
approach replacement-level fertility.
5. Socioeconomic development and population program
effort. Studies have shown that the pace of fertility
decline is related to the level of socioeconomic
development of a country and the amount of effort
put into the family planning program. These studies are
summarized in Table 1, which shows the decline in the
TFR as a function of these two factors. This experience
can be used to develop realistic assumptions about the
rate at which fertility could decline in the future in any
given country. Table 1 shows that the most rapid
fertility declines experienced between 1978 and 1990
were for countries with strong family planning programs
and high levels of socioeconomic development during
the 1980s.

14
Table 1: Declines in TFR in 1990s by Level of Program Effort During 1994-1999 and
Socioeconomic Setting
Program Effort, 1994-1999
Overall
Socioeconomic VeryWeak/ Averag
Setting Strong Moderate Weak None e
High Mexico 0.37 Chile 0.10 Brazil 0.25 Costa Rica 0.36
Mauritius 0.24 Trinidad 0.45 Uruguay 0.09
Colombia 0.21 and Tobago Venezuela 0.31
Jamaica 0.26 Panama 0.08 Argentina 0.21
Korea, Rep 0.18 Uzbekistan 0.72 Kazakhstan 0.36
Malaysia 0.36 Average 0.34 Average 0.27 High
Average 0.27 0.28

Upper Middle Tunisia 0.81 Dominican 0.28 Ecuador 0.42 Guyana 0.10
Sri Lanka 0.30 Republic Namibia 0.45 Mongolia 0.70
Iran 1.80 Syria 0.79 Lebanon 0.54 Average 0.40
Peru 0.50 Turkey 0.40 Jordan 0.76
Thailand 0.15 Algeria 0.97 Oman 1.05
Average 0.71 Egypt 0.49 Paraguay 0.38
El Salvador 0.35 Nicaragua 0.49 Upper
Philippines 0.50 Honduras 0.62 Middle
South Africa 0.50 Average 0.59 0.58
Average 0.54

Lower Middle India 0.35 Ghana 0.70 Bolivia 0.48 Congo -

Zimbabwe 0.60 Senegal 0.66 Zambia 0.20 Nigeria 0.46
Morocco 0.55 Pakistan 0.35 Guatemala 0.47 Mauritania 0.10
Kenya 0.80 Lesotho 0.38 Cameroon 0.60 Average 0.19
China 0.12 Average 0.52 Côte d'Ivoire 0.70
Indonesia 0.40 Average 0.49 Lower
Vietnam 0.80 Middle
Average 0.52 0.46

Low Bangladesh 0.45 Mali - Uganda - Benin 0.40

Tanzania 0.40 Malawi 0.30 Ethiopia 0.30
Nepal 0.30 Niger - Mozambique 0.30
Average 0.23 Central Madagascar 0.10
African Chad -
Republic 0.30 Yemen 0.50
Haiti 0.40 Laos 0.50
Average 0.20 Sudan 0.40
Cambodia 0.20 Low
Average 0.30 0.27

Overall Average 0.49 0.44 0.44 0.28 0.41

Source: Ross and Stover, 2001.

14
2. The Age Distribution of Fertility
In addition to the TFR, the age distribution of fertility is also
required to make a population projection. In DemProj, this
information is entered as the percentage of lifetime fertility
that occurs in the five-year age groups 15-19, 20-24, 25-29,
30-34, 35-39, 40-44, and 45-49.
Age at childbearing is relevant for reproductive health. The
age of the mother may be of particular concern to the well-
being of the infant, with younger ages being associated with
risks of premature delivery or protracted labor, and older
ages with congenital malformations. The mother herself also
may be adversely affected by extremes in age of
childbearing. Women who bear children at younger ages
may be physically and socially unprepared, and older
mothers may face aggravated anemia or hemorrhagic
complications.

Use of Fertility Models in Population Projections

As noted earlier, in DemProj, model schedules are used to
generate age patterns in conjunction with levels of fertility,
mortality, and migration.
Projected fertility is affected by certain age characteristics.
For example, appropriate fertility rates need to be assigned
by age group, as those groups vary in size, which contributes
to the size of the population being projected to the next
time period. In addition, some implications of population
projections follow from the age of the mothers as they bear
children.

Fertility Surveys
Information on the initial age distribution of fertility is usually
available from national fertility surveys. It may be reported
as age-specific fertility rates (the number of live births per
1000 women in the age group) rather than as the percent
distribution of fertility. Data on age-specific fertility rates can
be converted to the required percentage distribution by
dividing each age-specific fertility rate by the sum of all the
age-specific fertility rates. This calculation is illustrated in
Table 2.

15
 Table 2. Calculation of the Distribution of Fertility by Age Group
from Age-Specific Fertility Rates

Age Age-Specific Fertility Percent Distribution

Group Rate for Fertility by Age
Group
15-19 84 10.9
20-24 202 26.3
25-29 203 26.4
30-34 143 18.6
30-34 97 12.6
40-44 34 4.4
45-49 5 0.7
Total 768 100.0

From a review of hundreds of fertility shapes, we know that

fertility patterns tend to concentrate and to shift to younger
years as fertility declines (Horne and El-Khorazaty, 1996). In
Figure 1, for example, the normalized fertility distributions for
Bangladesh show a switch from a traditional plateau
distribution in 1975 to a rather peaked distribution in 1993-94.
For Taiwan, Figure 2 shows a shift from a homogenous
distribution to one where fertility occurs almost entirely
among women in their twenties.

16
Figure 1: Shift in Fertility Distributions -
Bangladesh: 1975
to 1993-94

percent distribution
30

25

20

15

10

0
15-19 20-24 25-29 30-34 35-39 40-44 45-49
age group

1975 BFS 93-94 DHS

Figure 2: Shift in Fertility Distributions - Taiwan:

1956
to 1983

17
 percent distribution
50

40

30

20

10

0
15-19 20-24 25-29 30-34 35-39 40-44 45-49
age group

1956 1983

18
 Table 3: UN Model Tables of the Age Distribution of Fertility

TFR 15-19 20-24 25-29 30-34 35-39 40-44 45-49 Total

Sub-Saharan Africa
2 8.2 35.4 29.9 17.4 7.2 1.7 0.1 100
3 14.0 31.1 24.7 16.6 9.2 3.6 0.6 100
4 14.9 25.9 22.1 17.1 11.7 6.4 1.8 100
5 16.1 25.4 22.0 17.0 11.6 6.2 1.6 100
6 16.4 24.7 22.1 17.3 11.7 6.2 1.5 100
7 14.7 23.5 21.9 17.9 12.8 7.2 2.0 100
Arab Countries
2 7.2 31.1 30.3 19.7 9.0 2.4 0.2 100
3 6.6 29.1 29.8 20.7 10.4 3.2 0.2 100
4 7.6 24.4 26.0 21.1 14.2 6.9 1.4 100
5 8.5 23.1 24.9 21.0 14.2 6.9 1.4 100
6 8.8 21.9 24.3 21.1 14.8 7.5 1.6 100
7 7.8 21.7 25.1 21.9 15.0 7.2 1.4 100
Asia
2 2.8 31.1 38.4 21.1 5.9 0.7 0.0 100
3 2.4 23.5 33.7 25.6 11.9 2.8 0.1 100
4 3.8 20.8 27.9 24.6 15.7 6.3 0.8 100
5 5.6 21.4 26.6 23.3 15.4 6.7 1.0 100
6 7.9 22.8 26.2 22.0 14.2 6.1 0.9 100
7 11.8 24.1 24.1 19.5 13.0 6.3 1.3 100
Average
2 6.1 32.5 32.9 19.4 7.4 1.6 0.1 100
3 7.7 27.9 29.4 21.0 10.5 3.2 0.3 100
4 8.8 23.7 25.3 20.9 13.9 6.5 1.3 100
5 10.1 23.3 24.5 20.4 13.7 6.6 1.3 100
6 11.0 23.1 24.2 20.1 13.6 6.6 1.3 100
7 11.4 23.1 23.7 19.8 13.6 6.9 1.6 100

Source: World Population Prospects, as assessed in 1973.

19
 Regionality and Model Fertility Patterns: the United
Nations Model Schedules
Aspects of fertility that can vary from one locale to another
include: how concentrated the fertility rates are, how fast
this concentration occurs, and how young a shift is
displayed. The United Nations Population Division has
developed regional patterns to describe the shift in
childbearing as fertility changes, using patterns labeled as
sub-Saharan Africa, Arab nations, and Asia. In Table 3, the
fertility distributions by various total fertility rates are given for
each of the four patterns. Figure 3 depicts the shifting
fertility distribution for the sub-Saharan Africa model for
changing fertility levels. It shows a relatively flat distribution
of fertility across ages for high level of TFR, and an
increasingly peaked distribution (reaching a maximum in the
age group 2024) when TFR is 2 or 3.

Figure 3: Fertility Distribution by TFR: UN Sub-

Saharan Africa Pattern

percent distribution
40

35

30

25

20

15

10

0
15-19 20-24 25-29 30-34 35-39 40-44 45-49
age group
Africa TFR=2 Africa TFR=3 Africa TFR=4
Africa TFR=5 Africa TFR=6 Africa TFR=7

20
The differences among the model fertility tables can be
seen in Figure 4. In this figure, the age pattern of fertility is
compared for all four tables at a TFR of four. The model
table for sub-Saharan Africa has a higher proportion of
fertility concentrated in the youngest age groups, while
the Asia table has the oldest pattern.

Figure 4: Comparison of Model Fertility Tables at a

TFR of 4.0

percent of lifetime fertility

30

25

20

15

10

0
15-19 20-24 25-29 30-34 35-39 40-44 45-49
age group

Sub-Saharan Africa Arab Asia Average

However, regions are not as well associated with fertility

patterns as they are with mortality patterns. Mortality
patterns by region have been employed long and
successfully to analyze and project mortality. Regional
mortality forces that tend to match up with geography are
climate, country proximity, culturally based sex roles, diet,
medical practices, political stability, etc. Fertility is subject to
relatively more forces that are primarily not biological and
therefore are harder to aggregate by geographic area than
is true for mortality.

Figures 5 and 6 illustrate the inconstancy of regional patterns

for projecting fertility. They show that the Asia pattern was
an adequate model for Bangladesh in 1975 when the TFR
was 6.3, but by 1989, when the TFR had dropped to 5.1, the
Asia model was far from adequate. Bangladesh’s status

21
 then was better matched by the sub-Saharan Africa
pattern.

Figure 5: Shift in Fertility Distributions – Bangladesh

and
UN Asia Pattern

Figure 6: Shift in Fertility Distributions – Bangladesh

and
UN Sub-Saharan Africa Pattern

22
In sum, DemProj users should not assume that because their
country lies geographically within Asia or sub-Saharan Africa,
or is an Arab country, they should necessarily choose that
respective region. Users should first check their country or
region’s initial fertility distribution against those given in Table
3. If these shapes do not fit, or the country lies outside of
those three regions, the Average pattern is recommended.

Combining Data from Fertility Surveys with Model

Tables
The most common situation is one where data on the
distribution of fertility by age are available for the base year.
However, most projections include changing TFR in future
years. The model tables are useful for estimating how the
fertility pattern will change as fertility changes. This raises the
problem of how to use the base year survey data but still
take advantage of the model tables for future values of TFR.
The simplest solution is to use the fertility survey data for the
base year and the model table values for the final year of
the projection or the year in which fertility stops changing.
Interpolating between these two points will yield a fertility
distribution that takes advantage of both methods. To get
the most out of this method, the interpolation for any
particular year should be based on how much the TFR has
changed within the projected range, and not on the year
itself.

23
 The Coale-Trussell Fertility Model
Another approach to using model fertility shapes in DemProj
is to use the Coale-Trussell model (Coale and Trussell, 1974
and 1978). This model takes into account the change in
marriage patterns, the degree of fertility control, and the
underlying natural fertility of a population. The main feature
that makes the model of interest to DemProj is that it is
relational: users supply an initial age distribution and desired
future total fertility rate, and a new age pattern of fertility is
generated which takes the characteristics and components
of the initial distribution into account. Figures 7 and 8 show
the shapes generated if we start with the rather different
shapes of Bangladesh’s 1975 and 1989 Fertility Surveys (BFSs).

24
Figure 7: ASFRs Generated by Coale-Trussell Model
Compared with Survey ASFRs: Bangladesh,
BFS 1975

(reference fertility schedule BFS 1975)

births per thousand women

300

250

200

150

100

50

0
15-19 20-24 25-29 30-34 35-39 40-44 45-49
age group

1975 BFS TFR= 6 TFR= 5 TFR = 4 TFR = 3

Figure 8: ASFRs Generated by Coale-Trussell

Model Compared with Survey ASFRs:
Bangladesh,

(reference fertility schedule BFS 1989)

births per thousand women

300

250

200

150

100

50

0
15-19 20-24 25-29 30-34 35-39 40-44 45-49
age group

1989 BFS TFR= 6 TFR= 5 TFR = 4 TFR = 3

25
 BFS 1989

The Coale-Trussell model is recommended for populations

whose initial fertility distributions do not resemble any
regional pattern, or have some idiosyncrasy. It performs best
in the medium run, if moderate levels of fertility are targeted;
or in the short run, regardless of levels of fertility. But it gives
too peaked results when low levels of fertility are projected
after a substantial period of fertility decline.

C. Mortality
Mortality is described in DemProj through two assumptions:
life expectancy at birth by sex, and a model life table of
age-specific mortality rates.

1. Life Expectancy at Birth

Base Year Estimates

Life expectancy at birth is the average number of years that a
cohort of people would live, subject to the prevailing age-
specific mortality rates. It is a useful measure that summarizes in
one indicator the effect of age-specific mortality patterns. Life
expectancy can be calculated from vital statistics on deaths if
The best source of reporting is complete. In the developing world, death
information on life registration is not usually complete enough to be used for this
expectancy will usually purpose. Estimates of life expectancy are usually derived
be national reports instead from large-scale surveys or censuses. The best source of
prepared by analyzing
information on life expectancy will usually be national reports
large-scale surveys or
censuses. prepared by analyzing these surveys. If national estimates are
not available, life expectancy estimates may be obtained from
a variety of other sources, including the United Nation’s World
Population Prospects or the Demographic Yearbook, the U.S.
Census Bureau’s World Population Profile, the World Population
Data Sheet of the Population Reference Bureau, or the World
Bank’s World Development Indicators.

Future Assumptions
An assumption about future levels of life expectancy at birth
is required for all population projections. There are several
options for setting the life expectancy assumption.
1. National projections. Many countries have official
population projections that include assumptions about
the future course of life expectancy, often with several
variants. If population projections are being made for

26
 planning purposes, it is often recommended that the
official assumptions and projections be used.
2. National goals. Many countries have national
population goals that include life expectancy. It is
often useful to use these goals as a starting point for
projections. In one projection it may be assumed that
the life expectancy goal has been achieved, while in
others model users may examine the effects of a delay
in achieving the goal. Sometimes goals are expressed
in terms of crude death rates or population growth
rates instead of life expectancy. In these cases,
different life expectancy assumptions may be tried to
discover a projection that is consistent with the national
goals for these other indicators.
3. United Nations and U.S. Census Bureau projections. The
population projections prepared by the United Nations
Population Division and reported in World Population
Prospects include assumptions about levels of life
expectancy for each country reported. These
assumptions may be used. The U.S. Census Bureau
projections also contain a set of life expectancy
assumptions.
4. Recent trends and international experience. If
information is available on life expectancy for several
years, it may be useful to analyze the trends and
develop a future assumption based on continuing past
trends. It should be noted, however, that past trends
cannot be expected to continue for very long into the
future. Life expectancy rarely increases at a constant
pace throughout an entire demographic transition.
Rates of decline are often slow at first, increase during
the middle of the transition, and slow again as they
approach high levels.
5. United Nations model schedule. In preparing its
population projections every two years, the United
Nations Population Division uses a model schedule of
changes in life expectancy. This schedule assumes that
life expectancy at birth, for both males and females,
increases by 2.0 to 2.5 years over each five-year period
when life expectancy is less than 60 and then increases
at a slower rate at higher levels. Table 4 shows the
working model used in the United Nations population
projections.

27
 Table 4: UN Working Model of Life Expectancy Improvement During
a Five-Year Period

Fast Rise Middle Rise Slow Rise

Initial Life
Expectancy Male Female Male Female Male Female
55.0-57.5 2.5 2.5 2.5 2.5 2.0 2.0
57.5-60.0 2.5 2.5 2.5 2.5 2.0 2.0
60.0-62.5 2.5 2.5 2.3 2.5 2.0 2.0
62.5-65.0 2.3 2.5 2.0 2.5 2.0 2.0
65.0-67.5 2.0 2.5 1.5 2.3 1.5 2.0
67.5-70.0 1.5 2.3 1.2 2.0 1.0 1.5
70.0-72.5 1.2 2.0 1.0 1.5 0.8 1.2
72.5-75.0 1.0 1.5 0.8 1.2 0.5 1.0
75.0-77.5 0.8 1.2 0.5 1.0 0.3 0.8
77.5-80.0 0.5 1.0 0.4 0.8 0.3 0.5
80.0-82.5 0.5 0.8 0.4 0.5 0.3 0.3
82.5-85.0 – 0.5 – 0.4 – 0.3
85.0-87.5 – 0.5 – 0.4 – 0.3

2. Life Expectancy and AIDS

In a number of countries, the AIDS epidemic has had a
significant impact on mortality. It affects both life
expectancy and the age and sex pattern of mortality. This
health concern raises two problems for population
projections. First, in countries with high HIV prevalence, the
future course of the AIDS epidemic will be the single largest
determinant of future life expectancy. Second, the age
pattern of mortality will depart significantly from the patterns
described in the model life tables discussed below.
Therefore, in countries with adult HIV prevalence greater
than a few percent, it is best to consider the effects of AIDS
explicitly in the population projection. These effects cannot
be incorporated simply by changing the life expectancy
assumption since the age pattern of mortality is also
affected (AIDS deaths are concentrated in the age groups
of 15-49). The recommended approach is to first develop a
population projection that ignores the effect of AIDS, then to
make assumptions about the future level of adult HIV
prevalence and let the computer program calculate the
effects of AIDS on the population projection. Such
projections can be prepared using DemProj and AIM, the
AIDS component of Spectrum. The methods and
assumptions to do this are described in a companion
manual, AIM: A Computer Program for Making HIV/AIDS
Projections and Examining the Social and Economic Impacts
of AIDS.

28
 Most of the steps required to prepare a population
projection that excludes the effects of AIDS are the same as
for any other projection. However, preparation of the life
expectancy assumption may be different. If AIDS has not
yet contributed significantly to mortality, then estimates of
current life expectancy can be used. However, in many
countries AIDS has already affected life expectancy. In
these cases, it is necessary to prepare an estimate of life
expectancy that excludes the impact of AIDS. This could be
done by removing AIDS deaths from the schedule of deaths
by age and recalculating life expectancy. However, AIDS
deaths are usually greatly underreported, making this
approach questionable.
An alternative approach is to start with an estimate of life
expectancy for some year before a significant number of
AIDS deaths occurred. Then estimate how life expectancy
would have changed from that year forward to the base
year of the projection in the absence of AIDS. This could be
done by extrapolating the historical trend or by using the
United Nations model schedule shown in Table 4. Life
expectancy in the absence of AIDS can be projected into
the future in a similar manner.
It should be noted that the It should be noted that the life expectancy assumptions
life expectancy assumptions reported in World Population Prospects and World Population
reported in World
Profiles include the effect of AIDS. In countries where AIDS is a
Population Prospects and
World Population Profiles serious problem, these life expectancy projections cannot be
include the effect of AIDS. used to create a population projection in the absence of AIDS.
Both organizations do produce special tables that show the
non-AIDS life expectancy assumptions that can be used in
DemProj.

3. Age-Specific Mortality

Model Mortality Tables

The mortality input to DemProj, life expectancy at birth,
indicates overall mortality in a population. But Demproj also
needs the pattern of mortality in order to produce mortality
rates by age group. Specifically, the rates required by
DemProj are survival ratios, which will survive one age group
into the next five-year group.
The majority of countries to which DemProj has been applied
have had no complete, empirical life tables⎯and life tables
are what yields survival ratios, or sx. Even if there were such
tables, generally little is known about how the pattern of
mortality would evolve, given projected changes in mortality
levels. In addition, entering data for a life table function for

29
 all 18 age groups is burdensome. So for both parsimony and
consistency with studied patterns, DemProj employs model
life tables.

Regional Model Life Tables

Two sets of model life tables are employed by DemProj: the
Coale-Demeny (Coale, Demeny, and Vaughan, 1983)
model tables and the United Nations tables for developing
countries (United Nations, 1982). These two sets differ in (1)
the algorithm they use to generate the mortality schedules,
and (2) the empirical data sets from which they were drawn
(Coale-Demeny: life tables from Europe and other
industrialized regions from the first half of the 20th century;
the United Nations: life tables from developing countries
from the second half of the 20th century).
But the two model sets are similar in one important respect:
they contain regional families that are distinguished by
underlying causes of death. Referring to the regions from
Europe whose life tables provided distinct patterns, the
Coale-Demeny families are known as: North, East, South,
and the non-idiosyncratic West. The United Nations families
name very specific regions—Latin America, South America,
Chile, South Asia, East Asia, plus the generic General.
Several aspects of these model schedules are highlighted in
Figures 9 through 11. In each, the probability of dying during
an age interval, qx , is plotted against age groups. (A semi-
log scale is used to highlight important differences.)

30
In Figure 9, the probabilities of dying are generated for the
Coale-Demeny West model (for females), for five mortality
levels. The range in mortality levels is from very low [e(0)=35]
to rather high [e(0)=75]. In this series, the effects of the
changing mortality levels are most noticeable in the
younger age groups.

Figure 9: Model Life Table Mortality Probabilities –

Coale-Demeny West Female Models Table:
Compared by Level of Life Expectancy

probability of dying
1

0.1

0.01

0.001
0 10 20 30 40 50 60 70 80
age group

75 65 55 45 35

31
 In Figure 10, the qx mortality probabilities are plotted for the
four Coale-Demeny regions for females having the same life
expectancy at birth: 45 years. The North pattern has the
highest death rates for infants and children; the West has
high death rates for age 15 and higher.

Figure 10: Model Life Table Mortality Probabilities –

Coale-Demeny Female Model Table
Families: Comparison by Family at e(0) =
45

probability of dying
1

0.1

0.01
0 10 20 30 40 50 60 70 80
age

South North East West

32
In Figure 11, the mortality probabilities are graphed for the
five United Nations families, again for females with a life
expectancy at birth of 45 years. There is greater
differentiation between United Nations than between
Coale-Demeny families. One extreme is given by the East
Asian pattern, which has a rather high level of adult
mortality and a relatively low level of under age 10 mortality.
The other extreme is the South Asian family, which has a
distinctly low adult mortality pattern (between ages 10 and
50), and a relatively high child mortality pattern.

Figure 11: Model Life Table Mortality Probabilities –

United Nations Female Model Table
Families:
Comparison by Family at e(0) = 45

probability of dying
1

0.1

0.01
0 10 20 30 40 50 60 70 80
age group
Chile South Asia Latin America
General East Asia

Selecting a Model Life Table

There are several approaches to determining the most
appropriate life table for a particular country or region in
terms of the mortality level and pattern.
1. The best way is to compare data on age-specific
mortality for that country with the mortality pattern at
the corresponding level of life expectancy for each of
the model tables. The appropriate table will be the
one that most closely matches the actual experience.

33
 2. The best model life table may already have been
determined by demographers from that country.
Consulting with them or reviewing reports or official
projections may indicate the model table that the
national demographers consider the best fit.
3. If neither of the first two options is possible, a third
option is to compare the crude death rate and infant
mortality rate that would result from choosing each
model life table with the actual estimates of these
indicators. The table that provides the closest fit may
be the most appropriate. The DemProj computer
program will generate corresponding crude death and
infant mortality rates for the first year of the projection
automatically as a guide to choosing the best table.

Modifying the Model Life Tables

Every effort has been made to make DemProj as easy to
use as possible. This ease of use comes at the expense of
some flexibility. There may be times when it is desirable to
determine an age-specific schedule of mortality without
using these model life tables. DemProj does not contain
an editor to allow interactive entry of custom tables of
age-specific mortality and the majority of the data
DemProj contains are survival probabilities, not mortality
rates. However, DemProj does allow for the use of a
custom life table. A custom table can be created by
editing the custom life table fie.
In a typical configuration, the computer files containing the
information on model life tables are contained in the
directory C:\Spectrum\DP. All model life table files for
females have the extension “.f” while all male files have the
extension “.m”. The Coale-Demeny tables are named
“cdnorth,” “cdsouth,” “cdwest” and “cdeast.” The United
Nations tables are named “unchile,” “unea,” “ungen,”
“unla,” and “unsa.” The custom tables are named custom.f
and custom.m. These files can be edited with any editor or
word processor that can produce ASCII or text files to enter
the desired patterns.
The format of the model life table files is as follows:
• Row 1: life expectancy at birth.
• Rows 2 through 81: one-year survival ratios for ages 0 to
79.

34
 • Row 82: the proportion of the age group 80+ that will
survive to the next year.
• Row 83: the infant mortality rate.
• Row 84: the child mortality rate (1-4).
In each row, the values in the columns correspond to the life
expectancy shown in that same column in the first row.
Any of the figures in these files, including life expectancy,
may be changed to produce a new life table. To produce
an entire schedule of age-specific mortality rates for a
particular projection, enter the life expectancy for each
year or five-year interval in the first row and the
corresponding survival rates in the rows below. The life
expectancy values do not need to be round numbers, nor
do they need to be equally spaced; however, they do need
to be arranged from the lowest value in the first column to
the highest value in the last column. Once the new values
are entered, save the table as a text file using a unique
name, and then specify the custom table and enter this
name when creating the DemProj projection file.

D. Migration
Migration refers to the number of migrants moving into or
out of the area for which the population projection is being
prepared. If the projection is for a country, then it is
international migration. If the projection area is a region or
city, then migration refers to people moving into or out of
the region or city.

In most cases, information Migration is specified through two inputs. The first is the net
on migration will come number of migrants, by sex and year. If the net flow is outward,
from local sources,
then net migration should be a negative number. If the net flow
usually studies based on a
national census. is inward, then it should be positive. In most cases, information
on migration will come from local sources, usually studies based
on a national census. The United Nations report World
Population Prospects does contain estimates and projections of
total net migration, but they are not disaggregated by sex.
The second component of the migration assumption is the
distribution of migrants by age for each sex. This information
also needs to come from national studies. There are no
simple model tables for patterns of migration by age;
however, the United Nations has devised an approach to
developing age and sex migration patterns. This approach
is based on model schedules of gross migration developed

35
 by Castro and Rogers (described in United Nations, 1989, pp.
65-69).
Age and sex patterns of Net international migration is not a major component of
migration vary population change in most countries. Often, migration can be
considerably.
ignored without a significant effect on the population
projection. However, for special areas, such as cities, and for
some countries, migration can be very important. Moreover,
age and sex patterns of migration vary considerably. In Nairobi,
for example, migrants to the city consist largely of young males
seeking work. In other cities, such as Kinshasa, migrants to the
city are composed primarily of entire families. In Jordan, there
was a significant outflow of migrants during the oil boom in the
Persian Gulf states of the 1970s and 1980s, but during the 1990s,
there was a net inflow of migrants as families returned to Jordan
due to reduced employment opportunities in the Gulf.

E. Current Population: Urban and Rural

DemProj can be used to make urban and rural population
projections along with the national projection. DemProj first
projects the national population and then distributes it
according to region. It does not make a separate
projection for each region. The benefit of this approach is
that separate estimates of TFR, life expectancy, and model
life tables are not required. The disadvantage is a lack of
flexibility in fully specifying the differences between regions.
If this flexibility is important, DemProj should be used to
project urban and rural regions separately.

There are two methods for developing the urban and rural
projections. For the first, the analyst provides an assumption
about the percentage of the total population that is urban
for each year of the projection. Then, rural population is
simply the difference between the total and urban
populations.
The second approach relies on a method developed and
used by the United Nations Population Division. This method
assumes that the percentage of the entire population that is
urban follows an S-shaped curve. Therefore, the rate of
urbanization will be slow at first, accelerate when the
percent urban reaches 30-70 percent, and then slow down
once urbanization reaches high levels. This approach
requires an estimate of the annual growth rates of the urban
and rural populations in the base year and the size of the
urban population by age and sex in the base year. This

36
 information should be readily available from national
censuses. A typical pattern of urbanization generated by
this approach is illustrated in Figure 12. A complete
description of the methodology used can be found in
Section V.G.
Figure 12: Typical Pattern of Urbanization

percent of population urban

100

80

60

40

20

0
0 25 50 75 100 125 150 175 200
time in years

F. Projection Outputs
DemProj will calculate and display the population size by
year. Projections can be examined in terms of total
population or population by age, sex, and region. In
addition, a number of demographic indicators can also be
displayed. A complete list of indicators available is given
below with definitions for those that are not obvious. Many
of these definitions are taken from the Population Handbook
(Population Reference Bureau, 1989); a complete glossary is
contained at the end of the manual.
DemProj indicators:
• Total population size
• Population aged 0-4
• Population aged 5-14
• Population aged 15-64
• Population aged 65+
• Total net international migration

37
 • Annual growth rate (GR). The rate at which the
population is increasing or decreasing in a given year
due to natural increase and net migration, expressed
as a percentage of the base population.
• Births. The total number of annual births.
• Child-woman ratio. The number of children under the
age of five per woman of childbearing age (15-49).
• Crude birth rate (CBR). The number of live births per
1,000 population in a given year.
• Crude death rate (CDR). The number of deaths per
1,000 population in a given year.
• Deaths. The total number of annual deaths.
• Defined age group. The size of the population in a user-
defined age group.
• Dependency ratio. The ratio of the economically
dependent part of the population (those aged 0-14
and 65 and over) to the productive part (those aged
15-64).
• Doubling time. The number of years it would take for
the population to double its current size at the current
annual rate of growth.
• Gross reproduction rate (GRR). The average number of
daughters that would be born to a woman (or a group
of women) during her lifetime if she passed through all
her childbearing years conforming to the age-specific
fertility rates of a given year. This is similar to the TFR
except that it counts only daughters.
• Infant mortality rate (IMR). The number of deaths to
infants under one year of age per 1,000 live births.
• Life expectancy [e(0)]. The average number of years a
person can expect to live based on the age-specific
death rates for a given year. This is the calculated life
expectancy at birth. If AIM is not being used, then this
number will be the same as the input life expectancy.
However, if AIM is being used, then the calculated life
expectancy will include the impact of AIDS deaths.
Therefore, it will be different from the input life
expectancy.
• Mean age of childbearing. The average age of
mothers at the time of birth.
• Median age. The age that divides a population into
two numerically equal groups.

38
• Net reproduction rate (NRR). The average number of
daughters that would be born to a woman (or a group
of women) during her lifetime if she passed through all
her childbearing years conforming to the age-specific
fertility rates and age-specific mortality rates of a given
year. This is similar to the GRR except that it includes
the effect of mortality that would cause some women
to die before completing their childbearing years.
• Rate of natural increase (RNI). The rate at which the
population is increasing or decreasing in a given year
due to the surplus or deficit of births over deaths,
expressed as a percentage of the base population.
• Sex ratio. The number of males per 100 females in a
population.
• Total fertility rate (TFR). The average number of children
that would be born alive to a woman (or a group of
women) during her lifetime if she were to pass through
all her childbearing years conforming to the age-
specific fertility rates of a given year.
• Under-five mortality rate (U5MR). The number of deaths
to children under the age of five per 1,000 live births.

39
40
IV. Program Tutorial
This tutorial covers the key steps in installing and running
Spectrum and DemProj. It assumes that you have a
computer running Windows 98 or later version and that you
are familiar with the basic operation of Windows programs
and terminology.

A. Before You Get Started

You will need to collect data and make certain decisions
before running the model. For example, to set the
projection parameters, you will need to decide

• The first year of the projection (usually the latest year for
which you have population data)
• The last year of the projection
• Whether you want to project the total population or
whether you want to break it down into urban and rural
segments
• Whether you want to enter population numbers in
terms of thousands or millions.

The data you will need include

• The first-year population, broken down by sex, and in age

categories of 0-4, 5-9, etc. until the final category of 80+
• Sex and age information in rural and urban categories
for the first year, if you want to break the projections
down into rural and urban populations
• Total fertility rate
• Age-specific fertility rates
• Sex ratio at birth (the number of male births per 100
female births); default is 105
• Life expectancy for the first year, male and female
• The model life table you want to use
• Net migrants per year by sex and age.

41
 These inputs are all described in Chapter III of this manual.
B. Installing the Spectrum Program

The Spectrum program is distributed on CD-ROMS or through

the Internet at [Link] [Link]. It must be
installed on a hard disk before it can be used. Spectrum will
operate on any computer running Windows 95 or later
version. It requires about 30MB of hard disk space.
To install the Spectrum program, follow the directions below.

Installing from a CD-ROM. Insert the CD-ROM into your CD-

ROM drive. The installation program should start
automatically. If it does not, Select “Start” from the task bar,
then select “Run” from the pop-up menu. In the dialogue
box that appears, click on Browse, and find the file
[Link]. Then press “Ok.”

Installing from the internet. Start your internet browser and

go to [Link]. Click on “Software” and then
“Spectrum”. Next click on “Spectrum download (single
executable file). From the dialogue box that appears next,
select “Save”. Select a location for the file. Once the file has
been downloaded, click on that file and the follow the
instructions.

C. Creating a New Projection

1. Starting the Spectrum Program
1. To start Spectrum:
1. Click the “Start” button on the task bar.
2. Select “Programs” from the pop-up menu.
3. Select “Spectrum” from the program menu.
Alternatively, you can use Windows Explorer to locate
the directory “c:\spectrum” and double click on the
file named “[Link]”.

2. Specifying the Projection Parameters in Projection

Manager
To create a new projection:
1. Select “File” from the menu bar.

42
2. From the pull-down menu that appears, choose “New
projection.”
Next the “Projection manager” dialogue box will appear
and will look like the following screen:

The following information is required to create a new

projection:
Projection title: Enter the title you wish to assign to this
projection. This title will be printed at the top of all printed
output and will be used to identify the projection if more
than one projection is loaded at a time.
Projection file name: You must also assign a file name to the
projection. This is the name that will be used to store all data
files associated with this projection. This name must obey
the usual DOS-based rules for file names, i.e., it should not be
more than eight characters in length and should not contain
any illegal characters (you may use letters, numbers, and
certain symbols such as a hyphen). Do not enter an
extension with the file name; the program will automatically
assign the extension “.pjn” to the projection file.

43
If a box is shown in gray, 1. Click on the “Projection file name” button to open an edit
you will not be able to box and type the file name.
change its contents. It
2. Click on “Ok” to accept or “Cancel” to reject.
means that another
projection has been loaded,
and the data must remain the First year: Click in the “First year” box to enter the first year of the
same. If you want to create projection.
an entirely new projection,
you should close the other Final year: Click in the “Final year” box to enter the final year of
projections, using “File” the projection. In most cases the final year should not be more
and “Close,” and then select than 50 years later than the first year. However, if Spectrum is
“File” and “New.” Users being used only to make a population projection (e.g., no other
may want to have several modules will be used), then projections up to 150 years in length
projections open in order to may be made.
examine the effects of
changing assumptions such Projection timeframe. The radio button (or option) labeled
as the final year total fertility “Standard demographic projection < = 50 years” will be
rate. selected by default. This is the normal mode of operation.
However, if you wish to make a population projection longer
than 50 years, then you may select the button for “Demography
projection > 50 years.” This choice will cause the DemProj
module to make the population projection in five-year intervals,
rather than in single years. In this case, a projection up to 150
years in length may be made.
Once all the information is Active modules. These radio buttons (or options) let you select
entered for this dialogue other Spectrum modules that will be used with the population
box, click on the “Ok”
projection. If none are selected, then only a population
button. You can always
return to this screen and projection will be made. Alternatively, you may select any or all
change some of the of the other modules to be included with your projection.
information by selecting
EasyProj. EasyProj is a special feature that allows you to use
“Edit” from the menu bar
and “Projection” from the data prepared by the United National Population Division and
pull-down menu. published in World Population Prospects. If you click on the
EasyProj button, the program will prompt you to select a country
and ask whether you want to use the UN low, medium, or high
projection assumptions. Once you click OK, the program will
load the base year population, the total fertility rate, the male
and female life expectancy, and the model life table from the
United Nations estimates and projections. By default, the
projection will assume zero net migration.

D. Entering the Base Year Data and

Projection Assumptions
1. Specifying New Demographic Projection Parameters

44
For readers who feel they need additional review or
explanations of the terms found in this section, Chapter III
and the glossary to this manual may be useful.
To enter the assumptions for the population projection,
select “Edit” from the menu bar and “Demography” from
the pull-down menu. A dialogue box like the one shown
below will be displayed.

The “Projection parameters” button lets you set the

configuration of the projection. Then you may enter the
assumptions by selecting the “Demographic data” button.

45
 First, select the button labeled “Projection parameters.” The
screen should look like this:

In the top left part of this box the first year, final year and
projection period are shown. These entries cannot be
changed here; they are shown for information only. Users
can change these parameters by selecting “Edit” from the
menu bar, and then “Projection.”
The other items in the dialogue box are as follows:
Currency Name: Type in the name to label the currency
units.
Urban/rural projection: Use these buttons to indicate
whether you want to include urban and rural projections as
part of your overall projection. The default is “Do not include
urban/rural projections.”
Scale: Scale refers to the scale used when base year
population figures are entered. Consider the case when the
population in the base year is 10 million. If you select a scale
of “Units,” then you should enter the base year population as
10000000. If you select a scale of “Thousands,” then you
should enter the base year population as 10000. If you
select a scale of “Millions,” then you should enter it as 10.
For most country applications, “Thousands” will be the best
choice. It will keep users from having to enter very long

46
numbers while still leaving enough digits for adequate
precision.
Once all the information has been entered, click on the
“Ok” button to close the dialogue box.

2. Entering the Base Year Data and Projection

Assumptions

To enter the base year data and projection assumptions:

1. Choose “Edit” from the menu bar.
2. Choose “Demography” from the pull-down menu.
3. Choose “Demographic data” from the “Demographic
data” dialogue box. This step will display an editor like
the one shown below.

For each of the inputs required for the population projection,

there is a tab near the top of the screen.
1. To enter data for any of these assumptions, click on the
appropriate tab to display the editor for that variable.
2. Then click anywhere inside the editor to make it active.

47
To enter data for any of Although you can enter the data in any order, it is generally
these assumptions, click on best to begin by entering the data on the first year population
the appropriate tab to
as this information is needed to calculate some of the indicators
display the editor for that
variable. Then click displayed in the “ASFR” and “Model life table” editors. Each of
anywhere inside the editor to the editors is explained in the following pages.
make it active.
3. About the Editors
Each editor in DemProj is similar. At the very top of the
screen, the variable name appears. At the bottom of the
screen are the special edit keys. “Duplicate” allows you to
copy information from one cell, column or row to another;
“Interpolate” to enter a beginning and ending number and
have the computer calculate numbers for the intervening
intervals; “Multiply” to multiply a cell, column or row by a
specific number; and “Source” to write notes indicating the
source of the data for future reference.
To use the “Duplicate” button,
1. Highlight (select) the range (column, row, or cells to be
affected). The first cell in the range should be the value
you want to copy.
2. Extend the range to the last year by using the mouse
(hold down the left button and drag the range) or the
keyboard (hold down the shift key and use the arrow
keys).
3. Click on the “Duplicate” key to copy the value at the
beginning of the range to all the other cells in the range.
To use the “Interpolate” button,
1. Enter the beginning and ending values in the
appropriate cells.
2. Highlight the entire range from beginning to end.
3. Click on the “Interpolate” key to have the values
interpolated and entered into each of the empty cells.
To use the “Multiply” button,
1. Highlight the range (column, row, or cells to be affected).
2. Enter the multiplier in the dialogue box.
3. Click “Ok” to accept. The entire range will be
multiplied by the designated number.
To use the “Source” button,
1. Click on the “Source” button to open a small word
processor window.

48
 2. Enter the source of the data and make any special
comments about the assumptions.
3. Click on “Close” to return to the editor.
This feature allows you to keep a record of the data sources
and assumptions as you make the projections. This source
information will be maintained with the data file and printed
whenever you print the projection summary. It is strongly
recommended that you use this feature to avoid later
confusion.
When you have finished entering all the necessary data for
the component into the editor,
1. Click the “Ok” button to return to the “Sector selection”
dialogue box.
2. Click the “Close” button to complete the editing
process, or
3. Click on another sector to edit the data for that sector.
The “Cancel” button allows you to exit the editor without
making any changes to the data for the sector.

First Year Population

DemProj does have range The population by age and sex in the base year is entered in
limits built into the this editor. The scale is shown at the top of the editor. In our
individual editors. For
example, the scale we chose was “Thousands.” Thus, a figure
example, the model will
accept estimates of life such as 43,230 would be entered as 43.23.
expectancy between 32.5 1. Enter the size of the population by age and sex in the
and 87.5. If you enter a
value such as 125, you
appropriate cells of the editor. The total number of males
will be prompted to and females will be automatically calculated and
change the number. displayed in the last row of each column.
2. If your projection includes an urban and rural
projection, enter the urban population size by age and
sex. If you are not including urban and rural
projections, these columns will not be displayed.
3. Check to be sure that the total male and female
populations are correct once you have entered all the
required information. If they are not, you have
probably entered a number incorrectly in one or more
of the cells.
4. Once all the data are entered correctly, click on the
tab “Total fertility rate” to move to this editor.

49
 Total Fertility Rate
This editor is used to enter the base year total fertility rates
(TFRs) and the assumed future TFRs. The first few years will be
shown on the screen. To use this editor,
1. Click somewhere inside the editor to make the scroll
bar appear.
2. Scroll to the right or left to see all the years.
3. Enter a TFR for every year of the projection.

If you want to keep the TFR constant at the base year value,
you can enter the base year value and then use the
“Duplicate” key; duplication may be done for any interval of
time. To duplicate a value:
1. Highlight (select) the range. The first cell in the range
should be the value you want to copy. Extend the
range to the last year using the mouse (hold down the
left button and drag the range to the right) or the
keyboard (hold down the shift key and use the right
arrow key to scroll to the right).
2. Then click on the “Duplicate” key. This step will copy
the value at the beginning of the range to all the other
cells in that range.

50
You may want to enter a TFR value for the first year and the
last year and interpolate between these values to fill in the
intervening years; interpolation may be done for any interval
of time. To do this,
1. Enter the TFR values in the first and last years.
2. Highlight (select) the entire range from the first to the
last year.
3. Press the “Interpolate” key. Values will be interpolated
and entered for each of the years between the first
and last.
When you have entered the information on the TFR, click the
“ASFR” tab to move to the next editor.

ASFR

This editor is used to set the distribution of fertility by age. It

will look similar to the screen shown below.

There are three ways to specify the age distribution of

fertility. You can use the United Nations fertility model tables,
enter your own distribution, or use the Coale-Trussell fertility
model (see the glossary or Chapter III for more about these
distributions).

51
 United Nations Model Tables. The default is to use the United
Nations model tables. Select an option by clicking on the
button next to the name of the table you want to use. To
the right of each table name you can see the crude birth
rate that would be calculated for the base year if you
selected that table. (This column will show CBR = 0.0 for all
tables if you have not yet entered the base year population
or TFR.) Data for the age distribution of fertility will be
calculated from the TFR and the selected model table only
when you leave the “Demographic data” editor.
Your Own Data. If you wish to enter your own data for the
age distribution of fertility, then click on the button “Enter
from keyboard.” The bottom part of the screen will change
to display a spreadsheet-type editor (see screen below).
Enter your data into the appropriate cells. The total for each
year will be calculated and displayed in the last row. The
total for each year should be equal to 100 before you leave
this editor. If even one column does not total to 100, you will
see a warning message when you try to exit this editor. The
message will say “At least one column total < 99.9. Return to
editor?” or “At least one column total > 100. Return to
editor?” If you get either of these messages you should click
on “Yes” to return to the editor and correct the inputs.

52
Coale-Trussell Model. If you wish to use the Coale-Trussell
model, then click on the button “Coale-Trussell.” The bottom
part of the screen will change to display a spreadsheet-type
editor (see sample screen below). Enter the age distribution
of fertility in the base year. The Input TFR will be imported
automatically from the “Total fertility rate” editor; a message
will warn you if the TFR calculated here does not match the
Input TFR.

The “Duplicate” and “Interpolate” buttons function in the

same manner as for the “First year population” editor.
You may see the values as a table or a chart. If you want to
see a table of the actual values that have been calculated
from the UN fertility model tables, click on the “Selected”
button. It will display a table of values for all years in the
projection:

53
 To see the values as a graph,
1. Click on the “Chart” button. This will show a graph of
first year and last year values (not shown).
2. To close the graph, click on the “X” icon in the upper
right part of the graph window (Windows 95) or the “-“
icon (Windows 3.1).
3. Click the “Close” button to return to the editor.

To compare the age-specific fertility distribution that would

result from selecting any of the UN or Coale-Trussell model
tables,
1. Click the “Comparison” button to see a table
comparing the age specific fertility distribution). This
table will show the base year.
2. Advance the table by clicking the “Next Year” button,
or return to earlier years by clicking the “Prev Year”
button.
3. To see a graph instead of a table, click the “Chart”
button (see screen below).
4. To close the graph, click on the “X” icon in the upper
right part of the graph window (Windows 95) or the “-“
icon (Windows 3.1).
5. Select “Close.”

54
Once you have successfully entered all the necessary data,
click on the “Sex ratio at birth” tab to move to this editor.

Sex Ratio at Birth

This sex ratio at birth (the number of male births per 100
female births) is set by default to 105. For most countries it
will be about 103 to 105. You may enter a new number in
the edit cell or accept the default value; a message
(warning) box will appear if an invalid number such as 300 is
entered.

55
 Once you have specified the birth ratio, click on the “Life
expectancy” tab to move to this editor.

Life Expectancy
The life expectancy editor (see sample screen below) is used
to enter the life expectancy at birth for both males and
females in the base year, and to enter assumed values for
future years. The first few years will be shown on the screen.
1. Click somewhere inside the editor to make the scroll
bar appear.
2. Scroll to the right or left to see all the years.
3. Enter life expectancy for every year of your projection.

56
If you want to keep the life expectancy constant at the
base year value, you can enter the base year value and
then use the “Duplicate” key. To copy a value,
1. Highlight (select) the range. The first cell in the range
should be the value you want to copy. Extend the
range to the last year using the mouse (hold down the
left button and drag the range to the right) or the
keyboard (hold down the shift key and use the right
arrow key to scroll to the right).
2. Then click on the “Duplicate” key. This step will copy
the value at the beginning of the range to all the other
cells in that range.
You may want to enter a life expectancy value for the first
year and the last year and interpolate between these values
to fill in the intervening years. To do this,
1. Enter the life expectancy values in the first and last
years.
2. Highlight (select) the entire range from the first to the
last year.
3. Press the “Interpolate” key. Values will be interpolated
and entered for each of the years between the first
and last.

57
 When you have entered the information on the life
expectancy, click the “Model life table” tab to move to this
editor.

Model Life Table

Once you click the “Model life table” tab, you should see a
dialogue box like the one shown below.

With this dialogue box you can specify which model life
table (see glossary) should be used for this projection.
Values for both the four Coale-Demeny model life table
families and five United Nations model families are shown.
The display also shows the crude death rate and the infant
mortality rate that would be calculated in the first year if you
chose the corresponding life table. Choose the appropriate
table by clicking on the name of the table. The age-specific
mortality rates needed by the program will be calculated
based on the life expectancy and model life table when
you leave the editor.
Once you have chosen the model life table, click on the
“International migration” tab to move to the next editor. If
you do not wish to consider international migration in this
projection, you may skip this editor.

58
If you want to use a custom life table, select “Custom” from
the list, then click on the “Custom” button at the bottom of
the screen and enter the name of the file containing the life
table information.

International Migration
International migration is specified through three input
screens. When you select the “International migration” tab
you will see the editor shown below.

In this editor you specify the number of net male and female
migrants. If the direction of the net migration is out of the
country, then the figures should be negative; if the direction
of the net migration is into the country, then the figures
should be positive. Enter figures for male and female
migration for all years. If you leave the figures set to zero
there will be no effect of international migration in the
projection.
Once you have specified the net number of migrants, you
need to specify the age distribution of these migrants. Click
on the “Male age distribution” button to enter the age
distribution for male migrants and the “Female age
distribution” button to enter the distribution for female
migrants.

59
 Regional Assumptions
If you are including an urban and rural projection, there will
be a tab for “Regional assumptions.” Select that tab and
you will see the “Regional assumptions” editor.
There are two methods for entering regional assumptions.
The editor will first appear with the button “Enter percentage
directly” chosen. In this editor you may enter the
percentage of the entire population that is urban for each
year in the projection. (Note: The term “regional” may
apply more broadly than to the rural/urban distinction. It
may refer to a geographic distinction [highlands/lowlands];
a cultural distinction [indigenous/foreign]; or a political
distinction [north/south], etc.)
To use the other method for making regional projections,
click the “Use growth rate difference” button. That editor
requires two pieces of information: the base year growth
rates of the urban and rural populations.

4. Leaving the Demographic Data Editors

Once you have entered all the necessary information,
1. Leave the demography editors by clicking on the “Ok”
button in any of the editors. When you click the “Ok”
button, the program will record your changes and
return to the “Demographic data” dialogue box.
2. Click on “Close” to keep your work, and you will return
to the main program. If you decide that you do not
want to keep the changes you have just made, click
the “Cancel” button in any editor. This will exit the
demography editors and restore all inputs to their
values before you entered the demography editor.
Any changes you made during the current editing
session will be lost.

5. Saving the Input Data

Once you have entered the projection assumptions, it is a
good idea to save the data onto your hard disk. To do this,
select “File” from the menu bar and “Save projection” from
the pull-down menu. The data will be saved using the file
name you specified earlier.

E. Making the Projection

Whenever you enter data for a new projection or edit the
assumptions, DemProj will note that the data have been

60
 changed. The next time you try to display an indicator it will
inform you that the data may have changed and ask if you
want to recalculate the projection. Normally, you should
answer “Yes” to this question. DemProj will then make the
population projection. This may take only a few seconds if
you are making only a population projection, or could take
somewhat longer if you are also making a projection
including AIDS, family planning or reproductive health.
Once the projection is made, you will not be asked again if
you want to project the population, unless you edit the
assumptions.

F. Examining the Output

To see the results of the projection, select “Display” from the
menu bar. From the pull-down menu select “Demography.”
You will then see another menu showing the seven
categories of indicators available:
• Population
• Fertility
• Mortality
• Vital events
• Ratios
• Age groups
• Summary

61
 Choose one of these categories and you will see one final
menu listing the indicators available in that category. Select
one of the indicators. Then you will see the display dialogue
box. It will look something like the one shown below.

62
The exact choices available will depend on the indicator
you have selected. For “Total population” you can set the
region to “Total,” “Urban,” or “Rural” if you are doing a
regional projection. The sex can be set to “Both,” “Male,” or
“Female.” The display interval will normally be in single years
but you can change it to display every five or ten years if
desired. The chart type is also set through this dialogue box.
Click on the button next to the type of display you want.
Normally the display will show all the years in the projection.
However, if you want to see only part of the projection, you
can change the final year by selecting a new final display
year from the “Final year” list box.

63
 Once you are satisfied with the type of display, click the
“Ok” button and the display will appear. It will look
something like the display shown below.

All the projections that are currently in use will be displayed

on the same graph.
You can change the configuration of the display by clicking
the “Configure” button. You can also change the type of
display by putting the mouse pointer anywhere inside the
chart and clicking with the right mouse button.
To close the display, click on the “Close” button. You do not
have to close the display immediately. You can choose to
display another indicator and it will appear on top of the first
display. The first display will be covered but it will still be
there. You can return to any previous display that you have
not closed by choosing “Window” from the menu bar and
selecting the name of the display from the pull-down menu.
From the “Window” selection you can also choose to tile or
cascade all the existing display windows.

64
1. Graphs and Bar Charts
Spectrum will display a variety of graphs and bar charts,
including:
• Line charts
• Two- and three-dimensional bar charts (column charts)
• Two- and three-dimensional horizontal bar charts
• Two- and three-dimensional overlap bar charts (bars for
multiple projections are shown on top of each other)
• Three-dimensional perspective bar charts
To print the active chart, select “File” from the menu bar
and “Print” from the pull-down menu.

2. Tables
Spectrum will also display data in the form of tables. In
tables, each projection that is in use will be displayed in a
separate column. You can scroll through the table to see all
the years using the PgUp and PgDn keys or the mouse.
To print a table, select “File” from the menu bar and “Print”
from the pull-down menu.

3. Displaying All Age Groups

If you wish to see the population size by age and sex,
choose “Display,” “Demography,” “Age groups,” and then
“All age groups.”
You can display the information as a table or as a
population pyramid (see glossary) showing either numbers of
people (“Pyramid (numbers)”) or the percent distribution by
age and sex (“Pyramid (percent)”).
The pyramid display always shows two pyramids. If you are
using a single projection, then the pyramid on the left will
always be for the base year. You can change the year for
the pyramid on the right by clicking one of the buttons at
the bottom of the screen to advance the pyramid one year
(“Next”), show the previous year (“Previous”), show the first
year (“First year”) or show the last year (“Last year”).
If you have two projections loaded, then the pyramid on the
left will display the first projection and the one on the right
will show the second projection. Both pyramids will display
the same year; using the “Next” and “Previous” buttons will
change both pyramids.

65
 If you have more than two projections loaded, you will be
asked to choose which two pyramids should be shown
before the pyramids appear.

4. Demographic Summary Table

The final display available in DemProj is a summary table
showing all the demographic indicators. You can scroll
through this page to see all the output. If you have more
than one projection loaded, the indicators for the second
projection will immediately follow the first. To print a table,
select “File” from the menu bar and “Print” from the pull-
down menu.

G. Saving the Projection

It is always a good idea to save the projection to your hard
drive whenever you make a change to any assumptions. To
save the projection without changing the name, choose
“File” from the menu bar and “Save projection” from the
pull-down menu.
To save the projection with a different name, choose “File”
from the menu bar and “Save projection as” from the pull-
down menu. You will then have a chance to specify a new
file name for the projection. Normally when you save the
projection with a new name, you should also change the
projection title. This will avoid confusion if you have both
projections loaded at the same time.

H. Opening an Existing Projection

If you have already created a DemProj projection or are
using a projection provided by someone else, you can
immediately load that projection.
1. Select “File” from the menu bar.
2. Select “Open projection” from the pull-down menu.
3. Select the file you wish to use and click the “Ok” button
to open the projection.
You can open more than one projection at a time. Just
repeat these steps to load a second or third projection.
When you have more than one projection loaded, all
projections will be displayed in the graphs and tables. The
number of projections you can load at any one time is
determined by the amount of available memory in your
computer.

66
 When you have more than one projection loaded, you will
be asked to choose a projection when performing certain
tasks, such as editing assumptions. The program will display
a list of the projection names and you may choose the
appropriate one from the list.

I. Closing a Projection
To close a projection that has already been opened,
1. Choose “File” from the menu bar, and
2. “Close projection” from the pull-down menu. If you
have more than one projection loaded, you will be
asked to select which projection should be closed.
Closing a projection just removes it from the computer’s
memory; it does not erase it from the hard disk. You can
open that projection again at any time.

L. Aggregating projections

If you want to add together several projections to produce

a total for a region, you can do that with the “Aggregate”
function. This function will read the projections you specify
and combine then into a single projection file. All number
indicators referring to numbers of people will be added
across all projections. Rates, such as TFR and life
expectancy, will be calculated as weighted averages.
Weighting will be done by an appropriate population. For
example, the TFR for the aggregate population will be
calculated as the weighted average of the TFRs for each of
the input projection, weighted with the number of women of
reproductive age (15-49). To use this feature, follow these
steps:

1. Select “File” and “Aggregate” from the Spectrum menu.

(Note: this command not available if you have any
opened projections. If the command is not available, first
select “File” and “Close” and close all projections.)
2. Click the “Add” button to select the projection files to
be aggregated. You can select as many files as you
want. Once you have selected all the projection files,
you can save the list, if you may want to use it again, by

67
 clicking the “Save” button. To remove files from the list,
click “Remove”. If you have previously created a file list,
you can use it by clicking the “Open” button and
selecting the file.
3. Once you have the files you want to aggregate, click
“Process”. Spectrum will read each of the files in the list
and create a new file. Specify a name for the new file.
4. Once you have completed processing and provided a
new file name, click the “Close” button.
5. You can now load the new file you just created and
examine any of the DemProj indicators.

68
69
V. Methodology

DemProj calculations are based on the standard cohort

component projection modified to produce a single-year
projection.

A. Projecting the population by single ages

The inputs to the demographic projection are:

• Pop5(a,s): Population by 5 year age groups (a) and

sex (s) in the base year
• TFR(t): Total fertility rate by year
• ASFD(a,t): Distribution of fertility by age by year
• SRB(t): Sex ratio at birth by year
• LEB(s,t): Life expectancy at birth with AIDS by sex and
year
• Model life table
• Migration(a,s,t): Net in-migrants by age, sex and time

We assume that the base population is a mid-year estimate

and that the rates (TFR, life expectancy and migration are
calendar year averages).

The first step is to split the base year population, which in 5

year age groups, into single ages. This is accomplished by
using the Beers procedure (Beers, 1945). This procedure uses
a series of polynomial equations to divide the population in
5 year age groups into single year age groups while
maintaining the population total and providing a smooth
transition from one age to the next. The produces a base
year population by single age: Pop(a,s,to).

The age distribution of migration is also specified in 5 year

age groups. Migration is also split into single age groups
using the Beers procedure.

Mortality is specified as life expectancy at birth by sex and a

model life table. The life tables provide survival rates by
single year of age for certain values of life expectancy, such

70
as 20, 25, 30, 25, and so on. DemProj interpolates between
these index values to find the age-specific survival rates for
the exact life expectancy specified for each year.

Model life tables are provided in five year age groups. These
are split into single age survival rates, S(a, a+1), using the life
table indicators for nLx (the number of person-years lived
between ages x and x+n by an initial cohort of 100,000
people) and lx (the number of survivors at age x out of an
original cohort of 100,000 people.

1. Spop(0) = 1L0
2. Spop(1) = (l1 + l2) / 2
3. Spop(2) = (l2 + l3) / 2
4. Spop(3) = (l3 + l4) / 2
5. Spop(4) = (l4 + l5) / 2
6. Spop(5..80+) calculated by applying Beers procedure
to nLx values
7. Calculate S(a,a+1) = Spop(a+1) / Spop(a)

The number of deaths from mid-year to mid-year is

calculated as:

Deaths(a,s,t-1,1) = (Pop(a-1,s,t-1) + [migration(a-1,s,t-

1)+migration(a-1,s,t))/2] * ( 1 – [S(a-1,a,t-1)
+ S(a-1,a,t))/2] )

The number of deaths during the calendar year t is:

[Deaths(a,s,t-1) + Deaths(a,s,t)]/2

The population is projected by age and sex for ages 0 to 79

as

Pop(a,s,t) = Pop(a-1,s,t-1) + [migration(a-1,s,t-1) +

migration(a-1,s,t)]/2 – deaths(a,s,t-1,t)

The number of births from mid-year to mid-year is

Births(t-1,1) = ∑a=15,49 [Pop(a,female,t-1) +

Pop(a,female,t)]/ 2 * [TFR(t-1) + TFR(t)] / 2 *
[ASFD(a,t-1) + ASFD(a,t)] / 2

71
 The number of births during the calendar year is

[Births(s,t-1) + Births(s-t)]/2

The population of age 0 is

Pop(0,s,t) = (Births(s,t-1) + Births(s,t))/2) * Spop(0)

72
Figure 13: Fertility Distribution by TFR: UN Sub-
Saharan Africa Pattern

percent distribution
40

35

30

25

20

15

10

0
15-19 20-24 25-29 30-34 35-39 40-44 45-49
age group
Africa TFR=2 Africa TFR=3 Africa TFR=4
Africa TFR=5 Africa TFR=6 Africa TFR=7

Figure 14: Fertility Distribution by TFR: UN Arab

Pattern

73
 percent distribution
35

30

25

20

15

10

0
15-19 20-24 25-29 30-34 35-39 40-44 45-49
age group
Arab TFR=2 Arab TFR=3 Arab TFR=4
Arab TFR=5 Arab TFR=6 Arab TFR=7

Figure 15: Fertility Distribution by TFR: UN Asia

Pattern

percent distribution
40

35

30

25

20

15

10

0
15-19 20-24 25-29 30-34 35-39 40-44 45-49
age group
Asia TFR=2 Asia TFR=3 Asia TFR=4
Asia TFR=5 Asia TFR=6 Asia TFR=7

74
B. Urban and Rural Projections
The methodology of projecting urban and rural populations
follows the United Nations method of growth rate difference.
In this method the size of the urban population is projected
as follows:
urbant = urbant-1 • ( totalt + URGD • ruralt-1) / totalt-1
where:
urbant = urban population at time t
rural = rural population
total = total population
URGD = urban rural growth rate difference.
This equation is actually a logistic function. The urbanization
rate will slow after the percent of the population which is
urban passes 50 percent.
The URGD is a combination of the difference in the urban
and rural growth rates actually experienced in the most
recent time period and the hypothetical difference
calculated with the equation

URGD' = 0.044 - 0.028 • initial fraction urban.

The rural population is calculated as the difference between

the total population and the urban population.
This method is described fully in Methods for Projections of
Urban and Rural Population, Manual VIII (United Nations,
1974).
The age distributions of the urban and rural populations are
also calculated as logistic functions. This procedure makes
use of the logit function. The logit of the percent urban is
defined to be 0.5• ln (percent urban/[1 - percent urban]).
In this approach the logit of the percent urban in each age
group is assumed, as a first approximation, to increase as the
logit of the total percent urban.
As an illustration, consider the steps in calculating the size of
the 10-14 urban age group in 1985. First we calculate the
increase in the logit of the percent of the total population
that is urban:

IUL = logit(% urban in 1985) - logit(% urban in 1980),

75
 where:
IUL = increase in urban logit.

Next, we assume that the logit of the percent of the 10-14

age group that is urban will increase by the same amount:

logit(% 10-14 urban in 1985) = logit(% 10-14 urban in 1980) + IUL.

The logit is then transformed back into a percentage and

the size of the 10-14 urban age group is calculated:

% 10-14 urban = ( e2 • logit[% 10-14 urban] ) / ( 1 + e2 • logit[% 10-14 urban] )

urban pop 10-14 = total pop 10-14 • (% 10-14 urban).

These calculations are carried out for each age group. Then
all the age groups are normalized to sum to the total urban
population. The size of the rural age groups is simply the size
of the total age group minus the size of the urban age
group.

76
VI. References
Barkalov, N. B. 1984. Modeling the Demographic Transition.
Moscow: Moscow University Press.

Barkalov, N. B. and S. Kirmeyer (n.d.). 1989. New Shapes for

Fertility Analysis. Unpublished.

Beers, H. S. 1945. “Six-Term Formula for Routine Actuarial

Interpolation,” The Record of the American Institute of
Actuaries, 34 Part I (69): 59-60.

Bos, E., My T. Vu, E. Massiah and R. A. Bulatao. 1994. World

Population Projections: 1994-95 Edition. Baltimore: John
Hopkins University Press.

Coale, A. J., P. Demeny and B. Vaughan. 1983. Regional

Model Life Tables and Stable Populations: Second Edition.
New York: Academic Press.

Coale, A. J. and T. J. Trussell. 1974. “Model Fertility Schedules:

Variations in the Age Structures of Childbearing in Human
Populations,” Population Index, 40 (2): 185-258.

Coale, A. J. and T. J. Trussell. 1978. “Technical Note: Finding

the Two Parameters That Specify a Model Schedule of
Marital Fertility,” Population Index, 44 (2): 203-213.

Coale, A. J. and J. Trussell. 1996. “The Development and Use

of Demographic Models,” Population Studies, 50: 469-484.

Horne, A. Dale and M. N. El-Khorazaty. 1996. “Childbearing

and Bongaarts Indices for Coale-Trussell’s Model Fertility
Schedules,” Genus, LII (1-2): 161-180.

Jamison, E. and F. Hobbs. 1994. World Population Profile:

1994. Washington, DC: U.S. Bureau of the Census.

Population Reference Bureau. 1989. Population Handbook:

International Edition. Second Edition. Washington, DC:
Population Reference Bureau.

77
 Stover, J. 1994. DemProj: A Demographic Model for
Development Planning, Version 3. Glastonbury, CT: The
Futures Group International.

United Nations. 1995. World Population Prospects: The 1994

Revision. Sales No. [Link].16. New York: United Nations.

United Nations. 1993. World Population Prospects: The 1992

Revision. Sales No. [Link].7. New York: United Nations.

United Nations. 1989. World Population Prospects: 1988.

Department of International Economic and Social Affairs.
Population Studies, No. 106. New York: United Nations.

United Nations. 1983. Manual X. Indirect Techniques for

Demographic Estimation. New York: United Nations.

United Nations. 1982. Model Life Tables for Developing

Countries. Department of International Economic and
Social Affairs. Population Studies, No. 77. New York: United
Nations.

United Nations. 1977. World Population Prospects as

Assessed in 1973. Department of Economic and Social
Affairs, Population Studies, No. 60. New York: United Nations.

United Nations. 1974. Methods for Projections of Urban and

Rural Population, Manual VIII. New York: Department of
Economic and Social Affairs, United Nations.

U.S. Bureau of the Census. 1994. World Population Profile:

1994. Washington, DC: U.S. Bureau of the Census.

Van de Walle, E. and Henry, L. 1982. Multilingual

Demographic Dictionary. International Union for the
Scientific Study of Population (IUSSP). Liège, Belgium: Ordina
Editions.

78
VII Glossary of Terms
. Many of the following terms were obtained
from the Population Reference Bureau's
Population Handbook (1989); others were adapted from the
International Union for Scientific Study of Population’s
(IUSSP’s) Multilingual Demographic Dictionary (Van de Walle
and Henry, 1982).
Age-specific fertility rate (ASFR). The number of births in a
year for a population of a given age and sex to the mid-
year population of that same sex and age group.
Annual growth rate (GR). The rate at which the population is
increasing or decreasing in a given year due to natural
increase and net migration, expressed as a percentage of
the base population.
ASCII (text) file. A file of text or data in a standardized,
character format, which is transportable between software
applications and operating systems.
Births. The total number of annual births.
Child-woman ratio. The number of children under the age
of five per woman of childbearing age (15-49).
Coale-Demeny model life tables. Families of life tables
derived from empirical life tables of Europe and other
industrialized nations of the world. They are based primarily
on data from the first half of the 20th century.
Coale-Trussell relational fertility model. This model
incorporates (1) the timing of marriage, (2) a population’s
natural fertility, and (3) the degree of fertility regulation in a
population. It permits users to generate a fertility schedule
from an empirical baseline and a specific total fertility rate.
Cohort. A group of persons who experience certain events
within a specified period of time, such as those who are born
or who are married in the same year.
Cohort component projection. A projection made by
subjecting all cohorts, on an annual or five-year basis, to
mortality and migration assumptions, and applying fertility
assumptions to women of reproductive age.

79
 Crude birth rate (CBR). The number of live births per 1,000
population in a given year.
Crude death rate (CDR). The number of deaths per 1,000
population in a given year.
Deaths. The total number of annual deaths.
Dependency ratio. The ratio of the economically
dependent part of the population (those aged 0-14 and 65
and over) to the productive part (those aged 15-64).
Dialogue box. A box permitting users to choose among a
limited number of options. The box is accompanied by text
elaborating upon those options.
Doubling time. The number of years it would take for the
population to double its current size at the current annual
rate of growth.
Fifth root. The fifth root of the number “n” is the number
which, multiplied by itself five times, exactly equals “n.” For
example, the fifth root of 243 is 3 (3x3x3x3x3).
Gross migration. The total number of people who move to
and depart from an area.
Gross reproduction rate (GRR). The average number of
daughters that would be born to a woman (or a group of
women) during her lifetime if she passed through all her
childbearing years conforming to the age-specific fertility
rates of a given year. This is similar to the TFR except that it
counts only daughters.
Infant mortality rate (IMR). The number of deaths to infants
under one year of age per 1,000 live births.
Initial age distribution. The age-sex distribution at the
beginning of a projection period.
Interpolation. Given two numbers that serve as boundary
points, it is possible to estimate the values that lie at intervals
between the two points. For example, if the total fertility rate
for a country or region was actually measured only in 1980
and in 1995, by assuming even increments from year to year,
it is possible to interpolate a TFR for each intervening year.
(Spectrum uses a linear form of interpolation so that the
difference between each annual value is the same. Other
nonlinear forms of interpolation also are possible, but are not
used in Spectrum.)
Life expectancy (e(0)). The average number of years a
newborn can expect to live based on the mortality
conditions at the time.

80
Life table. A table of values based on a series of related
functions having to do with survivorship over intervals of
time.
Logit. A function of the real number x (where x must be
between 0 and 1), which is defined as logit(x) = 2 ln(x/(1-x)).
Mean age of childbearing. The average age of mothers at
the time of birth.
Median age. The age that divides a population into two
numerically equal groups.
Model. Computer system designed to demonstrate the
probable effect of two or more variables that might be
brought to bear on an outcome. Such models can reduce
the effort required to manipulate these factors and present
the results in an accessible format.
Model life table. A table of values based on a series of
related functions having to do with survivorship over intervals
of time.
Module. Synonym for “model.”
Net migration. The number of people who move to an area
minus those who move away.
Net reproduction rate (NRR). The average number of
daughters that would be born to a woman (or a group of
women) during her lifetime if she passed through all her
childbearing years conforming to the age-specific fertility
rates and age-specific mortality rates of a given year. This is
similar to the GRR except that it includes the effect of
mortality that would cause some women to die before
completing their childbearing years.
Normalized. The transformation of a series of data points to
a percent distribution summing to 100 percent.
Population aged 0-4. The portion of the population under
age five. It is often of interest as it may be subject to high
mortality, and consequently may require a large amount of
health care relative to other age groups.
Population aged 5-14. The portion of the population which is
generally considered to be of school age.
Population aged 15-64. The portion of the population which
is generally considered to be of employment age.
Population aged 65+. The portion of the population which is
generally considered to be of post-employment age.

81
 Population projection. Computations depicting the future
course of a population’s size, its structure, and its interaction
with dynamics such as fertility, mortality, and migration. The
projection is constructed based on assumptions about the
future course of those population dynamics.
Population pyramid. A graphic that portrays an age-sex
distribution. It is so named because of its shape.
Pop-up menu. A menu from which users can select items or
actions. Pop-up menus can appear anywhere on the
screen.
Probability of dying (qx). The probability that an individual
(or group of individuals) having lived to exact age x will die
before reaching exact age x+5.
Probability of survival (sx). This term is the complement to
the probability of dying. It refers to the probability that an
individual who has reached exact age x will go on to live to
exact age x+5.
Pull-down menu. A menu opened by clicking on key words
at the top edge of the screen. Pull-down menus allow users
to select operations.
Quinquinnial. Five-year age groups. The basis for most
cohort component projections is five-year age groups over
five-year intervals.
Radio button. These buttons emulate raised buttons on early
radios, which were punched to select radio stations. The
graphically portrayed raised “radio buttons” on interfaces
permit users to select among at least three alternatives.
Rate of natural increase (RNI). The rate at which the
population is increasing or decreasing in a given year due to
the surplus or deficit of births over deaths, expressed as a
percentage of the base population.
Semi-log graph. A graph of data points for which one axis is
graduated logarithmically, and the other is not, almost
always being arithmetic instead.
Sex ratio. The number of males per 100 females in a
population.
Survival ratio. The proportion of the population of a
particular age that survives to the next age in the next year.
Synthetic (hypothetical) cohort. A cross-section of the
population whose various demographic experiences (such
as mortality chances) are consolidated into summary
indicators, as if to a true (birth) cohort.

82
Total fertility rate (TFR). The average number of children that
would be born alive to a woman (or a group of women)
during her lifetime if she were to pass through all her
childbearing years conforming to the age-specific fertility
rates of a given year.
Under five mortality rate (U5MR). The number of deaths to
children under the age of five per 1,000 live births.
United Nations model fertility schedules. Five regional
schedules prescribed by total fertility rate.
United Nations model life tables for developing countries. A
set of families of life table schedules derived from empirical
life tables from developing countries, based on data from
the second half of the 20th century.
Vital statistics. Statistics on demographic events—births,
deaths, marriages, and divorces—which are recorded
through an event registration system.

83
84
VII Acronyms and
Abbreviations
I.
AIDS acquired immune deficiency syndrome
AIM AIDS Impact Model
ASFR age-specific fertility rate
CBR crude birth rate
CDC U.S. Centers for Disease Control and Prevention
CDR crude death rate
CPS Contraceptive Prevalence Surveys
DHS Demographic and Health Surveys
e(0) life expectancy
GDP gross domestic product
GNP gross national product
GR annual growth rate
GRR gross reproduction rate
HIV human immunodeficiency virus
IMR infant mortality rate
NRR net reproduction rate
RAPID Resources for the Awareness of Population
Impacts on Development [project]
RNI rate of natural increase
TFR total fertility rate
U5MR under-five mortality rate
UN United Nations
USAID United States Agency for International
Development
WFS World Fertility Surveys

85
86
 Appendix A

The Coale-Trussell Relational

Fertility Model
Appendix A: The Coale-Trussell
Relational Fertility Model

A.1 Description of the Model

This model, first introduced in 1974 (Coale and Trussell, 1974,

1978; United Nations, 1983), is the most widely used model of
the age composition of fertility. The model decomposes
age-specific fertility rates into three factors corresponding to
basic fertility determinants:

• natural fertility: hypothetical fertility that might exist in

the absence of birth control, if all women were in
sexual unions during their entire reproductive spans.
• birth control: deliberate control over childbearing by
way of contraception and/or abortion.
• cohabitation (consensual or marital): time spent by
women within sexual unions, with time in unions being
shortened due to premarital sexual abstinence, spousal
separation, or dissolution of the union.

The model formalizes the relationship between age-specific

fertility and its determinants in a very simplified form
(Barkalov and Kirmeyer, n.d. [1989]). It assumes that:

• natural fertility within unions is proportional to a certain

age schedule that is approximately the same for
different populations;
• the intensity of birth control is also proportional to a
standard age schedule; and
• the age shape of the proportion currently in unions is
similar to the age-specific proportion of ever-married
individuals in a female population.

A-1
 Formally, according to the model, any set of the
age-specific fertility rates, fx, is graduated as follows:

fx = Gx •M•nx• e− m• v x

Elements of this formula are the following:

Functions nx and vx are model standard schedules constant

over populations:

nx is the standard schedule of natural fertility;

vx is a standard pattern of birth control impact on
fertility.

In addition,
m is the model parameter of birth control.

It might considered an index of Αbirth control≅ measuring

the degree of fertility reduction below the natural level. The
shape of vx, however, leads to the conclusion that the birth
control factor represents stopping behavior (a policy to
terminate childbearing after a certain parity) rather than
spacing behavior (a policy to postpone childbearing).

M is the level parameter.

Since natural fertility may vary significantly from population

to population, the parameter M might be explained as a
natural fertility level, influenced in a population of women in
union primarily by breastfeeding. However, it also represents
the intensity of fertility regulation applied within young ages
(postponing, basically) where vx is set to 0. And thus, values
of M exceeding 1 are quite normal (see Coale and Trussell,
1978). In addition, the parameter M partially reflects the
average impact of marriage/union interruption.

Gx is the model proportion who ever married.

A-2
 Gx is formalized, based on a standard density function,
which in turn takes two parameters:

• the singulate mean age at marriage (SMAM), which is

the arithmetic mean age at first marriage; and
• the initial nuptial age (a0), which is the age at which a
significant number of sexual unions start.

The value of a0 has a narrow range of reasonable variation.

Usually it runs from 12 (some south Asian and Latin American
countries) to 17 (some European countries).

A.2 Employment of the Model in Projections

The Coale-Trussell model is simplified in DemProj when used

in the projection of fertility schedules. It is assumed that in a
projection period, the fertility change which occurs would
affect primarily the stopping patternsΧor, for higher-order
births, the birth-spacing pattern. And it assumes that the
marriage pattern and the pattern of spacing for lower-order
births would not be altered. Thus, the model would
demonstrate a relationship between the projected age-
specific fertility rates, fa, and the (probably empirical)
reference fertility rates, f0,a, as:

fa = f0,a • (e− va ).

Here, va is the Coale-Trussell standard age-specific schedule

of fertility control (values given in Table A.1) and Δ is the
parameter of change in the projection period. If this
parameter were known, then the projected age-specific
fertility rates could be uniquely generated (and their values
would always be non-negative). Target aggregate fertility
(the total fertility rate, or TFR) is represented as:

TFR = ∑ fo,a • (e
a
− va
).

A-3
 Table A.1: Coale-Trussell Fertility Control Schedule (vx)

Age Group Fertility Control Schedule

15-19 -0.069
20-24 -0.069
25-29 -0.279
30-34 -0.667
35-39 -1.042
40-44 -1.414
45-49 -1.671

If the proposed TFR were not too small when compared to

the original TFR, the equation would have a unique solution,
permitting projection of fertility rates for a broad range of
levels. The solution could be found with any nonlinear
method. A bisection with inverse quadratic interpolation
was used for the algorithm implemented into DemProj
(Barkalov, 1984). The approximation for the change
parameter, Δ, is found by letting

H1 = ∑ f o,a •v a

and

H 2 = ∑ fo,a • v a 2
a

where H1 and H2 are the arguments from the quadratic

equation that are used to solve for Δ, the parameter of
change over the projection period.

Then when the original TFR is greater than the projected TFR,

2
H1 - (2 • D • H2 + H1 )
r= ,
H2 < 0

A-4
 where D is the difference between the projected TFR and
the TFR from the empirical source (such as a survey).

A.3 An Application to Bangladesh

As an example, the model was employed to project two sets
of fertility rates from Bangladesh. The first set came from the
Bangladesh Fertility Survey of 1975, when Bangladeshi fertility
was high and had the characteristic plateau shape (see
Figure A.1) of high-fertility behavior. The second set of
reference fertility rates came from the Bangladesh Fertility
Survey of 1989, following a substantial decline in fertility and
transition to a peaked shape of fertility rates (see Figure A.2).
These empirical rates are given in the upper part of Table
A.2.

The results of the fertility schedule projection are rather

different. They allude to the key feature of this type of
fertility model: It is relational. Thus, the projected schedules
bear a relationship to the components of the empirical
reference schedule. In the case of the 1975 schedule (as
noted above), it has the typical plateau shape of high-
fertility countries. But it is atypical in that the pattern has a
leftward "tilt," with higher (relative) concentration than usual
for younger-aged fertility. As a result, the Coale-Trussell
projected pattern is more peaked at younger fertility than
are the shapes generated with reference to the 1989 data.
Because they have fewer idiosyncracies, then, the reference
data of 1989 provide a good set of projections to use in
matching the next round of data collected in Bangladesh.

A-5
 Figure A.1: ASFRs Generated by Coale-Trussell
Model Compared with Survey ASFRs:
Bangladesh,
BFS 1975

(reference fertility schedule BFS 1975)

births per thousand women

300

250

200

150

100

50

0
15-19 20-24 25-29 30-34 35-39 40-44 45-49
age group

1975 BFS TFR= 6 TFR= 5 TFR = 4 TFR = 3

Figure A.2: ASFRs Generated by Coale-Trussell

Model Compared with Survey ASFRs:
Bangladesh,
BFS 1989

A-6
 (reference fertility schedule BFS 1989)

births per thousand women

300

250

200

150

100

50

0
15-19 20-24 25-29 30-34 35-39 40-44 45-49
age group

1989 BFS TFR= 6 TFR= 5 TFR = 4 TFR = 3

Table A.2: Bangladesh: ASFRs from Various Surveys, and Projected Rates

ASFR, by Source
1993-94
Age Group 1975 BFS 1989 BFS 1991 CPS DHS
15-19 109 182 179 140
20-24 289 260 230 196
25-29 291 225 188 148
30-34 250 169 129 105
35-39 185 114 78 56
40-44 107 56 36 19
45-49 35 18 13 14

Empirical TFR: 6.33 5.12 4.27 3.39

Reference Fertility: 1975 Bangladesh Fertility Survey

15-19 108 105 101 95

20-24 287 279 269 251

25-29 283 254 216 164

30-34 234 181 123 63

35-39 166 111 61 22

40-44 93 54 24 6

45-49 30 16 6 1

Projected TFR: 6 5 4 3

A-7
 ASFR, by Source
1993-94
Age Group 1975 BFS 1989 BFS 1991 CPS DHS
(Coale-Trussell model)

Reference Fertility: 1989 Bangladesh Fertility Survey

15-19 161 181 174 161

20-24 266 259 248 230

25-29 248 222 186 136

30-34 213 163 108 51

35-39 163 108 57 17

40-44 91 52 22 4

45-49 32 16 6 1

Projected TFR: 6 5 4 3
(Coale-Trussell model)

A-8
A-