POPULATION REFERENCE BUREAU, INC. 1755 MASSACHUSETTS AVE., N.W., WASHINGTON, D. C. 20036 (202) 232-2288
POPULATION STATISTICS: WHAT DO THEY MEAN?
An exchange of letters took place recently in
the Portland (Oregon) Press between a local resi-
dent and a representative of the Portland chapter
of Zero Population Growth, Inc. At one point in
the exchange the citizen asked: "Shall we use
ZPG's statistics or those used from the Census
Bureau...?" The ZPG man was quick to respond
that he used Census Bureau statistics, too.
The episode is typical of the confusion that
is generated when complex, controversial subjects
are discussed in terms of numbers. Statistics are
powerful weapons in a controversy; all too often
they are used improperly, inaccurately, and even
dishonestly. Those who wish to understand a sub-
ject, and to judge the merits of opposing views,
must be able to spot the occasions when statis-
tics are used incorrectly. Those who wish to take
part in, or report on, a controversial subject have a
further task: to make sure the numbers they quote
are both accurate and appropriate to the point
The major purpose of most population statis-
tics is to describe how populations have changed and
are changing, and--most difficult but also most inter-
esting--how they will change in the future. Demog-
raphers have developed a number of measures to
describe the behavior of populations. Each has ad-
vantages--and defects. This Profile outlines the most
common measures of population dynamics, and in-
dicates how they can most profitably be used.
There is another side to demographic statis-
tics that, unless it is continually kept in mind, can
also be misleading. That is the question of accuracy.
Demographers are constantly aware that the num-
bers they use may not be very reliable, but some-
times laymen forget it. Even in advanced countries
where statistics have been collected for many years,
the accuracy of numbers that represent the behav-
ior of millions of people often is open to question.
And sometimes, especially for indicators based on
sophisticated computations, available figures can
A good way to illustrate both the quality of
demographic data and some of the major demogra-
phic measures is to describe how one of the simp-
lest figures--the total national population--is devel-
oped. The basic measures used for estimating total
population come from the NATIONAL CENSUS.
Most countries today--there are some notable ex-
ceptions including the world's most populous coun-
try, China--conduct population censuses on a fairly
regular basis. Some censuses collect information on
a broad range of subjects, while others are aimed
simply at counting the number of people in the
country at a given time.
If a national census involves a genuine effort
to count every person, rather than just estimating
populations on a regional basis, the most common
error is likely to result from the fact that some peo-
ple just didn't get counted.
The size of the undercount may be estimated
by conducting a post-census survey in which a rep-
resentative sample is asked whether they were count-
ed during the census. The proportion of those
sampled who had not been counted gives an esti-
mate of the size of the undercount. In the 1960
U.S. Census, for example, the undercount was esti-
mated to be 3 percent, or almost 6 million
people. The estimate of the 1970 census under-
count has not yet been made by the Census Bureau.
A census gives more or less accurate estimate
of the number of people in a country at one point
in time. But censuses are taken too seldom to be
satisfactory for many uses. Often, estimates are
needed of the population at times between censuses.
One way of making such estimates is to use
the INTERCENSAL GROWTH RATE. This fig -
ure is computed by subtracting the population fig-
ure of the previous census from that of the current
census, which yields the intercensal population
growth. In the United States, for example, the
1970 Census population was 203 million, and the
1960 population was 179 million.
INTERCENSAL = [1970 POPULATION]
GROWTH [1960 POPULATION]
This figure is converted to an average annual
growth RATE by dividing it by the number of
years between censuses (10 years in the case of the
U.S.) and dividing the result by the total popula-
tion of the earlier census year:
AVERAGE YEARS BETWEEN CENSUSES
GROWTH RATE EARLIER (1960) POPULATION
= 1.3 PERCENT PER YEAR
The growth rate can be used to make a rough
estimate of the population in the years between
censuses. If the population had continued to grow
at 1.3 percent per year through 1971, then the
population for that year would be 1.3 percent grea-
ter than the 1970 figure, for a total of 207 million.
This estimate is based on a statistical project-
ion of the 1970 population in which it is assumed
that the population continued to grow at 1.3 per-
cent. Most such projections of where a population
is going are based on an assumption of this sort:
that the rate of change will remain the same as it
was when last measured (or that it will vary in a
certain way). Rates can change unpredictably,
however. A given rate is accurate only at the time
it is measured. Any extension beyond that period,
especially for more than a short time, is open to
Another way to estimate population between
censuses involves the use of three other measures:
The CRUDE BIRTH RATE measures the
number of babies born in one year for each 1,000
persons in the population at the midpoint of that
BIRTH RATE = NUMBER OF BIRTHS PER YEAR 1000
Similarly, the CRUDE DEATH RATE is the
number of deaths in one year per 1,000 population:
NUMBER OF DEATHS PER YEAR
DEATH RATE = X 1000
NET MIGRATION is the difference between
the number of people who enter the country in one
year (immigration) and the number who leave
NET MIGRATION = IMMIGRATION EMIGRATION
Net migration can be either positive, with
more immigrants than emigrants, or negative, with
more people leaving than entering.
All three of these measures are complicated
and difficult to obtain, especially for the less devel-
oped countries. They require that up-to-date birth
and death records be kept on the local level, and
that the results be forwarded to a central authority
for integration on a national basis. Even in the
United States, it was not until 1933 that all the
states joined the federal system of birth and death
registration, and many countries still do not have
Still, even partial registration can be used to
estimate birth and death rates, and together these
two indicators give a measure of the NATURAL
INCREASE of a country (excluding migration):
RATE OF NATURAL [BIRTH RATE DEATH RATE]
Since birth and death rates are measured as so-
many per 1,000 population, the difference is divid-
ed by 10 to yield increase per 100 population, or
In the United States in 1970, the birth rate
was 18 per 1,000, and the death rate was 9 per
1,000. That means that the rate of natural increase
was (18-9)/10 = 0.9 percent.
~Lj o ^ .< /.
Net migration can be converted to a "pr-
1,000" figure and added into the equation to yield
the GROWTH RATE:
BIRTH_ DEATH + MIGRATION x 1000
L RATE RATE j + X 1000
GROWTH RATE R POPULATION
In the United States in 1970, net migration
was 400,000, or 2 per 1,000 population. So the
total growth rate was (18-9 + 2)/10 = 1.1 percent.
This figure can be compared with the average
annual Intercensal Growth Rate computed earlier.
That figure was 1.3 percent; so the annual popu-
lation growth rate in the United States has dropped
compared with the average rate during the 1960s.
For projections of population growth over
more than a year or two, demographers like to
have even more detailed indicators than births and
deaths. Most of these measures depend on a know-
ledge of the AGE STRUCTURE of a population.
Age structure has significant effects on popu-
lation growth. Basically, the rate of natural increase
depends on two factors: the rate at which women
in the reproductive age group are having babies,
called FERTILITY, and the proportion of women
who are in the child-bearing period of their lives.
In the United States,for instance,the children born
during the high-fertility years of the 1950s, more
numerous than those born in the previous decade,
will be entering the ranks of parents during the
1970s and 1980s. Even if fertility remains at its
present low level, the number of babies born per
1,000 total population--the birth rate--is likely to
be higher. Fertility would have to drop below pre-
sent levels for the birth rate to remain unchanged.
Fertility is measured in several different ways.
Each has advantages and disavantages, which makes
it more or less useful in projecting future popula-
GENERAL FERTILITY is the number of
children born each year per 1,000 women in the
reproductive age group (15 to 44 years in the Uni-
ted States). This measure takes the age structure
into account to a certain extent, unlike the birth
rate. In 1971 the general fertility rate was 82.3
births per 1,000 women. If general fertility stayed
at that level for 30 years, the length of the repro-
ductive cycle, that measure would correspond to
30X82.3=2469 births per 1,000 women, or 2.47
births per woman. But that is a highly artificial
measure, since it assumes constant fertility and
does not reflect changes in the age structure. For
this reason, while the general fertility rate is used
to measure changes in current fertility, other indi-
cators are computed to reflect longer term fertility
The AGE SPECIFIC FERTILITY RATE elim-
inates the effect of age structure entirely; it is the
number of births per year to 1,000 women of a
particular age. It can be computed (if the data are
available) for each single year of age during the re-
productive years, but it is usually computed for
five-year age groups: women 15-19 years old, 20-
24 years, 25-29 years, and so on.
The TOTAL FERTILITY RATE is based on
age-specific rates. It measures the total number of
children 1,000 women would have if they passed
through their reproductive years with the age-spe-
cific fertility of a particular year. For example, in
1968 the fertility rate for women aged 15-19 years
in the U.S. was 66 births per 1,000; for women aged
20-24 it was 167, and for women aged 25-29 it was
To compute the total fertility ratefor 1968,a de-
mographer would assume that 1,000 women would
have 66 births per year between the times when
they were 15 and 19, 167 births per year between
the ages 20 to 24, 140 births per year between ages
25 and 29, and so on, corresponding to the fertility
rates that existed for each age group in that one
Naturally, no group of 1,000 women is going
to experience exactly the fertility pattern assumed
in computing the total fertility rate. Age-specific
fertility is likely to change over the 30 years requir-
ed for a group of women to pass through their fer-
tile period. So the total fertility rate is hypotheti-
cal: It is another of those measures that need the
warning phrase: "If present rates continue".
There is a fertility measure that, unlike the to-
tal fertility rate, measures the number of children a
group of 1,000 women have actually had. That is
called the COMPLETED FERTILITY RATE. It
measures the total number of children born to
women who reach the end of their reproductive
cycle in the year the measure is taken. In the
United States in 1968, the completed fertility rate
for women aged 44 was 2.7 per 1,000 women.
At first glance it might seem that completed
fertility would be a much more reliable indicator
than the unreal figure represented by the total
fertility rate. The difficulty is that most children
are born to women in their 20s; so the completed
fertility figure measures the child-bearing behavior
of women who had most of their children 20 years
earlier. This makes it much less useful for estima-
ting what fertility behavior is like at present, or is
going to be like in the future.
ZERO GROWTH-THE REPLACEMENT LEVEL
Much of the discussion going on about popu-
lation today is concerned with the question of pop-
ulation growth, and especially on the question of
zero population growth. Using the indicators des-
cribed in the previous section, what can be said
about halting population growth in this country?
Some recent reports, drawing on declines in
the birth rate and in fertility, suggest that zero
population growth might be just around the corner.
But a further look at the statistics, in the light of
their limitations, shows that this possibility is
What is zero growth? Basically, it is a condition
under which the birth rate is equal to the death
rate (ignoring the effects of net migration).
Obviously, with the current 1971 birth rate
at 17 per 1,000 and the death rate at 9 oer 1.000.
Completed fertility for the cohort of women
who came out of their childbearing years in 1970s
(born 1925-1929) will be over 3. Quite clearly,
this cohort was nowhere near the replacement level
of fertility. But it has already been noted that the
completed fertility applies to women who were
doing most of their childbearing 15 or 20 years
earlier. It can tell what fertility behavior was in
the past, but it is not very informative about pres-
ent fertility behavior, and almost useless in pro-
jecting future behavior.
The measure of total fertility would seem to
be more useful, since it deals with current fertility
patterns. In computing total fertility, it is assumed
that women just entering the reproductive years
will follow the age specific fertility pattern for a
given year. If this age specific pattern remains the
same throughout the 30 years of their reproductive
cycle, these women would have a completed fertil-
ity equal to the total fertility that was computed
for that given year.
In 1968, the total fertility rate was 2.5. Pro-
visional figures for 1969 and 1970-based on 1968
age specific fertility rates-are 2.4 and 2.5. In 1957,
at the height of the baby boom,the total fertility
rate was 3.7. Looking at this decline, there is a
temptation to say that fertility is close to the re-
placement level of 2.11-particularly since general
fertility dropped in 1971.
the country is nowhere near ZPG. Its rate of natu- But there are two reasons for resisting this
ral increase is 8 per 1,000 or 0.8 percent per year 4 temptation. In the first place, total ertility is
a net increase in the U.S. population of about 1.7 only hythetical. It is based on the assume on
million people per year. In fact, as far as is known, a f tili will r a nhged or30 yers.
birth rates have never declined to close to the level But age-specific fertility could go up during that
of the death rate in the United States, even during period; it could go down. Age at marriage could
the low fertility period of the 1930s. change; so could the spacing of children born.
It could increase at the higher age levels and de-
But birth and death rates, although they crease at the younger ages, or vice-versa. Fertility
accurately reflect present population conditions, behavior is so greatly influenced by complex social,
are poor predictors of the future. This is because economic and psychological factors that predicting
they are so sensitive to variations in the age its future course is extremely chancy.
structure, and because fertility behavior can change The other reason for hesitancy is that the
id Iy. -total fertility rate is out of date. For several reasons,
Replacement Level of Fertility particularly budgetary ones, the National Center
for Health Statistics has fallen behind in computing
So the next step is to look at fertility rates. data on age-specific fertility, so that the latest avail-
Demographers have calculated that if women have able figures are for 1968 (with provisional 1969
a completed fertility rate of 2.11 children, there and 1970 figures). The lag in this important indica-
will be enough births to replace the parents and tor is particularly unfortunate because of the con-
compensate for premature deaths. This figure,2.11, troversy currently surrounding population ques-
is called the REPLACEMENT LEVEL of fertility, tions.
The general fertility rate, the number of child-
ren born each year per 1,000 reproductive-age
women is more current: The provisional figure for
1971, 82.3, has already been published. General
fertility, as has been shown, is a better indicator
of childbearing activity than the birth rate, be-
cause it takes the age structure into account. But
if total fertility is a chancy tool for estimating fu-
CHILDREN EVER BORN TO AGE 40, FOR WOMEN BORN 1880-1929
Totol births per weaon
to age 40
4.01 1 I I
0 1 I 1.- 1I I I, I-
1880- 1885- t890- 1895- 1)0- 1905- 1910- 1915- 1920-
1884 1889 1894 -1899 1 4 1909 1914 1919 1924
Stoe of Birth
Source: National Center for Health Statistics
Completed Fertility decreased for women who did most of their childbearing in the 1930s,
increased for post-World-War II mothers.
TOTAL FERTILITY AND GENERAL FERTILITY, 1915-1971
Source: National Center for Health Statistics
The decrease in fertility since the baby boom peak of 1957 parallels the earlier decline of
the 1920s and early 1930s.
ture population trends, the general fertility rate
is even more unreliable.
U.S. POPULATION-WHERE IT'S AT
The previous section was concerned with the
limitations on what can be gleaned from popula-
tion statistics. The general conclusion was that it
is impossible to say with certainty at this time
how close fertility is to the replacement level, and
most unwise to predict when replacement will be
reached. It is even less advisable to predict how
soon population growth will end.
But some conclusions can indeed be drawn
about current fertility behavior, based on data
available. At the least, what is known about the
present can be compared with the fuller details
that have been collected for earlier periods.
Completed fertility measures the childbearing
behavior of women coming out of their reproduct-
ive years. The most convenient figure to use is the
total number of children born to women up to the
age of 40, since few children are born beyond that
age. The curve shows that women who were born in
1885 to 1889, and who were at peak fertility just
before World War I, were averaging more than three
births each by age 40. Twenty years later, in the
1930's, women in their peak fertility years were
bearing fewer children: The average fertility by age
40 for this group was a little more than two births
per woman. Early in the 1950s, fertility was up
again; the group that was in its 20s in those years
reached an average of three births per woman at
The trend traced by the completed fertility
curve is also reflected in the total fertility indica-
tor, which was higher than three children per
woman in the early 1920s, dipped to 2.2 children
in the late 1930s, and rose to 3.7 children in the
late 1950s. Since then, the total fertility rate shows
a decline; in 1968, the latest figure available, it was
2.48 children per woman.
The general fertility rate similarly reflects these
changes. In the early 1920s it was well above 100
births per 1,000 women in the reproductive age
groups. It fell below 80 during the 1930s, then rose
to a peak of 121 in the late 1950s. Since then it
has decreased, reaching 82.3 in 1971.
These trends show that fertility has declined
dramatically in the 14 years since general fertility
peaked in 1957. But a look at history shows that
the decline was equally dramatic in the 12 years
from 1921 to 1933. General fertility has not yet
reached the low point of the 1930s, when it re-
mained below 80 per 1,000 for eight years.
What about replacement fertility and zero
population growth? The total fertility rate for
the low years of the 1930s was very close to the
replacement level. Despite these low levels, how-
ever, the country was never close to zero popula-
tion growth. The baby boom that took place in the
1940s and 1950s, and continued well into the
1960s, sent the population growth rate up a steep
Nor is the country close to zero growth today.
Fertility has not fallen to the levels of the 1930s;
even if it does, it will take many years of sustained
low fertility to overcome the effects of the rela-
tively large number of people now entering the
reproductive period of their lives.
The decline in fertility since the late 1950s
has not been ignored by those attempting to pre-
dict future U.S. population trends. Rapid changes
in fertility place a heavy burden on the economy
and on the society. The sharp increase in child-
bearing during the baby boom years put a severe
strain on schools, once these children reached their
sixth year; likewise, a substantial decrease in the
young age groups would require adjustments in
education, health and other services.
As a result, there has been much controversy
over the significance of the fertility decline of the
1960s, accompanied by some dubious manipulation
of population statistics on the part of a few partici-
pants. One highly publicized report by a private
research firm extended the current fertility decline
downward through the 1970s and announced that
the nation was in danger of "Instant ZPG"-vague-
ly defined as zero growth within a decade or a few
Warnings of this kind are contradicted by the
population projections released by the U.S. Cen-
sus Bureau. Their main technique in making pro-
jections-the current series includes four, labeled
Series B, C,D and E-is to assume that fertility
willfollow various paths, and to compute how pop-
ulation will change, given the present age structure,
certain factors of mortality and immigration levels.
Series B, the highest fertility assumption, is based
on a level of 3.1 children per woman; Series E, the
lowest, assumes that fertility will decline gradually
to reach replacement, 2.11, in the year 2000. An-
other Series, labelled X, is the same as Series E but
assumes that net immigration will be zero.
Under Series X assumptions, zero growth
would not be reached until the year 2037, at which
time U.S. population would be 276 million, 67
million greater than at present.
Below Replacement Fertility
What if fertility dropped below replacement?
Such fertility behavior has never happened before,
but there is nothing to prevent such an event, and
there are a number of forces, including better con-
traceptives, easier access to abortion, economic
factors, and environmental concerns, that could
push it down.
In response to such speculation the Census
Bureau has issued a new series of "illustrative" pro-
jections with lower fertility assumptions than Series
E. There are 12 projections in the new series, six
assuming no net immigration, and six assuming an
annual net immigration of 400,000 (the same as
for Series B through E).
The Census Bureau has not issued these new
projections as a substitute for the Series B-through-
E curves. Rather, they are illustrations of what
could be expected under what seem to be rather
unrealistic assumptions (although some of them
PROJECTED POPULATION GROWTH, VARYING FERTILITY ASSUMPTIONS, TO 2030
Population in millions
Source: U.S. Bureau of the Census
Even if fertility drops well below replacement, as in the Census Bureau's Series V projection,
population growth will continue well into the 21st century.
would appear to be more realistic than the Series B
projection, with its assumption of a fertility of 3.1
Two of the new projections, Series T and Ser-
ies W, assume that fertility will drop to replace-
ment by 1980 and 1970 respectively. Series V
assumes that fertility will continue its downward
course to reach a level of 1.5 by 1980, and then in-
crease to reach replacement by the year 2000.
These projections show that even if fertility
continues to decrease considerably below replace-
ment, as in Series V, population growth would con-
tinue well into the 21st century, although it would
have slowed substantially by the year 2000. Assum-
ing no net immigration, population under Series
V assumptions would be 240 million in 2000, and
254 million in 2030. With annual net immigration
of 400,000, population would be 255 million by
the year 2000 and 287 million by 2030, according
to the Series V projection.
The future course of population growth is
vitally important to the economic and social well-
being of America. Providing public services such as
schools and health care; predicting future markets
for everything from toys to automobiles and hous-
ing; planning for urban growth-all these activities
are affected by how many people there are, and
how many there will be.
But projecting future population, as this
Profile shows, is not an easy task. The spread be-
tween the Census Bureau's Series B projection and
its "illustrative" Series V curve is almost 100 mill-
ion people by the year 2000--less than 30 years in
the future. The spread becomes almost 200 million
--closeto the present total U.S. population ---by the
year 2020. Both of these extreme projections are
unlikely, but neither is unreasonable. And their
reasonableness makes .the planning task of those
responsible for preparing for the future much
The projections prepared by the Census Bur-
eau, and the analysis presented here, show one fact:
that the threat of "Instant ZPG" is a remote one.
Population growth is extremely likely to continue
for many years, even if fertility rates remain at
present levels or drop even further.
At the same time, there is strong evidence that
population growth is likely to come to an end in
the future, perhaps within the lifetime of many of
those being born today. Although there will be
many adjustments to be made in the economy and
the society as a result of this trend, there will be
time to adjust to them. And there are also many
benefits to be derived from a reduction in popula-
As the Commission on Population Growth and
the American Future put it in its final report of
"The nation should welcome and plan for a
stabilized population." 0
READINGS IN POPULATION
Brown, Harrison and Edward Hutchings, Jr. Are our Descendants Doomed? New York: The Viking Press,
1972. 377 pp. $12.50.
Twelve papers presented at a California Institute of Technology Conference on the impact of tech-
nological change on the growth and concentration of human populations, discussing religious and
cultural aspects of population control, developments in contraceptive research, UN's role in pop-
ulation/family planning programs, historical changes in the balance of births and deaths, and the
relationship of economic development to population growth and environmental change.
Brubaker, Sterling. To Live on Earth: Man and His Environment in Perspective. Baltimore: The Johns
Hopkins Press, 1972. 202 pp. $6.95.
A text on the impact of economic, technological and demographic growth on the environment, dis-
cussing areas of concern, future prospects, options and possible solutions.
Brunn, Stanley D. Urbanization in Developing Countries: An International Bibliography. East Lansing:
Latin American Studies Center and The Center for Urban Affairs, Michigan State University, 1971,
693 pp. $8.00.
Over 7,000 multilingual, bibliographic entries on political, social, economic, health, planning and
community development and housing aspects of urbanization in Africa, Asia and Latin America.
The material is listed by region, subregion and country, and includes a subject index.
Commission on Population Growth and the American Future. Population and the American Future.
New York: New American Library, 1972. 362 pp. $1.50.
A paperback advance release of the report of the historic Commission's two-year study on the econom-
ic social and environmental implications of US population growth. Included are all dissenting state-
ments by individual Commission members, references and lists of the more than 100 commissioned research
papers and witnesses in national public hearings.
Lee, Luke T. and Arthur Larson, Eds. Population and Law. Durham, North Carolina: Rule of Law Press,
1971. 452 pp. $19.38. Twelve structured surveys and evaluation of legislation affecting population,
dealing with birth control, family planning education and services, marriage, divorce and economic
factors related to family, in Asia, the Middle East and Europe; also describes the activities of the
United Nations and other international agencies in the field of population.
Petersen, William, Ed. Readings in Population. New York: The Macmillan Company, 1972. 483 pp.$5.95.
Forty selections by experts in their respective fields on various demographic topics, including popula-
tion growth, theories, age and sex, classification of residence, urbanization, migration, health and mor-
tality, fertility and population policy.
Shryock, Henry S. and Jacob S. Siegel. The Methods and Materials of Demography. Washington, D.C.: U.S.
Government Printing Office, 1972. 959 pp. $7.00.
A two-volume work designed as a text for courses on demographic methods, and as a research tool for
the professional, giving information on population data gathering techniques, classification, tabulation
and summarizing measures used to reveal population dynamics and composition, with special emphasis
on statistics available for underdeveloped nations and applicable methodology for these areas. 1
TO RECEIVE PRB PUBLICATIONS REGULARLY, PLEASE FILL OUT AND MAIL THIS ..
MEMBERSHIP/SUBSCRIPTION ORDER FORM
* Population Bulletin-Comprehensive publications focusing on important aspects and consequences of population
* PRB Selection-Occasional short essays reprinted from existing articles or abridged from speeches.
* Population Profile-Concise, illustrated analyses of significant population trends.
* World Population Data Sheet-An annual 11" x 17/2" chart providing detailed population information-for 145
The Bulletin is issued six times a year. PRB Selections, Profiles, or Data Sheet are issued between Bulletins.
O Annual Contributor ($15 or more) E Annual Member ($8) O Special Membership for Teachers ($5) ] Special Membership for Students ($5)
O Special Subscription for Libraries ($5) O Check for $ ....--------- is enclosed. O Bill me
Name of School (for Teachers, Students or Libraries)
Street and No.
City State Zip
(Send to: Population Reference Bureau, 1755 Massachusetts Ave., N.W., Washington, D. C. 20036.)