THE RECENT SHARP DECLINE IN FERTILITY IN BRAZIL:
ECONOMIC BOOM, SOCIAL INEQUALITY AND BABY BUST*
Jose Alberto M. de Carvalho,
Paulo de T. Almeida Paiva
and Donald R. Sawyer**
Working Paper N! 8
THE RECENT SHARP DECLINE IN FERTILITY IN BRAZIL:
ECONOMIC BOOM, SOCIAL INEQUALITY AND BABY BUST*
Jose Alberto M. de Carvalho,
Paulo de T. Almeida Paiva
and Donald R. Sawyer**
Working Paper N! 8
* This is a preliminary version of a chapter in W. Parker Mauldin (editor),
Fertility Decline in Developing Countries: Case Studies, which will be pub-
ished in 1982.
** Members of the Economic Demography program at the Center for Regional
Planning and Development (CEDEPLAR) of the Federal University of Minas Gerais.
Although the paper was written jointly, Jose Alberto M. de Carvalho was pri-
marily responsible for the demographic analysis, Paulo de T. A. Paiva for the
economic analysis and Donald R. Sawyer for preparation of the final version.
Lea Melo da Silva participated in the research but was unable to continue for
reasons of health. The principal sources are papers by Carvalho, 1980, Paiva,
1980, and Berqu6, 1980, cited in the bibliography.
In spite of a sharp decline in mortality levels since 1930, as
well as relatively rapid economic growth, mainly due to intense indus-
trialization, the fertility level in Brazil remained more or less con-
stant until the beginning of the 1960's. This seems, at first sight,
to indicate that economic growth is n6t sufficient to bring about a
decline in fertility. On the other hand, the 1970 census and PNAD
(National Household Sample Survey) data for the following years sug-
gest a rapid fertility decline after 1965, a new period of rapid eco-
What might have occurred to account for this change in the repro-
ductive behavior of the population? This paper attempts to describe
the Brazilian fertility decline and suggest possible explanations, on
the basis of data which are presently available. Regional estimates
of fertility levels since 1930 and an analysis of the extent of the
recent decline are presented in Section I. Section II deals with
intermediate variables, population policies and some possible compo-
sition effects. Section III includes a discussion of the possible
impact of increased economic inequality on fertility.
Since 1980 census data are not yet available, the conclusions
drawn in this paper should be considered only as working hypotheses
subject to further investigation.
I. Demographic Trends in Brazil
The population of Brazil has grown at rates above 22 per year since
the final decade of the nineteenth century, reaching a total of 93 mil-
lion inhabitants in 1970 and approximately 119 million in 1980. Natural
increase has been responsible for most of the growth. International mi-
gration from Europe reached a peak during the second half of nineteenth
century, providing a significant contribution to growth in the last three
decades of the last and first four decades of this century. Since 1940
the role of immigration in Brazil's population growth has been minimal,
the key variables having been fertility and mortality.1
Although the principal purpose of this paper is to describe and
discuss an apparent drop in fertility in the second half of the 1960's
and first half of the 1970's, it would be useful to place recent changes
in the context of fertility and mortality trends since 1930. In a coun-
try with the continental dimensions of Brazil, with considerable divers-
ity, a clearer understanding is obtained by breaking down national aver-
ages into regional levels.
Since Brazil is a country in which the vital statistics system is
still not reasonably complete, except in a few limited areas, researchers
are obliged to rely upon census data for the estimation of national or
regional fertility and mortality levels, using indirect techniques. This
is not the place to explain such techniques or discuss possible distor-
tions. Other studies show that, in the case of Brazil, they give reason-
ably good results, especially when, as in this article, the purpose is
to establish comparisons over time, rather than identifying absolute
Fertility estimates were obtained by applying the Brass fertility
technique, and mortality estimates by use of the Brass childhood mortal-
ity technique. Life expectancies at birth were obtained through logit
transformations, using the childhood mortality estimates of Brazil and
the life tables for Mexico as standards.3
A. Regional Trends in Fertility and Mortality in Brazil, 1930--70
Estimates of fertility (total fertility rate) and natality (crude
birth rate) for 10 regions in the decades following 1930, 1940 and 19604
are shown in Table 1.
BRAZIL AND REGIONS TOTAL FERTILITY RATES (PER WOMAN) AND
CRUDE BIRTH RATES (PER THOUSAND) 1930-1940 TO 1960-1970
Total fertility rates Crude birth rates
Region 1960-70 1960-70
1930-40 1940-50 1960-70 190-0 1930-40 1940-50 1960-70 1940-50
Amaz8nia 6.9 7.3 8.1 1.10 50.2 50.6 49.3 0.97
Northern Northeast 7.0 7.0 7.3 1.04 51.1 49.4 45.3 0.91
Central Northeast 7.9 7.7 7.8 1.01 55.5 55.1 51.1 0.92
Southern Northeast 6.9 7.3 7.6 1.04 50.3 52.9 49.4 0.93
Minas 7.2 6.8 6.5 0.95 50.4 48.6 42.4 0.87
Rio de Janeiro 4.2 4.0 4.0 1.00 33.7 33.0 29.3 0.88
Sao Paulo 5.6 5.1 4.2 0.82 41.8 39.3 31.9 0.81
Parana 5.9 5.9 6.5 1.10 42.1 42.9 43.2 1.01
Extreme South 6.2 6.2 5.1 0.82 43.0 43.6 32.6 .0.75
Vest 6.2 6.4 6.6 1.03 44.3 45.8 44.2 0.97
BRAZIL 6.5 6.3 5.8 0.92 47.2 46.5 41.1 0.88
Source: 1930-40 and 1940-50 from CARVALHO, 1974, p. 408; 1960-1970 from CARVALHO, 1978, passim.
States and territories of the ten regions are: Amaz8nia: Acre, Amazonas, Amapi, Para, Rond8nia and Roraima;
northern Northeast: Maranhao, Piaui; Central Northeast: Ceara, Rio Grande do Norte, Paraiba, Pernambuco,
Fernando de Noronha, Alagoas; Southern Northeast: Sergipe, Bahia; Minas: Minas Gerais, Espirito Santo;
Rio: Rio de Janeiro; Sao Paulo: Sao Paulo; Parana: Parana; Extreme South: Santa Catarina, Rio Grande do
Sul; West: Goias, Mato Grosso do Sul, Mato Grosso, Federal District. (See map.)
From the 1940's to the 1960's, the estimated fertility for the
country as a whole decreased 8%. This modest decrease over the 20 year
period is due to downward trends in Minas, Slo Paulo and the South of
Brazil, all of which are developed areas. The drop in the estimates for
Sio Paulo and the South reached 18%. Fertility rates in Rio, a develop-
ed region which included the national capital for most of that period,
did not drop, but maintained the low level, by Brazilian standards, al-
ready reached in the 1940's. In the six other regions, which are among
the least developed (the three Northeastern regions) or in frontier
areas (Amazonia, West Region and Parani), the estimated level of fertil-
ity remained constant or increased, reaching 10% in the cases of Amaz6nia
In the same period, the estimated crude birth rate decreased in
all regions except Parana, where there was an increase of only 1%. In
the regions where the total fertility rate decreased, the percentage
drop in the crude birth rate was even greater. Since during the period
in question there was no significant change in the fertility distribu-
tion,5 the age structure in all regions changed in such a way as to de-
press the birth rate, assuming constant fertility. Thus, while the
estimated fertility for Brazil as a whole decreased by only 8%, the
estimated natality fell by 12%.
Estimates of life expectancy at birth (e ) for the decades follow-
ing 1930, 1940 and 1960 are shown in Table 2. For the country as a whole,
8 increased from 41.2 to 53.7 years, an average annual gain of 0.42
years. Thus, there was a significant decline in mortality, although the
level remained relatively high, by international standards, especially
in the Northeastern region (see Map, page 6 ). Regional differences
are illustrated by the fact that for the 1960-70 decade, while infant
mortality in Brazil was estimated at 91 per thousand live births, it
reached 189 in Rio Grande do Norte, 162 in Paraiba and Pernambuco, 153
in Alagoas and 150 in Ceara, states of the Central Northeast region.6
By comparison, the infant mortality rate in Argentina in 1965 was 56.9,7
comparable only to Rio Grande do Sul, the state which at that time had
the lowest mortality levels in Brazil.
Despite rapid economic growth in Brazil during the post-war period
and explicit policies aimed at development of the Northeast, life expectancy
BRAZIL AND REGIONS
EXPECTANCY AT BIRTH
REGIONS 1930-40 1940-50 1960-70 1
Amaz8nia 39.8 42.7 53.8 1.26
Northern Northeast 40.0 43.7 50.6 1.16
Central Northeast 34.7 34.0 44.2 1.30
Southern Northeast 38.3 39.2 49.9 1.27
Minas 43.0 46.1 55.6 1.21
Rio 44.5 48.7 57.7 1.18
SAo Paulo 42.7 49.4 58.8 1.19
Parana 43.9 45.9 57.0 1.24
Extreme South 51.0 55.3 62.6 1.13
West 46.9 49.8 57.9 1.16
BRAZIL 41.2 43.6 53.7 1.23
p. 419; 1960-70
SOURCE: 1930-40 and 1940-50 from CARVALHO (1974),
from CARVALHO (1978), passim.
BRAZIL AND ITS TEN REGIONS AS OF 1970
S '- mapoa
-- l NORTHERN NORTHEAST
... 1 --" l --^ ---- ---_CENTRAL
PAmazonas oro] P rd- NORTHEAST
-- Ceoro Rio Grnde
tw- -- ---- do Norte
S-. -- _Piu Pauraiba
"Uclu -7" Pr perr. ambuco
ondni -- EAlogoos
IMato Grosso------ -_" _
to Grosso. .- D.F NORTHEAST
I- E-- spirito Santo
So Paul RIO
Rio de Janeiro
PARANA Parano SAO PAULO
PEnsoNS PER K 2(1,701 ,o Grande EXTREME
I--- I pe /Kmt do Sul SOUTH
S1- O10 per / Km
3*l"- 20 per / Kmg
201-- 50per / Kmn
1001-- 200 per Km2
The ten major regions ore in bold type and
the states and territories in copitolizod words
Source! Afir..r.1o ..9g do GIash,1 11957)
of the Central Northeast region in the 1960's was approximately equal to
that of the rest of Brazil in the 1930's. As a rule, the greatest rela-
tive gains in life expectancy between the 1940's and the 1960's occurred
in the regions which had the highest levels of mortality.
With regard to fertility and mortality differentials by urban or
rural residence, the 1970 census revealed the following patterns:
1. Rural fertility was higher than urban fertility in all states.
For the country as a whole the total fertility rate for the urban popula-
tion was 4.7, and for the rural population, 7.6.8
2. There was no such clear pattern for mortality levels at the
state or regional level. Using 1950 census data, Mortara had already
found higher urban mortality in several states. When the 1970 data
are broken down by family income levels, a clear pattern of rural-urban
differentials emerges. Among the lowest income groups, urban mortality
was greater than rural mortality in all of the ten regions, while rural
mortality was greater in the middle and high-income groups.1
B. The Recent Drop in Fertility
The results of the 1970 census show, as mentioned previously, that
some regions had already undergone significant fertility decline by that
time. Analysis at the state level, with application of Brass techniques
to the urban population in 1970, produced clear evidence, in almost all
of the 25 states and territories, of rapid fertility decline in urban
areas. While the P./F. series, i.e. the values of the quotient of mean
parity and accumulated current fertility, generally decreased rapidly
for the rural areas, as one would expect, they were stable or increased,
with very few exceptions, for urban areas.1 This was a good indication
of sizable fertility decline among the urban population. In Slo Paulo
state and South regions, the P./F. series also indicated that rural fertil-
During the 1970's, the Brazilian census bureau (Funda9go IBGE) car-
ried out a series of surveys, called PNADs (National Household Sample
Surveys), which include questions that permit the estimation of fertil-
ity. The sampling plans based on the 1970 census, were drawn up so that
results would be representativeat the level of seven regions. In the
first PNADs, the Amazon and West regions (except the Federal District)were
not included; later, the urban areas of these regions were surveyed. For
our purposes, only 1972 and 1976 PNAD data are used, together with the 1970
census. Fertility estimates based on the PNAD surveys suggest a sharp
decline in fertility during the decade of the 1970s.
SDoubts may arise with regard to representativeness of the 1976 PNAD
results, since the census tracts included in the survey were selected bas-
ed on the 1970 census and were not changed. Given the marked mobility of
the Brazilian population during the decades, with large-scale rural-urban
migration and rapid growth of the larger cities, it is possible that the
migrant population is under-represented. In any case, the data from 1972
and 1976 should be representative, at the very least, of the tracts which
were surveyed. Since the tendencies observed are very clear, it seems
reasonable to assume that something quite similar occurred on the level
of the regions and the nation as a whole.
Estimated total fertility rates (TFR) for 1970, 1972 and 1976 are
shown in Table 3. Since the estimates are controlled by the level of ac-
cumulated fertility (parity) of women aged 20-24, they refer, grosso modo,
to the average for 1965-70, 1967-72 and 1971-76. It should be remembered
that the estimates for Brazil as a whole do not include Amazonia and the
West regions (except the Federal District) for 1967-72, and that they do
not include the rural population of these regions for 1965-70 and 1971-76.
These differences in coverage should not cause serious distortions. In
1970, the total population of Amazonia and the West regions, not including
the Federal District, was only 8.7% of Brazil's total population. For exam-
ple, the inclusion of the rural population of those regions would have little
effect on the TFR estimate for 1965-70, increasing it from 5.5 to 5.6.
When the estimates for 1965-70 are compared with those for 1967-72
(see Table 3), the fertility decline is small, except for Minas and the
Federal District, where the data indicate drops of 8% and 15%, respectively.
In these two cases, there may have been sampling problems for so rapid a
change to appear. However, when the 1971-76 estimates are compared with
those of 1967-72 (last column), there are large differences between
estimates based on the two PNADs, indicating considerable decline for
all regions, including Minas and the Federal District. These data are
a more reliable indicator of fertility decline between the latter
BRAZIL AND REGIONS ESTIMATED TOTAL FERTILITY RATES AS OF THE
1970 CENSUS AND 1972 AND 1976 PNAD SURVEYS
PNAD 1971-76 1971-76
REGIONS* 1965-70 1967-72 1971-76 1971-76 1971-76
REGIONS* 1965-70 1967-72
Northeast 7.5 7.4 6.0 0.80 0.81
Minas 6.0 5.5 4.2 0.70 0.76
Rio 3.7 3.6 2.7 0.73 0.75
Sio Paulo 3.8 3.7 2.9 0.76 0.78
South 5.2 5.0 3.9 0.75 0.78
Federal District 5.4 4.6 3.7 0.69 0.80
BRAZIL 5.5a 5.3b 4.2a 0.76 0.79
ACADEMY OF SCIENCES (1979). In CARVALHO, J.A.M. (1980),
* Minas, Rio and Slo Paulo same as
excluded, except as noted below.
Northeast Northern Northeast,
South Extreme South, Paran
Federal District part of the We
Tables 1 and 2. Amaz6nia and West
Other regions regrouped as follows:
Central Northeast, Southern Northeast.
a Excluding rural population of Amaz8nia and West, but including Federal
b Excluding Amaz6nia and West, but including Federal District.
half of the 1960's and the first half of the 1970's, since the estimates
are based on the same census tracts. Since the variation in the esti-
mates occurs in the same direction and in similar proportions in all
regions, it is probably not due to sampling errors within the census
tracts used for the series of PNADs or to the fact that, because of mi-
gration, these tracts may not have been representative of the regional
populations in 1976.
All regions experienced a very sharp drop in fertility over the
four-year period, ranging from 25% in the region of Rio de Janeiro to
19% in the Northeast. Comparing the results of Table 3 with those of
Table 1, it can be seen that:
1. Although the three regions making up the Northeast, which had
high levels of fertility, did not show signs of a fertility decline be-
fore 1960-70, during the first five years of the 1970's they followed
the general trend of declining fertility, at a slightly lower rate.
2. The region of Rio de Janeiro, which had had the lowest and
most stable fertility level since the 1940's, experienced the greatest
relative decline in fertility in the country between the two five-year
3. The region of Minas, which had a fertility level above the
national average until the 1960's, began to experience a slight decline
in the 1960's and then a sharp drop between the two periods under scrutiny.
4. The South and Slo Paulo regions, which had experienced the
greatest decline in the country between the 1940's and the 1960's, also
experienced a sharp drop in fertility between the end of the 1960's and
the beginning of the 1970's.
5. The country as a whole, which between the 1940's and 1960's had
experienced a decline of only 8% in its fertility level, experienced a
sharp drop of around 20% between 1967-72, and 1971-76 that is, a span
of only about four years.
Barring large sampling errors within the census tracts of the PNADs
or serious problems in the representativeness of the tracts in relation
to the regional populations in 1976, it seems safe to conclude that be-
tween the end of the 1960s and the start of the 1970s, Brazil began to
experience a new demographic tendency, characterized by increasingly lower
Aside from data provided by the PNADs, during the 70's there was
no other source deemed to be representative of the country as a whole,
or of one or more of its regions, that would permit the calculation of
fertility estimates. However, the National Study of Human Reproduction
of the Brazilian Center for Analysis and Planning.(CEBRAP) surveyed seven
counties of Brazil, covering only rural areas in two, urban areas in three
and both types in the other two. The counties were chosen following the
development of a rural and urban typology for the country, based princi-
pally on prevailing forms of organization of production. Sio Jose dos
Campos is a heavily industrialized city, near the Metropolitan Area of
Sio Paulo; Sertiozinho is a county in the countryside of the State of
Sio Paulo, with an economy based on sugar-cane plantations; Conceigao
di Araguaia is in an important agricultural frontier area in the Amazon
region; Santa Cruz do Sul is located in the South and has an economy
centered in tobacco production on medium and small-size farms; Parnaiba
is located in the poor State of Piaui, with its agricultural production
based mostly on pre-capitalist forms; Recife is the most important city
of the Northeast, with a population in 1970 of around 1,000,000 and an
economy based mainly on the tertiary sector activities, particularly
public services and commerce, with an important share of its population
working in the "informal sector"; and Cachoeiro do Itapemirim is a
medium-size city in the state of Espirito Santo with a dynamic economy
based on co merce, services and industry.12
The fertility estimates in the areas surveyed by CEBRAP's National
Study of Human Reproduction are shown in Table 4, which also includes
additional information that will be discussed later on. Since detailed
life history data were collected, it is possible to estimate fertility
for 1965, 1970 and 1975.
It is important to call attention to the fact that any comparison
between the estimates of Table 4 and those of Table 3 requires consider-
able caution, since the areas represented and the techniques of estima-
tion are different in the National Study of Human Reproduction the
estimate is direct, based on life history questionnaires. For each
county in Table 4, comparison between the estimates for the three years
BRAZIL DISTRIBUTION OF AGE-SPECIFIC FERTILITY RATES (ASFR) ACCORDING TO
1972 and 1976 PNADs
Age of Women
SOURCE: NATIONAL ACADEMY OF SCIENCES, 1979.
BRAZIL ESTIMATED TOTAL FERTILITY RATES,* PREVALENCE OF CONTRACEPTIVE USE AND MEAN DURATION OF
BREASTFEEDING IN COUNTRIES INCLUDED IN THE NATIONAL STUDY OF HUMAN REPRODUCTION
1975 Prevalence of Mean Duration of
COUNTRIES 1965 1970 1975 1965 Contraceptive Breastfeeding
Use (%)** (Months)**
Conceicio do Araguaia
Sio Jose dos Campos
Santa Cruz do Sul
SOURCE: BERQUO, Elza S. (1980)
Numbers in parentheses refer to the number of women.
* Based on all women in the sample.
** Based on a subsample of women who, had
one livebirth during the 1970-75 period.
shown should indicate fertility change, although subject to memory lapses
and sampling errors, principally in view of the relatively small number
of women included in the samples.
With the exception of Sertiozinho and Parnaiba, both of which are
rural, in all the other areas surveyed, the estimates indicate a signi-
ficant drop in fertility, including the rural areas of Santa Cruz do Sul
and Conceigio do Araguaia; in fact, in the former there was a sharp de-
cline of around 28%. Thus, the estimates based on the National Study of
Human Reproduction are consistent with the substantial decline in fertil-
ity in Brazil shown by the analyses based on the PNADs.
One must await results of the 1980 demographic census for the final
word on fertility changes during the 1970's. There is some skepticism
in Brazil regarding the magnitude of the fertility decline, which was
not expected to be so large. In any case, the PNADs estimates as well
as the CEBRAP study are consistent internally and with each other and
seem to reveal a picture of the overall fertility trend in the country.
Figure 1 shows the fertility distribution curves according to the
PNADs of 1972 and 1976. The specific fertility rates declined by 6% in
the group of women between 15 and 19 years of age; 18% in the group of
20 to 24 years; 24% in the groups of 25 to 29 years and 30 to 34 years;
26% in the groups of 35 to 39 and 40 to 44 years, and 33% in the group
of 45 to 49 years. It can be seen that the specific rates fell for all
age groups, dropping relatively more among the older groups, as would
be expected based on the experience of those countries which already
have undergone their demographic transitions.
II. The Spread of Fertility Control
In order to understand the recent drop in fertility it is import-
ant to have some notion of the mechanisms by which it occurred (nuptial-
ity or control of marital fertility); the possible role of family plan-
ning programs and of the distribution of the decline by regions, includ-
ing urban-rural residence and income level. While data on these ques-
tions are fragmentary, some inferences can be made, as a preliminary
guide to further discussion and investigation.
A. Some Evidence on Intermediate Variables
There is no survey similar to the World Fertility Survey in Brazil,
which would be representative on a national basis, and would provide infor-
mation on intermediate fertility variables. Several local surveys and
studies are reviewed here in order to provide some idea of how selected
intermediate variables may have influenced the sharp fertility decline
in Brazil, especially the role of nuptiality, contraceptive use, breast-
feeding and abortion.
Table 5 shows the average ages at which women in each region enter
the first conjugal union, be it civil, religious or consensual, according
to the censuses of 1960 and 1970 and the PNAD of 1976. Between 1960 and
1976, there was an increase of around one year for all regions. This in-
crease was not dramatic and, although it certainly may have had an impact
on fertility, it obviously was not the principal intervening cause of the
recent sharp drop in fertility in the country. By the very pattern of
the fertility decline, according to Figure 1, it can be concluded that,
in general, the population controlled fertility within marriage.
Since it covers such differentiated areas of the country, CEBRAP's
National Study of Human Reproduction provides valuable data on the pre-
valence of contraceptive use. Unfortunately, there is only information
for the 1970-75 period, making it impossible to analyze trends over time.
Table 4 shows the percentage of women that used contraceptives before
giving birth to the last child as well as the average duration of breast-
feeding. The data in the table refer to only those women in the sample
who had had at least one liveborn child between 1970 and 1975. Prevalence
in the use of contraceptives is low in the rural areas of Parnaiba and
Conceiclo do Araguaia, which, according to Table 4, had a decline of 9%
in fertility (Conceicgo) or even an increase (Parnaiba). The rural areas
of Sertiozinho and Santa Cruz do Sul exhibited high prevalence rates,
approximating the rates of the urban areas, which leads one to suspect
that the apparent increase in fertility in Sertiozinho between 1960 and
1970 is due to underestimation of total fertility in 1965. There is a
large variation regarding the average duration of breastfeeding, with a
much shorter duration in the urban areas of the poor Northeast (Parnaiba
and Recife) and a relatively longer duration in Slo Jose dos Campos, an
BRAZIL AND REGIONS MEAN AGE AT FIRST
1960, 1970 AND 1976
PNAD REGIONS 1960 1970 1976
Northeast 22.2 22.9 23.1
Minas 21.9 23.4 23.8
Rio 23.0 23.8 23.9
Slo Paulo 22.2 23.3 23.6
South 21.7 22.2 22.7
Federal District 20.6 23.3 23.7
Amazonia and West** 21.3 22.1 22.9***
BRAZIL 22.2 23.0 23.3***
SOURCE: ALTMANN, Ana M. G. WONG, Laura R. (1980), p. 21
* Including consensual unions. Estimated using Hajnal method.
** Excluding the Federal District.
*** Excluding the rural population of Amazonia and West regions.
In 1978 a contraceptive use survey was undertaken in the State of
Sgo Paulo entitled "Sao Paulo State contraceptive Prevalence Survey". 13
Results are representative for the capital (the municipio of Sao Paulo),
other urban areas of the State and the rural area. A previous fertility
study, carried out in 1965, included data on, among other things, the
prevalence of contraceptives in the "district" of Sao Paulo, thus making
possible comparisons for the state capital.4 (In 1970 more than 95.0%
of the population of the municipio of Slo Paulo lived in the "district").
Table 6 shows the distribution, in 1978, of the use of contracep-
tives among married women between 15 and 44 years of age in the three
population groups of the State of Sio Paulo. There is no appreciable
difference of the percentage of women using contraceptives between the
urban and rural areas. It is surprising to find high prevalence of
contraceptives among women in the rural areas, indicating generalized
use throughout the entire state. In all three areas the methods most
used were oral pills and sterilization, accounting for about 70% of those
women using contraceptives. Withdrawal, condom and rhythm were also
used but a lesser degree. The high prevalence of sterilization is sur-
prising since it requires the participation of medical and hospital
services and is formally prohibited by Brazilian legislation, except in
cases of medical indications based exclusively on health reasons.
The proportion of married women currently using contraception
rises from 63.9% to 73.1% if infertile and currently pregnant women
are excluded. The survey does not show large variation in contracep-
tive use by level of education, varying from 59.6% among married women
with incomplete elementary schooling, to 64.8% among those with complete
elementary schooling, to 67.9% among those with five or more years of
schooling. The variation is a little larger when controlled by income
group, being 47.4% among the poorest and 66.9% among the wealthiest.
In the same survey only 1.3% of those women not currently using contra-
ceptives declared religious reasons for not using them.15
Table 7 presents for the district of Slo Paulo in 1965 and for the
municipio of Sio Paulo in 1978 the percentages of married women using
the pill, condom or who were sterilized, showing an enormous increase
in the use of the pill, an increase in the proportion of sterilized
women and a drop of almost 50% of the proportion using the condom.
STATE OF SAO PAULO PERCENTAGE OF CURRENTLY MARRIED WOMEN*
AGED 15-44 CURRENTLY USING CONTRACEPTION- 1978
Municipio Other Urban Total for
Method of Areas Rural Areas State of
S So Paulo Areas Sio Paulo
Orals 30.0 26.7 27.0 27.8
Sterilization 13.9 18.0 14.1 16.1
Withdrawal 5.9 7.3 10.5 7.3
Condom 6.9 7.3 3.3 6.6
Rhythm 4.5 6.2 3.2 5.2
Other Methods 2.2 0.5 0.5 0.9
Currently Using 63.4 66.0 58.6 63.9
Not Currently Using 36.6 34.0 41.4 36.1
TOTAL 100.0 100.0 100.0 100.0
Number of Cases 546 600 734 1,880
SOURCE: NAKAMURA, Milton S. and FONSECA, Joaquim P. B. 1979 (Table 8).
* Including those instable consensual unions.
MUNICfPIO OF SAO PAULO PERCENTAGE OF MARRIED WOMEN* AGED
15 49 (1965) OR 15 44 (1978) USING ORAL PILLS,
CONDOM OR STERILIZATION
Method 1965 (District) 1978 (Municipio)
Orals 6.0 30.0
Condom 12.0 6.9
Sterilization 7.0 13.9
TOTAL 25.0 50.8
* Including consensual unions.
- BERQUd, Elza S. et al., eds. 1977
- NAKAMURA, Milton S. and FONSECA, Joaquim P. B. (1979).
in BERQUd, Elza S. (1980). Table 8.
Even though the data for 1965 and 1978 shown in Table 8 are not
strictly comparable, it seems clear that if in 1965 the incidence of
sterilization was much more prevalent among women having higher educa-
tional levels, in 1978, sterilization was more common among women with
a lower level of education. If one bears in mind that these data refer
to a stock variable, it can be concluded that between 1965 and 1978 the
sterilization rates among less educated (i.e., low-income) women, if
compared to those of the other groups, increased at a still higher rate
than might initially appear on the basis of the table.
Abortion is practically illegal in Brazil, being accepted only in
a few exceptional cases. Consequently, no national statistics are avail-
able which would permit evaluation of abortion trends. It is common
knowledge that abortion is frequently used as a means of preventing un-
desired births and it is widely believed, including among physicians,
that the practice of abortion is increasing. It seems that such an
increase is significant mainly among unmarried young females, partly
due to greater sexual freedom, which is not accompanied by adequate
familiarity with contraceptive methods. In spite of the lack of official
statistics, it can be seen that an increasing proportion of beds in matern-
ity wards for the poor, especially in the major urban centers, are occupied
by women under medical treatment as a consequence of abortion complica-
tions. These abortions are carried out under extremely unfavorable con-
Legalization of abortion has been increasingly debated in the mass
media, including the national television networks. Estimates ranging
from 1 to 3 million abortions per year have been used as an argument
favoring legalization.6 The sources of information are not known, but
the magnitude of the numbers suggests that abortion has attracted consider-
able attention in a country where the annual average number of births
was around 3 million in the 70's.
A question on abortion, without discrimination of whether sponta-
neous (miscarriage) or induced was included in the life history question-
naire of CEBRAP's National Study of Human Reproduction. The results,
shown in Table 9, indicate that a sharp increase in induced abortions
is likely to have occurred between 1969 and 1975, since the frequency
SAO PAULO PREVALENCE OF STERILIZATION AMONG MARRIED WOMEN
AGED 15-49 (1965) OR 15--44 (1978), BY EDUCATIONAL LEVEL
Educational level 1965 (District)* 1978 (State)**
(0--3 years) 4.1 18.6
(4 years) 6.7 14.3
(5 + years) 11.2 14.9
SOURCE: 1965 SZMRECSANYI, Maria I.Q.F. (1977), pp. 308 and 335.
1978 NAKAMURA, Milton S. and FONSECA, Joaquim P. B. (1979),
* Ever married women, including consensual unions.
** Currently married women, including consensual unions.
NUMBER OF CUMULATED INDUCED AND SPONTANEOUS ABORTIONS PER 100
WOMEN OVER 15 YEARS OF AGE IN 1960 AND 1975 IN SELECTED COUNTIES
Counties 1960 1975
ConceiqAo do Araguaia
- rural 3.0 6.4
- rural 19.7 31.8
- urban 25.3 33.9
- urban 21.6 31.8
Cachoeira do Itapemirim
- urban 6.3 9.4
Sio Jose dos Campos
- urban 25.3 38.2
- rural 14.4 20.0
Santa Cruz do Sul
- urban 2.4 6.0
- rural 1.9 2.8
SOURCE: BERQUO, Elza S. (1980). Table 5.
of miscarriages has probably fallen due to an overall fertility and mor-
tality decline in the country in the period. Nevertheless, these data
should be interpreted with caution. As the 1960 estimates refer to women
still alive at the time of the survey (1976-78), memory errors in the
1960 estimates are probably greater than those in the 1975 estimates,
since they refer to abortions that would have occurred more than 15 years
before the survey.7 Furthermore, the women above 15 years of age in
1960 had a lower average age than those above age 15 in 1975.
The Sao Paulo Fertility Study and the Contraceptive Prevalance
Survey allow indirect comparison on abortion in the municipio of Sao
Paulo in 1965 and 1970, as can be seen in Table 10. Unfortunately, as
with the 1965 information, data of the second study do not discriminate
between induced and spontaneous abortion. In the 1965 data, some of the
women that had had at least one spontaneous abortion had probably also
undergone induced abortion, and the percentages cannot be added. Never-
theless, according to the table, the percentage of women that had at
least one spontaneous abortion by 1965 (27.56%) is higher than that of
women who had either type of abortion by 1978 (24.1%). Even though there
was probably a decline in spontaneous abortions between 1965 and 1978, the
data from Table 10 do not allow any conclusion indicating that induced abortion
has risen in SEo Paulo. If any inference could be drawn, it would be in the
opposite direction, i.e., a decline in induced abortion.
As the quality of responses to questions on abortion is likely to
vary greatly from survey to survey, mainly due to the high sensitivity
of the subject in the country, it may be preferable to accept the well-
founded opinions of persons dealing more directly with the problem,
like doctors and social workers, who report an increasing prevalence
of induced abortion in Brazil.
The data on intermediate variables presented in this section can-
not be generalized directly to the whole country, because of the hetero-
geneity among Brazil's regions. Nevertheless, given the small difference
in the State of Sao Paulo in contraceptive prevalence between rural and
urban areas, between women according to the level of schooling and in-
come, and also in view of the results of the National Study of Human
Reproduction in seven of the nine areas analyzed, it appears that the
widespread use of contraceptives is the principal intermediate variable
MUNICIPIO OF SaO PAULO PERCENTAGE OF EVER MARRIED* WOMEN AGED
15--49 (1965) OR 15-44 (1978) WITH AT LEAST ONE ABORTION
Abortion 1965 (District) 1978 (Municipio)
Induced 11.54 -
Spontaneous or Induced -24.3
SOURCE: 1965 LEVY, Maria S. F. and MOLEDO, Rosalba (1977), p. 365.
1978 NAKAMURA, Milton S. and FONSECA, Joaquim P. B. (1979).
* Including consensual unions.
to explain the recent sharp fertility decline in the country, although it
is not possible to evaluate accurately the impact of abortion.
B. Policies Directly Affecting Fertility
Brazil has never had an official family planning program. Until
1974, when the Bucharest Conference was held, the government's position,
at least in its official statements, was pronatalist, based on arguments
which included availability of land and other natural resources, national
power and strength, and geopolitics. This does not mean, however, that
various means of birth control could not be used by a growing proportion
of the population.
In spite of formal restriction on distribution of contraceptives,
their availability in practice can be attributed in part to the fact that,
at first, modern methods were used primarily by privileged groups, and in
part to the fact that the official position was undergoing a process of
reformulation. A new attitude was revealed at Bucharest, where it was
recognized that the government should promote access to information and
to contraceptives among all segments of the population. This policy began
to be implemented in the Maternal and Infant Health Program in 1978 and
forms part of the Prev-Saide Program, announced in 1980. 18
Although the Catholic Church is officially opposed to artificial
birth control, its position has become less rigid, especially at the level
of individual counseling by the clergy. The Church's position was pro-
bably more important at the level of policy-making than among the popula-
tion at large.
A private organization called the Brazilian Association for the
Welfare of the Family (BEMFAM), operating with the support of the Inter-
national Planned Parenthood Federation and other international sources,
has actively promoted family planning in Brazil. In addition to lobby-
ing and publicity efforts, it has signed contracts with state governments
to set up family planning programs through a system of community based
distribution. The states where it operates programs are Pernambuco,
Rio Grande do Norte, Paraiba and Alagoas, in the Northeast, as well as
Parana, in the South. Although BEMFAM's programs may have contributed
to the recent drop in fertility, their impact would have been felt in
well-localized areas, particularly in the Northeast, while the drop
occurred in all six of the regions covered by the PNAD.
The relative insignificance of family planning programs in Brazil
is reflected in the score that Mauldin and Berelson gave for the country's
"family planning effort" in 1972.19 On a scale of 0 to 32, Brazil receiv-
ed a score of 0. This may be too low, but it is not far from the truth.
The occurrence of fertility reduction practically without family
planning programs makes Brazil an important case for the discussion of
the effects of institutional efforts on fertility. Contrary to the
analysis by Tsui and Bogue, who used inadequate fertility data, Brazil's
fertility did decrease, without an institutional effort, and should not
be used to contrast with countries in which fertility declined because
of a strong institutional effort without a strong economic performance.
C. Possible Effects of Redistribution of the Population by Place of
Residence or Income Levels
It was seen above that the percentage decrease in the total fertil-
ity rate (TFR) between the last half of the 1960's and first half of the
1970's was on the order of 20%, a sharp drop compared with the decline
of only 8% between 1940-50 and 1960-70. It is not clear whether the
drop took place through changes in composition of the population or de-
cline in specific rates for different segments. Because of economic
and social inequality within and among Brazil's diverse regions and marked
fertility differentials among different sub-groups of the population,
it could be thought that high rates of social and spatial mobility in
the period under scrutiny might account for a sizable part of the de-
cline. Specifically, the migrants to the more developed regions or from
rural to urban areas, as well as new members of the growing "middle
class" (discussed below), might have assimilated the lower fertility
levels characteristic of the areas or strata which they entered. While
such changes would have involved modification of behavior, it is inter-
esting for present purposes to separate them from changes in specific
rates which may have occurred within the different sub-groups and con-
sider them compositional effects. The most relevant effects are those
concerned with inter-regional migration, rural-urban migration and
mobility from lower to higher income strata.
Inter-regional migration from the less developed to the more devel-
oped regions, even though it may actually raise the average fertility
level of the areas of destination in the short run, tends with time to
favor lower fertility among the migrants. If migrant women are more
fertile than those who stay in the regions of origin, average fertility
in those regions may also decline, although to the extent that population
pressures are diminished, out-migration may also contribute to the main-
tenance of high fertility.
Redistribution among Brazil's regions between 1960 and 1970 was
not profound. The developed regions included 66.1% of the total popula-
tion in 1970 as compared to 64.2% in 1960, a gain of only 1.9%. The
relative uniformity of fertility decline among the six PNAD regions for
which there is information, including large drops in the less developed
areas, provides another indication that migration to the more developed
areas was not responsible, through a composition effect, for the over-
all decline in the period studied. However, the various indirect effects
of interregional migration through selectivity, return migration, different
rates of assimilation, etc., deserve further study. The possible contri-
butions of migration to Amazonia and the West to a decline in the areas
of origin and to the maintenance of fertility levels for the country as
a whole should also be taken into account once more complete data are
The PNAD data are not suited for separate estimates of rural and
urban fertility, but the 1970 census shows that rural fertility (TFR- 4.6)
was about 60% higher, on the average, than urban fertility (TFR 4.7).
If migrants adopted urban fertility levels soon after arrival, the
rural-urban migration which occurred in the period in question could
have contributed to the drop in fertility. However, the volume of mi-
gration was not sufficient to have much impact. Even when the 1970
urban and rural TFRs are held constant, the observed increase in the
urban population from 58.3% of all families in 1970 to 67.0% of all
families in 1976 would have decreased Brazil's TFR by only 4%.
Redistribution of the population by income levels is more difficult
to interpret. Estimated TFRs for the rural and urban populations by
level of family income, based on the 1970 census, are shown in Table 11.
BRAZIL ESTIMATED TOTAL FERTILITY RATES BY MONTHLY HOUSEHOLD INCOME
AND BY URBAN AND RURAL RESIDENCE 1970
Res e Less than Cr551 to Cr101 to Cr5151 to Cr5301 to Over
Residence C c C c rTotal
Cr 50 Cr$ 100 Cr 150 CrS300 cr$ 500 Cr04 500
Rural 7.73 7.89 7.65 7.99 7.70 5.86 7.73
Urban 6.61 6.96 7.10 5.95 4.84 3.10 4.83.
Total 7.39 7.62 7.47 6.72 5.37 3.28 5.95
Percentage of Families 6.2 17.1 13.5 27.6 14.5 21.1 100.0
SOURCE: CARVALHO, Jose A. M. and PAIVA, Paulo T. (1976), pp. 36-37.
The monthly household income refers to 1970 and the fertility levels to approximately the 1960-70
Not including people not living in households or living in households with no declared income.
It can be seen that in neither the rural nor the urban populations is
there a strictly linear relationship between income and fertility. Al-
though the overall trend is for fertility to decrease with income, it
increases in the lowest income categories, and the relationship is some-
what erratic for the rural population.
It can be seen in Table 11 that in 1970 only 21.1% of the house-
holds with monetary income received above Cr$500, which at that time
was a little more than 2.5 times the highest regional minimum wage.
IBGE has established a poverty level of 3.5 minimum wages per family.
The great majority of the Brazilian population lives below this poverty
line. In 1975, for example, 33.0% of the families in the Rio de Janeiro
region lived in absolute poverty, 46.2% in Minas Gerais and 81.8% in the
Northeast. 2 Thus, in spite of rapid growth of the "middle class," the
poor population continued to predominate.
When one considers possible composition effects of redistribution
of the population among income strata, particularly the possible depres-
sive effect on aggregate fertility levels of expansion of the "middle
class," three points should be taken into consideration: 1) Even if
the "middle class" increased rapidly in size, though social mobility, it
remained small, with limited impact on the population as a whole. 2)
Fertility differentials between adjacent strata, where mobility is most
likely to occur, are not large. 3) There is less margin for change in
fertility among the middle strata, where fertility is already relatively
Summing up possible composition effects, it seems that even though
fertility differentials among regions or between rural and urban areas
are large, the volume of migration in the short period under considera-
tion was too small to account for a large drop. Analysis of redistribu-
tion of the population among income levels is more complex. It is possible
that the rapid expansion of the higher strata produced an appreciable
composition effect. Still, it must be remembered that, even with rapid
rates of growth of the "middle class," the absolute volume of population
involved is small, and that the upwardly mobile may not have high fertil-
It should also be borne in mind that, in reality, redistribution
of the population in space or among income levels takes time to produce
effects on fertility, and that the period in question is quite short,
four to six years, depending on the calculation. Furthermore, mobility
occurred during the period, on the average, at the mid-point, leaving
very little time for readjustment of fertility, which is measured taking
into account the experience over a period of 5 or 10 years.
The general conclusion that can be drawn is that it is very unlikely
that redistribution of the population into geographical areas or income
strata with lower fertility, while certainly important in long-term trends,
could account for a drop on the order of 20% in a few years. A reduction
of this magnitude could hardly have occurred only through expansion of
the segments of the population which already had relatively low fertility
or through intensification of the decline in these segments. The data
which are available on fertility by region show that TFRs in the less-
developed regions with high fertility dropped almost as much as fertility
in the more-developed regions, which already had relatively low levels.
Data are not available on changes in fertilityby rural-urban residence
or by income level, but the magnitude and rapidity of the decline, even
in regions which are predominantly rural and quite poor, suggest that
explanations of the recent drop should not be restricted to the segments
of the population which have relatively low fertility, but should focus
mainly on the large mass which has, at least until recently, possessed
very high fertility.
III. Economic Basis for the Sharp Drop in Fertility
In the late 1970's and early 1960's, when the fertility decline
in Brazil accelerated, there was also an acceleration or deeping of
various broad social and economic changes which had begun in previous
decades as part of the process of industrialization and urbanization
which has characterized the country's development since at least the
1930's. For the most part, detailed data are not yet available for
the decade of the 1970's, but because of the time lags which are normal-
ly involved in fertility responses, use of data for the analysis of
changes during the 1960's is not entirely inappropriate to outline the
nature and distribution of economic changes among sub-groups of the
population which might have influenced fertility between 1965 and 1976.
The conventional measure of economic growth, per capital income, grew
from $392.25 in 1965 to $698.28 in 1974 (calculated in 1970 dollars), an
increase of 6.6% per year.22 Other changes were important if less dramatic.
According to the census, the urban population grew from 45% of the total
in 1960 to 56% in 1970. The average annual rate of growth for the urban
population in the 1960's was 5.1%, compared to only 0.6% for the rural
population. Between 1950 and 1960 the growth rate was highest in cities
with more than 500,000 inhabitants (6.3%), while between 1960 and 1970
the most rapid growth (8.7%) occurred in cities between 50,001 and 500,000
inhabitants. This change in the distribution of growth rates by size
category of cities indicates a widening of the base of the urban network
in Brazil, with a larger part of the population living in middle-sized
cities in various parts of the country. In 1960, 54.0% of the labor force
was engaged in agricultural activities, while by 1970, the proportion of
the labor force in agriculture dropped to 44.6%.24 Agriculture's share
of domestic income followed the same tendency, dropping from 26.7% in
1960 to 17.1% in 1969.25 This reflects an increase in the contribution
of the industrial and service sectors to domestic income and to labor
absorption. As can be seen in Table 12, the percentage of both males and
females in the 15-24 age bracket who were attending school grew rapidly
from 1960 to 1970. In the same period, female participation in the labor
force grew from 16.5% to 18.4%.26
These data would seem to indicate that Brazil underwent a typical
fertility decline associated with processes of economic growth, urbaniza-
tion, industrialization, modernization, etc., as suggested by the demo-
graphic transition model. The implication is that, as a growing proportion
of the population became integrated into the "modern" sector and gained
access to the benefits of development, it adapted its fertility behavior
by reducing the number of children. However, the surprising rate and
extent of the recent decline, as shown above, suggest that this interpreta-
tion must be qualified. Although a small but growing part of the Brazilian
population participated in the benefits of rapid economic growth in the
period in question and adopted new patterns of fertility, there is also
evidence of considerable fertility decline among regions and sectors of
the population who participated only marginally, if at all, in the benefits,
and whose situation in some respects may have even deteriorated. The
BRAZIL PROPORTIONS ENROLLED IN SCHOOL, BY AGE GROUP,
AGE 15-24, 1960 AND 1970
1960 1970 1960 1970
15 19 19.3 38.7 15.7 35.4
20 24 5.5 17.7 3.0 13.7
SOURCE: Censos Demograficos, 1960, 1970.
fertility estimates presented above show that there was a sharp reversal
of the previous upward tendency in fertility in the Northeast, without
evidence of substantial abatement of regional disparities. In the more
developed regions, on the other hand, acceleration of the fertility de-
cline already under way was so rapid that it could hardly have occurred
without involving the low-income population. Thus, the apparent associa-
tion between economic growth and fertility decline on the aggregate level
should be probed in order to discover the distinct mechanisms which may
be at work among different sub-groups of the population. This is difficult
with the information presently available, but some insights can be gain-
ed by turning attention to the specific characteristics of recent economic
growth in Brazil which may have led to the onset of a fertility decline,
especially among the low-income population, in the absence of national
birth control policies.
A. Inequalities in the Recent Economic Growth of Brazil
Brazil's economy grew rapidly in the entire post-war period. The
potential annual growth rate of Brazil's GNP is estimated at 7.08%,27
greater than that of the United States,8 and well above the rate of
demographic growth of about 2.5%. The growth of the actual product,
however, followed a pronounced cyclical pattern. Between 1962 and 1967
it grew at annual rates below 5% (the low point being 1.3% in 1963), while
the price index rose steadily, with inflation reaching 87% in 1964.29 The early
1960's was a period of political instability, culminating in 1964 in the over-
throw of the elected government. Various anti-inflationary measures
were then taken, including a rigid policy of wage controls, the prohibi-
tion of strikes and collective bargaining, the effects of which will be
discussed below. The rate of increase in prices diminished, and after
1968 the actual GNP began to grow at rates around 10% a year. The indus-
trial product grew at about 11% per year and the agricultural product
grew at rates of about 6% per year.30 Agricultural growth between 1968
and 1976 was led by export crops such as coffee and soybeans. While
the average annual growth of export crops (except coffee) was 9.8%, the
growth rate for domestic consumption crops was only 3.3%.31 As will be
shown below, the process of proletarianization of the agricultural labor
force intensified during the 1970's, when temporary wage-labor was substi-
tuted for resident or permanent labor partly as an unintended result of
the extension of labor legislation to cover rural workers.
The higher rate of growth of manufactured goods led to an increase
in their contribution to the domestic product. The growth of production
of durable consumer goods is especially noteworthy, and is responsible,
according to some analyses, for the economic recuperation of the late
1960's.32 These new consumer items were bought by a growing domestic
market constituted by a "middle class" with increasingly diversified
The improved purchasing power of this group can be detected in
data on concentration of individual income from the 1960 and 1970 censuses
(see Table 13). The 20% of the population that received 54.5% of the
income in 1960, received 63.2% in 1970. The percentage increase in real
average income was much greater at higher income levels, especially the
top 20%. On the other hand, the 50% at the base of the income pyramid
gained an insignificant increase of 1% in their real income between the
two censuses. This differential growth broadened the gap between the
strata with the highest and the lowest incomes. Other more refined
indicators show that there was actually a drop in real individual income
for urban workers between 1962 and 1972. This will be discussed below.
In addition to income data, there are other indicators that suggest
that the "middle class" expanded. Table 14, which shows employment in
administrative and technical or scientific occupations, gives an indirect
indication of growth of these strata. Between 1960 and 1970, the labor
force employed in these occupations grew at an average annual rate of
5.7%, more than double the rate of growth of the population (2.8%), and
well above the rate of growth of the non-agricultural labor force (3.7%).
Thus, it seems that in the 1960's, in addition to concentration of income
in the top 20%, there was also an increase in the relative size of the
B. Economic Trends Favoring Fertility Reduction Among the Low-Income
Although data are not available on fertility changes over time for
different income levels, occupational groups or other socio-economic strata,
the estimates based on the 1970 census, presented above, show that fertility
BRAZIL DISTRIBUTION OF INCOME PERCENT CHANGE IN
IN MEAN REAL INCOME, 1960 1970
Percent of Economically Percent of Income (2) (3) Percent Increase in
Active Population 1960 1970 Mean Real Income
(1) (2) (3) (4) (5)
Lower 50 17.7 13.7 -4.0 1
10 7.5 6.2 -1.3 8
10 9.0 7.2 -1.8 3
10 11.3 9.6 -1.7 10
10 15.6 14.8 -0.8 23
10 38.9 48.4 +9.5 6.1
Upper 5 27.4 36.2 +8.8 72
SOURCE: Hoffman, 1975
In WOOD, Charles
H. and Carvalho, Jose A. M. (1980), p. 19.
BRAZIL TYPICAL MIDDLE CLASS OCCUPATIONS, 1960 AND 1970
Occupations 1960 1970 Rate of Growth
(A) 1,821,285 2,997,023 5.1%
Professional and Technical
Occupations (B) 705,881 1,386,635 6.9%
"Middle Class" Occupations*
( C = A + B ) 2,527,166 4,383,658 5.7%
SOURCE: Censos Demograficos, 1960, 1970.
* Total of Managerial and Professional and Technical occupations.
is lowest among the higher income strata in urban areas. The process by
which such strata reduced or continued to reduce their fertility apparen-
tly accounts for the general downward tendency in Brazil in recent decades.
After about 1965, there was an acceleration of the decline for the country
as a whole, which affected all regions (at least those covered by the
PNAD surveys) and which involved the low income population in urban and
rural areas. The available evidence on intermediate variables suggests
that the decline was due primarily to deliberate control of marital fertil-
ity. Such deliberate control is a new phenomenon in Brazilian demographic
history, leading to a large drop in a short period of time. This reduc-
tion occurred during a period in which, in spite of rapid economic growth
at the aggregate level, the economic position of the low-income population
grew worse in relative terms, and in absolute terms as well, at least among
the urban poor.
The Brazilian experience thus casts doubt on some of the prevailing
thinking about the determinants of fertility declines. In the first place,
the poor population seems to have responded to environmental conditions
without the direct intervention of family planning policies. Secondly,
rapid economic growth may, in some cases, have contributed to the fertility
decline through deterioration of living conditions as inequality increased,
rather than through higher levels of income, schooling, etc. that are
commonly supposed to accompany urbanization and modernization. Such a
rapid drop suggests an abrupt change in family size patterns among a large
segment of the population rather then the simple continuation of a secular
trend. In this context, the problem is to explain why the pattern changed,
in other words, why low-income families began to control their fertility.
We have sought an answer in the historical conditions which determine the
standard of living of a population, the ways in which individuals earn
their livelihood, and in the changes in the composition of their consumer
basket. This question has been framed by Easterlin in terms of the inter-
play between aspirations and resources:
One of the determinants of the fertility of a married
couple is the tension between the couple's material aspira-
tions and their resources what might be termed the 'rela-
tive affluence' of the couple. If a couple's resources are
abundant relative to their aspirations, they will feel freer
to have children. If their resources are scarce relative to
aspirations, they will be hesitant about having children.34
If the recent decline in fertility in Brazil is placed in the frame-
work of the conflict between resources and aspirations, it may be possible
to account for changes in the low-income population as well as in the
The aspirations of the "middle class" seem to have grown more than
their income. When rapid economic growth resumed in 1968, there was an
increase in the demand for skilled labor, the scarcity of which35 led to
wage increases well above increases in the cost of living. This situa-
tion differed from the economic expansion of the first half of the 1950's,
when unskilled labor received real wage increases above those of the
Thus, in the recent boom, additional goods and services were placed
within the reach of the "middle class". The National Housing Bank made it
possible to purchase homes with long-term mortgages (up to 25 years), which,
because of indexation, kept a sizable part of family income committed for
long periods. The very rapid growth of production of consumer durables
(automobiles, electrical appliances, electronics, etc.), together with the
expansion of consumer credit, led to a growing diversification of "middle
class" consumption patterns. Increasing numbers of students were admitted
at all levels, especially colleges, which multiplied most rapidly in the
private sector, stimulated by an ideology of social mobility. These are
examples of how a growing number of families after the first half of the
1960's took on financial commitments that were not common in previous gene-
rations. Although there are few quantitative data, it is quite clear that
the new "middle class" profoundly changed its tastes, compared to its
parents' generation, especially when the parents lived in rural areas or
small cities. Such changes in consumption patterns almost certainly in-
duced members of the new "middle class" to have fewer children or to post-
For the low-income population, the question is not only one of how
much was earned, but how resources were acquired and used and what were
the implications for their standard of living. The majority of the popula-
tion remained outside capitalist relations of production until the 1960's.
Labor legislation covered only a small part of the urban labor force and
it was not applied in the countryside.37 Without this protection, rural
workers were not guaranteed the official minimum wages adopted in the
1930's to cover the expenses of an individual worker for food, clothing,
shelter, hygiene and transportation.38 The majority of Brazil's popula-
tion lived outside of the monetary economy, with a standard of living
close to the level of physiological subsistence, with food, housing and
shelter produced outside the market, i.e., in domestic production. Pos-
sible results of the expansion of wage-labor on fertility will be discus-
sed below, especially with regard to the relative price of goods in the
basic consumer basket.
Deterioration of the standard of living will be analyzed in the
same perspectives. Decline in workers' real wages during the 1960's and
early 1970's and rising relative prices for food stuffs crops during the
1970's reduced the purchasing power of low-income families, holding const-
ant the composition of their consumer basket.
It thus seems that the process of proletarianization from the 1960's
onward favored a reduction in family size and that the decline in the
standard of living of the urban poor also induced such families to use
deliberate fertility control in order to alleviate, at least partially,
pressure on their resources.
1. Increasing Proletarianization
In the European experience, the process of proletarianization appears
to have had a positive influence on fertility levels, principally because
in the peasant communities the occurrence of marriages was affected by
economic opportunities and fertility levels were subject to a certain
form of social control.39 Tilly observed that:
... the European peasant household (the peasant house-
hold or the peasant community, or both) operated as an ef-
fective population-control mechanism, closely matching the
opportunities for marriage and procreation to the number of
persons the land could support 0
He found that the expansion of wage labor weakened the system of
population control.41 Tilly goes on, estrapolating this interpretation
for the contemporary world, suggesting that:
the poor, economically dependent populations of the world
are repeating the demographic experiences of the proleta-
rian segments of Western rural populations under conditions
of more complete proletarianization and more thorough
penetration of capitalism than occurred in the rural
The Brazilian situation appears to be moving in the opposite direc-
tion of that suggested by Tilly, that is, the process of proletarianization
in Brazil may be producing a drop in fertility.
The typical organization of agricultural labor which is not totally
wage labor, in the Northeast and the Center-South regions of Brazil, differs
in various aspects from the peasant organization in Western Europe during
the pre-industrial period. In the first place, the large territorial expanse
of Brazil in relation to population size made economic growth possible under
conditions characterized by abundant land.43 In the second half of the 19th
century, the only region which appears to have been subject to demographic
pressure was the Northeast.4
In the second place, high mortality levels up until the 1930s probably
contributed to the formation of families at younger ages. Mortality estimates
of the Brazilian population indicate that life expectancy at birth at the
end of the 19th century was between 27.3 and 33.9 years. From the economic
point of view, given this mortality level, the peasant family could only
propagate if the father, upon dying, already had a grown son to take his
place in the production of food.
Thus, for the estimated level of mortality for the second half of
the 19th century, the Brazilian population had to have a pattern of marriage
,at a very young age, for both men and women. Unfortunately, there are no
available data for Brazil as a whole which show the average age of women
upon marriage during the second half of the 19th century.
In the third place, non-wage labor was not an institution which
acted counter to capitalist development. On the contrary, subsistence
activity has been responsible for a large part of food production and was
even incorporated into the organization of labor in the production of ex-
port crops (coffee and sugar). A classical example is the organization
of labor under the system of colonato. According to Holloway, in the
colonato the family income of the colono was composed of a monetary
part and a non-monetary part. The monetary wages were derived from
these sources: 1) the yearly cultivation of a given number of coffee
trees, which varied with family size; 2) the coffee harvest, 'the contract
for which was also made with the family of the colono and was paid accord-
ing to the acreage worked; and 3) daily labor on the coffee plantation,
which from the point of view of the monetary income of the colono, was
The non-monetary part was derived from the assignment of land for
the colono to grow his crops, which included corn, beans and domestic animals;
and from free shelter.
Since for the most part the principal items of the subsistence
basket were either produced domestically food or did not represent
a financial cost for the family housing there does not appear to have
been economic pressure for reduction in family size. On the contrary,
under this system, family income could only grow as family size increased.
On the one hand, the labor contract was with the entire family rather than
just with the colono, and the monetary wages were based on tasks (number
of coffee trees and acreage worked). Therefore, the larger the family,
the greater the monetary wages. On the other hand, the larger family had
comparative advantages in domestic food production, through economies of
scale and through the possibility of dividing labor. In both cases, the
marginal costs of domestic production were reduced as family size increased.48
This type of organization of agricultural labor is common through-
out Brazil, although with different characteristics from region to region,
and has continued up to the present. One of its most common forms is share-
cropping (parceria). In a recent study on sharecropping and family size,
Oz6rio de Almeida concluded:
... the sharecropping system in the Northeast constitutes
an institutional arrangement which regulates a specific
set of transactions between sharecroppers and landowners.
Landless peasants gain access to the land and to services
of food storage during times of need. Landowners, for their
part, obtain access to family labor thanks to the super-
imposition of the life cycles of individual members. Indebted-
ness is intrinsic to this relationship... The short-term ef-
fect consists of providing an incentive to increase the con-
tribution of family labor and specialization in the commercial
harvest. Its long-term impact apparently encourages increase
in family size, both through the reduction in costs of large
families as well as by the increase in benefits not only for
the sharecroppers, but also for the landowners. Consequently,
the sharecropping system appears to be linked to high
In short, the organization of non-wage labor in Brazil appears
to have contributed significantly to the establishment and continuance
of high fertility patterns. Table 15 shows that non-wage labor has been
the predominant form of agricultural labor in almost every region up until
1970.50 In 1970, wage labor constituted 50% of the male labor force in
agriculture in only two regions Sgo Paulo and Rio de Janeiro which
were also the regions which, in that year, exhibited the lowest fertility
The hypothesis suggested above is that the pattern of high fertil-
ity in Brazil many have been determined by the specific conditions of the
organization of domestic production under which large family size contri-
buted to lower the costs of subsistence of the labor force. The items
which constituted the largest share of the family consumer basket were
food and shelter. The former was produced almost entirely by the family
and the latter did not involve significant financial costs for rural labor-
ers. Given the high share of domestic production in the consumer basket
of the poorest families, price variations had little effect on the standard
of living of this population over time. The expansion in subsistence
production which accompanied the development of capitalism in the first
half of the 20th century, the growth in food production and the expansion
of agricultural frontiers probably contributed enormously to the relative
stability of national fertility levels.
The new phase of Brazilian capitalist growth after 1965 brought
various modifications to this picture, the most important among which,
with regard to fertility levels, was the intensification of the process
of proletarianization. This process meant the loss of the benefits of
domestic food production, making a large number of children not only
burdensome, but also unfeasible from the economic point of view. Thus,
contrary to orthodox Malthusian predictions, poor families adjusted to
the new objective conditions by reducing births through contraception.
Economic evidence supports this interpretation. As was seen in
Table 15, the demographic censuses indicate that there was a trend indicat-
ing a reduction of wage laborers in agricultural production in Brazil from
BRAZIL AND REGIONS WAGE LABOR AS PERCENTAGE OF THE LABOR FORCE; 1940 TO 1970
AGRICULTURAL ACTIVITIES NON-AGRICULTURAL ACTIVITIES
1940 1950 1960 1970 1940 1950 1960 1970
Brazil 35.36 34.88 26.74 26.31 72.15 73.40 73.25 77.12
Sao Paulo (VI) 50.56 47.38 51.00 77.21 78.12 80.38
Rio (VII) 49.21 49.67 50.68 82.41 81.81 83.15
Extreme South (IX) 19.59 16.62 15.30 72.28 72.48 76.59
Minas (V) 49.70 34.83 32.36 70.63 70.39 74.12
Parana (VIII) 36.92 26.02 20.84 72.59 68.55 72.34
Amazonia (I) 19.80 16.65 13.21 69.93 66.36 71.26
West (X) 23.45 22.26 22.49 61.33 66.91 69.09
Central Northeast (III) 33.04 25.60 29.95 66.35 66.85 69.74
Southern Northeast (IV) 28.18 19.11 23.24 64.55 61.18 69.11
Northern Northeast (II) 9.98 5.88 5.07 54.04 50.86 55.03
Brazil 20.29 24.43 17.04 17.07 62.74 81.10 75.65 83.95
Sao Paulo (VI) 33.31 40.62 54.40 88.68 88.57 90.19
Rio (VII) 38.59 44.15 44.03 89.11 88.70 89.89
Extreme South (IX) 6.25 5.02 3.13 82.22 82.52 87.48
Minas (V) 38.25 26.67 25.72 83.70 74.86 82.16
Parana (VIII) 19.60 10.01 9.66 83.00 81.42 87.30
Amazonia (I) 9.90 6.63 7.68 75.82 69.06 80.76
West (X) 16.66 15.28 11.29 73.38 65.62 77.42
Central Northeast (III) 31.85 21.21 29.59 67.67 54.30 69.92
Southern Northeast (IV) 20.55 11.93 16.13 64.38 56.97 71.34
Northern Northeast (II) 3.98 3.40 2.23 69.36 44.73 62.23
SOURCE: Censos Demogrificos 1940,
1950, 1960, 1970.
BRAZIL WAGE LABOR AS A PERCENTAGE OF THE LABOR FORCE, 1940 TO 1970
Rate of Change
1940 1950 1960 1970 1972 1976 1977 50-70 72-77
A All Activities
Total 45,6 50,6 47,9 55,2 55,5 61,6 62,6 0.87 2.40
Men 46,3 48,4 45.7 51,2 55,1 60,4 62,5 0.56 2.52
Women 42,9 64,0 58,0 70,2 56,5 64,7 62,7 0.92 2.08
Total 33,3 34,1 25,8 25,4 28,1 32,5 35,4 -2.92 4.60
Men 35,4 34,9 26,7 26,3 31,4 35,7 39,8 -2.81 4.72
Women 20.3 24,4 17,0 17,1 16,6 20,8 22,2 -3.52 5.77
Total 69,4 75,4 73,9 79,2 75,0 78,2 78,2 0.49 0.84
Men 72,2 73,4 73.3 77,1 74,8 76,8 77,5 0.49 0.71
Women 62,7 81,1 75,7 84,0 75,3 80,9 79,6 0.35 1.11
SOURCE: 1940 to 1970 -
1972 to 1977 -
1950 and 1970. This drop is not related to variations in the relative
share of women in the labor force, since it occurred both in the female
and male populations. Between 1972 and 1977 there was a reversal of this
downward trend, according to the data of the PNADs. The annual drop, in
20 years, of 2.9% in the relative share of wage labor in agricultural
production (from 1950 to 1970), was transformed into an annual increase
of 4.6% in only 5 years (from 1972 to 1977). In the total labor force
(Table 16) wage labor, as a percentage of the labor force, increased
0.9% per year from 1950 to 1970 and 2.4% per year from 1972 to 2977. Al-
though demographic census data cannot be compared with data from the
PNADs with regard to the labor force, the fact that, in five years during
the 1970s, the relative share of wage labor grew much more rapidly than
in the previous 20 years covered by the demographic censuses, seems to be
a strong indication.
The prominence of the process of proletarianization in Brazilian
agriculture in the period after 1965 can also be found in the following
analysis regarding the State of Sao Paulo:
The actively employed population in Sao Paulo agricul-
ture, the most developed in the country, and which has the
best available statistical data, fell by 37% during the
period 1964-75, principally through the elimination of the
resident the colono. The number of resident workers on
the properties was reduced by half, while the number of non-
residents increased by more than 40%. In large part, they
were the same laborers, but they were employed as volantes*
(day laborers, or laborers employed on a monthly basis) and 51
lived in nearby towns, under extremely precarious conditions.
This process of proletarianization destroyed a system of relations
that encouraged a pattern of high fertility. In the first place, this
process has meant the monetization of the labor force, eliminating advant-
ages of scale and division of labor derived from larger families. This
has resulted also in an increase in the costs of subsistence of the labor
force in various ways: (i) directly through the incorporation of a mone-
tary cost in several items of the consumer basket, such as housing;
(ii) by the increase in the cost of food purchased in the market, where
the profits of middlemen are added on at various stages of marketing;
* Translator's note: volantes non-resident laborers living in
nearby towns and commuting to the place of work on a daily basis.
or (iii) by the increased cost of food subject to oligopoly controlled prices.
In other words, in order to maintain the same consumer basket, the proletarian-
ized family has to work a greater number of hours. In the second place, the
labor contract became individual, rather than collective. With this transforma-
tion, uncertainty emerged regarding children's contributions to the family income.
Furthermore, if the child's wage was lower than his production under the subsist-
ence system, there would be a loss, when there was a job for him. Finally, if
in the process of proletarianization there was migration to urban centers, the
subsistence consumer basket was probably altered by the addition of new items,
such as transportation, electricity and water.
Insofar as domestic production is eliminated and the cost of sub-
sistence of the labor force increases, the process of proletarianization
appears to have acted as a disincentive to large families. This is a ge-
neric proposition which should be qualified in specific contexts. Unfor-
tunately, there are no data currently available, apart from those present-
ed in this paper, to examine this hypothesis in greater detail.
2. Deterioration of Living Standards
As seen in the previous discussion, the Brazilian economy experi-
enced a rapid surge in economic growth between 1968 and 1974. During the
same period, however, wage readjustments for unskilled laborers did not
keep pace with increases in the cost of living, thereby resulting in a
drop in real wages. Table 17 includes several indicators of real wages
for unskilled urban laborers. The minimum wage series, perhaps the most
representative of the mass of wage laborers, exhibits a diminishing trend
beginning in 1962 and lasting until the beginning of the 1970s. The series
for laborers in the civil construction industry suggest a drop in the
1970s, as well.
The minimum wage series by itself would only be sufficient to
indicate a fall in the standard of living of unskilled urban laborers
during the 1960s. Bacha, however, using other indicators of urban wages,
concludes that the "standard of living of unskilled urban laborers probably
grew worse"52 during the period between 1966 and 1977. It becomes neces-
sary to ascertain the degree to which the minimum wage prevails in the
total labor force in Brazil. Although there is currently a controversy
INDICATORS OF REAL WAGES OF UNSKILLED URBAN WORKERS IN BRAZIL
MINIMUM MEDIAN ASSISTANTS BRICKLAYERS'
WAGE (a) WAGE (b) WAGE (c) WAGE (c)
1961 118 102.0
1962 98 95.1
1963 91 91.4
1965 87 87.1
1966 78 84.8
1967 76 88.2
1968 75 97.6
1969 72 99.4 101.0 104.0
1970 73 100.0 100.0 100.0
1971 72 93.0 94.5 94.5
1972 73 89.2 99.1 89.9
1973 74 84.5 83.3 85.0
1974 83.3 89.5
1975 88.4 90.9
1976 86.8 83.6
1977 92.6 91.3
(a) Base year 1952, Souza and Baltar, 1979, p. 658.
(b) Median real wage for industry in the city of Rio de Janeiro, base year
1970. Bacha, 1979, p. 610.
(c) Real wage of assistants (serventes) and bricklayers (pedreiros) in
civil construction, base year 1970. Bacha, 1979, p. 618.
in the specialized literature regarding the importance of the minimum wage
as the "standard of remuneration of urban labor",53 PNAD data for large re-
gions such as Minas and the Northeast show that majority of workers receive
up to one minimum wage.54 In Sao Paulo, however, the data indicate a fall
in the proportion of the labor force earning up to one minimum wage.
Questions can also be raised with regard to the significance of the
minimum wage in terms of purchasing power. It was mentioned above that the
official minimum wage in Brazil was set in terms of the physiological needs
of an individual worker, not of his family. Consequently, a family with an
income of only one or two minimum wages would have to resort, necessarily,
to non-monetary sources of income in order to survive. This fact complicates
interpretation of trends in family income using census data, since the census
does not include non-monetary income. Furthermore, the PNAD definition of
income is somewhat different from that of the census.
It should be noted that a decline in the real wages of skilled work-
ers does not imply, necessarily, that family income also decreased, since the
decline may have forced more family members to seek employment. This seems
to be what occurred in Brazil.5 Although there are no data on family income
in 1960, data from 1970 and 1976 permit comparison of family income with indi-
vidual income, provided that care is taken with the difference in definitions
used in the census and in the PNAD survey. While the Gini index of concentra-
tion of individual income increased from 0.562 in 1970 to 0.589 in 1976, the
same index for family income decreased from 0.564 to 0.561. This reflects the
fact that, in spite of growing concentration of individual income in the 1970's,
families maintained their relative positions by increasing the amount of wage
labor performed, especially by family members other than the principal bread-
Data on infant mortality provide another indication of deterioration
of the living standards of the low-income population. According to Wood and
Sawyer,57 the infant mortality rate in Sio Paulo increased between 1961 and
1973, and in Belo Horizonte, between 1960 and 1973. These are two of the
three largest urban centers of the country, where the benefits of rapid eco-
nomic growth were concentrated, but there was also visible deterioration of
the quality of life of the low-income population.
Another important indicator of the lowering of living standards
among the poorest urban populations in Brazil, is the behavior of the rela-
tive prices of food staples. During the period 1965-75 the growth of food
production for domestic consumption dropped considerably, as can be seen in
STable 18. In several states, the rates of growth of production of beans,
rice and manioc were actually negative between 1965 and 1975.58
BRAZIL ANNUAL GEOMETRIC GROWTH RATES OF PRODUCTION OF RICE,
BEANS, POTATOES, CORN AND MANIOC, PERIODS FROM
1954-56 TO 1974-76
Products 1954-56 1964-66
Rice 6.4 1.8
Beans 3.8 0.6
Potatoes 4.3 3.6
Corn 4.9 5.0
Manioc 5.2 0.04
SOURCES: Rice and beans Lemos and Servilha, 1979, pp. 48-49.
products 1954 to 1966 from Conjuntura Economica, 26, 11, Nov.
p. 5; 1974 to 1976 from "Negocios em Exame," 1980, p. 60.
Barros and Graham observed that, during the period 1968-76, the
growth of agricultural production in the South region of the country "oc-
curred much more among export commodities" and the productivity of the
"domestic sector" declined.59 Lemos suggested that there was a negative
relationship between the production of manioc and the acceleration of the
gross domestic product growth rate.
What is important for the argumentation developed in this paper is
the fact that the domestic supply of food staples declined precisely at
the time that there was a sharp drop in fertility. Growth of the wage-
earning population during the period together with a relative reduction
in the food supply probably pushed food prices higher. This would contri-
bute to an increasein the cost of subsistence of agricultural laborers during
this period, as well as in the cost of food for urban laborers. Barros and
Graham showed that, in the city of Sao Paulo, food costs, which until 1972
had generally risen less than inflation, began to rise at a faster rate
from 1973 to 1975.61 Alves and Vieira, in a study on the trend in the
pattern of food consumption of the population of the city of Sao Paulo
in the last four decades, noted that there was an "increase in the relative
share of the minimum wage expended on food."62
Excluding the year 1950, when the minimum wage was the same as in 1945
(after 1945, the minimum was readjusted for the first time in 1952), in
the last three decades the percentage of the minimum wage required for the
purchase of the minimum requirement of calories and proteins in Sao Paulo
reached its maximum point in 1975 (See Table 19). From 1965 to 1975 the
cost of calories in relation to the minimum wage rose by 60%, while the
relative cost of proteins rose by 89%. Alves and Vieira concluded that
MINIMUM WAGES NECESSARY TO PURCHASE MINIMUM DAILY REQUIRE-
MENTS OF CALORIES AND PROTEINS FOR ONE ADULT 1939 1975
1939-40 1945 1950 1955 1960 1965 1970 1975
Minimum wages 0.209 0.261 0.419 0.174 0.201 0.197 0.279 0.315
Minimum wages 0.123 0.161 0.269 0.114 0.140 0.111 0.185 0.210
SOURCE: Instituto de Ppsquisas Economicas, cited in Alves and Vieira, 1978,
"... throughout almost the entire period, there was no consistent gap be-
tween the food price index and the consumer price index. Only in the first
half of the present decade have food prices risen at a faster rate than the
cost of living... The explanation for the relative increase of the cost
of food consumption in the second half of the 1960s was the wage policy which
kept increases in the minimum wage below price increases.64
These findings, however, are with reference only to the city of Sao
Paulo. Table 20 shows the wholesale price trend of agricultural and manu-
factured products during the period from 1948 to 1976. During the period
from 1955 to 1963 the wholesale prices of agricultural products tended to
rise less than the prices of manufactured products. This trend was rever-
sed in the period following 1964. It can be inferred that the sharper rise
in the wholesale prices of agricultural products in the latter period may
have also caused an increase in retail prices, after a certain lag. If this
is actually what happened, the relative cost of food increased for the wage-
earning population of Brazil. It is clear that regional studies would be
needed to evaluate the differential impact of these price increases.
Nevertheless, from the aggregate point of view, it can be affirmed
that, in the period between 1965 and 1975, Brazil experienced a drop in
the growth of food production for domestic consumption and a rise in rela-
tive food prices. At the same time, between 1964 and 1972, the real minimum
C. Hypotheses on the Acceleration of the Fertility Decline
Beginning in the second half of the 1960s, there appears to have been
a very sharp acceleration in the falling trend in fertility levels of the
Brazilian population, the initiation of which was already apparent in sev-
eral regions Minas, Rio de Janeiro and Sao Paulo since the 1930s. This
acceleration occurred mainly due to increased use of contraception. From
an economic point of view, there seem to have been two principal causes for
the acceleration in the decline in fertility: one structural the intensi-
fication in the process of proletarianization; and the other cyclical the
fall in the standard of living among large segments of the population.
In contrast to western Europe, the process of proletarianization in
Brazil may have disrupted previous social relations of production in such
way as to discourage higher fertility. As was seen above, the process of
ANNUAL PERCENT CHANGE IN PRICE INDICES OF AGRICULTURAL AND MANUFACTURED PRODUCTS
Agricul Manu- Agricul Manu-
tural fractured Difference tural fractured Difference
Products Products (3)=(1)-(2) r Products Products (3)=(1)-(2)
(1) (2) (1) (2)
1949 16.1 3.7 + 1964 99.4 83.4 +
1950 20.0 4.7 + 1965 42.4 61.4
1951 19.2 18.0 + 1966 42.0 32.4 +
1952 15.1 7.6 + 1967 24.5 25.6
1953 13.1 15.9 1968 17.2 30.4
1954 29.8 31.3 1969 21.6 20.2 +
1955 11.5 13.4 1970 28.9 17.2 +
1956 16.0 24.1 1971 25.1 17.5 +
1957 8.4 17.8 1972 22.3 16.0 +
1958 8.6 16.8 1973 19.1 14.9 +
1959 32.2 43.5 1974 29.2 29.4
1960 38.3 23.2 + 1975 24.1 29.2
1961 34.8 42.6 1976 59.0 36.0 +
1962 60.1 45.2 + 1977 49.5 39.2 +
1963 65.4 82.8
* Price indices refer to aggregate supply
26 of Conjuntura Econ6mica tables).
concept of Fundagao Getalio Vargas estimates (columns 17 and
Source: Conjuntura Econ6mica, several issues.
proletarianization increased the cost of subsistence of the labor force
through the substitution of foodstuffs purchased in the market for domes-
tically produced foodstuffs, even though this did not involve a change in
the composition of the consumer basket. In addition, other items placed
new burdens on the family, such as housing and transportation. These
transformations had the effect of discouraging large families. The divi-
sion of family labor and economies of scale no longer necessarily increased
family income; rather they resulted in increased costs of subsistence,
since food costs rosemore than among families which were not fully proleta-
rianized. With regard to rural laborers, the analysis of the trend of
real wages does not reveal a fall in the standard of living in the 1970s.
However, according to Bacha's estimates, the increase in the real wage
of day laborers and permanent workers in relation to the 1970 level in
16 Brazilian states only began in 1974.6 The real wage of these two cate-
gories of laborers fell between 1967 and 1973, a period in which food pro-
duction also fell. The combination of the process of rural proletarianiza-
tion with the relative drop in the real wage of laborers between 1967 and
1973 may have caused a drop in fertility in the poor rural population.
With regard to lower-income urban laborers, the process of proleta-
rianization appears to have had a smaller effect as compared to rural
laborers, as seen by the relative growth in wage labor (Table 15). Never-
theless, their standard of living undoubtedly fell during the period, due
to wage policy up to 1973 and due to the relative increase in food prices
during the 1970s, which made the continuation of the old fertility patterns
If this scenario corresponds to what actually happened, the Brazilian
experience does not confirm Tilly's hypothesis, mentioned previously, but
on the contrary suggests that the expansion of capitalism and of proletaria-
nization are creating conditions adverse to large families in Brazil. The
Brazilian experience is equally not confirming Malthusian predictions. The
drop in the standard of living did not take place due to excessive popula-
tion growth. The findings indicate that this drop occurred due to institu-
tional control of wage readjustments and to the increase in the relative
price of foods, both of which are independent of the rate of population
growth. The increase in the costs of subsistence of the labor force resulting
from the process of proletarianization also cannot be imputed to popula-
The Brazilian experience appears to point to a situation in which
the population is adjusting to historical conditions determined by the
process of proletarianization and to the standard of living.
If, on the one hand, the effects of the process of proletarianiza-
tion indicate a permanent change in the fertility pattern of the country,
on the other hand, the effects of the standard of living can be reversible.
Thus, an increase in the standard of living might lead to a net increase
in fertility, especially because 43% of the drop in the total fertility
rates between 1970 and 1975 occurred in the group of women from 20 to 29
years of age, among whom, presumably, fertility control took place prior
to the completion of the desired family size. An improvement in the
standard of living and more favorable conditions in the labor market could
induce these women to have more children, just as the younger cohorts have,
on the average, more children.66 However, it is very unlikely that Brazil
will return to fertility patterns established under conditions of subsistence
The purpose of this paper has been to present evidence of rapid de-
cline in fertility throughout Brazil in the late 1960s and early 1970s,
to examine the intermediate variables that may have contributed to this
decline, to consider the possible impact of family planning programs or
policies, and to suggest how Brazil's pattern of unequal economic growth
may have created the basic conditions favoring smaller families both
among the "middle class" and the low-income population.
The available data indicate a drop on the order of 20% in fertility
levels (TFR's) in all regions covered by the PNAD household survey, includ-
ing those regions which before the 1960s had constant or even rising lev-
els. It would be practically impossible for a reduction of this magnitude
to occur without involving the low-income population, which makes up the
great majority of Brazil's population.
The limited data presently available indicate widespread use of
contraceptives, including the rural population, with predominance of oral
pills and sterilization. Data on abortion are inconclusive, there being
a widespread opinion that the practice has become more common. Between
1960 and 1976, there was a small increase, of about one year, in women's
age at first conjugal union, which would not explain a sharp drop in fertility.
It is important-to note that this drop occurred in a country in which
there was no official family planning policy, of either outright control or
of an educational nature, although in a few states especially in the North-
east, the Brazilian Association for the Welfare of the Family (BEMFAM) has
The relatively affluent "middle class" was doubtlessly responsible
for part of the decline, during a period in which it received a growing
share of the national income, which was already quite concentrated, and
during which a variety of consumer durables became available.
The most important aspect of the decline, however, seems to be the
change in behavior of the low-income population. The Brazilian experience
shows that unequal income distribution does not necessarily retard a decline
in fertility, at least in a situation where the economic basis for the live-
lihood of the masses has been so profoundly altered as in Brazil. On the
one hand, intensification of the process of proletarianization drastically
reduced the contribution of domestic production to the consumption of a
large part of the Brazilian population. Thus, it became more difficult
to raise large families because children ceased to participate, for the
most part, in the production of the various items consumed directly by
the family, and because the prices of these goods on the market were higher
and grew faster than wages. Low-income families in Brazil have a standard
of living very close to subsistence. In the 1960s there was a decrease in
the real wages of unskilled urban workers, which could only be compensated
for by greater participation in labor markets, especially by additional fam-
ily members. In the 1970s, in spite of a small increase in real wages, the
prices of foodstuffs increased more rapidly than the prices of manufactured
goods, further depressing the standard of living of the low-income popula-
tion, and making it less feasible to maintain large families.
1 For a discussion of historical trends of
the population of Brazil, see Merrick and Graham, 1979.
2 See for example Mortara, 1958; Santos, 1975; J. Carvalho,1973;
3 See Carvalho, 1974, pp. 415-418.
4 Due to problems in the 1960 census, estimates are not
presented for the decade following 1960. Estimates correspond
approximately to the decade preceding the census year since
the fertility level is defined by the parity of women aged
20-29 in the census.
See Carvalho, 1974, p. 409.
Carvalho, 1978, passim.
United Nations, 1969, P. 577.
1Carvalho and Wood, 1978.
For an explanation of the Pi/Fi series, see Brass, et al.,
1968, Chapter 3.
1For detailed description of the study and the selected areas,
see the series of volumes under the title Estudos de Popula-
ao published by CEBRAP, 1978-1980.
Nakamura and Fonseca, 1979.
1Berquo, et al., eds., 1977.
Nakamura and Fonseca, 1979, passim.
"Denunciamos Essa Hipocrisia." Em Tempo, 1980, p. 8.
It is a well-known fact in demography that older women
underreport the number of liveborn children they have had.
It is probable, therefore, that underreporting of mis-
carriages or induced abortions which occurred in the distant
past would be even greater.
For a review of changes in population policies in Brazil, see
Mauldin and Berelsonl978.
2Tsui and Bogue, 1978.
21 Fundaggo IBGE, 1979, p. 215.
Pundaco IBGE, 1979, p. 192,using 1970 exchange rate of
23 Fundaco IBGE, 1979, p. 44.
24 errick and Graham, 1979, p. 162.
2Conjuntura Econ8mica, 25, 9 (Sept. 1971), p. 102;and 27, 12
(Dec. 1973), p. 7.
Female labor force participation refers to the percentage of
the female population age 10 or more which is engaged in the
labor force. Merrick and Graham, 1979, p. 165.
The potential growth rate is the growth rate of the GNP
estimated on the basis of peak years (1950, 1961 and 1974).
See Bacha, 1976, pp. 17-18.
Branson. 1972, p. 7.
For critical analysis of the Brazilian economy in the post-
war period, see, among others, Bacha, 1976, pp. 17-39;
Furtado, 1972; Fishlow, 1973; Tavares, 1975.
For real product indices bysector see Conjuntura Econrmica 27,
12 (Dec. 1973), p. 5.
1 For an analysis of Brazilian agriculture in this period, see,
"among others, Barros and Graham, 1978.
See Suzigan, et al., 1974.
Tavares, 1975, p. 202.
Easterlin, 1976, p. 417.
L. Carvalho, 1973, p. 17.
According to Macedo and Garcia, 1978, p. 18, coverage by the
minimum wage is limited because many workers with monetary
income receive less than the official minimum wage. According
to the authors, this is due to the "existence of income
derived from other sources which are not restricted by the
legal minimum, non-compliance with the law, existence of non-
-monetary .income and informal work relations which by their
very nature are not covered by legal restrictions."
38 For a discussion of the economic significance of the official
minimum wage, see Oliveira, 1975, PP. 11-12, and Werneck
Vianna, 1977, especially p. 113.
39 Among the vast literature, see especially Wrigley, 1969, and
Tilly, 1978, p. 22.
4Tilly, 1978, p. 22.
Tilly, 1978, p. 22.
43 This aspect is emphasized in practically all of the economic
histories of Brazil.See, among others, the classic study by
44 Furtado, 1961, p. 154.
Merrick and Graham, 1979, p. 42.
4Oliveira, 1975 and 1977.
An assumption implicit in this argument is that the area of
land provided to the colono for domestic production was large
enough to avoid diminishing returnsfor families with, for
example, 7 to 10 children.
Almeida, 1977, p. 330.
These data are from the demographic censuses and should be
interpreted with caution, especially with regard to enumera-
tion of women in non-wage agricultural labor and to the
definition of wage-labor, which refers here to workers
enumerated as "employees" in the census.
51 "Negbcios em Exame", 1980, p. 60. See also Saint, 1980.
Bacha, 1979, p. 627. The indicators Bacha used for wages of
unskilled urban workers were wages for assistants (serventes)
and bricklayers (pedreiros) in civil construction and median
monthly wages in industry.
See for example Macedo and Garcia, 1978, and Souza and Baltar,
54 Macedo and Garcia, 1978, p. 10. According to PNAD data, the
percentages of employees receiving up to one minimum wage for
Brazil as a whole declined from 43,2% in 1972 to 39.7% in
1972, while the percentage receiving up to two minimum wages
declined from 75.4% to 70.6% in the same period.
55 According to data cited in Macedo and Garcia, 1978, pp. 40-41,
the percentage of family members engaged in the labor market
in the municipio of Sao Paulo in 1971-72 increased with per
capital family income from 0 to 1.5 minimum wages. Such
increase was due to the participation of women and children.
56 Fundagio IBGE, 1979, pp. 196 and 207.
57According to Sawyer, 1980, p. 193, infant mortality rates in
Sao Paulo increased from 59.3 per thousand live births in
1960-64 to 86.9 in 1970-73. Wood, 1977, p. 58, shows that
infant mortality in Belo Horizonte increased from 74.2 in 1960
to 124.8 in 1973. In both cities the infant mortality curves
are almost mirror images of real wage curves.
From 1964-66 to 1974-76, the annual rate of growth for produce
tion of rice was negative in the states of Goias, Minas Gerais
and Sao Paulo and for production of beans in Rio Grande do Sul,
Goias and Sao Paulo, according to Lemos and Servilha, 1979,
pp. 48-49. From 1966-1968 to 1976-78, manioc production
presented negative growth rates for Brazil as a whole and for
the states of Rio Grande do Sul, Santa Catarina, Minas Gerais,
Sao Paulo, Sergipe and Parana, according to estimates provided
by Mauricio Lemos in personal communication.
Barros and Graham, 1978, p. 702.
This observation was made by Mauricio Lemos in his research
project on manioc production, now in its final stages.
Barros and Graham, 1978, p. 712.
6Alves and Vieira, 1978, p. 748.
Alves and Vieira, 1978, p. 750, emphasis added.
Alves and Vieira, 1978, p. 750.
6Bacha, 1979, p. 623.
A mechanism similar to that of the "baby boom" in the post-
-war years in advanced capitalist societies, in the interpre-
tation of Richard Easterlin. See, for example, Easterlin and
Almeida, Anna Luiza Ozbrio de. 1977. "Parceria e Tamanho de Fami
lia no Nordeste Brasileiro." Pesquisa e Planejamento Economi-
Qo, 7, 2, agosto, pp. 291-331.
Altmann, Ana Maria Goldani and Laura R. Wong. 1980. "Padrbes e
Tendencias de Nupcialidade no Brasil." Paper presented at
the III Encontro Nacional de Estudos Populacionais, Associa-
c9o Brasileira de Estudos Populacionais, Aguas de Sao Pedro,
13 a 17 de outubro de 1980.
Alves, Edgard L. G. and Jose L. T. M. Vieira. 1978. "Evolugao do
Padrao do Consumo Alimentar da Populaco da Cidade de Sao
Paulo." Pesquisa e Planejamento Econ8mico, 8, 3, dezembro,
Bacha, Edmar. 1976. Os Mitos de uma D&cada. Rio de Janeiro: Paz
Bacha, Edmar L. 1979. "Crescimento Urbano, Salarios Urbanos e Ru
rais: o Caso do Brasil." Pesquisa e Planejamento Economico,
9, 3, dezembro, pp. 585-628.
Barros, Jose R, M. de and Douglas H. Graham. 1978. "A Agricultu-
ra Brasileira e o Problema da Produgao de Alimentos." Pes-
quisa e Planejamento Econ8mico, 8, 3, dezembro, pp. 695-726.
Berqu6, Elza S. 1977. "A Fecundidade Rural-Urbana dos Estados
SBrasileiros em 1970." Revista Brasileira de Estatistica, 38,
151, jul.-set., pp. 251-303.
Berqu6,.Elza S. 1980. "Algumas Indaga(ces sobre a Recente Queda
da Fecundidade no Brasil." Paper presented in the VI Reuniao
do Grupo de Trabalho sobre o Processo de Reprodugao da Popu-
lac5o, Comissao de Popula(co e Desenvolvimento da CLACSO,
Teres6polis, abril de 1980.
Berqub, Elza S., Maria Coleta A. F. de Oliveira and Candido P. F.
de Camargo, eds. 1977. A Fecundidade em Sao Paulo: Caracte-
risticas Demograficas, Biol6gicas e Sbcio-Econ6micas. Sao
Paulo: CEBRAP, Editora Brasileira de Ciencias.
Branson, William H. 1972. Macroeconomics: Theory and Practice.
New York: Harper & Row.
Brass, William, et al. 1968. The Demography of Tropical Africa.
Princeton: Princeton University Press.
Carvalho, Jose A. M. 1973. Analysis of Regional Trends in
Fertility, Mortality and Migration in Brazil: 1950-1970.
Ph.D. dissertation, London School of Economics.
Carvalho, Josh A. M. 1974. "Regional Trends in Fertility and
Mortality in Brazil." Population Studies, 28, 3, December,
Carvalho, Jose A. M. 1978. Fecundidade e Mortalidade no Brasil -
1960/70. Belo Horizonte: CEDEPLAR.
Carvalho, Jose A. M. 1980. "Evolugao Demografica Recente no
Brasil." Pesquisa e Planejamento Econ8mico, 10, 2, agosto,
Carvalho, Jose Alberto M. and Paulo T. de A'. Paiva. 1976. "Estru-
tura de Renda e Padrbes de Fecundidade no Brasil." In Manoel
Costa, ed. Fecundidade: Padries Brasileiros. Rio de Janeiro:
Altiva, pp. 21-38.
Carvalho, Josh A. M. and Charles H. Wood. 1978. "Mortality,
Income Distribution, and Rural-Urban Residence in Brazil.
Population and Development Review, 4, 3, Sept., pp. 405-420.
Carvalho, Livio W. R. de. 1973. Principios e Aplicacao da Poli-
tica Salarial P6s-1964. Brasilia: Departamento de Economia,
Institute de Ci@ncias Humanas, Universidade de Brasilia.
Textos para Discussao, 9.
CEBRAP. 1978-1980. Estudos de Populacao. Sao Paulo. Volumes I to
"Denunciamos Essa Hipocrisia." 1980. Em Tempo, 3, 98, 24 de ja-
neiro, pp. 8-9.
Easterlin, Richard A. 1976. "The Conflict Between Aspirations
and Resources." Population and Development Review 2, 3 and
4, .Spt.- Dec., pp. 417-425.
Easterlin, Richard and Gretchen A.. Condran. 1976. "A liote or the 59.
Recent Fertility Swing in Australia, Canada, England and
wales, and the United States." In Hamish Richards, ed.
Population, Factor Movements and Economic Development.
Cardiff: University of Wales Press,-pp. 139-151.
Fishlow, Albert. 1973. "Some Reflections on Post-1974 Brazilian
Economic Policy." In Alfred Stepan, ed. Authoritarian Brazil.
New Haven: Yale University Press, pp. 69-118.
Fundacao IBGE. 1979. Indicadores Sociais, Relat6rio 1979. Rio de
Furtado, Celso. 1961. Formaqco Econ6mica do Brasil, 4 ed. Rio de
Janeiro: Fundo de Cultura. Trans. The Economic Growth of
Brazil: A Survey from Colonial to Modern Times. Berkeley and
Los Angeles: University of California Press. 1963.
Furtado, Celso. 1972. Analise do 'Modelo' Brasileiro. Rio de Ja-
neiro. Civilizagao Brasileira.
Habakkuk, H. J. 1972. Population Grouth and Economic Develomer.t.
Leicester: Leicester University Press.
Hoffman, Rodolfo. 1975. "Tendencias da Distribuicao da Renda no
Brasil e suas Relacoes com o Desenvolvimento Ecocnimco."
In R. Tolipan and A. C. Tinelli, eds. A Controversia sobre
Distribuicao de Renda e Desenvolvimento. Rio de Janeiro: Zahar.
Holloway, Thomas H. 1972. "CondiCqes do Mercado de Trabalho e
Organizacqo do Trabalho nas Plantacqes na Economia Cafeeira
de Sao Paulo, 1885-1915 Uma Analise Preliminar." Estudos
Econ6micos, 2, 6, December, pp. 145-180.
Lemos, Mauricio Borges and Valdemar Servilha. 1979. Formas de
Organizaqco da Producqo de Arroz e FeijAo no Brasil. Brasi-
Levy,Maria Stella F. and Rosalba Moledo. 1977. "Caracterizacao
Skcio-Demografica de Tres Coortes." In Elza S. Berqub, et
al., eds. A Fecundidade em Sao Paulo: Caracteristicas Demo-
graficas, Biolbgicas e Sbcio-Econ8micas. S3o Paulo: CEBRAP,
Editora Brasileira de CiEncias, pp. 337-372.
Macedo, Roberto and Manoel Garcia. 1978. "ObservacSes sobre a Poll-
tica Brasileira de Salfrio Minimo." Sao Paulo: FIPE/USP, mimeo.
Mauldin, W. Parker and Bernard Berelson. 1978. "Condiciones del
Decenso de la Fecundidad en los Paises en Desarrollo, 1965-
-1975." Estudios de Poblacibn, III, 1-6, enero-junio,
Merrick, Thomas W. and Douglas H. Graham. 1979. Population and
Economic Development in Brazil 1800 to the Present.
Baltimore and London: Johns Hopkins University Press.
Mortara, Giorgio. 1958. "A Fecundidade das Mulheres e a Sobrevi-
vencia dos Filhos nas Populagbes Urbanas e Rurais no Brasil."
In IBGE, Pesquisas sobre a Fecundidade das Mulheres e a So-
brevivencia dos Filhos no Brasil. Rio de Janeiro. Estatistica
g, Demografica, 25.
Nakamura, Milton S. and Joaquim P. Barreto Fonseca. 1979. "Sao
t Paulo State Contraceptive Prevalence Survey." Campinas: PESMI,
Pontificia Universidade Catolica de Campinas.
National Academy of Sciences. 1979. Preliminary Report of the
Panel on Brazil. Washington, unpublished preliminary version.
"' Negbcios em Exame." 1980. Brasil em Exame, marco, Edicao Especial.
Oliveira, Francisco de. 1975. "A Economic Brasileira: Critica A
Razao Dualista." Selecbes CEBRAP, 1. Sao Paulo: CEBRAP,
Oliveira, Francisco de. 1977. A Economia da Dependencia Imperfei-
ta. Rio de Janeiro: Graal.
Paiva, Paulo de T. Almeida. 1980. "Algumas Hip6teses sobre as Rela-
q6es entire Proletarizaqio e Fecundidade no Brasil." In Anais do
VIII Encontro Nacional de Economia. Nova Friburgo: Associaqio
Nacional de Centros de Pds-Graduaqio em Economia, pp. 331- 380.
Saint, William. 1980. "Mao-de-obra Volante na Agricultura Brasi-
leira: uma Revisao da Bibliografia." Pesquisa e Planejamento
,on Econ8mico, 10, 2, agosto, pp. 503-526.
Santos, Jair L. 1975. "Analise das VariAveis Demograficas atra-
v6s dos Resultados dos Censos de 1940, 1950 e 1960." Sao
Sawyer, Diana Oya. 1980. Mortality-Fertility Relationships
through Historical Socio-Economic Change: the Case of Sao
Paulo, Brazil. Doctoral thesis, Harvard University.
Souza, Amaury de. Forthcoming. "Populacgo e Politica Populacio-
nal no Brasil: uma Resenha de Estudos Recentes." Informativo
e Bibliografico de Ciencias Sociais, 22.
Souza, Paulo Renato de. 1980."Salario e Mao-de-Obra Excedente."
CEBRAP, Valor, Forca de Trabalho e Acumulacao Monopolista.
Sio Paulo. Estudos CEBRAP no. 25, pp. 67-112.
Souza, Paulo Renato de and Paulo R. Baltar. 1979. "Salario Mi-
nimo e Taxa de Salario no Brasil." Pesquisa e Planejamento
.Econ8mico, 9, 3, dezembro, pp. 629-660.
Suzigan, Wilson, et al. 1974. Crescimento Industrial no Brasil:
Incentives e Desempenho Recente. Rio de Janeiro: INPES/IPEA.
Relatbrio de Pesquisa, 26.
Szmrecsanyi, Maria Irene de Q. F. 1977. "Instrucao da Mulher e
Compqrtamento Reprodutivo." In Elza S. Berqu6, et al., eds.
A Fecundidade em Sao Paulo: Caracteristicas Demograficas,
Biol6gicas e Sbcio-Econ8micas. Sao Paulo: CEBRAP, Editora
Brasileira de Ciencias, pp. 289-336.
Tavares, Maria da Conceicao. 1975. Da Substituicio de Importa-
9ces ao Capitalismo Financeiro, 4 ed. Rio de Janeiro: Zahar.
Tilly, Charles, ed. 1978. Historical Studies of Changing
Fertility. Princeton: Princeton University Press.
Tsui, Amy Ong and Donald J. Bogue. 1978."Declining World Fer-
tility: Trends, Causes, Implications." Population Bulletin,
"24 Anos de Economia Brasileira Estatisticas BAsicas." 1972.
Conjuntura Econ8mica, 26, 11, novembro, Suplemento Especial.
United Nations. 1969. Demographic Year Book, 1969. New York:
Viana, Luiz Werneck. 1977. "Leis Sociais e Demografia." Estudos
CEBRAP, 21, julho-set., pp. 93-147.
Wood, Charles H. 1977. "Infant Mortality Trends and Capitalist
Development in Brazil: the Case of So Paulo and Belo Hori-
zonte." Latin American Perspectives, 4, 4, Pall, pp. 56-64.
Wood, Charles H. and Jose A. M. de Carvalho. 1980. "Population
Growth and the Distribution of Household Income: the Case of
Brazil." Gainesville and Belo Horizonte, mimeo.
Wrigley, E. A. 1969. Population and History, New York: World