You are on page 1of 18

Population Studies

ISSN: 0032-4728 (Print) 1477-4747 (Online) Journal homepage: http://www.tandfonline.com/loi/rpst20

Latin American Households in Comparative


Perspective
Susan De Vos
To cite this article: Susan De Vos (1987) Latin American Households in Comparative
Perspective, Population Studies, 41:3, 501-517, DOI: 10.1080/0032472031000143026
To link to this article: http://dx.doi.org/10.1080/0032472031000143026

Published online: 04 Jun 2010.

Submit your article to this journal

Article views: 50

View related articles

Citing articles: 9 View citing articles

Full Terms & Conditions of access and use can be found at


http://www.tandfonline.com/action/journalInformation?journalCode=rpst20
Download by: [Universidad Autonoma de Barcelona]

Date: 06 November 2015, At: 08:52

Population Studies, 41 (1987), 501 517


Printed in Great Britain

Latin American Households in


Comparative Perspective*
SUSAN

DE VOSf

Downloaded by [Universidad Autonoma de Barcelona] at 08:52 06 November 2015

INTRODUCTION
A d i s t u r b i n g element in recent theorizing a b o u t the role o f the h o u s e h o l d in
i n d u s t r i a l i z a t i o n o r d e p e n d e n t d e v e l o p m e n t I is the fact t h a t little is k n o w n a b o u t the
c o m p o s i t i o n o r o r g a n i z a t i o n o f h o u s e h o l d s in m o s t areas o f the d e v e l o p i n g world,
including L a t i n A m e r i c a . ~ Is it r e a s o n a b l e to theorize by using a n u c l e a r m o d e l t h a t
e m p h a s i z e s a m a n , his wife a n d children 3 o r is it m o r e r e a s o n a b l e to use a m o d e l o f a n
e x t e n d e d h o u s e h o l d t h a t includes o t h e r kin as well ?4 A r e h o u s e h o l d s h e a d e d by w o m e n ,
either single-parent h o u s e h o l d s o r e x t e n d e d h o u s e h o l d s , c o m m o n e n o u g h to be included
in a n y general t h e o r y ? H o w c o m m o n are u n r e l a t e d servants, lodgers o r b o a r d e r s ?
D u r i n g the early 1960s, one o f W i l l i a m G o o d e ' s 5 m a j o r conclusions a b o u t family
c h a n g e a r o u n d the w o r l d was that m o d e r n i z a t i o n was a s s o c i a t e d with m o v e m e n t
t o w a r d s a conjugal, e g a l i t a r i a n family system. This w o u l d i m p l y that theories o f social
r e p r o d u c t i o n t h a t used a nuclear family m o d e l might, indeed, be a p p r o p r i a t e . H o w e v e r ,
a n u m b e r o f E u r o p e a n social historians b e g a n to argue d u r i n g the late 1960s, on the
basis o f new historical evidence, t h a t the W e s t e r n family was basically c o n j u g a l even
before i n d u s t r i a l i z a t i o n , ~ a n d t h a t the r e l a t i o n s h i p between i n d u s t r i a l i z a t i o n a n d family
o r h o u s e h o l d structure will be different elsewhere. 7 In p r e - i n d u s t r i a l times they argue, the
h o u s e h o l d f o r m a t i o n system o f N o r t h w e s t e r n E u r o p e was distinct in t h a t the age at
m a r r i a g e o f b o t h m e n a n d w o m e n was relatively high, a c o n j u g a l couple h a d to establish
* The author gratefully acknowledges permission from the statistical offices of Colombia, Costa Rica, the
Dominican Republic, Mexico, Panama and Peru to use World Fertility Survey data to study household
structure in Latin America. The project was funded by NICHD under grant HD18788. Mr Beverly Rowe,
former Head of the Computer and Archive Division of the World Fertility Survey, facilitated the project's
access to the data. Cheryl Knobeloch and Roger Wojtkiewicz provided assistance in data handling. Shirley
Mellema and Valerie Bower provided secretarial assistance. Thomas Burch and an anonymous reviewer
provided helpful comments on earlier versions of this paper. Facilities of the Center for Demography and
Ecology of the University of Wisconsin-Madison, funded by NICHD Center Grant HDO5876, are also
gratefully acknowledged.
i" Research Associate at the Center for Demography and Ecology, University of Wisconsin-Madison, 1180
Observatory Drive, Madison, Wisconsin 53706, U.S.A.
i Joan Smith, Immanuel Wallerstein and Hans-Dieter Evers (eds.), Households and the World-Economy
(Beverly Hills, Ca. : Sage Publications, 1984).
Thomas K. Burch, 'Household and family demography: a bibliographic essay', Population Index, 45
(1979), pp. 173-195.
a Friedrich Engels and Karl Marx, 'The basis of the family', Chapter 6 in Neil J. Smelser (ed.), Karl Marx
on Society and Social Change (Chicago : University of Chicago Press, 1973).
4 Robert Hackenberg et al., 'The urban household in dependent development', Chapter 8 in Robert
McC. Netting et al., Households: Comparative and Historical Studies o f the Domestic Group (Berkeley,
University of California Press, 1984).
5 William J. Goode, Worm Revolution and Family Patterns (New York: The Free Press, 1963).
e Peter Laslett, ' Introduction ', Chapter ! in Household and Family in Past Time (Cambridge: Cambridge
University Press, 1972).
7 Peter Laslett and other social historians have used evidence on household composition to discuss problems
about family structure. Although the family and the household should not be equated, the nature of the
household can provide indications about the family.
501

Downloaded by [Universidad Autonoma de Barcelona] at 08:52 06 November 2015

502

SUSAN DE VOS

their o w n h o u s e h o l d u p o n m a r r i a g e s a n d h o u s e h o l d s t e n d e d to be simple. 9 It is also


o b s e r v e d t h a t m a n y rural p r e - i n d u s t r i a l h o u s e h o l d s included m e m b e r s w h o were n o t
related to the h o u s e h o l d head, b u t were y o u n g p e o p l e p e r f o r m i n g a p e r i o d o f ' s e r v i c e '
before they m a r r i e d a n d established h o u s e h o l d s o f their own. 1' n This implies t h a t the
n u c l e a r family m o d e l b a s e d on W e s t e r n t h e o r y m a y not be a p p r o p r i a t e .
C o m m o n to b o t h the m o d e r n i z a t i o n a n d W e s t e r n family a r g u m e n t s however, is the
fact t h a t neither is b a s e d u p o n m u c h i n f o r m a t i o n a b o u t L a t i n A m e r i c a . There was
insufficient i n f o r m a t i o n for G o o d e TM to include the region in his treatise on family
change, while the E u r o p e a n social historians discuss India, C h i n a a n d J a p a n b u t have
neglected a c o m p a r i s o n between N o r t h w e s t E u r o p e a n d L a t i n A m e r i c a 3 ~' 14 A r g u m e n t s
a b o u t the distinctive e c o n o m i c contexts o f c o l o n i a l i s m a n d d e p e n d e n t c a p i t a l i s m in
L a t i n A m e r i c a t~ also fail to p r o v i d e evidence o n w h a t the f a m i l y o r h o u s e h o l d f o r m is.
This is n o t surprising since m a n y censuses f r o m the region still d o n o t c o n t a i n even the
simplest i n f o r m a t i o n o n h o u s e h o l d s , a n d detailed i n f o r m a t i o n on h o u s e h o l d t y p e is
a l m o s t entirely lacking (for a recent s u m m a r y , see Torrado~S).
The p u r p o s e o f this p a p e r is to e x a m i n e survey d a t a o n h o u s e h o l d c o m p o s i t i o n in six
c o u n t r i e s t h a t c a n lead to tentative suggestions a b o u t the similarities a n d distinctions
between a L a t i n A m e r i c a n h o u s e h o l d system a n d t h a t o f N o r t h w e s t e r n E u r o p e o r
elsewhere. 17 T o d o so, I e x a m i n e several d i m e n s i o n s suggested b y s c h o l a r s : (1) h o u s e h o l d
complexity, (2) age at m a r r i a g e a n d the s e p a r a t e residence o f conjugal couples, (3) the
p r o p e n s i t y to have u n r e l a t e d h o u s e h o l d m e m b e r s a n d (4) female headship. I use d a t a
a John Hajnal, 'Two kinds of preindustrial household formation system', Population and Development
Review, 8 (1982), pp. 449~94.

9 Peter Laslett, 'Characteristics of the Western family considered over time', pp. 12-50, in Peter Laslett,
Family Life and Illicit Love in Earlier Generations (Cambridge: Cambridge University Press, 1977).
10 Hajnal, loc. cir. in footnote 8.
11 Richard Wall, 'The age at leaving home', Journal of Family History, 3 (1978), pp. 181-202.
1~ Goode, op. cit. in footnote 5.
is Laslett, op. cit. in footnote 6.
14 Hajnal, loc. cit. in footnote 8.

15 Francesca M. Cancian, Louis Wolf Goodman and Peter H. Smith, 'Capitalism, industrialization, and
kinship in Latin America: major issues', Journal of Family History, 3 (1978), pp. 319-336.
1~ Susana Torrado, 'Estrategias familiares de vida en America Latina: la familia como unidad de
investigacion censal (primera parte)', Notas de Poblacion, no. 26, pp. 55-106; and 'Familiares de vida en
America Latina: la familia como unidad de investigacion censal (segnnda parte)', Notas de Poblacion, no. 27,
1981.
17 The issue of conceptualization and definition plagues any comparative study that attempts to find
common patterns within diverse social contexts. It has been argued that it is meaningless or futile to try to
compare households in different societies because a single concept of 'household' is inappropriate in all
contexts, or because the household is properly considered in terms of function rather than form. (See e.g.
Benjamin N. F. White, 'Rural household studies in anthropological perspective', pp. 3-25, in Hans P.
Binswanger et al. (eds.), Rural Household Studies in Asia (Singapore: Singapore University Press, 1980); Hymie
Rubenstein, 'Caribbean family and household organization: some conceptual clarifications', Journal of
Comparative Family Studies, 14 (1983), pp. 283-298; Richard Wilk and Robert McC. Netting, 'Households:
changing forms and functions ', in Robert McC. Netting et al., Households: Comparative and Historical Studies
of the Domestic Group (Berkeley, University of California Press, 1984).) Yet the 'household' is at the core
of much social theory that attempts to generalize about social behaviour beyond the confines of one culture
with one technology. For no other reason than this, we must develop standard measures, conscious that the
more an analysis adheres to a particular scheme of analytic categories, the more likely it is to be comparable,
and the less likely it is to adhere to actual behaviour. Conversely, the more strictly it adheres to specific folk
categories, the less likely it is to be comparable (Eugene A. Hammel, 'On the *** of studying household form
and function', Chapter 2 in Robert McC. Netting et al., Households: Comparative and Historical Studies of the
Domestic Group (Berkeley, University of California Press, 1984). Demographers define the household in
terms of the co-residential unit, recognizing that all household members thus defined may not share a common
budget or take their meals together in all circumstances, and that significant transfers may exist between
households (Richard Wall, 'Introduction', in Richard Wall (ed.), Family Forms in Historic Europe
(Cambridge: Cambridge University Press, 1983).

LATIN AMERICAN HOUSEHOLDS IN COMPARATIVE PERSPECTIVE

503

from Mexico, Costa Rica, the Dominican Republic, is Panama, Colombia and Peru
gathered in the middle 1970s.
BACKGROUND

Downloaded by [Universidad Autonoma de Barcelona] at 08:52 06 November 2015

Factors underlying household organization in Latin America

F o u r major factors underlying the organization of households in Latin America are the
idealization of the patriarchal extended household, the idealization of the roles of men
and women, marital patterns (high ages at marriage for both men and women, the
c o m m o n custom of consensual union and high marital instability) and c o m m o n ruralurban migration a m o n g young women. The first three factors are rooted in the nature
of the Spanish conquest in Latin America, while the fourth has been linked to Latin
America's intermediate level of development.
The Spaniards brought with them from traditional pre-industrial Southern Europe
the notion that an older male should preside over a family that extended beyond his own
nuclear unit to include married sons and other kin. 1~'~ The older male was supposed to
head an economic unit of production and consumption, and he was supposed to have
authority over the life and death of his wife, children and grandchildren. Although this
ideal was shared by all social classes, it seems to have been actualized most closely
throughout Latin America by the landed gentry of European ancestry.
Related to the idea of the patriarchal family was a set of ideas about the ideal roles
of men and women. The ideal m a n was supposed to be forceful, daring and virile
whereas the ideal woman was supposed to be submissive and oriented towards her
family. Whereas the virile man could engage in sexual 'exploits' outside marriage, a
w o m a n was expected to be chaste before and faithful within marriage. 21 The ' m a c h o '
idea is consistent with the formation of casual sexual unions, whether or not a man could
or would support a family.
A third factor underlying household organization in Latin America is the region's
distinctive marital pattern. As in Western Europe, ages at first marriage are high for both
men and women, about 22 years for women and 26 years for men. ~2'~3 (Consensual
union is considered to be a type of marriage for this purpose.) Unlike in Western Europe,
however, consensual union instead of civil or religious marriage is quite common. 24
M a n y consensual unions are stable, especially after children have been born; the spouses
simply wanted to avoid paying the relatively high cost o f a wedding. They m a y live with
is I include the Dominican Republic in Latin America, although others might consider it part of the
Caribbean. My reason is that it shares a Spanish heritage with other countries in Latin America, unlike many
countries in the Caribbean.
~9 Howard I. Blutstein, J. David Edwards, Kathryn T. Johnston, David S. Morrise and James D. Rudolph,
Colombia: A Country Study (Washington, D.C. : U.S. Government Printing Office, 1983).
zv Thomas E Well, Jan Knippers Black, Howard I. Blutstein, Kathryn T. Johnston and David S.
McMorris, Mexico: A Country Study (Washington, D.C. : U.S. Government Printing Office, 1982).
31 Richard F. Nyrop (ed.), Panama: A Country Study (Washington, D.C.: U.S. Government Printing
Office, 1980).
32 Mohammed Kabir, 'The demographic characteristics of household populations', WFS Comparative
Studies, No. 6, 1980.
23 Jacob S. Siegel, 'El hogar y la familia en la formulacion de programas de vivenda', Estadistica, June
(1963). Reprinted in Thomas Burch et al. (eds.), La familia eomo unidad de estudio demografico (San Jose,
Costa Rica: Centro Latinoamericano de Demografia (CELADE), 1963).
24 See also Susan De Vos, 'Using world fertility survey data to study household composition: Latin
America', CDE Working paper 85-22, University of Wisconsin, Madison: Center for Demography and
Ecology, 1985.
17

L P S 41

504

SUSAN DE VOS

Table 1. Singulate mean age at marriage ( S M A M ) for women and men, and percentage o f
women aged 35-64 who are married or divorced/separated in six Latin American and other
selected countries, by age
Per cent
divorced/separated

Per cent married

Downloaded by [Universidad Autonoma de Barcelona] at 08:52 06 November 2015

SMAM*
F/M
35-44t
Unweighted average
for six countries:~
Colombia
Costa Rica
Dominican Republic
Mexico
Panama
Peru
United States 1975
Netherlands 1975
Spain 1978
Poland 1974
Japan 1975

45-54

55~4

35-44

45-54

55~4

--

77

68

53

12

12

22/26
22/26
20/25
22/24
22/26
23/26

74
74
75
82
77
81

62
68
64
73
68
75

47
56
49
57
49
58

5
9
16
6
13
7

8
10
21
8
19
7

7
10
23
7
19
8

------

80
89
87
87
90

78
83
81
79
82

68
72
68
62
65

13
4
1
5
3

10
3
1
4
4

8
3
1
3
3

Source: World Fertility'Survey household files for the six Latin American countries gathered in the middle
1970s and United Nations 1982 Demographic Yearbook, Table 40.
* S M A M for women and m e n refers to the singulate mean age at marriage. S M A M is calculated from a
cross-sectional distribution o f marital status, by age. ' T h e S M A M is computationally straightforward, but
stringent assumptions are required, namely, that there are no differentials by marital status in mortality and
migration and, especially, that the nuptiality patterns have not changed. W h e n these assumptions are met, the
cross-sectional percentages single can be taken to represent the experience of actual cohorts.' The figures are
from Kabir, loc. cir. p. 47, in footnote 22.
~" Ages.
~: Figures for Colombia and Peru are based on weighted counts.

one set of parents until children are born, and establish their own household
thereafter. 25.26
Other consensual unions are not stable, however, and marital instability in Latin
America is relatively high (see Table 1). For instance, an average of 77 and 68 per cent
of women aged 35~14 and 45-54 respectively were currently married in the six Latin
American countries, compared with between 85 and 90 per cent and 81-85 per cent in
the Netherlands, Spain, Poland and Japan. A relatively high proportion of women in
these age groups were divorced or separated. Although the ideal may be to marry, it is
not uncommon in the lower classes for households to be matrifocal, and children to be
'fathered in a series of free unions in which men move in temporarily with the
mother'. 27-~9 Such instability could result in a high incidence of households headed by
women or in a high incidence of extended households that would be quite different from
the idealized patriarchal extended household.
A fourth factor influencing household organization in much of Latin America is the
high rate of rural-urban migration. 3 One consequence of this migration is that wives
~5 Carmen Diana Deere, ' T h e differentiation of the peasantry and family structure: a Peruvian case study',

Journal o f Family History, 3 (1978), pp. 422--438.


2e Carl Kendall, 'Female-headed households and domestic organization in San Isidro, Guatemala: a test
of Hammel and Laslett's comparative typology', Journal of Comparative Family Studies, 9 (1978), pp. 129141.
27 Blutstein et al., op. cit. p. 118, in footnote 19.
2a See also Weil et aL, op. cit. in footnote 20.
29 Winifred Weekes-Vagliani, ' D o m i n i c a n Republic', pp. 291-327, in Women in Development (Paris:
Organization for Economic Co-operation and Development Centre Studies, 1980).
30 Carmen A. Miro and Joseph E. Potter, Population Policy: Research Priorities in the Developing Worm
(London : Frances Pinter Publishers, 1980).

LATIN AMERICAN HOUSEHOLDS IN COMPARATIVE PERSPECTIVE

505

and children m a y be left behind in rural areas as husbands seek work in urban areas. 31
Another consequence is that young unmarried women m a y migrate to cities in search
of employment because in rural areas they may have 'few opportunities of recreation
and almost no possibility of w o r k ' . 3~ Sometimes, young unmarried rural women are sent
to the city by a family that expects remittances from her wages. 33 It is c o m m o n for these
young unmarried migrants to become domestic servants in urban households.

Downloaded by [Universidad Autonoma de Barcelona] at 08:52 06 November 2015

C o m m o n a l i t y w i t h & diversity

The contrast between Latin America and elsewhere is the main focus of this paper, but
it is also important to acknowledge the cultural and socio-economic diversity between
different countries in Latin America. Each country considered in this study - Mexico,
Costa Rica, the Dominican Republic, Panama, Colombia and Peru - has been subjected
to a different mixture of Indian, Afro-American and European influences and populations. F o r instance, Peru still contains a significant proportion of Indians in its population while in the Dominican Republic there are no Indians but m a n y mulattoes. 34,3~
There are also significant Afro-American populations in P a n a m a and Colombia. 36,37
Mestizos predominate in Mexico, Costa Rica, Panama, Colombia and Peru. 38
Probably the best illustration of the idea of simultaneous commonality and diversity
is given by the composition of the adult population by marital status (see Table 1).
Compared with such countries as the Netherlands, Spain and Poland, the proportion of
women aged 35-64 who are separated or divorced is high in all six Latin American
countries. At the same time, the difference between different countries is large.
Colombia, Mexico and Peru tend to be at the lower end of a range with a minimum of
around six per cent, while the Dominican Republic and Panama tend to be at the upper
end with 13 per cent or more.
Another example is the gross national product per head (GNP). In all six countries
G N P was below the world average of U.S. $2,754 in 1981, but exceeded that for the 'less
developed' world of $728. At the same time, the G N P in Peru was half that in Mexico,
$1,122 compared with $2,250. 39 The countries have been classed as intermediate in the
stage reached in the demographic transition, except for Peru which has been classed
a m o n g the least advanced. 4
Our sample does not include the more socio-economically advanced countries of
Latin America such as Brazil, Argentina or Chile. However, such cultural factors as
marital patterns do not appear to be associated with economic indicators, implying that
the six countries can help to represent the diversity of Latin America. For instance,
although Mexico and Peru are at opposite ends of the range of gross national product
per head, the proportions of women aged 35-64 years who are divorced or separated,
31 Deere, loc. cit. in footnote 25.
39 Douglas Butterworth and John K. Chance, Latin American Urbanization (Cambridge: Cambridge
University Press, 1981), paraphrasing M. Margulis, Migracion y Marginalidad en la Sociedad Argentina
(Buenos Aires, 1968).
33 Butterworth and Chance, op. cit. p. 57, in footnote 32.
34 Richard F. Nyrop (ed.), Peru: A Country Study (Washinton, D.C.: Government Printing Office,
1980).
35 Thomas E. Weil, Jan Knippers Black, Howard I. Blutstein, Kathryn T. Johnston, David S. McMorris
and Frederick P. Munson, Area Handbook for the Dominican Republic (Washinton, D.C. : U.S. Government
Printing Office, 1973).
39 Blutstein et aL, op. cit. in footnote 19.
37 Nyrop op. cit. in footnote 21.
33 See also Weil et al., op. cit. in footnote 20.
39 Population Reference Bureau. Data Sheet 1983 (New York: PRB, 1983).
40 Miro and Potter, op. cit. p. 45, in footnote 30.

506

S U S A N DE VOS

are similar in both (Table 1). Although GNP per head is similar in Colombia and the
Dominican Republic they are at opposite ends of the range of the proportion of women
aged 35-64 years who are divorced or separated. Also, the singulate mean ages at first
marriage appear similar in all the countries (Table 1).

Downloaded by [Universidad Autonoma de Barcelona] at 08:52 06 November 2015

DATA

Since data on household composition are scarce in censuses in Latin American


countries, information for this study comes from six World Fertility Survey household
samples gathered in the middle 1970s (1975 to 1977). The de jure samples consisted of
9,647; 4,235; 10,685; 12,945; 4,724; and 7,204 households containing 54,846; 23,024;
57,008; 72,694; 23,177; and 39,531 individuals of all ages and both sexes in Colombia,
Costa Rica, the Dominican Republic, Mexico, Panama and Peru, respectively.
Data for different countries were recoded into a relatively standard format. In this
process, original information on relationship to household head was converted into a
standard variable referring to the individual's generation in the household relative to the
head (e.g. whether of the parental, head's or child's generation). Though generally
beneficial, this standardization resulted in the loss of some information regarding
extended kin. For instance, no differentiation was made between children, nieces and
nephews, as all would be coded as belonging to the child's generation. Hence, I had to
assume that an individual in the child's generation was, in fact, the child of the
household head, that an individual in the parent's generation was actually the parent
(in-law) of the head and so forth, instead of a more distant relation. Even with this
assumption, I shall show that extended family households are numerous in the six Latin
American countries.
Lodgers are not considered part of the same household if they do not take their meals
with the people with whom they reside. 41'4~ Thus age-specific headship rates based on
these data are higher than they would be if lodgers were considered part of the same
household; the proportion of people who are recorded as unrelated to the household
head is relatively lower; and the proportions in solitary or 'no family' households are
relatively greater. More information on the data can be found in De Vos 43 and
Kabir. 44
HOUSEHOLD

COMPLEXITY

One of the features thought to distinguish a Western household system from systems
elsewhere was the relatively low level of household complexity in the West. 4~ Most
households contained a simple family of parent(s) and child(ren). How does Latin
America fit into a continuum of complexity ?

4x World Fertility Survey,' Interviewers' Instructions', Basic Documentation, no. 6 (Voorburg, Netherlands:
International Statistical Institute, 1975).
~ ' M e m b e r s of a household live together and eat together. Hence a household is not necessarily a dwelling
or a family (though in m a n y cases it is). For example, servants or friends living with the family are members
of the household but m a y not be family members. Also, especially in urban areas, there m a y often be more
than one household in a single dwelling' (World Fertility Survey, loc. cit. p. 11, in footnote 41).
4a De Vos, loc. cir. in footnote 24.
44 Kabir, loc. cit. in footnote 22.
45 Laslett, op. cit. in footnote 9.

L A T I N A M E R I C A N H O U S E H O L D S IN C O M P A R A T I V E P E R S P E C T I V E

507

Downloaded by [Universidad Autonoma de Barcelona] at 08:52 06 November 2015

The ratio of adults per household


Since direct information on household complexity is unavailable for most populations,
the adult per household (A/H) ratio is often used to compare household complexity
around the world. This index only requires information on the number of households in
a population and the number of adults. Preferable to an index of mean household
size4e (see footnote 42), it is predicated on the idea that the least complex households
tend to consist of a solitary adult, whereas extended households usually contain three or
more adults. Although the A / H ratio ignores the average number of children per
household, it is sensitive to the age composition of the adult population. Thus
comparisons are best undertaken with age-standardized ratios. 47 1 have standardized by
using two ages for the beginning of adult status, 15 years and 25 years. Fifteen years is
the age commonly used, but an older age would seem reasonable in the Latin American
populations given the relatively late age at first marriage. 4s
Table 2 contains information on the average household size, the crude A / H ratio
(using the 15+ definition for adults), and two age-standardized A / H ratios
corresponding to different beginning ages for adults for the six Latin American
countries, and several other countries for comparative purposes. The average household
size and the crude A / H ratios in the six Latin American countries were larger than in
either the United States or Ireland. For instance, the mean household size was around
5.4 compared to 3.2 in the United States and 4.1 in Ireland. The crude A / H ratio was
around 3.0 compared to 2.3 and 2.8. When standardized for age however, the A / H
ratios of the six Latin American countries were intermediate between those of countries
of low and high complexity. On one side were the Netherlands and the United States
with (15+) A / H ratios of 2.5 and 2.6, on the other Japan and Ireland with relatively
high A / H ratios of 3.1 and 3.4. It is known that stem household arrangements are
common in these two countries. 49'50 Ratios in the six Latin American countries ranged
from 2.7 to 3.0. As a check, in Puerto Rico a ratio of 2.7 was found. When the adult
population is defined as consisting of individuals 25 years old and older, the differences
are reduced but the same pattern emerges.
4~ Early attempts at comparison used mean household size to indicate complexity, reasoning that larger
households tend to be more complex. This proved to be an unsatisfactory technique for comparison because
there were relatively small differences between average household sizes in very different types of societies (e.g.
Laslett, loc. cit. in footnote 6) and differences in household size are determined primarily by fertility, not the
level of complexity. (See also Thomas K. Burch, 'Some demographic determinants of average household size:
an analytic approach', Demography, 7 (1970), pp. 61-69.
~7 Thomas K. Burch, 'The index of overall headship: a simple measure of household complexity
standardized for age and sex', Demography, 17 (1980), pp. 25-37. In addition, Burch has developed an overall
index of headship which indirectly standardizes for age composition, thus widening the availability of a
summary measure to countries with few household data. I do not use the measure in this paper, because I can
directly standardize the A / H ratio with the available age-specific household headship rates.
4s The ratio can also be standardized for the marital distribution of the adult population where the
necessary data exist (Thomas K. Burch, Shiva S. Halli, Ashok Madan, Kauser Thomas and Lokky Wai,
'Measure of household composition and headship based on aggregate, routine census data', in John
Bongaarts, Thomas K. Burch and Kenneth W. Wachter (eds.), Family Demography: Methods and Their
Applications (Oxford: Oxford University Press, 1987). Unfortunately, the necessary data are rarely available.
40 Michael Gordon, Brendan Whelan and Richard Vaughan, 'Old age and loss of household headship: a
national Irish study', Journal of Marriage and the Family, 43 (1981), pp. 741-747.
50 S. Philip Morgan and Kiyosi Hirosima, 'The persistence of extended family residence in Japan',
American Sociological Review, 48 (1983), pp. 269-281.

508

SUSAN DE VOS

T a b l e 2. Mean household size (MHHS), crude ratio of adults per household ( A / H ) and

standardized ratios of adults per household (St. A / H ) for six Latin American and for
selected other countries in comparative perspective
St. A/HI"

Downloaded by [Universidad Autonoma de Barcelona] at 08:52 06 November 2015

Area
Netherlands (1970)
United States (1970)
Puerto Rico (1970)
Japan (1970)
Ireland (1971)
Six Latin-American countries,
unweighted average:~
Colombia
Costa Rica
Dominican Republic
Mexico
Panama
Peru

Crude*
MHHS* A/H
(15 +)

(25 +)

-3.2
--4.1
5.4

-2.3
--2.8
3.0

2.6
2.5
2.7
3.1
3.4
2.8

1.9
1.9
2.0
2.3
2.5
2.1

5.6
5.4
5.3
5.6
4.9
5.4

3.2
3.2
2.9
3.0
2.8
3.1

2.9
2.9
2.7
2.8
2.8
3.0

2.1
2.1
2.0
2.1
2.0
2.1

* Data for mean household size and the crude ratio of adults per household for the United States and
Ireland come from a communication from Thomas K. Burch, Department of Sociology, University of Western
Ontario, Canada.
t Data for the standardization of A/H ratios for the Netherlands, the United States, Puerto Rico, Ireland
and Japan come from the age-specifichousehold headship rates published by the United Nations, loc. cit. in
footnote 58; and from the stable population age distribution with a gross reproduction index of 2.5 and life
expectancy at birth of 50 years, published in Bureh, loc. cit. in footnote 46.
J/ Data for the six Latin American countries come from the household samples of the World Fertility
Survey.
Figures for Colombia and Peru are based on weighted counts.

The distribution of households by type


A l t h o u g h less c o m m o n l y available for most areas of the world, it is necessary to a u g m e n t
i n f o r m a t i o n o n A / H ratios with distributions that show the p r o p o r t i o n s o f household
by type, because the A / H ratio is only a n indirect measure o f household complexity. I
use a typology o f household c o m p o s i t i o n suggested for historical a n d c o m p a r a t i v e
household studies by Eugene H a m m e l a n d Peter Laslett. ~a I n its most basic form, the
typology consists of five h o u s e h o l d categories: solitary, n o family, simple family,
extended family a n d multiple family. Solitary households consist o f only one individual,
whereas ' n o f a m i l y ' households c o n t a i n individuals who are n o t related to each other.
Centred a r o u n d the presence of conjugal units, simple families consist o f parent(s) a n d
children, whereas extended families consist of related individuals who do n o t all belong
to the same conjugal u n i t (see T a b l e 3). 5~ A special case o f the extended family
household, multiple family households, c o n t a i n s two or more conjugal units, s3
The basic typology can be sub-divided further when there are e n o u g h cases. H a m m e l
61 EugeneA. Hammel and Peter Laslett, 'Comparing household structure over time and between cultures',

Comparative Studies in Society and History, 16 (1974), pp. 73-109.


53 There seems to be confusion over the categorization of households that contain relatives but no conjugal
unit, for instance, two unmarried siblings. While such households are sometimes considered 'no family'
households, they should be considered 'extended' households, because the household members are kin.
Furthermore, their relationship is no more distant than that between a grandparent and grandchild. In this
study, households composed of such elements without a conjugal couple are always considered extended.
58 The basic typology does not consider the existence of household members who are unrelated to the
household head such as servants. Since the pattern in Latin America proves to be quite different from that of
pre-industrial Western Europe or of India or China, I augment the basic typology by considering whether
households contain an unrelated member.

LATIN

AMERICAN

HOUSEHOLDS

IN COMPARATIVE

PERSPECTIVE

509

Table 3. Comparing household distributions from a number of historical samples and six
Latin American countries in the middle 1970s (percentages)
Simple

Downloaded by [Universidad Autonoma de Barcelona] at 08:52 06 November 2015

Sample

No
family

Solitary

Bristol
(Rhode Island)
Ealing (1599)
(England)
Ealing (1861)
Longuenesse
(France 1778)
Belgrade
(Serbia 1733/4)
Bertalia (Italy 1880)
Nishinomiya
(Japan 1713)
Yokouchi
(Japan 1846)
Colombiat
Costa Rica
Dominican Republic
Mexico
Panama
Peru1"

Complex

Husband
and
Single
wife
parent

' Extended' ' Multiple'


(no or one (two or more
couple)*
couples)
Total

Sample
size

89

(2)

(0)

(72)

12

65

13

(6)

(2)

(85)

.
1

.
l

61

15

(19)
(19)

(2)
(3)

21
22

(209)
(66)

59

(15)

(14)

29

(273)

2
7

0
0

54
32

8
11

(17)
(29)

(16)
(21)

33
50

(347)
(132)

(14)

(24)

39

(107)

(27)
(24)
(34)
(19)
(28)
(26)

(3)
(4)
(2)
(6)
(4)
(6)

30
28
35
24
31
31

(9,647)
(4,235)
(10,685)
(12,945)
(4,725)
(7,204)

.
5
4
8
4
9
7

.
1.0
0.9
0.9
0.6
1.0
0.5

.
55
58
48
64
50
54

9
9
8
7
9
8

Sources: Laslett, op. cit. pp. 61, 85, in footnote 6; Kertzer (David I. Kertzer, Family Life in Central Italy,
1880-1910: Sharecropping, Wage Labor, and Coresidence (Rutgers University Press, New Brunswick, New
Jersey, 1984, p. 63); World Fertility Survey household samples.
* The' no-family' category does not include households with siblings, as in Laslett's original typology. Such
households are in the extended household category.
t Figures for Colombia and Peru are based on weighted counts.

and Laslett s4 suggest 19 possible sub-divisions, but different researchers tend to


emphasize different sub-categories. In this study, simple family households consist of:
(a) husband-wife households with or without children, and (b) single-parent households.
Extended households are divided into (a) 'special' households containing relatives but
without a conjugal couple, 5s (b) extended family households with one couple (e.g. a
nuclear family with an unmarried elderly parent). The purpose of this latter sub-division
is to discern the incidence of co-residence among non-married kin which might otherwise
be overlooked if the primacy of the nuclear family or the conjugal unit were
emphasized. 5~
Consistent with the proportion indicated by the A / H ratios, the proportion of
complex households in the six countries of Latin America was intermediate between that
found in historical samples for the West and Japan (Column 5 of Table 3). Whereas the
proportion of complex households did not exceed 21 per cent of the households in the
pre-industrial North American or Western European samples and never dropped below
39 per cent in the two Japanese samples, the proportion in the six Latin American
countries ranged between 25 and 36 per cent. This is roughly the range for the samples
from Southern and Eastern Europe (Bertalia, Italy and Belgrade, Serbia).
54 Hammel and Laslett, loc. cit. p. 96, in footnote 51.
5s These 'special' extended households are sometimes considered 'no family' households.
5e See also De Vos, loc. tit. in footnote 24.

510

SUSAN DE VOS

T a b l e 4. A decomposition of extended households* by whether extended laterally, vertically

or a combination of both: several historical samples and six Latin American countries in
the middle 1970s
Type of extension (per cent)

Downloaded by [Universidad Autonoma de Barcelona] at 08:52 06 November 2015

Sample
Ealing
(England)
Longuenesse
(France)
Belgrade
(Serbia)
Nishinomiya
(Japan)
Bristol
(American colonial)
Colombiat
Costa Rica
Dominican Republic
Mexico
Panama
Perut

Total

Per cent
of all
households

Lateral

Vertical

Combination

(29)

(57)

(14)

(7)

(27)

(64)

(9)

(11)

17

38

58

79

29

19

76

63

48

(0)

(100)

(0)

(2)

25
20
17
17
17
16

43
44
42
41
44
42

32
36
41
42
39
42

2,960
1,205
3,756
3,176
1,486
2,264

31
28
35
24
31
31

Source: Laslett, op. cit. in footnote 6, p. 85, Table 1.15, World Fertility Household Samples for Mexico,
Costa Rica, Dominican Republic, Panama, Colombia and Peru.
* Extended households are those in which kin who are no longer members of the same conjugal unit coreside. In Latin America, this includes 'no family' households consisting of siblings or other relatives. In
laterally extended households there are only one or two (proximate) generations in which the extra kin are of
the same generation(s) as the members of the central (nuclear) unit. Vertically extended households contain
three or more generations of kin, or two non-proximate generations, but no lateral extensions. Households
with both lateral and vertical extensions are considered to contain a 'combination' of those extensions - e.g.
a household with a widowed sister of the head and a widowed mother of the head.
t Figures for Colombia and Peru are based on weighted counts.

T h e r e a s o n for the a m o u n t o f h o u s e h o l d extension in L a t i n A m e r i c a , c o m p a r e d to


S o u t h e r n a n d E a s t e r n E u r o p e o r J a p a n , a p p e a r s to be different however. U n l i k e the
o t h e r samples, for instance, the p r o p o r t i o n o f multiple family h o u s e h o l d s with two o r
m o r e conjugal couples was relatively small in the L a t i n A m e r i c a n samples (Table 3). I n
a d d i t i o n , a l t h o u g h this is less firmly b a s e d empirically, h o u s e h o l d extension in L a t i n
A m e r i c a a p p e a r s to have involved m o r e c o m b i n a t i o n s o f vertical a n d lateral extension.
I a m c o n s t r a i n e d by the r a r i t y o f d e s c r i p t i o n s o f h o u s e h o l d c o m p o s i t i o n o f this type to
c o n s i d e r again the five h o u s e h o l d samples c o n t a i n e d in Laslett's 1972 s t u d y s7 (see T a b l e
4). In c o n t r a s t to the other unfortuiaately small samples, lower p r o p o r t i o n s o f the
c o m p l e x h o u s e h o l d s in the six L a t i n A m e r i c a n c o u n t r i e s were e x t e n d e d o n l y vertically,
a n d larger p r o p o r t i o n s were e x t e n d e d b o t h laterally a n d vertically, t h a n a p p e a r s to have
been the case in the o t h e r samples. A s u b s t a n t i a l p r o p o r t i o n , r a n g i n g f r o m one-sixth to
o n e - f o u r t h , o f the c o m p l e x h o u s e h o l d s in the L a t i n countries were laterally e x t e n d e d
only.
The reasons for the difference in h o u s e h o l d c o m p l e x i t y between L a t i n A m e r i c a a n d
elsewhere a p p e a r s to be a c o m b i n a t i o n o f two p a t t e r n s : (1) the relative i n d e p e n d e n c e
o f c o n j u g a l units t o g e t h e r with the tendency for conjugal couples to extend their
h o u s e h o l d s by including u n m a r r i e d relatives, a n d (2) the tendency for m a n y h o u s e h o l d s
h e a d e d b y w o m e n to be e x t e n d e d as well. E a c h o f these issues will be c o n s i d e r e d in
turn.
57 Laslett, op. cit. in footnote 6.

LATIN AMERICAN HOUSEHOLDS IN COMPARATIVE PERSPECTIVE

511

Downloaded by [Universidad Autonoma de Barcelona] at 08:52 06 November 2015

MEN'S HEADSHIP AND THE SEPARATE RESIDENCE OF CONJUGAL COUPLES


Like W e s t e r n h o u s e h o l d s , h o u s e h o l d s in L a t i n A m e r i c a rarely c o n t a i n m o r e t h a n one
c o n j u g a l couple. T h e m o s t c o m m o n exceptions to this rule occur when the c o u p l e are
y o u n g a n d d o n o t yet have a n y children, o r when they are old. Even u n d e r these
circumstances however, couples t e n d to live in their o w n h o u s e h o l d . T h e clearest
i n d i c a t o r o f this t h a t can be c o m p a r e d with o t h e r countries with the a p p r o p r i a t e d a t a
is the age-specific h e a d s h i p rate o f men. T h e ' r a t e ' shown in T a b l e 5, refers to the
p r o p o r t i o n o f m e n in an age g r o u p w h o h e a d their o w n h o u s e h o l d . (Recall t h a t lodgers
are n o t c o n s i d e r e d h o u s e h o l d m e m b e r s if they d o n o t t a k e meals with the others. T h u s
h e a d s h i p rates for L a t i n A m e r i c a n countries are higher t h a n if lodgers were c o n s i d e r e d
m e m b e r s o f the h o u s e h o l d . )
T a b l e 5. M e n ' s age-specific headship rates in comparative perspective
Age...
Unweighted sample average*
High-income averaget
Low-income averaget
Colombia (1976)~:
Costa Rica (1976)
Dominican Republic (1975)
Mexico (1976/7)
Panama (1976)
Peru (1977):~
Puerto Rico (1970)
Netherlands (1970)
United States (1970)
Spain (1970)
Japan (1970)
Ireland (1971)

15-24

25-34

35-44

45-54

55-64

65 +

11
12
14
10
10
11
14
10
9
14
18
21
3
11
6

65
71
59
64
65
64
71
62
62
72
82
84
59
62
52

86
88
81
85
84
87
89
84
88
87
92
92
83
83
73

92
92
86
90
92
90
93
89
94
90
95
93
90
91
81

91
92
87
89
88
91
93
90
92
91
95
94
91
92
84

83
82
79
80
80
84
86
85
84
85
86
87
79
68
75

* For the six Latin American countries.


t Dates for countries with high incomes and low incomes per head refer to around 1960, and are presented
in United Nations, op. cit. in footnote 58, p. 14, Table 2. Estimates for the four developed countries come from
United Nations, op. cit. in footnote 58, Tables 7 and 9.
:~ Rates for Colombia and Peru are based on weighted counts.
I n the first p a n e l o f T a b l e 5 we s h o w the u n w e i g h t e d average for the six L a t i n
A m e r i c a n c o u n t r i e s a n d a v e r a g e h e a d s h i p rates for high- a n d l o w - i n c o m e c o u n t r i e s
a r o u n d 1960 as r e p o r t e d by the U n i t e d N a t i o n s . 5a I n each case, a l t h o u g h the n u m b e r s
differ, h e a d s h i p rates are very low for m e n between the ages o f 15 a n d 24, b u t rise r a p i d l y
to a p e a k at 4 5 - 6 4 years o f age. T h e rates for m e n aged 65 years a n d o l d e r are s o m e w h a t
lower t h a n for those 5 5 - 6 4 y e a r s o f age. T h e rate for 15-24 y e a r o l d m e n is influenced
b y age at m a r r i a g e a n d achieving i n d e p e n d e n c e f r o m the p a r e n t a l h o m e , while t h a t for
m e n aged 65 years a n d o l d e r is affected b y retirement. 59 M o s t striking however, is t h a t
the average h e a d s h i p rates for L a t i n A m e r i c a a r e close to those for h i g h - i n c o m e
c o u n t r i e s in 1960.
T h e m i d d l e p a n e l o f T a b l e 5 c o n t a i n s age-specific h e a d s h i p rates for m e n in i n d i v i d u a l
L a t i n A m e r i c a n countries. Panel 3 o f T a b l e 5 c o n t a i n s h e a d s h i p rates in 1970 for low5s United Nations, 'Estimates and Projections of the Number of Households by Country, 1975-2000' (New
York: ESA/P/WF., U.N., 1981).
59 United Nations, 'Use of model headship rates', chapter vi in Methods of Projecting Households and
Families, 1973.

512

SUSAN

DE VOS

Table 6. Distribution of households headed by married men and formerly married women
for six Latin American countries in the middle 1970s (percentages)
Household type
Simple f a m i l y

Non-family*

Husband/
wife

Extended family

Single
parent

No
couple

One or more
couples

Sample
size

Downloaded by [Universidad Autonoma de Barcelona] at 08:52 06 November 2015

Headed by married men


Unweighted averaget

--

74

--

--

26

Colombia
Costa Rica
Dominican Republic
Mexico
Panama
Peru

-------

73
76
69
80
75
71

-------

-------

27
24
31
20
25
29

Unweighted averaget

14

--

42

36

Colombia
Costa Rica
Dominican Republic
Mexico
Panama
Peru

8
10
17
17
15
14

-------

47
49
33
43
36
43

37
33
44
30
43
31

8
8
6
10
6
13

7,971
3,229
7,362
10,294
3,109
5,389

Headed by formerly married women

1,679
726
2,217
1,535
847
926

* N o n - f a m i l y h o u s e h o l d s a r e ' s o l i t a r y ' o r ' n o f a m i l y ' h o u s e h o l d s . See text.


t Figures for Colombia and Peru are based on weighted counts.

complexity countries like the United States and the Netherlands, and for the relatively
high-complexity countries like Japan and Ireland (1971), and for Spain and Puerto Rico.
Again, men's headship rates in the Latin American countries tend to be closer to those
in the Netherlands and the United States than in Ireland or Japan (or Spain).
Although comparable headship rates are not available for the other countries, the
rates in Latin America are even higher when computed for married men only, indicating
that married men almost always head their own households rather than share a
household with another married man. In general, 93 per cent or more of the married men
over 35 years old headed their own households, comparable to the proportions reported
by Hajnalr0 for Western Europe in the past (figures not shown; see De Vosel). The
proportion, still high, tended to be only slightly lower for married men aged 25-34.
Headship rates for married men aged 15-24, came to only 72 or below, suggesting that
a minority of young couples reside in a parental household. Since the average marriage
age of men tended to exceed 25 (except in Mexico where it was 24.4), s2 early marriage
may be more liable to be associated with dependent living arrangements.
While married men are likely to head their own households, these are commonly
extended, rather than consisting of simple husband and wife only. One indicator, shown
in the first panel of Table 6, is the distribution of households by type headed by married
men. On average, one-quarter of these households are extended. Also, age-specific
headship rates of formerly-married men and women of around 60-70 per cent, are
clearly lower than those for married men, indicating that formerly-married adults often
60 H a j n a l , loc. cit. in f o o t n o t e 8.
el See also S u s a n D e V o s , ' L a t i n A m e r i c a n h o u s e h o l d s in c o m p a r a t i v e p e r s p e c t i v e ', C D E W o r k i n g P a p e r
8 5 - 1 6 , U n i v e r s i t y o f W i s c o n s i n , M a d i s o n : C e n t e r f o r D e m o g r a p h y a n d E c o l o g y , 1985.
e2 K a b i r , loc. cit. in f o o t n o t e 22.

LATIN AMERICAN HOUSEHOLDS IN COMPARATIVE PERSPECTIVE

513

j o i n a h o u s e h o l d h e a d e d b y a n o t h e r (figures n o t shown). ~s'e4 E q u a l l y i m p o r t a n t ,


however, is the fact t h a t m a n y o f the f o r m e r l y m a r r i e d also h e a d their own households.

Downloaded by [Universidad Autonoma de Barcelona] at 08:52 06 November 2015

HOUSEHOLDS HEADED BY WOMEN


In c o n t r a s t to the s i t u a t i o n in historical E u r o p e where h o u s e h o l d s h e a d e d by w o m e n are
either n o t m e n t i o n e d as a n i m p o r t a n t type e5,66 o r are e s t i m a t e d to be relatively
u n c o m m o n , 67 they are c o m m o n in the six c o u n t r i e s o f L a t i n A m e r i c a s a m p l e d here,
r a n g i n g f r o m 14 to 21 p e r cent o f all h o u s e h o l d s ( T a b l e 7). In a cross-section, these
h o u s e h o l d s c o n t a i n e d between seven a n d 15 p e r cent o f all children u n d e r 15 years o f
age. N o d o u b t , a n even greater p r o p o r t i o n o f children will have lived in a household
h e a d e d b y a w o m a n at s o m e p o i n t in their lives.
It is difficult to o b t a i n s a t i s f a c t o r y c o m p a r a t i v e d a t a on h o u s e h o l d s h e a d e d by w o m e n
b e c a u s e the c o m m o n l y used age-specific h e a d s h i p rate for w o m e n includes b o t h m a r r i e d
a n d u n m a r r i e d w o m e n , a l t h o u g h the f o r m e r are a l m o s t never c o n s i d e r e d household
heads. Still, a p e r u s a l o f such overall rates p r o v i d e s i m p o r t a n t i n f o r m a t i o n (see Table 8).
First, w o m e n ' s h e a d s h i p is higher on average in l o w - i n c o m e countries, as c o m p u t e d by
the U n i t e d N a t i o n s for 1960, a n d o n average in the L a t i n A m e r i c a n countries in the
m i d d l e 1970s, t h a n in the h i g h - i n c o m e c o u n t r i e s in 1960. A l t h o u g h higher levels o f
w i d o w h o o d p r o b a b l y c o n t r i b u t e d significantly to these figures for the low-income
c o u n t r i e s it r e m a i n s true t h a t m a n y w o m e n (between one-fifth a n d o n e - q u a r t e r o f those
aged 45-54) h e a d e d their o w n h o u s e h o l d s . Secondly, there was a definite age p a t t e r n to
h e a d s h i p t h a t was the same in m o s t regions o r countries, in spite o f differences in level:
the rates rise with age in c o n t r a s t to the curvilinear p a t t e r n for men, reflecting the
increase in d i s r u p t e d m a r r i a g e with age. 6s'69 Thirdly, the o n l y age g r o u p s in which the
T a b l e 7. Proportion o f households headed by a woman in the middle 1970s,
six Latin American countries
Country
Colombia*
Costa Rica
Dominican Republic
Mexico
Panama
Peru*

Per cent

Total no. of
households

17.4
17.1
20.8
13.6
20.2
14.4

9,647
4,235
10,683
12,945
4,725
7,204

Source: World Fertility Survey household samples.


* Figures for Colombia and Peru are based on weighted counts.
6a See De Vos, loc. cit. in footnote 61.
64 For any age group, the number of formerly married women greatly exceeded that of formerly married
men. Once this is taken into consideration however, there seems to have been little difference between the
headship rates of formerly married men and formerly married women under 35 years of age, or between the
ages of 45 and 54 years. Formerly married women aged 35-44 were more likely to head their own households
than formerly married men of the same ages, while the opposite was true for those 55 years old or older. This
may be related to differences in number of children by age, and to differences in who lives with children after
separation or divorce. For more on this, see De Vos, loc. cit. in footnote 61.
e5 See, for example, Hajnal, loc. cit. in footnote 8.
66 See, for example, Laslett, op. cit. in footnote 6.
67 See, for example, Richard Wall, 'Woman alone in English society', Annales de D~mographie Historique,
(1981), pp. 303-317.
6s United Nations, op. cit. chapter x, in footnote 59.
69 Japan is a notable exception to the general pattern, in that the rate for women aged 65 and older is
actually lower than that for the 55-64 year old group.

514

SUSAN DE VOS

Table 8. Women's age-specific headship rates for six Latin American countries in

comparative perspective

Downloaded by [Universidad Autonoma de Barcelona] at 08:52 06 November 2015

Age...

15-24

25-34

35-44

45-54

55-64

65 +

Unweighted sample average*


High-income average
Low-income average

1.5
3.2
2.5

6.5
5.9
8.6

12.8
8.6
16.1

21.5
15.5
24.7

29.7
25.3
32.6

36.3
39.3
37.6

Colombiat
Costa Rica
Dominican Republic
Mexico
Panama
Perut

1.5
1.5
2.5
0.8
1.3
1.1

5.7
6.0
9.8
3.9
8.8
4.5

13.4
12.2
17.3
9.0
15.7
9.0

23.3
20.2
27.8
16.5
24.5
16.6

28.6
26.5
37.2
25.9
33.5
26.3

31.4
35.0
47.3
33.3
42.3
28.3

Puerto Rico (1970)


Netherlands (1970)
United States (1970)
Spain (1970)
Japan (1970)
Ireland (1971)

2.1
4.7
5.0
0.5
3.9
1.9

8.2
5.8
11.4
1.6
4.3
4.0

14.4
7.0
12.9
3.7
7.9
5.9

19.2
12.3
16.4
10.0
15.9
13.8

25.4
23.7
26.1
20.1
18.5
28.4

35.8
43.6
42.2
31.1
13.2
37.3

Sources: 'Estimates and Projections of the Number of Households by Country, 1975-2000', United
Nations, op. cit., Tables 2, 7 and 9, in footnote 58; and World Fertility Survey household samples for Colombia,
Costa Rica, Dominican Republic, Mexico, Panama and Peru.
* For the six Latin American countries.
~" Figures for Colombia and Peru are based on weighted counts.

women's headship rates were higher for low-complexity, high-income countries, were
those for women 15-24 years of age and 65 years and above.
When age-specific headship rates are computed for formerly married women in the
Latin American countries (figures not shown), 7 it is clear that a majority of formerlymarried women aged 35 and older head their own households. Headship was highest for
the 45-54 year-old group, where the average proportion heading their own households
was 71 per cent. Although it might be conjectured that most of the households headed
by formerly married women would be single-parent households, many were in fact
extended-family households (Table 6). 71 Most of the extended households headed by
formerly married women did not contain a conjugal couple (Table 6).
I N D I V I D U A L S U N R E L A T E D TO THE H O U S E H O L D HEAD

Another dimension of household organization which is different for Western Europe


and the 'East', and where the pattern in Latin America is again different concerns the
prevalence of individuals living in households headed by a non-relative. There were
many servants, boarders and lodgers in pre-industrial Western Europe, but very few in
India or China. For example, HajnaF 2 has presented data from several rural areas in
Iceland (1729), Norway (1801), Flanders (1814) and England (1599-1796) to support the
assertion that 'servants were numerous, apparently always constituting at least six per
cent, and usually over ten per cent, of the total population', and that' almost all servants
were unmarried and most of them were young (usually between 10 and 30 years of age)'.
Furthermore, he calculated that there were 100 or more servants or other non-relatives
7o See De Vos, loc. cit. in footnote 61.
71 About one-third of the households headed by formerly married men were also extended, while non-family
and single-parent households made up the other two-thirds.
72 Hajnal, loc. cit. p. 471, in footnote 8.

Downloaded by [Universidad Autonoma de Barcelona] at 08:52 06 November 2015

LATIN AMERICAN HOUSEHOLDS IN COMPARATIVE PERSPECTIVE

515

p e r 100 h o u s e h o l d s in r u r a l a r e a s o f W e s t e r n E u r o p e in p r e - i n d u s t r i a l times c o m p a r e d
to o n l y six in r u r a l C h i n a o r India. 73
D o m e s t i c service is fairly c o m m o n in L a t i n A m e r i c a too, where this is t h o u g h t to
indicate an i n t e r m e d i a t e level o f d e v e l o p m e n t . 74 C o n t r a r y to the p a t t e r n in p r e - i n d u s t r i a l
W e s t e r n E u r o p e however, where m a l e ' s e r v a n t s ' were c o m m o n in r u r a l areas, d o m e s t i c
servants in L a t i n A m e r i c a tend o n l y to be y o u n g u n m a r r i e d w o m e n in u r b a n areas.
T h r e e i n d i c a t o r s o f this p a t t e r n are the p r o p o r t i o n o f all h o u s e h o l d s t h a t c o n t a i n an
u n r e l a t e d m e m b e r , the n u m b e r o f u n r e l a t e d i n d i v i d u a l s per 100 h o u s e h o l d s , a n d the
p r o p o r t i o n o f a n y age a n d sex g r o u p t h a t consists o f u n r e l a t e d individuals. 75
Between f o u r a n d twelve p e r cent o f the h o u s e h o l d s in the six L a t i n A m e r i c a n
c o u n t r i e s c o n t a i n e d at least one m e m b e r w h o was n o t related to the h o u s e h o l d h e a d (see
T a b l e 9). Since this is m o r e c o m m o n in u r b a n t h a n in rural areas, the p r o p o r t i o n is
higher if only the m a j o r m e t r o p o l i t a n a r e a o f the c o u n t r y is considered. F o r instance, 18
p e r cent o f the h o u s e h o l d s in Bogot/t, c o n t a i n e d a m e m b e r w h o was n o t related to the
h o u s e h o l d head. T h e lowest p r o p o r t i o n was eight p e r cent o f the h o u s e h o l d s in M e x i c o
City (Table 9).
H a j n a l ve f o u n d t h a t there were 104 u n r e l a t e d individuals per 100 h o u s e h o l d s in his
s a m p l e for rural D e n m a r k 1787-1801. In c o n t r a s t , there were o n l y six u n r e l a t e d
i n d i v i d u a l s p e r 100 h o u s e h o l d s in r u r a l I n d i a in 1951. T h e c o m p a r a b l e figures for all the
h o u s e h o l d s in the six samples f r o m L a t i n A m e r i c a r a n g e d f r o m 7 to 18. T h e figure climbs
to between 9 a n d 24 p e r 100 h o u s e h o l d s when the m a j o r m e t r o p o l i t a n a r e a s are

T a b l e 9. Percentage o f households with an unrelated individual, and number o f


unrelated individuals per 100 households
Percentage of households
with a member unrelated
to the head

Colombia
Costa Rica
Dominican Republic
Mexico
Panama
Peru
Rural Denmark
Rural India (1951):~

Number of unrelated
individuals per
100 households

Total

Capital
cityt

Total

Capital
cityt

12.1
9.3
10.2
4.5
6.4
7.2
---

18.1
16.1
15.2
8.2
10.6
12.4
---

18
7
15
7
9
11
---

24
9
23
13
15
15
104
6

Source: Data for rural Denmark come from Hajnal, loc. cit. in footnote 8, Table 1, in and refer to 26
parishes from 1787 to 1801.
Data for Latin America come from World Fertility Survey household samples. Data for Colombia and Peru
are based on weighted counts.
~" For Colombia, Bogot/t; Costa Rica, San Jos6; Dominican Republic, all urban areas; Mexico, Mexico
City; Panama, Panama City; Peru, Lima.
:~ Data for India come from Hajnal, loc. cit. in footnote 8, Table 3, and refer to the Census of India 1951,
vol. 1, India, Part la, Demographic Tables, table C.I(ii).
73 See also Wall, op. cit. in footnote 17.
74 Butterworth and Chance, op. cit. in footnote 32.
75 In the Dominican Republic, it was not possible to distinguish between the capital city of Santo Domingo
and other urban areas. Such a distinction was possible in the other countries.
7n Hajnal, loc. cit. in footnote 8.

Downloaded by [Universidad Autonoma de Barcelona] at 08:52 06 November 2015

516

S U S A N DE VOS

considered. 7v In Colombia and the Dominican Republic especially, the presence of nonrelated individuals in urban households appears important.
According to Hajnal's 78 European samples, the highest proportion of individuals
unrelated to the household head was among men aged 20-24 in a sample of Flemish
villages, at 48 per cent. In the Latin American samples, the highest proportion was found
for women aged 15-19 years in BogotA. Seventeen per cent of the women in this age
group were not related to the household head, compared to only two per cent of the men
in this age group (figures not shown). The proportions were even higher for unmarried
women, among whom fully 20 per cent at ages 20-24 years were unrelated to the
household head in Colombia, Costa Rica and Panama; proportions in other countries
were similar, ranging from 15 to 18 per cent.
SUMMARY AND DISCUSSION
Large gaps exist in our ability to compare the household and family systems in different
areas of the world, because of the scarcity of adequate data. This situation has led to
conflicting ideas about the probable impact of modernization or development on the
family. If the experience of Northwestern Europe could be used to predict the nature of
the family elsewhere, then models in which the nuclear household plays a dominant role
in social reproduction might be valid. Yet arguments based on recent historical evidence
from the West lead us to question the validity of applying such a model. Several social
historians have argued that the household formation system of pre-industrial
Northwestern Europe was distinct from others because of late ages at marriage, the
separate residence of conjugal couples and the circulation of young unmarried
individuals in another's household during periods of service. Latin America is rarely
included in a debate that usually focuses on differences between Western families and
those of the East - i.e. of India or China.
The omission of Latin America is serious, because more than in any other part of the
developing world, Latin American societies have been influenced deeply by European
customs. In this paper we have tried to use several dimensions of household organization
to place Latin American households in a comparative perspective.
A major finding was that Latin American populations share with pre-industrial
Europe the custom of a late age at marriage, even when consensual union is considered
as a form of marriage. Also like pre-industrial Western Europe, there was rarely more
than one conjugal couple in a household. The most common exceptions occurred among
married couples who were either very young or old.
Another major finding, that distinguishes Latin American households, is that their
complexity was intermediate between that of Northwestern Europe and Japan. This was
indicated both indirectly in terms of the average number of adults per household (Table
2) and directly by the proportion of households with members of more than one conjugal
unit (Table 3). There appeared to be two reasons for this: the tendency for conjugal
couples to accept unmarried relatives into their households, and the tendency for many
households headed by women to be extended (Table 6).
A second distinction found between the household organization of pre-industrial
77 Note that an individual is considereda householdmember only if he/she habitually takes meals with coresidents. Otherwise, that person is considereda separate household.
78 Hajnal, loc. cit. in footnote 8.
70 Susan De Vos and Alberto PaUoni, 'Formal models and methods for the analysis of kinship and
household organization', CDE Working Paper 84-30, University of Wisconsin, Madison: Center for
Demography and Ecology, 1984.

Downloaded by [Universidad Autonoma de Barcelona] at 08:52 06 November 2015

LATIN AMERICAN HOUSEHOLDS IN COMPARATIVE PERSPECTIVE

517

Northwestern Europe and present-day Latin America is the low proportion of


households headed by women in the former area compared to their numerical
importance in the latter. Between 14 and 21 per cent of the households in the sample of
six Latin American countries were headed by women. This appears to be one result of
the relatively high marital instability in Latin America, compared to the West in the past,
although households headed by women are becoming increasingly common to-day in
countries like the United States (Table 1).
A third distinction between household composition in Latin America and the preindustrial West concerns the existence of young people who were not related to the
household head. In the West, young unmarried adults of both sexes often became
servants in others' households before they themselves married and formed a household
of their own. In Latin America, in contrast, there appears to be a concentration of
unrelated individuals only among young unmarried women in urban areas. Men, or
people in rural areas do not commonly live in households with non-related persons.
The implication of these findings is that the nuclear-family household is not the best
focus for theories of social reproduction in Latin America, because the extended family
household is important there, whether or not it contains a husband/wife nucleus. The
findings also imply that there is a Latin American household form that needs to be
distinguished from the rather simplistic dichotomy of households into a 'Western' or
'Eastern' form. These conclusions must remain tentative for a couple of reasons.
First, observations about household composition can only lead to indirect inferences
about rules of household formation because there is no direct correspondence between
the observed configurations and such rules. Rather, household composition at any point
in time is the product of socio-economic conditions, the availability of kin with whom
to co-reside, and only thirdly, household formation rules. TM For example, living in an
extended family household may only be a survival strategy of the poor that would not
be preferred if people could afford separate residences at a certain standard of living, s
Secondly, the countries for which data are reported here are in a state of rapid
political, economic and social transition. No doubt, the nature of the family and
household is changing too, so that this cross-sectional view relates only to a point in a
dynamic situation. We cannot present a picture of change, but only infer that theories
of change based on an incorrect view of the household are bound to be flawed.

s0 See also Michael Anderson, 'Family, household and the industrial revolution', Chapter 6, in Michael
Anderson (ed.), Sociology of the Family (London : Penguin Books, 1971).

You might also like