WOMEN IN MIGRATION:
K... -
r;
PREPARED BY
: INTERNATIONAL CENTER FOR RESEARCH ON WOMEN
1010 16TH STREET N.W., WASHINGTON, D.C. 20036
FOR: OFFICE OF.WOMEN IN DEVELOPMENT -
; ''AGENCY FOR INTERNATIONAL DEVELOPMENT
WASHINGTON' D.C. 20523
:GRANT No. AID/OTR-G-1592 ?
JUNE 1979 :
A THIRD WORLD FOCUS,,
*l: ,- .-,. '' *''
::':
i' i
-
.n~
i':.
:' ''
:
r
...
-.,i ;. ~, --.~
-,
.i
; -
;-
II~..
~-~
r
:;
-1
:`;;": '.
r
r
II
'-
41r
.--.-
.i
? i '- .~
I-I- ...-
''
'i 1
.-;
-~
.. :
:1
-;--
--
i
II :
. .- -:=-
--
il..--
r
:-
.,..
;.' 5---i
-~ '-;~
"
I r
- .. -
1
r
c:~ -i`'.. :.;
;
~--1
,. -,
-
WOMEN IN MIGRATION:
A THIRD WORLD FOCUS
by,
International Center for Research on Women*
* Nadia Youssef, Mayra Buvinic, and Ayse Kudat (coordinating authors)
with Jennifer Sebstad and Barbara Von Elm (contributing authors)
The views and interpretations in this publication are those of the
author and should not be attributed to the Agency for International
Development.
AID/otr-G-1592
TABLE OF CONTENTS
PAGE
I. INTRODUCTION 1
II METHODOLOGY
a. Data Sources 7
b. An Index of Sex Differences in Migration Trends 10
III WHERE ARE THEY?: MIGRATION PATTERNS
a. Results: Sex Differences in Regional Migration
Trends 16
b. International Migration 42
IV WHO ARE THEY?: CHARACTERISTICS OF WOMEN MIGRANTS 67
V WHY DO THEY MIGRATE?: FACTORS EXPLAINING THE
MIGRATION OF WOMEN 82
VI ECONOMICS OF MIGRATION
a. What is the Economic Situation of Women Migrants? 98
b. What is the Impact of Male Migration Upon Rural Women? 115
VII IMPACTS OF MIGRATION ON FAMILY STRUCTURE 123
VIII IMPLICATIONS FOR POLICY 143
TABLES and FIGURES
BIBLIOGRAPHY
ANNEX A: Participants of the Policy Roundtable on "Migration
and Women"
ACKNOWLEDGEMENTS
This study was made possible by Grant No. AID/otr-G-1592 from the
Bureau for Program and Policy Coordination (PPC/WID) of the Agency for
International Development. We are grateful to Arvonne Fraser, WID
Coordinator for her interest throughout the study. Eduardo Arriaga,
Special Assistant for International Demographic Methods, U.S. Bureau of
the Census, helped us develop the migration index. Special appreciation
is extended to him. Sincere thanks go to Judith Johnson for her excellent
editorial work and to Alissa Okuneff, for her assistance in the literature
review. We are also indebted to James Brown, Economist, Bureau of the
Near East, Agency for International Development, and William P. McGreevey
Program Director, Population Study Center, Battelle,for their help in
facilitating the initial stages of this work.
TABLES and FIGURES
Tables
la Sex Differentials in Outmigration from Rural Areas in
Third World Countries by Region and Indicating Total (TMI),
Female-Dominated (FMI), and Male-Dominated (MMI) Outmigration.
lb Sex Differentials in Outmigration from Rural Areas in
Third World Countries by Region and by Age.
2 Standardized Ratio of Urban to Rural Population in Africa,
by Country, Age and Sex.
3 Immigrants to Australia, Canada and USA from a Selected
Number of Developing Countries.
4 Regional Comparisons of Index of Femaleness in Long Term
Immigration and Emigration.
4a Male and Female Immigrants in Selected Countries Around 1965.
4b Age and Sex-Specific Long-Term Emigration from Selected
Countries.
5 Sex-Specific Migration to Two Oil Producing Countries.
6 Composition of Turkish Migration Stream by Year and Sex.
7 Unemployment by Migrant Status and Regions
Brazil, 1970.
8 Economically Active Population by Type of Activity
Metropolitan Sao Paulo, Brazil, 1970.
9 Occupational Status of Economically Active Recent Migrants
and Residents by Sex and Destination: Colombia, 1964
(In Percentages).
10 Percent Distribution of Incomes of the Economically Active
Populations by Sex and Migration Status: Metropolitan Regions,
Brazil, 1970.
Figures
1 Deviations of the "Observed" Rural Female to Male Ratio from
the "Expected" Female to Male Ratio for Africa by Country and
by Five Year Age Groups.
Figures (continued)
2 Deviations of the "Observed" Rural Female to
from the "Expected" Female to Male Ratio for
Country and by Five Year Age Groups.
3 Deviations of the "Observed" Rural Female to
from the "Expected" Female to Male Ratio for
Central America and the Caribbean.
Male Ratio
Asia by
Male Ratio
Latin America:
4 Deviations of the "Observed" Rural Female to Male Ratio
from the "Expected" Female to Male Ratio for Latin America:
South America by Country and by Five Year Age Groups.
5 Deviations of the "Observed" Rural Female to Male Ratio
from the "Expected" Female to Male Ratio for the Middle East
by Country and by Five Year Age Groups.
The title of this report will lead many to question--justifiably--the
reasons for singling out or isolating for study such a large group of
people in the Third World who may have in common only the fact that they
are women. What does the sex variable represent, and does it add anything
to our understanding of development processes in general and migration
trends in particular?
Recent evidence on the economic participation of women sheds light
on the first question, for women in Third World countries are often
among the poorest of the poor. Contrary to conventional statistics
(and wisdom), the economic participation of these women in both the
traditional and modern sectors is substantial, but it is also different
from that of men, for women are generally overrepresented in those
economic sectors with low productivity and meager earnings. If, in
addition, they have to bear the main economic responsibility for their
families, theirs are the households with the least economic resources.
Women-headed households in the Third World are numerous and their
numbers are growing. One of the factors contributing to the establishment
and perpetuation of these households seems to be sex-specific migration;
either men move in search of jobs and leave women and children behind,
or women migrate autonomously and establish their own households.
This link between women-headed households in the Third World and
migration was the first in a chain of factors that motivated this report.
The absence of any analysis of the existing studies of autonomous
women migrants was the second. The third, and most important factor
was the apparent contradiction between recent data on the magnitude
of migration moves of autonomous women, on the one hand, and the lack of
acknowledgment of such data in the migration literature, on the
other. The fourth was our awareness of the importance of informing
those engaged in formulating development programs and policy of, at
the very least, the overall magnitudes and trends in Third World countries.
Trends. It is a well-known and accepted fact that women predominate
in rural to urban migration in many countries in Latin America and the
Caribbean. In the Dominican Republic, for instance, the new urban
influx of women is almost four times as great as that of men; of the
total population of Bogota, Colombia, in 1973, 51 percent of the women
and 45 percent of the men had come from other parts of the country; and
in Jamaica, the ratio of women to men has long been conspicuously
higher in urban than in rural areas. The movement of Latin American
women to metropolitan areas began in the early 1960's and it continues
today. The pervasiveness of this trend led, ten years after it began,
to the first studies focusing on the migration of women; while the
evidence is still very incomplete, a fairly consistent socioeconomic
profile of the migrant woman in the region is beginning to emerge.
Puch less is known about the trend toward increasing migration of
women in Asia and Africa that has taken place in the last two decades.
Moves of autonomous women migrants to selected Asian cities -- Bangkok,
Hong Kong, Manila, and Delhi -- intensified in the 1960's. In Thailand
between 1960 and 1972, the number of male recent migrants increased by
112 percent, while that of women increased by 142 percent. In Bangkok,
in the 1960s there were more male than female migrants, but that was
reversed in the 1970s. While women always predominated among migrants to
Manila, their predominance increased in the 1960s. That decade also
witnessed increasing "autonomous" migration of rural women to Hong Kong
and to Delhi. Pt the national level in India, the volume of female
migration exceeds that of men in rural-urban moves as well as in
moves from small towns to cities. In the 1960s there also began to
be a greater equalization of what had been a very unbalanced, male-
dominated sex ratio in West African cities; the cause of this trend was
increasing migration to urban areas of young West African women who
chose to remain in the cities rather than go back to their villages to
marry. In the Middle Eastern countries, the more striking event during
the 1960s was the emigration of Turkish wonei to West Germany as part
of the internationalization of labor.
The increasing magnitude of the autonomous migration of women in the
Third World emphasizes the fact that an understanding of migration
patterns and trends needs analysis by sex differentials. In spite of this,
studies of migration continue to associate female migration with
marriage. It is difficult to know whether this is a cause of or a
consequence of the lack of data. The result is the prevalent assumption
that the migration of women has no economic motivations or implications,
a view which obscures the link between migration factors and economic
conditions in developing countries. The data presented in this report
will show that there are serious economic dimensions to women's migration
patterns and that women migrants -- married and autonomous -- should
be studied as a category separate from male migrants and non-migrant women.
When women migrants become a focus for research, then migration models
will be designed that include the urban informal sector and women's role
therein as a factor in the migration equation and that show the relation-
ships among migration, the size of the urban job market, urban employment, a
and unemployment.
Problems of Conceptualization. One major problem facing those
who want to study women migrants is the difficulty of obtaining
reliable data. In many developing countries, accurate and sex-disag-
gregated population statistics are not available. Indeed, many countries
do not collect statistics on migrancy status, so that information
on migration must be derived indirectly from whatever demographic
data are collected. Sex ratios, urban and rural growth rates, fertility
patterns, all give us some clues to rates of migration. But they give
little indication of the direction, composition, or structural factors
related to migrant flows, particularly where women are concerned.
We had to deal with several other problems when we began our
study of migrant women in the Third World. A major problem is that
using available data it is difficult to distinguish between autonomous
female migration as distinct from that of accompanying migration
(wives moving with or following husbands). Where sex differences are
outlined in data sources, only rarely are data on marital status
variables also available. What little information there is suggests
that marital status and age are both factors in migration decisions, but
that whether they are positive or negative factors depends on the stage
reached in the life cycle at the time of the decision. A related problem
is the inability to discriminate between those married women who migrate
for economic reasons (even if they do precede, accompany or follow their
husbands) and those who do not and, more generally, the inability to link
the economic behavior and/or motivations of migrant women with their (de
jure and de facto) marital status.
A third problem is that all "types" of migration are lumped together
under one heading. It may be possible to find the total female migrant
rate in a given country but not to distinguish among the particular types
of movements that are involved. migrants (men and women) can be subdivided
into the following typology : seasonal agricultural migrants, short-term J
migrants, inter-rural migrants, inter-urban migrants, rural-urban migrants,
interregional migrants, and international migrants. Except for those in
the interregional and international groups it is difficult to discover
how many migrants are in each of the.categories. There is virtually no
information on the number of seasonal women migrants, for instance,
despite their apparent magnitude and it cannot be assumed that this category
is captured by the internal migration statistics.
Because of these shortcomings in the data, our analysis will be
concerned mostly with women migrants in urban areas. When focusing on
r!uch of what this report says can apply equally to men migrants. Its
emphasis is, however, on the behaviors of women migrants that are different
from those of men migrants and/or that respond to particular conditions
affecting women but not men.
specific variables (i.e., women's age and their economic situation) we
usually cannot distinguish between autonomous women migrants and those
who migrate as dependents since these variables are rarely cross tabulated
by marital status. In regards to women left behind as a consequence of
male emigration, this analysis examines the scant data there are on the
negative impact of male outmigration on rural women and on agricultural
productivity. Little attention has been given to the social and economic
context in which the women left behind function or to how they cope with
these dislocations. The data are limited, but some suggestive trends can
be drawn from them.
Throughout the report we have attempted to answer, or at least provide
partial answers to, a number of questions :
-Who are the women migrants and how do they differ from male migrants
in terms of age and other demographic characteristics?
-Why and under what socioeconomic conditions do women migrate? What
are the factors that motivate them to move from one place to the other?
-What are the characteristics of women's socioeconomic condition
in the place of destination? How do these differ from male migrants and
from non-migrant women? What are the reasons for this difference?
-Who are the women left behind? What is the economic and familial 4
context in which they function?
We have also tried to provide an idea of the magnitude of women
migrants in Third World countries by developing an index of sex differences
in outmigration trends for forty-six countries.
This report does not deal with the migration trends of Third World
women into the United States since this topic is currently receiving
research attention from scholars (Chaney and Safa amongst others).
II. METHODOLOGY
a. Data Sources
There are obvious data related problems which make it difficult
to assess the magnitude of internal and international migration for
both men and women in the Third World. Unrecorded "illegal" migra-
tion and seasonal migration figure prominently among the categories
for which data is lacking.
There are only two publications that attempt to quantify sex
differences in migration: The Handbook of International Data on
Women (Boulding et al; 1976), and Trends and Characteristics of
International Migration Since 1950 (U.N. Forthcoming).
Boulding's book ranks countries according to an "index of femaleness"
in long-term immigration and emigration. The index -- which is ex-
pressed in terms of the ratio of women to the total number of non-
residents and emigrants -- identifies countries where the immigrant/
emigrant population is predominantly women and others where it is mini-
mal. Because data are available for only a\ limited number of coun-
tries, the index is useful primarily for what it shows about the
differences among countries in the degree of female participation in
the migratory process.
The U.N. publication gives annual statistics of immigrants and
emigrants classified by sex and age, and census data on the foreign-
born or alien population classified by sex and age and,where avail-
able, by period of immigration.
In addition to these sources there are some national censuses
which include in their published volumes tabulations comparing birth-
place to current residence, and in some instances, sex differences
with respect to demographic characteristics of migrant groups.
While these three sources are important in establishing the mag-
nitude of women's participation in both internal and international
migration, they are not helpful in distinguishing 'autonomous' women
migrants from those who move with spouse or parents, since cross-class-
ification by marital status of migrants is not provided (with
very few exceptions). Likewise, amongst married women the extent of
migration which is induced by economic reasons is obscured.
There are some regional, country, and community studies that
complement the international statistics. Such studies are heavily
weighted towards countries in the Latin American region, where
attention has been drawn to the predominance of rural-urban female
migration, particularly the concentration of women in capital-city
areas (Castro et al, 1978; Connell, 1976; Elizaga, 1972; Elton, 1974; Fox
and Hugert, 1977; Herold, 1978; Jelin, 1977; Kemper, 1977; Rengert, 1978;
Standing, 1978d) In addition censuses in Latin American countries cover
data on migrants by sex more comprehensively than those in any other
region. (it is not clear whether the studies have caused governments to
collect better data or the data have caused the studies.)
In Africa, attention has been almost exclusively focused on the
predominance of male migrants and the points of origin and desti-
nation. The "women left behind" have received only passing mention.
The first systematic attempt at addressing this issue is, of course,
the seminal work of Esther Boserup -- Woman's Role in Economic Deve-
lopment -- which deals with the consequences for women of male mi-
gration -- the increased work, the economic burdens and the related
effects on agricultural productivity. It was the latter point, in
particular, that sparkled the interest of development planners. The
increase in women migrants to urban centers in Africa has not yet re-
ceived widespread systematic attention, though some anthropologists have
researched the situation for West Africa (Little, 1973; Sudarkasa, 1977).
There have also been some studies of sex differences in rural and urban
migration in Asian countries (Goldstein and Tirasawat, 1977; Pernia,
1977; Piampiti (n.d.); Sallaf, 1976; Singh, 1978a). Several have probed
into the questions of which women are being affected by this process and
in what ways. (Pernia, Sallaf, and Singh to cite only a few.)
Data on Turkish women migrants to West Germany have provided the
basis for studies of international migration (Abadan-Unat, 1977; Kudat,1975a;
1975b; Kudat, et al, 1976, 1979). A recent development of interest has
been the export of Arab nationals from the labor-abundant Arab countries
to the oil-producing countries that face labor shortages, but the impact
of the move on the women left behind has not been addressed except in
the case of Yemen. Given that the proportion of the exported male labor
amounts to 30 per cent of the total economically active male population
in that country, and that some specific villages are virtually depleted
of active males, women in Yemen have been assuming in some instances,
the burden of most agricultural tasks (McClelland, 1978; Ross, 1977).
Increasingly, then, Third World countries are experiencing
significant movements in female migration both as a percentage of
total migrants and in absolute numbers. It is unfortunate that
available data and, indeed, many country and community studies do
not provide indications of how many autonomous women migrants are
involved in this trend. The linkage between economic conditions
and female migrancy is also missing because of the lack of infor-
mation on women involved in seasonal migration.
b. An Index of Sex Differences in Migration Trends
In the absence of actual figures on the number of migrants by
sex, sex ratios representing the proportion of males to females re-
siding in rural and/or urban areas are frequently used as indicators
of migratory moves by sex. To identify countries in the Third World
that have autonomous migrants of both sexes and give rough estimates
of the magnitudes of these autonomous migration trends, we constructed
a sex differential (or "femaleness") migration index composed of both
an observed (or actual) and an expected sex ratio. Population figures
from national censuses for forty-six Third World countries were used
to calculate the index.Y/
The observed sex ratio (y) is defined as the number of females
per 100 males for all five-year age groups between the ages of 15 and
64 in rural areas. For each five-year cohort:
rural females X 10
Rural males
2/ U.N. 1977, Demographic Yearbook, 1976.
Other things being equal, we assume that when y is greater than
100 -- that is, more than 100 females per 100 males reside in rural
areas -- there is male-dominated outmigration. Conversely, when y
is smaller than 100 we assume female-dominated outmigration from rural
areas. If y Is equal to or close to 100 -- that is, the numbers of fe-
males and males in the rural area are equal -- there are three possible
explanations: 1) There is no outmigration from the rural area, 2) there
is family outmigration, or 3) males and females outmigrate independ-
ently but in the same numbers. We focus on rural outmigration rather
than urban immigration both because rural sex ratios are less likely
than urban ones to be distorted by international immigration and be-
cause the magnitude of urban to rural migration, when compared with
rural to urban moves, is generally not large enough to have a major
distortion effect on the rural population ratios.
The sex ratios by age groups help to identify outmigration that
is selective by both sex and age and indicates when in their life
cycle women and men migrate independently. Further, analysis of sex
ratios by age which leaves out people between the ages 1 to 15 reduces
some of the possible biases due to sex differences in fertility and
mortality rates, which tend to be large in many developing countries.
To further reduce biases due to age-specific sex differences in fer-
tility and mortality rates, we introduced a second ratio into the mi-
gration index, the expected sex ratio (y).
This expected ratio is defined as the expected number of females per
100 males that would result in a population where only sex differentials
at birth and mortality are operating (not migration) for all 5 year age
I
groups between the ages of 15 and 64. Life tables by sex and age
from the U.S. Bureau of the Census series on Country Demographic
Profiles, which take into account morality differentials but where
migration is not a factor, are used to calculate y for each five-year
group where:
f L female x 100
L male
L stands for life table estimated population for a five-year age group
and f is a constant included to account for the sex ratio at birth and
represents the number of females per 100 males. Assuming a differen-
tial of 105 males for every 100 females; f equals 0.952.
If y-y is different from zero, that is, if the observed sex ratio
is larger or smaller than the expected sex ratio, we can safely assume
that these deviations from the expected value are due neither to mor-
tality nor to the differential sex ratio at birth; we further assume
that the main factor accounting for such deviations is sex differen-
tial migration. It is also assumed that biases due to sex differences
in age misreporting and underenumeration are not large enough to sig-
nificantly alter the observed sex ratios. Given these assumptions,
we define and interpret the deviations from the expected sex ratio
values as follows:
1. The male dominant migration index (MMI), indicates the extent
to which men migrate more than women from rural areas and is represented
when the sum of the deviations from the age-specific expected ratios
3/ See explanatory notes at end of chapter.
is positive (i.e., more females than males, y-y is greater than 0).
2. The female dominant migration index (FMI) indicates the extent
to which more women than men outmigrate from rural areas and is re-
presented when the sum of the deviations from the age specific ex-
pected ratios is negative (i.e., more males than females, y-y is less
than 0).
3. The sum of the two indices, MMI and FMI (using absolute num-
bers), is the total sex differential migration index (TMI), and it is
an indicator of the extent of overall sex differential outmigration
from rural areas within a country. (TMI = MMI + FMI)
The migration index is relative and yields information only on the
extent to which more women than men (or vice versa) migrate within
specific age groups. It cannot yield information on the absolute mag-
nitude of migration for each sex. Therefore, a score of zero in the
female or male migration index for any country does not necessarily
mean that no women or men outmigrate either alone or with their fami-
lies; rather, it merely indicates that within the specific age group
women do not have a greater or lesser propensity to migrate from rural
areas than men.
To compensate for this "loss" of information we have used, as com-
plementary information, the proportion of urban to rural population
by sex and age in each country. Independently for each sex, we ob-
tained urban to rural ratios for the population in every five-year
age group between the ages of 15 and 64; in addition, we obtained the
urban to rural population ratio for all ages, and compared this total
with the age specific urban/rural ratio. The working hypothesis for
this comparison is that if there is no migration from rural areas to
urban areas, the age-specific urban/rural ratios will be similar to
the total urban/rural ratio; if there is rural outmigration, however,
the age-specific urban to rural ratio will tend to be larger than
the total ratio. Thus, a country with a FMI close to zero might still
have outmigration of women from rural areas, if the urban to rural
ratio for the female population within specific age groups is sig-
nigicantly larger than the total urban to rural ratio for the female
population. If the same pattern for the male population in the coun-
try is found, this suggests that the low value of the total migration
index, (i.e. no sex differences in migration) is due to significant
outmigration from rural areas of both men and women rather than to no
outmigration.
In order to compare the age-specific urban/rural ratios between
men and women and between countries (given the different levels of
urbanization across countries), the age specific urban/rural ratios
were standarized for each sex by dividing these ratios by the urban/
rural ratio for the whole population (multiplied by 100). The age-
specific urban/rural ratios presented in this report are the standardized
scores. Because this standardization procedure makes difficult any
interpretation of age-specific urban/rural ratios that are smaller
than the total urban/rural ratio, they are not included in our ana-
lysis. In short, they should not be interpreted as an indication of
urban to rural migration.
Additionally, we have assumed that the distribution of the posi-
tive urban/rural ratios for those age groups analyzed is not altered
by rural/urban fertility differentials.
15
Explanatory Notes:
While it would have been desirable to use information for
rural populations, only life table estimates for whole country popu-
lations were available.
Estimated life tables were not available for every country.
In Africa tables were available for only Kenya (1969) and Ghana (1970).
Except for Kenya, all the African countries listed in our index
used the age-specific averages for these two countries.
For Asia, estimates were available for India (1969); Indonesia
(1961-71); Korea (1966); Malaysia (1970); Nepal (1974-76); Philippines
(1969-71); and Thailand (1970). For the two countries without
estimates (Bangladesh and Pakistan) we used data for India which was
thought to most closely approximate these countries' sex-specific
mortality rates.
For Central America estimated life tables were available for
Costa Rica (1972-74); Guatemala (1970-72); Honduras (1974); Mexico
(1970); and Panama (1969-70). Age-specific averages of estimates for
these countries were substituted for other Central American and
Caribbean countries for which data were not available.
For South America estimated life table ratios were available only
for Brazil (1967) and Chile (1969-70). The averages of the age-
specific expected ratios were used for all other South American
Countries listed in Table 1.
Estimated life tables for Middle Eastern countries were not
available through the U.S. Bureau of the Census. However, data for
the East Bank of Jordan (1972) was obtained from "A Study of Mortality
in Jordan with Special Reference to Infant Mortality," by Dr. M.
Sivamurthy and Abdul Rahim A. Ma'ayta. The expected sex ratios
used for all Middle East countries were derived from these data.
Below is a summary of the computations.
5Fx f.5LF
y- -- X 100 LM X 100
5 x 5 x
M.I N y y > 0
FMI = y y 0
TMI = MM1 + FMI
FU
M5Ux
F5R
s 5 x s tfR
females -- X 100 males = x" X 100
FU MU
t
FRt MR
Where: x = specific age group; F = female.population; M male population;
U urban; R = rural; L = life table function; s = standardized ratio; and
t total sex specific population.
III. WHERE ARE THEY?: MIGRATION PATTERNS
a. Results: Sex Differences in Regional Migration Trends:
Africa. Using population data for fourteen African coun-
tries, the Total Migration Index (TMI) shows that, of the four major
'/ regions in the Third World, Africa has the highest level of sex dif-
ferential outmigration from rural areas. The average Total Migra-
tion Index (which is the sum of all age specific female and male
dominated outmigration) is 187 for the African countries while it is
110 and 111 for South and Central America and the Caribbean respec-
tively, 103 for the Middle East, and 75 for Asia. Only three African
countries, Ethiopia, Lesotho and Mauritania, show total migration
indexes below 100 suggesting that in these countries there is little
or no sex differential migration and, perhaps, little outmigration
4/
from rural areas in general (see Table 1).- We will come back to the
general question of outmigration for these three countries later on.
Figure 1 reveals a prevailing and quite consistent pattern in
most of the other countries in the African region. With the exception
of Libya, outmigration from rural areas is heavily male dominated,
especially in the three age groups between the ages of twenty and
thirty-four. Further, South Africa and Botswana (and, to a lesser
extent, Tanzania) show no female dominated migration in any age group.
These results are consistent with the evidence in the migration litera-
ture which indicates predominant male outmigration for wage labor in
urban areas and for work in the mines (Lesotho presents a discordant
finding; because of sex-segregated work in the mines, the patterns
4/ Tables 1, lb, and 2 are at the end of the chapter. Figures 1 through
Share based on data which is included in Table lb.
17
Figure 1. Deviations of the "Observed" Rural Female to Male Ratio from the
"Expected" Female to Male Ratio for Africa by Country and by Five Year Age Groups.
Deviations from expected
105 rural ratio of women per
100 100 men K
100
95 *** Kenya 1969
90 oooooo Lesotho 1972
S \ -- -Libya 1973
85 -- xxxxx Mauritania 1975
80 -- imem Morocco 1971
\ ........ Rwanda 1970
75 *"*S.. u* South Africa 1970
70 0t 3= Tanzania 1967
Botswana 1971
65 --
i **
60 -- .
55--
45 -
430 ...**a.
35.- / \ \ '
30'-- J **. \
25 I
+(MMI) 20 -
-10 i O 4.. "
-IO 1/5 \-< .k %
-10
-(FMI) -20 .. ***
-25 -
-30 -
-35 -
-45 --- Five Year
15-19 20-24 25-29 30.34 35-39 40-44 45-49 50-54 55-59 60-64 Age Groups
Source: Rural Population data for the "observed" sex ratios were obtained from the UN Demographic Yearbook. 1976 Date noted for
each country refers to year data were collected Data for the "expected" sex ratios were derwived from the U.S Bureau of the
Census estimated life table values
Note: Positive deviations from the "'expected" sex ratio indicate male dominated rural out-migration (MMI); negative deviations
reveal female dominated rural out-migration (FMI)
should have been similar to those of South Africa).
As Figure 1 also shows, except for South Africa, Botswana and
Lesotho, all other countries show female dominated outmigration in
the 50-54 and on age groups; in many countries, this outmigration
starts by age 45. An immediate question arising is that perhaps
these data reflect sex differentials in mortality rather than mi-
gration. While it cannot be completely discarded, this explanation
loses ground since the migration index compares the observed (ac-
tual)sex ratio for the specific age group with an expected ratio
based on estimated life table values that control for mortality dif-
ferentials. The recent evidence, moreover, increasingly supports
this pattern of female dominated outmigration from rural areas in
'Africa -- outmigration that takes place when women become widows or
V/
separate/divorce and find themselves with no means of economic sup-
port in the rural environment. Figure 1 further reveals female do-
minated outmigration for the ages 15-19 for Kenya, Lesotho, Morocco,
Rwanda, and Libya. Libya's general outmigration pattern deviates
most within the African context. For unknown reasons it shows female
dominated outmigration for all age groups (we found no literature on
women's migration in Libya;as such, it emerges as a country in need
of sex-specific migration research.)
The standardized urban/rural ratios give additional information
on migration of both men and women (migrating jointly or autonomously)
that is not reflected in the index because the sex ratios only yield
relative or differential magnitudes. They show that, in most countries,
there is significant rural to urban migration of both women and men
between the ages of twenty and thirty-four -- a large portion of
which probably is family migration (see Table 2). These urbanward
trends are present in Mauritania for men and women between the age
of twenty-five and thirty-nine, which indicates that there is rural to
urban migration in this country that was not picked up by the index.
This is probably because men and women migrate in similar numbers.
The urban-rural ratios for Ethiopia, however, are more difficult to
interpret. They show stable rates for women but do show rural/urban
differentials for men between the ages of twenty-five and forty-four.
This should have been picked up by the index. As we already men-
tioned, the other country that has non-interpretable data is Lesotho.
The general possibility of increasing migration among autonomous young
African women to the cities defines a central issue for policy and
program formulation.
Asia. Asia, represented in this analysis by nine countries, has
the lowest regional sex differential in rural outmigration (see Table 1).
Two countries, Nepal and Indonesia, deviate from this pattern re-
vealing comparatively high levels of male dominated outmigration from
rural areas between the ages of 20 to 34. Nepal also reveals male
dominated outmigration in the older age groups -- ages 50 and over.
Except for this country, and to a lesser extent for Thailand (where in
general the least amount of sex differential outmigration is shown),
all other countries show small but consistent women dominated out-
migration in the older age groups -- a "milder" version of the pattern
reflected in the analysis of African countries (see Figure 2).
20
Figure 2. Deviations of the "Observed" Rural Female to Male Ratio from the
"Expected" Female to Male Ratio for Asia by Country and by Five Year Age Groups.
105
100
95
90
85
80
75
70
65
60
55
50
45
40
35
30
25
.(MMI) 20
15
10
5
0
-5
-10
-15
-(FMI) -20
-25
-30
-35
-40
-45
Deviations from expected
rural ratio of women per
100 men
*o*e Bangladesh 1974
xxxxx India 1969
ooooc Indonesia 1971
===- Korea 1966
........... Malaysia 1970
---- Nepal 1976
- Pakistan 1969
t*** Philippines 1971
i-,Thailand 1970
I
I
I
1 S. I
5-19 20-2 I2 I
15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64
Five Year ...
Age Groups
Source Rural Population data for the observed" sex ratios were obtained from the UN Demographir Yearbook. 1976 Date noted for
each country refers to year data were collected Data for the "expected sex ratios were derived from the U S Bureau of the
Census estimated life table values
Note Positive deviations from the expected sex ratio indicate male dominated rural out-migration (MMI) negaltve deviations
reveal female dominalea rural out-migration (FMli
r
~-
~
Half of the countries show slight female dominated outmigration
for young women between the ages of 15 and 19. All Asian countries
show a consistent trend of male dominated outmigration between the
ages of 25 and 34. These two groups of people probably outmigrate
from rural areas -- the young women going to the cities -- mostly
for work related reasons. The literature indicates that jobs in the
city are a predominant reason for the high number of young women
migrants to Seoul, Korea; as Figure 2 shows, the analysis confirms
this finding. The analysis of sex differentials also corroborates
the well known fact of significant women outmigration in the Philippines.
This is the only country in the Asian region that shows a high fe-
male dominated outmigration index and almost no male dominated moves
(see Table 1).
Data in Table 2 show that,with the exception of Bangaladesh
(which shows no substantial outmigration) and Pakistan (which shows
only male outmigration), the other seven Asian countries indicate /
high levels of rural to urban migration of both young men and women
between the ages of 15 and 24. The literature suggests that some of
these young women migrants move to urban areas with or following their
husbands while others move to find jobs in the cities.
Latin America: Central America and the Caribbean. Seven Central
American countries plus Cuba, Haiti, and Puerto Rico, all show
an extremely high and consistent pattern of female dominated
outmigration from rural areas (See Table 1). Costa Rica,
Figure 3. Deviations of the "Observed" Rural Female to Male Ratio from the
"Expected" Female to Male Ratio for Latin America; Central America and the
Caribbean.
105 Deviations from expected
S rural ratio of women per
100 -
100 men Costa Rica 1973
95 -- oooooCuba 1970
90 ~~~=o Dominican Rep. 1970
90-
-Guatemala 1973
85 -- Haiti 1974
S........ Honduras 1973
80 -
Im Mexico 1976
75 -- ,*- Panama 1976
xxxxx Puerto Rico 1970
70 -
65
60
55
50
45
40
35-
30
25
+(MMI) 20 -- *
15 .
5J-
,xxx xxxxxxx
^^"*f-- -------- n, \ ___ A .- ---,-
.0 P .... %
N- .............
5 V t'v, --.
-15 cc ***t **
-(FMI) -20 c*** ..
-25 -
.30- 0oo
00
-35 C
.40 i I I I I coo
40 I I I I Five Year
-45 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 Age Groups
Source: Rural Population data for the observed" sex ratios were obtained from the UNDemographic Yearbook. 197 Date notedfor
each country refers to year data were collected Data for the "expected- sex ratios were derived from the US Bureau of the
Census estimated life table values
Note: Positive deviations from the "expected" sex ratio indicate male dominated rural out-migration (MMI. negative deviations
reveal female dominated rural out-migration (FMI)
Cuba, Honduras, Mexico and Panama reveal female dominated outmigration
for all age groups. The remaining countries, interestingly enough,
show neutral or slightly male-dominated outmigration within the younger
age groups, especially between the ages of 20 and 29 (see Figure 3).
Except for Haiti in particular, and to a lesser extent for Mexico
and Panama, the urban/rural ratios indicate that the migration of women tothe
urban areas increases steadily with age. Haiti and Panama show very
high rural to urban migration rates for young women ages 15-19. The
urban/rural ratios show rural to urban migration occurring with less
intensity for men than for women; where it does occur, it tends to
concentrate in the 20 to 29 age groups (Table 2). This combination
of findings suggests that the male dominated outmigration picked up
by the index is international (labod migration rather than rural to
urban migration within the country of origin. The evidence from the
literature confirms the general pattern for the region: women pre-
dominate in rural outmigration movements at all ages; where men pre-
dominate,it is a result of international rather than national urbanward
migration. The analysis further suggests that for women in the younger
versus the older age groups, there is significant outmigration to
other rural areas or other countries, rather than mostly national ur-
banward moves. Many of these younger women probably move with or
following their mates; yet many others probably move autonomously.
South America. The Seven South American countries show a con-
sistent trend but with extremely quantitative variations. Except for
Guyana, six South American countries repeat the regional trend for
Figure 4. Deviations of the "Observed" Rural Female to Male Ratio from the
"Expected" Female to Male Ratio for Latin Americas South America by Country
and by Five Year Age Groups.
105 Deviations from expected
rural ratio of women per
100 100 men
95 -
xxxxx Bolivia 1972
90
SBrazil 1970
85 ***. Chile 1967
0 ooooo Ecuador 1974
80
==-= Guyana 1970
75 ...... Paraguay 1972
70 Peru 1973
65
60
55 -
50
45
40
35
30
25 -
+(MMI) 20
15
10
., _XX X ..-.
-5
-5 2,-00 C*oo Ai
-10 \ .> *..
-35
-40 I I I I I I I
SI I I II Five Year .
15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 Age Groups
Source- Rural Population data for the "observed" sex ratios were obtained from the UN Demographic Yearbook 1976 Date notedfor
each country refers to year data were collected Data for the "expected" sex ratios were derived from the US Bureau of tne
Census estimated life table values
Note: Positive deviations from the "expected" sex ratio indicate male dominated rural out-migration (MMI). negative deviations
reveal female dominated rural out-migration (FMI)
Central America and the Caribbean and show female dominated rural
outmigration in all or almost all the age groups. Unlike any other
region reviewed here, however, the extent of sex differential in-
migration varies from an extremely high value for Chile to an ex-
tremely low one for Peru (a TMI of 235 for Chile and of only 28 for
Peru -- as shown in Table 1).
Figure 4 shows particularly high sex differentials in rural out-
migration for the older age groups for all countries. Apart from
Guyana, only Bolivia and Peru show slight male dominated outmigration
from rural areas; Bolivia reveals this dominant male outmigration for
the age groups between 15 and 34, and Peru only for the 30 to 34
age group. The urban/rural ratios confirm the urbanward migration
of women. They also show a rural to urban migration of men for all
countries (except Guyana) occurring especially between the ages of
20 to 34 (Table 2). Table 2 also indicates rural to urban migration
in Peru (not picked up in the sex differential index) of men between
the ages of 20 to 34 and of women between the ages of 15 to 29.
Paraguay, another country with low sex differentials, reveals in the
urban/rural ratio urbanward migration of both men and women, especially
in the very young and very old age groups. -
Middle East. The Middle East follows Asia in yielding compara-
tively low sex differentials in rural outmigration -- as shown by the
average total migration index (TMI) forfive countries in the region
(see Table 1). According to this index, Turkey has the greatest ex-
tent of sex differential in outmigration, which is heavily male
Figure 5. Deviations of the "Observed" Rural Female to Male Ratio from the
"Expected" Female to Male Ratio for the Middle East by Country and by Five Year Age
Groups.
105 Deviations from expected
100 rural ratio of women per
100 men.
95
90
85 ........ Iran 1971
80 -ooooo Iraq 1973
SLebanon 1970
75 ==== Syria 1970
70 -- Turkey 1967
65
60
55
50-
45 \
I \
40- I
35-- \
30--
25 I o
+(MMI) 20 I c0
oo o
.10 0 / o. *
0
0 1 o -- -o ____.___, -
-15 -
-(FMI) -20 o 00
.25 -o
-30 -
.35 -
-40 -
45 I I I I I I Five Year .-- -
15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60.64 Age Groups
Source: Rural Population data for the "observed" sex ratios were obtained from the UN Demographic Yearbook. 1976 Date noted for
each country refers to year data were collected Data for the "expected" sex ratios were derived from the US Bureau of the
Census estimated life table values
Note: Positive deviations fiom the "expected" sex ratio indicate male dominated rural out-migration (MMI). negative deviations
reveal female dominated rural out-migration (FMI)
dominated while Iran shows the least amount of sex differential
5/
rural outmigration.
As it can be observed in Figure 5, it is harder to identify
consistentage specific patterns of sex differential rural outmigra-
tion for this region than,for instance,it is for Africa or Central
America. It may well be that the absence of a regional pattern is due
to the small number of cases (countries). An additional problem with
interpreting the Middle Eastern sex differential rates is that they
are based on Census data collected prior to the oil related labor
migration of the 1970s. This labor migration has probably altered
substantially the sex differentials in rural outmigration.
Given these data limitations, the more consistent regional trends
are male dominated rural outmigration between the ages 25 to 34 for
five countries (Iran being an exception) and female dominated rural
outmigration from age 45 onwards also for five countries (Turkey being
an exception in this case). This latter finding is probably reflect-
ing outmigration of widows while the former is a function of labor
related male migration.
Iran shows a small but quite consistent pattern of female domi-
nated outmigration across age groups. The urban/rural ratios ad-
ditionally show rural to urban migration of both men and women especial-
5/ The urban/rural ratios for Turkey indicate higher male than
female migration to the cities. However, the gap between the
two is smaller than is indicated by the MMI and FMI of this
country. The small urban/rural ratios for males may reflect
heavy male outmigration from urban areas to other countries --
men who are replaced by male migrants from rural areas.
6_/ However, this may well be a function of high female under-
enumeration in the rural areas of this country.
28
ly between the ages of 30 to 54, which probably indicates family out-
migration.
The urban/rural ratios for the other countries in the region
also show rural to urban migration for both sexes in the middle age
groups. However, the age ranges in which men and women are most apt
to migrate differ from country to country.
Table la: Sex Differentials in Outmigration from Rural Areas in Third World
Countries by Region and Indicating Total (TMI) Female-Dominated (FMI) and
Male-Dominated
(MMI) Outmiqration
AFRICA
x (SD)
Benin
Botswana
Ethiopia
Kenya
Lesotho
Liberia
Libyan Arab Republic
Mauritania
Morocco
Rwanda
South Africa
Southern Rhodesia
Tanzania
Uganda
TMI FMI MMI
187 (123)
a/
187
429
95
126
13
391
137
52
163
238
337
209
136
104
59 (45)
48
0
94
27
13
117
137
30
84
91
0
92
20
66
128 (132)
139
429
1
99
0
274
0
22
79
147
337
117
116
38
Source: United Nations Department of Economic and Social Affairs, 1977.
Demographic Yearbook 1976. (Table 7); and United States Bureau
of the Census, c. 1960-1975. Country Demographic Profiles for:
Kenya, Ghana, India, Indonesia, Korea, Malaysia, Nepal, Philippines,
Thailand, Costa Rica, Guatemala, Honduras, Mexico, Panama,
Brazil, and Chile.
a/ TMi = (FMI + MMI1. For each country FMI values are the sum over five year
age groups between the ages 15-64 of the observed negative deviations from the
expected number of rural women per 100 rural men. MI1 values are the sum over
five year age groups between the ages 15-64 of the observed positive deviations
from the expected number of rural women per 100 rural men.
,
30
Table la(Continued): Sex Differentials in Outmigration from Rural Areas
in Third World Countries by Region and Indicating Total (TMI), Female-
Dominated (FMI), and Male-Dominated (IMI) Outmigration
TMI FMI MMI
ASIA
x (SD) 75 (29) 21 (18) 54 (38)
Bangladesh 54 11 43
India 72 18 54
Indonesia 101 13 88
Korea 59 25 34
Malaysia 80 32 48
Nepal 138 1 137
Pakistan 70 32 38
Philippines 64 57 7
Thailand 38 0 38
MIDDLE EAST
7 (SD) 103 (23) 49 (17) 55 (37)
Iran 76 64 12
Iraq 122 57 65
Lebanon 81 51 30
Syria 111 53 58
Turkey 127 19 108
Source: Ibid.
31
Table 1 a(Continued): Sex Differentials in Outmigration from Rural Areas
in Third World Countries by Region and Indicating Total (TMI), Female-
Dominated (FMI), and Male-Dominated (MMI) Outmigration
TMI FMI MMI
LATIN AMERICA:
CENTRAL AMERICA &
THE CARIBBEAN
x (SD) 111 (43) 95 (49) 14 (24)
Costa Rica 104 104 0
Cuba 200 200 0
Dominican Republic 142 135 7
El Salvador 74 60 14
Guatemala 82 73 9
Haiti 114 53 61
Honduras 59 59 0
Mexico 78 78 0
Nicaragua 79 76 3
Panama 160 160 0 .
Puerto Rico 116 51 65
SOUTH AMERICA
x (SD) 110 (65) 94 (77) 16 (33)
Bolivia 113 96 17
Brazil 134 134 0
Chile 235 235 0
Ecuador 103 103 0
Guyana 97 5 92
Paraguay 63 62 1
Peru 28 23 5
Source: Ibid
IS 0 -
7 c- -1
;I N 2
kn wi
0 C 0 0 0 C
o c o om ua
x
g 2 s 8 8
en mU Pn 1
0 0% 0 ~ = "
1 7I m I u C f
i % C% C" l
N O
m C0 0 C0 % a,'
-r |
- oI 0
N. N N
0% 0% 0' 0
0
C
VI
0%
- I
.LI:
I--
o0 -
I 10 ^ 0 0 %0A 0
ca T gc
-
-I
.30-
.B0
-A
J'
41-.
0 C
4-15
VIS
U-
>Y,
41*
e L
.0
US
L 41
.0 C
^ Oc
=5^
.5~4
- *0 I.
0)~1
o L~
C O)
LO
-I '
55
1C~
3A5 6
0 AU
L U
41 LY
- C~
u Xc
-*0 L1
z- 8
5-?*
o0 .
C >C LA
CO
LA 41
ctc
41 L -
- J
1- 10 l
.- 0 *0
cgx
--4~1
'C C LI
- -0
uge
AU
e 041
&>- 4
A- C -
0N o f 0 in o ) 0 I
6 - -
ci cc co G C s
0%
S 0 0 0 0 0 0
N 0I L -C
i 0 0 0 0
a, 0 V1 ON 0
Sj N ( I
C7
S I I I 1 1 I --
*O N co cm
f x S m co C el
V)
t1 0% M 0 0 C0
0 m 0 i m 0 h 0 0% 0
IO
SN 0 0 --
I 0 in I
0 ^0 -
Uo foI U 0 o 0
S* 0 0 0 0 0 0 0
% % % 0% 0% 0% 0
IC
~O IC
U IC
1.1.1 &
H
o
I.
IA
0
L
*0
a,
II
u
trt
C
0,
U
0.
0
Ift
w
C)
VI
C,
I,
C,
01
01
IO
0%
C,
IC
N
1s
0%0)
*1
I)CC
(0
fsc
CIAC
o -
Y)0
"S 3
S
>.0
i^C
gC
S0I
ii
-w u
*I-
IC
ifl0
OIL
-I 0
- Cl
IVI tf
N 0
N N
4 4
0 0
CN .4 0%
- t N N
S %. ,
.4
fm 4 0 O w e
- fi N p r r.-)
%o 0 sN cf Ch w 5 Ns 4
c0 9% 40 0% c c0 0' 0% o 9% o
S I Cv 1 t I I i S
^ r ^ 4 4 N ,-
un
4~~~7 0 0 % 45 N 0
o 4 0 0 00 00 0 00 0%0
611 -z
NO C- c9 w 04 9
61Q 0 0 % c f c -
:14 -004
LM) M r f" ON ew w f" Cc qO m
o1 4^ O -^ 4^ 4 9S 9' 4 -
c' % 0%s' 0 0 -
If r o4 c o < u) m ^
45.~ -
Y = -N 0 V 40 M N 0% -
C: c ^ c c w
0 0% m 0% m 0% C0% C, % O m 0
c c r a, 0 C, Cc 0
Fl,
6M wf ON a, 0% N 0 F, 0 45 0,
oh 0% 0% 0
0%' a, m0 a, %0% a, o0 m0 a 0
rw
0 ol a 0 -7 a, 9
C cl trt ^ r s D
Fo a, or c0 a 40 C7% F f, co
>-^ S- -. -^ -^ -^ -^ ^
0% -.
CCL
N II6
9% % 9% % 40 9% 0% 9% 9%
g. 0- Nr -/ N0 -n 0^ N/ 0) 1% 9 N
=^ o 4 ^ o- o9 4 r 4 -
-r -r ^
M -
9% % 9 9 40 9% = 9 9
419 ** 0% 0% 0% 0% 0%i 0% 0' 0% 0% y
o 0f 0% 0% % 0% 0% 0% 0 o% oy 0
uj 0s -- 4S' 45. F, 4 0 40 0
0 ~ ~ 0 0%J 0%t 0%^ 0% 0% 0%t 0% 0% 0%0%.
=4 0
45 .5 -^ eo s IA ^ o
0 4
45 C 45 U
45fs rsj V 45 45
.4i 91 0 r B ( 0
c^ c^ o c^ y y t ^ c
5 1hI h h h
5
uj uj a: L e~
Br u e c r
"J = C S 3 r o ^ r
c i- u i- j ( uc
0,
0I
e>
f<.
45-
c
Q
-oc
-Cr
ow
*0*
<- .>
0%
- .C
3
u-
0*J
0'
SL-
-L
-5g
i u
-C
0v>
OC
-r..
0
0B 3
c,,
45=
3
f t
540
- 0%
I L
CM 95
<- 0
- U
*c -
"nU
- 0
C.5
a IA
c-
H0
S'5
vtm
-sL
30
- u
i?
- OJ
v> ^
u e
&*
A q 0. 10
0%> 0
f S
m 10 A -
0% 0
- N~ -
- -
5 CO 1
COi 0 0
C'J 0 I
CO 0%X
>l f> IO (y w in r
w 1 - -I -
CO s-
d> r.i % 0 O O O O % -- 0
0 co CO C O cc oO CO 0
7 --
1 0
- .5. 10 0 It -^ 0^ 0
A 2.. 0% 0% 0 0 0 CO 0 0
N1~ N Q~ IA CO a~n00%
C en 0 0n c 0% e 0n
C7, co ON IA Ch 10 0%
5. W0 C C O 0 0 CO 0n 0%
~~~~: m ^cc*tf r ^
w LA 0
s f cocs m v' a aii
ne
o 0% CO c 0 0 CO CM O O0
o Co er 0 A IA N oN .5 -
> r s 0 I -
S CO O 0 0% CO CO m 0%
COw co N 0 C % IA 0
S C O CO 0 N 0 -O
SO 0 0 OC C0 PO
mA m5-CON 0% -
o 0 0T 0 C
NJ wi 1O7 C 10 I% 10 m C0
C-71 co
^- N- -^o
CO -' -r ^ ej ^ ^ t ^ ^
6A w M m CO C % m % N N
i 1. CO O 0' 0 0%- O CO 0 0'
0% 0 0 0D m 0 0 0
0 Co C, Ca 0 7 = O N
N T -- -
IA X CO COt IA. -~ N
FS *
-A CO 0% oj N CO 10
I 0% e. CO 0 0 0 0
* CO CO N' 1% 0% 10
Ift uj 0't 0' 0% 0% 0' 0% 0't 0% ff%
O 10 0 ~' N mI N\ 0' N* N
0 ^ 0 0% 0 ^ 0% 0 CO 0% 0
"*- .r..rr_' _i
- 0
0% 0% 0%
S.
#5 01
- a
fiJ
0 0 IV
1 i ^
- CO 0 C
0
c
0 -* '
.5 0. c
-e 0. .5
c 0. -:
= c- c i
'0
N U
I0%
i!
10,
0%01
50
f^ -
.5
o'
0
4,
4,4,
.0'0
-'.
10 -O
^1
-
C0%
l,~
0-
L.
0
05
-0
0U
*- -'
s3
ii
-I.
19
-c
e re
OISU
&i
Ir I
0, 10 4,
0-O
In4
B)3
4S 4,.
- 0.4,
L=?
'0.5
4:- xi
i (
- 4, 0.
J2 -
-In.0 0.
- k.
Ic" S
IUJ *J.5
~0I
z1
5!4
,,
'0, .
0 0 .
N
0 0
0 0*
- 0 0
0 0
0 0
0 N
-0.
aN
0. 0
0 0
- N-
- I
^
0 0
r- P-S
N CO
- 0
0 0
0. N
0 0.
!- 0 c N 0
II a
i 00 00 CV r
0 0. 0 0 0 N 0 co 0 cO N Os
0 .0
0 0
S 0o
0 03
- N
I I
0 0
c C
a, CO
0.
0 '
UM USl
0 0
0 0
I I 2 2 2 2 2 2 2 2 2 2 2 2 2
e N S e C 0 4 ow
a a t a o O
ico Ch, c ^ 'S 3 '^
r -^ ^^
*Ir r 0 0-
cc wc
g| 0 kn > 0 co o CM SS US 0
a S N Il g
6j - -
sl 8 ,E
wj s a--
S ri 5; p o 3- S 5 5 5 5 3 5 3 5
N^ Nf ^ l NS N N N N N0 Nf NB N NA
o o M m 0 0 0 0 0 0 0 0
C' s 0. E a c a % a cl, m cy O a
kl -kn CP- -
m cl m &I Ol a, mo -l &. a,*m f- n
i 0 0f 0' 0 0 T 0' 0f 0^ 0 0^ 0 0 0^ 0
0r c% 0 C', N C% a% C0 US r' a, N %
Xi 8t Ef wf O O W ff' O f ^ ftO
*f US 0. N 0. N 0. 0n 0 w -E 0c 0
D NO 0 N^ US -^ NC I 0 N 0 US N ) ?- S
I 0 - - -
> 0 0 0 0 0 0 0 0 0 0 0 0 0
U, In
N US r N^ 0 0 0- "S '0 0 -. aC -a
0, N r 0. No US 0. 0 N ^' uS N 0
^ 0 0 0 O ^ C ^ 0 0 ^ 0 ^ ^0 0 0 i0
^ 0 c ^ 0 f 0 f 0 ^ 0 0 0^ 07 0. 00 0^
C vC
e, 0 I
4: 4: -a
a,
-c u
a? 4:
aU a, -s
CI U C
aC. -- U ~ C L aB 'C
0 C o c
== I~ 'C =
~ ^ ^ So o t r
- o 1-1 r >- 3
a.
-a-.
-I aO
-i
4:0.
'-'0
~S-
5-g
-C:
.4Y
us:
- a.
CCC
Le
U- 0
.i r fe
^C?.
- 0ai
<5o^
U~ U-4
-C- 0
-us
Ua.a
-: C 4,
-x
a us
4,
s
4:0
-
L' a L
*A L
CJ 4J
Oja
VL
f u
ye s c
's'
=>^-:
fcte
C
40-
03
2Y-s
QI e /
0
^ c
U
0"
C
-- -
I,
o w
E
U CI
a s
a, 0.
4s .4
= 4
0
a,'s
c i
Ci
I S
nC.
'-^
=s ^
g ^
I.L.
C,
w- uS
I. 0
4 '
'C C,
c C
C -
0 ft
S
a,
C.-
C^.
fa C
a,
4:
C)
- 0.
4:
- t'
a,
I..
SU:
'C
4:S
CI
- 41
a.-
I
i
i
O-
r-
r*r^
. %. =* C
N L1i c. O ,
= un r a
C\J?
--N N
07 -7 N~
It= h(C h
I*- w
PS = = C)'C
C*I r !% m. P
nL. f-=C* Oc
%= .c
L~C ~*C
en
i M
i-- CC -o h- Ch -C
Lf C^J -- C' -rS -5 --Q
Im-
l Cc M *- (." pr'r --C .
N- IC.
i LC ~~ ~=
|co n O N ~
C
0
111
re
S- c
<
C1
0.
Ek
ere
- 3
= ro
^r.
tIl
d,
u
c
" 5=
0t)
.
rer
CC: CCI NC OC= -- -cO
c.c Ot 00' Nc 0= nE -~
% n u C e- eun nf- CO .
L;S LL- LE LI' L L.' Lh_ LE L.Z LU
---" In ^I ;
s
rz c1
X _'
I
=o- r'j I -r 0
6n ^- 6n
CN
7>
-----------
-
C c~ a := oo < i. r^ *i
X
Ln V
oe v
S-
.,-
C (U C
L C TO C
C
E fm
o *-
-c- *'-x
e- *- C3
CL C
9 -
- E c
S0 t S-
C
>. 'o & v
r- E 0
_c C
=u
S3 .
rr d
WmI 0)
0. .
0-
u -
0 0
jre c'
*^ r i
5C3 o <
<-> 3 e _
e a)=- s
o ca L
0
N"C
OEN
0I
CCI-
-0
*-> zO f, %C C % rN.-
r"^ IC rj n%- r~-
-- N C 0 NC
n- <- C <. r"- -^-
Vc 0 \i
c CS Ml C S".
00 00 -0 SC
- c a 0N N- --
0'r- ~ C a ? N= NL N= ~- 0?h -~
C%. ( C~ J- -C c 0Zc -0 -
Si
- -
I i
-9
>B X
- 0
=1
10
-9
. C'
c >
Li
cI U
2!5
S- t-
>.
*^
0
0
-J
w
N
~s
"3
. -
CC
0-j c6
si
Z-J
LO lNCh eC NC
Lii
I-
0?
C
SI ,- -N -S II 221-
SN. ,- N -C --
C 0
I j.. "- ,, -O.. 0 0 0 .0
NJ- N -- 0 0- C -- -0 -0
NJ
n I Ln
S- cc cj -rM C- C- C :
I -0 I> NO ei r c- c C-C -- C 0-0
ci
~Ii
0
'51
fl
-
I -
I
CO
eC3- -~L~ -- -N
S-0
---
N- -= OC C --
L < -- -
S 1 -- -- -1 -
-i N- -- C --
N -0
Si 5S %.: .= S
J-- -
C'
--- C CC' CO N ~ 2-
^~ ~ ~~~~f -T-------------
rn I,- -tr %Z Lp
1L6. -
^ i-* a- ~> *
CS uS iS -C AS -
=! 5- o
*-) i/i D
I I : E
I- -, -C !
C I
e ~- c~ or
-0
IsO ?
iS l.
3^ C'
CN- C -
N CM Clc
-I
^
C,
s.;
Sl ',
iii
| i
=1
!
=i
- I
gI
ai
ij
a;
'B
t-
-s *
-v -
S- L:
=3 ., s-I
O = I
S- ; I- =
>>(= = o
00 t- ; ^1
O: K
0-
U- L. -7 "- -* -
u- ,,,-- o
: --! -
6rr,,
-II-
c~ -- -~ --
N)- i
u N '- ( N
c -
N C,,.;
S--C
I -C -CU Cr) 3 -
13 2r
- | 00
S-- --
a2 ES S
MI L
C,- l
h ~h 0zC 4C'*
I ; "-= S t"2 = c 0," 0 0,= -
|- -- --
ON ; CN cr c
LO,
0 C- -
I
kZ0 c I- L! C- S- -
- j
m C 5 co t- en
iZ :i CL
I -
2 i
lo
-0o S
C C
Co X
C- .
~h CC? GC~ ~~
~~ ~h ~3L? ~
Oi
V
I
L
b. International Migration:
There are pronounced data limitations which severely hinder a
satisfactory description of women's involvement in international mi-
gration:
1. Sex specific data are scarce and unreliable;
2. When available these data do not distinguish between mi-
gration of autonomous women and migration of dependent
women.
3. These data are based upon the immigrant stock, rather than
emigrant stock, immigrant and/or emigrant flows.
4. Available stock data on immigrants do not reflect migrant
characteristics such as age, labor force participation,
incomes, types of employment and recency of migration.
5. Comparative data on the stock or flow characteristics of
internal and international immigrants are lacking.
6. Most immigration statistics on women do not incorporate
breakdowns by region, ethnic group or national origin, nor
do they always distinguish between tourists and migrants.
This problem is particularly visible with regard to emigra-
tion statistics.
7. When number of migrants are given for a specific time period,
it is difficult to interpret the data. For instance, when
for a given country total migrants for the period 1960-70
are specified as "x", it is unclear whether "x" excludes
migrants from previous periods, whether it is a yearly aver-
age flow or if it is the cumulative stock (excluding returns).
8. Although the data compiled by labor importing developed
countries are more detailed, they cover the migrant female
workers more consistently than they do migrant dependent
women.
9. Even in the case of migration statistics of developed coun-
tries only rough estimates for illegal migrants are avail-
able, and such estimates are not sex-specific.
The general belief that women's participation in international
migration is not significant can now be challenged. Women not only
join the labor migration movements in significant numbers autonomously
but also accompany their families, and join the labor force immediately
upon arrival or later. For instance, between 1960 and 1974, 1,512,200
migrants arrived in the US from just three Latin American countries:
7/
Cuba, Mexico and the West Indies (U.N. Forthcoming). -More than
half (52%.) of thesemi grants were-women. The labor force participation
of the women over 16 years of age was 48%. To give yet another illustra-
tion based upon Table 3, twelve developing countries supplied 108,738
women to three developed countries in the periods specified.
7/ If illegal migration could have been included, this figure would
have been much higher.
Immigrants to Australia, Canada and USA from a
Countries.
Selected Number of Develooina
Sending
Couatry
Egypt
South Africa
India
Lebanon
Turkey
Total # of
immigrants (000)
15.9
20.3
24.6
31.2
16.0
Sex ratio
114
110
107
155
115
Total # of
females (000)
7.4
9.9
11 .8
14.5
7.4
Labor force b
participation-
To Canada
(1970-74)
To USA
(1960-74)
Hong Kong
India
Phillipines
Cuba
Mexico
West Indies
China
Hong Kong
India
Japan
Korea
Philippines
-/ The sex ratios used in
males per 100 females.
2/ As percentage of women
this section are obtained by calculating
over 16 years of age.
the number of
Source: Based upon statistics included in U.N.,Trends and Characteristics of Internationa'
Migration Since 1950, Forthcoming.
TABLE 3:
Receiving
Country
Australia
(1960-73)
38.7
38.1
27.7
349.8
731.9
430.5
167.5
42.6
88.7
63.5
121.4
226.0
114
141
77
87
101
80
89
96
125
29
54
68
18.1
15.8
15.6
187.0
365.8
239.1
88.6
21.2
39.4
43.2
78.8
134.5
I ....
The Handbook of International Data on Women (Boulding et al, 1976)
offers female indices for long term immigration and emigration based upon the
1970 UN Demographic Yearbook. Long term migrants were defined as people
leaving their countries for more than a year during the period 1962-1969.
The index designed by Boulding et al showed the percentage of women among
all migrants. Accordingly, maximum, minimum, and mean female participa-
tion by continent were calculated for 44 countries.
The regional indexes are given in Table 4 to illustrate the extent of
female participation in international migration. For instance, in Africa,
where female participation in international migration is a highly neglected
phenomena in the literature, 34 per cent of all recorded immigrants in
African countries were women. Additionally, 43 per cent of all emigrants
leaving African countries for residence in another country for more than
a year were women.
Although these data are used primarily as a way of illustrating
country differences, the wide variations in the magnitude of international
migration may further hinder the usefulness of the data. For instance,
while in Trinidad 71 per cent of all migrants are women, they add to a
total of 50 women. The absolute number of women involved in Cyprus is
9 while in the United States it is 231,825. The Federal Republic of
Germany, on the other hand, has a relatively low female participation
in immigration (36 per cent), but such a rate involves 229,326 women.
There are significant regional variations in women's participation
in international migration. However, such variation is reduced when the
recency of a given country's involvement in such migration is taken into
r-_ en co 0m
CDi 0
or c<
LO L
*A *
i- -
mL LA
(\i p
in co
Ln ,
0 co
9-. r. a*
L r- CD
0 0 0
o k
U U
-nCM
r S.- L
t o
c -j LaJl
V, ID1
C
'Vr
0)
3
Ln 0 C
CM %0 r-
tr qr 4cr
en R> LO
f- aN C)J
o.- o' 0m
cr LA o.
oo m \D
ko CC
*0
Ln
%0
Ln
I0
cu
un
.9- to
4-Lu
0)
I->
oB
rer
1-re
0)
ere
m
S-
r- I
3
01
r
41
c,
ra
0.)
m
0
cn
4I-
0
C
0
0.e
0 u
lz
CM
.9
m c
> E
cm L
Cl L
cv U
o Uj
x
Cl 0.
o ?O
S-
r S-
C,
'V
0 )E
9- 3
10 -
0 0)
o9- -g
y. CQ
0 L"
'0
01
*i- '-0
01-) C
.C.C
(0 'O
= c'
'V'
a. -c o
0 0
cI re
ro .C
0
.1 VI
3 C
a)
0.) 0.'.=-
S ..
o -^0
/ 'Ve^*-
consideration. During the initial phases of labor movements from one
country to another, female participation is often low. At subsequent
phases, however, this participation increases due both to autonomous
and to dependent female emigration.
A comparison of immigration and emigration data yields significant
differences in the magnitude of international population movements as a
whole (U.N. Forthcoming). The most important reason for the differences
in these statistics is that immigration figures refer to the stock of
migrants while emigration figures are expressed yearly. As an example of
these differences, in total there are approximately 215,000 Turkish
migrant women workers abroad, whereas in 1978 the total number of female
emigrant workers were just a few hundred.
As far as the push factors are concerned, there are no basic /
differences between the internal and international migration of women.
Economic needs are the main cause of departure for external labor markets.
The examination of the pull factors, however, will yield a critical
difference: often, the migration of women, whether employed or dependent,
is controlled by explicit policies of the host countries. Depending upon
these policies, the extent and nature of the labor force participation of
immigrant women differs, both cross-sectionally and historically.
In West Germany, as well as in the remaining labor importing countries
of Western Europe, more detailed statistics (surveys and censuses) are
available for migrant men and women of different nationalities. These
statistics reveal population characteristics of the migrant groups (sex,
age, marital status), as well as more detailed characteristics of the
workers (sex, age, length of residence, sector and type of emp1:-:;ent,
unemployment status, housing conditions, location of child, motivations,
return patterns and intentions). Indeed, many of the survey statistics avail-
able for foreign workers in Europe incorporate a comparison of the two sexes.
But the main focus is on the workers rather than on the population in general.
For the purpose of this report, however, we have not gone into a survey of
this literature. Nor have we compiled the various statistics on foreign
populations available through the host countries. Similarly, we have not
reviewed the existing immigration literature on the foreign populations in
the U.S., Canada, Australia, and New Zealand. As these provide very extensive
coverage of women migrants, further research utilizing the resources should
be pursued. in what follows, we present an effort to illustrate some of the
major inter- and intracontinental trends in international migration of women
as a framework and starting point for further research.
Having reviewed some of the general trends and data problems a very
brief description of the regional patterns of female international migration
will be given.
Africa. In Africa, the greatest number of emigrants originate from
North Africa and, particularly, from the Mahgreb countries--Algeria, Morocco,
and Tunisia. An estimated 1.4 million people from these countries were
living in Europe in 1974 (U.N., Forthcoming). The sex ratio for Mahgreb
migrants was high in all age groups over 15,'as is typical for Africa. In
some of the receiving African countries, the effects of male dominated migra-
tion can already be seen. For instance, the recent flow of workers from
Mahgreb countries to Libya raised the sex ratio in the latter country from
108 in 1964 to 114 in 1973.
8/ The sex ratios have been improving in Africa's emigration countries. For
instance, Mauritius has had significant emigration during the 1960s and 1970s
which was characterized by a high sex ratio during the initial phases but which
proceeded into family migration in the latter half of the 1960s, thus improving
its sex ratio. This indicates a high rate of female migration in the 1970s.
Intra-continental migration in Africa has greater significance than
inter-continental migration. Because African states often divide tribes,
much of the migration appears to be inter-regional and, sometimes, intra-
tribal. Although the migratory movements follow rules set during the
colonial times, some of the new states have set new rules restricting
the type and duration of cross-boundary migrations for economic and
political reasons. These seasonal and fluctuating movements are generally
male dominated. The sex ratios for some of the African States
show that Uganda, Ghana, South Africa, Southern Rhodesia and the
United Republic of Tanzania receive the largest number of migrants.
Malawi and Togo, where the sex ratio of the migrant stock favor women,
are exceptions to the African pattern (Table 4a).
Because individual countries follow different practices vis-a-vis
immigration, age specific sex ratios manifest visible differences. For
instance, the 'ex-age pyramid of foreign-born Africans in South Africa
provides an example of labor migration in an extreme form with practically
no accompanying family members... Since the custom is to recruit male
workers for short-term contracts and then replace them with new migrants,
the high percentage of male foreign-born in the 20-24 and 25-29 age old
group are not surprising." (UN, Forthcoming; 220)
The percentage of immigrants under 15 years of age from other African
countries was 16 in Ghana, 10.2 in Southern Rhodesia, 27.1 in Zambia and
1.2 in South Africa. These figures clearly show a family oriented migra-
tion to Zambia as opposed to a labor oriented migration to South Africa
(Table 4b).
TABLE 4a: Male and Female Immigrants in Selected Countries around 1965
in in Males per
(000's) (000's) 100 Females
Males Females
Africa (foreign born populations)
Gambia (1963) 20.7 14.9 139
Ghana (1970) 210.8 139.0 152
Kenya (1969) 86.6 72.1 120
Liberia (1962) 19.6 12.0 163
Malawi (1966) 143.5 151.0 95
Senegal (1960) 99.0 70.7 140
South Africa (1970) 443.0 47.1 941
South Rhodesia (1969) 240.5 99.0 243
Swaziland (1966) 20.4 19.2 106
Togo (1970) 69.3 74.3 93
Uganda (1969) 462.1 289.6 160
U. K. Tanzania (1967) 244.5 207.2 118
Zambia (1969) 104.2 90.8 115
SUB-TOTAL 2,161.2 1,266.9 127
Immigrants per Country 166.2 97.4
TABLE 4a (Continued)
in in Males per
(000's) (000's) 100 Females
Males Females
Asia:
Bahrain (1971) 26.5 11.3 234
Hong-Kong (1971) 874 842.1 104
Kuwait (1971) 244.4 146.9 166
Malaysia (1970) 422.4 342.0 123
Nepal (1971) 123.5 214.0 58
Singapore )1970) 276.0 252.1 109
Sri Lanka (1963) 152.5 95.7 159
Thailand (1970) 213.7 135.9 157
SUB-TOTAL 2,333.0 2.040.0 139
Immigrants per Country 291.6 255.6
in in Males per Females froi
(000's) (000's) 100 Females other Latin
Males Females American
Countries
Latin America
Argentina (1970) 1,151.8 1,041.5 111 127.5
Brazil (1970) 671.4 557.8 120 35.0
Venezuela (1971) 318.9 277.6 115 120.0
SUB-TOTAL 2,142.1 1,876.9 114 282.5
Immigrants per
Country 714.0 625.6 94.1
TABLE 4b: Age and Sex-specific Long-term Emigration from Selected Countries
Ages
male
female
males
per 100
females
Africa
Angola (1972)
Botswana (1971)
Mauritius (1971)
South Africa (1970)
South Rhodesia (1970)
Total Females (14,480)
m
f
m/f
m
f
m/f
m
f
m/f
m
f
m/f
m
f
m/f
0-14
1,306
920
141
300
150
200
273
178
153
1,429
1,256
113
658
652
100
3,156
15-49
2,319
2,415
95
6,290
1,780
353
719
878
81
2,627
2,402
109
1,724
1,769
97
9,244
Latin America
Costa Rica (1971)
Trinidad and Tabago
(1970)
Total Females (38,047)
865
745
116
1,160
420
276
86
127
67
458
503
91
259
285
90
2,080
m
f
m/f
m
f
m/f
5,534
4,306
128
1,050
1,030
101
5,336
67,087
24,093
278
1,990
2,230
89
26,323
8,120 -
6,128
132
220
260
84
6,388
Source: (Compiled from UN (n.d.) Emigration Statistics)
a) Only countries indicating long-term emigrants are chosen
b) Developed countries have been excluded
On the whole, emigration from Africa is male dominated. However,
sex ratios are lower among North African migrants in European countries
and the rate of female labor force participation is highest among them.
There is a concentration of migration in France with migrant women
occupying the low skilled factory and service jobs. Informal sector
domestic employment is also widespread. Few African women migrate to
Libya or to other oil producing Middle East countries, either as workers
or as dependents. Migration to other African countries is also res-
tricted, both in volume and in nature: when women migrate, it is often
as dependents rather than as workers.
Latin America. Five types of international migration have been
observed in Latin America (Breton, 1976). First, there are frontier
workers who are fully integrated into the labor market of a foreign
country but still live at home. Second, there are seasonal agricultural
workers migrating for several months at a time--often irregularly,
depending upon labor needs. Third, there are short-term temporary
migrants who leave their homes for several months each year. Fourth,
there are long-term temporary workers who usually work in a country
for several years under special bilateral agreements and, fifth, there
are long-term or permanent workers who settle with their families in
a country while maintaining their own nationality. Over all, the
migration of unskilled workers leaving one rural area for another
predominates.
Intra- and intercontinental migration in South America, which has
been small in magnitude compared to overseas migration, has increased
since 1950 as a result of population growth, widening disparities in
economic development and/or improved communications. "In 1975, the
total number of intracontinental migrants and their families settled
or working abroad is thought to have been of the order of 5 million,
of whom rather more than 3 million were migrant workers properly
speaking--of both sexes and all ages--around 400,000 were frontier
workers, and over 1,500,000 were members of their families." (Breton,
1976:340).
Focusing on the total international migrant population in Latin
America, it is observed that they are predominantly in the 30-40 age
group and they are older than both the internal migrants on the con-
tinent and the international migrants in Europe. Among the inter-
national migrants, 55 per cent are men. Additionally, there were
country differences in the sex ratios favoring, for instance, females
in Venezuela. "There is often a family marked propensity to emigrate
among the unmarried, but in the high emigration countries and those
where the phenomenon assumes the proportion of agricultural settlement
in the true sense the great majority of migrants are married." (Breton,
1976:344). As expected, the education and training of international
migrants are lower than the national averages of the host country, but
their labor force participation is significantly higher than that of
the natives.
Many of the Latin American countries, and particularly Argentina
and Brazil, received large numbers of migrants of European origin (Table 4a);
These migrants differ from migrants from other Latin American countries in
many respects, including their sex composition (U.N., Forthcoming). Although
such comparisons will not be detailed here, we should point out that
there are significant country differences. In the 1970s Argentine males
predominated among migrants from other Latin American countries and
females predominated among recent Spanish immigrants. The reverse holds
for Venezuela for the same period. Again, these could be characteristic
of certain periods and not necessarily of consistent trends.
Another significant international migration pattern of Latin
America is migration to the USA. This pattern has changed its character
over the years from a male dominated movement to a female dominated one.
The participation of Latin American women in all types of inter-
national migration is a significant and widespread phenomenon, especially
when compared to the behavior of females in the Middle East and Africa.
In addition, the labor force participation of Latin American women in
other countries is also high. The jobs held by these women vary with
the type of migration and with the type of structure of the host economy.
Asia. International migration which originates from Asia also has
a long tradition. Countries such as India and Pakistan have been con-
tinuous sources of migrants for many decades. Studies reveal little
about the Asian migrant women, their numbers, characteristics, and
problems. It seems, however, that a large bulk of the Asian emigrant
women leaving their countries for Europe, Great Britain, the USA, and
Canada become permanent settlers in these countries (Table 3 ).
Added to these traditional migrant groups is a significant number of
women from Asian countries affected by wars. Both of these two broad
categories of women have been left out of this paper, since a satis-
factory coverage of their problems would require intensive research
in different directions than those of this report.
Briefly, the predominance of males is also observed among Asian
emigrants to other countries, especially among the migrants from
India, Pakistan, Bangladesh, Hong Kong, and Sri Lanka. The reverse is
true for recent migrants from Malaysia. However, even among the male
dominated movements, the female ratio has been increasing. The sex
ratios of immigrants in Southeast Asia, for instance, have improved,
particularly after the limitations these countries have put on immigra-
tion. Many Asian nations accommodate large numbers of female migrants from
other countries. For instance, in the 1960s, there were a total of 2,040,000
female immigrants in seven Asian countries, an average of 251,600 in each of
them (Table 4a). The sex ratios were high, with the exception of Nepal. The
index of femaleness in long-term emigration also showed a far greater variation
in Asia than in other continents (see Table 4).
However, research which goes beyond simple ratio statistics is
particularly needed for Asia, as the intensity of inter- and intra-
continental migration of many Asian countries has already reached
significant proportions and is likely to gain further momentum as
political pressures increase.
Middle East. Recent international migration trends within the
Middle East are due largely to the 1973-74 oil price increase and the
concomitant increase in the demand for labor within the oil rich
countries of the region (Choucri, 1977). In six oil producing countries
in 1975, there were a total of 1,236,800 migrant workers from Arab
countries, 291,200 from Asia, and 86,900 from Iran, Turkey and Africa.
Including the Europeans (2 per cent of all migrant workers), these
oil producing countries employed over 1.6 million workers from other
countries. "In 1975, there were over two-and-a-half million Arab
workers and their dependents living in Arab States... In early 1975...
there were 1,570,000 Arab workers living abroad... and over 2,500,000
migrants for employment in the Arab Near East." (Birks and Sinclair,
1979:1).
Migration to the oil producing countries is predominantly male
oriented. The total sex-ratios of migrants to Bahrain and Kuwait is
234 and 166 males per 100 females, respectively. An extremely large
sex ratio is found for Iranians (978) in Kuwait, which indicates that
almost no Iranian women migrated to this country. Also apparent is
that few females from Oman have moved to the nearby Gulf States, as
indicated by the high female sex ratios. (See Table 5). Birks and Sinclair
(1977b)give several reasons for the low number of females migrating
to these countries from Oman. First, housing is difficult to obtain
for families in oil producing regions and, when found, is expensive.
Sex-SDecific Migration to Two Oil Producing Countries
Number of Number of
Receiving Sending Male Migrants Female Migrants Males per
Country Country (in 000's) (in 000's) 100 Females
Bahrain
India
Iran
Oman
Pakistan
Total
Kuwait
Egypt
India
Iran
Iraq
Jordan
Lebanon
Oman
Pakistan
Syria
Total
4,0
3.5
3.5
3.3
150
217
733
162
234
133
157
978
152
118
126
555
179
121
-2.7
1.6
1.3
2.1
11 .3
13.0
6.8
3.6
15.5
67.8
11.2
2.2
5.3
10.0
147.0
26.5.
17.0
10.5
35.5
23.6
80.0
14.0
12.4
5.4
17.2
277.0
Source: U.N. Trends and
forthcoming. P. 239.
Characteristics of International
MiaratioW Since ].950.
- -
--- ti n Si c ]-.950,
Table 5 :
Second, women generally do not migrate to areas which are close in
proximity while males more often commute to work on a weekly or short-
term basis.
While the sex ratios for Jordan, Lebanon and Egypt indicate male
dominated migration trends, they are far more equitable than other
Arab countries. In fact, according to a USAID report (1977) in 1975
there were as many Palestinian and Jordanian females as males in
Kuwait. This is due to the fact that migrants from Jordan and Lebanon
are more apt to be permanent migrants and, therefore, bring their
wives and children with them.(Clarke, 1977). Family movement to Kuwait
also occurs among Egyptians, but it is more common among professional
and highly skilled migrants than among lower skilled migrants who are
more likely to migrate alone (Birks and Sinclair, 1978a).
Migrants from Jordan, Lebanon and Egypt are usually preferred
because they are generally better educated, highly trained and speak
Arabic (Birks and Sinclair, 1977a). However, the oil-rich countries
are increasingly becoming concerned about the high percentage of non-
nationals within their borders and, therefore, are trying to dis-
courage permanent migration as well as the migration of dependent
family members (Clarke,1977). For this reason, more Asians are being
recruited to work temporarily in the Gulf States.
Between 1971 and 1977, the number of workers from Asia increased
by 276 per cent. In this latter year, two-thirds of all foreign
workers in Bahrain were from Asian countries--particularly from India
and Pakistan.
Saudi Arabia has increasingly relied on countries such as Japan
to send skilled labor for short periods of time in an attempt to
discourage migrants from remaining permanently. In the United Arab
Emirates in 1976, Asians accounted for 69 per cent of the expatriate
work force--again, mostly from India and Pakistan, but some from other
Asian countries as well. The number of Asian workers also increased
substantially in Kuwait in recent years, albeit, not nearly to the
degree found in either the United Arab Emirates or Bahrain.
There is also heavy migration among North African Arabs to Arab
countries other than the Gulf States. For example, Egyptians are
increasingly migrating to Jordan (approximately 6,000 were expatriates
in 1977) to replace the Jordanian work force who have migrated to
Saudi Arabia and the Gulf States. Interestingly, the sex ratio among
Egyptians migrating to Jordan is 100. Apparently, Egyptian women are
more likely to find employment in Jordan than elsewhere, which is an
incentive for them to migrate (Birks and Sinclair, 1978a). There is a still
larger trend of migrants leaving Egypt for the Sudan. Again, the sex ratio
is equitable (101.4). Birks and Sinclair (1978a) state that the
low cost of travel is an incentive for professional and skilled workers
to bring their wives and families with them, even though employment
opportunities for women are low in this country. Most Egyptians,
however, migrate temporarily to Libya (between 275,000 and 380,000). There
is no information regarding the extent to which these migrants are male. The
movement from Tunisia to Libya is also male dominated, although smaller in scale.
There is also migration from the Middle Eastern countries to other
regions, particularly to Europe. Among the various Middle Eastern countries,
Turkey has exported greater amounts of labor to Western Europe than others.
Outmigration from this country started in the early 1960s, reached its peak in
1972-73 and decreased significantly after the so-called energy crisis of
November, 1974. Female participation in Turkish international migration is a
relatively recent phenomenon as compared to that of males. Because the immigra-
tion of dependent family members was discouraged for much of the period between
1960 and 1979, the labor force participation rate of Turkish women abroad has
been very high. By the time the migratory stock abroad reached its peak in the
early 1970s, almost a quarter of all Turkish workers were women. Focusing on
the yearly inflow the sex ratio has changed sporadically from as low as 6.6 to
as high as 39.5 (See Table 6). But information in the stock about that per-
centage of women in the labor force increased steadily from 6.8 per cent in
1960 to 26 per cent in 1975 (Abadan-Unat, 1977). In addition to the working
women abroad, the numbers of dependent wives also increased, and the great
majority of the workers are now accompanied by their wives, whether or not
they work. In September 1977, there were 1,118,000 Turks in West Germany
of whom 443,100 were females. Although net immigration of the Turks to
this country has been negative since the end of 1974, that of females has
been positive due to family reunions. (Statistiches Bundesamt, Statistiches
Jahrbuch 1978, Wiesbaden, 1978). For instance, in 1976 the net immigration
of Turkish men to West Germany was -25,844 and of women +765.
Table 6: Composition of Turkish Migration Stream by Year and Sex
Number of Percent Migrants Percent Migrants
Year Migrants who are male who are female
1966 34,410 76.9 23.1
1967 8,947 60.5 39.5
1968 43,097 73.8 16.2
1969 103,005 80.0 20.0
1970 128,395 84.0 16.0
1971 87,563 83.9 16.1
1972 84,589 78.1 21.9
1973 134,934 80.1 20.2
1974 19,073 93.4 6.6
1975 (half year) 1,796 84.4 11.6
Source: Migration News, no. 4, 1976.
International migrant women, when compared to all other categories
of migrant women, are more visible-- economically, socially, and politically.
Their problems can also be more clearly identified. However, there are
significant differences between regions in this regard. The visibility of
migrant women from the Middle East and Northern Africa in the Western
European countries has been amply documented. Such is also true for the
inter-continental movements of the Asian and Latin American women. Yet, as
we have pointed out, women's share in intra-continental migration is steadily
increasing. Whether such a trend in international migration makes the
immigrants less visible is yet to be studied together with its implications.
However, the greater visibility of international migrants should make this
group particularly attractive for policy makers.
Because there are many different types of international migration
(seasonal, fluctuating, permanent, temporary, inter- and intracontinental)
it is difficult to make generalizations concerning the socio-economic charac-
teristics of the female migrants. A focus on the migrant stock in Europe
from the Middle East, North Africa, and Asia shows a heavier concentration
in the 20-30 year age group. The age selectivity of the host labor markets
is also reflected in some of the small scale studies of returnees. In this
type of migration is predominantly a labor migration, men and women are
allowed to enter the host countries at their early working ages and are
pushed out of the market before they become 40. For instance, the examina-
tion of the age pyramid for Turkish workers abroad shows that less than 5%
of the immigrants are over 45 years of age. The dependent population below
15 is also small in comparison with the working age group. The age
structure of the returnees as compared with the stock reveals an older
population.
The limited evidence on international migration yields stronger
selectivity to be at play. For instance, international migrants in
Europe, as compared with internal migrants of a given country, are younger,
more educated, more skilled, with greater working experience, with
greater exposure to urbanization (i.e. originating to a lesser degree
from rural areas). The selectivity is particularly strong for women,
not so much when compared to migrant men but especially, when compared
to the overall characteristics of native women (Kudat, et al, 1976).
The general observation that migrant women are greater participants
of the labor force than native women also needs further qualification
in the case of international migration to Western Europe or other developed
areas of the World where female participation in the labor force is high.
For instance, what distinguishes migrant women from native women in
Europe is not so much their rate of economic participation (which is lower
in the former case for some ethnic groups) but the type of participation.
In both cases, excluding illegal migration, labor force participation is
mainly in the formal sector. But migrant women go directly into the
lowest paying, lowest skill industrial and service jobs, with little or
no previous training and with little on-the-job training opportunities.
Those illegal migrants and wives are also taking marginal domestic
employment. Studies based upon the European experience also show women
to be the first targets for dismissals in economic crises. Moreover,
cultural and language barriers, as well as their lack of previous work
and unionization experience, make it harder for women in international
markets to join labor struggles or to obtain unemployment benefits even
when they are legally allowed.
The discussions presented on the effects of male migration on
women left behind are largely applicable to the women left behind by
international migrants. Although the remittances may be greater, it is
unclear whether they are able to receive the savings directly. The
effects of greater distance and longer periods of separation on the
regularity of remittances are also unknown
Nevertheless, the migration of women across national boundaries
is thus becoming a widespread phenomenon even in continents where the
rate of female participation is low. The available information also
indicates the increasing labor force participation of international
migrant women. However, there is very little information on the living
and working patterns of these women, and little comparative evidence
showing the effects of the immigration and emigration patterns of these
countries. Since the magnitude of international population movements
(which primarily involves low skilled labor) is likely to increase and
since problems of integration for migrant women are reported to be
numerous, research and policy which focuses on international migrants
is long overdue. Such research should consider the implications for
development of the return migration of men, and, especially, of women.
These women contribute to their countries economies through their
66
remittances as men do. They also help in the diffusion of new technology
and other innovations. However, women differ in a critical dimension
from men; the rate of return migration among women is lower and consequently,
many of the effects of international migration that emerge as a result of
returnees to either the country of origin or, more importantly, to the
community of origin, are less likely to be manifested differently for
women than for men.
IV. WHO ARE THEY?: CHARACTERISTICS OF WOMEN MIGRANTS
Age
A universal feature of migration is that it varies with age and,
particularly in the case of women, with stages in the life cycle. A
review of country-specific studies enables us to identify three age
groups at which mobility among women appears to be the highest: the
adolescent, the early twenties and the over 50+ age groups. These
differ somewhat from the general age patterns of male migrants; for
women, the age patterns also vary according to destination.
It is important to note that the age at which women migrate and
the age distribution of the stock of migrants in any given area are
two different concepts, but available data are not detailed enough
to allow us to distinguish between the two. Generally, the older a
migration process, the more similar is the age distribution of
migrants and native residents. When a migration stream is new, the
age composition of the migrant stock is skewed; as more people stay
in the receiving areas the distribution becomes more even.
Those migrating to urban areas, particularly to capital cities,
are predominantly women in their late teens and early twenties, which
may reflect the fact that young (single) girls can more easily secure
employment--as domestic laborers--than can young men of the same age.
This pattern is often thought of as the Latin American experience, but
that is only because there is more data available for that region. As
countries in other regions begin to focus on the migration issue, it
is highly probable that similar patterns will emerge.
The overriding tendency for female migrants to be younger than male
migrants is borne out by studies in Colombia, Jamaica, El Salvador,
Brasil, Bangladesh, and Thailand. For instance, in Colombia in 1970,
56 per cent of all migrants to the capital, Bogota, were women, and
young women outnumbered men in the 10-19 age group by a ratio of ten to
six. This age group accounts for 38.4 per cent of all females in
migration during 1970-75 (Lubell and McCallum, 1978). In fact, women pre-
dominated in all migrant age cohorts except the 30 to 39 age group, and
54 per cent of all migrants to other urban areas were women. In movements
to rural areas, by contrast, men predominated in all age groups, without
exception (Martine, 1975; Lewin and Romani, 1977). In rural Mexico, women
typically leave for the city between the ages of 15 and 19 (Weist, 1973).
In Chile, 50 per cent of the women coming into Santiago were between the ages
of 14 and 25; among these 70 per cent arrived on their own (i.e., were not
dependents) (Elton, 1974). In the capital cities of Jamaica and El Salvador,
the mean age of women migrants is between 15 and 19; for men, it is 20 and 24.
A study in Brazil that traced migration streams to six major cities (metro-
politan areas) showed that women migrants in Sao Paulo were typically three
years younger (20.3) than the men (23.3).
Most female migrants to urban Bangladesh are between the ages of 15 and
19, and in that age range they outnumber the male migrants. The female rate
continues to be high for ages 20-24. At most other ages, the number of male
migrants is greater (Ruzika and Chowdhury, 1978).
Perhaps a better way of demonstrating the significance of the
effect of the youthfulness of a migrant population upon the age structure
of an area is to examine the potentially active age cohorts. Lubell
and McCallum (1978) in comparing the age distribution of native and
migrant males and females in Bogota, Colombia, point out that in the
potentially active 15-29 age group, 40 percent are migrant women and
only 20 percent are native resident women. There are also twice as
many migrant women than native women in the 30-44 age group.
Goldstein's analysis of the Thai data (1973) allows for further
specification regarding the age selectivity factor in relation to
different types of mobility experienced by men and women. In general,
Thai women migrate at an older age than do Latin American and Bangladesh
women--that is, in their mid- and late twenties. In some instances,
women migrants are older than their male counterparts. It is not possible
to ascertain whether this later age reflects migration taking place
with families instead of migration by individuals.
Among rural-urban migrants in Thailand the highest percentage of
women are in the 20-24 age group, the next highest are 25-34, and after
that the number of female migrants drops drastically until older ages.
The same pattern is observed for men.
Peak migration from rural areas into the capital city area, Bangkok,
for men is between the ages of 20 and 34. It is highest for women in
the 15-24 age group (Goldstein, 1973). A breakdown by marital status of
female migrants into Bangkok shows that 44 percent are single, 52 percent
are married, and 8 percent are widowed or divorced (Piampiti, n.d.).
However, women who migrate from rural areas into other urban provinces
are usually older, and there the peak migration age is 20-24.
With respect to inter-urban mobility, the levels of female
mobility are 31 percent for ages under 15 and 56 percent for ages
20-24, declining to 25 percent among those 45 and over (Goldstein, 1973).
Inter-urban migration occurs most frequently for Bangkok women at ages
25-34; for women in other urban areas, the peak is reached earlier,
at ages 20-24. Goldstein interprets such movements as reflecting
change in residence at marriage. Women in Bangkok tend to marry at a
later age than women in other urban locations, which would explain
the differentials in the peak ages (Goldstein, 1973).
There is increasing evidence that women migrate at both extremes
/ of the age hierarchy. In some countries they are more numerous than
men in both the youngest and the oldest cohorts; in some they outnumber
men in the oldest cohorts only. The female/male difference is parti-
cularly accentuated in the 50 and over cohorts, To
migrate at this age is a distinctive female characteristic; it is
particularly striking in moves to capital city areas (Colombia, Mexico,
Nigeria, Thailand).
In contrast to the men migrating to Ibadan, most women migrating
S on their own are nearing 50, are widowed or divorced, and have functioned
as heads of household (Sudarkasa, 1977). Migration rates clearly
increase among rural women moving to Bangkok at the upper end of the
age hierarchy, i.e., for women aged 65 and over. The same is not true
for men. Older women (65+) also display high rates (25%) of intra-
urban mobility. The patterns reflect movements associated with widow-
hood: women leave their homes to live with their children. Female
migrants to Bogota outnumber males in all age groups except the 30-39
cohort; but the sex differential is particularly salient in the 50 and
over age groups. These differentials in the older age groups are
interesting. It has been speculated that they are due to mortality
differentials by sex in areas of origin and that women, particularly
after they are widowed, join family members who had previously migrated
or seek employment in domestic service when their children become
independent (Martine, 1975b).
Marital Status
The most common practice in the migration literature has been
to treat the status of female migrants as "accompanying wives"--assuming
that women were involved in family migration only--and/or to emphasize
the temporary migration of young, single, economically active women
who show high participation in urban domestic service occupations. The
autonomous migration of women has been largely ignored. In order to
research the extent of its magnitude, information sources are needed
that identify women migrants by age, marital status, and fertility
both in their current place of residence and at the time they migrate.
Such data would yield insight into:
a) the extent of autonomous female migration; and,
b) the influence that marital status per se may have either in
inducing or deterring migration, depending on the stage of life
cycle involved.
On the basis of the limited information that is available,
however, we can make some statements about the marital and household
status of women migrants:
1) Men and women who migrate to urban areas are predominantly
young and single. The larger the city of destination, the greater
the tendency for women migrants to be single. Among those who move
to rural areas, the men are again usually single, but the women are
more likely to move in this direction when accompanying their spouses
(Martine, 1975b).
2) Recent findings on the participation of women aged 50 and over
in the migration process are significant. Widowed women in particular
surface as highly mobile in both rural-urban and inter-urban migration.
The same is probably true of divorced women, who may in fact migrate
at an earlier age than widows. Unfortunately, census categories often
combine widows and divorced and separated women into one category;
because the widowed tend to be more numerous in absolute numbers
than the divorced (or separated), the tendency is to single out the
group's behavior pattern as typical of widows only. At the same time,
in countries where data is disaggregated by marital status there are
very high percentages of divorced women among migrants. This may only
be partly due to higher incidence of divorce in cities, but it also
reflects the predominance of divorced women in internal migration
(Youssef, 1973).
3) Information is not widely available on the distribution by
sex and migration status of heads of household. Some of the data
suggests that female headship, whatever the reason for it, is a con-
dition sufficient to bring about the migration of women. There is a
strong tendency for female heads of household--especially those in
older age cohorts--to be associated with migration to urban areas in
Mexico (Weist, 1973), Thailand (Goldstein, 1973), and Ghana (Sudarkasa,
1977). Yet there are other indications in the younger age cohorts that
single women (mothers) who head households may also play a vital role in
the rural-urban migratory process. There is specific reference to the
exit of single rural women because of pregnancy (Castro et al, 1978).
4) Studies among young single female migrants in the city indicate
that the majority do not intend to return to the point of origin (Castro
et al, 1978). This is reflected in several instances in the low repre-
sentation of single women in return-migration.
5) Capital city areas are particularly attractive to women (and
men) who do not have family ties--the single, the widowed, the divorced,
and th separated --although by far the largest number of migrants are
10/
single.- A study in Bangkok shows that 44 per cent of all women migrants
were single, eight per cent were divorced and widowed. The young unmarried
come to the city for socioeconomic reasons and rely on family and friends
who already live in the city (Piampiti, n.d.). In Kingston, Jamaica, the
majority of women migrants are single. They have come to the capital alone
9/This would have enabled us to estimate the magnitude of female heads of
household as participants in the migratory flow.
10/Both the Latin American and the Thai patterns indicate that migrants
into capital city areas have higher proportions of unmarried (single,
separated, or widowed) men and women than does the resident population
of either the capital city or other urban areas (Martine, 1975: Goldstein,
1974).
and depend on the income they earn for subsistence. They remain single,
even after giving birth to several children (Standing, 1978d).
Of the women aged 14 and over who go to Santiago, 49 per cent are
by themselves; of those who go to Lima, 62 per cent are alone. Most (70
per cent) are single and, in each city, are dependent upon what they earn
for survival (Elizaga, 1972). The smaller the place of origin, the greater
the probability that single women migrate to the capital by themselves.
Among migrants coming from areas with less than 5,000 inhabitants, 59 per
cent in the case of Santiago and 62 per cent in the case of Lima had come
on their own (Elton, 1974).
Education
"It is frequently assumed that higher education per se may
serve as a stimulus to migration, both through the greater per-
ception of new vistas as a result of more education and because
of the need to move to a different location where special skills
resulting from more education can be better utilized." (Goldstein,
1973).
The data available do not support the assumption that level of
migration is directly related to level of education insofar as women are
concerned. Rather, they point to a low level of selectivity, probably
because most migrants from rural areas, particularly those who are
women, have not had much education.
Two types of data sets are available, although each is limited in
its country coverage. There is some information on the educational
status of women migrants and how it compares to male migrants and to
native women in the receiving area. The general picture that emerges
from these comparisons is that the average education of migrants is
low and that for women migrants it is lower than for male migrants.
While the difference between the education of recent migrant males and
that of native urban resident males is slight, discrepancy between the
two corresponding groups of women is considerable. Specific country studies
in Chile (Bustamante, 1978), India (Zachariah, n.d.), Brazil (Castro
et al, 1978), Indonesia (Sethuraman, 1976), and Turkey (Kudat,et al, 1976)
confirm the educationally disadvantaged position of the female migrant
in relation to her male counterpart. Among lifetime migrants in India,
58 per cent of the women, as compared to 35 per cent of the men, are
illiterate. Among those who are educated, male migrants have received
significantly more education than women migrants.
Singh (1978a:352) reports for India that:
"...national level data regarding educational levels and
work force participation rates of migrants reveal that the
majority of female workers are illiterate and that there are
practically no jobs pursued by women at all between those
which require no education and those which require high levels
of education. Significantly,the illiterate, unskilled migrant
women of India seem to have greater ease in finding employment
than those with some education."
Singh further points out the influence of education on work among
poor Indian migrants by comparing educational levels of workers and non-
workers. Among males, literacy rates are similar for workers and non-
workers (71 per cent). But among women, workers had lower literacy
levels (46.5 per cent) than nonworkers (65.2 per cent). The percentage
of male migrants who had acquired literacy skills outside of the formal
educational system was higher for both workers and nonworkers (25 per
cent and 34 per cent, respectively) than it was for both categories of
women. For example, only 7.9 per cent of working women had become
literate through non-formal training, as compared to 28.4 per cent of
nonworking women. The data consistently show that it is the women with
the least amount of education who are most likely to work (Singh, 1978a).
The explanation for this pattern is probably not that these women are
more likely to find work, but rather that they are willing to take any
work that is available.
In Lebanon, data on the educational level of migrants coming into
Beirut and its suburbs showed significant discrepancies between the
sexes. In 1971 among migrants aged 15-44, 20 per cent of the men and
47 per cent of the women had not attended school (Tabarrah, 1976). In
a study conducted in Jakarta in the early 1970s, it was noted that 75
per cent of the migrants had less than six years of schooling, the rest
wereilliterate. Women migrants had far lower educational levels than
men, but this did not affect their employability or their perception of
the beneficial aspects of migration. As was found in India, the lower
educational level of migrants in Jakarta, the higher the probability
of their working, and the more positive the feeling that they were
better off than before (Sethuraman, 1976).
The literature also points out the discrepancy in educational
standards between migrant women and native women in the urban receiving
area. Again, it is the migrant women who have an educational dis-
advantage. By contrast, the comparison between educational levels of
male migrants and those of males who are urban residents shows only
slight differences or none at all.
In Brazil sex differences in education are not marked among urban
residents; in fact, in some areas, education levels for women are
distinctly nigher than those of men. Among the low-income urban classes,
women seem to have an educational advantage with respect to exposure to
and/or completion of primary and secondary schooling (Castro et al, 1978).
A comparison of educational levels of migrants and those of the urban
resident population in various metropolitan regions of Brazil showed
male migrants to be at a slight disadvantage to male residents only in
Sao Paulo, whereas the discrepancy between the two groups of women was /
considerable (Castro et al, 1978; Elton, 1974).
In two villages in Ecuador, in which migration was found to be
positively related to education, the educational level of migrant males
and urban males was found to be roughly similar. Migrant women, however,
had received significantly less education than urban women in the receiving
area (Scrimshaw, n.d.). A comparison of the level of schooling of migrant and
resident women aged 25-35 in Kingston, showed that 73 per cent of the
migrant women, but only 50 per cent of the nonmigrant women,had received
only one to three years of education. Fifty per cent of the urban resident
women and 27 per cent of the migrants had had nine or more years of
schooling (Standing, 1978d). In Beirut, the illiteracy rate of womcn
migrants aged 15-44 was 47 per cent as compared to 25 per cent among the
women urban residents (Tabarrah, 1976).
In her analysis of Chilean data, Herold(1978) distinguishes between
types of migrants and specific areas of urban destination and challenges
the assumption that female migration in Latin America is characterized
by low-status women moving to major cities where they become prime
examples of social and economic marginality. Her data indicate that this
V pattern applies only to migration to the capital and to some other
cities; it is not characteristic of migration to all urban areas,
particularly in more recent years. When educational levels are con-
trolled for rates, there is a positive association between level of
education and female rates of migration for all recent migrant types
to urban destinations. In the aggregate, total recent migrants in
Santiago show lower educational levels than the native population. The
differential is reversed for women migrants in other urban areas, however,
with total recent migrants having clearly higher educational levels than
the resident population (Herold, 1978).
Destination
/ Ravenstein's principle that women who migrate usually do so over
short distances is confirmed in some cases. It is not clear, however,
what the influence of marital condition (rather than sex) is in explaining
the choice of destination since there is not sufficient data on marital
status. The data that is available in general shows that married migrants
of either sex travel shorter distances than those who are unmarried (Castro
et al, 1978).
Some findings do also in'ic te theat- nn -igrate further t.`a wcmen.
Sudarkasa, (1977) found t'i; :.o -. re ijc i. ster: Africz. in Argentina,
:t has been found that :.;-::;. :.-- :aliy travel shorter distances on their
fi-st cmve: \,h eczs only 17 per cent of male migrants in a recent study
t-a'.'eel' t: areas less tihn 100 km. away, 57 per cent of the women did
so (Connell et al, 1976). Women in Brazil predominated in intrastate
migration; with respect to interstate migration, women outnumbered men in
moves to urban areas (Castro et al, 1978). In India, women tend to migrate
to areas close to their point of origin; male migrants predominate in long-
distance moves (Zachariah, n.d.). In Colombia women outnumber men in short-
distance moves and in long-distance moves from rural areas, while men out-
number women in distant moves from urban areas (Perez, 1976).
A recent stu.y in, Maila indicates a changing trend. Whereas in the
1960s the.'e ,-ere fe'.: se.: differences among Philippine migrants in terms of
distai'ce t; e'led ('er'y, 1974), a decade later, women in Manila dominate
i:; mi~r-.or flc;.s ii;'cl'.'ig greater distances (Smith, 1978).
The i. -i'-tic fcr .:n-ong far away is based on both social and
ec:nomi: re:- s. heroesas f.' .'n m grants a specific occupation at
des in-ticn is n~i-e i-no:rtdat than the size of the destination itself,
f- r vw"e r.!-::-ic'-ints the size of destination is more crjticalbecause
of the varic:y f pa s.sible occupations available. Pernia (nd.) reports
for the Phi-lipines t--t place of destination had no significant effect
for the Philippines that place of destination had no significant effect
on men's decision to migrate. For single women and those who were heads
of household, the size of place of destination was found to be significant
at the .01 level.
Mexican women (and men) tend to migrate to metropolitan areas and
larger cities and avoid the smaller ones (Cornelius, 1975). Chilean
women are more prone to migrate to large urban centers (Valparaiso and
Santiago); whereas more men migrants go to the far removed provinces such
as Tarapaca and Magallanes (Bustamente, 1978).
Herold (1979) argues that it is women among the poor
who are first movers that are attracted to the capital city areas. She
hypothesizes that:
"...these women would have less knowledge about alternative
destinations and must continue to rely on the traditional
information network which is transmitted primarily through kin
presently residing in Santiago or that the capital continues
to have the best job market for these women."
However, if one takes the individual's total history of migration, a
more complex picture evolves and one which suggests that women may be
involved considerably more than men in a step-wise migration process, even
over generations. This is strongly suggested by findings from El Salvador,
1i/
Mexico, Ecuador, Chile, and Brazil, Migration histories of women migrants
11/ Studies of Latin American migration show a step-wise migration in which
many migrants first locate in an intermediate location. Weist (1973) maintains
that the typical migration pattern in Mexico is "from farm or hamlet to town,
and from town to city (principally Mexico City)..." (p. 182). Others found
that males were more likely to migrate in a step-wise pattern than were females.
Scrimshaw (n.d.) found that, while 46 per cent of male migrants lived in a town
or city other than place of origin prior to moving to Quayaquil, Ecuador, this
was true for only 32 per cent of the women who moved there. Similarly, Elton
(1974) found that, of the migrants to San Salvador, 70 per cent of the females
moved directly to the city while less than 63 per cent of the males did so.
She also found that 11 per cent of the male migrants in San Salvador had pre-
viously migrated to places that were smaller than their hometowns, but only
5 per cent of the female migrants did so. Again, Elton pointed to economic
opportunities as the motivating force behind this migration pattern.
to San Salvador and Santiago indicate that rural women move to small towns
in their first move, and in the following generation move to the capital
areas. In each country considerably fewer females than males migrated
directly to San Salvador and Santiago from rural areas. In San Salvador
as many as 55 per cent of the migrant women, as compared to only 11 per
cent of the migrant men, had lived in places smaller than their places
of birth before moving to the capital. This is despite the fact that there
is only a slight difference between the sexes with respect to their birth-
place (Elton, 1974). There is an important exceptionhowever. The younger
the rural migrant, the greater the chances that she will come directly into
the capital city area. In Santiago, for example, among migrant women who
fall within the mode age group (15-19) the percentage who came to the
capital city area directly from a rural area is higher (18 per cent) than
the corresponding percentage among the male migrants (11 per cent) who fell
within the mode age group which for men is the 20-24 age group. Step
migration is less frequent in Colombia. Only 35 per cent of migrants in
Bogota had moved to some other place prior to coming to the capital city;
51 per cent had come directly. The data does not, however, point out the
sex differences involved (Lubell, M McCallum, 1978).
In India, it appears that women migrants outnumber the men in small
cities and villages; males predominate in migratory movements which
involve long distances. Again, it is not clear to what extent this
pattern is determined by actual consideration of the distance factor (i.e.,
do women select small towns/villages because these are less distant from
their place of origin?) or whether it is influenced by family migration
(i.e., marriage migration is more common in villages and small towns)
(Zachariah,n.d.)
Each of the above factors, whether they relate to the poverty of the
area of origin, to the perceived availability of opportunities in areas of
destination, or to the characteristics of migrants themselves, are important.
to understanding migration patterns, most particularly the ways in which
female migrants differ from male migrants. Such considerations are valid
not only for the autonomous movements of single, widowed, divorced, and
separated women who are heads of household, but also for the women who
migrate with or follow after their husbands.
It is conceivable that all "push" factors and many "pull" factors
apply equally to different types of migration--be they international,
inter-regional, or intra/inter-urban--and that it may not be necessary
to identify different explanatory factors for each type of migration.
V. WHY DO WOMEN MIGRATE?: FACTORS EXPLAINING THE MIGRATION OF WOMEN
Until very recently, marriage was the main factor singled out to
explain the migration of women. Women in the Third World migrated with
their spouses for, it was assumed, the same reason as their husbands.
If a woman migrated alone, it was only to follow or to find a husband
(Elton, 1974; Thadani and Todaro, 1978). There was no room in the
migration literature for the autonomous migration of women for motives
other than mating, despite increasing common knowledge to the contrary.
This largely untested attribution of "marriage only" motives to
the migration of women is in part due to the invisibility of (or lack
of data on) women as economic producers and an overemphasis on their
roles as reproducers and homemakers. It naturally led scholars and
development experts to overlook any socioeconomic significance of
female migration and, thus, to dismiss the importance of analyzing sex
difference in motives and determinants of migratory patterns. For ins-
tance, in the case of India, A. Singh (1978a)argues that the well-known
fact that the volume of female rural migration far surpassed that of
male migration has been dismissed as a reflection of the custom of
marrying outside a woman's village of origin.
It is true that in developing countries many women have migrated
and still do migrate, at least ostensibly, for marriage purposes.
Evoking marriage, however, as the factor accounting for the moves of
such large numbers of people can only obscure our understanding of
the economic and social factors that affect women and men migrants
differently. More importantly perhaps, it yields no information
useful for policy or program formulation.
Women themselves may report that they migrate for marriage
reasons only because it is one of the few culturally sanctioned
explanations or rationalizations for their autonomous migration. More
generally, women seem to underreport the economic reasons for their
moves. Analyzing migratory movements in Subsahara Africa from rural
areas to primate cities, K. Little (1973) observes that, while both
men and women move to improve their socioeconomic status, women express
this motivation in a different, more personal language which reveals
a sex difference in attitude only. In her sample of women migrants to
Gaborne, Botswana, Bryant (1977) observes that, in one interview, 41
per cent of the women said they had come to Gaborne to find a job, while
50 per cent said they had come to join a relative. Yet six months later,
51 per cent said they had come because of a job, and only 37 per cent
said they came to join a relative. Bryant attributes these differences
to the interview situation that in the first case led women to feel social
pressures and thus give socially sanctioned responses. (They were inter-
viewed by men and in front of the whole household.)
Ideal data to explain the migration of women would compare character-
istics of women migrants with those of women staying in the place of origin
and/or native women in the place of destination as well as with those of
male migrants. Equally important, these comparisons should be based on
a valid assessment of women's (and men's) actual economic behavior. Cur-
rently, sex differences in, for instance, the association between wages
and migration may simply be a result of absence of reliable data on
female wages, as Schultz (1971) pointed out in a study of internal mig-
ration in Colombia that yielded a significant association between wages
and migration for men but not for women. The problems of measuring workers'-
participation in, as well as the wage value of, subsistence agriculture
and work in the informal sector are well known. It is also becoming well
known that women are overrepresented in these two areas of economic activity.
While perhaps less well known, recent evidence also shows that there are
gross underestimations of women's participation, as wage laborers, in
modern sector agricultural activities (Deere, 1979; Buvinic, 1978).
Just as improved measures of women's economic behavior are needed
in order to explain the migration of women, explanatory factors of the
economic behavior of women, both at points of origin and destination,
are needed in models constructed to explain migration patterns. A
model recently formulated by Thadani and Todaro (1978) is such an attempt.
To explain the internal migration of autonomous women, they propose modi-
fying the male model to include:
a) actual rural-urban wage differentials and a measure of sex
discrimination in the modern sector (by measuring the probability of
employment) in the factor assessing employment/income differentials
in the formal sector, and
b) a factor measuring employment/income differentials in the in-
formal sector.
In addition, they include two marriage factors -- one accounting
for a normative pressure to marry (marriage for its own sake) and the
other responding to women's aspirations for economic mobility through
marriage (operationalized as the probability of marrying males in the
formal sector). They also include a sex role constraint and a "residual"
factor.
The section below will review recent evidence on factors affecting
the migration of women bearing in mind the theoretical and data limita-
tions just mentioned.
The Apparent Reasons
The available evidence consistently shows sex differences in the
(verbal) response autonomous migrants give to explain their own decisions
to migrate. Across countries and over time, men's reasons for migration
Share pr dominantly work-related. Women's reasons are less consistent
I
S(over time and across countries) and often include marriage and family
Kas well as work reasons. A survey of moves in and out among more than
two hundred villages in Bangladesh indicates that men move mainly
because of work and/or living conditions (57 per cent and 89 per cent
of the independent moves in and out, respectively). Women, on the other
hand, move most often as a result of marriage or marriage breakdowns
(63 per cent and 67 per cent of their independent moves in and out,
respectively). The same survey reveals a very high divorce rate in this
region that affects women specifically; there are 2.7 times more divorced
women than divorced men (Ruzicka and Chowdhury,1978). In a survey of
a large sample of migrants to Lima, Peru, 53 per cent of the men and 30
per cent of the women gave economic reason for their moves, while almost
half of the women and only one in six males gave family reasons (Macisco,
1975). Half of the women in a sample of migrants to Lagos, Nigeria gave
accompanying or following their husbands as motives. Only 8 per cent
said they came for work or education-related reasons (Lucas, 1974).
While family and marriage are often mentioned by women, economic
reasons increasingly are also being given. Forty per cent of the women
in a sample of migrants to Bangkok, half of those in a sample migrating
to Kingston, and 81 per cent of a smaller sample of women migrants to
the slums of New Delhi gave employment as the main reason for their move
(Piampiti, n.d.; Standing, 1978b; Singh, 1978a ). As has already been
mentioned, evidence from women migrants to Gaborne, Botswana, and also
to Lagos, Nigeria, suggests that women may underreport the economic reasons
behind their moves (see Bryant, 1977; Lucas, 1974).
Women also verbalize freedom from traditional norms and restrictions
in the village as a main reason for their moves to urban areas. Little reports
this as the case for many African women, especially women who have unhappy
or barren marriages. Connell et al (1976) find that women among the Baoule
in Ivory Coast migrate as "an act of defiance against men", and Castro et al,
(1978) find that many young women migrate to urban areas in Brazil after having
lost their virginity. There is no evidence of men giving similar "freedom"
reasons for their moves, and the possibility exist that these women migrate
not to attain freedom but because they are forced to leave when they break
socially defined codes, which tend to be harsher for women than for men.
However, it should be kept in mind that motives of women migrants may differ from
those of men only in their expression; the reasons given need not correspond
with the real reasons for migration.
Underlying Socioeconomic Factors
To pinpoint socioeconomic factors that affect the migration of the
sexes differently, this section will review regional economic factors
explaining women's migration as well as factors that may restrict the
migration of one sex but not the other. The relative mobility of the
sexes in different geographical areas depends on the relative economic
responsibility carried by men and women, the relative availability of
alternative jobs for the two sexes, and economic as well as noneconomic
restrictions on women.
Women migrants in Latin America and the Caribbean. In the last two
decades women in the region have predominated in rural to urban migration
flows. They are both young (10-19 age cohort) and old,single, less educated
than their male counterparts, and generally also less educated than native
urban women. They tend to migrate to the larger cities and metropolitan
areas, whether they have moved in stagewise fashion or directly from the
rural region of origin.
Parallels have been drawn between historical internal migration
patterns in the region and those of the United States; these stages can
be related to different levels of economic development. In the first
stage, more males than females migrate and migration is seasonal or
residence at the destination lasts only a year or two. During the second
stage, more families migrate, and more migrants intend to stay for several
years or until they retire, if not permanently. Finally, during the third
stage, more females migrate (Elton, 1974)..
There is wide agreement that economic factors determine this third
migratory stage. Women's high rates of rural outmigration are attributed
to their displacement from subsistence agriculture as land consolidation,
agricultural mechanization, and the growth of wage employment reduce
women's productive role and leave them increasingly dependent on men's
insecure income. In conditions of strictly limited access to cultivable
land, population growth has added to fragmentation of land ownership and,
thus, to stagnant incomes. This general pattern of stagnant and declining
rural living standards, common to many economies in which capitalist growth
is occurring, has meant lack of jobs for young women in agriculture as well
as decreased opportunities to earn even low incomes (Standing, 1978d).
On the other hand, the large metropolitan areas offer these women
work in either domestic service or the informal sector. As low paid as
these jobs may be, it is argued that since young girls are not needed to
help in either agricultural work or in household tasks, poor families
may maximize potential resources such as wealth, income, and employment
opportunities of all family members by sending their young daughters to
town to become domestic workers, even if only for room and board (Boserup,
1970; Jelin, 1977).
The data available is quite consistent in supporting this "pull"
argument. In fact, pull factors seem to outweigh all others in explain-
ing women's mobility. In Chile, the correlation between urban population
and migratory pull is higher for women than it is for men (Bustamante,
1978); in Peru, pull factors appear more important than push factors to
explain the predominantly female migratory flows to urban areas in the
1960s (Macisco, 1975).
The employment patterns of female migrants in the large Latin
American cities--their high participation in sectors of low productivity
and wages as domestic work and other personal services--suggests urban-
ization rather than industrialization as the structural factor "pulling"
women to the cities. The available evidence supports this suggestion.
Data from Chile shows that women migrants are more attracted than men
by urban areas that provide health, housing, and basic education infra-
structure; that is, when compared to men, they appear to migrate Dot only
/ for the jobs the city offers but also for the infrastructure and services
of urban environments (Bustamante, 1978). However, although they are attra-
cted by such services, they usually do not benefit from them because they
cannot find work in areas where services are available.
Historical labor force participation data from Brazil and from
Colombia indicates that men's, but not women's, participation is directly
related to industrialization levels (Lewin et al, 1977; Leon de Leal, 1977).
Large metropolitan areas in Brazil show an inverse relationship between
level of economic development and the proportion of female population
living in the area (the urban sex ratios); further, it is in the least
developed metropolitan areas that women most often are employed in the
tertiary sector of the economy, especially in the category of personal
12 '
services (Lewin et al, 1977)- The Brazil data suggests that women
migrants may end up in urban environments with low levels of industria-
lization, limited opportunities for productive employment and low or
inaccesible levels of urban services. The Chile data confirm this.
Urban areas with better health and housing infrastructures "hold" on
to migrant men more than to migrant women; when compared to women, men
leave sooner those provinces with less health and housing infrastructures.
The explanation may lie in the types of employment offered to men and women.
Men are placed in urban areas in the high capital technological sector
associated with high earnings as well as good health and housing services.
Men's jobs stabilize men in areas with good infrastructure facilities,
1-'That is, there is a significant negative correlation between the
economic development of different metropolitan areas and female labor
force participation in services.
while women, who migrated in the first place to these areas because of
more housing and health facilities, obtain jobs that prevent their access
to these urban benefits (Bustamante, 1978).
Although the "pull" factors have been largely confirmed, recent
findings bring into question the "push" factors widely used to explain
the women's outmigration from rural areas in Latin America and the
Caribbean. The commonly held assumption is that women's role in agri-
culture in the region is low and/or restricted to the subsistence sector.
Data from Brazil, Colombia, and Honduras, however, challenge census re-
porting and show that a significant proportion of wage labor in current
commercial agriculture is women's labor (Lewin et al, 1977; Deere, 1979;
Buvinic, 1978). Moreover, the Brazil data, show a positive association
between expansion of small landholdings (through colonization and sub-
divisions) in the 1950-60 decade and growth in women's labor force parti-
cipation in agriculture. It does not seem, therefore, that women migrate
to the cities in the region because they have no access to wage earning
jobs in agriculture or because fragmentation of landownership has dis-
placed them from agriculture. It is highly probable that rural/urban wage /
differentials still play an important part in women's rural outmigration,
even if they are agricultural wage laborersA. An additional central
factor may be rural/urban differentials in infrastructure especially
housing, education, and child care--which is one of the bases for our
hypothesis that a substantial proportion of those women migrants may be
In fact, a logical prediction is a much higher probability of out-
migration for rural women who have access to cash earnings than for those
who do not. The only exception would be the migration of young girls
many of whom are sent to the city to reside with relatives and/or "god
parents" (J.elin, 1977).
de facto heads of household with one or more children to support.
Sex differences in migratory patterns in Sub-Sahara Africa. In
order to explain migratory patterns in Africa, an analysis must be made
of the sex-specific factors that prevent (as well as those that promote)
the migration of women. It is by now well known that economic policies
introduced in the early part of this century by colonial regimes
triggered a vast migration of rural African men to plantations and urban
areas in search of work that would provide them with cash incomes. The
overall redirection of economic activity from precolonial production and
trade to export oriented production and commerce, the introduction of
goods and services that had to be purchased with cash, and the imposition
of compulsory labor laws caused the migration of, for instance, West
Africans from the interior to the coastal administrative/commercial centers.
It also reinforced the customary wide difference in marriage age of young
men and girls in African villages. The recruitment for wage labor of
males between the ages of 20-35 left a high village ratio of women to men
in those age groups; women waited and married the older men who had re-
turned from wage labor (see Boserup, 1970; Sudarkasa, 1977). Women gener- -
ally did not migrate with the men, not only because of labor policy
restrictions but also--and more importantly--because women had had a sig-
nificant part in pre-colonial subsistence agriculture and remained in
charge of subsistence crops in the village. (Boserup, 1970; Tinker, 1976).
This pattern of highly male selective seasonal or nonpermanent
migration continues today, especially in South African countries where
government restrictions do not permit males to be accompanied by their
families. Recent UN estimates place up to nine males for every female
in some mining towns in Lesotho and Botswana, among others. Mueller
(1977) estimates that the average miner spends 35 per cent of his work-
ing life in the mines. Social restrictions against the migration of
women from rural to urban areas are also mentioned in migration studies
for selected African countries (i.e. Zambia and Kenya), although such
restrictions are not present in other countries (i.e., Ghana and Nigeria)
(Caldwell, 1968; Levine, 1966). In Zambia, until 1916 it was illegal
for a single women to migrate to town without permission of the native
commissioner (Schuster, 1979).Little (1973) interprets the enactment of
this law as an attempt to preserve tribal stability and induce the return
of migrant men. One of the reasons given for women's reluctance to move
to town now is the fear of being labeled a prostitute (Schuster, 1979).
Men perpetuate this restriction by not marrying urban women, but
returning to the villages in order to find wives. On the other hand,
Levine (1966) reports positive reactions to the migration of women in
northern Nigeria, where women's market and trading activities require
mobility.
Women, however, are starting to leave the rural areas in some
African countries. A significant proportion of single women started
migrating to Lagos, Nigeria, after the 1966-67 Civil War; by 1973, for
every woman aged 25-29 who grew up in Lagos State there were three mig-
rants of the same age (Levine, 1966). Caldwell observes a significant
tendency toward a greater equalization of the sexes in the 1960s
in West African non-primate cities. He attributes these demographic
changes to increased employment opportunities for women in the cities
and a vast improvement in the system of roads and transportation
(Caldwell, 1975). Bryant (1977), Schuster(1979), and Sudarkasa. (1977)
also cite the employment opportunities that African cities are offering
women; not surprisingly, as in the Latin American case, they are
domestic work. Also, as in the Latin American case, the migration of
autonomous African women seems to occur in two extreme age cohorts, the
very young and women over fifty years of age. For the very young, the
"pull" factors identified are jobs as domestics and marriage motives. The
"push" factors are further deteriorating economic conditions in rural
areas, coupled with heavy agricultural burdens for women and the severe
shortage of males of marriagable ages within certain villages and/or
status groups as a result of previous patterns of male migration (Little,
1973).
For the very old, the "pull" factors are jobs in the cities. The
"push" factor again is increasing rural poverty, particularly for widowed
or divorced women with dependents and without the traditional economic
support they used to have.
An additional push factor given is the increased willingness of
farmers to hire women at less than male wages (Connell et al, 1976), which
suggests that, as in Latin America, more women than is currently
thought may be participating as wage labor in commercial agriculture.
Sex differences in migration patterns in North Africa and the
Middle East. The existing migration literature reports few women
migrants in North Africa and the Middle East, at least in internal
migratory flows. Islam has been widely thought to restric women's