Sex Preference and the Value of Sons and Daughters in Nepal

Author(s): Yagya B. Karki

Source: Studies in Family Planning , May - Jun., 1988, Vol. 19, No. 3 (May - Jun., 1988),
pp. 169-178
Published by: Population Council

Stable URL: https://www.jstor.org/stable/1966752

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

Population Council is collaborating with JSTOR to digitize, preserve and extend access to
Studies in Family Planning

This content downloaded from

103.173.2.149 on Sat, 11 May 2024 01:48:49 +00:00
All use subject to https://about.jstor.org/terms
 Sex Preference and the Value of Sons
and Daughters in Nepal

Yagya B. Karki

Sex preference and the specific value of sons and daughters to parents in Nepal are examnined
using rural and urban survey data fromn 1979. Ideal family size among all respondents zvas, onl
average, thlree children, with two sons and one daughter the preferred sex coinpositionl for about
90 percent of all respondents. Among those who reported current contraceptive use, the mean
number of living sons was hiigher than the mean number of living daughters for all respondenits.
Most couples have at least one son before they adopt contraception; respondents hlad, on average,
three to four births before adopting contraception. Sons are preferred to daughters by Nepalese
parents mainly for socioeconomic and religious reasons, as opposed to the economnic reasons
reported elsewhlere in many developing societies. The findings indicate that the economic inotive
for hlaving bothi sons and daughters may be weakening in Nepal, but that a preference for sons
does exist. (STUDIES IN FAMILY PLANNING 1988; 19, 3:169-178)

Sex preferences, and preference for sons in particular,
as a major determinant of family size, have been well
documented in several Asian societies.' The Value of
Children (VOC) study on six countries in Asia and the
Pacific, for example, reports that family size is influenced
by son preference (Arnold et al., 1975). This report cites,
among other factors, continuity of the family name, re-
ligious rituals, and ancestor worship as important rea-
sons for wanting sons in Taiwan, South Korea, and
Thailand. In addition to wanting sons for economic rea-
sons, the rural respondents in Thailand also cited reli-
gious reasons for wanting sons. However, the VOC
studies concentrated only on urban and semi-urban
areas, although the latter were reported as rural areas
(Ware, 1978). Also, although these studies document son
preference, they do not explain much about how such
attitudes and beliefs actually affect fertility behavior.
A few studies, in addition to documenting the im-
portance of son preference, try to show its impact on
fertility. A cross-cultural comparative study (Freedman
and Coombs, 1974) on son preference based on second-
ary data shows stated attitudes that indicate a clear pref-
erence for sons in Korea and Taiwan, and to a lesser de-
gree in India. The impact of this attitude on fertility
Yagya B. Karki, Ph.D. is Director, Worldview International
Foundation/Nepal, P.O. Box 2912, Tangal, Kathmandu,
Nepal. At the time this paper was written, he was a
consultant in demography at the National Commission on
Population, His Majesty's Government of Nepal,
Kathmandu.
behavior, however, was insignificant. Also, because the
study was exploratory and the data analyzed were not
collected specifically to examine son preference and fer-
tility behavior, it was not possible to look at the impact
of son preference on fertility.
A number of studies conducted in India do look at
son preference and fertility. In some Indian communities
sons are highly prized; the birth of a son may be an oc-
casion for celebration, whereas a new daughter is almost
cause for commiseration (Gideon, 1962). Furthermore,
when times are difficult, male children are given better
treatment than female children; mostly boys attend
schools, and boys are better fed and better clothed. De-
spite this, not much evidence exists of any influence on
fertility in societies that favor male children (Cassen,
1980).
One way to look at the impact of son preference
on fertility would be to examine birth order data by sex
of children, for couples of completed fertility. If son
preference has any influence on fertility, couples who
have sons first would be expected to curtail childbearing
sooner than those who have daughters first. One study
that carried out work along these lines in India, however,
found no such effect (Repetto, 1972). In a nationwide
sample of about 25,000 couples in India in 1970, the ma-
jority wanted children of both sexes but expressed a
modest preference for sons (ORG, 1971). The parity pro-
gression ratios (the probability of going on to parity x +
1 from x) of these couples were little influenced by the
sex of children already born, although it was found that
couples who had only daughters were the least likely to
want no more children and wanted the largest number

Volume 19 Number 3 May/June 1988 169

This content downloaded from

103.173.2.149 on Sat, 11 May 2024 01:48:49 +00:00
All use subject to https://about.jstor.org/terms
 of additional children, particularly sons (Sharma and
Jain, 1974). Overall, the analysis shows that ideals about
the desirability of sons do not greatly affect reproductive
performance, but among contraceptive users, particu-
larly acceptors of terminal methods, some direct corre-
lation is seen between the number of surviving sons and
family planning acceptance (Sharma and Jain, 1974: 96).
Apparently, the effect of son preference on fertility
can be observed only when a high level of contraceptive
practice exists. This was the case, for instance, in South
Korea, where the effect of son preference showed up
only after the third birth (Park, 1978), when parents be-
gan to practice contraception. Similar findings have been
reported in Taiwan (Heer and Wu, 1974).
The Nepal Fertility Survey (Nepal FP/MCH Project,
1977) has produced a few tables that show a strong in-
fluence of son preference on the number of children de-
sired. For example, it was found that women with one
son and one daughter were almost three times more likely
to want no more children as compared with women who
had two living daughters and no son (27.4 percent vs.
10.0 percent). However, this study lacks adequate in-
formation to explain why son preference appears to be
so strong in the analysis. Besides, the study was not de-
signed specifically to explore sex preference in Nepal.
Utility of Children
A population prefers one sex to the other generally be-
cause of some definite "utilities" associated with the
preferred sex. These utilities, however, vary from one
society to another and also within different segments of
the same society. In the context of the Hindu society of
India, Mamdani (1973) writes that the first duty of a mar-
ried woman is to become a mother; her dharma (duty) is
to bear children, particularly sons. Mamdani appears to
overlook the noneconomic value of sons because, in his
study, the ultimate goal of having sons is to acquire eco-
nomic prosperity. This is illustrated in the following re-
marks made by an elderly farmer to Mamdani:
You were trying to convince me in 1960 that I should not
have any more sons. Now, you see, I have six sons and
two daughters and I sit at home in leisure. They are grown
up and they bring me money (1973: 109).
It may be that wealthy people, particularly those with
land, want many children for economic reasons. Pun-
jab-the state that Mamdani studied-is the richest one
in India. Mamdani's quotation certainly does not speak
for the vast majority of poor Indians, though, because it
is very unlikely that their children will be as economically
rewarding as the children of Mamdani's well-off farmer.
A recent study (Vlassoff and Vlassoff, 1980) of Mahar-
ashtra villages, where 87 percent of the villagers depend
on farming for their livelihood, has produced evidence
that having sons was no guarantee even for old age se-
170 Studies in Family Planning
This content downloaded from
103.173.2.149 on Sat, 11 May 2024 01:48:49 +00:00
All use subject to https://about.jstor.org/terms
curity for parents, let alone for economic prosperity. This
study casts doubts on the view that sons represent "poor
man's capital."
Nag et al. (1978), analyzing time budget data on the
activities of children aged six years and over from villages
in Java and Nepal, report that children are economically
useful. However, Vlassoff (1979) has come up with evi-
dence from a village in Maharastra State, India that eco-
nomic costs outweigh the benefits of children up to age
16. Still, he says, children are expected to work because:
(a) it serves to strengthen the bond between parents and
children; (b) it provides a release of energy in an envi-
ronment where recreational facilities for children are al-
most nonexistent; and finally, (c) it fosters the learning
of agricultural skills through a period of lengthy ap-
prenticeship.
Study Objectives
Using primary data from the West Central hills of Nepal
and employing both qualitative and quantitative meth-
ods, the questions addressed in this study are:
1 What sex do the respondents prefer their child(ren)
to be?
2 How strong is sex preference?
3 What are the most important utilities of sons and
daughters to parents?
4 Does stated sex preference show up in actual be-
havior? For example, how-if at all-does it affect
fertility and contraceptive practice?
It is hoped that the findings will add substantially
to the already existing knowledge of social demography
in general and to Nepalese demography in particular.
This study is of great import to the latter, because con-
tinuing and rapid population growth in Nepal2 is already
having serious deleterious effects on the ecological, so-
cial, and economic environment. The problems are par-
ticularly severe in the hills, where traditional farming
practice is forcing peasants to work increasingly steeper
slopes that are often unfit for sustained farming, even
with the elaborate terracing practiced there. An increas-
ing demand for fodder and fuelwood is diminishing the
supply of ground-holding trees, causing frequent land-
slides that, in turn, destroy lives, homes, and crops
throughout the hills (Eckholm, 1976). Such deteriorating
conditions of the hills are forcing people to outmigrate
continuously.
Data and Methods
The data come from two sample surveys conducted in
Gorkha and Pokhara in 1979. Both Gorkha, a rural dis-
trict, and Pokhara, a town, lie in the West Central mid-
hills area of Nepal. The questionnaire used in the surveys
consisted of three main parts. Part 1 comprised census-
type questions; Part 2 elicited information on fertility and
on infant and child mortality from 15-49-year-old ever-
married women; and Part 3 focused on the Value of Chil-
dren items, such as advantages and disadvantages of
having children, ideal and desired family sizes, satis-
faction from having children, birth control information,
perceived economic values and costs of children, and
information on the socioeconomic background of re-
spondents.3 Well trained interviewers were employed
to administer the questionnaire. In the rural area, only
males conducted the interviews, as it was not possible
to recruit a sufficient number of females, partly because
the job involved spending nights away from home. The
Nepal Fertility Survey (Nepal FP/MCH Project, 1977) also
employed only male interviewers in the hills. Altogether,
3 supervisors and 18 interviewers worked for two and a
half months in the rural area. In the urban area, four
females interviewed the wives, and three males-in-
cluding one supervisor-interviewed the husbands.
In Gorkha, out of 47 village panchayats (a politico-
administrative unit roughly equivalent to a county in
Great Britain), five were selected on the basis of proba-
bilities proportional to size. Nine village panchayats were
not included in the sample universe because of difficult
access, as they were above 10,000 feet in altitude. Census-
type information was collected for every second house-
hold, which produced a population of 11,542 individuals
representing 7 percent of the total population of 164,858
of the sampled universe. Part 2 of the questionnaire was
administered to ever-married women aged 15-49 in every
household, while Part 3 was administered to wives aged
15-39 with at least one living child and whose husbands
were at home at the time of the survey. Consequently,
of the total 1,909 sampled households only 669 house-
holds had women eligible for Part 3 of the questionnaire
and, of these, 683 wives were successfully interviewed.
Only 15-39-year-old women were interviewed, because
this is the age when both partners are most sexually active
and also when they give the most attention to the values
and costs of bearing and rearing children. In addition to
the wives, 216 rural husbands from every third house-
hold were interviewed. The response rate among the
wives was 99 percent and among their husbands it was
98 percent.
Pokhara town, comprising 16 wards, had a total
population of 36,189 in 1979. For sampling purposes,
wards 1, 2, 3, 4, 9, and 12 were classed as highly urban-
ized, and the other ten wards were classed as less ur-
banized. Ward 3 was sampled from the highly urbanized
groups, and wards 11, 13, and 16 from the less urbanized
group, on the basis of probabilities proportional size. The
survey questionnaire was administered to every second
household in ward 3, but in order to make the sample
sizes of the two groups uniform the wards from the less
urbanized area were stratified. Thus, demographic in-
This content downloaded from
103.173.2.149 on Sat, 11 May 2024 01:48:49 +00:00
All use subject to https://about.jstor.org/terms
formation was collected on a total of 1,863 individuals
constituting about 5 percent of the total population of
the town.
In Pokhara, a census-type questionnaire was ad-
ministered to 291 households containing 343 ever-mar-
ried women aged 15-49, and from these, 188 wives were
interviewed for the VOC information. Only 50 percent
of husbands were administered the VOC questionnaire
in the town. The response rate was 93 percent for wives
and 91 percent for husbands. Interviews in both areas
were completed, on average, in 52 minutes.
The Setting
Based primarily on subsistence agriculture, Gorkha is
typical, in many respects, of the mid-hills area of Nepal.
In Pokhara, about one-half of the population lives on
trade and commerce and the other half on farming. The
villagers in Gorkha produce two crops a year on dry land
and normally one crop a year on irrigated land. They
depend entirely on the monsoon rains for irrigation.
In the Gorkha sample almost every family had some
land to work, the average being 1.4 acres per household,
whereas the corresponding figure for the urban area was
0.4. However, in Pokhara about 57 percent of the people
interviewed reported nonagricultural occupations as
opposed to only 2 percent in the rural areas.
Bullocks provide the main motive power for culti-
vation. Apart from plowing, all other work needed for
farming is done by human beings. Without human labor,
which includes the work of men, women, and children,
the rural economy cannot be sustained. In addition to
land, the villagers also own animals; the mean number
of cattle, which includes cows, oxen, and buffalo, is about
six per family, and the mean number of goats and sheep
per family is about three.
The household is the primary unit of production for
the family's own consumption. In most villages, paid
labor is scarce. Instead, mostly reciprocal labor arrange-
ments (parma) and communal activities are practiced in
the hills, with the result that very little individual wealth
is accumulated.These arrangements are tantamount to
a form of communal insurance against a single household
being left destitute as a result of a natural disaster (for
example, a landslide), and also enable poor families to
conduct such events as weddings and funerals. The
goods required for running the local economy are man-
ufactured by the household itself through cloth weaving,
house building, and the manufacture of rice stalk mats,
bamboo baskets, winnowing fans, and the like, activities
that usually take place when the agricultural season is
over or slow.
Just as in many other less developed countries, sig-
nificant migration takes place from the hills to other areas
for employment. Many adult males seek employment
outside the country, leaving their families to work land
Volume 19 Number 3 May/June 1988 171
 on their own. This is common even in the urban areas.
About 31 percent of 15-49-year-old rural married women
reported that their husbands were away from home at
the time of the survey; the corresponding figure for the
urban area was 20 percent. However, this outlet for the
burgeoning population is narrowing both in India and
South East Asia. Since the majority of the migrants come
from the hills, the dwindling employment opportunities
outside Nepal have important implications for the hills
population.
According to the present survey, about 15 percent
of rural and 55 percent of urban respondents were lit-
erate. The national literacy rate for Nepal around 1979
was 19 percent (World Bank, 1981). The rural district un-
der study had no electricity at the time of the survey, but
a road joined its district headquarters (located about 1,200
meters above sea level) with the main roadway, which
passed along its southern border. It also had 10 high
schools, 170 primary schools, and 2 hospitals. Most
modern amenities are foreign to the villagers of Gorkha.
Pokhara, on the other hand, is the location of one of five
regional headquarters of the Western Development Re-
gion in Nepal. It is probably the second most important
valley in Nepal attracting foreign tourists.
The total fertility rate (TFR) in 1979 was higher in
the rural area (6.4) than in the urban area (4.3) (Karki,
1982). Using the Nepal Fertility Survey data, the national
TFR was found to be 6.3 in 1976 (Goldman et al., 1979).
As expected, the infant mortality rate (IMR) was higher
in the rural area (130 per 1,000 live births) than in the
urban area (60 per 1,000) (Karki, 1982). Urban-rural
mortality statistics are not available but the national IMR
in 1976 was 134 per 1,000 live births (Central Bureau of
Statistics, 1978).
Almost no one in the rural population, if the statistics
are taken at face value, practices contraception; only 15
percent reported having ever heard of family planning
in the rural area, though nearly 60 percent knew about
family planning in the urban area. Current contraceptive
use was, however, low in both areas: about 29 percent
(N = 27) of the husbands and only 11 percent (N = 20) of
the wives reported current contraceptive use in the urban
area. The corresponding figures were 7 percent for hus-
bands (N = 15) and 7 percent for wives (N = 48) in the
rural area (Karki, 1982). The extent to which the urban
wives were reluctant to reveal contraceptive use appears
quite considerable, and it is probable that the rural wives
were equally if not more reluctant to discuss their con-
traceptive behavior. It may be that the husbands were
also concealing information about their contraceptive
use. All of the respondents may have been similarly hes-
itant to discuss their knowledge of family planning.
Since in the Nepalese Hindu society extramarital sex
is very uncommon, it follows that when husbands are
practicing contraception, their wives must also be doing
so. Since so many more husbands than wives reported
current contraceptive use, it is plausible to assume that
172 Studies in Family Planning
This content downloaded from
103.173.2.149 on Sat, 11 May 2024 01:48:49 +00:00
All use subject to https://about.jstor.org/terms
the Nepalese women interviewed did not want to discuss
their sexual behavior in any context. Campbell et al.
(1979) and Bennett (1976) make special reference to the
Bahun-Chhetri community, who are orthodox Hindu.
One of the main problems is that the whole idea of family
planning is a very sensitive matter in Nepalese society,
particularly among women. It is said that when a woman
talks openly about family planning, she is seen as morally
degraded, and her fidelity to her husband becomes sus-
pect.
Campbell et al. (1979) argued that one reason for
such a low level of family planning knowledge among
the NFS informants was that the interviewers were
males. Interestingly, however, the degree of complete-
ness of reporting on current contraceptive use was no
better among urban as compared to rural wives in the
present study, even though urban wives were inter-
viewed by females.
Nepalese society is male-dominated and wives
hardly ever become involved in matters outside the
home. Moreover, since revealing information about
family planning is likely to make most wives anxious
about their relationship with their husbands, they do
not usually talk about it openly unless they obtain their
husband's consent. It is also reported by Campbell et al.
(1979) that most women would not want to talk about
family planning in the presence of others, particularly
the elderly members of the family, to whom they must
pay respect. In view of such constraints, every effort was
made in the present study to interview each person alone.
Also, the respondents were told that the present study
had nothing to do with the government's family planning
program in Kathmandu. Yet, the interviewers got the
impression that the respondents, particularly the wives,
did not want to talk about family planning.
Results
Sex Preference
The questionnaire contained both open-ended and
structured questions on sex preference. The set of open-
ended questions was on ideal family size and compo-
Table 1 Response to an open-ended question about ideal
family size and composition
Wives' response Husbands' response
Variable Rural Urban Rural Urban
Desired (mean)
Family size 3.2 2.9 3.2 3.2
Number of sons 2.0 1.7 2.1 1.9
Number of daughtersa 1.2 1.2 1.2 1.3
Numberof respondents 683 188 216 92
aThose who did not want any daughters included 43 rural wives and 3 urban
wives, and 13 rural husbands and 6 urban husbands.
Table 2 Preferred sex of children and family composition The set of structured questions required respondents
among Nepalese couples to choose their ideal combination of sexes if they could
Percent of responses
have only three children. When asked what their first
Wives Husbands
choice was, respondents indicated a strong preference
for sons (Table 2), with about 90 percent of all respon-
Sex composition Rural Urban Rural Urban
dents choosing two boys and one girl. Respondents were
First choice
Three girlsa 0 1 0 0 then asked to choose the next best combination, and the
One boy, two girls 5 1 3 7
majority selected one boy and two girls. Finally, those
Two boys, one girl 86 98 89 93
Three boysa 9 1 8 0 respondents who had not already chosen either three
Total percent 100 100 100 100 sons or three daughters as their optimal family compo-
(Number of respondents) (683) (188) (216) (92)
sition were asked to choose between the two. As ex-
Second choice
Three girlsa 1 1 0 0 pected, the overwhelming majority chose all sons. This
One boy, two girls 50 68 55 66
clearly indicates a strong preference for sons, a preference
Two boys, one girl 5 1 3 7
Three boysa 44 31 42 26 that persists even when the respondents are broken
No response 1 0 0 0
down by rural and urban areas and by sex.
Total percent 100 100 100 100
(Number of respondents) (620) (186) (198) (92)
For comparisons with other Asian societies, the re-
Third choice sponses shown in Table 2 are again classified, in Table
Three girls 2 2 3 9
3, by sex bias scales (Coombs et al., 1975), ranging from
Three boys 96 97 96 88
No response 1 1 2 3 one (girl preference) to seven (boy preference). The other
Total percent 100 100 100 100 societies included in Table 3 are Taiwan, South Korea,
(Number of respondents) (340) (127) (115) (67)
and the Philippines (Coombs and Sun, 1978). The strong
Note: Percentages may not sum to 100 because of rounding. son preference in Nepal is very much on a par with that
aNot asked further.
in Taiwan and in South Korea. The Philippines offers a
marked contrast, where as much girl as boy preference
is found.
sition. The respondents were asked what their ideal Table 3 shows that husbands express no preference
family size and sex composition would be if they could for girls at all, whereas some wives, albeit a very small
start their family building all over again. The mean ideal proportion, do indicate some preference for girls. The
size is about three (Table 1). Preference for sons over rural respondents appear to show stronger son prefer-
daughters is slightly stronger among the rural respon- ence (IS6-7 = 48 percent) than their urban counterparts
dents, but the mean number of sons desired in both (IS6-7 =29 percent). If son preference affects fertility,
groups is about two, and the mean number of daughters then it is likely that the rural population will ultimately
desired is about one. have larger families than the urban ones.

Table 3 Sex bias distribution among Nepalese couples

Coombs IS scalea

Girl preference Balance Boy preference Total

Respondents ISi IS2 IS3 IS4 IS5 IS6 IS7 % N

Wives
Rural 0 1 c 5 45 39 10 100 659
Urban c c 0 2 66 31 c 100 186
Allb c 1 C 4 50 37 8 100 845
Husbands
Rural 0 0 0 4 49 38 8 100 213
Urban 0 0 0 7 68 25 0 100 87
Allb 0 0 0 5 54 35 6 100 300
Rural husbands
andwives 0 c c 5 46 39 9 100 872
Urban husbands
and wives c c 0 4 66 29 c 100 273
All respondentsb c c c 5 51 36 7 100 1,145
Taiwan 0 0 1 11 54 32 2 100 5,463
Korea 0 1 0 4 4 49 5 100 d
Philippines 5 8 24 33 20 7 3 100 d

Note: Total percent may not add to 100 because of roundi

aThe Coombs IS preference for sex of children scale rang
with IS4 indicating a clear preference for balance of s
to give unbiased population estimates. cLess than 0.5 percent. dNot available.
Source: C.H. Coombs, L.C. Coombs, and G.H. McCleland, "Preference scales for number and sex of children," Population Studies
29,2 (1975):273-298; L.C. Coombs and T.H. Sun, "Family composition preference in a developing culture: The case of Taiwan,"
Population Studies 32,1 (1978):43-64.

Volume 19 Number 3 May/June 1988 173

This content downloaded from

103.173.2.149 on Sat, 11 May 2024 01:48:49 +00:00
All use subject to https://about.jstor.org/terms
 Reasons for Wanting Sons and Daughters Karki, 1982: 73). Thus, the old age security value of sons
(socioeconomic) and the religious value (noneconomic)
The analysis so far demonstrates that all respondents
appear to be the most important dimensions of the value
have a preference for sons, but that few would be satisfied
of children in the mid-hills populations of Nepal. The
with sons alone. In order, therefore, to examine desired
old age security value of sons also was seen in a study
family composition further, separate open-ended ques-
conducted in the neighborhood of Kathmandu city, al-
tions were asked on reasons for wanting sons and
though only a minority felt that sons provide adequate
daughters. This was done after the respondents had in-
security for old age (Schuler and Goldstein, 1986).
dicated the sex composition of their ideal family.

Reasons for Wanting Daughters

Reasons for Wanting Sons
About 5 percent of all respondents did not want any
All respondents who wanted at least one son in their
daughters (see Table 1, note a) in their ideal family. The
ideal family were asked to give at least two reasons for
rest were asked, "For you, what are the most important
wanting sons. The question was, "For you, what are the
reasons for wanting a daughter?" (The interviewers were
most important reasons for wanting a son?" (Interview-
instructed to probe for two reasons.)
ers were instructed to probe for two reasons.) The first
In their study on the work activities of children in a
reasons given are shown in Table 4. The responses have
village in Nepal, Nag et al. (1978) reported that girls are
been broadly divided into economic, socioeconomic, and
more useful than boys in the household and familial
sociopsychological and cultural reasons. The old age se-
work. While this may be true, the reasons for wanting
curity value of children has been categorized as socio-
daughters given as first responses to an open-ended
economic because help in old age can be both economic
question (shown in Table 5) do not appear to support
and social. For poor parents, help from children is pri-
the high economic value of daughters. The second re-
marily economic. For parents who are not constrained
sponses, not shown, do not indicate the economic im-
financially, though, help in old age is noneconomic, often
portance of daughters either (Karki, 1982: 76). Both the
taking the form of companionship, for example, or emo-
first and second responses suggest the importance of
tional support. The reasons given for wanting sons in
noneconomic reasons for wanting daughters. This pat-
Table 4 are primarily noneconomic for the rural respon-
tern is much more pronounced among rural than urban
dents, but the urban respondents more often gave so-
respondents. It is only the urban wives who realize the
cioeconomic reasons, particularly the old age value of
practical utility of girls, although slightly more urban than
sons. In the rural area, the religious value of sons was
rural husbands mentioned this reason for wanting girls,
emphasized more, which is not surprising because re-
as well.
ligion has a great influence on the traditional way of life
A very high proportion of rural respondents stressed
in the rural population. With increasing urbanization,
the religious value of daughters. The birth of a daughter,
the religious influence may gradually erode.
in some Nepalese communities, is hailed as an event that
Nearly 90 percent of the respondents gave a second
will bring prosperity (Thapa, 1980), and some religious
reason for wanting sons. The second responses, not
occasions require the presence of females. For instance,
shown in the table, reverse the picture because the re-
Tihar, the second most important festival of the Hindus,
spondents who gave a noneconomic reason first for
requires the presence of one's sisters to make it a happy
wanting sons gave a socioeconomic reason second (see

Table 4 First responses to an open-ended question on the

reasons for wanting sons Table 5 First responses to an open-ended question on the
reasons for wanting daughters
Percent
Percent
Wives Husbands
Wives Husbands
Response Rural Urban Rural Urban
Response Rural Urban Rural Urban
Economic reasons 10 5 11 10
Practical help 3 2 4 2 Economic reasons 10 44 16 17
Inherit family debts and assets 4 2 4 8 Practical help 10 44 16 17
Sons earn income 3 1 3 0 Socioeconomic reasons 3 6 2 2
Socioeconomic reasons 24 47 18 47 Old age security 3 6 2 2
Old age security 24 47 18 47 Sociopsychological and
Sociopsychological and cultural reasons 86 51 81 80
cultural reasons 67 48 71 43 For religious occasions 53 25 52 37
Religious ritual reasons 60 47 64 27 Avoid loneliness 13 19 12 26
Family continuity 5 1 6 13 Make relatives 19 7 15 17
Happiness/satisfaction 1 0 1 2 Othera 1 0 2 0
Social status/sons' achievements 1 0 0 1 Total 100 100 100 100
Total 100 100 100 100 (Number of respondents)3)(63) (175) (201n) (86)
(Number of respondents) (683) (188) (216) (92)
Note: Total percents may not add to 100 due to rounding.
Note: Total percents may not add to 100 due to rounding. 'NO specific reason for wanting girls; just want girls.

174 Studies in Family Planning

This content downloaded from

103.173.2.149 on Sat, 11 May 2024 01:48:49 +00:00
All use subject to https://about.jstor.org/terms
Table 6 Respondents who want no more children, by number of living children and living sons

Wives Husbands

Rural Urban Rural Urban

No. of No. of % want % want % want % want
living living No. of no more No. of no more No. of no more No. of no more
children sons all wives children all wives children all husbands children all husbands children
1 0 73 8.2 28 7.1 30 3.3 10 10.0
1 75 20.0 23 4.3 29 3.4 9 0.0
Total 148 14.2 51 5.9 59 3.4 19 5.3
2 0 26 15.4 8 37.5 4 25.0 0 0.0
1 81 82.7 39 61.5 22 77.3 18 44.4
2 35 74.3 13 84.6 10 50.0 4 100.0
Total 142 68.3 60 63.3 36 63.9 22 54.5
3 0 14 7.1 4 50.0 6 16.7 3 0.0
1 57 86.0 12 75.0 15 93.3 7 100.0
2 65 96.9 19 94.7 19 100.0 14 92.9
3 23 95.7 0 0.0 7 85.7 0 0.0
Total 159 84.9 35 82.9 47 85.1 24 83.3
4or
more 0 3 33.3 1 100.0 0 0.0 1 100.0
1 40 95.0 5 80.0 13 100.0 3 66.7
2 69 98.6 17 100.0 23 95.7 11 100.0
3 66 100.0 11 100.0 18 100.0 6 100.0
4or
more 56 98.2 8 100.0 20 100.0 6 100.0
Total 234 97.4 42 97.6 74 98.6 27 96.3
Total 683 70.4 188 59.0 216 63.9 92 64.1

occasion. The sisters ceremonially offer Tiko (a mark on
the forehead) to one another and garlands to their broth-
ers, and daan (gifts) are exchanged (Department of Tour-
ism, 1980). If there are no sisters in the family, the broth-
ers request women or girls from the neighborhood to
put Tiko on their foreheads and receive a benediction,
but that kind of arrangement is not really satisfactory.
There are probably more festivals in Nepal than
anywhere else in the world (Anderson, 1971), and for
proper observation of almost all of them the presence of
a young unmarried girl is necessary. On such occasions,
by offering Tiko and some daan in the form of metal (par-
ticularly copper coins) to a young unmarried girl, a per-
son is believed to earn punya (religious merit) for the next
life. Parents in the neighborhood of Kathmandu city also
stress that they are supposed to support their daughters
and receive nothing but love and religious merit in return
(Schuler and Goldstein, 1986: 68).
The perception that daughters are valuable for avoid-
ing loneliness is stronger among urban as compared to
rural parents. More urban than rural respondents men-
tioned this as a value of daughters. Girls, it is felt, are
more playful and enjoyable as children than boys.
Another important reason given for wanting daugh-
ters is that they increase the size of the extended family
by adding relatives. The urban wives seem to attach less
importance to this than the rural wives, but a slightly
greater proportion of urban as compared to rural hus-
bands feel that this is an important reason for having
daughters. Relatives can also be added to the family
through sons, but in the Hindu society of the Nepalese,
gaining relatives through daughters has a more signif-
icant connotation. According to the Hindu scriptures,
parents earn religious merit when they perform Kanya-
daan (give a virgin daughter to a boy in another family).
This is probably one reason for early and universal female
marriage in Nepal.
Effect of Current Sex Composition of
Children on Fertility
The effect of sex preference on fertility is examined by
considering the percent of respondents who said that
they want no more children (Table 6). In Nepal, 60-70
percent of respondents wanted no more children, and
the percent wanting no more children increases with the
number of living children. Within each living children
group, the sex composition of the family influences future
childbearing desires. For instance, among respondents
with two living children, the proportion wanting no more
children among those who have one son is two to three
times higher than the proportion among those with no
sons. In larger families, the differences are even greater.
In the group of respondents with three children, over
five times as many of those with one son want no more
children as compared with women who have no sons.
Furthermore, respondents with two or more living sons
are much more likely to want no more children than those
with one son. Despite small numbers, these relationships
generally hold for respondents of both sexes and in both
areas (Table 6).
Effect of Son Preference on Contraception
The effect of son preference on contraceptive use is in-
vestigated by considering the timing of first use and the

Volume 19 Number 3 May/June 1988 175

This content downloaded from

103.173.2.149 on Sat, 11 May 2024 01:48:49 +00:00
All use subject to https://about.jstor.org/terms
Table 7 Parity at first contraceptive use among Nepalese Table 8 Respondents who reported having living sons when
couples they first practiced contraception

Percent responding Percent

Number of live births prior Wives Husbands Number of Wives Husbands

to contraceptive use Rural Urban Rural Urban living sons Rural Urban Rural Urban

1 10 5 13 4 0 2 0 0 0
2 13 15 13 26 1 30 35 47 44
3 21 45 7 44 2 35 55 13 48
4 15 25 20 15 3 17 10 7 7
5 15 10 13 7 4ormore 17 0 33 0
6+ 27 0 33 4 Total percent 100 100 100 100
Total percent 100 100 100 100
Total number of respondents 48 20 15 27
Total number of respondents 48 20 15 27 Mean number of living sons 2.2 1.8 2.4 1.6
Mean number of live births
before adopting contraception 4.1 3.2 4.3 3.1 Note: Total percents may not add to 100 due to rounding.

Note: Total percents may not add to 100 due to rounding.

variations in percent users. Despite the fact that the in-

formation on contraception may be biased, as discussed Table 9 Respondents who reported having living daughters
earlier, it seems worthwhile to investigate when a couple when they first practiced contraception
first practices contraception and whether their decision Percent
is influenced by family size composition. Nepalese cou-
ples generally believe that family planning should begin Number of Wives Husbands
living daughters Rural Urban Rural Urban
only after they achieve their desired family size and sex
0 21 10 13 11
composition. 1 25 50 27 48
2 30 30 40 33
To most, family planning means stopping child-
3 21 10 7 4
bearing completely rather than spacing births. This is 4ormore 4 0 13 4

probably why people who are using birth control before Total percent 100 100 100 100

their first birth may be especially reluctant to admit it. Total number of respondents 48 20 15 27
Mean number of living daughters 1.7 1.4 1.9 1.4
No respondents reported using any form of birth control
before the birth of a first child (Table 7). In addition, the Note: Total percents may not add to 100 due to rounding.

Table 10 Respondents who ever practiced contraception, by number of living children

and living sons

Husbands who ever used

Wives who ever used contraceptives contraceptives
Number of Number of Rural Urban Rural Urban
living children living sons N % N % N % N %
1 0 73 1.4 28 0.0 30 0.0 10 0.0
1 75 5.3 23 4.4 29 6.9 9 11.1
Total 148 3.4 51 2.0 59 3.4 19 5.3
2 0 26 0.0 8 0.0 4 0.0 0 0.0
1 81 3.7 39 7.7 22 9.1 18 33.3
2 35 14.3 13 7.7 10 0.0 4 50.0
Total 142 5.6 60 6.7 36 5.6 22 36.4
3 0 14 0.0 4 0.0 6 0.0 3 0.0
1 57 7.0 12 16.7 15 6.7 7 71.4
2 65 6.2 19 31.6 19 0.0 14 42.9
3 23 4.4 0 0.0 7 0.0 0 0.0
Total 159 5.7 35 22.9 47 2.1 24 45.8
4or
more 0 3 0.0 1 0.0 0 0.0 1 0.0
1 40 7.5 5 20.0 13 15.4 3 0.0
2 69 11.6 17 23.5 23 8.7 1 1 45.5
3 66 10.6 11 18.2 18 5.6 6 33.3
4 or
more 56 14.3 8 0.0 20 25.0 6 0.0
Total 234 11.1 42 16.6 74 13.5 27 25.9
Total 683 7.0 188 10.6 216 6.9 92 29.4

Note: The denominators used to calculate

176 Studies in Family Planning

This content downloaded from

103.173.2.149 on Sat, 11 May 2024 01:48:49 +00:00
All use subject to https://about.jstor.org/terms
rural respondents had, on average, approximately four
births before adopting contraception. Urban couples had
about three births before beginning family planning.
The family sex composition among the respondents
who ever practiced birth control reveals that most couples
have at least one living son (Table 8) before practicing
birth control. Some respondents in both rural and urban
areas had practiced birth control without having a living
daughter (Table 9) but only one had done so without
having a living son. Despite the small numbers the pat-
tern is similar to that in a sample of 214 women studied
in a Pokhara sterilization camp (Miller and Joshi, 1974).
From this analysis, it appears that it would be quite dif-
ficult, if not impossible, to persuade most couples to
adopt family planning if they did not already have living
sons.
The data presented so far do not reflect the extent
of contraceptive use and how it is affected by the number
and sex composition of living children. This is done by
considering the percent of respondents who practice
contraception (Table 10). A small proportion of Nepalese
couples practice contraception, and because of the small
numbers involved, there is no clear pattern. However,
respondents without a living son do not practice con-
traception at all, except for one woman in the rural area,
which might be a reporting error. Apparently, even those
respondents who have one child start thinking about
birth control measures, if the child is a son.
Summary and Discussion
This small-scale study and a large nationally represen-
tative sample survey have indicated strong evidence of
son preference in Nepalese society. The young couples
(i.e., female respondents aged 15-39) in Nepal do not,
however, seem to desire many (i.e., more than three)
sons, even though they value them very highly. They
do not want only sons, either. Their ideal family size is
three, and their ideal sex composition of children is two
sons and one daughter. Most desire at least one living
child of each sex.
Nepalese parents prefer sons to daughters because
of their culture and the various roles that sons play in
their family life. It is only one's own son who can perform
death and post-death rituals to ensure that the gate of
heaven will be opened for parents. In addition, sons
continue the family name and provide support in old
age. For these and other reasons, it is not surprising to
often hear people say, "Let it be late, but let it be a son. "
When respondents were asked what their ideal
family size and sex composition would be if they were
to start their family life all over again, about 6 percent
of rural and 3 percent of urban couples did not want
any girls at all. The rest, however, who wanted at least
one girl in their ideal family, said that they wanted girls
This content downloaded from
103.173.2.149 on Sat, 11 May 2024 01:48:49 +00:00
All use subject to https://about.jstor.org/terms
primarily for religious occasions and not for econom-
ic benefits. This was more strongly expressed in the ru-
ral area.
The economic motive to have many children may be
changing even in the rural areas of Nepal. The existence
of such a pattern is supported by the family size data:
both the mean ideal family size of about three and the
mean desired family size (includes number of living chil-
dren, current pregnancy, and number of additional chil-
dren wanted) of about four are low for a population that,
for the most part, does not practice contraception. Also,
the reasons expressed for wanting sons and daughters
support the suggestion that the economic motive for
having children may be weakening.
The author's own observation of the deteriorating
hills environment and the perceived economic costs of
raising children was articulated by the respondents
(Karki, 1982: Chapter 5), who spoke in support of a shift
toward smaller family size. The economic situation of
the people in the hills has, unfortunately, become worse
in recent years. Due to the recession of the forests every
year, it takes markedly longer to fetch wood and water.
Also, the villagers now feel the shortage of good agri-
cultural land. The very high rise in the value of land has
made its purchase by villagers practically impossible be-
cause of its prohibitive cost. Those families who had good
land are now only moderately well off, because as gen-
erations mature, inheritance customs continually divide
large estates between several sons; thus, more sons
means less land for each. Many villagers, therefore, link
the poverty of the people to large family size.
Family size preferences and the decision to practice
contraception are both affected by son preference-that
is, the desire to have at least one living son. The pro-
portion of couples with at least one living son who desire
additional children is very low compared to their coun-
terparts with no son. Also, virtually no couples without
sons, in both the urban and in the rural samples, were
practicing contraception. Among those who reported
current contraceptive use, the mean number of living
sons was higher than the mean number of living daugh-
ters.
Although Nepalese couples express a strong desire
for sons and their intentions with regard to future family
building are affected accordingly, such a phenomenon
still does not account for low contraceptive use because,
even among those who have two or more sons, the pro-
portion practicing contraception is much smaller than
the proportion wanting no more children. A large gap
exists between the desire to cease childbearing and any
action to realize that goal. Moreover, the percent of cou-
ples who do not have a son decreases quite rapidly as
the number of living children increases. Thus, only 10
percent of respondents with three children have no sons.
It is this small group whose contraceptive and fertility
behavior would be affected by the lack of a son.
Volume 19 Number 3 May/June 1988 177
Notes
This paper is based on the author's Ph.D. thesis in demography, "Fer-
tility and the Value of Children: A Study of Rural and Urban Populations
in Nepal" (Karki, 1982). Research costs were covered by a generous
grant from the Ford Foundation. The financial support is gratefully
acknowledged. The author is grateful to the late Professor D.V. Glass
of the London School of Economics, who was his supervisor initially,
for providing much encouragement and support during the planning
of the project. The author also wishes to thank the referees and members
of the Editorial Committee of Sttudies in Family Planniing for their helpful
comments.
An earlier version of this paper was presented in a lecture to the
staff of the Centre for Population Studies, London School of Hygiene
and Tropical Medicine, February 1982.
1 For a review of the literature on preference for sons around the world,
Karki, Y.B. 1982. "Fertility and the value of children: A study of rural
see Williamson (1976).
2 The population of Nepal has grown from 8.3 million in 1952/54 to
9.4 million in 1961, 11.6 million in 1971, and 15.0 million in 1981.
The average annual rates of growth in percentages are 1.6, 2.1, and
2.7 during the 1952/54-61, 1961-71, and 1971-81 census periods, re-
spectively. See Central Bureau of Statistics (1985).
3 For more, see Chapter 3 in Karki (1982).
4 See Chapter 5 in Karki (1982).
References
Anderson, Mary M. 1971. The Festivals of Nepal. London.
Arnold, F., R. Bulatao, C. Buripakdi, B.J. Chung, J.T. Fawcett, T. Iritani,
S.J. Lee, and T.S. Wu. 1975. The Value of Children: A Cross-National
Study. Voltum2e 1, Introduction and Comparative Analysis. Honolulu:
East-West Population Institute.
Bennett, Lynn. 1976. "Sex and motherhood among the Brahmans and
countries." Studies in Famiilly Planning 3, 4: 70-76.
Chhetris of East Central Nepal." Journal of the Institutte of Nepal anid
Asian Studies (Special issue). Contributions to Nepalese Studies,
Volume 3. Kathmandu: Tribhuban University.
Campbell, J. Gabriel, Ramesh Shrestha, and Linda Stone. 1979. T/le
Use and Misuse of Social Science Research in Nepal. Kathmandu:
Research Centre for Nepal and Asian Studies, Tribhuban University.
Cassen, R.H. 1980. India: Population, Economny, Society. New York:
Holmes and Meier.
Central Bureau of Statistics. 1978. The Demtiographic Samriple Suirvelj of Nepal.
Kathmandu: National Planning Commission Secretariat, His Maj-
esty's Government of Nepal.
. 1985. Intercemisal Changes of Somne Key Censtis Variables, Nepal,
1952/54-1981. Kathmandu: His Majesty's Government of Nepal.
Coombs, C.H., L.C. Coombs, and G.H. McCleland. 1975. "Preference
scales for number and sex of children." Population Studies 29, 2: 273-
298.
Coombs, L.C. and T.H. Sun. 1978. "Family composition preference in
a developing culture: The case of Taiwan." Population Studies 32, 1:
43-64.
Department of Tourism. 1980. Nepal: Pacific Travel News. Kathmandu:
His Majesty's Government of Nepal.
178 Studies in Family Planning
This content downloaded from
103.173.2.149 on Sat, 11 May 2024 01:48:49 +00:00
All use subject to https://about.jstor.org/terms
Eckholm, E.P. 1976. Losing Ground: Environmental Stress and World Food
Prospects. New York; W. W. Norton Company.
Freedman, R. and L.C. Coombs. 1974. Cross-Cultuiral Comparisons: Data
oni Two Factors in Fertility Behavior. New York: The Population Coun-
cil.
Gideon, H. 1962. "A baby is born in the Punjab.'' American Anthropologist
64.
Goldman, Noreen, Ansley J. Coale, and Maine Weinstein. 1979. "The
quality of data in the Nepal Fertility Survey." WFS Scientific Reports.
No. 6. Voorburg: International Statistical Institute.
Heer, D. and H. Wu. 1974. "The effect of infant and child mortality
and preference for sons upon fertility and family planning behavior
and attitudes in Taiwan." In Population and Social Change in South
East Asia. Ed. J.F. Kanter. Lexington: Lexington Books.
and urban populations in Nepal." Ph.D. thesis, London School of
Economics.
Mamdani, M. 1973. The Myth of Population Control: Familly Caste and Class
in an Indian Village. New York: Monthly Review Press.
Miller, R. and P. L. Joshi. 1974. Laparoscopic Sterilization in Camp Setting:
The Nepal Experience. Washington, DC: Agency for International
Development.
Nag, M., B. White, and R.C. Peet. 1978. "An anthropological approach
to the study of the economic value of children in Java and Nepal."
Cuirrenit Anthropology 19: 293-306.
Nepal Family Planning/Maternal-Child Health (FP/MCH) Project. 1977.
Nepal Fertility Survey,. First Report (1976). Kathmandu: Ministry of
Health.
Operations Research Group (ORG). 1971. An All-India Survey of Fanily
Planning Practices. Baroda: ORG.
Park, Chai Bin. 1978. The Fourth Korean Child: The Effect of Son Preference
on Subsequent Fertility. Report No. 87. Honolulu: East-West Pop-
ulation Institute.
Repetto, R. 1972. "Son preference and fertility behavior in developing
Schuler, Sidney Ruth and Melvyn C. Goldstein. 1986. "Family planning
in Nepal from the user's and nonuser's perspectives." Studies in
Family Platnning 17, 2: 66-77.
Sharma, D.V.N. and A.K. Jain. 1974. "Preference about sex of children
and use of contraception among women wanting no more children
in India." Dem1ography India 3, 1: 81-104.
Thapa, Dhruba B.S. 1980. "Law and population." In Population of Nepal.
Country Monograph Series, No. 6. Bangkok: Economic and Social
Commission for Asia and the Pacific, United Nations.
Vlassoff, M. 1979. 1980. "Labor demand and economic utility of chil-
dren's case study in rural India." Population Studies 33, 3: 415-428.
Vlassoff, M. and C. Vlassoff. 1980. "Old age security and the utility
of children in rural India." Population Studies 34, 3: 487-499.
Ware, Helen. 1978. The Economic Value of Children in Asia and Africa:
Comparative Perspective. Papers of the East-West Population Institute,
No. 50. Honolulu: East-West Population Institute.
Williamson, Nancy E. 1976. Sons or Daughters: A Cross-Cultural Survey,
of Parental Preferences. Beverly Hills: Sage Publications.
World Bank. 1981. World Development Report, 1981. Washington, DC:
World Bank.