The missing girls


back to issue

THE deficit of women in India’s population has been documented ever since the first decennial enumeration of people was conducted in the late 19th century. Over the span of more than a century, the deficit has progressively increased as evident from the sex ratio of the population; the number of women per 1000 men steadily declined from 972 in 1901 to 933 in 2001. India shares with China (and other South Asian countries, with the exception of Sri Lanka) this phenomenon of deficit of women in the population. Both the Oriental societies are patrilineal, exhibit strong son preference, and men enjoy higher status relative to women. Throughout the rest of the world, women outnumber men by three to five per cent.

The deficit of women in India and the possible factors responsible for it have aroused attention among demographers, social scientists and women activists who have tried to understand the phenomenon in terms of under-enumeration of women in the census counts, sex-selective migration, sex ratio at birth, as well as sex differentials in mortality. Historically, under-enumeration, especially of child brides in certain regions where such a custom is observed, finds favour with many analysts of census data as a key factor accounting for deficit of girls aged 10-14 years.

On the other hand, there is no evidence to support the likelihood of sex differentials in migration (implying greater outmigration of women) or greater than the usual masculinity of sex ratio at birth. What has been convincingly demonstrated is that the primary factor contributing to the deficit of women in India is the anomalous higher mortality among women compared to men. Indian women continue to experience higher mortality than men from age 12 months to almost up to the end of the reproductive period. Again, elsewhere in the world, women generally experience lower mortality than men at almost all ages such that the life expectancy at birth of women is greater by five to eight years compared to that of men (Visaria 2002).

This sex differential in mortality in India is a result of the discriminatory treatment received by girls and women, which more than offsets their natural advantage over men. Within India, both the social practices and cultural ethos that undervalue women are stronger in some regions than in others. In an almost contiguous belt extending from northwest of India to parts of Rajasthan, Gujarat and Maharashtra, the undervaluation of women is evident in the sex ratio of their population and in their juvenile sex ratio. In fact, an increase in the deficit of young girls noted in the Censuses of 1981 and 2001 is indicative of a strong possibility that the traditional methods of neglect of female children is increasingly being replaced by not allowing female children to be born.



If only all births in the country or a region were registered, it would be possible to calculate the sex ratio at birth and surmise the extent to which female births are prevented from occurring, since sex ratio at birth is biologically determined and globally ranges between 102 and 107 male births to 100 female births. Unless there is a conscious effort at intervention by human beings, the sex ratio at birth is most unlikely to change significantly over a long period of time. However, in India except for states like Kerala, Tamil Nadu and Goa, where registration of births is nearly total, elsewhere births are far from systematically or fully registered. A significant proportion of births occurring at home go unregistered.

In the absence of accurate information on such vital events, we have to depend on the decennial censuses for data on the number of children in the age group 0-6 by sex and region (up to district level) to estimate juvenile sex ratios. Other things being equal, the juvenile sex ratio does not undergo significant changes over time. In India, with a somewhat faster decline in female child mortality compared to male child mortality, the juvenile sex ratio should over time become more favourable to girls. However, contrary to this expectation, in the contiguous region from north to west of the country, the deficit of girls increased (and not decreased) between 1981 and 2001.



This is also the region where historically, deficit of women in the total population has been quite marked. The adverse juvenile sex ratio in itself was no surprise. What triggered the alarm bell was that despite overall improvements in the mortality situation, and a greater increase in life expectancy of women compared to that of men indicating that women have gained more than men from improved health care, the deficit of girls increased as shown in Table 1.



Sex Ratio (females per 1000 males) of Population for all age groups and for 0-6 age group for India and select states, 1981-2001

State/Census Year





All ages

0-6 years

All ages

0-6 years

All ages

0-6 years

All India







Himachal Pradesh











































Between 1981 and 2001, the sex ratio of the total population at all India level as well as in the six states in the northern and western parts of the country remained virtually the same with the exception of Gujarat where the deficit of women somewhat increased during this period. At the same time the deficit of young girls, which was not evident in 1981 except in the traditionally and historically masculine states of Haryana and Punjab, became quite stark by 2001 in Himachal Pradesh, Gujarat, Rajasthan and Maharashtra. The deficit of girls further increased in Haryana and Punjab.



In fact, according to the 2001 Census, there were 49 districts in the country where for every 1000 male children aged 0-6 years there were less than 850 female children. A majority (38) of these districts were located in just three northern and western states of Punjab, Haryana and Gujarat (Census of India 2001). The decline of 60 to 83 points in the juvenile sex ratio between 1991 and 2001, or in a span of just one decade observed in many of these districts, cannot be explained solely by the discrimination against girls that has been practiced in this region for several decades because at no other time in the history of census taking in the region has the sex ratio of children declined so drastically.

The distortion in the sex ratio was brought out starkly in an analysis of the data from the second National Family Health Survey (NFHS) conducted in 1998-99 by Arnold, Kishor and Roy (2002). They showed that at an all India level, the male to female sex ratio of the last births was 1434 (or 697 girls for every 1000 boys) among currently married women who did not want any more children, which was much higher than the sex ratio of 1069 for all the other births. However, there were significant inter-state variations and in states of Haryana, Punjab and Gujarat the strong son preference was manifested in the sex ratio of last births, which ranged between 1752 and 2173, implying that for every 1000 girls who were last births there were more than 1750 boys who were last births, reflecting a strong effect of gender preference on reproductive behaviour.



In a recently completed study in Mehsana district in Gujarat and Kurukshetra district in Haryana, undertaken with the support of HealthWatch Trust, it was observed that the last births had a stronger preponderance of boys than all other births. Although more than twice as many boys as girls were reported among the last births by most groups of women, among those women who belonged to upper castes, whose families were landed and who were literate, there were more than 240 males for every 100 girls in the last births (Visaria 2003). This distortion was very likely due to the use of sex-selective abortions which helped parents get rid of unwanted daughters, or due to avoiding having children once the minimum desired number of sons were born. In either case, the preference for sons was evident in the behaviour of the couples.

In the Gujarat and Haryana study it was further noted that as the birth order increased, the preponderance of male children increased. Although the sex ratio of the first birth was greater than the normal acceptable range of 104-107 boys per 100 girls, by the time women had their third or higher parity child, the chance of that birth to be a male birth was greater by 30 to 50 per cent. The preponderance of boys among the second and the third child was much greater for women who were educated beyond primary level, women who were not engaged in any economic activity or who reported themselves as housewives, women who belonged to upper castes and those whose families were landed.

The overall assertion made by many that the sex ratio is much more adverse among the economically and educationally well-off population groups compared to others does indeed hold true according to this study.



The neglect of and discriminatory behaviour against girls leading to excess female mortality has been widely documented in several studies (Visaria 1971, Miller 1989, Das Gupta 1987, Kishor 1995). But the recent increase in the juvenile sex ratios discussed above has very likely resulted from the rapid spread of ultrasound and amniocentesis tests for sex determination in many parts of the country, followed by sex-selective abortions. Because of the simplicity of these tests and their easy availability on the one hand and strong son preference on the other, female-specific abortions appear to have become popular and widely used.

It is important to understand the emergence of this phenomenon in a wider perspective. India pioneered in legalizing induced abortion under the medical termination of pregnancy (MTP) Act, 1971 that specifies the reasons for which an abortion can legally be performed. The act specifies clearly who can legally perform the abortions and the kind of facilities in which they can be carried out. The stipulated conditions are such that abortions performed by trained doctors who are not registered in facilities specifically approved for abortion services are termed illegal. According to Chhabra and Nuna (1993), illegal abortions in India may be 8 to 11 times as high as legal abortions.

While the intention is to provide women with safe, legal, timely abortion services, given the stringent nature of the MTP Act, many safe abortions may be classified as not legal. At the same time, the availability of and access to legal abortion services is so limited for women living in remote rural areas that in the three decades since the passing of the act, many abortions not only take place outside its ambit but are often performed in unsafe conditions, leading to post-abortion complications and sometimes death.



Abortion can be legally availed if a pregnancy carries the risk of grave physical injury to woman, endangers her mental health, when pregnancy results from a contraceptive failure, from rape, or when it is likely to result in the birth of a child with physical or mental abnormalities. Methods to detect deformities in the foetus, such as amniocentesis and sonography that use ultrasound technology providing valuable and early information on a range of physical problems, have become available in the country, thanks largely to the private medical practitioners who are eager to use newer technologies for diagnosis. However, technologies that help detect physical or mental abnormalities in the unborn child can also identify the sex of the foetus at no extra cost or effort.

There was increasing indirect evidence from some parts of India that termination of pregnancies was resorted to not for the reasons stated under the MTP Act, but because there is a strong son preference leading to female-selective abortions. The gender bias was flagrantly aided by a combination of medical technology that helped detect the sex of the foetus on the one hand and the liberal abortion law that helped couples to abort the female foetus. In view of this, the Indian government, responding to the petition made by NGOs and women’s groups, passed an act prohibiting the practice of prenatal diagnosis of sex of the foetus (Pre Natal Diagnostic Techniques – PNDT – Act of 1994).



Under the act individual practitioners, clinics or centres cannot conduct tests to determine the sex of the foetus or inform the couples about it. However, in spite of putting monitoring systems in place both at the state and central levels, and with the act in place for 6-8 years at the time of the 2001 Census, it is evident that in many places the act has been violated with impunity. Since the two activities of sex detection of the foetus and abortion need not be linked at the stage of using the services, it has become possible to evade the law in connivance with the clinics having ultrasound facilities and doing sonography.

Judging by the hoardings even in small towns and the regular advertisements in local newspapers and magazines, it was evident that clinics conducting sex determination tests had mushroomed in many towns in the states of the northwestern belt before the passing of the PNDT Act in 1994. The open advertisements have now disappeared but the lucrative practice seems to flourish unabated by simply going underground as evident from the continued decline in the sex ratio of children 0-6 years of age.

Anecdotal evidence suggests that strong competition has reportedly led to a reduction in charges for availing these services, which has worked to the advantage of potential clients. Easy access is, to a certain extent, a response to an increasing demand and female foeticide apparently has replaced the old tradition of culture of neglect of the girl child, practice of infanticide among certain communities and sex differentials in the provision of medical care.

Although the release of the 2001 Census results has sparked serious concern about the widespread use of ultrasound and amniocentesis tests to detect the sex of the foetus followed by sex-selective abortions, our understanding of many issues around this practice at the level of the household or from the perspective of women who undergo such abortions is extremely limited. It is also limited about what actually compels couples or their families to resort to such a practice, who the real decision-makers in the family are, what impact aborting a female foetus has on the physical or mental health of the woman who typically undergoes abortion in the second trimester of her pregnancy.

Our understanding of how the interlinkages of sex-selective abortion and decline in fertility or the desired number of children are perceived and articulated remains limited. The question often raised is: does the desire for fewer children compel parents to produce children of the sex that they want or that conform to the societal norms and regulate their fertility behaviour accordingly? The qualitative data collected by conducting 44 focus group discussions, involving more than 400 women belonging to diverse socio-economic and educational groups in rural Gujarat and Haryana, has provided insight on some of these issues.



During the discussions with women both in Gujarat and Haryana, it was clear that a majority of women accepted the outcome of the first pregnancy – whether it was a boy or a girl. However, if the first-born child was a daughter, then the upper caste women were overtly or covertly pressurized to ensure that the second and/or the third child was a boy and to take appropriate measures. Although this pressure was lower among the lower castes, many among them have either started emulating women from the upper castes or thinking the same way.

Thus, son preference was evident among all social groups in both Gujarat and Haryana even when the desired number of children had come down to two or three. No group of women indicated that they would want more than two or three children. They came up with fairly rational explanations about why many children are not desired in the present times and situation. However, in spite of wanting fewer children, women candidly admitted preference for male children.



In order to minimize the influence of other members of the family on their decision, we asked women to imagine a hypothetical situation of having complete freedom to choose the number and sex composition of their children. Among those wanting three children, the overwhelming response was for two sons and one daughter. Similarly, some who indicated that they would like to have only two children preferred at least one of them to be a son. If, however, the two children turned out to be girls then they would almost certainly opt for a third child in the hope that it would be a boy. Women did discuss the possibility that not all sons support parents in their old age, and yet the desire for a son was very strong among women of all social groups.

As one backward community woman in Gujarat put it: ‘Yes, we wait for the son. We must have a son, howsoever he may turn out to be. We would always hope for a son. After all, the daughter will go away after her marriage. The son will stay with us and take care of us.’

Women from the upper castes which practice dowry (Chaudhary in Haryana and Chaudhary Patel in Gujarat), even voiced that if the first child born to them was a boy, they would be satisfied with just one child. The menace of the dowry system was a strong deterrent to having girls along with a fear that the daughter might be sent back to the parental house if her in-laws were not satisfied with the dowry or for any other reason.

‘There is trouble for daughters. They may find a good family or a bad family after their marriage. They [daughters] may come back home. If they have trouble with their in-laws, they may be sent back. In earlier times, the women used to do backbreaking labour, look after the cattle after their marriage. These days girls do not do like to do that. If there is an economic problem, the in-laws will send the girl back to her parental home. So, for the parents, a girl always remains a reason for tension’ (Patel woman from Gujarat).

‘A girl requires a dowry for her marriage, which is a cause for anxiety. Finding a suitable groom and hoping that she will settle down happily in her new home is always a source of worry for parents’ (a woman from Haryana).



That this near universal desire for more sons than daughters does get translated in actual behaviour was evident from the sex ratio of live births discussed earlier. In the focus group discussions too, women from all communities categorically indicated that if the first born child was a daughter, then the couple would want to and do find out the sex of the next child. Women knew where to go for sex determination tests, how much the tests cost, and so on. They were aware that such tests were not done in public hospitals. One had to go to private facilities, a majority of which according to them also provided abortion services. Women were able to describe the sex determination procedure quite accurately.

Women also indicated that after the birth of a daughter when they became pregnant again, there was some pressure from the elders in the family to ensure that the next child be a boy. Women themselves also wanted to produce a son. There is a deep internalization of patriarchal values that are linked to their sense of security. The son preference was internalized to such an extent that women had no hesitation in saying that they would want the sex of the foetus to be known if they had already given birth to one daughter.



Although most of them had to consult and get permission from their husbands (partly because the sex determination test involved a cost of few hundred rupees), they themselves saw nothing wrong in finding out the sex of the foetus. As articulated by a Kshatriya woman from Gujarat or a Chaudhary woman from Haryana: ‘We have to go for test if the first child is a girl. If we don’t go for the test, we may end up giving birth to three or more daughters in the false hope of getting a son.’ ‘Women definitely get the test done… if it is a girl they abort the foetus and if it is a boy, they keep the baby. Everybody knows about the test… the women themselves want to know whether they are carrying a male or a female child.’

Although the parents or parents-in-law of the women probably had given birth to several children, it appears that they do not wish their daughters-in-law to do so. As the women indicated, the facilities (for sex determination and abortion) did not exist in earlier times and so the parents had no choice but to have several children. The mother-in-law would even suggest that the daughter-in-law get the sex determination test done, especially after producing one daughter. The parents of the woman, however, had no say in the matter except for wishing that their daughters give birth to at least one son because their wellbeing and status in the families of the in-laws depended to a great extent on bearing sons.

‘Mothers-in-law have changed with the times. They are also aware of the rise in prices. They may have raised several children in their time, but it’s difficult to raise more children today’ (Backward caste woman from Haryana).

‘If we already have one son and one daughter, the in-laws would ask us to go in for a test and if it is a daughter, they would even ask us to have an abortion’ (Chaudhary Patel woman from Gujarat).



When women were asked about the decision-making process if the foetus was that of a female child, the overwhelming response was that after one or two daughters if the woman was found to be pregnant with another girl, the pressure on her to abort was enormous from the extended conjugal family. Women indicated that the decision to abort a female foetus was almost entirely that of their husbands and/or mothers-in-law. By themselves, women could not take the decision to go in for abortion. Women who have virtually no decision making power apparently accept whatever is desired by their conjugal family including husbands and go along with the decision made for them by others.

However, we observed some differences between women belonging to higher social groups and those who belonged to scheduled caste and other backward communities with regard to the influence of the in-laws in these matters. The high caste women had to inform and consult their in-laws but the low caste women had to obtain the consent of only their husbands for abortion. The influence of the extended joint family was not so strong on the decision of the women from lower caste groups.

‘A woman cannot take a decision on abortion on her own. If the husband does not want a daughter then he would ask us to go in for abortion. And if he wants a daughter, then we keep the daughter. If the husband is ready to support us and stand by us, we can be firm and go for abortion or not for abortion. In any case we need to consult our husbands’ (Backward caste woman from Gujarat).

‘If the first two children are girls and the third one too is a girl then we need to take the permission of the elders to go in for abortion. We have to follow the advice of the elders’ (Patel woman from Gujarat).

Women also reported that sometimes they themselves desire to abort a female foetus because they already have one or two daughters. This feeling was stronger among women belonging to social groups such as Patel and Kshatriya who value sons much more than daughters. Although they themselves, without much hesitation, would opt for abortion, they still would have to get the permission of the elders of the family to exercise their wish.



The analysis clearly points to an all-pervasive collusion of social norms and technology. On the one hand son preference is strongly entrenched in Indian society, especially in the northwestern region, and on the other the wellbeing and status of girls is so precarious once they are married that couples avoid having girls at all costs. Facilities conducting sex detection tests with ultrasound machines have proliferated and are found even in some of the relatively large villages.

Despite spread of schooling among girls in recent decades, the patriarchal social structure survives. Women derive value and status only as mothers of sons. Their happiness and social status in the conjugal homes is dependent on producing sons. Women have internalized these roles and values to such an extent that even when they say that daughters take better care of parents or are more emotionally attached to the mothers, these statements ring hollow because in fact more sons than daughters are desired. In the pursuit of sons they have become, with some pressure from the families, consumers of the new ultrasound technology which allows them to choose and bear sons. The liberal abortion law of our country also provides an opportunity to abort the child of unwanted sex.



The shift to small family size, more recently evident in India, has not however been accompanied by a simultaneous shift in the economic and social pressures to have sons and avoid daughters. As stated by women both in Gujarat and Haryana, they desire and want fewer children while ensuring that at least one if not two of them are sons. This has also led to increased acceptance and use of sex selection tests to achieve parental preferences to have sons while not exceeding the desired number of children.

At the same time, the awareness about a ban on sex determination tests is fairly widespread among women in our study area. Many women also felt that the ban should be removed and couples should have the choice to decide the sex composition of their children. Women were well aware that the services are easily available from private providers and are within easy access. Government legislation against the use of ultrasound technology for sex detection has only driven it underground and raised the cost but it is extensively available and used for sex detection.



The cost is still affordable and in any case, as many respondents indicated, the cost of the test and related abortion is much lower than the cost of providing dowry and other life long presentations to a daughter after marriage. As one of our researchers pointed out: ‘The alarm bells ringing in the corridors of power about the missing girls in the demographic picture do no find an echo in the dusty bylanes of the villages of these districts’ (Chaudhury 2003).

The combined effect of patriarchal structure and values and the practice of getting rid of daughters is such that certain social groups in both Gujarat and Haryana have started facing a deficit of brides for their sons. Reportedly, a few men are forced to remain bachelors and for some others brides are being brought or bought by paying bride price from scheduled tribes and other groups from far away places, including other states. We have no hard evidence on the extent of this practice but it may become a lesson in social integration. However, despite the deficit of women, whose impact is being felt when procuring brides, the social norms have not yet undergone a shift.

As is evident, legislation banning the use of sex-determination tests has thus far not succeeded in deterring couples from seeking these tests or preventing medical practitioners from performing them. The prevalent social norms and practices do raise a number of questions. Does passing a national legislation to regulate prenatal diagnostic technologies and their misuse constitute an answer? Thus far, the law has been largely ineffective but will regulatory mechanisms clamped at all levels or better implementation prevent its misuse? Will impounding ultrasound machines in unregistered clinics and demanding detailed registers about their use in authorized clinics help in reducing their use for sex detection of foetus?

We need a concerted effort to address the bias against girls at the source and alter the underlying conditions that promote sex-selective abortions. However, this is an uphill task and every action and every group that can address it will contribute to improving the status of women in our society.



F. Arnold, S. Kishor and T.K. Roy (2002), ‘Sex Selective Abortions in India’, Population and Development Review 28(4), December, 759-785.

R. Chhabra and S.C. Nuna (1993), Abortion in India: An Overview. Veerendra Printers, New Delhi.

Manisha Chaudhury (2003), ‘Sex-selective Abortion in Haryana’, HealthWatch Trust (typescript), New Delhi.

Census of India (2001), Series-1, India, Provisional Population Totals, Paper 1 of 2001, Supplement, District Totals, Office of the Registrar General of India, New Delhi.

M. Das Gupta (1987), ‘Selective Discrimination Against Female Children in Rural Punjab, India’, Population and Development Review 13(1), 77-100.

Sunita Kishor (1995), ‘Gender Differentials in Child Mortality: A Review of Evidence’, in Monica Das Gupta, Lincoln C. Chen and T. N. Krishnan (eds.), Women’s Health in India: Risk and Vulnerability, Oxford University Press, Bombay.

B.D. Miller (1989), ‘Changing Patterns of Juvenile Sex Ratios in Rural India, 1961-1971’, Economic and Political Weekly 22, 2 June, 1229-1236.

L. Visaria (2002), ‘Deficit of Women in India: Magnitude, Trends, Regional Variations and Determinants’, The National Medical Journal of India 15, Supplement 1, 2002, 19-25.

L. Visaria (2003), ‘Sex Selective Abortion in the States of Gujarat and Haryana: Some Empirical Evidence’, HealthWatch Trust (typescript), New Delhi.

P. Visaria (1971), The Sex Ratio of the Population of India, Census of India 1961, Monograph No. 10, Office of the Registrar General, India, New Delhi.