by Amartya Sen
The concept of ‘missing women’ was devised to give some idea of the enormity of this phenomenon by focusing on the women who are simply not there, primarily due to unusually high mortality compared with normal female-male mortality ratios. The methodology involves finding some way – unavoidably rather rough – of estimating the quantitative difference between the actual number of women in these countries, and the number we could expect to see had the gender ratio in survival been similar, in these countries, to that in other regions of the world where there is not such a significant anti-female bias in health care and other social influences relevant for survival.*
For example, if we take the ratio of women to men in sub-Saharan Africa as the standard (on the grounds that there is relatively little bias against women in health care, social status and mortality rates in sub-Saharan Africa, even though the absolute mortality rates are very high for both men and women), then its female-male ratio of 1.022 can be used to calculate the number of missing women in women-short countries. With India’s female-male ratio of 0.93, there is a ‘relative deficit’ of women of 9 per cent of the Indian male population between that ratio and the standard used for comparison, that is, 1.02 (based on the sub-Saharan African ratio). This already yields a figure of 37 million missing women in India in 1986 (when I first did the estimation).4 Similarly, the use of the same ‘standard’ of 1.02 (based on the sub-Saharan African ratio) produces a figure of 44 million for missing women in China at the same point of time. By adding these country estimates together, it soon emerges that for the world as a whole the magnitude of shortfall of women easily exceeds 100 million, already ‘missing’ in 1986 (since then the numbers have grown with the growth of absolute size of the respective populations).5 Other standards and different demographic procedures can also be used, as has been done by Ansley Coale and Stephan Klasen. These procedures yield somewhat different numbers, but invariably very large ones. For example, Klasen’s total number for the earlier period (as in the estimates by Coale and Sen) was about 80 million missing women. For more recent years, Stephan Klasen’s method yields numbers that are larger than 100 million.6 Anti-female bias in care and mortality imposes a massive penalty across the world against the survival of women.
Natality Inequality and an Indian Divide
Even within the demographic domain, gender inequality can manifest itself not just in the old form of mortality asymmetry, but also in the new form of sex-specific abortions aimed at eliminating female fetuses. This ‘natality inequality’ reflects the fact that many parents want the newborn to be a boy rather than a girl, given a general preference for boys in many male-dominated societies. The availability of modern techniques to determine the gender of the fetus has made such sex-selective abortion possible and easy, and it has become very common in many societies. It is particularly prevalent in east Asia, in China and South Korea in particular, but there is evidence that it also occurs to a statistically identifiable extent in Singapore and Taiwan. It is beginning to emerge as a significant phenomenon in India and south Asia as well.
This ‘high-tech sexism’ has changed – and is continuing to change – the female–male ratios at birth. Compared with the biologically determined standard ratio of about 95 girls being born per 100 boys (which is the ratio we observe in Europe and North America since sex-specific abortion does not exist as a significant phenomenon there), Singapore and Taiwan have 92, South Korea 88, and China a mere 86 girls born per 100 boys.
Given the incompleteness of birth registration in India, it is not easy to get exactly comparable female-male ratios at birth, but the ratio of females to males among children can serve, inter alia, as a guide to the natality bias as well, even though differences in child mortality rates of females and males can also influence these statistics. Indeed, both these phenomena (sex-selective abortions and female disadvantage in mortality of children) reflect anti-female bias, and they can certainly work together. As far as Indian statistics are concerned overall, even though mortality rates of boys and girls are now very close to each other, nevertheless the female-male ratio of the population under age 6 has fallen from 94.5 girls per 100 boys in 1991 (which was thus much in line with the ratio in Europe and North America) to 92.7 girls per 100 boys in 2001. This drop basically reflects the spread of sex-selective abortions and natality inequality, rather than any rise in the mortality of female children relative to male children.
There is a remarkable regional pattern associated with this new phenomenon in India. There has been, in fact, little (or no) decline in some parts of the country, mainly in the east and south, but it has fallen sharply in other regions, mainly in the north and west of India. There have been, for example, extremely sharp declines in the female-male ratios of children in Punjab, Haryana, Gujarat and Maharashtra, and significantly low ratios can be observed in most other states in the north and west of the country.
Because of the legitimate fear that sex-selective abortions might occur in India and serve as a new vehicle of India’s traditional anti-female bias, some years ago the Indian parliament banned the use of sex determination techniques for fetuses, except when as a by-product of some necessary medical investigation. But it appears that the enforcement of this law has been comprehensively neglected.
Consolation might be sought in the fact that even the latest ratio of females to males among Indian children (93 girls per 100 boys), while lower than the figure ten years ago, is still much higher than South Korea’s female–male ratio among children of 88 girls and China’s 86 girls per 100 boys.7 There are, however, additional grounds for concern, which require us to go beyond the current all-India average of female–male ratio of children. First, there are big variations within India, and the all-India average hides the fact that there are several states – in the north and west of India – where the female-male ratio for children is very much lower than the Indian average, and lower even than the Chinese and Korean numbers. Second, it must be asked whether these are ‘early days’ and whether – as the technology of sex determination becomes more widely available across India – the Indian ratio will continue to fall, catching up with – and perhaps even going below – the Korean and Chinese numbers.
There appears to be something of a social divide at this time running right across India and splitting the country effectively into two contiguous halves, in the extent of anti-female bias in natality and post-natality mortality. Since more boys than girls are born everywhere in the world for biological reasons, we must use as our comparative standard not a one-to-one ratio, but the proportions that can be observed in advanced industrial countries (in Europe and North America, for example), where sex-selective abortion is not a significant phenomenon. The female–male ratio for the 0–5 age group is 94.8 in Germany, 95.0 in the United Kingdom, and 95.7 in the United States, and perhaps we can sensibly pick the German ratio of 94.8 as the ‘cut-off point’ below which significant anti-female intervention can be suspected.
This dividing line produces a remarkable geographical split of the country. There are states in the north and west, led by Punjab, Haryana, Delhi and Gujarat, where the female-male ratio of children is very substantially below the benchmark figure (with ratios between 79.3 and 87.8). Other states in these regions also have ratios significantly below the dividing line of 94.8 girls per 100 boys, such as Himachal Pradesh, Madhya Pradesh, Rajasthan, Uttar Pradesh, Maharashtra, Jammu and Kashmir, and Bihar.8 On the other side of the divide, the states in the east and south of India tend to have female-male ratios that are above the benchmark line (that is, 94.8 girls per 100 boys, taken as our cut-off standard): with Kerala, Andhra Pradesh, West Bengal and Assam (each between 96.3 and 96.6 girls), and also Orissa, Karnataka and the north-eastern states to the east of Bangladesh (Meghalaya, Mizoram, Manipur, Nagaland, Arunachal Pradesh). The country looks split in the middle, falling neatly into two divergent segments.9
A partial exception to this sharp pattern of twofold regional split is provided by the southern st
ate of Tamil Nadu, where the female-male ratio is just below 94. As it happens, Tamil Nadu’s female-child ratio is still higher than the ratio of any state in the north and west, which form the ‘deficit list’, but nevertheless Tamil Nadu’s ratio is somewhat below the German cut-off point. However, the astonishing finding is not that one particular state seems to provide a marginal misfit, but that the vast majority of Indian states fall firmly into two contiguous halves, classified broadly into the north and west on one side and the south and east on the other. Indeed – and this is quite remarkable – every state in the north and west10 has a strictly lower female-male ratio of children than every state in the east and south (even Tamil Nadu fits into this classification), and the country stands firmly and sharply divided.
It may be asked whether the female-male ratio in child mortality is also similarly divisive. There is a statistical connection there, but nevertheless the pattern of female-male ratio of the number of children (which incorporates the impact of sex-specific abortion) produces a much sharper regional classification than does the female-male ratio of mortality of children, even though the two are also fairly strongly correlated. The female-male ratio in child mortality varies between, at one end, 0.91 in West Bengal and 0.93 in Kerala (in favour of girls) in the southern and eastern regions, on one side, to 1.3 in Punjab, Haryana and Uttar Pradesh, at the other, with high ratios (against girls) also in Gujarat, Bihar and Rajasthan, in the northern and western regions.
The empirical contrasts make it difficult to escape the conclusion that the north and west have clear characteristics of anti-female bias, influencing the composition of children, in a way that is not quite present – at least, not yet – in most of the east and south. Also, the incidence of sex-specific abortions cannot be explained by the availability of medical resources for determining the sex of the fetus, that is, by the presumption that the states that have more sex-selective abortions are the ones in which medical facilities are more developed and thus usable for this purpose. For example, Kerala and West Bengal in the non-deficit list, both with a ratio of 96.3 girls to 100 boys (comfortably higher than the benchmark cut-off of 94.8), have at least as many medical facilities as exist in such low-female-ratio states as Madhya Pradesh or Rajasthan. So the availability of medical opportunities cannot provide an adequate explanation – quite the contrary – and we have to look at factors that go beyond the supply side of medical technology. If facilities specifically for sex identification and subsequent abortion have been more extensively developed in some states than in others, the explanation has to be sought largely on the demand side, and not in terms of general development of medical opportunities.
Furthermore, the contrast does not seem to have any immediate and clearly explicable economic connection. The states with strong anti-female bias include rich ones (Punjab and Haryana) as well as poor (Madhya Pradesh and Uttar Pradesh), and fast-growing states (Gujarat and Maharashtra) as well as growth failures (Bihar and Uttar Pradesh). It is thus clear that we have to look beyond material prosperity or economic success or GNP growth into broadly cultural and social influences.
A variety of potential connections must be considered here, and there is a need for fresh research to explain the link between these demographic features and the subject matter of social anthropology and cultural studies.11 There is, of course, a wealth of substantial social anthropological studies, undertaken in the past, that have examined regional contrasts within India.12 These would certainly prove useful, though they must be further extended, especially since the regional division of India appears rather different in the new light of natality inequality. Indeed, the division between the north and west on the one hand and the east and south on the other is essentially different from the well-known traditional division between the broad north and the broad south that has received much attention already in Indian social anthropology, at least since Irawati Karve’s pioneering work, Kinship Organization in India.
However, before I end this section of my essay, I must also sound two notes of caution, concerning the temptation to take the observable pattern of regional contrast to be something bigger and sharper than it, quite conceivably, might eventually prove to be. First, any cultural resistance in the east and south to selective abortion against girls might not be immutable, and we must at least consider the possibility that, while the east and south seem more egalitarian in this respect at present, there can be a gradual spread of new practices – slower than in the north and west, but ultimately similarly pervasive. These are, of course, really pessimistic fears, but it is important to avoid complacency about the future of gender equity in the eastern and southern regions of India even as far as natality asymmetry is concerned. If counteracting measures are to be used, they must be considered even for those areas of India which seem relatively secure at the moment.
Having said that, however, it is certainly appropriate to note the observed regional division, from the gender perspective, between the north and west on one side and the east and south on the other. The causal antecedents of this division undoubtedly deserve investigative attention. This is also broadly, though not exactly, in line with the relatively lower relative mortality of female children vis-à-vis male children, in the east and south, compared with the north and west. The female-male ratio in child mortality varies, as was noted earlier, from as low as 0.91 in West Bengal and 0.93 in Kerala, on one side, all the way to 1.30 in Punjab, Haryana and Uttar Pradesh, on the other.
Second, even in the eastern and southern states in which the overall female-male ratio is still within the European range, there are signs that urban centres have, by and large, a somewhat lower ratio than rural parts of those very states. For example, the female-children ratio per 100 boys is 93.7 in urban Orissa as opposed to the ratio for rural Orissa of 95.4. Karnataka’s urban ratio is 93.9, compared with its rural ratio of 95.4. West Bengal’s urban ratio of 94.8, while much the same as the German cut-off line, is still below its rural ratio of 96.7. There may be some evidence here of the use of sex-selective abortions in the urban areas of the east and south as well, even though they seem far less frequent than in the north and west. Indeed, while the eastern and southern urban ratios of girls to boys (such as 93.7 in Orissa, 93.9 in Kerala and 94.8 in West Bengal) are typically lower than the corresponding rural ratios, they are still close to – or even similar to – the international benchmark (94.8) and very much higher than the urban ratios in the north and west, such as 86.6 in Delhi, 84.4 in Chandigarh, 82.7 in Gujarat, 80.9 in Haryana and 78.9 in Punjab.
It is not easy to settle, without further scrutiny, how deep these regional or cultural influences are. But the remarkable geographical division of India into two halves in terms of female–male ratio among children (reflecting the combined influence of the inequality generated by sex-specific abortions and differential post-natal mortality) does call for acknowledgement and further analysis. It will also be extremely important to keep a close watch on whether the incidence of sex-specific abortions significantly increases in the states in the east and south, where they are at this time relatively uncommon.
Unequal Facilities
Natality inequality and survival disparity both have important demographic features, which means that gender bias can be identified, in these cases, on the basis of demographic statistics. However, even when the statistics of life and death do not show much – or any – anti-female bias, there are other ways in which women can have less than a square deal. For example, there are many countries in Asia and Africa, and in parts of Latin America, where girls have far less opportunity for schooling than boys. This is certainly true of most parts of India, and even more so in Pakistan, though the situation in Bangladesh, while still quite unequal, seems to be changing fairly rapidly. Inequality in schooling has far-reaching consequences for the fabric of society and it can profoundly influence many different aspects of gender inequality and also deprivation in general for men as well as women (on which more presently).
There are also other basic facilities that are often asymmetrically distributed. For example, the opportunity to enter politics or commerce may be particularly restricted for women. There are inequalities also in social participation, especially when women are confined to their homes and incarcerated within traditional family lives. These can impose significant handicaps for both the well-being and the agency of women, and they can, as I shall argue presently, have far-reaching social consequences – well beyond the immediate deprivation they directly reflect. Furthermore, even when there is relatively little difference in basic facilities including schooling, the opportunities of special facilities, such as higher education or technical training, may be far fewer for young women than for young men.
Traditionally, this type of asymmetry has been linked to the superficially innocuous idea that the respective ‘provinces’ of men and women are different. This thesis has been championed in different forms over the centuries, and has had much implicit as well as explicit following. It was presented with particular directness in England by the Reverend James Fordyce in his Sermons to Young Women (1766), a book which, as Mary Wollstonecraft noted in her A Vindication of the Rights of Women (1792), had been ‘long made a part of woman’s library’. Fordyce warned the young women to whom his sermons were addressed against ‘those masculine women that would plead for your sharing any part of their province with us’, identifying the province of men as including not only ‘war’, but also ‘commerce, politics, exercises of strength and dexterity, abstract philosophy and all the abstruser sciences’.13