This column presents results of a study that breaks down “missing women” by age across the Indian states. It illustrates that Indian women face the risk of excess mortality at every stage of their lives, and attempts to explain excess female deaths in India after birth.
Recent events in New Delhi — shocking in their in-your-face barbarism but otherwise all too frequent — have brought home the fact that the treatment of females in India knows little discrimination when it comes to age. That mistreatment might begin at birth, but it does not let up in adolescence and continues unrelentingly into widowhood.
Yet it is not easy to get a firm handle on the extent to which women are mistreated at different age groups. In part, this is because direct statistical evidence is very hard to come by. We can illustrate this using rape. Well-known international media outlets haven’t tired of highlighting India’s 24,000 reported rape cases per year. For instance, a recent CNN report calls this a “shockingly high rate of violent crime”. That may well be, but for every 100,000 of India’s population that number works out to about 1.8 in 2010. The corresponding numbers in the United States stand at around 27.3; in England and Wales around 28.8. The definition of rape differs across countries, and that has something to do with this extraordinary differential, but so might the extent of under reporting. It is all a bit of a mystery, because underreporting of rape in the US and the UK is also known to be extremely high.
Criminal acts such as rape, sex-selective abortion, dowry killings or the treatment of widows -- these are dark inequities that are shrouded in the even darker ambiguities of inadequate data. But there are other ways of checking if the numbers “add up”. Consider, for instance, the phenomenon of “missing women”, brought to public attention by the pioneering work of Ansley Coale (1991) and Amartya Sen (1990).
Relative to developed countries, there are far fewer women than men in India. Estimates suggest that among the stock of women who could potentially be alive today, 25 to 50 million are demographically “missing” (Sen 1990, 1992). With rape and (more generally) mistreatment of women at all ages regularly flying under the statistical radar of data availability, deliberate sex selection at birth has been widely regarded as a key explanation. For example, Jha et al.(2006) estimate that as many as 10 million female foetuses have been aborted over a 20 year period. Other research focuses on early childhood and the possibility that young girls are systematically less cared for. Indeed new estimates from the 2011 Census reveal that the child sex ratio of boys to girls, aged 0 to 6, is at its highest level since independence. But a full age decomposition of missing women tells us much more.
In our recent paper, we provide such a breakdown of missing women by age across the states of India (Anderson and Ray 2012). We leave out the methodological details here, but roughly speaking, it involves a comparison of relative gender-based death rates in India with the corresponding ratios in developed countries, and thereby extracting an estimate of excess female mortality (or "missing women").
We do not dispute the existence of severe gender bias at young ages. But our computations yield some striking findings that go beyond infancy.
First, the vast majority of missing women in India are of adult age. Our estimates demonstrate that roughly 12% of missing women are found at birth, 25% die in childhood, 18% at the reproductive ages, and 45% die at older ages. One possible reaction to this observation is that excess female mortality is higher in adulthood simply because overall mortality, excess or not, is higher in adulthood. But we argue in our paper that this is not the whole story. Relative excess female mortality, by which we mean excess female mortality per female death, is highest in the reproductive years and at older ages.
Variation across states
Second, there is significant variation in the distribution of missing women by age across different states. Haryana, Rajasthan and Punjab typify what is perhaps the generic viewpoint regarding excess female mortality: that the bulk of missing females are missing at birth or at young ages. Over two-thirds of the missing women are missing by the age of 15 in these states. For Punjab, close to 60% of the excess female mortality is at birth, and for Haryana and Rajasthan, the numbers are well over 40%. These are disturbing numbers that reflect the conventional wisdom on missing women.
But Haryana, Punjab and Rajasthan account for well under 15% of the national total for missing women. Their profile is emphatically not the case for India as a whole. For all other states, the majority of missing women die in adulthood (older than age 15). Among this set of states, there is variation as well. The southern states (Kerala, Tamil Nadu, and Andhra Pradesh) have the lowest number of missing women and make up only 10% of the national total. The remaining bulk of the missing women are distributed throughout the rest of the country. One group of states (Madhya Pradesh, Maharashtra, Bihar, and Assam), which comprise roughly 37% of the total missing women in the country, have the highest excess female mortality in the country at all ages and the relative differences are largest at adult ages. Another group of states (West Bengal, Orissa, Himachal Pradesh, and Uttar Pradesh), which make up roughly 33% of the national total, have lower excess female mortality at the younger ages, but high excess female mortality at adult ages.
What explains India’s missing women?
Because there is so much state-wise variation in the distribution of missing women across age groups, it is difficult to provide a clear explanation for missing women in India. The traditional explanation for missing women, a strong preference for the birth of a son, most likely drives a significant proportion of missing women in Punjab and Haryana where biased sex ratios at birth are undeniably present. But the explanation for excess female deaths after birth is far from clear. We can speculate on possible contributing factors.
We find significant excess female mortality in childhood everywhere except for the southern states (Kerala and Tamil Nadu have particularly low numbers of missing girls). Our earlier work demonstrated that the bulk of missing girls in India who die prematurely in childhood, do so when they are under 5 (Anderson and Ray 2010). The excess mortality comes from infectious and parasitic diseases; largely vaccine-preventable or diarrhoea-related diseases that are typically linked to poor sanitary conditions. It is difficult to disentangle gender discrimination from other biological, social, environmental, behavioural, or economic factors that might explain excess female deaths from these diseases. For instance, it is conceivable that parents are more likely to give clean water to young boys compared to girls, but it is also possible that girls are inherently more susceptible to certain diseases. There is some evidence of gender bias with regard to the immunisation of young children (Pande and Yazbeck 2003). However, to conclude that this unequal treatment by parents explains all of the missing girls from vaccine-preventable diseases requires further research.
We find significant excess female mortality at the reproductive ages, particularly in the central and north eastern states. Our earlier work demonstrated that there were two key causes behind these excess deaths of women in India at this stage in their lives. The first was maternal mortality. The other, more ominously, was "injuries". In particular, for the year 2000, we estimated that excess female deaths for women in India from this cause exceeded 225,000, a number that dwarfs our overall maternal mortality estimates of 130,000 each year.
The excess female deaths from "injuries" (while admittedly indirect) would appear to be an indicator of overt violence against women. For the country as a whole, fire-related death is a leading cause; each year over 100,000 women are killed by fires alone. These excess deaths could well be associated with the custom of dowry which has been linked to bride-burning and dowry-death if promised dowry payments are not forthcoming. Rape, for all its horror, is the tip of the iceberg.
Comparing the different states, excess female mortality at older ages (older than 45) is highest in the north eastern states and lowest in the southern ones. Our earlier work demonstrated that the excess burden in this category falls mainly on non-communicable diseases. Cardiovascular disease is particularly implicated. In India, women die at a rate closer to men from cardiovascular disease relative to developed countries. There is an entire array of hypotheses to explain the phenomenon. It could be genetic - for instance, the recently discovered “heart disease gene” so prevalent in South Asia may be equally present in males and females, thereby lowering the gender skew in incidence. Lifestyle differences by gender may be important: diet, attention to personal health and well-being, and so on. Or it may truly be lack of overt "similar care": women seek or receive medical care less often in developing countries, or may be subject to greater stress.
Excess female mortality at all ages
More research is needed to specifically identify the different pathways which explain the state-wise variation in the distribution of missing women across India. What our research makes clear is that Indian women face the risk of excess mortality at every stage of their lives and that excess female mortality in adulthood is as serious a problem in India as missing girls who are never born or die prematurely in childhood.
- Anderson, Siwan and Debraj Ray (2010) “Missing Women: Age and Disease”. Review of Economic Studies, 77, 1262-1300.
- Anderson, Siwan and Debraj Ray (2012) “The Age Distribution of Missing Women in India”. Economic and Political Weekly, December 1, 87-95.
- Coale, Ansley (1991) ``Excess female mortality and the balance of the sexes in the population: An estimate of the number of `missing females" Population and Development Review 17(3), 517-523.
- Jha, P., R. Kumar, P. Vasa, N. Dhingra, D. Thiruchelvam, and R. Moineddin (2006) “Low male-to-female sex ratio of children born in India: national survey of 1.1 million households”. Lancet, 367, 211-18.
- Sen, Amartya (1990) “More Than 100 Million Women Are Missing”. The New York Review of Books, 37(20), December 20, 1990.
- Sen, Amartya (1992) “Missing women”. British Medical Journal, 304 (March), 587-588.