1.231, 1.838 and 1.167 for mothers who received ultrasound imaging as part of
a pre-natal check-up (Arnold et al.,2002). Despite a 1996 law that made it ille-
gal for the sex of children to be divulged to parents, and a 2001 Supreme Court
order for states to enforce the ban on using ultrasound to determine the sex of a
foetus, estimates of sex-selective abortions range from 100,000 to 500,000 per
annum (Arnold et al., 2002; Bhat, 2006; Jha et al., 2006). Ninety per cent of the
estimated annual total of around 5 to 6 million abortions were performed in
unregistered (illegal) facilities. The exact number motivated by sex selection can
only be guessed at. A survey in Haryana found that 18 per cent of women had had
abortions, of which more than one-third were for sex selection (Unisa et al.,
2007).
Sex-selective abortions are only a partial explanation for the imbalance in
the sex ratio. Excess female mortality in India continues until the age of 35
(Ravindran, 1995). Female mortality soon after birth (female infanticide)
continues to be the dominant way to remove female children in India. Rough
estimates for 1981–91 indicate that there were up to four times as many excess
female deaths taking place after birth as before birth (Das Gupta et al.,
1998:90). The male/female ratio in the 0–6 age group has continued to rise
(Dyson,2001). Except for Kerala,the cohort of young children has become
more male in every state. Between 1982–83 and 1992–93, the all-India child
mortality was 43 per cent higher for girls than for boys; in Haryana and Punjab
it was 135 and 81 per cent higher, respectively (Arnold et al., 2002:304). Much
fieldwork shows that female infanticide is prevalent in contemporary India
(Gardner, 2003). In the late 1980s almost 10 per cent of female births in Tamil
Nadu resulted in infanticide (George et al., 1992). This practice continued into
the 1990s (Chunkath and Athreye, 1997; Sudha and Rajan, 1999). In 2000, in
the Salem district of Tamil Nadu, 42 per cent of infant deaths were reported to
be due to ‘social reasons’ (Srinivasan and Bedi, 2007:859). In the 0–6 age
range the sex ratio in the same district had reached 1.175 boys to every girl.
This was despite the fact that Tamil Nadu has relatively high literacy, low
infant mortality, good education and basic health care.
Excess mortality among girl children also results from long-term neglect and
discrimination. The intra-household allocation of ‘life-sustaining’ resources can
be thought of as an investment decision. Because male children, later in life, are
more likely to get (better) paid work, allocating resources to them will maximize
the return on the household investment. After the neonatal period, excess female
infant mortality can stem from biases in the household allocation of survival-
related goods, such as nutrition and medical care. In India, there is noticeable
discrimination against girls in breastfeeding and in access to nutritious foods
such as milk and fats (Das Gupta, 1987; Pebley and Amin, 1991). Provision of
medical care is also significantly biased in favour of boys. Expenditure on
medical care and clothing in the Punjab is significantly higher for boys
(Alderman and Gerler 1997). The maximum differentials in the allocation of
medical care occur in the first two years of life, the period in which most child
deaths take place (Das Gupta 1987). In South Asia, boys are more likely to be
96 Sources of Growth in the Modern World Economy since 1950