Missing women and India’s religious demography

 However, a curious paradox in Indian economic demography, rarely commented upon, is that while Muslim fertility in is considerably higher than Hindu fertility, infant mortality among Muslims is considerably lower than among other groups. Motivated by this paradox, this paper put forward a new theory that provides an explanation for larger Muslim, relative to Hindu, fertility.

The economics literature on gender bias and son preference was linked directly with the discussion of religion and fertility. This paper argued that bringing together the notion of ‘son preference’ and the complementary concept of ‘daughter aversion’ provides insights: just as sons bring ‘benefits’ to their parents, daughters impose ‘costs’ and complementing a desire to have sons is a desire not to have daughters.

Consequently, the desire for sons increases family size while the fear of daughters limits it. A formal model, in which these two countervailing forces act so as to determine equilibrium family size and composition, was developed. The hypothesis was then tested using quantitative evidence based upon an empirical dataset from .

The data used were a nationally representative sample of 10,548 rural Indian women who had adopted a terminal method of contraception and who had therefore attained their equilibrium family size and composition. A number of Poisson regression models were estimated on the number of living children, the number of infant deaths, and the sex ratio.

The econometric study undertaken here shows that in the sample, Muslims had statistically the same degree of son preference as Hindus but a significantly lower degree of daughter aversion. These facts were shown theoretically to be sufficient to result in a larger average (equilibrium) family size for Muslims than for Hindus. The curious paradox about religion and reproduction in is that groups that have higher fertility, such as the Muslims, also display lower degrees of daughter aversion. The thrust of the econometric analysis went on to explain the number of sons and daughters to these women in terms of their personal and household characteristics.

The analysis showed that after controlling for other factors at the level of the individual, household, the village, and the region, there are differences by religion and caste in the determinants of the numbers of living children. Another important point to emerge from the analysis was that, after controlling for other factors, Muslim women had a smaller number of both male and female infant deaths compared to Hindus. For Hindu women, the increased likelihood of terminating fertility after the birth of a son was nearly three times the increased likelihood of terminating fertility after the birth of a daughter. By contrasts, the increase in the likelihood of Muslim women terminating their fertility after the birth of a son or daughter was substantially lower than the corresponding values for Hindus. This suggests empirically as well, that there is lower daughter aversion among Muslims. This was reflected also in the considerable differences by religion in the sex ratio at birth and in the sex ratio for currently living children, between Hindus and Muslims.

Other analysis presented in the paper included a logit model of the demand for contraception adopted by the groups under study. The demand for contraception was influenced by a range of economic characteristics. It was shown that the community effect on the demand for contraception operated entirely through differences between Hindu, Muslim and Dalit women in the change in their demand for contraception, following an additional daughter or son.

An important finding of the study is that there were no significant community effects operating over and above these ‘numbers based’ effects. Collectively, the findings from the analysis suggest that counter to popular belief that associates higher Muslim fertility with pronatalist tendencies within Islam, higher Muslim fertility in may in reality be related to gender bias, in particular the significantly lower levels of daughter aversion among this community.

The study concludes that notions of son-preference and daughter aversion may be significant in explaining, theoretically and empirically, intergroup differences in fertility by religion and caste in . Moreover, and paradoxically, the reason for the higher fertility of Muslim, relative to Hindu, women may lie in daughters being more welcome in Muslim than in Hindu families and ipso facto in the relatively better treatment that girls receive at the hands of Muslim parents. More research is needed on the complex interactions between religion, gender bias, and fertility behaviour, both in and elsewhere.