Religion and son preference in India and Bangladesh: Three essays on comparing Hindus and Muslims on son preference and sex differentials in child health

While the existence of son preference in south Asia is well-known, a gap in our understanding of the determinants of son preference is potential differences between religious groups. In this dissertation, I examine whether Hindus and Muslims in India and Bangladesh differ in terms of son preference. I find low daughter discrimination among Muslims and significant son preference among Hindus. I first analyze preferences for the ideal number and sex of children in India, and compare them to actual fertility behaviors that serve as a measure of sex selective abortion. I find that Muslim women are less likely to report a preference for sons in their ideal fertility responses. Analysis of parity-specific births conditional on the sex composition of previous children reveals that the odds of male births are higher than female births for only Hindus and specifically when the previously born children are only girls. In Chapter 2, I extend the analysis to stunting and childhood immunization. I find that Hindu girls are worse off compared to Hindu boys in terms of stunting when their older siblings are also girls. However, there are no sex differentials in immunization, which suggests that while Hindu girls are disadvantaged in terms of long-term intra-household access to nutrition, girls are not discriminated against, in either Hindu or Muslim families, when it comes to availing health services through a fixed number of low-cost or free events. In Chapter 3, I examine whether a group’s majority/minority status influences son preferences by comparing Hindu-majority and Muslim-minority India with Muslim-majority and Hindu-minority Bangladesh. Overall I do not find evidence for son preference among Muslims. In India, Hindus exhibit son preference in Hindu-majority clusters but not in Hindu-minority clusters. In Bangladesh, Hindus exhibit son preference in Hindu-minority areas but not Hindu-majority areas. This suggests that traditional, gender-biased norms prevail for a group with a majority at both the community and national levels. In Indian Hindu-minority clusters, the unique social and cultural environment with more gender-equitable norms influences Hindus. In Hindu-minority areas in Hindu-minority Bangladesh, traditional social norms may be reinforced through a greater threat perception and closely-knit networks