Working Paper: CEPR ID: DP16604
Authors: Sonia Bhalotra; Lakshmi Iyer; Irma Clots Figueras
Abstract: Debates around abortion typically invoke religion and politics but there is no causal evidence ofthe impact of politician religion on abortion. Leveraging quasi-random variation in politicianreligion generated by close elections in India and controlling for the party affiliation of politicians,we find lower rates of sex-selective abortion in districts won by Muslim state legislators, consistentwith a higher reported aversion to abortion among Muslims compared to Hindus. The competinghypothesis that this reflects weaker son preference among Muslims is undermined by statedpreference data and by demonstrating that fertility and girl-biased infant mortality increase inMuslim-won districts.
Keywords: Religion; Politician Identity; Abortion; Sex-Selection; India; Gender
JEL Codes: I15; J13; O15; P16
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Election of a Muslim legislator (K16) | Lower rates of sex-selective abortion (J79) |
| Election of a Muslim legislator (K16) | Higher probability of a girl birth at higher orders (J13) |
| Election of a Muslim legislator (K16) | Increase in probability of higher order births (J19) |
| Election of a Muslim legislator (K16) | Higher share of girls in neonatal and infant deaths (J13) |
| Election of a Muslim legislator (K16) | Reduction in sex-selective abortions (J13) |