Debates surrounding abortion invoke both religion and politics. This column examines whether the religious identity of legislators influences abortion rates in the districts in which they are elected, conditional upon their party affiliation. It finds that the election of one additional Muslim legislator in a district leads to an increase of 1.79 percentage points in the probability that a girl is born.
Debates surrounding abortion often invoke both religion and politics. Abortion is a highly politicised issue, with governments often seeking to amend abortion legislation in line with the preferences of elected leaders or the sentiments of the electorate. Yet, there is little research that establishes a causal link between leader preferences and abortion outcomes. In a new research paper (Bhalotra, Clots-Figueras and Iyer 2018), we examine whether the religious identity of legislators influences abortion rates in the districts in which they are elected, conditional upon their party affiliation. We examine sex-selective abortion in India, a phenomenon that has increased sharply over the past two decades, and that policy interventions have largely failed to address. A natural candidate explanation for this is insufficient political commitment. We investigate the role of preferences of state legislators. In particular, since Muslims express greater opposition to abortion than Hindus, we ask whether Muslim legislators are more effective at reducing sex-selective abortion. We find that they are. This is a striking result, given that Muslims are a minority religious group in India, albeit the largest, constituting 14.2% of the population in 2011.
Religious identity of legislators and sex ratio at birth
To conduct this analysis, we created a new database on the religious identity of candidates for state assembly elections. Our data covers India’s 16 largest states that contain more than 95% of the country’s population, over the period 1980-2010. Since electoral data in India do not identify candidate religion, we identify Muslim legislators and candidates based on their name1. Since it is difficult to obtain reliable data on abortion, and sex-selective abortions in particular are likely to be under-reported, we used the sex ratio at birth (that is, the probability that a birth is female) as a marker of sex-selective abortion. The data on sex ratios is obtained from the 1998-99 wave of the National Family Health Survey that also contains district identifier.
So as to isolate the role of politician preferences from voter preferences, we compare outcomes in districts where Muslims narrowly won elections against non-Muslims to those in districts where Muslims narrowly lost elections against non-Muslims. Our identifying assumption is that the identity of the winner in close elections (where the margin between winner and runner-up is less than 3% of the total votes cast) is quasi-random. We verify that districts where Muslims win vs. districts where they lose in close races are similar across a range of demographic characteristics as well as in terms of their recent political history.
We find that the election of one additional Muslim legislator in a district leads to a statistically significant increase of 1.79 percentage points in the probability that a girl is born. This effect emerges at third birth order, which is when the tension between continuing fertility and conducting sex-selective abortion becomes most acute, given that three is close to the median of desired fertility in India. The identified impact of legislator religion on the sex ratio at birth is largest exactly where previous research2 has shown that the underlying tendency to commit female foeticide is greatest: at higher birth orders (three and higher), in religious groups that practice more sex selections (Hindus and Sikhs, particularly those with a first-born daughter) and in states with more entrenched gender bias. Importantly, we do not find any impact of Muslim legislators on gender of birth before 1985 – this acts as a placebo test since sex-selective abortion was not feasible before ultrasound technology for prenatal sex detection became widely available from the mid-1980s onwards.
We also find that fertility is higher under Muslim legislators, consistent with continued fertility being an alternative means of obtaining the desired number of sons. The presence of an additional Muslim legislator leads to an increase of 2.02 percentage points in the probability of having a third or higher order birth. Thus the increase in fertility is almost completely matched in size by the increased probability that a girl child is born, consistent with the increase in fertility reflecting the reduction in girl abortion. Importantly, we find no evidence that reduced prenatal sex-selection is substituted by postnatal neglect. Indeed, infant mortality rates of higher-birth-order girls relative to boys are lower under Muslim legislators.
Aversion to abortion, not pro-woman preference
Our findings are consistent with Muslim legislators curbing sex-selective abortion because of a religious aversion to abortion. Using the World Values Survey 1990, we document that Muslims are significantly less likely to support abortion being practiced under a range of conditions. We also find that the impact of Muslim legislators on gender of birth is somewhat higher in periods after the implementation of the Pre-Conception and Pre-Natal Diagnostic Techniques (PC&PNDT) Act, 1994. This legislation made it an offence to conduct prenatal sex detection, imposing penalties on both citizens and medical providers for violating the guidelines. Our findings suggest that better enforcement of the prenatal sex determination ban may be a mechanism by which Muslim legislators enacted their preferences against abortion. This would appear to be the first evidence that religious preferences, embodied in political leaders, influence abortion-related outcomes3. We find limited support for alternative explanations such as Muslims having lower son preference or generally espousing more pro-female policies. In particular, the reported desire for sons relative to daughters is not significantly different between Muslim and Hindu women, and reported crimes against women do not vary with the religion of legislators.
- Two independent teams coded the religion based on names, and disagreements between the two teams’ classification were resolved by the authors on a case-by-case basis. After this procedure, we remained doubtful of the religious identity of less than 0.5% of candidate names, and assigned them as “non-Muslim” as a tie-breaking rule.
- See, among others, Almond et al. (2013), Bhalotra and Cochrane (2010), and Jha et al. (2011).
- The plausibility of our findings is enhanced by recent evidence that leverages the visit of the Pope to Brazil in 1991 to illustrate that religious preferences of a leader can influence fertility outcomes even in a relatively short time frame (Bassi and Rasul 2017).
- Almond, Douglas, Lena Edlund and Kevin Milligan (2013), “Son Preference and the Persistence of Culture: Evidence from South and East Asian Immigrants to Canada”, Population and Development Review, 39(1):75–95.
- Bassi, Victor and Imran Rasul (2017), “Persuasion: A Case Study of Papal Influences on Fertility Related Beliefs and Behavior”, American Economic Review: Applied Economics, 9(4):250-302. Available here.
- Bhalotra, S, I Clots-Figueras and L Iyer (2018), ‘Religion and Abortion: The Role of Politician Identity’, IZA Discussion Paper 11292.
- Bhalotra, S and T Cochrane (2010), ‘Where Have All the Young Girls Gone? Identification of Sex Selection in India’, IZA Discussion Paper No. 5381.
- Jha, Prabhat, Maya A Kesler, Rajesh Kumar, Faujdar Ram, Usha Ram, Lukasz Aleksandrowicz, Diego G Bassani, Shailaja Chandra and Jayant K. Banthia (2011), “Trends in selective abortions of girls in India: Analysis of nationally representative birth histories from 1990 to 2005 and census data from 1991 to 2011”, The Lancet, 377(9781):1921–1928.