Sex selection and health at birth among Indian immigrants

Highlights

• I document strongly son-biased sex ratios at birth for Indian immigrants in Spain.

• Children of Indian immigrants have poor health, relative to other immigrants.

• I find no gender gap in infant health among the children of Indian immigrants.

• The lack of a gender gap in health is likely driven by abortion of unwanted girls.

Abstract

I use birth-certificate data for Spain to document extremely son-biased sex ratios at birth among Indian immigrants (122 boys per 100 girls), especially at higher parities. I also show that the children of Indian immigrants display poor health outcomes during infancy. For instance, almost 10% of boys with Indian parents are born prematurely, compared with 6% of boys with native parents. However, there is no evidence of a gender gap in infant health among the children of Indian immigrants. I provide evidence suggesting that the poor outcomes of Indian children at birth may be attributed to the low endowments of Indian mothers, while the absence of a gender gap may be driven by the fact that the parents who would invest less in girls are less likely to carry the pregnancies of girls to term (more likely to practice sex-selective abortion), combined with the lower cost of prenatal investments in Spain (compared with India).

Introduction

In two influential papers in the early 1990’s, Sen, 1990, Sen, 1992 raised attention to the high male to female ratios prevalent in India and other Asian and North-African countries, attributing them to excess mortality among women, which he suggested might derive from the comparative neglect of female health. Since then, a new trend emerged in the same countries, with increasingly son-biased sex ratios at birth. Sen (2003) attributed this new trend to parental preferences for boys, combined with the availability of sex determination technologies and sex-selective abortion. These findings were confirmed in later work by Jha et al. (2006), Bhalotra and Cochrane (2010) and Bongaarts and Guilmoto (2015), among others. Recent estimates suggest that since 1990, the number of “missing women” in the world has risen by 43 percent to 126 million in 2010, and is projected to continue increasing (Bongaarts and Guilmoto, 2015).

In this paper, I study sex ratios at birth and newborn health for the children of Indian immigrants in Spain. Spain is both a country with a large (if recent) immigrant population, and one with high-quality administrative data from birth and death registers. I first confirm the results in previous studies regarding sex ratios at birth in other host countries (Dubuc and Coleman, 2007; Abrevaya, 2009; Almond and Edlund, 2008; Almond et al., 2013). I show evidence of severely son-biased sex ratios at birth among Indian immigrants, using high-quality birth register data for Spain. This bias is found to increase with birth-order, and is much more pronounced for India than for any other source country. I am able to rule out that the “missing girls” can be traced to differential stillbirth rates: I find that the pregnancies of Indian mothers are not significantly more likely to end in stillbirth than other immigrant groups, and Indian women carrying girls are no more likely to suffer a stillbirth than those carrying boys.

Then, I take advantage of the rich measures of health at birth to study the health of the children of Indian immigrants at birth. I show that the children of Indian parents are (much) less healthy than natives as well as other immigrant groups at birth (in terms of prematurity rates, birth-weight, etc), but there is no gender gap: newborn Indian girls are no less healthy than boys. This pattern remains when I analyze infant and child mortality. To my knowledge, this is the first estimate of the gender gap in health at birth among the children of Indian immigrants anywhere.

I then interpret the documented health gaps in terms of a combination of selection via abortion of girls, the poor socioeconomic (and, plausibly, health) status of Indian mothers, and differential prenatal investments in boys versus girls. First, I show that the poor health of Indian boys is likely a direct result of the poor endowments of India-born mothers. Then, I show evidence consistent with the absence of a gender gap in infant health among the children of Indian parents being plausibly driven by positive selection in (potential) pre-natal investments, such that the parents with a stronger preference for boys, who would thus invest less in girls, are less likely to carry the pregnancy of a girl to term. It is also likely that the lower cost of prenatal investments in Spain (compared with India) mitigates the gender gap in parental investments.

This study links the literature on sex ratios at birth (in India or among Indian immigrants) with the one on the health of Indian children, and the high mortality rates among Indian girls, focusing on health at birth. Recent studies have documented the poor health of Indian children, relative to international standards (and to countries with similar income levels), see for example Deaton and Drèze (2009) and Tarozzi, 2008, Tarozzi, 2012. The limited available evidence indicates that this gap is probably not due to genetic factors (Tarozzi, 2008). The poor health of Indian mothers is almost certainly a contributing factor (Coffey, 2015). Recent work by Jayachandran and Pande (2015) also suggests that at least some of the gap might be explained by parental preferences for the first son (so that girls and higher-parity boys receive lower parental investments and are thus less healthy).

While a number of papers failed to find direct evidence of lower parental (postnatal) investments in girls in India (such as nutrient intake), recent work suggests that some of the mortality gender gap is probably related to girls getting breastfed less than boys (Jayachandran and Kuziemko, 2011; Barcellos et al., 2014), as well as to gender gaps in vaccination and malnutrition (Oster, 2009a, Oster, 2009b; Aurino, 2017). Recent evidence also points to the role played by differential prenatal investments, such as tetanus shots during pregnancy (Bharadwaj and Lakdawala, 2013).

Another related strand of literature has studied Indian immigrants residing in other countries. Male to female sex ratios at birth have been found to be high (and increasing in recent decades) among Indian immigrants in the UK (Dubuc and Coleman, 2007), the US (Almond and Edlund, 2008; Abrevaya, 2009) and Canada (Almond et al., 2013). As is the case in India, the son-biased sex ratios are driven by higher-order births and families with no previous boys.

Not so much is known about the health of children of Indian descent residing in rich countries (at birth or later on), although some (limited) evidence has been provided of assimilation in health with time in the host country (Proos et al., 1992). Tarozzi (2008) shows that ethnic Indian children in the UK are as healthy as native children, with some evidence of assimilation with time in the host country, although the data limitations are severe (his sample includes only 471 ethnic Indian children, aged 2–18).

Also due to data limitations, little is known about the gender gap in health at birth (and its determinants) among Indian children (in India or elsewhere). This is due to the long-recognized low quality of birth registers in India, as well as the absence of large-sample nationally representative data sets that include measures of health at birth (such as birth-weight). Probably the best available data source is the National Family Health Survey (NFHS), which included almost 9,000 observations of children under age 1 (less than 500 under one month) in the 2005–06 wave. Birth-weight is reported for children under age 5, but the non-response rate is very high (66%). Data limitations thus make it hard to study the influence of pre-natal versus post-natal parental investments as contributors to the neonatal and later health outcomes of Indian boys and girls.

I thus contribute to this literature by providing reliable estimates of the gender gap in health at birth among Indian children, while linking them to sex ratios at birth, infant and child mortality, and parental endowments and prenatal investments.

I also add to a small number of recent papers (Anukriti et al., 2016; Hu and Schlosser, 2015; Lin et al., 2014) that highlight the fact that selective abortion of girls may have direct implications for the (relative) health of (surviving) girls. The strategy to link sex ratios at birth to the health of girls in those papers is to compare children in time periods and/or regions with lower versus higher incidence of sex selective abortion. Instead, I use as “controls” groups of parents with no son preference (as reflected in balanced sex ratios at birth), based on country of birth. I am also able to study health outcomes not available in previous studies, such as prematurity and birth-weight, which have been shown to have important long-term consequences, and which are only affected by pre-natal (versus post-natal) parental investments.

The remainder of the paper is organized as follows. Section 2 describes the data sources. Section 3 documents sex ratios at birth among Indian immigrants, as well as natives and other immigrant groups. Then, Section 4 analyzes the gender gap in a number of measures of health at birth, for children with native versus Indian parents. Section 5 interprets the health results as a function of parental endowments, investments, and selection, taking into account that the families that end a pregnancy when carrying a girl may not be random, and Section 6 concludes.