How can Indians reduce their son bias
As we recover from the economic consequences of Covid-19 with the hope of building back better, one of the most pressing needs is to ensure a better life for India’s women. While the unequal treatment of women relative to men takes many forms, this piece focuses on just one: The differential treatment of sons relative to daughters in early childhood that sows in the minds of both boys and girls, at a very early age, an acceptance of continued unequal treatment by society in the years to come.
Son preference in India is not just pervasive but persistent. Data from India’s censuses reveal that the sex ratio amongst children, the number of girls per 1000 boys in the 0-6 year age group, has been falling steadily. From 927 in 2001, this ratio declined to 914 in 2011, despite high rates of economic growth in the country. Supporting the lack of correlation between incomes and sex ratios, sex ratios are far below the national average in some of India’s most prosperous states (Gujarat and Maharashtra, at 886 and 883 respectively in 2011). Data from the 2019-20 round of the National Family Health Survey (NFHS), currently available only for a few states, reveal continued decline. In rural Bihar, for example, child sex ratios have fallen from 934 in 2015-16 to 903 in 2019-20.
Son preference also shows up in outcomes other than fertility, such as in schooling investments. We recently conducted a survey of approximately 25,000 relatively poor households in rural areas of 9 of India’s poorest states to evaluate a national programme that supports the formation of Self Help Groups (SHGs) for women’s financial inclusion. Data from this survey reveal that while gender differences in years of completed schooling for school-age children have been all but eliminated, significant differences exist in the probability of sons being enrolled in private primary schools relative to daughters. The percentage of sons between the ages of 7 and 14 enrolled in private schools, even in this sample of relatively poor households, is 17%. The corresponding percentage for daughters is 12%.
Academic theories suggest that improvements in women’s access to resources will reduce son preference. The argument is that mothers have a stronger preference than fathers for fewer children, for greater investment in the schooling of children, and for greater equality in expenditures on sons and daughters. However, husbands’ economic dominance implies that household decisions primarily reflect their preferences rather than those of the mother. Improvements in women’s access to loans and their earning potential will increase their bargaining power vis-à-vis their husbands, resulting in outcomes more closely aligned with women’s preferences. With this increase in “voice,” we should observe not only greater investments in children’s health and schooling, but also greater gender equality in these outcomes.
It is worth noting that this argument rests on improvements in women’s earning potential, regardless of whether she chooses to work or not. That is, it applies regardless of the decline in women’s labour force participation that the economy has witnessed since 2004. What matters is women’s potential to contribute to household income, a potential that has been increasing not just because of the significant gains in schooling attainment but also because of substantial gains in women’s ownership of assets. For example, in rural Bihar, women’s ownership of mobile phones increased from 41% in the 2015-16 NFHS survey to 51% in 2019-20, while the percentage of women reporting a bank or savings account that they used increased from 26% to 77% in this period. Supporting the hypothesis that it is improvement in potential earning power that matters, the NFHS data for rural Bihar also report a steady improvement in women’s role in household decisions, with 87% of currently married women in the state reporting being involved in 3 or more household decisions in the 2019-20 survey, compared to 75% in the 2015-16 survey.
The persistence of son preference despite this improvement in women’s economic status suggests that the hypothesis that such improvements will reduce gender inequalities in the next generation is false. In recent work, we examined this hypothesis in the context of rural women’s growing membership in SHGs and the consequent improvement in their independent access to loans. To identify the causal effects of SHG membership, we compared private school enrolments of two adjacent cohorts: those who were above the age of nine at the time when their mothers joined the SHG and for who choices regarding private-versus- government (primary) school had already been made, and those who were nine years old or younger. Given delayed school entry, this younger sample could still have enrolment choices affected by SHGs.
The figures below graphically document our results. The first figure is for sons, and shows that those under the age of nine at the time of SHG enrolment, to the left of the “cut-off” age or vertical line, have a significant increase in the probability of private enrolment relative to cohorts who just missed this cut-off to the right. The second figure, for girls, graphically reveals how their private school enrolments were completely unaffected by their mothers’ improved access to financial resources.
The striking feature of this evidence is not the differential treatment of sons relative to daughters that it reveals, but rather that differential treatment increases with women’s access to resources.
Our causal analysis accords with NFHS data that suggest greater son preference amongst women relative to men. For high poverty states, these data reveal that in states such as Bihar, Uttar Pradesh and Jharkhand, the proportion of rural women who state that they prefer more sons than daughters is 37.1%, 31.3% and 27.1% respectively. The corresponding percentages for men are 30.4%, 27.9% and 22.5%. The same data also reveal that women are far more accepting even of social norms that justify violence against women. Respondents were asked if it was acceptable to beat a wife if she exhibited a set of seven behaviours such as neglecting house work, arguing with her husband or showing disrespect for her in-laws. Across these seven behaviours, a greater percentage of women state that wife beating is justified. In rural India, 54% of women state that it is justified for at least one of these behaviours while 44% of men do so. Comparing responses to the same set of questions in the previous round of the NFHS (NFHS-3) reveals little change in attitudes, despite significant improvements in women’s education and wealth over the course of this decade.
Why should mothers systematically display greater son preference than fathers? Traditional explanations for son preference emphasise parents’ need for support in old age, given low incomes in prime-earning years that make it difficult to save for old age, and social norms which dictate that such support be given by sons. What has not been as well recognised is that the longer life expectancy of women generates a greater need for old age support for women and correspondingly higher levels of dependence on sons. In our data, 34% of women over the age of 50 are widows. This percentage is just 10% for men. Young mothers draw on their own experiences to form their expectations of their needs for old age support: Of women between the ages of 20-40, 19% report that their mothers are still living while their fathers are deceased. In contrast, only 7% of women in this same age group report their fathers being alive while their mothers are deceased.
Women are not just at greater risk in old age, they are also more vulnerable to income shocks in any period. Given current differences in women and men’s income earning potential, households respond to sharp increases in food prices or weather-induced reductions in agricultural output by ensuring the food needs of men over women. In these periods, mothers benefit more than fathers from the availability of remittances from sons.
This suggests that son preference is a veil: The real difference between fathers and mothers is in the priority they place on saving for old age and precautionary savings against income shortfalls, priorities that reflect their differential exposure to risk. When investing in sons represents the best means of insuring expenditure against income downfalls, this difference in saving propensities gets reflected in mothers’ displaying a greater preference for sons. Investing in sons’ education is estimated to provide a return of between 8 to 12%, which is generally a higher return than small farmers can expect from their agricultural landholdings or which the landless can earn from any financial asset. If education levels are high enough to ensure a salaried job, then the rate of return is not just higher but also safer.
The persistence of son preference, despite improvements in economic growth and in women’s economic status, thus reflects the fact that investments in sons currently provide the best means for women to ensure their own future well-being. Thus, when provided opportunities to improve their economic standing, women use these opportunities to further invest in sons and perpetuate norms that sustain gender inequalities. This creates a classic trap: Improvements in women’s income exacerbate gender inequalities in the next generation, reducing the effectiveness of government policies aimed at redressing these inequalities.
Escaping these traps requires paying attention to magnitudes. While Improvements in women’s access to loans and to income-earning opportunities do help to redress women’s greater vulnerability to income shocks and to improve their role in household decisions, reduced dependence on sons requires improvements in household incomes that are large enough to enable savings against income shocks and to provide for old age. While current strategies have focused on ensuring women’s access to relatively small loans, the emphasis now needs to shift to programmes that increase women’s incomes. Recent initiatives of the Government to promote women’s enterprises, not just their access to savings and loans, may generate real change in the years to come.
(The study has been authored by Dr Anjini Kochar who teaches at Stanford University and is a senior fellow at the International Initiative for Impact Evaluation and others)