Gist of Economy Survey 2017-18 [Chapter 7]: Gender and Son Meta-Preference: Is Development Itself an Antidote?

Gist of Economy Survey 2017-18 [Chapter 7]: Gender and Son Meta-Preference: Is Development Itself an Antidote?


  • Benefits of gender equality:
    • The intrinsic values of gender equality are uncontestable.
    • But now there is growing evidence that there can also be significant gains in economic growth if women acquire greater personal agency, assume political power and attain public status, and participate equally in the labor force.
    • In developing countries, working women also invest more in the schooling of their children.
    • Recently at Davos, IMF chief Christian Lagarde, quoting IMF research, said that women’s participation in the workforce to the level of men can boost the Indian economy by 27 percent.
  • Methodological problem afflicting assessments relating to gender:
    • The problem is one of conflating “development time” and “chronological time.”
    • Gender indexes such as the Global Gender Gap Index of the World Economic Forum (WEF) or the Gender Inequality Index (GII) of the United Nations Development Program (UNDP) rank countries in chronological time.
    • But such simple cross-sectional comparisons are prone to a potential flaw. The role of women evolves with development.
    • Scandinavia in the early 1900s was demonstrably less well-disposed to women than Scandinavia today, and possibly less well-disposed than countries today that have attained a level of development not dissimilar to Scandinavia in the early 1900s.
    • Thus cross-sectional comparisons could be misleading: a case of passing judgment in “chronological time” oblivious of “development time.”
    • Invoking “development time” is not to dismiss “chronological time” rather, policy-making should be informed by both perspectives.
    • Urgency of action should spring from assessments in chronological time but that must be leavened by the understanding that comes from assessments in development time.
    • This distinction is crucial for another reason: if a country’s performance is atypical in development time, the policy strategy will have to be far different from that if a country’s performance is typical. In the former, bleaker case, development itself cannot be counted upon to improve the role and status of women. The burden on government, civil society, and other stakeholders will correspondingly be greater.
  • Gender equality is an multidimensional issue. Accordingly, assessments in this chapter are made on three specific dimensions of gender:
    • Agency relate to women’s ability to make decisions on reproduction, spending on themselves, spending on their households, and their own mobility and health.
    • Attitudes relate to attitudes about violence against women/wives, and the ideal number of daughters preferred relative to the ideal number of sons.
    • Outcomes relate to son preference (measured by sex ratio of last child), female employment, choice of contraception, education levels, age at marriage, age at first childbirth, and physical or sexual violence experienced by women.
  • The analysis is based on the Demographic and Health Survey (DHS) datasets from 1980 to 2016. The survey has datasets at household level. The India National Family Health Survey (NFHS) 2015-16, which feeds into the DHS survey.
  • Findings:
    • On 14 out of 17 indicators relating to agency, attitude, and outcomes, India’s score has improved over time.
    • On 7 of them, the improvement is such that in the most recent period India’s performance is better than or at par with that of other countries, accounting for the level of development.
      • The progress is most notable in the agency women have in decision-making regarding, household purchases and visiting family and relatives.
      • There has been a decline in the experience of physical and sexual violence.
      • Education levels of women have improved dramatically but incommensurate with development.
    • On 10 of 17 indicators, India has some distance to traverse to catch up with its cohort of countries.
      • For example, women’s employment has declined over chronological time, and to a much greater extent, in development time (discussed later)
      • Another such area is in the use of female contraception (discussed later).
    • Son preference and Son meta-preference:
      • In the area called “son preference”, development is not proving to be an antidote. Son preference giving rise to sex selective abortion and differential survival has led to skewed sex ratios at birth and beyond, leading to estimates of 63 million “missing” women.
      • There is another phenomenon of son meta-preference which involves parents adopting fertility “stopping rules” – having children until the desired number of sons are born.
      • This meta-preference leads naturally to the notional category of “unwanted” girls which is estimated at over 21 million.


India’s Performance

  • On 12 out of 17 variables, average levels in India have improved over time.
    • For example, 62.3 percent of women in India were involved in decisions about their own health in 2005-06, which increased to 74.5 percent in 2015-16.
    • The percentage of women who did not experience physical or emotional violence increased from 63 percent to 71 percent.
    • The median age at first childbirth increased by 1.3 years over ten years.
  • Contraception:
    • India has some distance to traverse on several dimensions to be on par with other countries in development time. One such dimension is the use of reversible contraception methods. For their level of wealth, the use of reversible contraception methods among Indian women is 51.6 percentage points lower than it should be.
    • The number of married women in India who do not use any contraception method is high (46.5 percent).
    • Among women using any contraception method at all, the percentage of Indian women using female-controlled reversible contraception is unusually low (32.8 percent).
    • These findings warrant attention: since not many women use methods of reversible contraception, they have little control over when they start having children, but only seem to have control over when they stop having children.
    • This could affect other milestones early on in a woman’s life; for example, women may not get the same access to employment that men do.
  • Women workforce:
    • Another finding relates to the percentage of women who work, which has indeed declined over time (from 36 percent of women being employed in 2005-06 to 24 percent of women being employed in 2015-16).
    • There is a long and contested literature on whether this decline is a cause for concern or will improve naturally with time and development.
    • There is the more general phenomenon of a U-shaped behavior of female labor force participation with respect to development.
    • India is on the downward part of the “U” but even more so than comparable countries.
    • On the supply side, increased incomes of men allows Indian women to withdraw from the labor force, thereby avoiding the stigma of working; higher education levels of women also allow them to pursue leisure and other non-work activities all of which reduce female labor force participation.
    • On the demand side, the structural transformation of Indian agriculture due to farm mechanization results in a lower demand for female agricultural laborers. Evidence also points to insufficient availability of the types of jobs that women say they would like to do—regular, part-time jobs which provide steady income and allow women to reconcile household duties with work—and types of sectors that draw in female workers.
    • This, interacted with safety concerns and social norms about household work and caring for children and elders ,militates against women’s mobility and participation in paid work
  • Finally, the sex ratio of last birth is biased against females and is lower by 9.5 percentage points in 2015-16 in comparison to other countries. And this has remained stagnant in the last decade. The sections on son preference and son meta–preference discuss it.

Is there a convergence effect?

  • Encouragingly, there is evidence of convergence.
    • On all but 2 measures, gender indicators improve as wealth increases.
    • From a development time perspective, nearly all gender dimensions respond to wealth to a greater extent in India than in other countries.
    • This implies that even where India is lagging, it can expect to catch up with other countries as the wealth of Indian households increases.
  • The only two cases where such a convergence effect is not visible and where India appears to be falling behind even in development time is on women’s employment and sex of last child.


  • In last decade (2005-6 to 2015-16), there is a “convergence” effect in that the poorer performers in the earlier period improve their score more over time.
  • North-Eastern states doing substantially better than the hinterland states even in development time; hinterland states are lagging, some associated with their level of development and some even beyond that; surprisingly, some southern states such as Andhra Pradesh and Tamil Nadu fare worse than expected given their level of development.
  • Most North-Eastern states (with the exception of Tripura and Arunachal Pradesh) and Goa are the best performers at all points of time. Kerala is the next best performer.
  • The lagging performers are Bihar, Rajasthan, Madhya Pradesh, Uttar Pradesh, Jharkhand and, surprisingly, Andhra Pradesh.
  • Delhi’s performance worsens in a decade 2005-06 to 2015-16.
  • Finally, out of 100, the distance of the Indian states from their absolute frontier can be assessed.
    • The worst Indian score is 57.6 (Bihar) and
    • the best is 81 (Sikkim)
    • with most of India scoring between 55 and 65.
  • Analyzing the gender score for Indian states in 2015 against log per capita income, North-Eastern states have much better gender scores given their levels of income. On the other hand, accounting for their levels of income, Andhra Pradesh, Haryana, Bihar and Tamil Nadu perform less well.


  • The biologically determined natural sex ratio at birth is 1.05 males for every female. Any significant deviation from this is on account of human intervention – specifically, sex-selective abortion.
  • India’s sex ratio during 1970 to 2014 increased substantially from 1060 to 1108 (male per 1000 female) whereas if development acted as an antidote, it should have led to improvements in the sex ratio.
  • The sex ratio for the different states in India in 1991 and in 2011:
    • There is a negative correlation between income and sex ratio (a worsening in development time).
    • Most striking is the performance of Punjab and Haryana where the sex ratio (0-6 years) is approaching 1200 males per 1000 females, even though they are amongst the richest states.
  • Several decades ago, Sen (1990), noting the skewed ratio of females to males, estimated that nearly 100 million women were missing in the world (almost 40 million in India alone).
  • The stock of missing women as of 2014 was nearly 63 million and more than 2 million women go missing across age groups every year (either due to sex selective abortion, disease, neglect, or inadequate nutrition).


  • While active sex selection via fetal abortions is widely prevalent, son preference can also manifest itself in a subtler form. Parents may choose to keep having children until they get the desired number of sons. This is called son “meta” preference.
  • A son “meta” preference – even though it does not lead to sex-selective abortion – may nevertheless be detrimental to female children because it may lead to fewer resources devoted to them.
  • One indicator that potentially gets at “meta” preference for a son is the sex ratio of the last child (SRLC).
  • India after outlawing sex selection (via the implementation of Pre-Natal Diagnostic Techniques (PNDT) Act, 1994) saw a stabilization of its sex ratio at birth. However, it is not clear whether it resulted from changes in societal preferences or due to increased state regulation of sex-detection technology. SRLC helps us better understand and decompose the underlying factors.
  • A preference for sons will manifest itself in the SRLC being heavily skewed in favor of boys.
  • An SRLC of close to 1.05:1 would imply that parents’ decisions to continue having children is uncorrelated with previous birth being a son or a daughter.
  • Fertility stopping rules:
    • Families continue to have children until they get the desired number of sons. This kind of fertility stopping rule will lead to skewed sex ratios but in different directions: skewed in favor of males if it is the last child, but in favor of females if it is not.
    • Where there is no such fertility stopping rule, the sex ratio will be 1.05 regardless of whether the child is the last one or not.
    • Families where a son is born are more likely to stop having children than families where a girl is born. This is suggestive of parents employing “stopping rules” – having children till a son is born and stopping thereafter.
      • The only exception to this pattern is with regards to the first child. Even parents who have a first-born son are likely to continue having children, which reflects a pure family size preference – Indian parents, on average, want to have at least two children.
  • Findings:
    • Meghalaya stands out as an ideal state because both sex ratio at birth and sex ratio of last child are close to the benchmark.
    • States such as Kerala, do not seem to practice sex selective abortions (since their sex ratios at birth are close to the biological benchmark) but indicate some son “meta” preference (skewed SRLC).
    • Punjab and Haryana, on the other hand, exhibit extremely high son preference and meta preference – the overall sex ratios are significantly above the biological benchmark, and the sex ratio of the last child is heavily male skewed, implying parents are unlikely to stop after having a daughter.
  • A number of reasons for such a son preference:
    • patrilocality (women having to move to husbands’ houses after marriage),
    • patrilineality (property passing on to sons rather than daughters),
    • dowry (which leads to extra costs of having girls),
    • oldage support from sons and rituals performed by sons.
  • Such meta preference gives rise to “unwanted” girls–girls whose parents wanted a boy, but instead had a girl. The number of unwanted girls as 21 million.
  • Tellingly, for example, skewed sex ratios characterize families of Indian origin, even in Canada.


  • Because the challenge is historical and longstanding, no one stakeholder is responsible for creating it or solving it. On gender, society as a whole—civil society, communities, households— and not just any government must reflect on a societal preference, even meta-preference for a son, which appears inoculated to development.
  • The state and all stakeholders have an important role to play in increasing opportunities available for women in education and employment.
  • Understanding the importance of its role, the government has
    • launched the Beti Bachao Beti Padhao and Sukanya Samridhi Yojana schemes,
    • made 26 weeks maternity leave mandatory for women employed in the public and private sectors.
    • Further, every establishment that has more than 50 employees is now required to offer creche facilities.
  • These steps will offer support to women in the workforce.
  • Just as India has committed to moving up the ranks in the ease of doing business indicators, it should perhaps do so on gender outcomes as well.

Click here to download Gist of Economic Survey 2017-18- Chapter 7

Leave a Reply