India’s missing women: what the 1,020 sex ratio headline hid
NFHS-5 found 1,020 women per 1,000 men, but the survey overcounts women and 9 million girls were lost to sex selection. The real sex ratio in India.
In December 1990 the economist Amartya Sen published an essay in The New York Review of Books with a blunt title: “More Than 100 Million Women Are Missing”. He meant that across Asia and North Africa, tens of millions more women should have been alive, given how many girls are born and how long women usually live. Neglect, unequal food and care and, increasingly, sex-selective abortion had made the difference. India was at the centre of that count. The next year’s Census recorded 927 women for every 1,000 men.
Thirty-one years later, in November 2021, a headline seemed to close the chapter. The fifth National Family Health Survey (NFHS-5) found 1,020 women for every 1,000 men, the first time any round of the survey had recorded more women than men. “India is no more a country of ‘missing women’,” IPB News wrote, as many outlets did. The sex ratio in India has improved. But the evidence gathered since then shows that the 1,020 figure says less than it seemed to, and that the girls who are never born are still missing.
Why 1,020 does not mean more women than men
NFHS is a household survey, not a census. It counts the people present in a household on the night before the interview. Nand Lal Mishra, a research fellow at the International Institute for Population Sciences, Mumbai, told Down To Earth that this method can overstate the share of women, because men who migrate for work are more likely to be away. He called reading the result as “more women than men” “completely misleading”, while crediting a real improvement over 30 years.
Jashodhara Dasgupta, who has long worked on sex ratios, made the same point to the magazine: “NFHS counts only certain women, who belong to specific demographic categories. There is a bias in it.” The survey also left out institutional homes such as hostels, where many single men live, and part of its fieldwork ran through the Covid-19 months, when migrant workers had returned to their villages.
A longer look at the data, published in January 2025 by Pramit Bhattacharya and Nandlal Mishra for Data For India, found that every one of the five NFHS rounds has produced a more female-skewed sex ratio than the Census or official projections. In NFHS-5, the gap appeared in all 21 large states, ranging from 22 more women per 1,000 men than projected in Chhattisgarh to 168 in Bihar. Their conclusion was that the NFHS-5 figure cannot be used to claim a sharp improvement over the Census. The last Census, in 2011, recorded 943 women per 1,000 men.
The number that matters: girls at birth
The figure that shows whether girls are being prevented from being born is the sex ratio at birth. NFHS-5 put it at 929 girls for every 1,000 boys born in the previous five years, up from 919 in NFHS-4, the Health Ministry said in a December 2021 statement. The Sample Registration System, which tracks births through the year, shows a similar slow climb: 899 in 2014 to 913 in 2021, All India Radio reported. With no sex selection, nature produces roughly 950 girls for every 1,000 boys.
Put in people rather than ratios, the gap is large. A 2022 Pew Research Center analysis, reported by Down To Earth, estimated that at least 9 million girls went “missing” to sex-selective abortion in India between 2000 and 2019. The ratio of boys to girls at birth had widened from about 105 to 100 in 1950 to 111 to 100 in 2011, before easing to 108 to 100 in NFHS-5. The share of women who said they wanted more sons than daughters fell across every major religious group over the same two decades.
The government’s own Economic Survey of 2017–18 reached a starker estimate. It put the number of missing women at about 63 million and described a further 21 million “unwanted” girls, born to parents who had wanted a son, the Observer Research Foundation noted. The Survey found that when the last child in a family is a boy, the ratio is skewed far more than the country average, because parents keep having children until a son arrives. In Haryana it estimated about 2,300 boys for every 1,000 girls among last-born children.
Why it persists
The reasons are well documented: sons are expected to carry the family name, inherit land and support parents in old age, while a daughter is too often seen through the cost of her wedding. Dowry was banned in 1961, but it lingers as “gifts”. The Pre-Conception and Pre-Natal Diagnostic Techniques (PCPNDT) Act, passed in 1994 and in force from 1996, makes it a crime to reveal or select a baby’s sex. Enforcement has been weak. By the end of 2013, 17 years after the law took effect, only 143 people had been convicted under it across India, from 1,833 registered cases, the Health Minister told the Rajya Sabha, Down To Earth reported. Uttar Pradesh and Bihar had reported no convictions at all.
There is also a gap between rural and urban India, as IPB News noted in 2021: NFHS-5 recorded 1,037 women per 1,000 men in villages but 985 in cities. Part of that is migration. IPB News also pointed to the easier access to sex-selection technology in towns.
What the authorities say
The Health Ministry highlighted the 10-point rise in the sex ratio at birth in its December 2021 statement on NFHS-5, and the Centre’s main campaign against sex selection, Beti Bachao Beti Padhao, has run since 2015. The improvement is real, and it shows in every major data source. The question is the pace: at the rate SRS has recorded since 2014, about two points a year, it would take well over a decade to reach the natural level.
The wider picture
Missing women are not only those never born. Sen’s original count included girls and women who died because they were fed less, taken to the doctor later or married too young. That last thread runs through IPB’s report on child marriage in the Sundarbans, where families facing disaster marry daughters off as one fewer mouth to feed. Where women do get money and a say, as in Bihar’s JEEViKA self-help groups, the picture of a daughter’s worth starts to change.
India’s last completed Census was in 2011, so the most reliable count of who is alive is 15 years old; the population count for the delayed next Census is scheduled for 2027. Until its data arrive, survey figures like 1,020 will keep being read as more than they are.
What would change it
The tools exist. A Census would give a full count. Strict enforcement of the PCPNDT Act, with audits of ultrasound clinics in districts where the birth ratio is worst, would make sex selection riskier. Equal inheritance in practice, not just in law, and old-age security that does not depend on a son, would weaken the reasons behind it. These decisions sit with the Union Health Ministry, state health departments that license and inspect clinics, and the Registrar General of India, which runs the Census.
Sen’s essay was written to make an invisible loss visible. A single good-looking ratio should not make it invisible again.
More stories beneath the headline in Third Eye. For daily news, follow IPB News.
Sources: IPB News, 30 November 2021 · Down To Earth, 25 November 2021 · Data For India, January 2025 · Ministry of Health and Family Welfare (PIB), 17 December 2021 · All India Radio, May 2025 · Down To Earth on the Pew Research Center analysis, September 2022 · Observer Research Foundation on the Economic Survey 2017–18 · Down To Earth on PCPNDT enforcement
