A child born in India today will grow up to be 49% as productive as she could have been with full health and a complete education. That is not a critic's estimate. That is the World Bank's own Human Capital Index, and it ranks India 116th out of 174 countries, behind Bangladesh, behind Indonesia, and in the index's first edition, behind the West Bank and Gaza, a literal active conflict zone. India has spent every education policy document since 1968 promising itself it would spend 6% of GDP on education. It currently spends about 3.5%. Fifty eight years is a long time to keep almost paying a bill.
Chapter 1How many schools has India actually closed, and can anyone agree on the number?
Depends who you ask, and the fact that nobody fully agrees is itself worth sitting with before the individual numbers.
Parliament was told in one reply that 14,910 government schools closed between 2019 and 2024. A separate compilation citing UDISE+ data puts the figure since 2017-18 at 87,430 schools, mostly government run. Another widely circulated tally, built from state level Right to Information data, claims 89,441 closures with Madhya Pradesh alone accounting for 29,410 of them. What is not in dispute, because it comes straight from the government's own most recent UDISE+ 2025-26 release, is the direction and the pace. Government school closures more than doubled in a single year, from 4,338 in 2024-25 to 8,077 in 2025-26. One widely quoted analysis of the same data put it more bluntly, thirteen schools shutting somewhere in India every single day, more than half of them in one state.
Behind the closures sits an even less flattering number. Over 65,000 government schools now have fewer than ten students enrolled, and more than 5,000 reported zero enrolment in a single academic year, disproportionately concentrated in Telangana and West Bengal. In Uttar Pradesh, 9,508 schools function with exactly one teacher, between them responsible for roughly 6.2 lakh children, which is less a school and more a single adult performing a hostage negotiation with an entire grade's curriculum. Total enrolment across Classes 1 to 12 fell to 232.89 million in 2024-25, the lowest since 2017-18, a net loss of 16 million children from the system over seven years even as India's school age population has not shrunk anywhere near that fast. Where did they go. Mostly to private unaided schools, whose enrolment rose by 4.8 million in a single year, meaning the families who could scrape together the fees left, and the government school network shrank around the families who could not.
Chapter 2Why do the exams keep getting cancelled, and who is actually paying for it?
Start with the number that should have been impossible. On 3 May 2026, over 2.27 million candidates sat NEET UG, India's sole entrance exam for every undergraduate medical seat in the country. On 12 May, the National Testing Agency cancelled the exam entirely, after it had already been conducted, the first time in the exam's history that this has happened, once investigators found a pre circulated guess paper matched the actual question paper closely enough to be more than coincidence. Two million, two hundred and seventy thousand young people who wanted to become doctors had to sit the exam again because the system meant to certify who gets to hold a scalpel could not keep its own question paper secret.
This was not an isolated embarrassment. It was the latest entry in a genuinely long list. NEET UG 2024 was compromised when the paper was allegedly photographed inside an NTA strongroom in Hazaribagh and sold for 30 to 50 lakh rupees more than a day before the exam, a leak the Supreme Court confirmed had directly benefited at least 155 candidates, several of whom mysteriously scored a perfect 720, before declining to order a full national retest. UGC NET 2024, the exam that decides who gets to become a university professor or research fellow, was cancelled the day after it was held when its paper turned up for sale on the dark web, affecting 9.08 lakh candidates who had to simply wait for a new date. In the same year, Uttar Pradesh scrapped its police recruitment exam after 48 lakh candidates had already sat it, and separately cancelled its RO and ARO civil service prelims after the paper went viral on social media before the exam even started. CUET UG saw protests and stone pelting at exam centres over leak allegations. CSIR UGC NET in Haryana faced fresh leak accusations in 2025. By one independent count, at least eight to twelve major national or state level recruitment and entrance exams were compromised by leaks or credible leak allegations between 2024 and 2026 alone, a rate that makes exam integrity failure less an incident and more a functioning, load bearing part of how Indian competitive exams now operate.
Every one of these numbers is a person, usually between seventeen and twenty five, who spent a year or three of their life and a family's savings on coaching, only to have the one gate they were told mattered most simply fail to function as a gate.
Chapter 3Do children who stay in school actually learn anything?
This is the question that should worry a country more than enrolment statistics, because enrolment has genuinely gone up while the thing enrolment is supposed to produce has not kept pace.
ASER 2024, Pratham Foundation's long running household survey, found 76.6% of Class 3 students nationally still could not read a Class 2 level text, an improvement from the pandemic trough but still meaning three in four eight year olds cannot read at the level expected of a seven year old. Only about half of Class 5 students, ten and eleven year olds who have already spent five years in school, can manage the same Class 2 level passage. Arithmetic tells the same story from a different angle, 66.3% of Class 3 students and roughly 70% of Class 5 students could not perform basic arithmetic operations expected at their grade. Push the age group up to 14 to 18 year olds, students old enough to be preparing for board exams and competitive entrances, and a quarter of them still cannot fluently read a Class 2 level text in their own regional language, despite 86.8% of that age group being formally enrolled somewhere.
There is a genuine, fair counterpoint here worth including rather than burying, because the data supports it. Government school reading scores for Class 3 nearly doubled between 2022 and 2024, rising from 16.3% to 23.4% able to read at grade level, government schools have in fact been closing the gap with private schools faster than private schools have been improving, likely helped by post pandemic recovery efforts and the NIPUN Bharat foundational literacy mission. The floor was catastrophic and the floor is now merely bad. That is real progress, and pretending otherwise would be its own kind of dishonesty. It does not change the fact that a system enrolling 97% of its children and still producing this outcome has a quality problem that access alone was never going to fix, a lesson India appears to be relearning once a decade since the Right to Education Act passed in 2009 promising exactly the universal access it delivered, without the learning it did not.
Chapter 4Is vocational training actually filling the gap for students who leave the academic track?
No, and the scale of the shortfall is almost comic once you see the international comparison next to it. Only about 4% of India's workforce, roughly 36 million people out of a working age population many multiples larger, has received any formal vocational training. Compare that to the United Kingdom at 68%, the United States at 52%, Germany at 75%, Japan at 80%, and South Korea at 96%, a country that essentially decided as a matter of national policy that almost nobody would enter the workforce unskilled.
The infrastructure that does exist is running at roughly half capacity. India's network of over 14,000 Industrial Training Institutes has around 25 lakh sanctioned seats, and actual enrolment sits near 12 lakh, a 48% utilisation rate for a system that is simultaneously described as too small to meet demand and too underused to justify its own footprint, both true at once because the courses on offer frequently do not match what local employers are actually hiring for. Of the young people who do complete ITI training, the World Bank's own February 2026 assessment found fewer than half actually secure a job placement, while youth account for a startling 72% of India's total unemployed population. The government's response, a genuinely large 60,000 crore rupee scheme to upgrade 1,000 ITIs, approved in May 2025, alongside an 830 million dollar World Bank loan for the same purpose approved this February, is real money aimed at a real problem. Whether it closes a gap this wide within a single scheme's lifetime is the more honest question nobody in a press release has to answer.
Chapter 5So what does all of this actually cost the country?
This is where the individual scandals stop being individual and start being a single, compounding number, because none of the sections above are separate problems, they are the same problem measured from four different angles, and the World Bank's Human Capital Index is the instrument built specifically to add them all together.
India's HCI score of 0.49 means the productivity lost is not a rounding error, it is roughly half of everything a fully healthy, fully educated Indian workforce could otherwise produce. The component numbers explain exactly where that loss comes from. A child who starts school at four can expect to complete 10.2 years of schooling by eighteen, a genuinely respectable access number that the enrolment statistics above bear out. Adjust those same 10.2 years for what the child actually learned while sitting in the classroom, using harmonised international test scores, and the figure collapses to just 5.8 learning adjusted years. Close to half of the time Indian children spend in a classroom is, by the World Bank's own measurement, generating close to zero usable learning. That gap, 10.2 years of attendance producing 5.8 years of actual education, is the single most precise number in this entire piece, because it is the exact place where enrolment triumphalism and the ASER reading crisis stop contradicting each other and turn out to be describing the same hole from opposite sides.
The World Bank has also run the counterfactual, and it is worth ending on because it is the rare number in this piece that points somewhere other than down. A simulation of raising India's effective years of quality schooling toward 12 years lifts the country's expanded Human Capital score from 159 to 174 and translates into a 14.9% rise in lifetime earnings for the next generation, not a rounding improvement, a genuine, quantified dividend sitting on the table. Compare that potential 14.9% earnings gain against the 6% of GDP this country has been formally promising itself in every education policy document since the Kothari Commission reported in 1966, restated in 1968, restated again in 1986, restated again in the National Education Policy of 2020, and never once actually delivered, currently sitting at roughly 3.5%. India is not short of a plan. It has had the same correct plan, gathering dust, for fifty eight years. What it has been short of, across every school that closed with a good building and no students, every exam paper that leaked for the price of a small flat, and every classroom that produced 10.2 years of attendance and 5.8 years of learning, is simply the will to pay for the plan it already wrote.