From the Knowledge Garden — where we nurture ideas that bridge the gaps between policy dreams and classroom realities

Meera is ten years old. She is sitting in a Class 5 classroom in a government school in rural Madhya Pradesh, one of the roughly 650,000 children a small army of citizen volunteers visits every couple of years to ask a single, deceptively simple question: can you read this?

The volunteer hands Meera a card. On it are four sentences, written at a Class 2 reading level , the kind of text an eight-year-old, not a ten-year-old, is expected to read fluently. Meera looks at it for a long moment. She sounds out the first word. Then the second. By the third sentence, she has stopped trying to read and started guessing, filling in words that look plausible from the shapes of the letters she recognises.

Meera is not failing a test that measures whether she is exceptional. She is failing a test that measures something much more basic: whether the last five years of her formal education have given her the single most foundational skill a school is meant to provide , the ability to read and understand a short sentence written in her own language.

She is not alone. She is, in fact, the statistical majority.

This is the story the annual ASER survey and the World Bank's Learning Poverty index have been quietly, rigorously documenting for over a decade , and it is a story that rarely makes the front page, because it lacks the drama of a scam, a scandal, or a school collapse. It is simply a slow, patient, devastating accumulation of ordinary school days that add up to less than they should.

This piece continues the Knowledge Garden's exploration of Indian K-12 education, alongside The Tuition Trap, DARPAN, ANKUR, SETU, and PRAYAS. If the Tuition Trap asked why India spends so much money making up for what schools don't deliver, this piece asks a harder, more foundational question: what, precisely, are schools not delivering , and why does it matter more than almost anything else we discuss in Indian education policy?

In 2019, the World Bank and UNESCO's Institute for Statistics introduced a new global metric, designed to cut through decades of education-policy conversations that focused almost entirely on enrolment , how many children are in school , while quietly ignoring a much more uncomfortable question: how many of them are actually learning anything once they're there.

Learning Poverty is defined as the inability to read and understand a simple, age-appropriate text by age 10.

It's a deliberately narrow, deliberately blunt measure, and that's precisely its power. It doesn't ask whether a child can recite the alphabet, or identify letters, or sound out individual words under supervision. It asks whether a ten-year-old, handed a short passage written in language appropriate for their age, can read it and tell you what it means , the single gateway skill on which every other subject, every future exam, and arguably every future economic opportunity depends.

The index combines two components: schooling deprivation (children who are out of school altogether, and are assumed unable to read) and learning deprivation (children who are in school, attending regularly, but still haven't reached minimum reading proficiency). This combination matters, because it prevents a country from looking good simply by getting children into classrooms , a trap many countries, India included, have historically fallen into when celebrating enrolment statistics as if enrolment were the same thing as education.

When the metric first launched globally in 2019, the number was already sobering: 53% of ten-year-olds across low- and middle-income countries could not read and understand a simple story by the end of primary school. By the time the World Bank, UNESCO, UNICEF, the UK's FCDO, USAID and the Gates Foundation jointly released their 2022 update , accounting for pandemic school closures , the global figure had climbed to a staggering 70%. Seven in ten of the world's ten-year-olds in developing economies, unable to read a simple sentence. The report's authors called it the worst shock to global learning in recorded history, and estimated that without serious intervention, this single generation of children stood to lose a combined $21 trillion in lifetime earnings , roughly 17% of 2022's entire global GDP, in present value terms.

India's own Learning Poverty rate, calculated using pre-pandemic data, stood at 56%, according to World Bank estimates , a fraction better than the South Asia regional average, and only marginally better than the average for lower-middle-income countries as a group. In plain terms: more than half of India's ten-year-olds, even before COVID-19 disrupted two years of schooling, could not read and understand a short, simple text appropriate for their age.

Let that sit for a moment against the country's more familiar education statistics , the ones that tend to dominate press releases and policy speeches: near-universal primary enrolment, rising literacy rates by the traditional Census definition, expanding digital infrastructure, a growing edtech sector valued in the billions. All of these are true. All of them coexist, uncomfortably, with the fact that a majority of Indian ten-year-olds cannot do the one thing a decade of schooling is supposed to guarantee.

This is not a contradiction. It's the entire point of why the Learning Poverty metric exists , to expose exactly the gap between access to school and learning inside it, a gap that India's own long-running domestic survey, ASER, has been documenting in granular, state-by-state, grade-by-grade detail since 2005, long before the World Bank gave the phenomenon a global name.

The Annual Status of Education Report, published by the NGO Pratham, is one of the largest citizen-led household surveys in the world , reaching roughly 650,000 children across more than 600 rural districts in India's most recent 2024 round. Unlike school-based assessments, which only capture children who are present and enrolled, ASER's volunteers go door to door, testing children in their own homes, capturing a truer picture of what an entire age cohort , not just the children a school system chooses to count , actually knows.

The core finding, repeated with remarkable consistency across nearly two decades of surveys, is this: a large share of Indian children move up through the primary grades without acquiring the reading and arithmetic skills those grades are supposed to certify.

In the 2022 survey, conducted as the country was still recovering from pandemic-era school closures, only 20.5% of Class 3 children nationally could read a Class 2-level text , a text two grade levels below what a Class 3 child should already have mastered, since the benchmark itself is already set one year behind. ASER 2022 also found that roughly 80% of Class 3 students in rural India could not read a Grade 2 text, and a similarly large share could not perform basic Grade 2-level arithmetic , figures that, read plainly, describe a foundational learning crisis affecting the vast majority of India's rural primary-school children, not a struggling minority.

The good news , and there is real good news here, worth stating clearly rather than burying under alarm , is that ASER 2024, released in January 2025, documented a genuine recovery. The percentage of Class 3 children able to read a Class 2-level text rose to 23.4%, up from 16.3% in the depths of the pandemic-affected 2022 survey, and now modestly ahead of the pre-pandemic 2018 benchmark of 20.9%. Class 5 children reading at the Class 2 level reached 44.8% in 2024, recovering from 38.5% in 2022 and nearly matching 2018's 44.2%. Basic arithmetic showed similar gains: Class 3 children able to perform subtraction rose to 33.7% in 2024, above both the 2022 figure of 25.9% and the pre-pandemic 2018 rate of 28.2%.

Government schools, notably, drove much of this recovery , outperforming the pace of improvement in private schools, a genuinely important and under-reported finding that complicates the common assumption that private schooling is automatically superior. ASER Centre's own director has attributed part of this gain to the increased national focus on foundational literacy and numeracy under NEP 2020 and the NIPUN Bharat mission, suggesting that targeted policy attention, when sustained, does move the needle.

But sit with the actual numbers again, stripped of the (genuinely encouraging) trend lines: in 2024, more than three out of every four Class 3 children in rural India still could not read a text meant for a child two years younger than them. More than half of Class 5 children , that is, ten- and eleven-year-olds, exactly the age band the World Bank's global metric targets , still could not read at the Class 2 level. Recovery from a crisis is not the same as solving it. A patient who has recovered from the worst of a fever is not yet healthy; they are simply less sick than they were.

It would be easy , and it is, in fact, the most common policy failure mode in Indian education discourse , to treat reading proficiency as one metric among many, sitting alongside enrolment ratios, teacher-student ratios, infrastructure indices, and digital-access statistics, all worthy of roughly equal attention. This is a category error, and understanding why is essential to understanding why this particular number deserves to anchor an entire policy conversation.

Reading is the gateway skill, not a subject. A child who cannot read fluently by age nine or ten does not simply fall behind in "language" class. They fall behind in mathematics, because word problems require reading comprehension. They fall behind in science, because textbooks are written in prose. They fall behind in social studies, in the exam instructions themselves, in every subsequent grade's curriculum, which assumes , reasonably, since it's written to a syllabus, not to an individual child , that foundational reading is already in place. Every year a child spends in school without closing this gap is a year spent building on a foundation that was never poured.

The gap compounds, silently, for years before it becomes visible. This is precisely the mechanism we described in our earlier Knowledge Garden piece on the post-school learning gap, and in The Tuition Trap: a Class 3 child who can't quite read at grade level isn't failing a test that anyone is watching closely. They're simply a little quieter in class, a little more reliant on guessing and memorisation, a little more dependent on rote repetition to get through exams that don't actually test comprehension. By the time this gap becomes undeniable , often around board exam years , it has had a decade to compound, and the interventions available at age 16 are vastly more expensive, in time and money, than the interventions that would have worked at age seven.

It is the single best predictor economists have found for lifetime earnings and national economic growth. This is not a sentimental claim; it is the World Bank's own stated rationale for building the Learning Poverty index in the first place, and for attaching a $21 trillion global price tag to the pandemic-era spike in the number. Human capital , the skills and knowledge a population actually carries, not merely the number of years they sat in a classroom, is now understood by development economists as a stronger predictor of a country's long-run growth trajectory than most traditional measures of physical infrastructure or capital investment. A nation with near-universal school enrolment but low reading proficiency at age ten is, in the language of this framework, accumulating years of schooling without accumulating the human capital those years are meant to produce.

It exposes the limits of every other education reform if left unaddressed. New curricula, better school infrastructure, expanded digital access, glossy edtech platforms, ambitious skilling missions aimed at India's demographic dividend , all of it is built on an assumption that children arriving at each stage can actually read the material in front of them. If that assumption is false for a majority of children at the foundational stage, every subsequent investment in the system operates at a fraction of its intended efficiency. This is, in a very real sense, the argument for why NEP 2020 made foundational literacy and numeracy , not board exam reform, not skilling, not technology , its explicit first priority, with a national mission (NIPUN Bharat) built specifically around achieving universal foundational literacy and numeracy for every child by Grade 3, by 2026–27.

Learning poverty in India is not evenly distributed, and the variation itself tells an important story.

State-level ASER data shows genuinely wide gaps in both starting levels and rates of improvement. States such as Gujarat and Uttar Pradesh posted double-digit percentage-point gains in government-school reading levels between 2022 and 2024, while states like Himachal Pradesh and Bihar saw far more modest gains of four to five percentage points over the same period , despite Himachal Pradesh traditionally being considered one of India's stronger-performing states on education indicators, a reminder that past performance doesn't guarantee continued momentum. Some states, including Punjab, Andhra Pradesh and Telangana, actually recorded declines in certain grade-level indicators even as the national picture improved , evidence that a national recovery narrative can mask serious regional backsliding that state-level policymakers need to address directly rather than assume is covered by the aggregate trend.

The gap between government and private schools adds another layer of nuance that complicates easy assumptions. Government school enrolment, which had risen sharply during the pandemic years (from 65.6% of rural 6–14-year-olds in 2018 to 72.9% in 2022, as cost-conscious families moved away from fee-paying private schools), has since reversed almost back to pre-pandemic levels, settling at 66.8% in 2024 as the broader economy recovered. Yet it was government schools, not private ones, that drove most of the reading and arithmetic recovery over the same period; private school performance remained largely flat between 2022 and 2024. This finding deserves far more attention in public discourse than it typically receives, since so much of the narrative around Indian schooling implicitly assumes private is synonymous with better, an assumption ASER's own data increasingly complicates.

And then there is the gap that predates any of this: the rural-urban divide, the gender gap (which persists in some regions and has closed or reversed in others, notably several southern states where girls now match or outperform boys on both reading and digital literacy indicators), and the gap between children whose parents are literate and can support learning at home versus those whose parents cannot , precisely the dynamic we described in The Tuition Trap as one of the core drivers of the demand for private tuition, and one that operates just as powerfully, if less visibly, at the foundational reading stage.

Conversations with teachers working directly in early-primary classrooms reveal a pattern that aligns closely with what the survey data shows, but adds texture the numbers alone cannot.

A Class 3 teacher in a government school in rural Uttar Pradesh, whose students have benefited from her state's above-average recent gains, described the turning point in her own classroom not as a new textbook or curriculum, but as a shift in daily practice: dedicated, protected reading time each morning, before any other subject, with no other academic pressure attached to it, just decoding and comprehension practice, at each child's actual level rather than the grade-assigned level. This is, in miniature, exactly the philosophy behind NIPUN Bharat's foundational literacy push, and exactly the kind of structural, in-school intervention we argued for in The Tuition Trap as the genuine alternative to private tuition , closing gaps at the source, inside the school day, rather than outsourcing the problem to whoever a family can afford to pay after hours.

A teacher-trainer working across several low-performing districts described a different, more sobering observation: many Class 5 teachers, faced with a classroom where a large share of children cannot yet read at grade level, default to the only strategy that lets the syllabus keep moving , reading aloud to the class and having children repeat or copy, rather than diagnosing and remediating the underlying decoding gap. This produces students who can appear, on the surface, to be "keeping up" , able to recite, to copy from the board, to pass a rote-memorisation-based unit test , while the actual, measurable skill the Learning Poverty index is designed to catch remains unbuilt. It is a vivid illustration of exactly the distinction our earlier piece "From Rote to Reasoning" drew between the appearance of learning and the substance of it.

This is not, encouragingly, an area where policymakers are working blind. NEP 2020 made foundational literacy and numeracy its explicit first pillar, ahead of every other reform in the document, on the theory, well-supported by the global evidence above, that no subsequent reform matters if this foundation isn't secured. NIPUN Bharat, launched in 2021, set the explicit national target of universal foundational literacy and numeracy by the end of Grade 3, by the 2026–27 academic year, backed by structured pedagogy, teacher training, and dedicated assessment tools rolled out at scale across states.

The ASER 2024 recovery, and particularly the finding that government schools improved faster than private ones over the 2022–24 period, is genuine evidence that this focus is producing measurable results, not merely generating policy documents. That said, "recovery to slightly above pre-pandemic levels" and "on track for universal foundational literacy by 2026–27" are two very different claims, and the gap between them, the gap between roughly 45% of Class 5 children reading at grade level and the 100% the national mission targets , remains enormous, with less than two years left on the mission's own timeline as of this writing.

What the state-level variation tells us, meanwhile, is that this is not primarily a resource-constraint problem in the way it's sometimes framed. States with comparable resource levels have produced very different rates of improvement, which points toward the quality and consistency of classroom-level pedagogical practice, the kind of structured, phonics-based, level-appropriate small-group reading instruction described by the Uttar Pradesh teacher above, as the more decisive lever, more than infrastructure spending alone. This aligns with the growing body of international evidence, referenced in ASER's own analysis, that structured pedagogy interventions, properly implemented and sustained over multiple years rather than launched as one-off pilots, are the most reliable lever available for closing foundational learning gaps at scale.

For parents: The single most useful thing a parent can do for a child's long-term academic trajectory is not enrolling them in the earliest possible tuition or the most prestigious available school, it is checking, directly and regularly, whether their child can actually read age-appropriate text with comprehension, not just recite or recognise it. A report card showing passing marks in "English" or the child's mother tongue is not the same information as a direct check of whether the child can read a simple, unfamiliar paragraph and explain what it means. This single, low-cost, five-minute check, repeated every few months through the early primary years, catches foundational gaps years before they become board-exam crises , and years before they become the kind of gap that a Tuition Trap-style coaching subscription is marketed to "solve."

For teachers and school leaders: Protected, daily reading time at each child's actual level , not the grade-assigned level, appears repeatedly, across both ASER's findings and international evidence, as one of the most effective and lowest-cost interventions available. This requires, in practice, a willingness to group children by current reading level rather than strictly by grade for at least part of the school day, a structural change that costs little in resources but requires real pedagogical will and, often, permission from school leadership to depart from rigid grade-based scheduling.

For school administrators: Treating foundational reading assessment as a routine, low-stakes, frequent diagnostic, rather than an annual, high-stakes exam, allows problems to surface and be addressed while they are still cheap to fix. The ASER model itself, simple, quick, one-on-one oral assessment, is a template any school can run internally at a fraction of the cost of a formal standardised test.

For edtech companies: The most valuable, most under-served market opportunity in Indian K-12 edtech is not another adaptive-learning platform for board-exam or competitive-exam preparation, aimed at families who can already afford tuition, it is genuinely effective, vernacular-language, low-bandwidth foundational reading tools aimed at exactly the age band (six to ten years) and exactly the population (rural, government-school, first-generation-literate households) that the Learning Poverty and ASER data show is least served today. This is precisely the argument we made in our earlier piece on the digital divide and AI as a potential equaliser , and the Learning Poverty data makes the case for it with unusual clarity and urgency.

For Ministry of Education stakeholders: ASER 2024's core finding, that government schools improved faster than private ones over the pandemic-recovery period, is genuinely good news for a mission (NIPUN Bharat) that operates primarily through the government school system, and deserves to be communicated far more prominently than it currently is, both to sustain political and public support for the mission and to correct a public narrative that too often assumes government schools are a lost cause. At the same time, the sharp state-level variation in improvement rates argues for far more granular, district-level accountability and support , the same struggling states and districts appear repeatedly across multiple ASER cycles, suggesting the current model of national-level target-setting needs a much more targeted, differentiated implementation strategy for the specific geographies where progress remains slowest.

For students, especially older students and student-teachers: For those training to become teachers, or considering it, foundational literacy instruction, teaching a child to read, not simply reciting a syllabus to a child who already can, is arguably the single highest-leverage skill in the entire Indian education system today, more consequential to national outcomes than almost any other specialisation a young teacher could choose to develop expertise in.

Meera, the ten-year-old at the start of this piece, is not a statistic in the way that phrase is usually meant, as something distant, aggregated, and abstract. She is one of tens of millions of specific children, in specific classrooms, on specific school days, each one representing an entirely fixable problem that has simply not yet been fixed for them, in time.

The remarkable thing about the Learning Poverty number, and about ASER's two decades of patient, granular data underneath it, is how solvable the underlying problem actually is, relative to almost everything else this Knowledge Garden series has explored, the funding structures of the coaching industry, the digital divide, the teacher shortage, the curriculum-skills mismatch. Teaching a child to read by age eight or nine is not a frontier scientific challenge. It's a well-understood pedagogical problem, with well-documented solutions, that mostly requires consistent, well-supported, appropriately-targeted classroom practice, sustained over several years, rather than a fundamentally new discovery.

That is precisely what makes the number so important to keep in front of every policymaker, every school leader, every parent, and every edtech founder building the next product for Indian classrooms: not because it is the most dramatic statistic in Indian education, but because it is, in a very real sense, the one everything else is built on top of. Every curriculum reform, every skilling mission, every ambition of Viksit Bharat by 2047 assumes a population that can read what's put in front of it. Right now, for a majority of India's ten-year-olds, that assumption doesn't yet hold, and until it does, it deserves to be the first number in the room, not the one nobody talks about.

Further reading from the Knowledge Garden: