English only · Odia translation in progress

The First Classroom

The bell has not rung yet at Kashipur Government Primary School, but the children are already here.

It is 9:15 on a Tuesday morning in Rayagada district, one of the poorest in Odisha, one of the poorest in India. The school building --- a two-room concrete structure with a tin roof --- sits at the edge of a Kondh tribal hamlet, across a dirt path from a paddy field that has been fallow since the family that owned it left for a brick kiln in Andhra Pradesh last October. They took their two children with them. They will not return until May, and neither will the children’s enrollment records.

Inside the first room, a blackboard still carries yesterday’s Odia alphabets, half-erased by a palm that gave up midway. Thirty-seven students from Classes I through III share this space. Most sit on the floor. A few share two wooden benches that arrived during a Mo School improvement drive in 2019. The second room belongs to Classes IV and V, but today it is empty because the teacher assigned to those classes has been on “training leave” for eleven days. No substitute was sent.

So one teacher manages all five classes.

She is 34, a Siksha Sahayak --- a contract teacher --- earning Rs 7,500 per month. She has a D.El.Ed degree from the DIET in Koraput. She commutes 14 kilometres on a scooter from the block headquarters because she cannot afford to live closer and no teacher housing exists. She arrived at 9:40. By the official schedule, classes began at 9:00.

At 11:30, the real event of the school day will happen. Three women from a local Mission Shakti Self-Help Group will arrive carrying steel containers with rice, dal, and a vegetable curry. The mid-day meal. For most children in this school, it is the most reliable meal of the day. It is also the reason several of them are here at all. Their parents did not send them to learn the alphabet. They sent them to eat.

By 1:00 PM, the school day is effectively over. The students who remain after the meal are the ones whose parents work in the fields nearby and have nowhere else to leave them. The teacher assigns the older ones to “help the younger ones” and begins filling out the attendance register, the UDISE+ data entry forms, and the midday meal consumption log. This administrative work --- mandated by a system that measures everything except learning --- takes the remaining forty minutes of her day.

This is the first classroom. Not the one in the policy document. Not the one in the UDISE+ dashboard. The one that actually exists.

And what happens here --- what children absorb or fail to absorb in these rooms across 61,565 schools in 30 districts --- determines everything that comes after. The secondary school teachers who will inherit these students, the colleges that will enroll them, the competitive exams they will attempt, the jobs they will or will not get, the migration decisions they will make, the wages they will earn or fail to earn for the rest of their lives.

In software, there is a principle so old it has become a cliche: garbage in, garbage out. The quality of the output can never exceed the quality of the input. No algorithm, however sophisticated, can rescue data that was corrupted at the point of entry. You can run it through a hundred processing stages, apply the most elegant transformations, build the most beautiful interfaces to display the results --- and the output will still be wrong, because the input was wrong.

Odisha’s primary education system is the input layer of the state’s human capital pipeline. Everything downstream --- secondary education, higher education, workforce skills, economic productivity, migration patterns, even political participation --- processes what this layer produces. And the data this layer produces is corrupted at the source.

Not because children cannot learn. Because the system designed to teach them has optimized for the wrong metric.

The Enrollment Miracle and the Learning Crisis

Here is a number that looks like success: 98%.

That is the approximate enrollment rate for children aged 6-14 in Odisha’s schools. By the standard the world uses to measure educational access --- are children in school? --- Odisha has effectively solved the problem. The Sarva Shiksha Abhiyan campaign, the Right to Education Act of 2009, the mid-day meal scheme, the construction of tens of thousands of school buildings across the state over two decades --- these policies achieved what would have seemed impossible in the 1990s. Nearly every child in Odisha is enrolled in a school.

Now here is a number that looks like failure: 75%.

That is the approximate proportion of Class 5 students in rural Odisha who cannot perform basic division --- a mathematical operation expected at the Class 3 level. According to ASER 2018 data, only 25.4% of Class 5 students in the state could divide. Another 24.5% could do subtraction but not division. The remaining half could, at best, recognize two-digit numbers.

These two numbers --- 98% enrollment and 75% inability to divide --- are not contradictions. They are the same system producing both outputs simultaneously. The system enrolled children successfully. It did not educate them.

This is the pattern that education economist Karthik Muralidharan calls “schooling without learning,” and it is the defining failure of Indian public education. But what makes Odisha’s version of this failure particularly instructive is the sheer range of outcomes within a single state.

Take reading. ASER 2018 measured the percentage of children in Classes III-V who could read a Standard II level Odia text --- essentially, whether a child two to four years into school could read simple sentences. In Jajpur district, 70.6% could. In Jagatsinghpur, Bargarh, Kendrapara, Ganjam, Khordha, and Nayagarh, the figure ranged from 60% to 70%. These are outcomes that, while imperfect, suggest a functioning system. Children are learning to read.

In Malkangiri, the figure was 14%.

Fourteen percent. Six out of seven children in Classes III through V could not read a sentence appropriate for a second-grader. In Koraput, Nuapada, and Rayagada, the numbers hovered below 20%. The gap between Jajpur (70.6%) and Malkangiri (14%) is not an educational disparity. It is an educational apartheid --- a 56 percentage point chasm within the boundaries of a single state, administered by the same government, funded by the same budget, operating under the same curriculum.

The gap between Khordha (86.88% literacy) and Nabarangpur (46.43% literacy) --- over 40 percentage points --- is equivalent to the gap between developed and least-developed nations on global literacy indices. The undivided Koraput region, now split into four districts, operates at educational levels comparable to sub-Saharan Africa. Not metaphorically. Statistically.

And then there is the board exam illusion. In 2024, the BSE Odisha Class 10 examination recorded a pass rate of 96.07%. Girls passed at 96.73%, boys at 95.39%. Khordha district hit 97.98%. Over 3,200 schools achieved a 100% pass rate. By the metric the system uses to evaluate itself --- board exam results --- Odisha’s schools are performing brilliantly.

But board exams and ASER assessments measure fundamentally different things. Board exams test whether students can reproduce memorized content under examination conditions. ASER tests whether children can actually read, comprehend, and compute. The divergence between 96% pass rates and 75% computational illiteracy at Class 5 is not a paradox. It is a system that has learned to produce passing certificates without producing learning.

In software terms, this is validation without verification. The system validates that a student completed the required process --- attended enough days, sat for the exam, produced answers on paper. It does not verify that the student actually acquired the knowledge the process was supposed to deliver. When you validate without verifying, you get records that look clean while the underlying data remains corrupt.

The ASER 2024 data offers a more hopeful reading. Odisha was named among states with more than a 10 percentage point increase in the proportion of Class III children able to read Class II level text between 2022 and 2024. This improvement was driven primarily by government schools, which matters enormously in a state where 82% of schools and 88% of enrollment (ages 6-14) are in the public system. But the improvement is a recovery from catastrophe, not an ascent to adequacy. The 2022 baseline was a post-pandemic crater --- schools had been closed for nearly two years, and government school children, lacking access to online learning, had fallen further behind than their private school peers. Recovering from that crater is necessary. It is not sufficient.

The fundamental question remains: why does a system that has succeeded at enrollment fail so comprehensively at learning?

The answer begins with teachers.

The Teacher Who Wasn’t There

Consider the arithmetic of a single-teacher school.

A teacher is assigned to a primary school in a tribal block in Kandhamal or Gajapati. She is responsible for Classes I through V --- five grade levels, each requiring different material, different pedagogical approaches, different developmental engagement. She has perhaps 30-50 students across these five classes, though the actual number who show up on any given day may range from 15 to 40 depending on the season, the weather, and whether parents have migrated.

She teaches. She also maintains the attendance register. She fills out the UDISE+ data forms. She manages the mid-day meal logistics. She conducts parent outreach. She participates in mandated training programs (for which she must travel to the block or district headquarters, leaving the school unstaffed). She serves election duty when called. She manages the school infrastructure --- reporting broken furniture, leaking roofs, non-functional toilets.

On days when she is actually present and teaching, she divides her attention across five grade levels in a single room. She writes the Class III lesson on the blackboard, assigns the Class I students to practice letters on their slates, asks the Class V students to read silently from their textbooks, and attempts to explain a concept to Class II while keeping an ear on whether the Class IV students are doing the arithmetic she assigned them. The pedagogical literature calls this “multi-grade teaching” and treats it as a specialized skill requiring specific training. The training she received at her DIET mentioned multi-grade techniques in a single module. She was never shown how to actually do it.

Now consider what happens when she is absent.

National studies have consistently found teacher absenteeism rates of 20-25% in government schools across India. The causes are structural, not moral: mandated training programs, administrative duties at the block office, election duty, personal leave in remote postings with no medical facilities nearby. When the sole teacher in a single-teacher school is absent, the school does not operate with reduced capacity. It shuts down. There is no substitute teacher system in most tribal blocks. There is no automated learning module that can fill the gap. The children who walked to school that morning --- some of them walking 2-3 kilometres through forest --- find a locked door or an empty classroom, and they go home.

Multiply this across hundreds of single-teacher schools in tribal and western Odisha, multiply by the 20-25% absence rate, multiply by the 200 working days in a school year, and you get a rough estimate: in the most vulnerable districts, the effective school year may be 150-160 days rather than 200. And on the days the school is open, the teacher is dividing her attention five ways.

Odisha’s Pupil-Teacher Ratio looks acceptable on paper. At the primary level, it averages 16:1, well within the RTE norm of 30:1. At upper primary, it is 15:1. But these averages are, to use a statistical metaphor the system might appreciate, a classic case of the statistician who drowned in a river with an average depth of two feet. Coastal districts like Jagatsinghpur and Kendrapara may have PTRs of 10-15:1 --- effectively overstaffed relative to their declining enrollment. Tribal districts like Malkangiri and Nabarangpur may have PTRs exceeding 40:1 in remote habitations. The teachers are where the students are not, and the students are where the teachers are not.

The vacancy numbers tell the story. In 2024, the state advertised over 14,000 teaching positions in a single recruitment cycle: 2,629 positions in ST/SC Development Department schools, 5,530 vacancies across the SSD system, 6,025 Leave Training Reserve positions for government secondary schools. An education system that needs to hire 14,000 teachers in one year is not experiencing a staffing gap. It is experiencing a staffing crisis.

And then there is the two-tier system. Odisha has for decades employed Siksha Sahayaks --- contract teachers hired at a fraction of regular teacher salaries. A Siksha Sahayak might earn Rs 5,000-8,000 per month for performing the same duties as a regular teacher earning Rs 25,000-50,000. They serve in the most remote and difficult postings --- the schools regular teachers refuse to accept. They have no job security, no pension, no career progression. They are, in effect, the education system’s version of the dadan laborer: someone who does the hardest work for the lowest pay, with the least protection.

This matters for learning outcomes not because contract teachers are inherently worse than regular teachers --- research on this is mixed --- but because a system that treats its frontline workers as disposable cannot expect those workers to deliver transformative results. The Siksha Sahayak in Rayagada earning Rs 7,500 a month knows that her counterpart in a Bhubaneswar government school earns four to five times as much for an easier job with better infrastructure, shorter commute distances, and students who come from literate homes. Morale is not a policy metric, but it affects every classroom interaction.

The deepest problem, though, is not absenteeism or vacancies or pay scales. It is training. The share of Odisha’s education spending allocated to teacher training is less than or equal to 1% of the total education budget. This places Odisha alongside Bihar and West Bengal as states with the lowest investment in teacher professional development in India. The 30 DIETs and 31 ETEIs across the state produce teachers who have been taught about teaching. They have not been taught how to teach children who arrive in Class I with no prior exposure to letters, numbers, or the Odia language itself --- children who speak Kui, Sora, Mundari, Santali, or one of dozens of tribal languages at home and encounter Odia for the first time in the classroom.

Teaching a first-generation learner whose mother tongue is not the medium of instruction is among the most difficult pedagogical challenges in education. It requires specific techniques, specific materials, and continuous support. The training system does not provide any of this at scale.

Here is the vicious cycle: undertrained teachers deliver poor instruction. Poor instruction produces weak learning outcomes. Weak learning outcomes are masked by generous board exam marking. High pass rates create the illusion that the system is working. The illusion removes the urgency for reform. Reform does not happen. Teachers remain undertrained. The cycle repeats.

In software development, this is called a silent failure --- a system that produces incorrect output without throwing an error. Silent failures are the most dangerous kind because they persist undetected, corrupting downstream processes while every dashboard shows green.

The Meal That Became the Policy

At 11:30 AM in government schools across Odisha, something interesting happens. Attendance goes up.

Studies have consistently shown that the mid-day meal is the single most effective enrollment and attendance intervention in Indian education. Not better teaching. Not new textbooks. Not computer labs. Food. The provision of one cooked meal per day to children in Classes I through VIII has done more to get children through school doors than any pedagogical innovation of the past three decades.

Odisha understands this better than most states, and it has built what is arguably the most innovative implementation of the PM POSHAN (Pradhan Mantri Poshan Shakti Nirman) scheme in the country. The state’s model connects the mid-day meal program to the Mission Shakti Self-Help Group network --- over 6,500 women SHGs cook and serve meals across schools, earning Rs 38,000-55,000 per month per group. This creates a dual benefit: nutrition for children and livelihood for women. It has recorded the highest hygiene and satisfaction scores in national evaluations.

The meals themselves cost Rs 7.64 per child per day at the primary level and Rs 10.94 at the upper primary level. Let those numbers settle. The entire food budget for a child’s school meal is less than the price of a cup of tea at a railway station. And yet this Rs 7.64 meal is, for millions of children in Odisha’s poorer districts, the primary reason they attend school.

In April 2025, Odisha extended the scheme to Classes IX and X, one of the few states to do so. This extension is not just about nutrition. It is an implicit acknowledgment that the secondary-level dropout crisis --- 15% of students leaving before completing Class 10 --- has an economic stomach as much as an academic mind. Feed students through Class 10, and you have a better chance of keeping them there.

The meal program also reveals something uncomfortable about what education actually is in large parts of rural India. For policy planners, education is about curriculum delivery, learning outcomes, and skill formation. For a parent in Nuapada or Bolangir deciding whether to send their child to school or take them to a brick kiln in Telangana, the calculation is simpler: the school feeds my child, and the brick kiln does not (or not adequately). The mid-day meal converted education from an abstract future benefit (your child will learn to read, which may someday improve their earning potential) into an immediate, tangible benefit (your child will eat today).

Tamil Nadu understood this principle decades before anyone else. The Chief Minister’s Nutritious Noon Meal Scheme launched in 1982 --- thirteen years before the national mid-day meal scheme began in 1995. That thirteen-year head start compounded. Tamil Nadu’s dropout rate at the secondary level is below 5%. Odisha’s is 15%. The difference is not entirely explained by the noon meal, but the noon meal was the foundation on which Tamil Nadu built everything else: free bus passes, free uniforms, free textbooks, free health check-ups. Each intervention removed one more barrier between a child and a classroom. Odisha is following the same playbook but started later and has not yet achieved the same coverage density.

The quality concerns are real. Food inflation squeezes schools into purchasing cheaper ingredients. Central and state funding arrives late in the financial year, forcing schools to manage on credit. In many schools across eastern India, including Odisha, rising inflation has led to lighter, less nutritious meals --- no fruit, no milk, sometimes just rice and watery dal. The Rs 7.64 per meal was recently raised from Rs 5.90, but food prices have risen faster. The gap between what the scheme promises and what the cooking pot contains grows wider each year.

And there is a deeper question the meal program raises: when food becomes the primary draw, what does that say about the education itself? A school where the mid-day meal is the main attraction is a school that has failed at its primary purpose. The meal should be a support system for learning, not a substitute for it. In too many schools in Odisha’s poorer districts, the meal is the school. Children come, eat, and leave. The learning that is supposed to happen around the meal is the afterthought.

This is not a criticism of the meal program. It is a recognition that the program has succeeded at something the teaching system has not: giving families a reason to engage with the institution. The question is whether the institution can now deliver on the promise that enrollment implies.

The Infrastructure of Absence

There is a reason girls drop out of school at secondary level in Odisha at a rate of 12.5%, significantly higher than the national average of 9.6%. It is not an abstract reason. It has plumbing.

In 2023-24, 611 schools in Odisha lacked girls’ toilets. This number had increased from 514 the previous year. The situation was getting worse, not better. Of the schools that do have toilets, a significant proportion lack running water, making them effectively unusable. When a girl hits puberty --- typically around Class 6 or 7 --- and her school has no functional toilet, she has three options: not attend school during her menstrual period (losing 3-5 days per month), hold it all day (a health risk), or stop going to school entirely.

Many choose the third option. The dropout data confirms it: girls’ dropout at the secondary level concentrates in the years when menstruation begins. The KGBV (Kasturba Gandhi Balika Vidyalaya) residential schools for disadvantaged girls have installed incinerators for menstrual hygiene management. These 182 schools serve 18,400 girls. They are proof that the problem can be solved. They are also proof that the state has solved it for 18,400 girls while leaving hundreds of thousands to manage on their own.

But the toilet problem is just the most viscerally understandable symptom of an infrastructure crisis that spans every dimension of a school’s physical environment. UDISE+ 2023-24 data for Odisha reveals:

  • 2,182 schools lacked electricity. Of these, 1,672 were government schools. Try teaching a class in a concrete room with no lights during the monsoon, when the sky is dark by 10 AM. Try showing a digital lesson in a school that cannot power a screen.

  • 23,387 schools --- 38% of all schools --- lacked access to tap water. Students in these schools drink from borewells, hand pumps, or open wells without treatment. In a state where waterborne diseases are a leading cause of child illness, this is not an infrastructure gap. It is a health hazard operating under the name of an educational institution.

  • Only 24.9% of schools had personal computers or desktops, against a national average of 32.4%. The digital divide is not a future problem. It is a present reality in which three-quarters of Odisha’s schools cannot participate in digital education at all.

  • Only 27.5% of schools had a playground. Among the lowest figures nationally. Physical education, which the curriculum mandates, is physically impossible to deliver in three-quarters of schools.

The CAG audit of 2018-23 found that 12% of primary schools, 24% of upper primary schools, 42% of secondary schools, and 57% of higher secondary schools had adverse student-classroom ratios. In 43% of audited schools, students were found sitting on the floor due to lack of furniture. Not in 1990. In the period from 2018 to 2023.

Consider what this means for a child in a tribal block of Koraput or Malkangiri. She walks 2-3 kilometres to a school that may or may not have a teacher present. If the teacher is there, she sits on a floor in a room without electricity, shares a space with two or three other grade levels, drinks water from an untreated source, has no access to a computer, no playground, and --- if she is a girl approaching adolescence --- may not have access to a functional toilet. The curriculum she is taught bears no relation to her home language. The blackboard still carries yesterday’s lesson because there is no time to prepare today’s.

Now compare this with the experience of a child in Khordha district, attending a school that has been transformed under the 5T High School initiative or benefited from Mo School Abhiyan alumni funding. That child may sit in a well-lit classroom with furniture, have access to a smart classroom, use a toilet with running water, and be taught by a well-paid regular teacher with reasonable classroom ratios.

These two children live in the same state, governed by the same Chief Minister, funded by the same budget. They might as well live in different countries.

The infrastructure gap maps almost perfectly onto the three Odishas that the data reveals: coastal (high performance), western (middle, migration-affected), and tribal/KBK (lowest performance). The same geography that determines economic outcomes, health outcomes, and political outcomes determines educational infrastructure. The Long Arc documented how April 1, 1936 --- the day Odisha became a province --- created a state out of two systems: the relatively administered British coastal districts and the barely governed princely states of the interior. Ninety years later, the education system still runs on that same faultline.

The Dropout Cliff

Children do not drop out of school gradually. They fall off cliffs.

The data reveals three distinct transition points where Odisha’s education system hemorrhages students:

The Class 5-to-Class 6 transition. This is where primary school ends and upper primary begins. In tribal and rural areas, the upper primary school may be in a different village, several kilometres away. The nearby primary school was merged or closed under rationalization policy. The child who walked 1 km to primary school now faces a 3-5 km journey, potentially through forest, across a river, or along a road with no public transport. For girls, the distance barrier compounds with safety concerns and the onset of puberty. The dropout rate at the upper primary level is 3.2%, still low in absolute terms but representing the first major filter.

The Class 8-to-Class 9 transition. This is where the RTE’s guarantee of free and compulsory education ends. Class 8 is the last grade covered by the RTE Act. Class 9 begins the secondary stage --- a different school in many cases, with fees, new textbooks, examination pressure, and a more demanding curriculum. For families on the economic margin, the question becomes: can we afford to keep this child in school for two more years, or do we need them earning? For boys, this is often where the dadan migration system absorbs them --- sardars recruit adolescents for brick kilns and construction. For girls, this is where child marriage most commonly intervenes. The secondary dropout rate is where the system breaks: approximately 15% in 2024-25, with boys at 17.3% and girls at 12.5%.

The Class 10-to-Class 11 transition. Even among students who survive to Class 10 and pass the board exam (96% pass rate), nearly 30% never enter Class 11. The CAG audit found a secondary-to-higher-secondary transition rate of only 70.3%. This means that almost one in three students who complete ten years of schooling walk away at the threshold of the final two years. The reasons are primarily economic: higher secondary education involves additional costs, and the opportunity cost of two more years without income is unsupportable for many families.

The cumulative effect: between 1.50 lakh and 5.47 lakh children in Classes I through XI dropped out annually during the CAG audit period (2018-23). The secondary dropout rate surged by 86% over that five-year period, reaching 17.7% in 2022-23. And 61,487 children aged 6-18 remained completely outside the formal education system --- never enrolled, or enrolled and so thoroughly lost that the system no longer tracks them.

The CAG audit asked families why children dropped out. The answers were stark: unwillingness to continue studies (39%), poverty-related factors (27%), marriage (9%), other causes including migration, distance, and disability (25%). The 39% who cited “unwillingness” deserve scrutiny. A child who has spent five or six years in a school where she did not learn to read, where the teacher was frequently absent, where she sat on the floor in a room without electricity, is not “unwilling” to learn. She has learned something more fundamental: that school is not a place where learning happens. Her unwillingness is the system’s outcome, misattributed as her personal failing.

The gendered dropout pattern is particularly instructive. Historically, girls dropped out at higher rates than boys in India. In Odisha today, at the secondary level, the pattern has reversed: boys drop out at 17.3% versus 12.5% for girls. This reversal is driven not by girls’ improved retention (though some programs have helped) but by the economic pull that draws adolescent boys into the labor market. When a family in Bolangir is offered an advance by a sardar to supply labor at a brick kiln, the first to go is the adolescent son. The dadan system that The Leaving documented in detail is not just a migration problem. It is an education terminator.

But the reversal obscures a more complex reality. Girls who drop out, drop out into invisibility --- into domestic work, caregiving, and early marriage. Boys who drop out enter the labor economy, however exploitatively. The state can track boys who migrate; it cannot easily track girls who vanish into the domestic sphere. The gendered education cliff that Women’s Odisha documented --- where female literacy jumped from 2.5% at independence to 64% today but with a persistent 25% dropout rate at secondary level --- operates as a distinct mechanism, driven less by economic pull and more by social architecture: the in-law hierarchy, the marriage market, the expectation that education for a girl is a luxury expendable at the first sign of economic stress.

The Three Odishas in One Classroom

The district-level data does not just show a gap. It shows three different education systems operating under one administrative umbrella.

Coastal Odisha --- Khordha, Jagatsinghpur, Cuttack, Kendrapara, Puri --- functions at a level roughly comparable to India’s better-performing states. Khordha has 86.88% literacy, near-universal enrollment, adequate teacher availability, functional infrastructure, and the highest Class 10 pass rate (97.98% in 2024). These districts benefit from urbanization, educated parent populations, better connectivity, and multiple school options (including a growing private sector). A child in a Bhubaneswar government school may not receive a world-class education, but she receives a functional one.

Western Odisha --- Bolangir, Nuapada, Kalahandi, Bargarh, Sonepur --- sits in the middle but with a distinctive problem: seasonal migration. The dadan system removes entire families from these districts for six to eight months of the year. A school in Nuapada may have 50 students enrolled in October. By November, after the sardars have come with their advance payments, 15-20 of those students are gone. They will return in May, having missed more than half the academic year, unable to catch up, carrying learning gaps that widen each year until the child simply stops returning. No educational intervention --- no improved curriculum, no better-trained teacher, no digital classroom --- can solve the problem of a student who is physically absent for most of the year.

Tribal/KBK Odisha --- Koraput, Malkangiri, Rayagada, Nabarangpur, Gajapati, Kandhamal --- operates at educational levels that the data should make impossible to ignore. Nabarangpur’s literacy rate of 46.43% and Malkangiri’s tribal literacy of 35.23% are not figures from the mid-20th century. They are from the 2011 Census, the most recent available. In the fifteen years since, there has been improvement, but the base is so low that improvement from 35% to, say, 50% still leaves half the tribal population unable to read.

In these districts, the education challenge is qualitatively different from the coastal one. Children arrive in Class I speaking Kui, Sora, Mundari, Santali, or one of dozens of tribal languages. They encounter Odia for the first time as the medium of instruction. They face, in effect, a second-language immersion curriculum without the pedagogical infrastructure that second-language immersion requires. Then, as they progress, they encounter English as a third language, with 22.9% of government school Class V students able to read English sentences compared to 68% in private schools. This triple language barrier --- tribal mother tongue at home, Odia in the classroom, English as the aspired-to language of economic mobility --- is a challenge that the parallel civilisation of Tribal Odisha documents as a specific, recurring pattern: the state imposes its categories and languages on communities that operated with different ones, and then measures the communities’ failure to meet the state’s standards as evidence of their backwardness.

The school rationalization policy compounded the problem. Since 2013, approximately 10,000 schools have been closed or merged. In November 2023, 7,478 additional schools were marked for closure due to low enrollment. The logic was efficiency: a school with 15 students and one teacher is expensive per pupil. Merge it with a larger school 3 km away, and the per-pupil cost drops. The spreadsheet looks better. But for a Kondh girl in a remote habitation, the difference between a school 1 km away and a school 4 km away is the difference between going to school and not going to school. Over 25,000 tribal habitats lacked schooling facilities. The rationalization policy, designed to improve efficiency, may have manufactured dropouts in the communities that could least afford them.

The new BJP government announced in 2024 that closed schools would be reopened “based on necessity and official reports on ground reality.” The extent of actual reopening remains to be seen. Schools are easier to close than to reopen --- the teacher has been transferred, the building has deteriorated, the community has adjusted to the school’s absence.

The Comparators: What Others Built

There is a question that the data forces: is this just how things are in a poor state? Is Odisha’s educational underperformance simply the inevitable consequence of low income, high poverty, and developmental lag?

Kerala answers the question with a definitive no.

IndicatorKeralaOdisha
Literacy Rate94%73%
Female Literacy92%64%
Dropout Rate (Secondary)<2%~15%
Schools with Playground>60%27.5%

The gap is enormous. But the gap is not permanent. It is the product of specific historical choices, made over specific timelines.

Kerala’s education advantage began in 1817, when Travancore’s Queen Regent issued a royal decree mandating public education --- over a century before Indian independence. The Sree Narayana Guru movement in the late 19th century made education a weapon against caste oppression, creating mass social demand for schooling. The EMS Namboodiripad government in 1957 implemented land reform that broke feudal control, freeing lower-caste families to invest in education. By the time India achieved independence, Travancore already had literacy rates comparable to European countries.

Odisha only became a province on April 1, 1936. That is an 119-year head start for Kerala’s educational project. Education compounding works across generations: literate parents create educated children who become literate parents who create more educated children. Each generation inherits not just its parents’ knowledge but their attitude toward knowledge. In Khordha, this generational compounding has been running for perhaps three generations. In Malkangiri’s tribal communities, it has been running for less than one.

Tamil Nadu tells a different story but arrives at the same lesson. Tamil Nadu launched its noon meal scheme in 1982, thirteen years before the national program. It placed a primary school within 1 km of 99% of habitations. It invested in teacher quality and accountability. It achieved political consensus across party lines: education was never a partisan battleground between the DMK and AIADMK. Both parties competed on who could deliver better educational outcomes. The result: secondary dropout below 5%, literacy at 80%, and a workforce that could support the auto-cluster industrialization that made Tamil Nadu one of India’s wealthiest states.

Himachal Pradesh achieved near-universal primary enrollment despite mountain terrain far more challenging than Odisha’s. It did so through teacher accountability, effective School Management Committees, and community ownership of education outcomes. Its Pupil-Teacher Ratios are among the most comfortable in India: 13:8:6:9 across primary, upper primary, secondary, and higher secondary levels.

Here is what these comparators share and Odisha lacks:

Sustained investment over decades, not policy cycles. Tamil Nadu’s noon meal started in 1982. Kerala’s education project started in the early 19th century. Odisha’s serious education push is approximately two decades old.

Universal physical access. Schools within walking distance of every habitation, no exceptions. Odisha’s 25,000+ tribal habitats without schooling facilities are the antithesis of this.

Teacher quality as a priority, not an afterthought. These states invested more than 1% of their education budgets in teacher training.

Social demand for education. In Kerala, education is a cultural value, not just a government program. In Odisha, the first-generation-educated phenomenon means education culture is still being constructed in many communities.

Political consensus. Education quality transcended partisanship. In Odisha, education policy shifts with governments --- Mo School Abhiyan becomes Panchasakha Sikhya Setu when the administration changes. The program may continue under a new name, but the signal to the bureaucracy is that education policy is disposable, subject to the next election.

The peer comparisons with Bihar, Jharkhand, and Chhattisgarh are perhaps more illuminating. Bihar, with 62% literacy and schools operating under tents, makes Odisha look better by comparison. Jharkhand mirrors Odisha’s tribal-mineral-education nexus exactly: districts with mineral wealth should have the resources for excellent schools via DMF funds, but instead show high dropout and poor learning. Chhattisgarh’s private school enrollment has reached 27.1% --- nearly double Odisha’s 16.1% --- reflecting faster erosion of faith in government schools.

Odisha is not yet Chhattisgarh. Its 82% government school share and 88% enrollment in government schools at ages 6-14 mean the public system still holds most of the student population. This is a window, not a permanent condition. If government school quality does not improve, the migration to private schools will accelerate, and the government system will become a school of last resort for families with no other option --- a trajectory every education researcher calls the “death spiral” of public schooling.

The ASER 2024 recovery data suggests the window is still open. The question is whether the system can sustain improvement or whether 2024’s gains will dissipate as attention shifts elsewhere.

The Spending Paradox

The obvious response to Odisha’s education crisis is: spend more money.

The data says the problem is more disturbing than insufficient spending. It is the inability to spend what has already been allocated.

Odisha’s education budget for 2026-27 is Rs 42,565 crore --- 13.7% of the total state budget. The School and Mass Education Department alone receives Rs 31,997.53 crore. This is not a trivial sum. At approximately 3.8% of GSDP, it is below the NEP 2020 recommendation of 6%, but comparable to what Kerala and Himachal Pradesh allocated when they were at similar stages of educational development.

The CAG audit of 2018-23 found that the Samagra Shiksha scheme --- the primary centrally-sponsored vehicle for school education --- had a utilization rate of only 44-50%. Roughly half the money allocated for education was not spent. Both central and state contributions were released “in considerable portions towards the end of the financial year,” making it impossible to spend them meaningfully before the year closed. The School and Mass Education department surrendered Rs 1,159.31 crore in 2019-20 alone. Over five years, the department surrendered more than Rs 11,000 crore in unspent funds.

Eleven thousand crore rupees. Returned to the treasury. While 2,182 schools operated without electricity. While 23,387 schools lacked tap water. While 611 schools had no girls’ toilets. While 25,000 tribal habitats had no school at all.

This is not frugality. It is institutional incapacity --- the inability of the bureaucratic machinery to convert budgetary allocation into physical outcomes. The reasons are familiar to anyone who has studied Indian state governance: late fund release, complex procurement rules, shortage of implementing officers at the block level, lack of coordination between departments, and the absence of accountability for non-spending. No officer is punished for failing to spend education funds. Many are punished for spending them “incorrectly.”

But the composition of spending is perhaps even more troubling than the volume. The vast majority of education spending in Odisha goes to teacher salaries and infrastructure construction. The share allocated to teacher training? Less than or equal to 1%. The share dedicated to quality improvement --- learning materials, assessment systems, remedial support, pedagogical innovation? Somewhere between 1% and 3%.

In software terms, this is like spending 95% of the development budget on hardware and office space, 1% on code review and testing, and then wondering why the software crashes. The system has invested in the physical container (school buildings, teacher salaries) while starving the actual content delivery (what teachers know, how they teach, what children learn). Buildings do not educate children. Teachers do. And untrained teachers in well-built schools produce the same outcome as untrained teachers in poorly-built ones: children who pass through the system without learning.

This is the spending paradox: Odisha allocates money, does not spend it, and the money it does spend goes to the wrong things. The system is underfunded (relative to NEP recommendations), under-spent (relative to its own allocations), and mis-spent (relative to what actually improves learning). It is three failures stacked on top of each other, and each one makes the others harder to fix.

The Reforms: How Far They Reach

Odisha is not ignorant of these problems. Multiple reform programs have launched in the past decade, each targeting a specific failure point. The question is not whether reforms exist. It is whether they reach the schools where the crisis is deepest.

Mo School Abhiyan / Panchasakha Sikhya Setu connects alumni to their former government schools. Alumni contribute funds, the government matches at 2x, and CSR adds more. Total committed: approximately Rs 797 crore across 40,855 schools. Thirty lakh alumni and community members have been connected to the platform.

The model is genuinely innovative. No other state has mobilized this scale of private funding for government school improvement. But it contains a structural bias: it favors schools in urban and semi-urban areas where alumni are prosperous, connected, and emotionally invested. The government school in Bhubaneswar whose alumni include software engineers in Bangalore and bureaucrats in Delhi will receive generous contributions. The school in a Malkangiri tribal hamlet whose alumni --- if they can even be called that, since many did not complete schooling --- work as migrant laborers earning daily wages will receive nothing. The program may inadvertently widen the coastal-tribal divide it should be narrowing.

Kasturba Gandhi Balika Vidyalaya (KGBV) provides residential schooling for girls from SC, ST, OBC, minority, and BPL families in educationally backward blocks. The 182 KGBVs across 23 districts serve 18,400 girls with free textbooks, uniforms, stipends, menstrual hygiene facilities, CCTV security, and self-defense training. The program is well-designed, well-targeted, and hopelessly limited in scale. At 18,400 girls in a state where the secondary dropout rate for girls is 12.5% and where NCRB records over 8,100 child marriage cases, the KGBV is a life raft for the few, not a solution for the many.

The Pathani Samanta Mathematics Talent Scholarship awards Rs 5,000 per month to 1,000 students studying in Odia medium higher secondary schools. At 1,000 scholarships from a pool of approximately 5 lakh Class 10 graduates, it rewards the top 0.2%. It is a scheme for the exceptional. The fundamental problem --- 75% of Class 5 students unable to divide --- is a problem of the ordinary. No scholarship scheme for the talented few can substitute for foundational learning for the many.

The Utkarsh Programme, a partnership between Transform Schools (an NGO) and the state government, targets Class 9 students in secondary schools with 200 hours of intensive foundational skill development over 69 school days. This is perhaps the most diagnostically honest intervention: it acknowledges that students arriving in Class 9 lack skills they should have acquired in Class 3, and it attempts to remediate the gap. But 200 hours cannot fully repair six years of accumulated learning deficit. The program is a bandage on a wound that began bleeding in Class 1.

The pattern across all these reforms is consistent: each one is well-conceived, each one addresses a real problem, and each one is insufficient in scale or scope to alter the system’s fundamental trajectory. They are point interventions in a system that requires structural transformation. They treat symptoms without addressing the root cause: a primary education system that does not teach.

Garbage In, Garbage Out: The System Architecture

Step back from the individual data points and look at the system as a whole. What is actually happening here?

Odisha operates an education pipeline that processes approximately 76 lakh students through 61,565 schools across 30 districts. The pipeline has clearly defined stages: primary (I-V), upper primary (VI-VIII), secondary (IX-X), and higher secondary (XI-XII). At each stage, a set of inputs (teachers, infrastructure, curriculum, time) is supposed to transform raw material (children’s minds) into outputs (educated, capable young adults).

The system measures itself at the output stage. Board exam pass rates --- 96% at Class 10, 81-87% at Class 12 --- suggest the pipeline is working well. Nearly every student who enters the final examination emerges with a passing certificate.

But the pipeline has a critical flaw. The input stage --- primary education --- is corrupted. Children enter Class 1 and emerge from Class 5 without foundational reading or arithmetic skills. This corrupted data then flows through every subsequent stage. Upper primary teachers cannot teach Class 6 material to students who cannot read. Secondary teachers cannot teach Class 9 concepts to students who cannot do Class 3 arithmetic. The system compensates not by fixing the input corruption but by lowering standards at each successive stage --- passing students forward, marking liberally, adjusting difficulty downward.

By the time a student reaches Class 10, she has been “processed” for ten years. The board exam certifies that processing occurred. It does not certify that learning occurred. The output --- a Class 10 pass certificate --- looks like clean data. It is, in fact, corrupted data with a clean label.

This is the classic garbage-in, garbage-out problem applied to human capital. And like its software counterpart, it has a specific architectural implication: the fix must happen at the input layer. No amount of secondary school reform, higher education investment, or skill development programming can fully compensate for foundational learning deficits created at the primary level. You cannot build a second floor on a cracked foundation by making the second floor stronger. You have to repair the foundation.

The research from high-performing states confirms this. Kerala’s advantage is not its universities (which are middling by national standards). It is its primary schools, where nearly every child learns to read, write, and compute by Class 5. Tamil Nadu’s advantage is not IIT Madras (which is a central institution). It is the noon meal that got children into primary school and the teacher quality that ensured they learned something once they arrived. Himachal Pradesh’s advantage is not its higher education sector. It is its commitment to universal, quality primary education despite mountain terrain.

Every state that solved the education problem solved it at the primary level. Every state that failed to solve it --- Bihar, Jharkhand, and yes, Odisha --- failed at the primary level and attempted to compensate further downstream.

The ASER 2024 data showing recovery in early-grade reading is, from this architectural perspective, the most important data point in years. If Odisha can genuinely improve foundational learning at Class III --- if the 10 percentage point improvement is sustained and deepened --- the effects will ripple through the entire pipeline. Better Class 3 readers become better Class 5 mathematicians who become better Class 8 science students who become genuinely prepared Class 10 examinees. The compound effect of fixing the input layer takes years to manifest, but when it does, it transforms every downstream outcome.

But if the 2024 improvement is a blip --- a temporary focus that dissipates as political attention shifts --- then the system will revert to its steady state: high enrollment, low learning, inflated pass rates, and generation after generation of young people who spent ten years in school and emerged without the skills the economy demands or the education their certificates claim.

What Compounds and What Corrodes

Education compounds. This is both its promise and its cruelty.

The promise: a child who learns to read in Class 2 can learn history in Class 6, analyze texts in Class 9, and engage with complex ideas in college. Each year of learning builds on the previous year. The returns are non-linear. A literate parent raises children who arrive at school already familiar with books, letters, and the idea that printed words carry meaning. Those children learn faster, creating a positive feedback loop that strengthens across generations. This is how Kerala got to 94% literacy: not through a single policy intervention but through 200 years of compounding.

The cruelty: the same compounding mechanism works in reverse. A child who does not learn to read in Class 2 falls behind in Class 3, further behind in Class 4, and is effectively lost by Class 5. The gap between her and her reading peers widens each year. By secondary school, she is sitting in a classroom where the teacher assumes foundational skills she never acquired, covering material she cannot access, in a language she may not fully command. Her “unwillingness to continue studies” --- the reason 39% of dropouts cite --- is not a character flaw. It is the rational response to an irrational situation: why sit in a room where nothing being said makes sense?

And the corrosion extends beyond the individual. A community where most adults are functionally illiterate cannot support a school. Parents who cannot read cannot help children with homework. They cannot evaluate whether the school is teaching effectively. They cannot engage with School Management Committees (which the RTE mandates) in any meaningful way. The CAG audit found that participatory planning was not followed in Odisha’s schools. This is unsurprising: how do you have participatory planning with parents who cannot participate because the system never educated them?

The first-generation learning challenge --- which Tribal Odisha faces at its starkest --- is essentially a cold-start problem in computing. A system that requires pre-existing data to function well has no data to work with on its first run. The primary school in Malkangiri receives children who have never seen a book, whose parents have never written a letter, who have no model of what “education” means beyond the physical act of going to a building. The school must not only teach content; it must teach the very idea that content can be taught and learned. This requires fundamentally different pedagogy from teaching children who arrive already knowing their letters.

The 30 DIETs and 31 ETEIs across Odisha do not train teachers for this challenge. They train teachers for a generalized classroom that assumes basic home preparation. The mismatch between what tribal primary schools need and what teacher training institutions produce is the deepest structural failure in the system --- deeper than funding, deeper than infrastructure, deeper than teacher vacancies. It is a design failure: the system was designed for a student population that does not exist in the districts where the system is needed most.

The Data Quality Problem

Here is one way to understand everything discussed in this chapter.

A state’s education system is an information processing machine. It takes in raw inputs (children with varying levels of home preparation, language ability, nutrition, and cognitive readiness), processes them through a series of stages (primary, upper primary, secondary, higher secondary), and produces outputs (literate, numerate young adults with certified qualifications).

The quality of the output depends on three things: the quality of the inputs, the quality of the processing, and the integrity of the measurement system that evaluates the output.

In Odisha, all three are compromised.

Input quality is uneven to the point of bimodal distribution. A child arriving in Class 1 in Khordha has literate parents, Odia as her mother tongue, prior exposure to an Anganwadi center, adequate nutrition, and a school within walking distance with electricity, water, and furniture. A child arriving in Class 1 in Nabarangpur has illiterate parents, a tribal language as her mother tongue, inconsistent Anganwadi exposure, chronic nutritional deficits, and a school that may lack a teacher, a toilet, and a ceiling that does not leak. These are not two points on a spectrum. They are two different starting conditions. The system treats them identically.

Processing quality is degraded at the foundational stage. Teachers are undertrained (1% of budget on training), frequently absent (20-25% nationally), often managing multiple grades alone, and sometimes earning exploitative wages as contract workers. The curriculum assumes home preparation that does not exist. The pedagogy assumes linguistic homogeneity that tribal classrooms do not have. The result: processing corrupts the data rather than refining it.

The measurement system validates rather than verifies. Board exams with 96% pass rates certify that students completed the process. ASER assessments reveal that the process did not produce learning. The system’s own measurement tools --- the ones it uses to evaluate itself --- are designed to produce reassuring numbers. The external measurement tools --- ASER, NAS --- produce alarming ones. When a system’s internal metrics and external metrics diverge this sharply, the internal metrics are lying.

A system with corrupted inputs, degraded processing, and dishonest measurement is a system that produces garbage while reporting gold. It is the education equivalent of a factory that ships defective products with quality-certified labels. The products look fine on the shelf. They fail when the customer tries to use them.

The “customer” in this case is the economy. When 75% of Class 5 students cannot divide, the workforce these students eventually join will be computationally illiterate. When only 22.9% of Class 5 government school students can read English sentences, the service sector --- which increasingly demands English --- will draw talent from the 10-16% who attended private school, leaving the 84-90% who attended government school locked out of the fastest-growing employment sectors. The confidence level here is high, perhaps 80-85%: the primary education system’s quality failure is a binding constraint on Odisha’s economic transformation, more fundamental than infrastructure deficits, investment climate, or industrial policy. You can build factories. You cannot staff them with workers who were never taught to learn.

This claim should be tested. It would be wrong if Odisha’s economy could grow rapidly on the basis of natural resources, infrastructure investment, and migration-sourced remittances without requiring a skilled local workforce. That scenario is possible but historically rare. Every economy that has sustained rapid growth --- South Korea, Taiwan, China, the Indian states of Tamil Nadu and Karnataka --- did so on the foundation of mass education. The counterexamples (Gulf states, some resource exporters) required imported labor precisely because their domestic education systems could not produce it.

The Window

Odisha’s education story is not yet determined. The system is failing, but it is failing in ways that are diagnosable, and some of the remedies are visible in the data.

The ASER 2024 recovery is genuine. Odisha’s government schools drove the improvement. This means the public system --- which serves 82-88% of students --- retains the capacity to improve when focused interventions are applied. The base is low, but the direction is upward.

The PM POSHAN mid-day meal program, particularly the SHG-led implementation model, is nationally recognized. It demonstrates that Odisha can build innovative delivery mechanisms when the political will exists. The program’s extension to Classes IX and X is a structural bet that nutrition can reduce secondary dropout. The bet is probably correct, though its effect will take years to manifest.

The pre-primary coverage is near-universal: over 95% of 3-year-olds enrolled, over 75% of children aged 3-4 in Anganwadi centers. If this early childhood exposure translates into better school readiness, the input quality at Class 1 will gradually improve.

But the window will not remain open indefinitely. Three forces are closing it:

The private school migration. At 10-16% private enrollment (ages 6-14), Odisha’s government school system still holds most students. But the national trend is unmistakable: government school enrollment nationally fell from 72.9% in 2022 to 66.8% in 2024. If Odisha follows this trajectory, the government system will lose its most motivated families first --- the ones with enough awareness and income to choose private school. The remaining government school population will become progressively more disadvantaged, creating the death spiral observed in UP and Rajasthan.

The English language barrier. As India’s economy becomes increasingly service-oriented and digitally driven, English proficiency becomes a gateway skill. With only 22.9% of Class V government school students able to read English sentences, the government school system is producing graduates who are locked out of the economy’s fastest-growing sectors. The longer this gap persists, the harder it becomes to close. This matters especially because language proficiency, unlike many other skills, compounds with early exposure. A child who begins learning English in Class 1 has a fundamentally different trajectory from one who begins in Class 6.

The demographic window. Odisha’s school-age population will not grow indefinitely. The window during which the state has a large youth cohort --- the demographic dividend --- is finite. If this cohort passes through the education system without acquiring foundational skills, the dividend becomes a liability: a large young population without the human capital to contribute productively to the economy.

The fix, as the architecture of the problem suggests, must begin at the primary level. Not exclusively --- secondary and higher secondary reforms matter too --- but the highest-leverage intervention is ensuring that every child who completes Class 5 can read, write, and compute at grade level. This is not a revolutionary insight. It is the oldest finding in education research: foundational skills determine everything that follows.

The Kashipur Government Primary School where this chapter began is where the churning starts. The single teacher managing five classes on Rs 7,500 a month, the children who came for the meal and may or may not encounter a lesson, the blackboard with yesterday’s algebra still smudged on its surface --- this is the input layer. This is where the data enters the system. This is where the corruption begins.

Fix this, and the system’s output transforms across a generation.

Fail to fix this, and every downstream investment --- the universities, the skill programs, the IT parks, the industrial corridors --- processes garbage data and produces certificates that look like gold and perform like lead.

The first classroom is the only classroom that matters.

Sources

UDISE+ Data:

ASER Reports:

NAS Data:

Board Examination Results:

CAG Audit:

Infrastructure Data:

Dropout and Transition:

Mid-Day Meal / PM POSHAN:

Migration and Dadan System:

Teacher Data:

School Programs:

School Rationalization:

Comparative Analysis:

Budget and Finance:

District Literacy Data:

RTE Implementation:

Academic Studies:

Source Research

The raw research that informs this series.