Who Grades the Grader

Pritiraj BrahmaFounding Partner, Pritiraj & Partners · 25 August 2026

On 1 February 2025 the CBI arrested the chairman and six members of a NAAC inspection committee. The FIR describes ₹1.8 crore sought for a score above 3.65 and an A++ grade. Nearly 900 assessors were then removed. The response treated the assessors as the fault. The fault is in a design that makes a grade worth buying and cheap to sell.

ProvenancePublished 25 August 2026. Accreditation coverage figures are from a Lok Sabha reply of 13 February 2023, against the institutional counts then current. Cycle-wise grade distribution is from the NAAC's own published report, current to 14 August 2025 and reported in October 2025. Criminal-case details are as stated in the CBI's public statement of 1-2 February 2025 and the FIR as reported; the matter is before the courts and no finding of guilt is implied.

The Central Bureau of Investigation stated on 1 February 2025 that it had arrested the chairman and six members of a National Assessment and Accreditation Council inspection committee, along with the vice-chancellor and two executives of the institution they were assessing, after searches at twenty locations recovered ₹37 lakh in cash, six laptops and a mobile phone. The FIR, as reported, records an initial demand of ₹1.8 crore for a score above 3.65 and an A++ grade: ₹20 lakh for the person drafting the report, ₹3 lakh and a laptop for each of six inspectors. Within a month the NAAC had removed roughly 900 assessors. The charges are before the courts and no one is guilty until a court says so. But the design question does not wait for the verdict.

The useful question after a case like this is never "who was corrupt". It is "what made the transaction rational for both parties". A grade was worth ₹1.8 crore to one side and, allegedly, saleable for a fraction of that by the other. Both halves of that sentence are facts about institutional design, and both are fixable without a single prosecution.

Five findings anchor this analysis:

1. Most of the system has never been graded at all. Answering the Lok Sabha on 13 February 2023, the Minister of State for Education stated that of 1,113 universities and 43,796 colleges, the NAAC had accredited 418 universities and 9,062 colleges; 34,734 colleges were operating without accreditation, along with 695 universities. A quality signal that covers roughly a fifth of institutions is not a quality signal. It is a distinction held by the institutions that chose to seek it.
2. Among those graded, the distribution is compressed and the grade carries little discriminating power. The NAAC's own report, current to 14 August 2025 and reported in October 2025, records 11,829 institutions accredited in their first cycle: 2,533 with grade A, 7,850 with B and 1,446 with C. Among universities the concentration is at the top — 301 of 561 universities hold an A. Among colleges it is at the middle — of 11,268, some 7,610 hold a B. When two-thirds of a graded population sits in one band, the grade tells an applicant, an employer or a funder very little, while remaining valuable enough to be worth a crore.
3. The integrity failure reached the panel, not the periphery. The arrests of February 2025 involved the chairman of an inspection committee and six of its members, including a serving professor of a central university acting as member coordinator, along with a former NAAC deputy adviser and a serving adviser named in the FIR. This was not a clerk. It was the assessment instrument itself.
4. The institutional response was a purge, not a redesign. Around 900 assessors were removed in the weeks following, on grounds reported to include inactivity, refusal of assignments, inadequate reports, and scores awarded without justification or on incomplete data. Removing 900 people who behaved badly inside a system is necessary. It does not alter the conditions that produced the behaviour, and the number itself is an indictment of the panel-selection process that appointed them.
5. The trainer layer that is meant to raise quality is itself assessed poorly. In a separate NAAC review of the Malaviya Mission Teacher Training Centres — the former UGC Academic Staff Colleges, of which there are 116 — 66 were assessed, of which 7 were labelled "non-performer" and 46, about 70 per cent, "under performer". The institutions charged with improving teaching quality are, on the accreditation body's own assessment, largely under-performing.

The design that made it rational

Four features of the accreditation architecture, each defensible on its own, combine into a system that invites exactly what happened.

It is voluntary. An institution chooses whether to be assessed. The consequence of never seeking accreditation is, for most colleges, nothing at all — 34,734 of them demonstrate this daily. So the population being graded is self-selected, and the institutions with most to hide are the ones least likely to appear.

It is episodic. Accreditation is valid for five years and turns on a peer team's visit. A judgment made in a few days binds for half a decade, which concentrates enormous value into a single, short, human interaction between a small group of visitors and an institution with a great deal at stake.

It is stated to be non-punitive. The NAAC's published philosophy describes assessment as being for continuous improvement "rather than being punitive or judgmental". As an educational sentiment this is admirable. As a regulatory posture it removes the downside of a poor grade, leaving only the upside of a good one — and a one-sided incentive is the definition of an inflationary one.

Its output is a letter, not a record. What reaches the public is A++, A, B or C. The peer team report — the evidence, the observations, the reasoning — is not the thing the market reads. A grade can be bought. A published body of evidence has to be manufactured, and manufacturing is harder, slower and more visible than negotiation.

The counter-case

The NAAC's defenders make three arguments, and the first is genuinely strong.

Participation has widened substantially. The 11,829 institutions accredited in cycle one is more than double the 5,505 in cycle two, which reflects both new institutions and existing ones entering the process for the first time. The Council reduced its fees and shortened the self-study manual for affiliated and constituent colleges specifically to bring smaller institutions in, and the numbers suggest that worked. An accreditation body that had made itself unaffordable would have covered even less of the system.

Second, reform was already in motion before the arrests. The Radhakrishnan Committee report on reforms in accreditation and ranking predates the case, and the Council has moved towards a binary and hybrid model with virtual verification of documentary claims — a change that directly attacks the visit-centred vulnerability the case exposed.

Third, in fairness, the case became public because it was detected and prosecuted. A system in which nothing is ever caught is not cleaner.

Each is true. None reaches the structural point. Wider participation in a compressed, voluntary, letter-grade system produces more institutions holding a signal that means less. Binary accreditation removes the incentive to buy the difference between A and A++, which is real progress, but it also removes the information that distinguishes them — unless what is published alongside the binary decision is the evidence itself.

What we would do

1. Publish the peer team report in full, with the grade. Not a summary, not a score sheet — the report. An institution that has genuinely built laboratories, libraries and faculty strength loses nothing by having that written down and read. An institution that has not, loses the ability to convert a private negotiation into a public letter. This single change does more against grade-buying than any number of assessor removals.
2. Separate verification of fact from judgment of quality. Faculty counts, student-teacher ratios, laboratory holdings, placement records and audited finances are verifiable against records the state already holds — AISHE returns, EPFO data, affiliating-university records, audited accounts. Verify them centrally, before any visit. Let the peer team judge only what requires judgment: teaching, research culture, governance. A visiting panel should never be the body that certifies a number that a database could confirm.
3. Publish the assessor roster, the assignment rule and the conflict declarations. Who is on the panel, how they were selected for this institution, and what interests they declared, published at the time of assignment rather than after an arrest. Random allocation from a large published pool, with declared conflicts excluded automatically, is standard practice in audit and is not difficult here.
4. Make it accreditation or disclosure, with a date certain. Voluntary assessment cannot govern a system in which four-fifths of institutions abstain. An institution that has not been assessed by a notified date should be required to say so, plainly, wherever it advertises admission. That is not a penalty; it is the minimum owed to a family committing four years and several lakh rupees.
5. Keep a public register of appeals, revisions and revocations. Every grade changed, every accreditation withdrawn, every complaint upheld — listed, dated and reasoned. A quality regulator with no visible record of ever having withdrawn anything is telling the market that its judgments do not move, which is the same as telling institutions that the grade, once obtained, is permanent property.

The arrests of February 2025 were treated, understandably, as a scandal about individuals. They were better read as an audit finding. A grade that is worth ₹1.8 crore, awarded in a short visit, by a panel whose composition is not published, to a self-selected fifth of the system, is not an accountability mechanism. It is an asset. India does not need harsher assessors. It needs to stop issuing an asset and start publishing a record.