What an Indian degree is actually worth

Reconciles NIRF’s published placement and salary disclosures against MoSPI PLFS survey microdata. Both sources are public; every figure is an aggregate. No Avanti student data and no personally identifiable information.

📄 Download the PDF · 📊 The workbook, every college · 📁 All data and charts

Built on public data: NIRF’s college scorecards, the government’s own labour force survey, and AISHE’s degree records. August 2026.

Every year about 7.78 million Indians finish an undergraduate degree.

In the first years after they leave, one in ten holds a formal job — regular salaried work with a contract or social security. Seven in a thousand hold one that pays ₹50,000 a month.

That is 55,108 good jobs, across the whole country, for a graduating cohort of 7.78 million.

India’s ranked colleges say they produced 83,764 of them.

The two numbers, by field

Measured over the three years from each course’s normal completion age, which is when a college’s placement figure is supposed to come true.

	Finish each year	In formal work		Earning ≥₹6 lakh
Engineering	784,517	271,802	34.6%	34,689	4.42%
Doctors (MBBS, BDS)	85,934	22,869	26.6%	8,586	9.99%
Everything else	6,905,549	521,569	7.6%	11,833	0.17%
All graduates	7,776,000	816,240	10.5%	55,108	0.71%

Everything else is arts, science, commerce, law, management, agriculture, nursing, pharmacy, architecture — 89% of everyone who graduates. Fewer than two in a thousand of them will be earning ₹50,000 a month in their first years out.

A further 9.3% of all graduates are earning informally — self-employed, or salaried with neither a contract nor benefits. That is real work, and it is counted, separately, because it is not what a campus placement is.

Placement reconciles. Salary does not.

NIRF’s ranked colleges file two things: how many students they placed, and what those students were paid. The first survives contact with the national data. The second does not.

Field		Ranked colleges claim	Exists nationally	Room left
Engineering	placed	159,731	271,802	112,071	0.59×
Engineering	≥₹6 lakh	62,612	34,689	−27,924	1.80×
Doctors	placed	7,443	22,869	15,426	0.33×
Doctors	≥₹6 lakh	3,748	8,586	4,838	0.44×
Everything else	placed	82,767	521,569	438,802	0.16×
Everything else	≥₹6 lakh	17,404	11,833	−5,571	1.47×

263 ranked engineering colleges claim 62,612 graduates earning ₹6 lakh or more. The country contains about 34,689 such jobs for engineering graduates at that stage. The claim is 1.8 times the national total — and that is after we have already cut every reported salary by 20%.

Their placement claims, by contrast, fit comfortably: 159,731 against a national 271,802, leaving 112,071 formal jobs for the 2,594 colleges NIRF does not rank. Nothing is impossible there.

This is exactly the shape of our own haircut decision, arrived at independently: discount NIRF’s salaries, don’t discount its placement rates. The gap is in what colleges say students are paid, not in how many got hired.

The over-claim is in the tail of the ranking, not the top

It would be easy, and wrong, to read the above as “NIRF is inflated”. Split the ranked engineering colleges into the audited public institutions and everyone else:

	Colleges	UG graduates	Jobs ≥₹6 lakh	Filed median
IIT	22	13,536	9,844	₹17.52L
NIT	28	19,128	10,991	₹10.28L
IIIT	5	1,122	788	₹11.00L
The public elite	55	33,786	21,623
Every other ranked college	208	191,227	40,990	₹5.00L

The IITs, NITs and IIITs account for 62% of every entry-level ₹6 lakh job in the country, from 4% of engineering graduates. They fit inside the national pool with room to spare for BITS and the strongest private colleges. Their placement reports are public and audited, and nothing about them strains belief.

The other 208 ranked colleges claim a further 40,990 — more, by themselves, than the entire national pool. That is where the arithmetic breaks.

The top of the ranking is credible. Its tail is not.

The exams

Sixteen lakh students sit JEE Main each year. Twenty-three lakh sit NEET. This is where they land.

JEE Main

Ends up	People	Share	In a job	≥₹6L	Placed	Median	Expected	Odds
Elite engineering	105,054	6.6%	81,047	52,966	77%	₹8.40L	₹6.48L	50.4%
Ranked, not elite	119,959	7.5%	78,684	9,647	66%	₹3.46L	₹2.27L	8.0%
Unranked engineering	559,504	35.0%	112,071	—	20%	—	—	—
No engineering seat	815,483	51.0%	61,593	1,397	8%	₹2.04L	₹0.15L	0.2%

NEET UG

Ends up	People	Share	In a job	≥₹6L	Placed	Median	Expected	Odds
Elite medical	6,257	0.3%	4,636	3,166	74%	₹8.48L	₹6.28L	50.6%
Ranked, not elite	6,426	0.3%	2,807	582	44%	₹3.84L	₹1.68L	9.1%
Unranked medical	73,251	3.2%	15,426	4,838	21%	₹4.82L	₹1.02L	6.6%
No medical seat	2,214,066	96.3%	167,226	3,794	8%	₹2.04L	₹0.15L	0.2%

Expected earnings is the median salary multiplied by the chance of being placed at all — what a student faces on the day they enrol, before they know whether they will be one of the lucky ones.

Half of everyone at an elite engineering college clears ₹6 lakh. One in twelve at the ranked colleges below them. Half the students who sit JEE Main get no engineering seat at all, and land on odds of one in five hundred.

The unranked engineering row is left blank above because the arithmetic that fills it is the subject of the section before this one: once ranked colleges have claimed 62,612 of a national 34,689, there is nothing left to allocate.

What everyone else does

Formal work is a minority outcome. Here is what the rest are doing, and what it pays a month.

	Formal	Informal salaried	Self-employed
Engineering	₹30,000 (n=304)	₹15,000 (n=34)	₹20,000 (n=24)
Doctors	₹40,000 (n=24)	₹21,700 (n=4)	₹25,000 (n=9)
Other health	₹15,180 (n=37)	₹12,000 (n=17)	₹17,000 (n=3)
Everything else	₹17,500 (n=1,057)	₹10,000 (n=935)	₹10,300 (n=695)

Informal work pays about half what formal work pays, in every field. Doctors are the one group whose informal work is decently paid, and even there it is a handful of people doing locum shifts, not private practice — at 24 to 26, 63% of them are still studying for postgraduate entrance.

Among engineering graduates who do reach formal work, 58% are in computer programming and consultancy at a median of ₹32,600 a month. That is the mass-recruiter band, and it is the single most important fact about Indian engineering employment: the industry that hires most engineers pays them ₹3.9 lakh, not ₹6 lakh.

---

Supporting analysis

The full method, every decision and its reasoning, and the bugs found along the way are in METHOD.md. The short version follows.

Where the numbers come from

Number	Source	Year
7,776,000 UG graduates a year	AISHE	2023-24
784,517 engineering graduates	AISHE, Engineering & Technology	2022-23
85,934 doctors (65,836 MBBS + 20,098 BDS)	AISHE, by programme	2023-24
2,857 engineering colleges	AICTE	2022-23
780 medical + 323 dental colleges	NMC seat matrix, DCI	2024-25
Placements and salaries at ranked colleges	NIRF scorecards	mostly 2025
Every rate — employment, earnings	PLFS microdata	CY2025
16 lakh JEE and 23 lakh NEET applicants	NTA, as supplied	2025-26

Counts come from AISHE, rates from PLFS, claims from NIRF. No source is used for what it is bad at.

Why the small national pool is believable

₹6 lakh a year is ₹50,000 a month. Among engineering graduates aged 22–24 who are in salaried work, that sits at the 90th percentile — the median is ₹28,000 and the 99th percentile ₹78,000. It is a genuinely top-decile fresher outcome, not a middle-class one.

The natural objection is that PLFS, a household survey, misses well-paid young engineers living in shared urban rentals. Three checks say it does not, and they are reproducible — check_pool_plausibility.py and check_plfs_pool_ci.py.

• The audited anchor. If PLFS were blind to elite young engineers, the IITs and NITs alone would overflow the national pool. They fill 62% of it.
• Coverage. 71% urban, and 58% of employed engineering graduates are in software. The frame is finding the right people; software simply pays ₹3.9 lakh at entry.
• Sampling. Bootstrapped over 916 first-stage units, because PLFS is clustered and treating people as independent understates the error by a factor of four. The rate is 4.42%, CI [2.19%, 7.19%]. Even the upper bound leaves the pool at 56,441 — short of NIRF’s 62,613.

We discount NIRF’s salaries by 20%, and its placement rates not at all

Colleges report their own placement salaries to NIRF, unaudited, and the figure feeds their ranking. Cost-to-company bundles what never reaches a payslip — employer provident fund is 5–6% of it, gratuity about 2%, unrealised variable pay 3–5%, annualised joining bonuses 1–3%. Eleven to sixteen percent before anyone exaggerates.

Three independent checks agree. Avanti’s own alumni started on 0.75× what their colleges filed. Delhi Technological University’s per-student NAAC list shows a ₹10.50 lakh median against ₹13.25 lakh filed to NIRF. And the arithmetic gets there from statutory contribution rates alone.

We apply no discount to placement rates. We looked for evidence and could not find it: NIRF’s internal arithmetic is consistent, AICTE’s placement series is unusable, and our own alumni turn out more employed than their colleges claim. An unevidenced haircut is not better than none. The reconciliation above is the vindication — placement fits, salary does not.

What a single median hides

NIRF publishes one median per college, not a distribution, so salaries within a college are modelled as log-normal around it. Treating a college as wholly above or below ₹6 lakh would credit a ₹4 lakh college with no good jobs at all.

This assumption is not load-bearing. Across spreads from σ=0.20 to σ=0.80 the ranked total moves between 58,000 and 69,000 and the over-claim stays between 1.67× and 1.98×. Removing the distribution entirely lands within 1.2% of the same answer. The result is a property of what colleges file, not of the curve laid over it.

A correction

An earlier version of this analysis reported that ranked colleges accounted for more than 100% of India’s good jobs. Part of that was our own error: the source table summed duplicate rows, so 54 institutes had every figure multiplied, including median salary, which cannot be summed. We caught it against the scorecards NIRF itself publishes, and the bug is filed.

The over-claim reported here is a different finding, reached after that fix, and it rests on matching the timing of the comparison rather than on the arithmetic that broke.

What we are least sure about

The doctors row rests on small cells — 94 people, 7 of them above ₹6 lakh. PLFS cannot separate MBBS holders from nurses on education, so doctors are identified by occupation, which makes both rates an upper bound. Directionally sound, not precise.

The exam funnels are indicative. Not every engineering entrant sits JEE Main, so the 51% with no engineering seat is overstated; that group also includes repeaters and people who never enrol.

None of this is causal. Elite colleges select hard on entrance rank, and their graduates would earn more than average whoever taught them. This describes who holds the jobs, not what produced them.

Files

Everything is at storage.googleapis.com/avantifellows-bq-assistant/analysis/nirf-college-employability/

File	What it is
nirf_college_employability.xlsx	Every college, course by course, with a link to its NIRF scorecard
tier_tables.csv	The tier tables, machine-readable
reconciliation.csv	Ranked claims against national totals
national_totals.csv	Formal and informal, by field
exam_funnels.csv	Where JEE and NEET applicants land

Method in METHOD.md, rebuild steps in TECHNICAL.md. The NAAC and NIRF documents behind the assumptions are catalogued in SOURCES.md, archived privately because the per-student placement lists name individual students.