What we know about young Indians' job prospects
Reads four analyses of public data against each other — NIRF’s institute disclosures, the NAAC register, AISHE/AICTE/NMC capacity, and MoSPI PLFS survey microdata. Every figure below is carried from a committed extract; this page runs no query of its own. No Avanti student data and no personally identifiable information.
The conclusion is not that the numbers are bad. It is that the numbers we have describe the best part of the sector, and nobody publishes the rest.
What we can actually see
Section titled “What we can actually see”Three public sources describe Indian higher education and they see different things. Read together they say something none of them says alone.
| Source | What it observes | The catch |
|---|---|---|
| PLFS | People. A household survey that knows what a graduate is doing years later. | Cannot name the college. Nobody opts in — which is its strength. |
| NIRF | Colleges. Institutions submit their own placement and salary figures. | Unaudited, self-reported, and only from colleges that chose to be ranked — 9.1% of engineering ones. |
| NAAC | Accreditation. A grade for process and infrastructure. | Not an outcome. And the register sheds its weakest members, so a snapshot flatters the sector. |
Two conclusions the data supports
Section titled “Two conclusions the data supports”One: what you study decides your prospects. How long you study barely does.
Section titled “One: what you study decides your prospects. How long you study barely does.”An engineering degree puts 52.6% of its holders in formal work — a written contract or social security. A general degree, held by 49.26 million people, manages 16.2%. A two-or-three-year technical diploma (30.8%) beats a general master’s (25.6%) — four more years of study, a worse result.
Two: everything we can see is the best case, and the part we cannot see is most of it.
Section titled “Two: everything we can see is the best case, and the part we cannot see is most of it.”For engineering, published college-level outcomes exist for 263 of 2,897 colleges — about one in 11, covering 27.0% of graduates. Those colleges chose to be ranked. And inside that visible slice, reported salaries fall steeply toward the bottom of the table, which means the unmeasured majority sits below a visible floor that is already low.
For most Indian engineering graduates nobody publishes what happens to them — and the closer a college sits to the edge of what is published, the worse its reported outcomes get.
Engineering, in full
Section titled “Engineering, in full”The largest field and the best covered, and still only about one college in 11 publishes what happens to its graduates.
Step one — we can see about one college in 11
Section titled “Step one — we can see about one college in 11”India has 2,897 AICTE-approved engineering colleges. Published college-level outcomes exist for 263 of them — every college NIRF has ever ranked, at its most recent report. That is 9.1% of colleges and 27.0% of graduates.
The ranked colleges are also about four times larger than average, which is why a tenth of the colleges hold a quarter of the students. And they applied to be ranked.
Step two — pay collapses toward the bottom, but placement does not
Section titled “Step two — pay collapses toward the bottom, but placement does not”| Rank band | Colleges | Graduates | Placed | Median salary |
|---|---|---|---|---|
| 1-10 | 10 | 8,921 | 76.0% | ₹19.5L |
| 11-25 | 15 | 25,282 | 80.8% | ₹13.8L |
| 26-50 | 25 | 31,811 | 67.1% | ₹9.0L |
| 51-100 | 68 | 53,415 | 70.1% | ₹8.1L |
| 101-150 | 51 | 39,289 | 73.1% | ₹4.5L |
| 151-200 | 94 | 63,689 | 67.6% | ₹4.0L |
Salary falls from ₹19.5 lakh at the top to ₹4.0 lakh at the bottom — 4.9×, dropping at every step. The share of graduates placed barely moves: 76.0% at the top, 67.6% at the bottom.
So the ranking sorts colleges on what graduates are paid, not on whether they are hired. Both numbers come off the same form, from the same college. A rank is a wage signal that is widely read as a job signal.
Step three — the two sources cannot simply be subtracted
Section titled “Step three — the two sources cannot simply be subtracted”It is tempting to put NIRF’s 70.8% placed beside PLFS’s 52.6% in formal work and hand the difference to the unranked colleges. That is invalid: NIRF counts a placement offer at graduation, PLFS counts a formal job among 21-to-34 year olds, and the two populations differ.
The comparison that does work is on well-paid jobs, where both sources can be put on the same footing. Across a graduating cohort of 7.78 million, PLFS finds 55,108 graduates in formal work paying ₹6 lakh a year or more. India’s ranked colleges claim 83,764 — 1.80× the entire national pool.
Placement counts, by contrast, reconcile: ranked colleges’ claims fit inside the national total at 0.59×. The over-claim is in salary, not in hiring.
Step four — either way you resolve that, the unseen majority looks bad
Section titled “Step four — either way you resolve that, the unseen majority looks bad”Two readings fit, and they are not equally flattering, but both point the same way.
- The salary figures are inflated. Then the published numbers — the best-case slice — are themselves too high, and the real distribution sits below what this page shows.
- The salary figures are honest. Then ranked colleges hold essentially every well-paid job in the country, and the 73.0% of graduates outside them hold almost none.
There is no third reading in which the unranked majority is doing well. The lowest visible band already reports ₹4.0 lakh, and the unmeasured colleges are, by construction, the ones that did not qualify to be measured.
What this argument does not establish. No unranked college is observed here — not one. We do not know their placement rates, and we deliberately do not project them: an earlier version of this work fitted five curves to rank-versus-salary and extended them past the ranking, and the five disagreed by nearly a factor of two while fitting the visible range about equally well. That projection was removed rather than published. Placement figures are also unaudited institutional self-reports, and NIRF publishes no definition of “placed”. See the full method, including what was retracted.
Deeper: the engineering discipline note for rank bands, gender and the colleges NIRF has dropped; the engineering rung for what PLFS sees, including that 52% of employed engineering graduates work in software rather than engineering.
Every qualification
Section titled “Every qualification”Fourteen rungs, from below middle school to a medical postgraduate. One survey, no institution opting in or out.
| Qualification | People 21–34 | In formal work | Most common job |
|---|---|---|---|
| Below middle school | 53.97 million | 1.7% | Skilled trades & production (44%) |
| Middle school | 50.82 million | 3.6% | Skilled trades & production (46%) |
| Secondary (10th) | 34.03 million | 6.0% | Skilled trades & production (46%) |
| Higher secondary (12th) | 47.85 million | 7.4% | Skilled trades & production (27%) |
| Diploma, technical | 4.06 million | 30.8% | Skilled trades & production (33%) |
| Diploma, non-technical | 2.85 million | 25.8% | Skilled trades & production (38%) |
| Degree, engineering | 5.42 million | 52.6% | Software & IT (52%) |
| Degree, medical | 850 thousand | 32.4% | Health & medical (76%) |
| Degree, other technical | 2.27 million | 24.9% | Teaching & education (25%) |
| Degree, general | 49.26 million | 16.2% | Clerical & office support (19%) |
| Postgraduate, general | 11.15 million | 25.6% | Teaching & education (23%) |
| Postgraduate, other technical | 1.64 million | 37.3% | Teaching & education (36%) |
| Postgraduate, engineering | 813 thousand | 53.1% | Software & IT (45%) |
| Postgraduate, medical | 200 thousand | 62.1% | Health & medical (80%) |
Field decides, not level. A technical diploma reaches 30.8% — above a general master’s at 25.6%, which takes four more years. Engineering reaches 52.6%; the general degree that 49.26 million people hold reaches 16.2%.
Formal work means a written contract or social security, not any paid activity. Most people outside it are working — informally, self-employed, or in unpaid family labour. Each rung carries the full breakdown.
Population counts are lower bounds. Survey weights sum to a Census-2011-anchored projection about a fifth below India’s current population. Three independent anchors put the correction at 1.17×–1.32×, and it is not uniform by age, so no single scalar is applied. No rate or comparison here is affected — numerator and denominator carry the same deflation.
Every ranked discipline
Section titled “Every ranked discipline”| Discipline | Level | Graduates ranked | Placed, per NIRF | Salary, top vs bottom band | Share of field seen |
|---|---|---|---|---|---|
| Engineering | UG | 107,526 | 70.8% | 4.9× | 27.0% |
| Management | PG | 31,949 | 88.1% | — | 11.3% |
| Medical | UG | 6,953 | 60.5% | 0.8× | 14.2% |
| Dental | UG | 2,624 | 59.5% | 1.3× | 22.5% |
| Pharmacy | UG | 7,282 | 44.0% | 1.8× | 9.2% |
| Law | UG | 6,132 | 69.2% | 2.4× | 4.3% |
Coverage varies twelvefold. Published outcomes exist for 50.2% of agriculture graduates and just 4.3% of law graduates. “The ranked sector” means something very different in each field.
Rank predicts pay in engineering and almost nowhere else. Engineering’s top-to-bottom salary gradient is far steeper than any other discipline’s, and in medicine it inverts — lower-ranked colleges report higher median salaries than the top ten.
These are institutional self-reports, unaudited, from colleges that chose to be ranked. Salary is a median of institute medians, so it weights a 60-graduate college the same as a 1,200-graduate one — a useful ordering of fields, a poor estimate of what a graduate earns.
Accreditation, and what it does not tell you
Section titled “Accreditation, and what it does not tell you”7,566 colleges hold a NAAC grade. The typical one scores 2.67 out of 4 — a B.
| Grade | CGPA band | Colleges | Share | Accreditation lapsed |
|---|---|---|---|---|
| A++ | 3.51–3.83 | 216 | 2.9% | 1.9% |
| A+ | 3.26–3.50 | 654 | 8.6% | 2.0% |
| A | 3.01–3.25 | 1,230 | 16.3% | 2.7% |
| B++ | 2.76–3.00 | 1,306 | 17.3% | 2.9% |
| B+ | 2.51–2.75 | 1,390 | 18.4% | 4.7% |
| B | 2.01–2.50 | 2,070 | 27.4% | 6.8% |
| C | 1.51–2.00 | 700 | 9.2% | 16.0% |
The register sheds its weakest members. A C-grade college is 9 times as likely to be sitting on a lapsed accreditation as an A++ one, and the gradient runs smoothly across all seven grades. Colleges that scored badly do not come back for the next cycle — so any snapshot of this register flatters the sector it is drawn from.
An A grade puts a college in the top quarter; A++ in the top thirtieth. Worth knowing, since “NAAC accredited” is the claim usually made and the grade is the fine print.
A grade is not an outcome. NAAC assesses process, infrastructure and governance, not what happens to graduates. And the grade carries no information the CGPA does not — it is a deterministic band of it, verified across every college. Full analysis.
Check our work
Section titled “Check our work”Every number here is produced by a committed query and verified by a script that fails loudly. Nothing is typed in by hand, and this page computes nothing of its own.
- The education ladder — fourteen qualifications, 263,446 people surveyed. Carries its own register of retracted claims.
- Outcomes by discipline — eight disciplines, rank bands, coverage against the regulator’s own college counts.
- The accreditation register — 7,566 colleges and 497 universities.
- The reconciliation — where NIRF and PLFS can be compared on the same footing, and where they cannot.
- What a degree pays — earnings by qualification, and the pooled four-year version.
What the checkers check. Every published cell against the CSV that produced it — the PLFS ladder alone has 2,005. And comparative claims, not just numbers: “the lowest placement rate” is asserted against the data, because a sentence can go stale while every cell in the table stays correct. The first draft of the discipline notes contained five false superlatives, each with correct cells.
What we decided we could not say. No projection into the unranked sector — five functional forms fitted the visible range about equally well and disagreed by nearly a factor of two once extended. No placement or pay by gender: NIRF publishes enrolment by sex and placement in total, never placement by sex. No causal claim anywhere — selection into engineering is not random. And no institution-level join between NAAC grades and NIRF ranks; it is possible and has not been done.