Recommendations to government: fixing educated unemployment through what the state already pays for

13 September 2026. The recommendations of The lost middle as they stand after four critiques (CRITIQUE.md) and two persona rounds (DEBATE.md). Each carries its instrument, who signs, what it costs, the evidence grade the labour economist attached, who gains and who loses, and the objection it had to survive. Costs are the Expenditure persona’s; figures marked (verify) are from persona memory, not from the folder’s data.

---

The problem, stated so that the recommendations follow from it

1. The unemployment is a queue, and the queue is an age window. Following the same men born 1994–97 across six PLFS releases, general graduates go from 43% unemployed at 23 to 8.5% at 29, engineers from 57% to 7%, polytechnic holders from 35% to 7%, Class-12 leavers from 23% to 3% (CRITIQUE.md §3.4). Cohort differences at fixed age are small. This is the exam-eligibility window of government recruitment, not a skills mismatch, and no training scheme touches it.
2. The degree is the cheapest ticket that keeps every door open. ₹5,000 a year at a government college, flexible attendance, Level 4–10 eligibility, PG and B.Ed continuation — and, for SC/ST/OBC households, entry to the one labour market where the private sector’s caste discount is void. A government polytechnic at ₹8,000 opens Level 6 and little else; a reimbursed private B.Tech costs the household nothing. Households are choosing correctly against the prices the state has set.
3. Class-12 passers and graduates are already buying trade training — alongside the credential, not instead of it. Four in five people aged 18–24 who report full-time formal vocational training hold Class 12 or a degree; one in seven stopped at or before Class 10 (§3.2). The “degree plus certificate” bundle exists; no funded institution issues it.
4. The middle credential with the clearest advantage in the data is that bundle. Among men 35–40, general graduates with formal training are in regular salaried work at 71% against 48% without, with a government-job rate of 29% against 16%; the advantage holds in every household- consumption quintile (§3.1, §3.3). The polytechnic’s lead over the plain degree is real but small (62% vs 48% at ₹29,000 vs ₹27,000) and too thin to rank once wealth is held constant. The graduate-level diploma is the strongest middle credential of all (83%, ₹55,000, men).
5. The public middle is funded as salaries for posts that are not filled. State technical- education budgets are 85–90% establishment; faculty vacancy of 50% is the norm; recruitment runs through state PSCs on 3–5-year cycles and stalls on roster litigation. Of the Centre’s ₹34,200 cr of new commitments to technical education for 2025–30, 88% is for ITIs; the scheme for the middle is budgeted at a third of its run-rate. States and the Centre together pay private B.Tech fees through reimbursement schemes and the Post-Matric Scholarship while a third of government polytechnic and engineering seats sit empty — in Tamil Nadu, Punjab and Rajasthan; in Bihar, Odisha and Kerala they are 71–77% full.
6. Every scheme for this tier has failed for one reason. The polytechnic sub-mission the skills ministry inherited in 2017 upgraded 105 of 500; STRIVE spent a third; PMKVY 4.0 spent 14%. Money arrived without the authority to hire, and lapsed.

What follows is not a new mission and takes nothing from ITIs.

The recommendations

1. Let the middle hire: a faculty-fill component inside MERITE

What. The Centre funds 100% of the salary of newly recruited regular faculty in government polytechnics and government engineering colleges for five years, tapering 100/100/75/50/25 to the state, on the 2009 pay-revision-assistance model (verify). Release is against appointment letters, not sanction orders. Disbursement-linked indicators are on state actions: posts sanctioned by cabinet; AICTE 7th CPC scales notified; recruitment completed within 18 months through university- style selection committees outside the PSC cycle; roster cleared. A professor-of-practice route at diploma level and an industry-secondment route (one day a week from partner plants) count toward fill. A written year-six absorption undertaking is a release condition: a state that will not sign it does not get the money.

Instrument and signatories. EFC note re-scoping MERITE; CCEA because the outlay changes. Secretary (Higher Education), Secretary (Expenditure), Finance Minister, Cabinet; state cabinets for post creation.

Cost. ~10,000 posts in year one at ~₹13 lakh loaded = ₹1,300 cr; scaling to ~25,000 posts, ₹9,700–10,800 cr over five years. The loaded figure is computed, not estimated: a polytechnic Lecturer enters at a basic of ₹56,100 (AICTE diploma pay regulations, 2019, Academic Level 9A), dearness allowance is 60% from 1 January 2026, and with House Rent Allowance and the employer’s pension share the annual cost runs ₹12.95 lakh in a Z-class town to ₹14.30 lakh in an X-class metro. An earlier draft used ~₹12 lakh from memory, which was 5–8% low.

Evidence grade. Established that the constraint binds (faculty vacancy, scheme lapse record); plausible that filling it changes outcomes (industry hires from ~40 polytechnics and says the difference is “people and proximity”, debate/r1_industry_hr.md); untested as a causal claim.

Gains / loses. Gains: every student in a government polytechnic or GEC, disproportionately in states that never adopted AICTE scales. Loses: no one directly; the state’s wage bill from year six.

Survived. “No salaries on central money” (the taper and the undertaking); “this is 2008 again” (absorption is a condition, not a surprise); roster litigation (mitigated by the practice and secondment routes; not solved); “the Finance Commission is the right route” (closed until the next award period; this is the fallback).

2. Move the subsidy, not the student

What. State fee-reimbursement schemes (AP and Telangana fee reimbursement, Maharashtra Rajarshi Shahu, Karnataka SC/ST/OBC, Tamil Nadu first-graduate) and the Centre’s Post-Matric Scholarship for SCs become portable to the government polytechnic diploma on the same terms as a degree, with a ₹10,000 completion bonus paid on NCrF credit and APAAR record; lateral-entry seats in government engineering colleges are reserved for diploma holders. Hostel and maintenance allowances at parity, because the wedge is cost of attendance as much as fees.

Instrument and signatories. MoSJE revision of PMS-SC guidelines (Secretary MoSJE, Expenditure); state cabinet decision and government resolution amending each reimbursement scheme (Chief Secretary, cabinet); AICTE lateral-entry regulation.

Cost. Substitution, so near-zero net; the bonus ~₹120–150 cr a year by reappropriation within PMS-SC and the state schemes.

Evidence grade. Established that the relative-price wedge exists (a reimbursed B.Tech at ₹0 against a diploma at ₹8,000, debate/r1_expenditure.md); plausible that portability moves enrolment (no Indian test; the only comparator is the Haryana/Gujarat Class-12 equivalence, untested).

Gains / loses. Gains: low-income Class-12 passers in states with reimbursement, and the government polytechnic. Loses: private engineering colleges whose reimbursement receipts are the business model — the political killer, and the reason the Minister sells it as “the scholarship follows the student”.

Survived. “Fees are not the explanation” (correct for the sticker price; the wedge is relative and includes attendance costs); Vidyalaxmi loans as the alternative (dropped: ticket sizes below what banks originate, a ₹28,000 job does not service debt, nursing already at 21% of education NPAs).

3. A trade inside every general degree, compulsory, on idle public capacity

What. Dual enrolment: every first-year student in a state-university general degree takes a 300–600-hour practical trade module delivered by the district’s ITI or polytechnic on its idle capacity, credited through the National Credit Framework on APAAR, assessed by a third party against an NCVT/NSQF standard industry will test hires against, and recognised by the state PSC for state posts. Compulsory across a college, never elective. Funded through PM-USHA, whose window is extended.

Instrument and signatories. UGC regulation under NCrF making the credits count toward the degree; PM-USHA component guideline (Secretary HE); joint state resolution from Higher Education and Skill departments assigning ITI/polytechnic hours; NCVET recognition; state PSC notification recognising the credit — a release condition, not an aspiration.

Cost. ₹5,000–8,000 a student for consumables and instructor time (verify); ~₹150 cr in year one at 2 lakh students, ~₹700 cr a year at 10 lakh; ~₹2,000 cr over five years, 60:40 within PM-USHA.

Evidence grade. The best-supported middle route in the folder: general graduate + vocational men 71% regular salaried vs 48%, government-job rate 29% vs 16%, women 30% vs 18%, and the advantage holds in every consumption quintile (14–37 points) — but today’s holders self-selected, so this is descriptive, not causal. Industry says it would hire the profile “in volume” for discom field supervision, solar EPC, QC, stores and IT-services support at ₹2.4–3.2 L (from experience, verify).

Gains / loses. Gains: women and the bottom half, who are in the general degree and nowhere else in tertiary education; dead ITI and polytechnic sections, which get a use. Loses: nothing the student holds — exam eligibility is untouched. Risk: caste-coding if elective (hence compulsory); time-to-degree.

Survived. B.Voc’s death (salary-only grants, contractual trainers, PSC non-recognition — each made a funding condition); industry’s RPL objection (hours floor, third-party assessment); the equity critique (compulsory, not opt-in).

4. A diploma-apprentice channel in Junior Engineer posts

What. Twenty per cent of JE-grade recruitment reserved for course-completed diploma apprentices, on the Railways’ existing 20% Course-Completed-Act-Apprentice model: Railway Board circular for RRBs, DPE office memorandum for CPSEs, RDSS condition for discoms, state service-rule amendments for PWDs, rural works, irrigation and urban bodies. The apprenticeship is the diploma’s final year and the NAPS period at once; industry partners sign a written ≥50% conversion-to-rolls floor for diploma apprentices (not for ITI apprentices, which industry will not sign).

Instrument and signatories. Chairman Railway Board; Secretary DPE; Secretary Power; state cabinets.

Cost. Nil.

Evidence grade. Established that the JE post is the diploma’s option value (CREDENTIALS_AND_ NCVT.md §2) and that B.Tech holders crowd diploma holders out of it (industry, Secretary); plausible that a channel raises take-up where discom/PWD/Railways hiring is active (Mangal’s result that rules move behaviour); untested.

Gains / loses. Gains: diploma holders, overwhelmingly male. Loses: graduates competing for the same Level 6 posts (more female, more bottom-half). This is zero-sum inside government hiring and the paper says so.

Survived. “Hold JE recruitment steady” (dropped — nobody’s lever; this is a share of whatever is recruited); higher-qualification litigation from B.Tech holders (flagged; the apprentice route is the one form with precedent); “the private market will not mirror it” (accepted; the JE post is where the option value lives).

5. Run PM-SETU as designed, and let states repurpose — take nothing from it

What. A fund-release clause in the PM-SETU guidelines: DLI #1 (employment outcomes) verified at spoke level, not hub average; no central release to a hub whose spokes are under 60% occupancy for two consecutive sessions; states may move up to 20% of their ₹20,000 cr share into polytechnic second shifts and dual-enrolment skill labs housed in merged or repurposed government ITIs; no government ITI is closed — merge or repurpose only. DGT–NIOS Class 12 recognised by every state board and PSC so the ITI is not a dead end for the Class-10 leaver.

Instrument and signatories. Secretary MSDE, Secretary Expenditure (release clause); state boards and PSCs (NIOS recognition).

Cost. Nil.

Evidence grade. Established: the scheme is a PforR with an outcome DLI and ~2% approved (HYPOTHESIS.md §2.7); the spoke-level baseline does not exist in any document read; the ₹10,000 cr industry share has no stated counting basis and industry says it will not be cash.

Gains / loses. Gains: the 1.3 crore of each cohort who leave at or before Class 10 — the ITI’s proper audience and the poorest slice — because the money that reaches them will have to convert. Loses: hubs with empty spokes. No inter-state reweighting, so no regressive transfer.

Survived. The devil’s advocate (this is the scheme’s own targeting, enforced — agreed); the Minister (merge, never close); the equity critique (no performance reweighting between states).

6. Build only where the seats are full

What. Second shifts; new branches (electronics, EV-solar, mechatronics, instrumentation) replacing dead civil, textile and printing sections inside the same institution, with staff transferred within the campus; hostels — women’s first — in states with 70%+ fill (Bihar, Odisha, Kerala). No new campuses this cycle. Branch-, shift- and district-level vacancy published before any seat is added anywhere.

Instrument and signatories. MERITE capital window, 60:40 (90:10 NE and hill), second tranche on occupancy; state cabinet for branch restructuring.

Cost. ~₹600 cr capital over five years (100 hostels at ~₹6 cr) plus state second- shift running costs.

Evidence grade. Established: the state fill split (REVERIFICATION.md §9) and the branch concentration of vacancy (CRITIQUE.md Appendix B); women 10% of polytechnic holders (§3.1) and the women’s hostels 351 of 499 built, upgrades 105 of 500 (Tenth Report, 6 August 2025).

Gains / loses. Gains: Class-12 passers in the East, and women. Loses: districts with no polytechnic, which the Minister wants built and Expenditure refuses (recorded as open).

Survived. “Fill, don’t build” (state-differentiated); industry’s “certificates nobody tests” (branches tied to the state JE channel and to faculty fill).

7. Measure by payroll, not placement

What. EPFO-UAN and ESIC linkage on APAAR IDs for every AICTE, NCVT and state-board completer, published by institution, disaggregated by sex and social group from the first release; AICTE to publish lateral-entry and completion counts; MoE’s Analysis of Budgeted Expenditure to report technical education by level (polytechnic / degree) and object (salary / non-salary); every scheme DLI, ITI dashboard and industry MoU to count 12-month EPFO retention, never day-of-drive placement.

Instrument and signatories. MoE–MSDE–MoLE MoU; AICTE Approval Process Handbook condition; DGT MIS field; DPDP consent at enrolment.

Cost. ~₹50 cr (verify).

Evidence grade. Established that placement is inflated (industry’s own account; DGT 51% unemployed-but-available beside 80–97% placement claims) and that the folder’s central ranking could not be trusted until disaggregated by sex (§3.1).

Survived. “A survey tracer run by a 64%-vacant AICTE” (linkage, not survey); the equity critique (disaggregation as a condition); the industry objection to publishing bad numbers (industry offered its UAN data in exchange for NCVT-equivalence of in-house training).

8. Queue hygiene — the ask no one at the table can deliver alone

What. Binding annual exam calendars and published vacancy counts for SSC, RRBs and state PSCs; one-time registration; a dated re-exam rule after a leak. This is the one lever on the mechanism itself: it shortens the wait at a fixed number of posts, and the 2024–26 protests were about it.

Instrument. Recruiting-agency rules and Chief Ministers’ offices; not a scheme, not a Secretary’s gift.

Evidence grade. Established that rules move behaviour (Mangal); untested for calendars.

Dropped from v1, and why

v1 item	Why it is gone
Diploma Mission of PM-SETU’s order	No mechanism, no absorptive capacity, ₹24,000 cr state share; the polytechnic premise is a 13-point, male-only, upper-bound advantage at equal pay.
Default two-year diploma offer to Class-12 passers	Tracking: steers the bottom half and SC/ST off the reservation channel, excludes women, creams the top decile into B.Tech. Politically dead in a week.
“Hold JE recruitment steady”	Within nobody’s power; would build a second queue.
National Class-12 equivalence for the diploma	A state-board act; the Minister will do it, the Centre cannot.
Vidyalaxmi loans for polytechnics	Wrong constraint, contingent liability, default risk in the bottom half.
Closing dead ITIs	Does not happen; merge or repurpose.
Performance reweighting of PM-SETU between states	Regressive: moves money to Maharashtra’s full ITIs and away from states where industry is not.
“Stop treating skills as a synonym for ITI”	A sentence, not an expenditure; its content is now recommendations 3 and 7.

What should be measured before anything larger is committed

From the labour economist’s list (debate/r1_labour_economist.md), the two already run are in CRITIQUE.md §3.3–3.4. Still to run: coarsened matching on state × sector × MPCE quintile × age × social group across pooled releases; the age-cap discontinuity in graduate unemployment by social group; a seat-fill regression on AICTE approval data × DTE faculty fill × reimbursement rules by branch and shift; a counselling information RCT on the Chakravorty design; the state-by-state reimbursement-to-private-B.Tech total, which no one has tabulated and which decides how much of recommendation 2 is reallocation.