The Queue and the Missing Chairs
India has built the world's most ambitious skilling bureaucracy. It has not yet built a labour market that rewards the skill. The error is conceptual: we treat training as the near-cause of employment. A certificate can move a young person forward in a queue; it cannot add a chair. It cannot create a vacancy, raise a firm's productivity, pay for a safe commute, or compel an employer to retain a new worker. Yet for a decade, we have measured success by the length of the queue—crores enrolled, lakhs certified—while avoiding the harder question: where are the jobs, and do they pay?
The evidence is not absent; it is inconvenient. The Union government reports over 2.27 crore beneficiaries across PMKVY, the Jan Shikshan Sansthan programme, and the National Apprenticeship Promotion Scheme. PMKVY 4.0 alone claims 27.08 lakh trained by December 2025, with ₹1,652.89 crore utilised. Scale is real. But scale is not capability, and a certificate is not a wage.
The Comptroller and Auditor General's performance audit of PMKVY 1.0–3.0 provides the hard accountability record. For 2015–24: ₹14,448.90 crore outlay, ₹10,194 crore released, ₹9,261 crore utilised, 1.10 crore candidates certified. Of 56.14 lakh audited certifications, 23.18 lakh—about 41%—were reported placed. Yet the audit found missing demand analysis, weak verification, extreme state variation, irregular placement evidence, delays of up to 1,270 days, and nearly ₹40 crore in unpaid benefits. The aggregate number hid the truth: placement ranged from 51.08% in PMKVY 2.0 to 13.47% in PMKVY 3.0; from 49% in Assam to 0.7% in Rajasthan.
This is the dichotomy in stark relief. We have created a certificate factory that runs at full capacity, while the job market remains a thin, precarious bridge.
The Arithmetic of Illusion
The first public mistake is binary: either the mission did nothing, or the number trained proves success. Both collapse distinct measures. "Trained" is an activity. "Placed" is a reported transition—sometimes a joining letter, sometimes three days of work, sometimes a family enterprise with zero revenue. "Employed" is a survey status. "Decent work"—the only outcome that matters—requires earnings, stability, safety, voice, and social protection.
The government's own data reveals the gap. A late-August 2026 PMKVY dashboard snapshot showed 2.85 million trained and 200,224 placements—roughly 7%. Officials counter with a 42.8% placement rate for PMKVY 3.0, but even that figure cannot tell us if the job lasted, if the skill was used, or if the wage covered the rent. The CAG's field checks are more damning than any percentage: in Kerala, only five of 14 sampled candidates claimed as placed by one company were confirmed employed; two other companies had none of the claimed 17 and 18 candidates in work. The ministry recovered ₹22.33 lakh and blacklisted the agency. This is not proof that every placement was false; it is proof that verification was never built into the payment architecture.
Meanwhile, the Periodic Labour Force Survey 2025 describes the destination. Of all workers aged 15 and above, 56.2% are self-employed, 23.6% are regular wage/salaried, and 20.2% are casual labourers. Among non-agricultural regular workers—the "formal" economy—58.2% have no written contract, 47.3% lack paid leave, and 51.7% lack specified social security. Average earnings: ₹22,699 monthly for regular workers, ₹14,861 gross for the self-employed (not profit), ₹453 daily for casual labour outside public works.
We are training people for a labour market that rewards presence more reliably than skill.
When Supply Forgets Demand
The mission optimised what the state could count: batches, enrolments, assessments, certificates. Jobs were left to the end, often as a provider-reported number. Training centres compete for seats, mobilise whoever is available, choose courses based on equipment, then search for a job after classroom work. Employers may validate curricula or hire at the end, but are rarely required to expose a vacancy, provide supervised work, or share training costs.
The CAG found no micro-level skill-gap information sufficient to guide training, no long-term national strategy to translate economic plans into occupational demand, and the ministry's own admission that selecting job roles without demand analysis was a primary cause of poor placement. The mismatch is visible in sector choices: construction represented 28.37% of estimated skill requirements but only 3.21% of fresh PMKVY training; electronics represented 4.2% of requirements but 16.87% of training. Courses were not selected from verified local or sectoral demand; they were selected from what providers could deliver.
Recognition of Prior Learning (RPL) reveals how budgets shape composition. As short-term training targets fell, lower-cost RPL targets rose. RPL is valuable when it formalises capability and improves safety; it is poor value when certification substitutes for upgrading, tools, or supervised practice. We are mass-producing paper recognition while the underlying capability remains unbuilt.
The expense problem is not "too much spending"; it is weak marginal value. Nearly one-fifth of surveyed candidates were unaware of their benefits. District Skill Committees sat on ₹32.21 crore in untransferred funds. An administrative system that rewards utilisation before outcomes creates a market for paperwork, not productivity.
The Thin Bridge of Apprenticeship
Short-term training has been asked to perform the work of a school, a vocational college, a workplace, and an employment exchange. It cannot. Tacit knowledge—pace, quality, safety, customer handling, team discipline, the use of real equipment—is learned by doing productive work under supervision. Employers learn whether a worker fits; a vacancy becomes visible; a wage and reference become possible.
India has expanded apprenticeships, but participation remains uneven and concentrated. NITI Aayog's apprenticeship review highlights regional disparities, low MSME participation, and high dropout. The problem is cost: firms receive the benefit of a trained worker but bear the cost of coaching, tools, and supervision. The German dual system—in which apprentices spend three to four days weekly at the workplace under contract, with inter-company centres providing practical training small firms cannot offer alone—offers a principle, not a template. India cannot import institutional trust or employer density overnight, but it can adapt through cluster-based consortia, shared labs, and common mentors.
Without this bridge, a certificate remains a weak signal, and firms face a risky new hire. We leave young people to sink or swim in an ocean of informal recruitment, while employers complain of "unemployability." The truth is simpler: we have built classrooms without workshops, and workshops without contracts.
The Geography of Exclusion
A vacancy in an industrial cluster is not accessible to a young woman who cannot safely travel or find childcare. It is not accessible to a rural worker who cannot finance relocation and the first month's expenses. Job information is unequal: better-connected graduates learn about openings through networks; first-generation workers encounter opaque recruiters and fraudulent promises.
The female data exposes the cruelty of gender-neutral targets. Female self-employment stands at 64.2%, but average gross earnings for self-employed women are ₹6,374—less than half the male average. Among casual labourers, women earn ₹324 daily versus ₹453 overall. An 80% female classroom with a 20% female retention rate is not inclusion; it is a marketing statistic. Programmes need safe transport, hostels, childcare, predictable shifts, anti-harassment enforcement, and recruitment close to where women live. These are not "soft issues"; they are labour-market infrastructure.
The Weak Premium: Why Certificates Don't Pay
Low pay has structural causes: output value, firm productivity, bargaining power, and enforcement of standards. A three-month course rarely changes these fundamentals. If a firm sells a low-margin product, uses obsolete equipment, or cannot scale, it cannot pay a high wage simply because a worker holds a new certificate. Training subsidies without productivity upgrading pay for a signal the market has little reason to price.
The Economic Survey 2024–25 reports that more than 53% of graduates and 36% of postgraduates are underemployed below their qualification level. Educated youth face high unemployment because education is expanding faster than suitable work. More credentials intensify disappointment when jobs lack the complexity to use them. Good vocational systems do not end at entry; they offer ladders from helper to operator, technician, supervisor, specialist. India's fragmented programmes stop at the first certificate, leaving workers to pay and search again, looking interchangeable to employers who have no reason to reward progression.
From Certificate Factory to Employment Infrastructure
The remedy is to turn skilling into employment infrastructure. Fund verified capability and retention, not enrolment. Make employer-linked apprenticeship the default bridge. Implement NITI Aayog's proposed Digital Skills Passport, Skills Digital Registries, NCrF portability, and outcome-based financing—but treat them as tools to build, not achievements to claim.
Rebuild the chain from demand backwards. For every publicly funded pathway, the district must answer: which employers need this work; how many openings exist in 12–24 months; what capability is actually used; what wage and conditions are offered; who finances supervised practice? If these answers are absent, the programme should be a small experiment, not a mass target.
Put work inside learning. Every substantial vocational pathway should contain a paid, supervised work component. A practical funding formula: release 30% for mobilisation, 30% for completion and verified work assessment, 20% at three months when the graduate is employed or earning independently, and 20% after six or twelve months with a wage and job-quality check. No provider should earn full payment merely by issuing certificates.
Make job creation an explicit industrial-policy output. Production-linked incentives, public procurement, and infrastructure spending should disclose jobs created, median pay, apprenticeship conversion, women's participation, and social-security coverage. The test is additionality: did public support create work that would not otherwise exist?
Turn MSMEs from training consumers into partners. Cluster-level compacts can combine committed vacancies, shared labs, public cost-sharing, and local solutions for transport and childcare. Firms should receive support for supervised learning and productivity upgrade, not a bounty for hiring briefly.
Build the worker layer. The Digital Skills Passport must show the task performed, equipment used, supervisor, hours of practice, employer outcome, wage band, and retention—not just a logo and QR code. It needs consent-based sharing, correction rights, and a hard distinction between "completed a course" and "demonstrated capability."
The Roadmap: Killable Programmes, Not Killable People
The operational shift is blunt. Freeze mass expansion of job roles without verified demand. Publish provider- and district-level cohort results. Independently verify employment claims. Clear DBT and certificate backlogs. Pilot outcome-linked payments in representative occupations, with wage thresholds fixed before enrolment and retention checks at 3, 6, and 12 months.
Create sector and cluster consortia; embed paid workplace learning with academic credit; fund shared labs and mentors for MSMEs. Implement the Skills Passport and NCrF pilots with privacy safeguards. Tie industrial incentives to additional jobs and median real earnings.
The governance rule: repeatedly failing providers or job roles lose public funding, while learners receive transfer options. Districts can experiment, but results must be comparable. A public dashboard should show one blunt measure: the cost per graduate still in a relevant job or viable enterprise after 12 months, above the published earnings threshold.
Conclusion: The Bridge Must Reach the Other Side
India does not have a skilling problem in isolation. It has a jobs architecture problem. The state built a bridgehead of courses, centres, dashboards, and certificates, but left the bridge unfinished. The CAG demands a reset of execution and financial accountability. PLFS shows why employment totals conceal low pay and insecurity. The Economic Survey and NITI Aayog offer a starting point for demand mapping, apprenticeships, and portable capability.
The next phase must be judged by fewer, harder questions. Did a real employer need this skill? Did the learner demonstrate it in practice? Did the job last? Did earnings rise? Could the worker progress? Did the firm become more productive? Did public support create additional work rather than relabel existing vulnerability?
If yes, scale aggressively. If no, redesign or stop, however attractive the enrolment number. The ambition is not to train India to wait more efficiently for jobs. It is to make learning, enterprise, and employment one system—one capable of creating enough chairs, paying for the work of sitting in them, and giving people a ladder after they arrive.
000
