The Bleeding Entrepreneur

1. The fourth batch #

Consider a founder. She is a composite, but she is not an unusual one.

She is forty-one. She spent eleven years as a process engineer in a large company before she left to build something of her own — a small precision components unit on the edge of an industrial estate, thirty-one people on the rolls, a customer list of four, three of whom pay late. The unit exists because she mortgaged a flat. There is no investor. There is no board. There is a bank, a personal guarantee, and an EMI that falls due on the fifth of every month regardless of what the fifth of that month happens to look like.

Six months ago she took on eight trainees. She trained them herself, because there is no one else who can — no HR department, no learning-and-development function, no training manager. She taught them in the evenings after production, three to four hours at a stretch, for the better part of five months. She paid them through it. She taught them the machines, the tolerances, the customer's quality protocol, the reason the protocol exists, the specific and unglamorous discipline of catching a fault at station three rather than at dispatch.

Of the eight, two were good. One of those two left eleven days after the training ended, for eighteen hundred rupees more a month. The other is still there and she is grateful for him every single day. Of the remaining six: three were pleasant, willing, and after five months of instruction still could not reliably do the work; two treated the stipend as an income while they looked for something else, which she knows because she was copied by accident on a mail; and one talked. He talked in every review meeting. He raised problems with real fluency, he identified bottlenecks, he had opinions about the layout, and over five months he produced almost nothing. When she finally put him on a target he resigned within the fortnight and, she assumes, is now explaining the bottlenecks to somebody else.

This is the fourth batch she has trained. She is now assembling the fifth.

Nothing in that account is remarkable. That is the entire point of it. Ask any group of small manufacturers, any WhatsApp group of clinic owners or agency founders or fabricators, and you will get the same story with different nouns. And you will notice something in the telling: it is delivered flatly, the way people describe weather. It has stopped being an outrage. It has become the cost of doing business.

It should not be the cost of doing business. And it is worth being precise about why — because the reflex explanation, that today's young people lack character, is both unkind and wrong, and it will not fix anything.

2. What a small Indian business actually looks like #

Before anything else, it is worth correcting a picture.

When the word "business" appears in Indian public conversation, the image it summons is a large one — a listed company, a promoter, a quarterly result, an aggressive lobbying position. That image informs how the country talks about business owners: as a class with surplus, from whom things may reasonably be extracted.

The statistical reality is almost the opposite. India has around 7.83 crore enterprises registered on the Udyam and Udyam Assist platforms as of February 2026, and the Ministry of MSME's own composition data tells the story in one line: of 7.61 crore registered enterprises in January 2026, 7.56 crore were micro, 4.89 lakh were small, and 36,816 — about five in every ten thousand — were medium. Over ninety-nine per cent of India's registered enterprise base is micro. (Ministry of MSME Annual Report 2025-26; PIB, 30 March 2026)

The measured picture is smaller still. The Annual Survey of Unincorporated Sector Enterprises for 2025 counted 7.92 crore establishments employing 12.81 crore people, producing gross value added of ₹19.93 lakh crore. Divide it out and the average Indian unincorporated enterprise generates about ₹2.52 lakh of value added in a year — roughly ₹21,000 a month. (MoSPI / PIB, 24 March 2026)

That is the population under discussion. Together this sector accounts for around 31.1% of India's GDP, 35.4% of manufacturing output and nearly half of exports (Economic Survey 2025-26, via PIB). It is the substrate the economy stands on. And it operates on margins that leave no room for error at all.

CRISIL Intelligence, modelling the effect of a prolonged West Asia crisis, projects MSME EBITDA margins falling to 5–5.5% in FY2027, down 50 to 100 basis points, and notes that MSMEs absorb such shocks far harder than large firms do (CRISIL, June 2026). ICRIER's field survey of 2,365 registered manufacturing MSMEs found an average borrowing cost of 12.27%, that more than 98% of enterprises said the loan they received was not enough, and that only 23% had ever secured a collateral-free loan (ICRIER, March 2025).

Hold those two numbers next to each other. A single-digit operating margin against a twelve per cent cost of money. That is the arithmetic the founder in section one is doing at two in the morning.

Now add what else sits on her.

Delayed payments #

The Economic Survey 2025-26 puts the sum locked up in overdue payments to MSMEs at approximately ₹8.1 lakh crore. Independent estimates from GAME, FISME and C2FO trace a declining but still enormous figure — ₹10.7 lakh crore in 2022, ₹7.34 lakh crore as of March 2024 — and find that micro enterprises wait roughly three times longer than larger firms. The statutory payment window is 45 days; the measured national average invoice cycle is 73 days (Recordent, Indian SME Receivables Report 2026). The redress mechanism exists and moves slowly: of roughly ₹55,244 crore claimed through the MSME Samadhaan portal since 2017, about ₹3,018 crore had been settled by mutual agreement and a further ₹14,638 crore disposed by Micro and Small Enterprise Facilitation Councils as of December 2025 — together under a third of the value claimed, eight years in.

Compliance #

TeamLease RegTech's 2025 study of a single-state, single-unit manufacturing MSME counted over 1,450 compliance obligations a year, 59 categories of inspector with rights of visit, 48 registers to maintain, and 486 clauses under which the proprietor can be imprisoned — two-thirds of them arising from labour law. The estimated annual cost of compliance alone: ₹13–17 lakh (TeamLease RegTech, June 2025). For a formal unit turning over a few crores at a single-digit margin, that is a meaningful fraction of the year's profit, paid before a single rupee reaches the owner. Across India's business laws as a whole, TeamLease and the Observer Research Foundation counted 26,134 imprisonment clauses, with nearly two in five mapped compliances carrying a custodial penalty (Jailed for Doing Business, 2022).

And when it ends #

India publishes no MSME mortality statistic. A business that simply stops trading leaves no trace in any register. What we have instead is the National Crime Records Bureau. Between 2019 and 2021, 53,661 self-employed persons died by suicide — 16,098, 17,332 and 20,231 in the three years (Ministry of Labour, citing NCRB, Lok Sabha, February 2023). In the last year for which the occupational category was cleanly reported, 7,990 people classified as business persons died by suicide; separately, and across all occupations, 4,970 deaths that year were attributed to bankruptcy or indebtedness (NCRB, ADSI 2018). The NCRB assigns a single cause per death, which almost certainly understates the economic dimension.

This is the population that the phrase "businesses should do more" is usually addressed to.

None of this is an argument that entrepreneurs are beyond criticism. Exploitative employers exist in India, and later in this piece I will argue that they should face consequences as sharp as any I propose for anyone else. The argument here is narrower and factual: the median Indian enterprise is not a source of surplus. It is a person with a loan, a thin margin, and thirty-one people depending on them.

3. Three ways the money leaves #

Against that balance sheet, place the specific losses that this piece is about. There are three, and they are distinct.

4. The other half of the ledger #

If the article stopped here it would be a complaint, and it would deserve the response it got.

Because everything above is true, and so is the following.

Pay has not moved. Entry-level engineering salaries at India's largest IT employers have been stuck at ₹3–3.1 lakh a year — a figure essentially unchanged in nominal terms for around a decade, which in real terms is a substantial cut (People Matters, 2025). The ILO and Institute for Human Development found that across 2012–2022, real wages of regular workers either remained stagnant or declined (India Employment Report 2024). Analysis of PLFS data by Data For India puts average salaried earnings at around ₹21,000 a month in 2024, with real wages down about 4% between 2012 and 2024. An employer who has not raised wages in five years and is angry about disloyalty is describing a transaction, not a betrayal.

Job quality is poor. Around 57% of India's salaried workers have no written employment contract. Roughly six to seven per cent of the total workforce holds a permanent salaried job. Around 90% of workers are informally employed. India's average working week is 46.7 hours, and 51% of the workforce works more than 49 hours a week — the second-highest share in the world, after Bhutan (ILO, 2024). A person with no contract, no notice protection and no social security who leaves without notice is behaving exactly as the arrangement taught them to.

Employers default too. Reporting on EPFO data indicates employer arrears to the provident fund organisation rose to roughly ₹26,000 crore by March 2024, with about ₹10,000 crore of it owed by some 2,400 employers each in arrears of ₹50 lakh or more — and the EPFO does not publish a comprehensive defaulter list (The Ken; the underlying figures warrant confirmation from EPFO's own disclosures). Money deducted from a salary and not remitted is not a compliance lapse. It is the employee's money.

And people mostly do not leave for money. Randstad's Workmonitor 2025 India findings: 67% would quit a job that lacked learning and development opportunities — the single highest quit-driver in the Indian data, and far above the 41% global figure. 60% would leave without a good rapport with their manager. 58% have already quit because of a toxic environment. (Randstad Workmonitor 2025)

Read that against the foundit finding — 86% still looking after a 20% raise — and the picture resolves. People are not leaving because ₹18,000 became ₹19,800. They are leaving because they cannot see where they are going, and no one has shown them.

Which leads to the most uncomfortable number in this section. The World Bank's Enterprise Survey of 9,376 Indian firms found that only 7.7% of Indian firms offer formal training — against a lower-middle-income country average of 14.1% (World Bank Enterprise Surveys, India 2022). NITI Aayog cites the same 7.7% against Thailand's 18%, Malaysia's 24% and Mexico's 38%. Meanwhile 80% of Indian employers report difficulty finding skilled talent — up 25 points in a decade (ManpowerGroup Talent Shortage 2025).

I want to be careful here, because for the specific founder in section one that statistic is unjust — she trains constantly, at ruinous personal cost, and she is in the 7.7%. But at the level of the system, the number is real, and it explains why the employee's complaint and the employer's complaint are both correct. Employers stopped training because trained people leave. People leave because nobody trains them or shows them a path. Each side's behaviour is a rational response to the other's, and the loop tightens every year.

So the honest framing is not employers versus employees. It is this: there are good employers and unreliable ones, and reliable workers and unreliable ones, and none of the four can prove which they are to any of the others. The good employer cannot credibly promise that this job will be different, because every job promises that. The reliable worker cannot credibly promise that she will stay and deliver, because every candidate promises that. Both are trapped inside the same failure — and it is a failure with a name.

5. The name for this problem #

In 1970 George Akerlof published a paper about used cars that turned out not to be about used cars. Its argument runs as follows. When the seller knows the quality of a good and the buyer does not, the buyer can only rationally pay a price reflecting the average quality on offer. That price is too low for a good unit and too high for a bad one. Owners of good units therefore withdraw from the market. Average quality falls. The price falls again. The process repeats.

Akerlof stated the consequence directly:

"The presence of people in the market who are willing to offer inferior goods tends to drive the market out of existence — as in the case of our automobile 'lemons.' It is this possibility that represents the major costs of dishonesty — for dishonest dealings tend to drive honest dealings out of the market."

— The Market for "Lemons": Quality Uncertainty and the Market Mechanism, Quarterly Journal of Economics, 1970

He shared the 2001 Nobel Prize for this, with Michael Spence and Joseph Stiglitz, "for their analyses of markets with asymmetric information."

This is the entrepreneur's problem, described thirty years before most of the people now living it were born. The employer cannot observe reliability at the point of hiring. She therefore prices for the average — a low starting salary, a probation period, minimal training investment, a reluctance to delegate anything that matters. The genuinely reliable candidate, offered average terms, either accepts them resentfully or goes elsewhere. Standards drift down. The next employer prices lower still.

And note what Akerlof himself pointed to as the counterweight. In the same paper, discussing the institutions that arise to offset quality uncertainty, he named the obvious ones: "The high school diploma, the baccalaureate degree, the Ph.D., even the Nobel Prize, to some degree, serve this function of certification." Credentials exist to solve exactly this.

Which brings us to Michael Spence, and to the condition that determines whether a credential works at all. In his 1973 paper on job market signalling, Spence set out the requirement:

"…a signal will not effectively distinguish one applicant from another, unless the costs of signaling are negatively correlated with productive capability."

In plain terms: a credential only carries information if it is harder for the wrong person to get than for the right one. If anyone can obtain it at the same cost, it says nothing.

That is not a signal. It is a receipt. And it puts us back in Akerlof's market — where the employer must assume the average, and the honest candidate is punished for the dishonest one's existence.

And it is not only the paper trail. The degree itself is a weak signal in a country where 4.50 crore students are enrolled in higher education and more than a crore graduate every year (AISHE 2023-24), and where employability assessments of that cohort return figures ranging from 42.6% (Mercer-Mettl, India's Graduate Skill Index 2025) to 56.35% (India Skills Report 2026) — commercial measures on non-representative samples that disagree with each other by nearly fourteen percentage points, which itself tells you how little the qualification settles.

What this does to Indian firms #

The consequences are not theoretical, and the best evidence is from India.

In a randomised trial with textile firms near Mumbai, Nicholas Bloom and co-authors documented what Indian owners actually do when they cannot verify trustworthiness. Their finding, verbatim:

"In every firm in our sample, before the treatment, only members of the owning family had positions with any real decision-making power over finance, purchasing, operations, or employment. Non-family members were given only lower-level managerial positions with authority only over basic day-to-day activities."

The stated reason was specific — "they were concerned if they let their plant managers procure yarn they may do so at inflated rates from friends and receive kickbacks" — and so was the institutional cause: "A key reason for this inability to decentralize appears to be the weak rule of law in India. Even if directors found managers stealing, their ability to successfully prosecute them and recover the assets is likely minimal because of the inefficiency of Indian courts." (Bloom, Eifert, Mahajan, McKenzie & Roberts, Does Management Matter? Evidence from India, QJE 2013)

Most strikingly, in explaining firm size, the number of adult male family members available to run things predicted size better than the quality of management did — removing it from the regression cut explanatory power more than three times as much.

The companion study makes the cost explicit across countries: a one-standard-deviation increase in generalised trust is associated with roughly 30% larger firms, because a CEO who can delegate does not have to personally supervise, and one who cannot must (Bloom, Sadun & Van Reenen, The Organization of Firms Across Countries, QJE 2012).

Now set that beside India's most-quoted structural puzzle. The Economic Survey 2018-19 identified India's "dwarfs" — firms with fewer than 100 workers that have existed for more than ten years. They are more than half of all organised manufacturing firms by number, but account for only 14% of employment and 8% of productivity. India is full of businesses that never grew.

We usually explain that with regulation and credit, and both explanations are correct. This is a third explanation, and it sits underneath the other two: many Indian firms are small because their founders cannot find enough people they can verify.

The problem is not a shortage of trust. It is the absence of any bridge from the stranger into the circle.

6. Why no single firm can fix it #

Here is the part that explains why every individual solution has failed.

Suppose our founder decides to solve this herself. She could hire more carefully, verify harder, design a better probation. Suppose it works and she discovers, after six months, that a particular young man is reliable, trainable and honest.

She now possesses valuable information. She cannot keep it. The moment his competence is established, it becomes visible to every other employer, and she has no way to capture the return on the six months of discovery she paid for.

Amanda Pallais demonstrated this experimentally and named it exactly: "Hiring inexperienced workers generates information about their abilities. If this information is public, workers obtain its benefits." In her field experiment, workers who were hired and publicly evaluated saw employment rise from 12% to 30% and earnings approximately triple. The social value created was so large that the benefits exceeded the total social cost even though the experimental jobs produced no usable output at all. The entire value was informational. (Pallais, Inefficient Hiring in Entry-Level Labor Markets, AER 2014)

Information about who is reliable is a public good. Individual firms therefore produce too little of it, and take too few chances on unproven people.

This is not a moral failure by employers. It is the predictable outcome of asking private actors to fund a public asset.

The same logic governs training. Gary Becker's foundational result was that general skills — the kind that make a person more valuable to every employer — raise the worker's outside wage one-for-one, so the firm cannot recover the investment. (The modern literature complicates this usefully: Acemoglu and Pischke show that labour-market frictions and employers' informational advantage over outsiders can make firms train more, not less. But that mechanism depends on the incumbent employer knowing something outsiders don't — which requires exactly the kind of stable, verifiable relationship that India's churn destroys.)

Which is why 7.7% of Indian firms train. Not because they are short-sighted. Because under these conditions, training is a transfer.

So: no amount of effort by any individual entrepreneur will fix this. She can be more diligent than every one of her competitors and she will still be paying for information she cannot keep, and still be pricing for the average because she cannot see past it. The problem is structural, and structural problems require shared institutions. That is not an evasion of responsibility. It is a description of what kind of thing this is.

7. Where the deficit is manufactured #

If the market cannot verify people, the obvious question is what the system upstream is producing and what it is certifying.

Start at the bottom. ASER's Beyond Basics survey of 34,745 young people aged 14 to 18 across 26 states found that only 43.3% could solve a three-digit-by-one-digit division problem — arithmetic taught in Standard III or IV. About a quarter could not read a Standard II text fluently in their own regional language. Among those who could manage subtraction, more than 60% could handle a household budget, but only about 37% could apply a discount, and only around 10% could calculate the repayment on a loan. (ASER 2023 Beyond Basics)

The government's own national assessment agrees. PARAKH's Rashtriya Sarvekshan 2024, covering 21.15 lakh students across 74,229 schools in 781 districts, records national average scores of 37% in Grade 9 mathematics, 40% in science and 40% in social science — with percentages at 28% and logical proofs at 29% (PARAKH / NCERT, July 2025).

And the ILO's employment report adds the detail most relevant to any actual office: 75% of Indian youth cannot send an email with an attachment; around 60% cannot copy and paste files; roughly 90% cannot perform basic spreadsheet operations.

These young people are not idle. They are enrolled — ASER records 98.1% enrolment among rural children aged 6 to 14. They are graduating: more than a crore of them a year. The system is not failing to process them. It is processing them and certifying an outcome that has not occurred.

The consequences show up exactly where you would predict — at the top of the education distribution.

• The ILO-IHD report found youth unemployment among graduates at 29.1% in 2022, against 3.4% among young people who cannot read or write — nine times higher. In India, unemployment rises with education.
• The share of educated people (secondary and above) among all unemployed persons rose from 54.2% in 2000 to 65.7% in 2022.
• The most recent PLFS shows real headline improvement — overall unemployment 3.1%, youth unemployment 9.9%, down from the year before — but unemployment among those with secondary education and above at 6.5%, still more than twice the national rate (PLFS Annual Report 2025).
• Azim Premji University's State of Working India 2026 puts graduate unemployment near 40% among 15–25 year-olds.
• And the Economic Survey 2024-25, drawing on work by the Institute for Competitiveness, found that more than 53% of graduates are working in jobs below their educational qualification — its underlying table puts the share of graduates in high-competency occupations at 8.25% (Economic Survey 2024-25, Chapter 12).

Meanwhile the machinery built to fix this has struggled. Only 4.9% of Indian youth aged 15–29 have received formal vocational or technical training, per the Economic Survey 2025-26 — a figure the government's own headline reporting raises to 26.1% by folding in informal and self-taught learning. NEP 2020 itself noted that under 5% of the Indian workforce aged 19–24 received formal vocational education, against 52% in the USA, 75% in Germany and 96% in South Korea, and set a target of 50% exposure by 2025. We are at 4.9%.

The flagship skilling scheme tells the same story with more precision. The Comptroller and Auditor General's performance audit of PMKVY (Report No. 20 of 2025) found that of 56.14 lakh candidates certified under short-term training, 23.18 lakh — 41% — were placed. It also found that 94.53% of PMKVY 2.0 and 3.0 participant records had invalid, missing or null bank account details; that 1,79,407 unique email addresses were used across 71,20,995 candidate records; that 6.54 lakh underage candidates and 13.33% lacking the required education were certified anyway; and that in Kerala, training partners produced incorrect placement documents as proof of placement. (CAG of India, Report No. 20 of 2025)

And then the detail that closes the circle. A Parliamentary Standing Committee reported in March 2025 that placement tracking has been delinked from PMKVY 4.0, leaving the scheme's most recent phase without published employment outcomes. The Committee's own view of what that costs was blunt: "the placement statistics is the real barometer for measuring the success of the scheme." It recommended that placement details be uploaded to the Skill India Digital Hub.

The apprenticeship route is similarly hollow. Of roughly 1.96 lakh establishments registered under the national apprenticeship scheme in FY2024-25, only about 51,000 were active; more than 74% of registered establishments remained inactive. Around 77 lakh apprenticeship vacancies have been created since inception and over 70 lakh went unfilled. The reason given for small firms staying out is not indifference: they find the process cumbersome.

To be careful and fair about what this evidence does and does not show: none of it says that India's education system produces dishonest people. That would be an unsupportable claim and I am not making it. What it shows is narrower and, for this argument, sufficient:

1. The system certifies outcomes that have not occurred. A certificate that does not correspond to a capability destroys the informational value of every certificate — including the honest ones. When PMKVY certifies underage and under-qualified candidates and stops measuring placement, and when a Grade 9 cohort averages 37% in mathematics and still progresses, the certificate stops separating types. This is Spence's condition failing at national scale.
2. It measures almost nothing that predicts workplace conduct. Not reliability, not follow-through, not the ability to be trained, not the willingness to finish something unsupervised. NEP 2020 said this about itself: that assessment must shift away from "summative and primarily tests rote memorization skills," and that exam culture is "replacing valuable time for true learning with excessive exam coaching."
3. It leaves young people unable to prove what they are. A capable, honest twenty-two-year-old from a Tier 3 college holds a degree that means little, a mark sheet that means less, and no third-party evidence of anything an employer actually needs to know.

The last of these is the one the industry conversation almost always misses, so I want to state it as directly as I can. The reliable young person is a victim of this system in exactly the same way the employer is. She is standing in a pool that the employer has correctly learned to distrust, holding documents that are indistinguishable from documents available for ₹7,500, being offered ₹3 lakh and a probation period she will be assumed to fail. Her honesty earns her nothing, because there is no mechanism through which it can be seen.

8. What actually works #

This is the encouraging part, and it is not speculative. The question — can a credible third-party signal repair a labour market wrecked by information failure? — has been tested with randomised controlled trials in developing economies, published in the leading economics journals. The answer is yes, and the results are large.

South Africa #

Carranza, Garlick, Orkin and Rankin worked with 6,891 jobseekers in Johannesburg. They offered free assessment of communication, concept formation, focus, grit, numeracy and planning, and returned the results as a branded, verifiable certificate. Effects: employment up about 5 percentage points on a 30.9% base — a 17% increase — and weekly earnings up 34%.

The critical finding is what happened to the control variant. Private, unbranded certificates — the same information, without the identifiable, verifiable mark — produced no detectable employment effect at all, and earnings effects less than half the size, a difference the authors find statistically significant. The verifiability is what does the work. (Carranza et al., Job Search and Hiring with Limited Information about Workseekers' Skills, AER 2022)

The study also produced a finding that should reframe this entire debate: asked to rank their own skills, only 8% of candidates were correct on all three assessments, and 22% overestimated themselves on all three. The ignorance is two-sided. Workers do not know what they are either.

Uganda #

And this is the closest analogue to India's small enterprises. Bassi and Nansamba ran a trial with 787 vocational graduates and 422 small and medium enterprises. The certificates graded workers on creativity, trustworthiness, willingness to help others, attendance, and communication — that is, literally certified reliability, disclosed during job interviews, against a placebo-certificate control.

Results: employed workers with certificates earned 11% more over the following two years. Overall employment probability did not rise — but matching improved: higher-skilled workers ended up with higher-ability managers. And the worker-side effects are the ones that matter most for this argument. Certified workers became 15% less likely to take casual work, 3.8 percentage points more likely to pursue further training, more inclined to seek out larger firms, and reported a 7% reduction in the difficulty of signalling their own skills.

Credible certification made people invest more in themselves.

(Bassi & Nansamba, Screening and Signalling Non-Cognitive Skills, Economic Journal 2022)

Ethiopia #

Abebe, Caria, Fafchamps, Falco, Franklin and Quinn worked with 3,052 young people in Addis Ababa. A job-application workshop that included skills testing, with results presented as a certificate, raised permanent employment by 6.9 percentage points from a 12% base — a 60% increase — and, four years later, treated individuals were earning 25% more. Cost: about $18.20 per person excluding one-off development costs — $35 all in. (Abebe et al., Review of Economic Studies, 2021)

Four conclusions follow, and each of them constrains what a real solution must look like:

1. Verifiable third-party signals work, and the effects are large enough to matter to a business.
2. The signal must be branded, identifiable and checkable. Anonymous versions failed.
3. What is signalled must include conduct, not only skill. Uganda certified attendance and trustworthiness — the very attributes Indian employers say they cannot find.
4. A credible signal changes the person, not just their price. Once effort was visible, people invested more in themselves. Which points at something important: much of what looks like a character problem may be a rational response to a world where effort is not observed.

And it explains the foundit paradox from section three. Eighty-six per cent are still looking after a twenty per cent raise because money is not what is missing. What is missing is a legible path — a way to know that this year's diligence will be visible next year, to someone, somewhere. Sixty-seven per cent of Indian employees say they would quit over the absence of learning and development. That is the same statement in different words.

9. What is being built #

Everything above defines the specification. A solution to India's problem would have to:

The Rebuild India Foundation and its movement, ARISE, are building precisely this. I should say plainly what it is and is not: it is early. The Foundation dates from 2021 and runs Noble International School — a free school for 160 children in Bagalur, Bengaluru, operating since 2021 — and ARISE was founded in 2025, with its Charter Convocation scheduled for October 2026 and its first elected club boards due in July 2027. It is a registered charitable trust with 80G, 12A and CSR-1 registration. It has not yet been through a decade of results and nobody should pretend otherwise.

What is worth attention is the architecture, because it maps onto the evidence with unusual exactness.

The signal cannot be purchased #

ARISE runs on a single currency: one point is one validated hour. Points come from two sources that must stay in balance — learning hours (sādhana) across five dimensions of personal development and four builder competencies, and service hours (sevā) delivered as real work, typically teaching at a Nurture Centre. Neither can be bought. The ladder runs twelve levels, from Aspirant at zero to Beacon at ten thousand points, on the deliberate logic of the ten-thousand-hour idea.

This is the Spence condition satisfied by construction. You cannot forge two thousand validated hours of teaching children in the evenings. The cost of the signal is steeply negatively correlated with the very attributes an employer wants, because the only way to accumulate it is to be the sort of person who turns up on a Tuesday when nobody is watching, for years.

Conduct is observed, not claimed #

Two hours of service a week, thirty minutes of learning a day, a weekly club meeting. The clubs are stratified by life stage — an Empowerment Club for undergraduates, a Community Club for graduates, a Professionals Club, a Leadership Club for entrepreneurs and leaders, and a RITE Club for teachers. Assessment happens through Vyakti Darpan, across the physical, intellectual, emotional, social and spiritual dimensions plus leadership and building capacity, with an explicit rule that no questionnaire decides anything by itself:

"A person is never admitted or refused by a questionnaire. Every decision that matters is made by people, against a written standard, with room for you to be heard."

An employer reading a candidate's record is therefore not reading a self-report or a purchased letter. They are reading a multi-year attendance history at work nobody was paid for.

Bad conduct carries consequences — with due process #

ARISE maintains a Professional Conduct Register. Entries reduce accumulated learning points on a graded scale — 10% for a first entry, 15% for a second, 25% for a third — and a fourth entry means no earn opportunities and no privileges for two years. Entries expire after twelve months (twenty-four for the fourth). Membership itself is never revoked and earned certifications are never reversed; only access is withdrawn.

This is the piece almost every skilling initiative in India lacks, and it is the piece that makes the rest of it mean anything. A credential with no downside is a participation certificate. This one can be devalued by conduct — which is the only reason an employer would ever believe it.

It is free, by rule, in both directions #

Associates pay nothing to join or to progress; the membership fee is nil "by rule and permanently." Clubs hold no funds. A hosting college signs an understanding, names a faculty advisor and provides a room — no fee, no funding. And for companies, the Earn Partner agreement is explicit: "No partner pays the movement, and the movement pays no partner."

And the employer's obligations are symmetric. This is the detail I find most persuasive, because it addresses the fairness problem head-on. An Earn Partner commits to three things: share all opportunities at all levels; evaluate every candidate presented; return feedback within one week. In exchange the company receives access to candidates, an attribute validation report on each one, and a mutual-validation licence. Onboarding runs through a seven-point due-diligence review covering legitimacy, values alignment, youth-protection capability, financial integrity, brand-use intent, data handling and excluded-sector screening, followed by a written memorandum reviewed by legal counsel.

Read that again from the employee's side. A company that will not evaluate candidates or will not return feedback within a week does not get to be a partner. A company that fails a values-alignment or financial-integrity review does not get in. The mechanism that makes good candidates visible is the same mechanism that makes good employers visible — and both can lose standing.

That is the symmetry this article has been arguing for since section four. The point was never to give employers a better filter. It was to build a system in which reliability is legible in both directions and unprofessionalism is expensive on both sides of the table.

A separate Privilege Partner track completes the incentive: brands pre-commit specific privileges to each performance band, honoured on presentation of credentials, with no associate data passing to the brand. Rising in the system produces tangible advantage in ordinary life — which is what it means, concretely, for good behaviour to be rewarded rather than merely praised.

Underneath all of it sits the part that addresses the upstream deficit rather than working around it. The Nurture Centres — free evening learning centres for children in classes 6 to 10, five days a week, roughly a hundred children each, run by club members as their service — teach personal excellence alongside scholastic and professional excellence, with two named adults present at all times and police clearance required before unsupervised contact with children. The associates are simultaneously building their own record and repairing, one centre at a time, the foundational learning gap that section seven documents. And the RITE Clubs work on teachers, on the reasoning that they shape everyone else.

Nothing here requires the education system to be fixed first. That is the practical virtue of it. It builds the missing verification layer alongside the system that failed to produce it, and it does so with the one resource India has in surplus: young people who are willing, and have nowhere to demonstrate it.

10. An invitation #

Let me close by returning the argument to where it started, and putting it more carefully than the usual version.

It is commonly said that the entrepreneur must be saved because the entrepreneur creates jobs. That is true but it is a transactional argument and it invites a transactional reply. The stronger claim is this:

The founder in section one cannot find people she can trust. A twenty-two-year-old three kilometres away cannot prove that she is one. Neither of them created the gap between them. Both are paying for it — she in five months of unrecoverable evenings, the young woman in a starting salary that assumes she is average and a probation that assumes she will fail. Every year the gap widens, the employer trains less, the candidate's word is worth less, and both sides conclude, reasonably, that the other cannot be relied upon.

That gap is not a moral condition. It is a missing institution.

Countries that built one — through apprenticeship systems, verified credentials, functioning references — did not do it because their people were better. They did it because somebody built the bridge.

India has not built it. It cannot be built by any single company, because the information it produces is a public good and no firm can capture the return. It will not be built quickly by the state, on the evidence of the last decade. It has to be built by the people who need it, together, and it has to be free, or the people who need it most will not be in it.

So the ask is not for sympathy, and it is not for money.

One last thing, addressed to the industry specifically.

There is a version of this argument that ends by blaming young Indians for what the system did to them, and it is a popular version because it is easy and because it flatters the person making it. I have tried not to write it. The data does not support it. A generation that cannot do long division at eighteen was not given the tools; a generation that changes jobs every fourteen months has watched entry-level pay stand still for a decade; a generation accused of not being trainable is being hired by an economy in which fewer than eight companies in a hundred train anybody.

And there is an equally popular version that ends by blaming employers, which is just as easy and just as wrong for the ninety-nine per cent of Indian enterprises that are micro, running on single-digit margins, borrowing at twelve per cent, waiting seventy-three days to be paid, and owing 1,450 compliances a year to a state that can imprison them for 486 of them.

Both of those articles are complaints. Neither of them changes anything.

Then the good employer can find the good employee, and the good employee can find the good employer, and dishonest dealing stops being the profitable strategy it currently is.

That is what Rebuild India and ARISE are attempting. It is early, it is small, and it is free. It will work in proportion to how many entrepreneurs decide it is theirs.