As this dispatch is written, thousands of young Indians are camped at Jantar Mantar and in the centre of Delhi, refusing to leave. A day ago they tried to march on Parliament and were met with tear gas and batons, and scores were injured. What put them on the street was a leaked examination. The leaked paper is not the real grievance. The examination itself is: for millions of them it has become the one narrow gate to a stable life, and that gate has quietly been put up for sale.

Look at who is camped there. They are not the country’s poor and unschooled. They are its graduates and its aspirants, the young people who did everything the system asked of them. That is what makes the anger so combustible, and it shows up in the single number that has run through the coverage: youth unemployment above 29%. The figure needs reading with care. It is not the headline unemployment rate, which is closer to 4% or 5% in the Periodic Labour Force Survey. The 29% is the graduate youth unemployment rate: of every hundred young Indians who hold a college degree and are looking for work, close to thirty cannot find any. The gap between that and the headline rate is the whole story. In India today, a degree does not protect you from joblessness. It raises the odds of it.

This dispatch explains that anger through the data behind it. India makes graduates far faster than it makes jobs, turned a rigged examination into the only bridge between the two, and then acted surprised when the young found the bridge was broken and for sale.

Start with the finding that inverts everything. The Azim Premji University State of Working India 2026 report puts graduate unemployment among 15-to 25-year-olds in the labour force at nearly 40%, and at 20% for those aged 25 to 29. The number that matters most is not the level but the persistence. Graduate unemployment in India has hovered between 35% and 40% across the entire period from 1983 to 2023. This is not a post-Covid shock or an artificial intelligence panic. It is a forty-year-old structural fact that the size of the current youth cohort has finally made impossible to ignore.

The composition of the unemployed has flipped. The same report shows that graduates as a share of unemployed 20-to 29-year-olds rose from 13% in 1983 to 67% in 2023. In absolute terms, 11 million of India’s 63 million young graduates were unemployed as of 2023. This essentially means that joblessness in India has climbed the education ladder.

The person standing in front of a water cannon is not the school dropout. He is the first graduate his family ever produced.

The arithmetic behind this is brutal and simple. The State of Working India 2026 report finds that between 2004 and 2023, India added roughly 5 million graduates every year and created work for only about 2.8 million of them. Of those, fewer than 2 million found salaried jobs. The surplus does not disappear. It accumulates, it ages, and it waits.

What the surplus finds, when it finds anything, is not what the degree promised. Tracking young graduate men for a year from the moment they call themselves unemployed, the report finds that half of them find some work within twelve months. Only 6.7% find permanent salaried employment, and only 3.7% find white-collar work. The stable, respectable, salaried life that the degree was supposed to buy materialises for fewer than one in fourteen. Everyone else gets informal work, contract work, or the gig economy.

This is why the examination matters so much, and why leaking it is an act of violence rather than mischief. Every gate a poor or middle-class Indian is told to aim for clears a fraction of the people who show up.

Consider the medical entrance. 23.3 lakh candidates sat NEET-UG 2024 for roughly 1.08 lakh MBBS seats. Fewer than one in twenty who appear will become a doctor. The engineering gate is narrower still. Of the 48,248 candidates who qualified for JEE Advanced in 2024, the IITs had 17,385 seats, which means barely 1% of the larger JEE pool converts into an IIT seat. The civil services are the steepest climb of all. The Union Public Service Commission received 9.92 lakh applications for the 2024 Civil Services Examination and finally recommended 1,009 candidates against 1,129 vacancies. That is a selection rate close to one in a thousand.

The recruitment exams that the non-elite actually depend on are worse. The Railway Recruitment Board's 2024 NTPC notification drew 1.21 crore applicants for 11,558 vacancies, a success rate of 0.09%. Group D drew over a crore applicants for around 32,000 posts. The Uttar Pradesh Police constable recruitment of 2024 attracted close to 48 lakh applicants for 60,244 posts.

Then there is the overqualification, which says everything. In 2024, the Haryana government hired contractual sweepers on 15,000 rupees a month, and 3.95 lakh people applied, roughly 40,000 of them graduates and another 6,000 holding post-graduate degrees. In Maharashtra, 10 lakh applicants competed for 4,600 posts of talathi, the lowest rung of the revenue service, and among them were engineers, MBAs and PhD holders. In 2025, the railways received 1.08 crore applications for 32,438 Level-1 jobs. None of this is new. In 2015, the Uttar Pradesh secretariat advertised 368 peon posts, for which the qualification was a fifth-standard pass and the ability to ride a bicycle, and drew 23 lakh applications, 255 of them from people holding doctorates. This is not a labour market clearing supply against demand. It is a nation of qualified people fighting over crumbs, in which a doctorate is an overqualification for carrying files.

In a normal job market, many employers are hiring, so missing one opening is a small setback and you apply to the next one. The exam economy has no next one. The same few thousand seats are the only way out, and most aspirants get just two or three attempts before age limits close the door. So when a paper leaks, it is not petty cheating at a game. It is the theft of a young person's one real chance, and they can only stand and watch it happen.

The obvious objection is that these young people should simply try the private sector. The Lokniti-CSDS Youth Study of 2021 explains why they do not. 55% of Indian youth said they would prefer a government job, down from 65% in the 2016 round, with another 24% preferring to start their own business. The private sector trails both by a wide margin.

They are not being irrational. The government job offers the one thing the rest of the economy has stopped offering, which is security. The Periodic Labour Force Survey, the government's own count, puts the average pay of a regular salaried worker at 20,702 rupees a month in 2023-24. Nominal salaries have roughly doubled since 2012, but once inflation is taken out, the average salaried income was lower in 2024 than it had been a decade earlier, and real wages for rural workers in 2023-24 sat below where they were six years before. The reward for winning is shrinking. Against stagnant wages and insecure private work, a permanent government post with a pension is not a lazy aspiration. It is the last stable thing on offer.

And here the state does something that turns bad luck into betrayal. It refuses to fill the very jobs it has already created. The most recent comprehensive count, from the government's own Pay Research Unit, found 9.79 lakh of the 40.35 lakh sanctioned posts in central departments lying vacant, close to one in four. Add the states and the combined figure crosses 60 lakh. Even the elite services run hollow. As of January 2024, 1,316 of 6,858 sanctioned IAS posts and 586 of 5,055 IPS posts were lying vacant. Tens of millions compete in a lottery for a category of job that the government is simultaneously leaving empty by the lakh. The scarcity is not natural. It is manufactured by administrative neglect.

Now to the money, because this is where the betrayal becomes measurable in rupees taken from families that could not spare them. Coaching is no longer a supplement to Indian education. It has quietly become the education, a parallel private system that families are forced to buy because the public one has failed to prepare their children for the only test that counts.

The scale is large. The Indian test-preparation industry is estimated at 54,000 to 58,000 crore rupees, growing at 15% to 20% a year, with around 7.1 crore students enrolled. A machine this size exists only because the examination is both the sole ladder and one the school system does not equip you to climb.

Households now fund education privately because the state does not fund it publicly. Ministry of Statistics data show household spending on education rising 4.6 times, from 1.8 lakh crore rupees in FY12 to 8.43 lakh crore in FY24.

The National Sample Survey's Comprehensive Modular Survey on Education (2025) found that households spent an average of ₹22,024 per student annually on school education. Of this, private coaching accounted for about 40.7%, underscoring how shadow education has become a major component of schooling costs. Coaching nearly doubles the bill, taking annual per-student spending from 13,051 rupees without it to 22,024 rupees with it.

The same NSSO survey shows that coaching consumes 44.7% of a rural household's education spending against 35.3% for an urban one. The families with the least to spare surrender the largest proportion of it. This is the sharpest point in the entire story. For a poor family, coaching is the only substitute for private schooling, the connections, and the English-medium fluency that richer families already own. The coaching bill is a poverty tax dressed up as opportunity.

Nowhere is this clearer than Kota. At its peak, the town drew 2 to 2.5 lakh students a year and generated 6,500 to 7,000 crore rupees, sustaining around 4,000 hostels, more than 40,000 paying-guest accommodations, and over a lakh jobs. A single year there costs a family between 2.5 and 5 lakh rupees once coaching fees are added to hostel and living costs, more than many households earn in a year. The model has since spread to Sikar, Hyderabad, Patna, Prayagraj, and the UPSC quarters of Mukherjee Nagar and Old Rajinder Nagar in Delhi, where three aspirants drowned in a flooded basement library in July 2024. Families mortgage land and stake a year or two of earnings on a NEET seat with a one-in-twenty chance. When the paper leaks, they have not lost a test. They have lost a slice of the family's lifetime savings.

The money is captured at the top. Allen, Aakash, FIITJEE and Resonance charge between 75,000 and 2 lakh rupees a year for classroom coaching. The sector now produces financial giants. PhysicsWallah's revenue grew 2.6 times to 1,940 crore rupees in FY24, and it listed on the stock exchange in November 2025 at a valuation of around 3.19 billion dollars, or 31,500 crore rupees, even while posting a net loss of 1,131 crore in the same year. Allen took 600 million dollars from Bodhi Tree at a valuation near 1.2 billion dollars, and Aakash was bought by Byju's for 950 million before Ranjan Pai's Manipal group took control of it. Byju's itself, once valued at 22 billion dollars, collapsed entirely, and FIITJEE shut its centres overnight in early 2025 after collecting fees of 1 to 5 lakh rupees a student, with refunds beginning only in 2026. This proves the point that anxiety has been turned into an asset class. Venture capital now underwrites businesses whose addressable market is the desperation of the middle class. Some of those bets blew up, but the investor's loss was never the family's gain. The student carries all the downside risk. The platform keeps the valuation.

Stack these together, and the shape of the anger becomes economic, not merely emotional. A family in the bottom half of the income distribution spends a rising share of a falling real income on coaching, for an examination with a success rate between 0.1% and 5%, gated by a system that leaks papers to those who can pay for them. These families did not simply lose hope. They lost money they did not have.

The reason the system produces this outcome is that India never built a public university system. It licensed a private credential market. The State of Working India 2026 report records that the number of higher education institutions rose from 1,644 at liberalisation to 69,534, and that roughly 80% of them are private. Enrolment expanded and looked like progress. The All India Survey on Higher Education 2021-22 puts the Gross Enrolment Ratio in higher education at 28.4%, up from 23.7% in 2014-15. Access widened while quality and public accountability hollowed out underneath it.

The classrooms tell the story. The State of Working India 2026 report finds public colleges running at 47 students per teacher and private colleges at 28, against the norm of 15 to 20. The universities that are supposed to anchor the system cannot find teachers, or will not hire them. There were over 5,400 faculty vacancies in central universities in 2024, of which fewer than 15% had been filled by early 2025, and roughly 56% of professorial posts in top universities lay vacant. The state asks its young to trust institutions it refuses to staff, and then wonders why they trust coaching centres instead.

The under-investment runs from the classroom to the laboratory. India spends around 4% of GDP on education, against the 6% target first set by the Kothari Commission in 1968 and re-promised in the National Education Policy of 2020. It has never been met. The Economic Survey 2024-25 records gross expenditure on research and development at 0.64% of GDP, against China's 2.68%. Because there is no research economy to absorb the educated, every ambition funnels back to the same few salaried and government gates. The bottleneck at the examination is manufactured upstream, in the budget.

The examination system that all of this feeds is demonstrably corrupt. A widely cited compilation counts more than 70 confirmed paper leaks across 15 states in seven years, disrupting around 1.7 crore aspirants. The state’s own legislature effectively admitted the scale of the rot when Parliament passed the Public Examinations (Prevention of Unfair Means) Act in 2024, prescribing five to ten years in prison and a minimum fine of one crore rupees for organised leaks. You do not write a law like that for an isolated scandal.

The 2024 leak that started the current movement was not isolated. NEET-UG 2024 was sat by over 24 lakh candidates when the paper was compromised, triggering a national crisis and a Supreme Court intervention that declined a full retest and reworked grace marks instead. When the single gate to an entire profession is broken, there is no other gate to walk through. Rig the door and you rig the future behind it.

At the base of this machine, young people are dying. The National Crime Records Bureau’s Accidental Deaths and Suicides in India report for 2023, the latest available, records students as 8.1 percent of the country’s 171,418 suicides, which is close to 13,900 young lives, up from 7.6 percent in 2022. Failure in examination was cited in 1.4 percent of all suicides. Kota remains the symbol the country cannot look away from. When a system is this high-stakes and this corrupt at the same time, a leak is an insult laid on top of genuine grief. That combination of loss and injustice is what fills a square.

The anger also has a class geography, and it runs in two directions. The India Employment Report 2024 records that one in three Indian youth were not in employment, education, or training in 2022, and that this group is made up overwhelmingly of young women. The protest at Jantar Mantar is the visible tip of a much larger and quieter despair: the discouraged and the excluded who have stopped appearing in the unemployment count altogether.

At the other end, the families who can afford an exit have already taken it. The stock of Indian students studying abroad rose from around 6.8 lakh in 2019 to somewhere between 13 and 19 lakh by 2024, and Reserve Bank data show education-related outward remittances climbing steeply, with roughly 1.76 lakh crore rupees sent abroad over the decade.

The poor fight over 368 clerical posts while the affluent buy their way out of the Indian system entirely. The examination economy is a trap only for those who cannot afford the door marked exit.

The clock on all of this is running out. The State of Working India 2026 report notes that India’s demographic dividend, the phase in which the working-age share of the population keeps rising, is expected to peak around 2030. There are 367 million youth today, of whom 263 million are already out of education and in the potential workforce. The window to turn that cohort into an economic dividend is narrow, and it is closing.

If the story so far is a system that cannot absorb the educated, the last part is worse. It is a state actively defunding the ladders that let the poor and minorities climb.

The University Grants Commission is the body that funds fellowships, scholarships and ordinary colleges. The interim Union Budget for 2024-25 cut its allocation from 6,409 crore rupees to 2,500 crore, a fall of around 61 percent and its lowest in four years, while overall higher-education grants fell by about 9,600 crore. In the same budget, the grant for central universities rose 28 percent to 15,472 crore. The money was not only cut. It was moved upward, out of the dispersed channel that reaches small colleges and scholarship-holders and into a handful of elite institutions.

The stipend that keeps most research scholars alive has been frozen for eighteen years. The UGC Non-NET fellowship, which pays a PhD scholar 8,000 rupees a month and an MPhil student 5,000, has not been revised since 2006. Over the same period the elite, NET-qualified Junior and Senior Research Fellowships were raised to 37,000 and 42,000 rupees. The mass of research scholars, drawn disproportionately from non-elite backgrounds, has watched its stipend lose most of its real value to inflation while the credentialed few at the top got raises. The two-tier logic of the examination runs straight through the research system.

The fellowships that do exist arrive late, when they arrive at all. Through 2024, junior research fellows reported going unpaid for months while disbursements stalled on portal errors and pending approvals, and the All India Research Scholars Association ran sustained protests. The people who did everything right, cleared the national test and enrolled for a doctorate, were left without a stipend for months.

For minority scholars, a whole ladder was removed. The Maulana Azad National Fellowship, launched in 2009, paid up to 25,000 rupees a month for five years to MPhil and PhD scholars from six notified minority communities. It was discontinued from 2022-23 on the stated ground that it overlapped other schemes, and the government has told Parliament there is no proposal to restore it. Below it, the pre-matric minority scholarship was cut from 1,425 crore rupees in 2022-23 to 433 crore in 2023-24, part of a 38 percent fall in the Ministry of Minority Affairs budget. The scholarships that keep the poorest and first-generation students inside the pipeline are being withdrawn at the very moment the private cost of staying in it is rising.

And the country that starves its scholars produces them anyway. India runs on 262 researchers per million people, a figure from the Department of Science and Technology’s own statistics for 2020, against China’s 1,849 and the United States’ 4,825. PhD enrolment rose 81 percent to 2.13 lakh by 2021-22. India mints more researchers each year and funds neither their stipends nor the research economy that might employ them. The under-investment is not an accident of one budget. It is the settled shape of the thing.

A young Indian is told from childhood that education is the escape, and the data says education now raises his odds of being unemployed. He is funnelled toward an examination that clears one in a thousand, because the state has left the private economy insecure and its own posts unfilled. His family pays a coaching industry a poverty tax for the privilege of trying. The classroom that was supposed to prepare him is running at 47 students per teacher, the scholarship that was supposed to carry him has been cut, and the stipend that was supposed to sustain his research has not moved since 2006. Then the one thing sold to him as fair, the examination, turns out to be for sale.

MoSPI, the ministry that runs India's official statistics, has put its databases behind a ‘connector’ that plugs straight into Claude. The big household consumption survey, the jobs survey, the national accounts and the price indices are now available in the same window of Claude where you chat with the AI tool, ready to be questioned instead of painstakingly extracted from PDF tables.

For anyone who works with Indian data, that removes the dullest part of the job: digging numbers out of PDF tables by hand. This essay came out of one such sprint of data scraping. I pulled the consumption survey to check a claim I had read too many times to keep believing, and the numbers took me somewhere I did not expect.

The claim

The claim is that India is getting more equal. The government’s statisticians brought back the big household consumption survey in 2022-23 after an eleven-year gap, and ran it again in 2023-24. Both rounds say the same thing.

People are spending more on average, the gap between town and village has shrunk, and the standard measure of the spending gap between rich and poor has fallen. SBI Research read the first round as a sign that poverty was down to low single digits. The economist Surjit Bhalla has argued for years that Indian poverty has all but vanished. The reading has hardened into common sense: a smaller spending gap means growth has finally reached the people at the bottom.

What the survey says when you line people up

Pull the same data apart by group, and it changes shape. I lined up rural households from the poorest five percent to the richest five percent and calculated how much each group's monthly per-capita consumption expenditure (MPCE) changed between the 2022–23 and 2023–24 Household Consumption Expenditure Surveys after adjusting for rural inflation. The figures that follow are my calculations from the published HCES datasets.

Prices in the countryside rose about 5.5 percent across the period. The poorest five percent raised their real spending by 15.8 percent in a single year. The richest five percent saw theirs fall by about 8 percent. Every group in between grew more slowly than the group just below it, in a line that barely wobbles from the bottom of the ladder to the top.

The poor are catching up, the rich are not running away, and the gap is closing on its own. Before accepting it, one question has to be answered, because the survey itself cannot answer it.

A poor family can spend more for two very different reasons. It can spend more because it earned more, which means it climbed. Or it can spend more because someone handed it more, which means it was helped. The survey records what people spent. It does not record where the money came from.

This essentially means the survey cannot, by itself, tell help apart from earning. Everything that matters here turns on that difference.

What work paid

The question from the last section was simple: did the poorest spend more because they earned more?

The wage data suggest otherwise.

Most poor rural households depend on casual labour, paid by the day. According to the Periodic Labour Force Survey, the average daily wage for rural casual workers rose from about ₹382 in 2022 to around ₹413 in 2024. That looks like progress at first glance. But prices also rose during those two years. Once inflation is taken into account, the purchasing power of a day’s wage barely changed. By 2025, after inflation had eased, real wages had only just returned to roughly where they had started.

Four years had passed, but a day's work bought little more than it did before.

Now compare that with the consumption figures.

The poorest rural households increased their real spending by around 16 percent in a single year. Yet the real wages they depend on barely moved.

That does not prove that none of the extra spending came from higher earnings. But wages alone cannot explain an increase of that size. Something else must have helped households spend more.

Economist Himanshu has been making a similar argument for years. Rural wage growth slowed sharply during the second half of the 2010s, especially for agricultural workers. The latest labour market data suggest that the pattern has continued into the 2020s. For many poor households, wages are no longer rising fast enough to explain improvements in living standards. If higher wages were not the main driver of higher consumption, two possibilities remain. Families either received additional support or they financed more spending through borrowing.

The next section looks at both.

The given and the borrowed

Start with what was given. The clearest case is grain. Since January 2023, the government has given roughly 800 million people five kilograms of rice or wheat each month, free, where before they paid a few rupees a kilo.

The survey captures part of this story. It assigns an estimated value to many goods that households receive free through government welfare programmes and counts them as part of household consumption. The survey itself shows how much government transfers matter. In 2023–24, the average rural person's monthly consumption expenditure was ₹4,122. Once MoSPI added the estimated value of goods received free through welfare programmes, the figure rose to ₹4,247, an increase of ₹125 per person per month. That additional consumption did not come from higher earnings. It came from goods provided by the state.

A free ration of five kilograms of grain has a very different value depending on who receives it. For a poor household, it makes up a meaningful share of the monthly food budget. For a richer household, it is worth much less in proportion to what they already spend. Assuming a market value of about ₹30 per kilogram, the ration is worth around eight percent of the monthly consumption of the poorest rural households but less than two percent for the richest. The subsidy is the same in rupees, but not in its impact.

The difference is even greater because many higher-income households fall outside the National Food Security Act and receive no subsidised grain. As a result, the programme benefits the bottom of the consumption distribution far more than the top. That alone can reduce measured consumption inequality, even if households’ labour incomes remain unchanged.

Free grain is only one form of support. Over the past few years, many state governments have also begun sending cash directly to women through schemes such as Ladli Behna, Majhi Ladki Bahin and Gruha Lakshmi. For millions of households, these transfers now supplement wages every month. They may not appear in wage statistics, but they increase what families can afford to consume. That is another reason why higher consumption does not necessarily mean higher earnings.

Borrowing may also explain part of the rise in consumption. During the survey years, households were saving much less than before. Net household financial savings fell to about five percent of GDP in 2022–23, one of the lowest levels seen in several decades. At the same time, household debt climbed to around 41 percent of GDP. The Reserve Bank has noted a steady rise in unsecured and consumption-related borrowing during this period. Some households, in other words, were financing today's spending with tomorrow's income.

Here, it is important to qualify that borrowing was only part of the story. By 2024–25, household savings had begun to recover, and the rise in debt had slowed. Much of the borrowing during the survey years appears to have been a response to a difficult period rather than a permanent change in household finances. Some of it also went towards buying homes or other assets rather than day-to-day consumption.

Government transfers are different. They are not temporary. Free food grain reaches around 800 million people every month under the National Food Security Act, and many states are expanding direct cash transfers. These programmes have become a permanent feature of India’s welfare system.

This is not an argument against welfare. Feeding people and protecting them from hardship are achievements in their own right. But they should not be confused with rising incomes or greater economic mobility. A floor that is supported by public transfers is not the same as a floor that is rising because people are earning more.

The next question, then, is whether that higher floor is helping people climb.

Does anyone climb?

A raised floor is real. It is also not a movement. The whole question of climbing is whether the people standing on that floor can leave it. So change the question. Stop asking how much the poorest spend, and start asking whether a child born at the bottom can expect to end up anywhere else.

Begin with the rung that is supposed to do the lifting, which is education. School is the thing every climbing story relies on, the machine meant to turn a poor childhood into a richer adulthood. In India today, it runs backwards. The more schooling a person has, the more likely they are to be out of work. Among people who never learned to read, measured unemployment is almost nothing, about 0.2 percent. Among graduates, it is 13 percent. Among postgraduates, 12 percent. Each step up the schooling ladder raises, rather than lowers, the odds of sitting at home without a job.

At first, this looks like a paradox. Why should unemployment be lower among people with less education than among graduates?

The answer is that the poorest cannot afford to be unemployed. They take whatever work they can find, however insecure or poorly paid, because they have no savings to fall back on. That keeps measured unemployment low but hides a great deal of informal and low-quality work.

Graduates are in a different position. Many can afford to wait longer in the hope of finding a better job that matches their education. Their higher unemployment is therefore not simply a sign of idleness. It reflects the gap between what a degree promises and what the labour market actually offers.

For the young the gap is wider. Among everyone aged fifteen to twenty-nine, even on the kindest measure, about one in ten who want work cannot find it, three times the figure for the population as a whole. The State of Working India 2026 report puts the rate for young graduates far higher, and finds that only a small share of them land a steady salaried job within a year of leaving college. A generation was told to stay in school longer than their parents had. It did. The jobs that were supposed to be waiting were not.

Go up one level, from the single career to the whole family line, and the story holds. The most careful study of climbing across generations in India, by the economists Sam Asher, Paul Novosad and Charlie Rafkin, tracks how far children get in school compared with their parents. It finds that a child born into the bottom half can expect to rise very little above where the parents stood, that this has not improved in the decades since the economy was opened up in 1991, and that for some groups, Muslims above all, it has gone backwards. Dalits have closed part of the old gap. The broad bottom has not. The floor of the handouts has been raised, but it is not a launch pad. People stand on it. They mostly do not climb off it.

There is, however, an important counterargument, one associated with Amartya Sen’s capability approach. Welfare transfers do more than raise consumption today. A child who eats better is more likely to stay healthy and learn well. Cash in a mother’s hands can mean better nutrition, schooling and healthcare. From this perspective, today’s welfare spending is an investment in tomorrow’s worker. The gains may simply take years to appear.

That is an important point. But it depends on one condition: those better-educated children must eventually find better jobs. That is where India’s problem lies.

India is not short of schools. It is short of good jobs. As the earlier figures showed, even graduates face high unemployment. If today’s graduates struggle to turn education into better livelihoods, there is little reason to assume that better-fed graduates twenty years from now will automatically fare differently.

This does not diminish the value of welfare. Free grain and cash transfers improve lives today, and that is reason enough to support them. But welfare alone cannot create economic mobility. It can prepare people to climb only if there is a ladder waiting at the top.

The floor and the ladder

The state has learned one thing about itself. Holding a floor up is easy. Free grain can be shipped out in an afternoon and sent out along a ration network that reaches almost every village. Cash can be dropped into a hundred million bank accounts at the press of a key. Building a ladder is hard. A market full of decent jobs has to be grown slowly, out of investment and skills and firms, and India has not grown one fast enough to absorb the people its schools keep turning out. Handed an easy tool and a hard one, the state reaches for the easy one. It holds the floor.

This is not a trick played on the poor, and it is not vote-buying dressed up. The handouts feed people who would otherwise go hungry, and that is worth doing. The danger is not the handouts. The danger is what they have been allowed to stand in for. When a smaller spending gap gets read as the poor climbing, the pressure comes off everything else. A government that can point to a closing gap has a reason not to ask why its graduates cannot find work, why a day’s labour buys no more than it did four years ago, why a child born at the bottom is still, a generation later, found at the bottom. A sign of progress quietly becomes a substitute for it.

And the floor is only as steady as the budget beneath it. Incomes were never made equal. The ladder was never built. What was made equal was the plate, and the plate is refilled each month by a decision that can be unmade. Take the money away, and the gap reopens, because nothing underneath it has changed.

India has learned how to stop its people from falling. That is no small thing. But stopping a fall is not the same as helping someone rise, and rising is the part the country has not managed, and may have stopped trying to manage, because falling is so much easier to prevent and the figures look so much better when it is. The floor is real. The people standing on it are not moving. A country that mistakes the one for the other will keep calling the absence of disaster a kind of justice, for as long as the figures let it.

After fourteen years working in and around the development sector, across non-profits, government partnerships, and policy research ecosystem, I came to a fairly uncomfortable conclusion. Most of what the sector calls development is not really development. It is the management of poverty rather than the end of it. You help people cross a threshold, report the number, and call it success. The economist Lant Pritchett has a name for this. He calls it kinky development.

For a long time, I thought kinky development was a good enough diagnosis for almost everything I had seen. Then I started paying closer attention to what India's state and the development sector are actually doing, and found something that does not quite fit.

Pritchett's argument is about how aid bureaucracies work. Development organisations survive on measurement, and measurement shaped by funding produces a very specific logic: optimise for the thing you can count, and downweight the things you cannot. A microfinance organisation gets renewed because its borrowers repaid loans, not because those borrowers built businesses. A women's livelihoods program gets celebrated because SHG membership doubled, not because wages in the district rose. An education program wins an award because test scores improved over the control group, not because the labour market gave students any reason to stay in school.

The funder-non-profit complex in India runs on this. The machine works as it does because decades of bilateral aid and philanthropic grant-making have made kinky outcomes the currency of survival. You report what you can defend. You can only defend what your funders recognise as progress. And progress, for a board with no patience for thirty-year timelines, means things that move within a three-year project cycle.

This is why the same pattern repeats across the sector. The enrolment drives, the self-help group membership counts, the form-filling marathons at block offices. No one in the funding chain is required to ask what the data points add up to: whether the woman’s income changed, whether her daughter will have different choices, whether the block has more productive businesses five years later. Those questions live in a different column, one nobody is paid to fill.

I read India through this lens for years. Then I noticed something the kinky development frame does not quite fit.

India’s welfare architecture is not a donor project. The schemes that reach the furthest corners of rural Chhattisgarh were not designed in Washington or Geneva. They were designed by state finance secretaries, passed in legislative assemblies, and launched before election cycles. The incentive driving threshold targeting here is not a three-year donor project cycle. It is a five-year electoral cycle. And unlike a misaligned donor incentive, which can be corrected by reforming the aid architecture, a democratic incentive cannot be reformed from outside. It can only be changed by the voters who created it. The machine generating kinky outcomes in India is not a foreign aid bureaucracy optimising for its own survival. It is a democracy optimising for its own.

That is a different problem. It requires a different name.

I want to call what India now practices Threshold Statecraft.

Kinky development, as Pritchett describes it, happens to countries. It is the unintended result of misaligned donor incentives imported from outside. You fix it by reforming the aid architecture, by changing how donors measure success, by making structural outcomes fundable. These are solvable problems, at least in principle.

Threshold Statecraft is not a pathology from outside. It is a choice from inside. It is a political technology, refined across many election cycles, for producing visible evidence of transformation without doing the institutional work that transformation actually requires.

Take the Lakhpati Didi scheme, which the government has positioned as its flagship program for rural women's economic empowerment. The scheme identifies Self-Help Group members who earn at least one lakh rupees a year through SHG-linked businesses, certifies them as Lakhpati Didis, and counts them as milestones. The original target was three crore such women. By early 2025, the government declared that target achieved ahead of schedule. A new target of six crore has been set for 2029-30.

The headline looks clean. But step back and look at the full picture. India has roughly ten crore women enrolled in SHGs today. Three crore of them have crossed the one lakh annual income threshold. The other seven crore have not. The scheme was declared successful when roughly 30 percent of the enrolled population had reached the milestone. The 70 percent who remain below it were folded quietly into the next target. The political logic of declaration is not subtle: you announce success when the number is large enough to be impressive, regardless of what happened to everyone else.

The women below the threshold are mostly doing what development economists call survival entrepreneurship: selling vegetables, making pickles, running small credit cycles within their SHGs. These activities matter. They provide a buffer against the worst poverty and give women a bit more negotiating power at home. But the gains stop there. They do not accumulate, they do not scale, and they do not produce the kind of business that generates a tax liability or hires a permanent worker.

Female labour force participation has risen from 23.3 percent in 2017-18 to 41.7 percent in 2023-24, a number the government cites frequently and correctly. But the Periodic Labour Force Survey shows that rural female FLFP remains overwhelmingly in agriculture and informal self-employment. The factory jobs that South Korea and China used in the 1970s and 1980s to build their female working classes, jobs that came with formal wages, provident fund contributions, and real upward mobility, have not appeared at the scale required. Manufacturing’s share of GDP fell from around 16 percent in 2011-12 to 12.5 percent in 2024. Its share of total employment fell from 12.8 percent in 2011-12 to 11.4 percent in 2023-24, even as the total workforce grew by tens of millions. The SHG economy is filling the space that industrialisation was supposed to fill.

Ladli Behna Yojana, the monthly cash transfer scheme launched by Madhya Pradesh in 2023, makes the logic of Threshold Statecraft even more visible, because it drops the fiction of enterprise altogether. Eligible women receive a direct monthly transfer of fifteen hundred rupees. Madhya Pradesh spent close to nineteen thousand crore rupees on the scheme in its first full year. According to PRS Legislative Research, 12 states are estimated to spend about ₹1.68 lakh crore in 2025–26 on unconditional cash transfer schemes for women. Six of those twelve states are running projected revenue deficits. The Reserve Bank of India has warned that the rapid growth of unconditional cash transfers is crowding out capital spending on the productive infrastructure that would reduce welfare dependency in the first place.

The evidence on direct cash transfers is generally strong. Reviews of randomized evaluations synthesized by J-PAL find that cash transfers consistently improve household consumption, food security, and economic well-being. In India, research by Karthik Muralidharan, Paul Niehaus, and Sandip Sukhtankar finds that direct cash transfers can effectively support household welfare while reducing delivery inefficiencies associated with traditional transfer systems. Women with more liquidity have better bargaining power at home. Children in recipient households show improved nutrition and school attendance. These are real gains, paid to real people, and they should not be dismissed.

But the politics of the transfer actively works against ever making it unnecessary. In Madhya Pradesh, the BJP won the 2023 state election, defying pre-poll projections that had read the anti-incumbency mood as decisive. The Ladli Behna scheme, with over one crore women enrolled well before polling day, was credited as the decisive factor across party lines. A constituency had been created whose material welfare was directly tied to the current government staying in power. The fiscal commitment this creates crowds out the capital spending that would otherwise go into industrial infrastructure, skill development, and labour market reform. States spending most aggressively on cash transfers are, in the same budget cycles, cutting allocations to roads, industrial parks, and polytechnic colleges.

This is the democratic loop that makes Threshold Statecraft a stable equilibrium rather than a transitional phase. India's informal economy employs roughly 90 percent of the workforce. That creates an electoral constituency with very rational preferences for immediate, tangible, cashable welfare. This constituency is too large to ignore and too precarious to wait for thirty-year industrial policy timelines to deliver. The politician who promises an industrial corridor in Vidarbha gets credit in fifteen years, perhaps. The politician who deposits money into Jan Dhan accounts before Diwali gets credit before November. The calculation is not irrational. It is a coherent response to the political environment that mass informal employment creates.

The cruelty of the loop is that it reinforces itself. Welfare transfers reduce the fiscal space for public investment that would decrease welfare dependency over time. The political bandwidth consumed by scheme management, enrollment drives, Aadhaar seeding, and beneficiary verification, leaves little room for the structural reforms that take years to show results and attract fierce opposition. Land reform, labour law simplification, the industrial policy required to make India a serious manufacturing destination: each of these produces diffuse gains and concentrated, vocal, well-organised opposition. The cash transfer produces diffuse costs and concentrated, visible, electorally useful gains.

Kinky development happens when aid bureaucracies accidentally optimise for the wrong thing. Threshold Statecraft happens when democratic governments deliberately optimise for the right thing, the thing that wins elections, and the two happen to coincide.

This is not an argument against welfare programs. The gains from cash transfers and SHG credit are real. The women receiving Ladli Behna transfers have somewhat better lives because of them. The Lakhpati Didi who crossed the one lakh threshold is not a fiction, and her crossing it matters.

The harder point runs in a different direction. A state that builds elaborate delivery machinery for SHG credit but cannot simplify its labour codes, that can track ten crore women through a direct benefit transfer system but cannot run a single-window industrial clearance process, has made a precise bet on where its comparative advantage lies. That bet is not irrational. Building the DBT infrastructure was genuinely hard. The Aadhaar-linked direct benefit transfer system has few parallels anywhere in the developing world.

But it has produced an administrative culture built for welfare distribution rather than for the institutional reforms that would reduce the demand for welfare in the first place. India has built genuine functional capacity and deployed it in service of a political logic that makes further transformation progressively harder. The state is not pretending to develop. It is developing, but along the dimension that reinforces the current equilibrium rather than disrupting it.

India’s Economic Survey of 2023-24 estimated that the country needs to create 78.5 lakh non-farm jobs annually until 2030 just to absorb new workforce entrants. The SHG economy, for all its scale, does not produce those jobs. Neither do cash transfers. Both are necessary as consumption floors. Neither is a substitute for the manufacturing and services density that would generate formal wages for the 78.5 lakh people entering the labour market every year.

A state that has mastered the threshold has learned something real about governance. It has not yet learned how to make itself unnecessary.

The question Pritchett’s framework opens but cannot close is this: can a democracy that has found a stable equilibrium at the threshold choose, against its own political logic, to push past it? The answer cannot come from development economics alone. It has to come from politics. That means it depends on whether the cost of staying at the threshold, in foregone productivity, in premature fiscal exhaustion, in the 78.5 lakh non-farm jobs that need to exist by 2030 but do not, eventually becomes legible to voters before it becomes unmanageable for the state.

That is a much harder problem than managing a threshold. It is also the only one that matters.

I built the map I always wanted

I write Gentle Leviathan, and over the years almost every essay has sent me back to the same handful of works like Smith, Marx, Keynes, Hayek, Rawls, Sen. I’d reread a paper to check a claim, trace an idea to its source, or argue with it on the page. After enough of these trips, one thing became obvious: these works live everywhere and nowhere. Scattered across decades, journals, and sources. There was no single place to find them, let alone see how they fit together.

So I built one.

The Canon brings together the 115 works that shaped economics and political philosophy and puts them under a single roof. Each one gets a plain-English summary, a difficulty rating so you know what you’re walking into, and a link to the source. No jargon, no gatekeeping.

And it’s more than a list it’s a map. Every work is connected to the ones it builds on, answers, or argues with, so you can actually see how the ideas relate: Nozick replying to Rawls, the social contract running from Hobbes to Kant, Hayek and his critics sitting side by side.

This is the resource I wish I’d had when I started writing. It’s free, and it’s here.

I finished J. Bradford DeLong’s Slouching Towards Utopia this week, and I have not been able to read the news the same way since. Not because the book made me optimistic. Because it handed me a mental model, and once you have the model, the current chaos stops looking random. The trade wars, the AI gold rush, the political anger spreading across democracies and all the obituaries of liberalism, they stop looking like separate crises and start looking like one crisis, playing out in slow motion, chapter by chapter, exactly as DeLong said it would.

That is what these books do. The grand narrative, the long view, the attempt to hold 140 years of economic history inside a single argument. I keep returning to this genre because nothing else clears the fog the way it does. I have been reading books like this for years. They are my weakness and my education. They do not answer specific questions. They give you a way of seeing, and once you see this way, you cannot unsee it. DeLong, a Berkeley economic historian who spent three decades on this book, may have written the clearest version of this kind of argument that I have read in a long time.

His central claim is simple and slightly shocking. 1870 is the most consequential year in human history since the invention of agriculture. Not 1914; not 1945; not 1989.

Before that year, the global economy grew at roughly 0.45% per year.

To understand what that number actually means, consider what it implies for a human life. At 0.45% annual growth, a child born in 1700 would die in conditions almost indistinguishable from those their grandparents had known. Material life barely changed across generations. Poverty was not a problem awaiting a solution. It was the default condition of human existence, the stable background against which everything else happened. The Malthusian trap, where any productivity gains were swallowed by population growth, had held humanity in place for millennia. Progress was not something ordinary people expected, because they had never had reason to expect it.

Then, in the span of a few decades, three things changed at once. And the world was never the same.

The first was the industrial research laboratory. Before this, invention was largely accidental, individual geniuses in workshops, lucky insights, breakthroughs that nobody had systematically set out to find. What the research laboratory did was turn invention into a process. Edison’s Menlo Park was the early model. Then came Bell Labs, the German chemical companies, the pharmaceutical research divisions of the early twentieth century. These were factories for discovery. Invention stopped being something that happened to individuals and started being something that organisations could produce on schedule. The pace at which new technologies appeared, and then diffused through economies, accelerated sharply and stayed accelerated.

The second was the modern corporation. For most of human history, businesses were small, local, and constrained by the limits of what a family or partnership could manage. The corporation smashed that constraint. It could aggregate capital from thousands of investors, employ tens of thousands of workers, operate across continents, and outlast its founders by generations. It was an organisational technology as much as an economic one. And it gave the research laboratory’s discoveries a vehicle to reach the world at scale.

The third was the collapse in the cost of moving goods and information across the planet. Steamships cut ocean freight costs dramatically. Telegraphs made coordination across distances possible for the first time. Railways tied together what had been isolated local economies. A genuinely global market came into existence, where prices, labour, goods, and ideas moved at speeds and volumes that would have been unimaginable a century earlier.

Put these three together and global growth jumped from 0.45% to roughly 2.1% per year. Sustained over 140 years, that shift is transformative in a way that is hard to fully absorb. By 2010, average global income per capita was approximately 8.8 times what it had been in 1870. Nearly nine times richer, on average, across the whole world. No civilisation before had come anywhere near this. No prior generation in all of human history had experienced anything like it.

Now hold that number in your head. Because this is where the book gets interesting, and uncomfortable.

If the twentieth century was this economically productive; if it made the world nine times richer, why did it also produce two world wars, the Holocaust, the Gulag, the Great Depression, Mao’s famines, and a distribution of all that new wealth so unequal that mass poverty persisted long after the tools to end it existed? Why did humanity get nine times richer and still ended up in a mess which we are currently in?

This is the central puzzle of the book. DeLong does not let you dismiss it.

He makes you sit with it, and then he offers an answer that runs through two thinkers whose ideas were, in his reading, the animating conflict of the entire era.

Friedrich Hayek, on one side. The Austrian economist and philosopher who believed, with genuine conviction, that free markets are the most powerful coordination mechanism humanity has ever built. The price system aggregates information that no central planner could ever gather or process. When prices move freely, they direct resources toward where they are most needed, without anyone having to issue a command or understand the whole picture. Any serious interference with this system does not just slow growth. It lays the groundwork for authoritarianism, because planning requires coercion, and coercion requires concentrated power. Hayek had watched fascism and Stalinism emerge in real time. He thought he understood the mechanism. The title of his most famous book says it plainly: the road from intervention to serfdom is shorter than anyone wants to admit, and it is paved with good intentions.

Karl Polanyi, on the other side. A Hungarian-born historian who published The Great Transformation in 1944, the same year Hayek published The Road to Serfdom. Two books, the same moment, opposite conclusions. Polanyi’s argument was this: markets are not natural. They do not arise from human nature or emerge spontaneously. They are constructed by governments, through specific legal choices and institutional arrangements that are always political and always contestable. And when markets expand without limit, they commodify people. They dissolve communities, strip away job security, eliminate the sense of belonging that human beings need to function. They produce not just inequality but social disintegration. Polanyi called the predictable political reaction the “double movement”: markets expand, and society organises to protect itself. Through unions, welfare states, minimum wage laws, social insurance schemes. Or, when those channels are blocked, through much more dangerous forms of politics.

The twentieth century, in DeLong's reading, was the large-scale, decades-long clash between these two worldviews played out across the globe. And the lesson it teaches is brutal in its clarity. When Hayek won completely, meaning when markets had free rein with no counterweight, the world got the speculative excess of the 1920s, the Great Depression of the 1930s, and the social fragility that contributed to the catastrophe of 1914. When Polanyi’s position was taken to its extreme, when markets were not just regulated but abolished entirely, the world got Soviet collectivisation, Mao’s Great Leap Forward, famines that killed tens of millions, and political systems built on systematic terror. Neither side, unchecked, was survivable.

The best outcome the century produced came when both sides were forced to live with each other.

The post-war North Atlantic settlement, from roughly the late 1940s through to the mid-1970s, was built on exactly this forced coexistence. Governments kept markets. They also built welfare states, regulated banks, used Keynesian demand management to prevent catastrophic economic collapses, and constructed systems of social insurance that gave workers a floor below which they could not fall.

DeLong calls this a “shotgun marriage” between Hayek and Polanyi, blessed by Keynes. Neither side chose the arrangement. Both were stuck with it, and the arrangement worked. The result was three decades of growth that was fast, broadly shared, and politically stable. Real wages rose across most of the income distribution. Unemployment stayed low. The institutions of international economic cooperation, the IMF, the World Bank, what eventually became the WTO, tried to extend this managed capitalism to the global level. It was, by any historical comparison, remarkable.

It had limits, and they matter. The settlement was built primarily for white male workers in wealthy countries. In the United States, much of the New Deal architecture was structured to exclude Black Americans. Women entered the formal economy on terms the system had never been designed to accommodate. The decolonising world was largely an afterthought to negotiations it had no part in shaping. The miracle was real. But it was partial, and the people left outside it were accumulating grievances that would eventually find expression.

The cracks showed in the 1970s. Stagflation, slow growth and high inflation appearing together, exposed genuine weaknesses in the Keynesian approach and handed ammunition to those who had always wanted to dismantle the post-war arrangement. Some reform was warranted. What came instead was not reform.

The neoliberal turn under Thatcher and Reagan was a different project. It did not fix the post-war model. It dismantled it. Labour protections were stripped out. Financial markets were deregulated. Capital was freed to move across borders in ways labour never was, and never would be. Tax systems were restructured in ways that shifted after-tax incomes sharply upward. The argument, offered sincerely by some, cynically by others, was that faster growth would benefit everyone. Growth did continue, in most places. The broad distribution did not follow.

What the neoliberal era actually constructed was a second Gilded Age. Wealth concentrated at the top at levels not seen since the 1920s. Median wages in most Western economies stagnated even as productivity kept rising. Workers were producing more and taking home roughly the same. The institutions, communities, and social bonds that the post-war settlement had sustained hollowed out. DeLong does not soften his verdict on this. Neoliberalism promised shared prosperity and delivered concentrated wealth. It made the rich considerably richer and left most people competing for a share of growth that was barely moving for them.

Then came 2008. The financial architecture built on three decades of deregulation just collapsed. And the political response, austerity across most of the Western world, compounded the disaster, sacrificing the recovery of ordinary households to protect the interests of creditors.

The slow, uneven recovery deepened the inequalities that had been building for thirty years. By 2010, when DeLong closes his narrative, the postwar consensus was gone, the neoliberal model was discredited but still structurally in place, and nothing coherent had emerged to replace either. The political anger that became Brexit, Trump and other illiberal political forces across the world were already forming underground, looking for a container that would hold it.

Reading all of this in 2026 is a specific kind of uncomfortable experience. DeLong ends at 2010, but the story has continued at pace and it has continued exactly as his framework would predict.

The US tariff regime is the most aggressive attack on global trade since the 1930s. DeLong identifies cheap, globalised trade as essential to post-1870 growth, not a policy preference, but a structural pillar of the modern economy. It was one of three foundations that made the world nine times richer than it was a century ago. What is happening now is not a technical disagreement between economists. It is a direct dismantling of an institutional architecture that made prosperity possible. When bets like this have been placed before, when nations have turned inward and fragmented global trade, the outcomes have been severe. The costs fall hardest on those with the least power to influence the decision.

The other pressure on DeLong’s framework is artificial intelligence. The scale of investment and expectation around AI invites real comparison to the emergence of the industrial research laboratory in the 1870s. What the laboratory did was make invention systematic, converting discovery from a series of lucky accidents into a repeatable, scalable process. If AI does something similar for scientific research broadly, compressing decades of discovery into years across medicine, materials science, and energy, we may be at another genuine inflection point in the history of human productive capacity.

But DeLong’s framework asks the right question about this. Technology, in his account, is never the whole story. The post-1870 growth acceleration generated enormous material wealth. What it did not automatically generate was the institutions to distribute that wealth. Those required political struggle, social settlements, and hard choices, and when the politics went wrong, the costs were catastrophic. The question about AI is not whether it will increase output. It probably will. The question is whether the political systems of the 2020s have the capacity and the will to build institutions that ensure those gains are distributed broadly, rather than captured by whoever owns the infrastructure. On current form, that question does not have a reassuring answer.

DeLong does not pretend it does. The word in the title is “slouch.” Not a march. Not a sprint. Not a collapse. A slouch, ungainly, intermittent, without clear direction. The material conditions for a genuinely good life for most people on earth have existed for decades. The political and institutional conditions to translate that material wealth into actual human flourishing have not kept pace. They have never kept pace. The gap between what is materially possible and what is politically achievable is, DeLong implies, the defining problem of this century as much as the last.

By the late 2010s, the Chennai automobile corridor had emerged as one of the largest auto manufacturing clusters in Asia, accounting for roughly one-third of India’s passenger vehicle production. Around Chennai, the Sriperumbudur–Oragadam corridor evolved into a dense manufacturing ecosystem where Hyundai, Ford, BMW, Renault-Nissan and hundreds of component suppliers clustered within the same industrial geography.

Ask most economists why global manufacturers chose this corner of India over Pune, over Ahmedabad, over the vast industrial heartland of the Gangetic plain, and they will reach for the predictable answers: coastal access, existing port infrastructure, a government that processes approvals without ransacking investors. These things matter. They explain the geography of the bet. But they do not explain what made the bet winnable.

The deeper answer begins in 1956.

That year, K. Kamaraj, Chief Minister of Madras State, expanded and institutionalized a school meal programme that had first been experimented with by the Madras Corporation in the 1920s. Kamaraj was not working from a theory of human capital accumulation. He was solving an immediate, ground-level problem: children from poor families were not attending school because their households could not spare them for a full day without a midday meal. The schools existed; attendance was low; the children were somewhere else. So he fed them.

What he could not have known was that this single administrative decision would set off a chain of consequences still visible in Tamil Nadu’s GDP data sixty years later. He had placed, without knowing it, the opening move of a 40-year development bet whose returns would not be measurable for a generation.

The standard theory of economic development, the model that has dominated the field since Arthur Lewis wrote it down in 1954, tells a clean sequential story. Economies begin with a large agricultural sector and a small modern sector that offers higher wages. Surplus agricultural labour migrates to factories. Productivity rises. The economy industrializes. Eventually, when wages are high enough, manufacturing itself yields to services. Agriculture shrinks, industry grows, services dominate. The sequence is roughly the story of Britain in the nineteenth century, of South Korea and Taiwan in the twentieth, of China in the twenty-first.

Tamil Nadu tells a different story, and the divergence is not incidental. Between 1980 and 2010, the state’s agriculture share of GSDP fell from 24 percent to 9 percent, a collapse that any Lewis-model economist would recognise as textbook structural transformation. Services captured much of the released output, growing from 41 percent to over 53 percent of the economy at current prices by 2024-25. But manufacturing never hollowed out. By 2024-25, manufacturing alone accounted for 24.5 percent of gross value added, making it the single largest subsector in the state’s economy, ahead of real estate, construction, and trade. Tamil Nadu simultaneously shed agricultural dominance and built the densest manufacturing base of any Indian state. These facts do not refuse to resolve into any standard template.

The employment data adds another layer of strangeness. Between 1983 and 2022, Tamil Nadu's agricultural workforce share fell from 58% to 29%, a drop of nearly 30 percentage points across four decades. Manufacturing absorbed almost none of this. By 2022-23, manufacturing employed 16.8% of the workforce — precisely the same share as in 1983-84. Despite manufacturing output growing at nearly three times the national average, despite Tamil Nadu becoming the country's largest factory employer by count, manufacturing employment as a share of the total workforce returned to where it started. The sector that did most of the absorbing was construction, which expanded from 3% of the workforce in 1983 to 18% by 2022. Services absorbed another 13 percentage points, growing from 22 to 35%. The development premium that Tamil Nadu now holds is, for a large share of its working population, a wage earned on a scaffolding.

India as a whole never completed the Lewis transition either. Dani Rodrik has called the country’s development path a case of premature deindustrialization: India moved from agriculture to services, largely skipping the manufacturing stage that historically built the middle class in East Asia. The national manufacturing share of GDP has barely moved in thirty years. Against this backdrop, Tamil Nadu is not merely an outlier state. It is India’s closest answer to the question of whether an Asian-style manufacturing-led development path was ever possible here, and what it would have taken to make it work.

Sitting with this matters, because it complicates the easy triumphalism that tends to attach itself to Tamil Nadu’s headline numbers. Real GSDP growth of 11.19% in 2024-25, the highest in India. Exports doubling from $26 billion to $52 billion between 2020 and 2025. Manufacturing growing at three times the national average. These are real and they are large. But they describe the upper end of a bifurcated labour market, one part of which went to college and now writes software or assembles cars in Chennai, and another part of which drives a lorry to a construction site and is still waiting for the dividend.

Keeping both of these in view is not a reason to dismiss the growth. It is a reason to ask a more careful question. What explains the upper end? How did Tamil Nadu, alone among India’s large states, produce a workforce capable of staffing the kind of manufacturing and services economy that global firms want to invest in? And why is it so hard for any other state to replicate?

Start with the demography. Tamil Nadu’s total fertility rate in 1971 was 3.9 children per woman, not dramatically below the national average of around 5.4. By 1991, it had fallen to 2.2, a drop of 1.7 points in a single decade, the sharpest single-decade fall of any Indian state at the time. Two years later, in 1993, Tamil Nadu became the second state in India to reach replacement fertility, after Kerala. Bihar’s TFR in 2019-21 was 3.0 children per woman. By 2023, Tamil Nadu’s SRS estimate had fallen to 1.3, a number that places it in the company of ageing East Asian economies, not developing South Asian ones.

Bihar in 2021 was where Tamil Nadu was in the late 1980s.

Demographic transitions follow the pattern of female education with a lag of roughly fifteen to twenty years. Tamil Nadu’s female literacy rate was 30.9% in 1971, not obviously superior to comparable states. But it climbed steadily: 40.4% in 1981, 51.3% in 1991, 73.4% by 2011. At each decennial milestone, Tamil Nadu was running ahead of the national average. The women who became literate through the 1970s and 1980s drove the fertility decline of the late 1980s and 1990s. Their smaller, better-nourished families then entered the workforce from the late 2000s onwards. This is the mechanism through which a school meal in 1956 eventually appears, many decades later, as a productivity figure in an Annual Survey of Industries.

But the lag is not thirty years in any precise sense, and this essay would be dishonest to pretend otherwise. It is better understood as a range. The noon-meal programme of 1956 to an IT-ready workforce in 2000 is approximately 44 years. The OBC education expansion that Dravidian governments pursued from 1967 to the replacement fertility milestone in 1993 is about 26 years. MGR's 1982 expansion of the meal scheme to all children aged 2 to 15 to the manufacturing surge visible in ASI data from 2014 is about 32 years. Somewhere between 25 and 45 years is not a precise law, but it is a consistent pattern across multiple independent data series.

The range tightens somewhat when you look at what happened after 1982. That year, MGR expanded the noon-meal programme from school-age children to all children between two and fifteen, including preschool-age children through anganwadis. By 1984 the programme covered roughly 6 crore children across the state. The policy is often described as welfare spending. In nutritional terms, it was an investment in cognitive capacity. Children who are not chronically hungry learn more, retain more, and arrive at school ready to be taught rather than distracted by hunger. The cohort of children who entered the programme between 1982 and 1985 were born between approximately 1969 and 1983. They entered the workforce between 2000 and 2010, which is precisely when Tamil Nadu's manufacturing employment and export data begins to show the acceleration that looks, from a distance, like liberalization's delayed reward. The timing is not a coincidence. It is the lag, made visible in data.

The academic literature supports it. Robert Barro's work for the OECD finds education-to-growth lags of 15 to 40 years are empirically robust across countries, with the clearest effects at 15 years for basic schooling quality and longer payoffs for higher education investment. The Tamil Nadu story resembles, in a different form, the broader insight that Drèze and Sen drew from Kerala: social investments often create economic capability long before they produce visible industrial outcomes. Tamil Nadu is not exceptional in the pattern. It is exceptional in having followed the pattern so consistently, across so many successive governments, for so many decades.

Long before India began speaking seriously about demographic dividends, Tamil Nadu had already built one of the country’s deepest higher-education ecosystems. By 2014–15, its higher education gross enrolment ratio was nearly double the national average. This near-doubling of the national rate was not produced by proximity to a single technology corridor, the way Bangalore inflated Karnataka’s numbers, or by private investment in elite institutions. It was the accumulated product of reservation policy, which created mass demand for college degrees among communities that had not previously accessed them, combined with a state government that built the supply of affordable colleges and polytechnics to meet that demand. The OBC reservation push from 1967 onwards worked precisely because the noon-meal programme of the 1950s had already raised baseline enrolment and literacy enough that a substantial cohort of backward-caste students could plausibly benefit from college access. The two policies were not designed as a sequence. They became one, across two different governments and two different decades, because each created the conditions the other needed.

By the time India liberalized its economy in 1991, Tamil Nadu had built up a structural head-start that no other large state, except Kerala, could match. In 1983-84, only 58% of Tamil Nadu’s workforce was in agriculture, against a national average of around 69%. The state had more workers already partially detached from farming, already semi-educated, already urbanized enough to move toward industrial employment. When the reforms opened the economy, Tamil Nadu had a labour supply that was genuinely ready for the opportunities that followed.

That readiness also showed up in wages, which is a detail that deserves more attention than it usually receives. By 2017, rural wage levels, both agricultural and non-farm, were among the highest in India. High wages in a labour-abundant country are generally the sign of a tight labour market, meaning that educated, mobile workers have alternatives and can bargain. Tamil Nadu’s workers had alternatives. The school meals, the reservations, the expanding public sector and then private sector employment: all of these had created a population with exit options. That is a different kind of economy from one where labour has nowhere to go.

The Chennai auto cluster is perhaps the clearest crystallization of Tamil Nadu’s long structural transformation. When Hyundai Motor India chose Sriperumbudur in 1996, it was not entering an industrial vacuum. It was locating into an ecosystem that had been built over decades: component manufacturers linked to the older engineering economy of Madras, skilled workers trained through the state’s dense network of ITIs and polytechnics, port and highway infrastructure connecting Chennai to global markets, and a state administration with long experience managing industrial labour relations. Once established, the cluster became self-reinforcing. Firms such as Ford Motor Company, BMW and Renault–Nissan expanded into the same corridor, alongside hundreds of component suppliers. By the late 2010s, the Chennai region accounted for roughly 30% of India’s automobile production and around 35% of its auto-component industry. These numbers were not the cause of Tamil Nadu’s industrial success. They were the consequence of decades of prior investment in human capability, infrastructure, welfare provisioning and institutional capacity that made such industrial concentration possible in the first place.

Tamil Nadu’s per capita net state domestic product in 2023-24 was Rs. 3,61,619, placing it third among major Indian states, behind Telangana (Rs. 3,87,623) and Karnataka (Rs. 3,80,906). Bihar’s was Rs. 69,321. The gap between Tamil Nadu and Bihar is 5.2 times. Telangana’s and Karnataka’s higher positions reflect heavily Hyderabad and Bangalore respectively — single metro IT agglomerations that dominate their state averages. Tamil Nadu’s third-place ranking reflects a broader base: factories, farms, IT companies, and construction operating simultaneously at relatively high productivity across a large and diverse workforce.

The employment and output data together reveal a genuinely uncomfortable pattern. Manufacturing generates 24.5% of gross value added but employs only 16.8% of workers and has not grown its workforce share in forty years. Agriculture employs 29% of workers but generates only 13% of output. Services account for 54% of output but only 35% of employment. The only sector where employment share consistently grew was construction, which now accounts for 18% of the workforce while producing around 12% of output. This is not the picture of a smoothly functioning labour market. It is the picture of two economies running in parallel: a high-productivity manufacturing and services economy that is adding value without adding proportional workers, and a construction economy that is absorbing the people the high-productivity economy cannot or will not employ.

Tamil Nadu’s headline growth numbers are real. But the workforce doing most of the absorbing is laying floors and pouring concrete, not writing software or assembling cars.

It is worth pausing here to acknowledge what this story does not say. The MIDS Working Paper by Kalaiyarasan A., the most systematic academic account of Tamil Nadu’s structural transformation, is explicit about the limits of any single-cause explanation. Tamil Nadu’s growth was also powered by Chennai’s coastal geography, by the agglomeration effects of the auto cluster once it became self-reinforcing, and by active state industrial policy from the 1990s that successfully competed for foreign direct investment. Human capital investment was necessary. It was not sufficient on its own, and no responsible account of Tamil Nadu pretends otherwise.

Gujarat offers the useful counterpoint. Gujarat has comparable manufacturing depth, a similar share of factories, and a manufacturing-to-GSDP ratio that exceeds Tamil Nadu’s. But Gujarat’s social indicators (female literacy, child health, malnutrition rates) have historically been weaker than its economic headline numbers suggest. Gujarat’s manufacturing is more capital-intensive and employs fewer workers per unit of output.

The point is not that Tamil Nadu’s model is superior in every dimension. The point is that the character of industrialization a state achieves, who it employs, how it pays them, how the gains distribute, reflects the social structure that preceded the industry. Tamil Nadu’s manufacturing is labor-intensive because Tamil Nadu’s labour is both abundant and relatively educated. You do not get that combination by accident.

The 30-year lag argument, in its optimistic version, sounds like a recipe: invest in education and nutrition today, and your economy will be transformed in three decades. Bihar should start now and expect its payoff by 2055. The implication seems hopeful. It is not.

The lag argument, properly understood, is not primarily about the delay between investment and payoff. It is about the political conditions required to make the investment in the first place. And those conditions cannot be created by policy design.

The noon-meal scheme of 1956 was not a technocratic intervention executed by a welfare-maximizing planner. It was the expression of a social movement. The Dravidian movement, rooted in Periyar’s Self-Respect Movement of the 1920s and 1930s, had built, over more than three decades, a political culture in Tamil Nadu that was unusually oriented toward the dignity and material welfare of backward castes and lower-income families. Kamaraj was a product of that culture. When he decided to feed schoolchildren, he was not solving a human capital optimization problem. He was honoring a political commitment to communities that had been systematically excluded from the economy’s gains.

The 1967 arrival of the DMK in state government extended this logic into education through aggressive OBC reservations and an expansion of higher education access. This was contested, often bitterly. But Tamil Nadu’s political system had developed, through decades of Dravidian competition, an internal mechanism for translating lower-caste aspiration into educational investment. The result was not merely that more children went to school. The result was that children of agricultural laborers, of weavers, of potters, went to school, and their families invested in that schooling because they could see a credible path, created by reservations, by mid-day meals, by an expanding state sector, from schooling to employment.

Bihar in 2026 does not have this. Bihar has seen caste-based political assertion, particularly through the Mandal era. But the character of its political economy is different in ways that matter. The competitive welfare politics that drove successive Tamil Nadu governments to outspend comparable states on school meals, on rural health workers, on primary schooling infrastructure: this is not an institutional design that can be imported. It emerged from a specific historical sequence: lower-caste social reform preceding political assertion, political assertion preceding state capture, state capture generating welfare investment, welfare investment generating human capital, human capital generating an economy capable of attracting global manufacturing.

That sequence cannot be reverse-engineered. No policy paper can create the pre-conditions that Tamil Nadu’s Dravidian movement built between 1920 and 1960. This is the most honest, and most dispiriting, reading of Tamil Nadu’s success: much of what Tamil Nadu did cannot be done on purpose. It can only be recognized, in retrospect, as the outcome of a political history that no one was managing as a development strategy, because no one at the time knew that is what it was.

This is what development economists mean when they speak of path-dependence, though they rarely say it with this directness. The path that Tamil Nadu walked was not a path that another state can step onto from a standing start. The pre-conditions for the investments were themselves historical. They required a social movement, a political transformation, and several decades of competitive electoral politics oriented toward the welfare of communities that had traditionally received nothing from the state. That is a very long pre-condition for a 40-year investment lag.

The uncomfortable question for the rest of India is not whether Tamil Nadu’s model can be copied. It cannot. The question is what it would mean to govern with a 40-year horizon in a political system designed for five-year cycles. Tamil Nadu’s politicians were not patient individuals. Kamaraj ran an intensely competitive political machine. MGR built a welfare state partly because he was also building an electoral coalition, and he knew that feeding a child was a vote that lasted two decades. The patience was systemic, not individual. It emerged from a political culture that had learned, through the Dravidian long march, to measure success by whether poor families could survive and accumulate, not by whether a GDP number looked good in an election year.

There is a particular kind of cognitive dissonance that shows up in economies that are growing fast without distributing that growth. The headline numbers look good. The quarterly GDP figures impress; the stock market climbs and the corporate earnings hit records. And then you look at what happened to the wages of the people who actually made the things and performed the services that generated all that output, and something does not add up.

India has been living inside this dissonance for at least a decade. Between FY22 and FY24, the economy grew at an average annual rate of approximately 8.3 percent. Over the same period, regular wages contracted by 0.07 percent. Not stagnated. Contracted. The distance between those two numbers is not a data anomaly or a measurement artifact. It is a structural fact about how India’s growth is organized and who it benefits.

This dispatch has a set of ten numbers, drawn from India’s most credible official and institutional sources, that describe this structural condition with some precision. Each tells a partial story. Together, they describe something larger: an economy that has grown impressively while keeping the people who power it precisely where they started.

The economy grew. Worker wages did not.

Between FY22 and FY24, India’s real GDP grew at approximately 8.3 percent annually according to NSO national accounts data, with FY24 alone recording 8.2 percent growth. Regular wages, over the same period, contracted by 0.07 percent, according to PLFS data cited in the Economic Survey 2025. Not stagnated. Contracted.

This figure comes from the government’s own Periodic Labor Force Survey. It is the government’s account of what happened to workers in one of the fastest-growing major economies on earth. The gap between the headline growth story and this number is as stark as any single statistic in Indian economic data. GDP growth, it turns out, is not the same thing as wage growth, and confusing the two has consequences that eventually show up on the streets.

A decade of near-zero real wage growth.

Zoom out further and the picture is bleaker still. According to SBI Research Ecowrap, real wage growth in India averaged 0.23 percent annually between FY17 and FY26. Over ten years, across the entire workforce, wages grew in real terms by a cumulative 2.3 percent. Inflation, in the meantime, eroded purchasing power far faster.

For a worker in a garment factory or a machine shop, this means that her salary in 2026, adjusted for what it actually buys, is barely distinguishable from her salary a decade ago. The dearness allowance system, which is meant to index wages to inflation through the Consumer Price Index for Industrial Workers, has not closed this gap. Minimum wages in several states were revised so infrequently, and by such small amounts, that they functioned more as a historical record of what workers once accepted than as any real floor on what they should be paid.

Male workers are poorer in real terms than they were in 2017–18.

The Economic Survey 2025, drawing directly on PLFS data, found that real monthly wages for male self-employed workers stood at Rs 8,591 in 2023–24 compared to Rs 9,454 in 2017–18. That is a decline of 9.1 percent over six years, in real terms, for one of the largest categories of working men in India.

Self-employment in India is not a choice for most of the people who do it. It is the residual category: what you fall into when the formal economy does not have room for you. The men in this category are not tech freelancers or consultants. They are vendors, small contractors, piece-rate workers, and the vast middle of India’s informal labour market. A 9.1 percent real wage decline for this group, during a period when India’s stock market repeatedly hit all-time highs, is not a minor data anomaly. It is a description of who the growth has not reached.

Nearly half of all wage earners do not receive the legal minimum wage.

India’s national floor-level minimum wage exists on paper and is regularly revised. What it has not managed to do is reach the workers it was designed to protect. According to the ILO India Wage Report and analysis of PLFS survey data, approximately 42 percent of all wage earners in India receive wages below the national minimum wage floor. For casual laborers, the share is close to 52 percent. For regular and salaried workers, it is around 40 percent.

This is not a story about the informal economy hiding from enforcement. A significant share of these violations occur in plain sight, in establishments large enough to be surveyed, in sectors covered by the Minimum Wages Act. The compliance gap is a governance failure as much as an economic one, and it means that the minimum wage, as an instrument, has never actually set the floor it was designed to set.

Four in ten manufacturing workers are contract workers, the highest share ever recorded.

In FY23–24, contract workers constituted 42 percent of employees in India’s organized manufacturing sector, according to the Annual Survey of Industries. This is the highest share since tracking began in 1997–98, and it represents roughly double the share from 1999–2000, when contract workers were around 20 percent of the formal manufacturing workforce.

This shift matters enormously. Contract workers are employed through third-party agencies, have short-term agreements with no guarantee of renewal, and are legally separated from the factory’s permanent workforce in ways that strip them of most protections. The steady contractualization of the organized sector has allowed firms to expand production without expanding their obligations. A significant share of India’s formal manufacturing workers are, in effect, formally employed under terms designed to keep them perpetually provisional.

Two workers, one production line, very different pay.

ASI 2021–22 data shows that directly employed workers in organized manufacturing received, on average, Rs 586 per day. Their contract counterparts received Rs 523. That is a 12 percent gap in base wages alone. Once paid leave, provident fund contributions, medical benefits, and job security are factored in, the divergence in total compensation rises to approximately 1.5 times.

Two workers stand next to each other on the same production line. One earns substantially more than the other, has access to leave encashment, social security contributions, and a path to permanence. The other has none of these. The production line does not distinguish between them; the contract does. This arrangement is entirely legal. It is also one reason why aggregate wage data consistently understates how badly the lower half of the formal manufacturing workforce is faring.

For forty years, productivity grew. Wages did not follow.

An ILO Asia-Pacific Working Paper on declining wage share in India’s organized manufacturing sector traced this divergence back to the mid-1980s. Until roughly 1985–86, labour productivity and wages moved together. After that point, they began to diverge, with productivity growing at a compound annual rate at least 2.5 percentage points faster than earnings per employee. The consequences show up clearly in the data: the share of wages in net value added by organized manufacturing firms fell from approximately 26 percent in 1999 to approximately 16 percent in 2022, a decline of ten percentage points over two decades.

This is not a story about lazy workers demanding more than they produce. It is a story about the distribution of productivity gains. Workers made India’s manufacturing sector more productive. The returns from that productivity accrued elsewhere: to capital, to shareholders, to the owners of the enterprises. This is not unusual in global terms, but the scale and duration of the divergence in India are notable, and they compound with each passing year.

Delhi and Noida are the same city, with a 75 percent wage gap.

Before the April 2026 protests forced a revision, the minimum wage for an unskilled worker in Noida, governed by UP rates, was Rs 11,313 per month. Across the administrative boundary in Delhi, the same worker would legally be entitled to Rs 19,846. Haryana, which triggered the crisis with its April 2026 revision, set its unskilled minimum at Rs 15,221, up from Rs 11,257. The two cities that matter most to this story, Delhi and Noida, share a skyline, a metro system, a cost of living, and a labour pool. They do not share a wage policy.

Workers commute across this boundary daily. They live in the same neighborhoods, pay the same rents, buy vegetables at the same prices. The minimum wage differential of approximately 75 percent between Delhi and Noida has no correspondence to any real difference in what it costs to live. It reflects the fragmented structure of India’s minimum wage system, where over 1,200 distinct minimum wages across states and industries create a patchwork in which geography determines wages more than skill, productivity, or any principle of adequacy.

Corporate profits grew 22 percent in FY24. Employment grew 1.5 percent.

The Economic Survey 2025 documented a striking divergence at the top of the corporate sector. Corporate profits grew by 22.3 percent in FY24. Employment across the same firms expanded by 1.5 percent. Between 2012 and 2022, IT sector CEO salaries rose by 835 percent; entry-level salaries at the same firms rose by roughly 45 percent.

This is the political economy of the wage crisis in miniature. The gains from India’s growth are not being distributed across the workforce. They are being concentrated at the top, in the form of profits and executive compensation, while the workers at the base of the pyramid receive wages that are, in real terms, barely higher than they were a decade ago. This concentration is not primarily a story about inequality between formal and informal workers, though that gap is real. It is a story about what happens inside formal employment, inside organized industry, inside the sector that is supposed to represent the promise of India’s manufacturing ambitions.

There is a version of India’s growth story in which the numbers above are a temporary condition. It’s the price of a particular phase of development, soon to be corrected as the economy matures and wages catch up. This is a coherent argument. It has historical precedent. South Korea and Taiwan both went through decades of labour suppression before rising wages eventually accompanied rising productivity.

The problem is that India is not following that script. In South Korea’s case, the wage share of manufacturing output held up during the high-growth decades; the gains were more broadly distributed even as the economy was rapidly industrializing. In India, the wage share has been falling for forty years. Contractualization has accelerated rather than reversed as the economy has grown. Minimum wage enforcement has not strengthened; it has remained porous. These are not the characteristics of a system moving toward correction. They are the characteristics of a system that has settled into a stable equilibrium in which labour bears the costs of growth and capital captures its rewards.

This matters beyond questions of fairness. An economy where wages do not grow is an economy where domestic consumption is structurally constrained. India cannot become a large consumer market, which every vision of its economic future requires, if the workers who produce its goods cannot afford to buy them. The wage crisis is not a social problem appended to an otherwise successful economic model. It is a structural limit on the model itself. The facts in this dispatch are a question about whether the growth India is so rightly proud of can be sustained, and for whom.

India’s Development is a Political Choice

After the first heavy rain in Delhi, the pattern is familiar. A newly built expressway floods at a low point, traffic slows to a crawl, and neighborhoods that sit next to a river depend on water tankers. At the same time, gated colonies switch to private generators because the grid has tripped. None of this is new. The city builds flyovers, expands metro lines, and digitizes services, yet basic systems like drainage, water, and electricity fail to work together. The problem is not the absence of capacity. It is that different parts of the system improve in isolation without forming a coherent whole. Each agency optimizes for its own mandate, each project is announced and executed as a discrete intervention, and yet the lived experience of the city remains unchanged. India’s economy shows a similar pattern. There are islands of competence—globally competitive firms, world-class infrastructure in pockets, high-end digital systems—but they do not add up to broad-based transformation.

Rohit Lamba, in his latest paper titled ‘ India’s Governing Trilemma and the Paradox of Economic Nationalism’, describes this through what he calls a governing trilemma: development, democracy, and civilizational assertion pull policy in different directions. Each of these is powerful on its own, and each is rooted in legitimate demands. Development requires sustained, high growth over long periods. It demands that labour move out of low-productivity sectors into more productive ones, that firms invest and scale, and that the state build infrastructure and human capital consistently over time. This is not a smooth process. It involves failures, reallocations, and long phases where results are not immediately visible. Democracy, by contrast, operates through constant feedback. Elections are frequent, coalitions are broad, and governments face pressure to deliver tangible outcomes quickly. Policies that produce visible benefits—subsidies, cash transfers, welfare schemes—become central because they respond directly to voter needs. Civilizational assertion introduces a third logic, one that frames economic policy in terms of identity, sovereignty, and historical continuity. It gives meaning and coherence to political choices, but it also reshapes how success and failure are interpreted.

Lamba argues that these three logics cannot be fully satisfied at the same time. Development requires patience and discipline, democracy demands immediacy and responsiveness, and civilizational assertion seeks alignment with a larger narrative that is not easily revised. Any two of these can coexist, but bringing all three together creates persistent tension. India has experienced different configurations over time. The reform period of the 1990s and early 2000s combined development with democracy while keeping civilizational politics relatively muted. That phase delivered higher growth and deeper integration with global markets, but it did not produce a durable political consensus. In contrast, East Asian countries combined development with nationalism while limiting democratic pressures, allowing governments to impose discipline, direct resources, and sustain long-term strategies. India’s current configuration combines democracy with civilizational assertion, but without the discipline required for rapid development. This combination is politically stable, electorally successful, and socially coherent, yet it produces only moderate growth and incomplete structural transformation.

At one level, this framework explains why outcomes look the way they do. But it can also create the impression that India is constrained by forces beyond its control. The more uncomfortable reading is that India is not trapped; it has settled into an equilibrium that works politically, even if it underperforms economically. This becomes clear when one examines how the system responds to policy failure. In a development-driven system, failure is a signal. Policies are adjusted, firms are forced to compete, and resources are reallocated toward more productive uses. Feedback is hard, often painful, but it drives learning. In India’s current configuration, failure is absorbed rather than corrected. Policies that underperform are often defended more strongly because they are tied to political constituencies and broader narratives of identity. What Lamba describes as a ratchet effect emerges, where each round of underperformance leads to greater insulation rather than recalibration. Over time, this creates a system that is stable but resistant to change.

Trade policy illustrates this dynamic with clarity. Over the past decade, India has increased tariffs across several sectors and chosen to stay out of major trade agreements. These decisions are not arbitrary. They benefit organized domestic groups that prefer protection from foreign competition and align with a broader narrative of self-reliance. The political logic is straightforward. The gains from protection are visible and concentrated, while the costs—higher prices, reduced competitiveness, weaker export performance—are diffuse and slower to emerge. From a developmental perspective, however, this creates a problem. Manufacturing growth depends on integration into global value chains, access to cheaper inputs, and exposure to competition. Protection weakens these channels. Yet when these policies fail to deliver strong manufacturing growth, the system does not move toward greater openness. Instead, protection is reinforced, and underperformance is reframed as evidence of the need for deeper self-reliance. Manufacturing remains stuck at around 15 to 17% of GDP, reflecting this unresolved tension.

Industrial policy reveals a similar pattern, even though the instruments have evolved. The License Raj relied on direct control and ultimately failed to produce a competitive industry. Liberalization replaced it with markets and integration, but industrial policy has re-emerged in a new form through the Production-Linked Incentive schemes. These schemes offer subsidies tied to incremental output across fourteen sectors, with a total allocation of ₹1.97 lakh crore. The design is more sophisticated than earlier approaches. It attempts to reward performance and avoid permanent support. Yet the same constraints reappear in implementation. Democratic pressures lead to the distribution of incentives across sectors and regions rather than concentrating them where success is most likely. There is also pressure to show quick, visible results, especially in terms of employment, even though early-stage manufacturing tends to be capital-intensive. By the fourth year, only about a third of the expected employment had been achieved, highlighting the mismatch between political timelines and economic processes.

Civilizational framing adds another layer of rigidity. Industrial policy is presented as part of a broader project of national revival, which raises expectations beyond what incremental policy can deliver. When outcomes fall short, the space for adjustment narrows because acknowledging limitations risks undermining the narrative itself. At the same time, the discipline that characterized successful industrial policy in East Asia is largely absent. There are no strong export conditions attached to subsidies, and firms can operate within protected domestic markets without being forced to compete globally. Complementary investments in infrastructure, skills, and regulatory reform often lag, limiting the effectiveness of incentives. The result is a pattern of partial success without deep transformation. Output increases in some sectors, but value addition remains limited, and ecosystems do not fully develop.

The contrast with the services sector helps clarify the mechanism. India’s IT and business services industries have grown rapidly, integrated with global markets, and built competitive firms. This success did not come from heavy protection or large subsidies. It emerged because the sector operated outside the most binding elements of the trilemma. It did not generate strong demands for protection, nor was it heavily embedded in civilizational narratives. Firms were exposed to global competition from the beginning and had to meet international standards to survive. This created strong feedback, where performance determined outcomes. Manufacturing, by contrast, sits at the centre of the trilemma. It is politically sensitive, symbolically loaded, and developmentally critical, which makes it harder to impose the discipline required for sustained transformation.

What emerges from these examples is not a series of isolated policy failures, but a consistent pattern. Democracy rewards immediacy, civilizational politics provides legitimacy, and together they reduce the pressure to pursue difficult, long-term reforms. The system settles into a balance where moderate growth is sufficient to maintain stability. The persistence of 5 to 6% growth is not accidental. It reflects a configuration that works politically even if it falls short economically.

The costs of this equilibrium accumulate gradually. Job creation remains limited, particularly in sectors that can absorb large numbers of workers. Manufacturing does not scale in a way that transforms the structure of the economy. Investments in human capital and institutional capacity lag what is required for sustained growth. Opportunities to integrate more deeply into global production networks are missed. These costs are not always visible in the short term, but they shape the trajectory of the economy over time.

Why India Cannot Follow the East Asian Script?

The comparative sections of Lamba’s paper deepen the argument by showing that India’s trajectory is not an aberration but the result of specific structural conditions that make certain development pathways unavailable. What looks like policy inconsistency begins to look more like constraint once placed in comparative perspective.

The first difference lies in timing. India became a full electoral democracy at extremely low levels of income. This is historically unusual. Most successful late developers, Japan, South Korea, Taiwan, and China, either delayed democracy or constrained it while undertaking rapid industrialization. This sequencing mattered because it allowed governments to prioritize capital accumulation over immediate consumption. They could suppress wages, direct credit, and impose discipline on firms without facing electoral backlash. India, by contrast, faced mass electoral competition from the beginning. Political incentives, therefore, tilted toward redistribution and visible welfare rather than long-term investment. This does not mean democracy makes development impossible, but it fundamentally alters the pace and nature of transformation. It introduces a bias toward immediacy that is difficult to reverse.

The second difference is the nature of society and state structure. India’s diversity is not just cultural; it is politically institutionalized. Federalism, caste cleavages, linguistic identities, and regional parties create multiple centers of power. Policy must be negotiated across these layers. This produces accommodation rather than coordination. In East Asia, relative social homogeneity made it easier to construct a unified national project around development. In China, diversity exists but is channeled through a highly centralized party-state that can enforce coherence. India’s structure does neither. It disperses authority while concentrating fiscal and administrative power at the center, creating a mismatch between decision-making and implementation. The result is a system where neither top-down discipline nor bottom-up experimentation works effectively.

A third distinction lies in India’s economic structure. Unlike East Asian economies, India did not follow the standard sequence of structural transformation from agriculture to manufacturing to services. It moved prematurely into a service-led economy. Services now account for the largest share of output, but they cannot absorb low-skilled labour at scale. Manufacturing, which historically served as the bridge for mass employment, remains underdeveloped. This creates a structural trap. Policies continue to emphasize manufacturing as the route to development, but the underlying conditions, such as skills, infrastructure, and global positioning, are not aligned in the same way they were for earlier industrialisers. At the same time, the sectors where India has succeeded, particularly services, are not politically central in the same way. This misalignment between economic reality and political focus is a core part of the constraint.

The fourth factor is historical timing in the global economy. East Asian development occurred in a very different international environment. Trade rules were looser, global markets more open to labor-intensive exports, and technological change more favorable to late industrialisers. Today, the space for replicating that model has narrowed. Global value chains are dominated by incumbents, especially China. Automation reduces the employment potential of manufacturing. Climate constraints limit the feasibility of carbon-intensive industrialization. Trade regimes impose restrictions on the kinds of industrial policy tools that earlier developers used freely. India is attempting late industrialization under far less permissive conditions, which raises the cost of errors and reduces the margin for experimentation.

These structural features explain why India cannot replicate the East Asian path, even if it wanted to. The East Asian model rested on three pillars: export discipline, technocratic insulation, and coordinated state capacity. Firms received support, but only if they performed in global markets. Governments could adjust policies without electoral pressure. And industrial policy was embedded within a broader ecosystem of infrastructure, skills, and institutional reform. India lacks each of these conditions in crucial ways. Democratic competition makes it difficult to withdraw support from failing sectors. Civilizational politics makes policy adjustment appear as an ideological retreat. And fragmented governance prevents the kind of coordinated push that East Asia achieved.

This is where Lamba’s idea of a default equilibrium becomes important. India has not failed to find the right policy mix; it has settled into one that is politically viable. This equilibrium combines democratic responsiveness with civilizational assertion, while moderating developmental ambition. It delivers electoral success, maintains social cohesion, and provides visible benefits through welfare and symbolic politics. But it does not generate the sustained, high growth required for structural transformation. Growth stabilizes at a middling level, manufacturing remains stagnant, and employment generation lags behind demographic needs.

What makes this equilibrium durable is not just that it works, but that it contains its own feedback loop. Underperformance does not trigger course correction. Instead, it strengthens the very forces that produced it. When growth disappoints, governments double down on welfare to manage democratic pressures. When economic outcomes fall short, civilizational narratives intensify to maintain legitimacy. Policies are reframed rather than revised. Over time, this creates a system where economic signals are filtered through political and ideological considerations, weakening the learning capacity.

The comparative insight, then, is not that India is doing the wrong things, but that it is operating under a different set of constraints than those assumed in standard development theory. The East Asian experience cannot be transplanted because the institutional and historical conditions that made it possible do not exist in India. Instead, India represents a different model: a large, diverse democracy attempting development in an era of constrained globalization, where identity and politics play a central role in shaping economic choices.

This is why Lamba’s argument ultimately lands on a more uncomfortable conclusion. India’s trajectory is not a temporary deviation that will correct itself with better policy. It is a stable configuration that reflects deeper structural realities. Unless one of the underlying forces like democratic pressures, civilizational assertion, or developmental urgency, changes in how it operates politically, the equilibrium is likely to persist.

The central point, then, is not that India is constrained by an impossible trilemma, but that it is making political choices about which trade-offs to accept. A different path would require accepting short-term costs in exchange for long-term gains. It would involve allowing greater competition, even when it disrupts existing interests, investing more consistently in capacity, and reframing economic openness as a source of strength rather than vulnerability. These are not purely technical decisions. They are political choices that require building new coalitions and sustaining them over time.

Until such a shift occurs, the pattern is likely to persist. Like Delhi after rain, the system will continue to improve in parts while failing to come together as a whole.

India’s politicians want young people to have more babies. And they want them to live in cities. They have not noticed that these two things are at war with each other.

Here is the contradiction in its starkest form.

In January 2025, the Andhra Pradesh government also proposed disqualifying people with fewer than two children from contesting local elections. Andhra Pradesh CM has also proposed a ₹25,000 cash incentive for couples having a second child or more, just months after the state repealed its two-child norm. Tamil Nadu’s MK Stalin, whose state’s TFR has now fallen to 1.4, one of the lowest ever recorded for an Indian state, has publicly urged couples to have children “immediately” after marriage, framing it as a political imperative ahead of delimitation.

Now, here is what those same young Indians are actually facing.

A standard 2BHK apartment in Dwarka, not Worli, not Vasant Vihar, just ordinary middle-class Dwarka, costs ₹1.2 crore or more. In Mumbai, the average home now costs 34 times the annual income of a typical buyer, making it more unaffordable than New York, London, or Hong Kong. A Mumbai professional who bought a 2BHK in 2007 paid 3.5 times his annual income. Today, that same flat costs 34 times. Between 2022 and 2024, the number of homes priced below ₹1 crore across India’s major cities fell by 36 per cent, from 3.1 lakh to 1.98 lakh units. The affordable segment is not struggling. It has ceased to exist.

In Bengaluru’s Sarjapur Road, where India’s IT economy runs its daily operations, rents for a 2BHK rose 66% in five years. In Mumbai, 41% of households live in slums. A Delhi government survey found that 60% of housing units in the capital have just one or two rooms, while the average household size is above five persons.

The politicians asking young Indians to have more children have not bothered to ask where those children are supposed to sleep.

There is no shortage of explanations for why Indians are having fewer children. There is, however, a striking shortage of attention paid to the most material explanation of all: they cannot afford the space. This dispatch makes a simple argument: India’s housing crisis and India’s fertility crisis are not two different problems. They are the same problem, running through the same households, in the same cities, at the same time.

India’s national TFR has fallen to 1.9. Urban India’s TFR now stands at 1.5 and has been below replacement since 2004. The most economically productive states have collapsed furthest. SRS 2023 reports Tamil Nadu and West Bengal at 1.3, Maharashtra at 1.4, and Kerala at 1.5. Only six states — Bihar, UP, MP, Rajasthan, Chhattisgarh, and Jharkhand — remain above replacement. In several of India’s most productive states, TFRs are now lower than in much of Western Europe. If these were independent countries, the numbers would be treated as a national emergency.

There is no shortage of studies on India’s fertility decline.

Studies across India, from Drèze and Murthi (2001) using district-level panel data to more recent analyses using NFHS waves, consistently find that female educational attainment is the most powerful predictor of low fertility at the state and district level.

The relationship between women’s education and lower fertility is one of the most robust findings in demographic and development research. The mechanisms operate through a set of mutually reinforcing economic, social, and institutional channels.

Start with the economic mechanism: opportunity cost. Education raises the potential value of women’s time. A woman with secondary or higher education has better prospects in skilled employment, so time spent on pregnancy, childcare, and domestic work carries a higher economic cost. In economic theory, particularly Gary Becker’s household model, this leads families to substitute the quality of children for quantity. Parents invest more resources, such as education, health, and opportunities, into fewer children, resulting in smaller family sizes. Education also delays marriage and first birth, directly reducing fertility. NFHS data in India consistently show that women with secondary or higher education marry several years later than women with little schooling, shortening the reproductive window.

A second mechanism is greater knowledge and control over reproduction. Education improves access to information about contraception, reproductive health, and family planning. Literate women are better able to engage with health systems and make deliberate decisions about spacing or limiting births, reducing unintended fertility.

Education also produces normative change. Exposure to schools, peer networks, media, and urban environments reshapes aspirations and social expectations. In agrarian settings, large families often function as labour and old-age security, but education weakens this logic. As aspirations for children’s education and mobility rise, the perceived cost of raising each child increases, reinforcing the preference for smaller families.

Finally, education improves child survival and health behaviour. Educated mothers are more likely to adopt practices such as vaccination, sanitation, and timely healthcare. When child mortality falls, families no longer need additional births as insurance against child loss, which further lowers fertility.

These mechanisms are amplified by urbanisation. Education often leads to migration into urban labour markets, where children are expensive to raise, and child labour has little economic value. In cities, children become primarily consumption and human-capital investments rather than productive assets, pushing fertility downward.

The final factor is the focus of this dispatch.

All of these explanations are real. This dispatch does not contest them. What it argues is that there is an unexplored missing variable, one that sits between all of them and amplifies every one - the sustained rise in the cost of housing in Indian cities. A revealing diagnostic: a scan of six high-ranking population journals published between 2021 and 2024 found that of 278 fertility studies, 58 focused on education, 62 on employment and income, and only four on housing.

The Great Indian Housing Catastrophe

Let’s explore what has actually happened to housing costs in India. The most rigorous long-run analysis comes from the CSEP working paper by Gupta, Agnihotri and George. House prices in India rose at 9.3% per year between 1991 and 2021, which is a nominal thirty-fold increase, roughly fifteen times in real terms. Gold appreciated by 9.2% over the same period. For the investor class, this was a bonanza. For the first-time buyer, like a young couple trying to raise a family, it was a slow-motion catastrophe.

The CSEP paper introduces a useful metric, Price-to-Income Ratio (PTI): how many years of household income does it take to buy a home? The global affordability benchmark is 5. India’s national average PTI is 11, more than twice the benchmark. Housing in major Indian cities is extremely expensive relative to incomes. In 2019, Mumbai’s price-to-income ratio (39) was far higher than cities like Singapore (22) and nearly three times New York (12). While rent accounts for about 6% of household expenditure in urban India on average, it rises to around 19.2% in large metropolitan districts such as Mumbai, Delhi, Kolkata, Chennai, Hyderabad, and Bengaluru.

The study compares India’s housing market with the rest of the world and shows how poor transparency pushes prices up. In countries with transparent real estate markets, homes cost about 8 times annual income on average, while in countries with low transparency, the ratio rises to around 14. India’s housing market faces several structural problems. Only about 28% of cities have approved master plans, and even these are often too vague or poorly funded to guide real development. This limits land supply and makes it unpredictable. The lack of transparency also benefits developers, who often earn profit margins above 20%, much higher than in many other industries. At the same time, real estate is widely used to park unaccounted money, which increases demand from wealthy buyers and keeps prices high even when economic fundamentals do not support them.

Recent data only confirms the trend is worsening. Between 2020 and 2024, household incomes in major Indian cities grew at a compounded annual rate of 5.4%. Property prices grew at 9.3%. The average EMI-to-income ratio nationally has risen from 46% in 2020 to 61% in 2024, as home loan interest rates climbed from 7.35% to 9.1%. International best practice says a household should not spend more than 30% of its income on housing. India’s urban average is now twice that.

The middle segment of India’s housing market, homes priced roughly between ₹40 lakh and ₹1 crore, has traditionally served as the backbone of urban housing demand. This price bracket caters to the salaried middle class: professionals, small entrepreneurs, and first-generation urban households aspiring to own a decent dwelling in cities. Yet recent data suggest that this segment is under increasing stress from both the supply and price sides. On the supply side, the number of homes priced below ₹1 crore has fallen sharply across India’s major cities. According to data from the real-estate analytics firm PropEquity, the supply of such homes across the top nine cities declined from about 3.1 lakh units in 2022 to roughly 1.99 lakh units in 2024, a fall of around 36% in just two years. This drop reflects a clear shift in developer strategy. Builders increasingly prefer launching premium and luxury projects, which offer higher margins and faster capital recovery, rather than mid-income housing, where price sensitivity among buyers is much higher.

Despite these pressures, demand in this segment remains strong. Mid-income homes continue to account for a significant share of housing transactions in India’s largest cities. Reports from Knight Frank and other property consultants show that homes in the ₹80 lakh to ₹1.5 crore bracket still constitute a substantial portion of residential sales, reflecting the continued purchasing power and aspirations of the urban middle class. However, the mismatch between strong demand and shrinking supply is gradually pushing the market upward toward higher price brackets.

This trend has important implications for urban housing affordability. As developers focus on luxury and high-margin projects, the stock of homes accessible to middle-income households is declining. Meanwhile, rising construction costs, land prices, and regulatory constraints are making it increasingly difficult to build mid-range housing profitably in major cities. The result is a structural squeeze: the segment of the market that once anchored urban homeownership is slowly eroding.

If this pattern continues, the middle class may find itself increasingly pushed out of formal housing markets in large cities. Many households may either delay homeownership, move to peripheral suburbs, or remain in rental accommodation for longer periods. In that sense, the shrinking supply of mid-range housing is not merely a real-estate market shift; it reflects a deeper transformation in the structure of India’s urban housing economy.

Here is a thought experiment.

Imagine you are a mid-level official in a state government, responsible for clearing a building permit for a small industrial unit. The paperwork is complete. The land is zoned correctly. Every box has been ticked. But you know that a CAG audit three years ago flagged a similar clearance in your department for a procedural irregularity. Nobody was corrupt. Nobody was formally punished. But the file sat on someone’s desk for years, and that officer’s career progression quietly stalled.

What do you do?

If you are rational, and most Indian bureaucrats are, you do not decide. You send the file upward. You ask for one more document. You wait for a clarification that may never come. You are not lazy. You are not malicious. You are doing exactly what the system has trained you to do: keep your head down and stay out of trouble.

Now multiply this across thousands of desks, dozens of ministries, 36 States and Union Territories, and tens of thousands of regulatory interactions every single day. What you get is not corruption. What you get is something more insidious, a state that has learned to avoid the judgment, experimentation, and course-correction that development actually requires.

This is the argument at the heart of Chapter 16 of India’s Economic Survey 2025-26. The chapter does not blame politicians or lazy bureaucrats. It identifies something deeper: India’s binding constraint today is not the absence of good ideas or money, but the incentives that shape how decisions get made when things are uncertain. In the language this newsletter has used before, it is the gap between what the state intends and what it can actually deliver; the gap between policy and capability.

Readers of this newsletter will recall the framework from Natarajan and Somanathan’s State Capability in India, a book we covered in Dispatch #99. Their central image was of a hose: even if you pump more water through it, a damaged hose loses most of it to leaks. More funding, more schemes, more ambition, none of it helps if the delivery mechanism is broken.

The Economic Survey builds on this and adds a crucial layer. It is not just that the hose is leaky. It is that the people holding the hose have been punished so often for the direction they aimed it, even when their aim was reasonable, that they have started pointing it at the ground.

The Survey calls this bureaucratic risk aversion. The system punishes visible action more than it punishes inaction. Doing something and getting it wrong carries far higher personal consequences than doing nothing and letting a problem drag on. This did not happen by accident. It was built slowly through decades of retrospective audit, vigilance inquiries, and judicial review that judges decisions by what happened rather than on what was knowable at the time they were made.

The result is what the Survey calls irreversibility creep, meaning temporary measures that become permanent, experiments that never get tried because exit is too costly, and officials who treat every decision as a potential enquiry. Entrepreneurial initiatives get measured against the wrong yardstick: compliance rates, procedural correctness, and revenue figures. Not learning. Not long-term impact. Not whether a reasonable bet made with imperfect information was the right call.

This is a problem money cannot fix. Good intentions cannot fix it either.

The Survey adds:

Temporary measures have often become permanent through what might be called irreversibility creep, raising the stakes of any experiment. Good-faith decisions are subjected to retrospective scrutiny through audits, vigilance processes, and judicial review, often without adequate recognition of the uncertainty under which they were taken. Entrepreneurial initiatives are evaluated using metrics suited to routine administration, such as compliance, revenue collection, and procedural correctness, rather than learning, capacity creation, or long-term impact.

There is a line in the Survey worth reading twice: “Indian institutions treat visible correction as evidence of incompetence rather than maturity.” What this means in practice is that a bureaucrat who changes course when the evidence changes or who says “we tried this and it didn’t work, let’s adapt” is seen as having failed, not as having learned. The signal is clear: be right the first time, or don’t move at all.

Compare this to how post-war Japan handled industrial policy. Policy failures were common and sometimes expensive. What mattered was not whether a particular bet paid off, but whether it generated insight for the next decision. An official could oversee an unsuccessful experiment and still get promoted if the reasoning was sound and the learning was documented. The question asked was not “did this work?” but “was this a reasonable call given what you knew?”

This is the distinction between institutional forgiveness and what currently exists in India. Forgiveness, the Survey is clear, does not mean leniency. It is not a pass for corruption or carelessness. It is the ability to clearly separate honest error from wrongdoing. High-capacity states draw this line sharply. Low-capacity states blur it, and when it blurs, fear replaces judgment across the whole system.

The Survey further explains:

No country that has successfully navigated structural transformation attempted to make its entire bureaucracy entrepreneurial. Equally, it may not be possible for the bureaucracy to be entrepreneurial all the time, because stability and predictability are also necessary traits of good administration. Instead, successful states created bounded institutional spaces - zones where experimentation was permitted, accountability rules were differentiated, and learning was explicit. Japan’s sectoral desks, Korea’s planning agencies, and Singapore’s regulatory institutions all functioned as such safe spaces.

Safe Spaces to Experiment — Korea Lesson

The Survey proposes a specific fix: entrepreneurial safe spaces, implying defined zones where experimentation is allowed, accountability rules match the actual task, and learning is built in from the start.

The logic is solid. No country that successfully transformed its economy made its entire bureaucracy entrepreneurial. Japan’s sectoral teams, Korea’s planning agencies, and Singapore’s regulatory bodies, none of which redesigned the whole state. Instead, they created specific pockets where different rules applied: decisions were shielded from retrospective blame, teams could run experiments and document what they learned, and accountability asked “was this well-reasoned?” rather than “did this succeed?”

Korea is the Survey’s main example. Its early export drive forced firms to meet tough foreign standards before domestic ones existed. This created useful discipline, less room for arbitrary decisions, and more pressure to learn. Over time, the Korean state invested in standards bodies and testing infrastructure, treating them not as bureaucratic overhead but as tools of industrial strategy. Korea moved from following other countries’ rules to writing its own. In many industries, the standards the world follows today are Korean.

For India, the Survey suggests: mission-based teams focused on manufacturing, energy, logistics, or urbanization; regulatory sandboxes beyond just fintech; legal protection for good-faith decisions; and review mechanisms that prioritize learning over blame.

According to the Survey, experimentation must pass these 3 smell tests:

1. bounded,

2. reversible, and

3. accountable

An unbounded experiment is recklessness. An irreversible one is irresponsible. And one that is unaccountable is just discretionary power with a better name.

Policy Hysteresis: When the State Gets Stuck

One of the more interesting arguments in the chapter is about policy hysteresis, meaning what happens when temporary measures become permanent, when benefits harden into entitlements, and when reversing a policy becomes politically or legally impossible.

The pattern is familiar. Once people reorganize their lives around a government intervention, say a subsidy, a tariff, a land-use exemption, or a large contract, unwinding it causes litigation, protests, and disruption. Officials anticipate this, so they avoid experimentation in the first place. Any experiment that works becomes a permanent commitment. Any experiment that fails becomes a personal liability. The safest move is not to experiment at all. The state learns to delay rather than decide.

The Survey’s solution is to build reversibility into policy design from the start: sunset clauses, staged rollouts, conditional approvals, and mandatory reviews after implementation. Hold officials accountable for what they knew at the time, not for what happened afterwards.

There is a line in the chapter that sums up this phenomenon: “A state that cannot reverse cannot learn; a state that cannot learn cannot adapt.” This is not a slogan. It is a description of how institutional decay works, not through sudden collapse, but through the slow accumulation of decisions that can never be revisited.

What the Survey does not fully confront is who pays the political cost of reversal. Sunset clauses sound sensible. But every scheme wound down in India has produced a constituency that showed up to fight it. The farmer who planned around a subsidy rate. The firm that built capacity, assuming a tariff would hold. The official who attached her department’s targets to a program that now needs scrapping. These are real people. Building reversibility into policy requires political leadership willing to take short-term pain for long-term institutional health.

Organizational Design: Nobody Owns the Outcome

Beyond incentives and accountability, the Survey identifies a structural problem with how Indian government organisations work: nobody is clearly responsible for the final result.

A function-based structure is an organizational system where people are grouped according to their specialization, for example, finance with finance, legal with legal, planning with planning, and so on. This setup has clear strengths. When professionals work alongside others in the same field, their expertise deepens, standards become consistent, and institutional memory builds over time. The organization feels stable, orderly, and technically strong.

The problem arises because real-world challenges are not neatly divided by function. Take urban air pollution in India. It is simultaneously a transport issue, an industrial regulation issue, a public health issue, an environmental issue, and a municipal governance issue. In a function-based system, each department handles its assigned part, but no single unit owns the outcome: cleaner air. Work gets done, meetings are held, files move, and responsibilities are technically fulfilled. Yet if pollution levels do not fall, accountability becomes blurry because the system was designed to manage specialized tasks, not integrated results. This is why people say tasks have owners, but outcomes do not.

Function-based structures are excellent at preserving expertise, but they struggle with cross-cutting problems that require coordination and clear ownership. The issue is not incompetence; it is architecture. And organizational architecture quietly shapes behavior, determining whether anyone can truly be responsible for results.

The Survey draws a parallel with how large companies reorganized over time. Firms built on strict functional lines like finance, operations, HR, and legal often found that departmental excellence did not add up to good overall performance once they grew or diversified. Many shifted toward models that kept functional expertise but added leaders explicitly responsible for end-to-end outcomes. Someone owned the result, with the authority to set priorities, make trade-offs, and report on progress.

The lesson for India’s public sector is simple: someone needs to own the outcome, not just a piece of the process.

This is not a glamorous reform. There are no press releases about outcome ownership. But this is where strategies die: sound in conception, fragmented in execution, with nobody clearly responsible for the distance between the two. India has produced dozens of inter-ministerial committees and task forces that met regularly, filled binders with minutes, and delivered approximately nothing; not because the people in them were incapable, but because nobody owned the outcome and everybody could block the action. The Survey names this clearly. What is less clear is how you fix it without confronting the departmental turf battles that the current structure was built to protect.

Private Sector’s Critical Role

The Survey describes large Indian corporates as stuck in an uncomfortable middle position. On one side, you have the companies of East Asia, like Korea’s Samsung and Hyundai, Japan’s Toyota and Sony, that pushed hard for better roads, better courts, better ports, and better-trained workers because their survival in global markets depended on it. When your competition is international, and your government cannot build reliable infrastructure, you have a very strong reason to demand a more capable state. On the other side, you have the private sector of mature Western economies, which operates at arm’s length from government, competes on its own merits, and does not particularly expect the state to bail it out when things go wrong.

India’s large corporations are neither. They do not face enough international competition to need a truly capable state. And they are close enough to the government to get protection, bailouts, and favorable contracts when they need them. So they have settled into a comfortable middle ground that is, not pushing for better institutions, but not quite operating without them either.

The mechanism by which this damages governance is worth understanding clearly. When large firms can make good money through connections rather than through competitiveness, through a well-timed tariff, a renegotiated government contract, a bank loan that gets quietly restructured, they stop demanding the things that a capable state needs to provide. Fast courts do not matter much if you can settle disputes through relationships. Predictable regulation does not matter much if you can get a quiet word with the ministry. Skilled workers at scale do not matter much if your business model does not require them. The result is that India’s most powerful economic actors, who could be the loudest voices for better institutions, instead become some of the loudest voices for more discretion. And discretion, over time, eats away at the rule-based institutions that a developing economy actually needs to grow.

The state, for its part, learns to manage the economy through deals rather than through rules. Every bailout, every forbearance, every protection extended to a connected firm teaches the next firm that the right strategy is not to compete better but to negotiate better. The loop closes on itself, and each round makes the next round harder to break.

The Survey draws on international comparisons to make the case for what a different kind of private sector looks like. Post-war America saw companies like General Motors, Ford, and Caterpillar supply the machinery that underpinned the Marshall Plan’s reconstruction of Europe. This arrangement was commercially profitable, but explicitly aligned with national strategic goals. Bell Labs, IBM, and RCA treated investment in semiconductors and computing as both a business strategy and a national mission. The executives who ran these companies in the 1950s and 1960s talked publicly about duties to the country, not just to shareholders. This was not just rhetoric. It shaped investment decisions, shaped where they hired, and shaped how they thought about what they were building.

Post-war West Germany built its recovery around what it called the Soziale Marktwirtschaft, or the social market economy. The idea was simple: firms should compete hard and win on merit, but they should also accept obligations to workers, communities, and the broader health of the economy. Firms like Siemens, BASF, and Volkswagen were not charity operations. They were intensely competitive. But they operated within a framework that treated a strong company and a healthy society as connected goals, not competing ones.

Japan’s major manufacturers accepted coordinated investment and export targets aligned with national economic goals. Korean chaebol (Samsung, Hyundai, LG) took on enormous financial risk in industries that required years of losses before they became profitable, accepting exposure to international competition as a necessary proof of national capability.

Taiwan’s economic transformation ran through dense networks of small and mid-sized firms that coordinated on standards and technology through business associations, treating the island’s development as a collective project rather than a zero-sum competition.

Singapore’s private sector internalized something even more fundamental: that a small city-state with no natural resources survives only through its reputation for reliability, and that every firm’s behavior in international markets either builds or erodes that reputation.

Across all of these very different cases, the common thread is not a specific policy or a particular model of state intervention. It is a shared understanding among business leaders that their firms are embedded in a national project; that profit and national interest are not naturally opposed but can, with the right incentives and the right culture, point in the same direction.

The Survey asks India’s private sector to develop a similar understanding. I think the ask is right. I also think it is far too gently made.

The Indian corporate sector has benefited from state support, protection, and financial forbearance on a very large scale for a very long time. Tariffs that kept foreign competition out. Public banks that restructured loans that should have been written off. Government contracts were awarded on terms that would not survive competitive scrutiny. Infrastructure built at public expense has raised the value of private assets. The relationship between the Indian state and Indian big business has been, for much of independent India’s history, a relationship of mutual accommodation. The state provided protection and access, and business provided political support and, in some cases, financing. What it has not been, for the most part, is a developmental partnership of the kind the Survey is describing.

Citizens, Shortcuts, and the Quiet Tax on Governance

The Survey identifies a third actor in its account of state capacity: ordinary citizens, and makes an argument that is analytically correct but politically uncomfortable in ways the document does not fully acknowledge.

Every rule that has to be policed rather than followed voluntarily drains administrative capacity. Every policy that requires coercion rather than willing compliance pulls the state away from more complex tasks. Where civic norms are weak, the state has to spend its limited energy on enforcement rather than on the problem-solving that development requires. How citizens behave is not a secondary or moral question. It shapes what the state can actually do.

The Survey’s most striking observation here is the gap between how Indians treat private spaces and public ones. Homes are kept with great care. Streets, drains, and railway tracks are treated as nobody’s problem. The Survey compares India to Sri Lanka and parts of Africa and Latin America, not because those places produce more careful individuals at home, but because they have extended the ethic of care into shared space. Where that shift happens, and where the state matches it with reliable services, public behavior changes. The difference is less about individual character than about whether public space feels worth caring for.

Here is the complication the Survey does not fully engage with. The citizens described above, those who jump queues and negotiate enforcement, are not being irrational. They are responding to a state that has historically been unreliable: one that collects taxes but does not fix the drain, that sets rules but enforces them selectively, and that demands patience in queues run by officials with no accountability for wait times. Civic behavior and state reliability are connected. Citizens who experience public systems as arbitrary learn that the rules are not worth following.

You fix it, slowly and with no shortcuts, by making public systems reliable enough that following the rules becomes the sensible thing to do. The Survey knows this. It says so, briefly. But it places too much weight on what citizens owe the state, and too little on what the state owes citizens first.

Deregulation as State-Building: What the Evidence Shows

The chapter closes with a detailed account of the Task Force on Compliance Reduction and Deregulation, set up in January 2025 under the Cabinet Secretary to drive simplification across all 36 States and Union Territories. The results, as of January 2026, are more encouraging than most observers would have expected.

Of the 828 priority reforms identified across 23 areas and all states, 630 (76%) had already been completed. Another 79 were underway. The five sectors covered were land use, building and construction, labour, utilities and permissions, and a broader set of overarching reforms.

Two case studies stand out. In the Andaman and Nicobar Islands, moving the Change in Land Use process online cleared hundreds of applications within months, created new tourism capacity, and improved credit access for households and small businesses. In Tripura, a comprehensive reform package helped convert commitments from the Rising Northeast Investors Summit 2025 into actual investment on the ground, with Tripura accounting for a meaningful share of what the whole Northeast region attracted.

What distinguishes this exercise from earlier deregulation drives is not just the number of reforms but how it was run: coordination across agencies, real-time monitoring through a digital platform, and working through problems with states rather than issuing instructions from Delhi. It was designed to be continuous, not a one-time push.

The core lesson is one the Survey has been building toward all along. Deregulation is usually framed as the state stepping back. The Survey inverts this correctly. When you remove low-value compliance burdens, you free up capacity for the coordination problems that only the state can solve. Deregulation is not a retreat. It is a reallocation of limited energy toward the places where it actually matters.

India’s software industry is the most instructive example of this logic. From the 1960s to the early 1980s, the electronics sector was tightly controlled through state enterprises, import substitution, and pervasive restrictions on private and foreign participation. Despite substantial public investment, the state could not make it work. The shift came in the mid-1980s, when officials in the Department of Electronics moved away from control toward facilitation. They recognized software as an industry, eased foreign exchange restrictions for exporters, invested in technical education, and created technology parks with reliable power and single-window clearances. In 1985, the government approved Citibank’s private satellite link for remote software services. Texas Instruments was permitted to set up a development center in Bangalore shortly after. N. Seshagiri, one of the officials driving this shift, later said candidly: “We broke 26 separate rules to accommodate Texas Instruments’ Bangalore subsidiary and are willing to break more.”

What followed was one of India’s few genuine development success stories, an industry built not by picking winners but by creating conditions in which many firms could compete, learn, and connect to global markets. India did this once, through officials willing to exercise judgment and bend the rules their own system had written. The Survey is asking whether India can do it again, this time, by changing the rules so the next generation of officials does not have to break them.

State capacity, as the chapter carefully argues, is not one reform, one dashboard, one app, or one charismatic administrator. It is an ecosystem outcome. It is shaped by how decisions are taken, how risks are managed, how failures are treated, how regulation is designed, how bureaucracies are organized around outcomes rather than files, and how incentives shape the everyday behavior of firms and citizens. Weakness in any one part quietly corrodes the rest. Capacity is alignment.

Equally important is the Survey’s refusal to romanticize the State as a lone actor. Capacity is co-produced. When firms compete through productivity rather than lobbying, they shift the State’s role from negotiator to referee. When citizens comply not because they fear punishment but because norms are internalized, enforcement burdens fall. Administrative energy is freed for coordination and delivery. Institutional performance improves without necessarily expanding government size. That is an underappreciated insight.

The discussion on compliance reduction and deregulation illustrates this beautifully. Deregulation, done seriously, is not about weakening the State. It is about strengthening it. When rules are simplified, responsibilities clarified, and processes made time-bound, the State moves from low-value policing to higher-value problem-solving. The goal is not fewer rules; it is better rules that free administrative bandwidth. In that sense, governance reform becomes state-building.

The chapter’s most provocative claim, however, is about exposure. Manufacturing, exports, and global value chain integration are not just growth strategies. They are institutional stress tests. They force systems to become predictable, reliable, and competitive. Protected ecosystems can hide inefficiency for decades. Global integration reveals it quickly. Strategic indispensability cannot be built in a sheltered room.

The world ahead will not offer certainty. Risk is structural. The advantage will lie with countries that can act before clarity arrives, correct course without paralysis, and align incentives across state, market, and society. That requires a bureaucracy comfortable with experimentation, firms willing to compete without protection, and citizens who treat public systems as shared assets rather than arenas for extraction.

State capacity, then, is not a technocratic footnote. It is the central political economy challenge of the next decade.

Every winter, Indian cities in the North turn unbreathable.

Schools shut. Hospitals fill up. Doctors issue advisories that sound less like public health guidance and more like evacuation notices. Vulnerable citizens are told, without embarrassment, to leave town if they can. A few hundred people gather at India Gate, holding placards and wearing masks, protesting the air that feels like an administrative failure.

Subscribe now

And then, quietly, it ends.

The smog thins. Another crisis takes its place. Social media moves on. Life resumes. Policy inertia remains intact.

The same pattern repeats elsewhere. In Indore, people die after drinking contaminated water. There is shock, outrage, and accusations. Online anger surges. Official explanations arrive late and sound rehearsed. Within days, the story slips out of public consciousness. Systems stay exactly as they were.

After horrific crimes against women, candles are lit. Marches are organized. Statements are issued. And then, once again, the moment passes. Institutions absorb the anger and continue largely unchanged.

This is not a story about ignorance. The Indian middle class knows something is wrong. Nor is it a story about indifference. The anger is real, sustained, and increasingly articulate. The puzzle lies elsewhere.

Why does so much awareness produce so little pressure? Why does a class that complains endlessly fail to force the system to listen?

In this dispatch, I lean into public choice theory to unpack this paradox. Drawing on concepts like the logic of collective action and the dynamics of diffused costs versus concentrated benefits, the essay shows how structural incentives, and not just apathy, explain why pollution, unsafe cities, and persistent governance failures endure despite loud middle-class frustration. The everyday anger must be seen as the predictable output of incentives, and we should ask what it would take to finally turn awareness into pressure that reshapes policy and institutions.

The Indian middle class today is not merely larger than it once was. It is qualitatively different.

It has travelled. It has seen cities that function. It has experienced public transport that works, streets that feel safe, and air that does not require apps to monitor survival. Governance is no longer an abstract promise; it is something encountered abroad and missed at home.

This exposure has altered expectations. Bad infrastructure no longer feels inevitable. Pollution no longer feels like the price of development. Women’s safety is no longer framed as a social flaw alone, but as a failure of policing, design, and enforcement.

The middle class has become a harsh evaluator of the state.

And yet, it has not become its adversary.

This is precisely the terrain that public choice theory helps illuminate. Rather than asking why citizens behave badly, it asks how rational individuals behave inside imperfect institutions.

It forces us to confront an uncomfortable possibility: that widespread middle-class frustration is not a sign of democratic failure due to apathy, but the predictable outcome of incentive structures that make coordination costly, confrontation risky, and private adaptation easier than collective action.

Public choice theory asks us to stop assuming that sensible individual choices add up to good collective outcomes. Often, they do the opposite. When opting out is easier than speaking up, systems quietly decay. And when organizing is costly while benefits are shared by all, even widespread anger fails to turn into collective action.

There seems to exist a structural tension at the heart of contemporary middle-class India: the sharper the evaluation of state failure, the greater the temptation to step around the state altogether. And every time that happens, the possibility of collective pressure weakens just a little more.

The Logic of Collective Action: Why Numbers Don’t Automatically Mean Power

The intuitive belief that large numbers automatically translate into political power is one of democracy’s most comforting myths. If millions suffer from the same problem, we assume, surely pressure will build. Surely mobilization will follow. Surely governments will be forced to respond.

Mancur Olson shattered this assumption with a simple but deeply unsettling insight: size, by itself, is often a liability rather than an advantage.

Olson’s core argument is not about apathy or moral weakness. It is about incentives. In large groups, the benefits of collective success are shared by everyone, regardless of who actually participates. But the costs of participation, which essentially include time spent organizing, energy expended protesting, and risks of visibility or confrontation, are borne individually. This asymmetry creates a structural imbalance. Each recognizes that their personal contribution is unlikely to determine the outcome, while their personal costs are immediate and certain.

The rational response, then, is to wait.

Everyone hopes that someone else will act. Everyone assumes that if the problem is serious enough, others will step forward. And because this logic applies to everyone simultaneously, collective action collapses before it even begins. What looks like indifference from the outside is, from the inside, a perfectly rational calculation.

This is not an abstract theoretical problem. It plays out daily in the political behavior of the Indian middle class.

Consider the nature of middle-class grievances. Clean air, safe streets, functional infrastructure; these are not niche demands benefiting a narrow constituency. They are broad public goods. Everyone benefits from them. And precisely because everyone benefits, no one has a strong incentive to personally bear the costs of fighting for them.

Unlike small, organized groups such as unions, professional associations, and industry lobbies, the middle class does not face a clear “do or die” moment tied to a single policy decision. Its losses are incremental rather than catastrophic. Pollution shortens lives, but rarely in ways that feel immediate or traceable. Infrastructure failure wastes hours, not livelihoods. Unsafe streets create fear, not necessarily confrontation for everyone at once. The harm is real, but it is dispersed over time and across people. This makes mobilization emotionally difficult and strategically unclear.

Without clear targets, mobilization loses direction. There is no single ministry to besiege, no specific official to corner, no deadline around which to organize sustained pressure. Responsibility is diffused across multiple agencies, levels of government, and informal power structures. Each actor can plausibly claim partial responsibility and therefore deny full accountability.

For the individual citizen, this diffusion produces a sense of futility. Attending a protest may feel morally satisfying, but it rarely produces visible returns. Filing complaints is time-consuming and often met with bureaucratic indifference. Sustained civic engagement demands persistence without guarantee. Non-participation, by contrast, carries no stigma and minimal cost.

Over time, people learn that silence is cheaper than confrontation, that private solutions are more reliable than public ones, and that outrage expressed online is safer than pressure applied offline. Frustration accumulates, but it accumulates privately. Anger circulates, but it does not consolidate.

Seen through this lens, the political weakness of the Indian middle class is not surprising. It is the predictable outcome of a large group pursuing diffuse public goods in an institutional environment that raises the cost of coordination while lowering the cost of exit. Numbers exist, but leverage does not. Awareness grows, but power remains elusive.

Two Ideas That Explain More Than They Should

To understand why the Indian middle class feels angry yet politically ineffective, we need to pause and understand two ideas from public choice theory. The first idea is called diffused costs and concentrated benefits. The second is its mirror image: concentrated costs and diffused benefits.

Once you see how these work, middle-class apathy begins to look less like a moral failure and more like a predictable outcome.

Let’s start with diffused costs and concentrated benefits.

This describes a situation where a large number of people each suffer a small loss (diffused costs), while a small group enjoys a large gain (concentrated benefits). The total damage to society may be enormous, but because the harm is spread thinly, no single person feels injured enough to fight back.

A simple example helps. Imagine a government policy that adds ₹100 a month to everyone’s electricity bill. For most households, this is annoying but not enraging. It does not trigger protests. It does not feel worth taking a day off work to oppose. But now imagine that this policy funnels crores of rupees to a small group of power producers. For them, the benefit is huge, visible, and worth defending aggressively.

Who will organize? Who will lobby? Who will show up, repeatedly, to protect their interests?

Public choice theory, associated with thinkers like James Buchanan and Gordon Tullock, tells us the answer is obvious. The group that gains a lot will fight hard. The group that loses a little will mostly stay quiet.

This is not because people are stupid or lazy. It is because the cost of mobilization is higher than the individual loss.

Now apply this to air pollution. Every middle-class household pays a small, invisible price: more coughs, slightly lower stamina, higher long-term health risks, and a bit more spending on purifiers or medicines. The damage is real, but it arrives slowly and unevenly. No single morning feels like the moment to revolt.

Meanwhile, lax enforcement benefits a small set of actors, construction firms, transport operators, contractors, intermediaries, who save real money every day regulations are ignored. For them, the incentive to resist change is strong and immediate.

The result is predictable. Pollution continues. Outrage spikes briefly. Nothing structural changes.

Now turn to the second idea: concentrated costs and diffused benefits.

This describes the opposite situation. Here, a small group has to bear a clear, immediate cost (concentrated costs), while the benefits are spread thinly across society over time (diffused benefits). These policies are often good for society—but politically difficult.

Think of stricter traffic enforcement. It may inconvenience certain groups, such as drivers who violate rules, transport operators who cut corners, and even local officials who benefit from discretion. These costs are concentrated and visible. People know exactly what they are losing.

But the benefits, like fewer accidents, safer streets, and better urban life, are spread across millions of people. No single individual can point to a specific gain and say, “This happened because of that reform.” The reward is collective, gradual, and hard to claim credit for.

From the perspective of a political or bureaucratic decision-maker, this is a bad deal. Why take on angry, organized losers today for benefits that are diffuse, delayed, and politically unattributable tomorrow?

So reforms stall. Or they are announced but weakly implemented. Or they appear briefly and then fade.

Now bring these two ideas together, and the Indian middle class comes sharply into focus.

Most middle-class grievances fall into these two traps.

Pollution, bad infrastructure, contaminated water, unreliable civic services, these impose diffused costs. Each family suffers a little. The pain accumulates, but it does not concentrate. As a result, anger spreads wide but thin. It produces complaints, not coalitions.

At the same time, fixing these problems usually involves concentrated costs for specific actors: stricter enforcement, loss of discretion, higher compliance costs, and disrupted informal arrangements. These actors are few, organized, and motivated. They push back hard.

The middle class sits awkwardly in between. It is large, vocal, and aware but structurally weak. Its losses are not sharp enough to force collective sacrifice. Its gains from reform are too diffused to generate sustained mobilization.

What looks like apathy is often a rational response to weak incentives.

People adapt instead. They buy private solutions. They move to better neighborhoods. They install purifiers. They rely on bottled water. They plan exits within the city, within the country, or abroad.

Each decision makes sense individually. Collectively, they drain pressure from the public system.

Seen this way, middle-class apathy is a misleading label. What we are really seeing is a mismatch between the scale of problems and the incentives for action.

The middle class is not indifferent. It is constrained. Its costs are diffused. Its benefits from reform are collective and delayed. The costs of mobilization are immediate and personal.

Public choice theory does not excuse this outcome, but it explains it. And once explained, the moral story changes. The question is no longer, “Why don’t people care enough?” It becomes, “What would make caring rational?”

Until that question is answered and until incentives are redesigned so that costs and benefits align with collective action, the cycle will continue. Anger will rise. Social media will flare. Private adaptations will multiply.

And the system will learn, once again, that it can survive middle-class frustration without having to respond to it.

Let’s understand this using women’s safety as an example.

Women’s safety appears, at first glance, to defy public choice logic. The moral stakes are high. Public outrage is intense. Media attention is immediate. Surely this is an issue capable of overcoming collective action problems.

And yet, change remains slow and uneven.

The reason lies in concentrated costs and diffused benefits. Meaningful improvements in women’s safety require actions that impose immediate, visible costs on political and bureaucratic actors: reallocating police resources, enforcing inconvenient laws, confronting informal power structures, and investing in urban design that yields no quick political dividends.

The benefits of these actions, however, are gradual and widely shared. Safer cities do not produce instant, attributable wins. From a political incentive perspective, this is a poor bargain: high upfront costs with uncertain, diffused rewards.

Public choice theory predicts what we observe. Symbolic actions substitute for structural reform. Committees are announced. Statements are issued. Surveillance is intensified briefly. But the bigger institutional changes that are costly, complex, and politically risky are postponed.

One of the most uncomfortable conclusions of this analysis is that the Indian middle class is a weak interest group. Its political engagement is primarily individualistic. Voting occurs periodically. Complaining is constant. Exit is increasingly common. Private schools, private healthcare, gated housing, and overseas mobility function as substitutes for public goods.

This pattern is illuminated by Exit, Voice, and Loyalty. When exit becomes easier than voice, systems decay quietly. Each act of private escape reduces the incentive to invest in collective repair. Over time, the capacity for collective pressure erodes.

The middle class does not lack opinions. It lacks organizational architecture.

Economic growth rarely comes from one single cause. It usually grows out of many layers of change, some easy to see, others quietly shaped by social and institutional forces over time. Universities are often praised as drivers of innovation, but their real impact on regional growth is usually slow and hard to see. Research across economics and economic history now shows that their value goes beyond teaching students. Universities help build knowledge ecosystems that shape how economies learn, adapt, and innovate over time.

This literature situates universities not as isolated producers of skills but as institutional nodes in broader systems of knowledge creation and diffusion. Their economic impact operates through a sequence that is increasingly well understood: universities accumulate human capital; that human capital enables innovation; innovation, when embedded locally, generates agglomeration and sustained growth. The power of this sequence lies not in any single link, but in its cumulative and path-dependent nature.

Universities are often seen as places that simply produce skilled workers, but their impact goes much deeper than that. Yes, they train people. But what they really create is a shared base of knowledge, habits, and ways of thinking that shape how a region responds to change. Human capital is not just about how many educated people live in a place; it is about how those people help societies learn, adapt, and solve problems over time. Earlier economic thinking treated education like another input, similar to machines or raw materials. Newer research shows something more subtle: education changes how institutions function, how ideas circulate, and how quickly societies can adjust when new technologies or opportunities emerge. In that sense, universities don’t just supply workers; they quietly reshape the capacity of an entire region to evolve.

Empirical research consistently shows that regions with higher concentrations of tertiary-educated workers experience faster long-term growth. Valero and Van Reenen’s cross-country study of over 1,500 regions demonstrates that increases in university density are associated with higher future GDP per capita, even after accounting for pre-existing economic conditions. Crucially, these effects materialize with a lag, suggesting that universities operate through slow-moving mechanisms rather than immediate demand stimulus.

What distinguishes universities from other forms of human capital investment is their durability. Graduates often remain embedded in local labor markets, shaping managerial practices, organizational routines, and professional norms. Over time, this accumulation changes the composition of local labor markets, tilting them toward problem-solving, abstraction, and technical coordination, capacities that become increasingly valuable as economies move up the value chain.

Human capital alone does not guarantee innovation. What matters is the interaction between skilled individuals, institutional environments, and opportunities for recombination. This is where universities function not merely as training centers but as epistemic hubs.

A substantial body of work shows that knowledge spillovers from universities are geographically bounded. Research by Jaffe, Trajtenberg, and Henderson famously demonstrated that patent citations cluster spatially, indicating that ideas travel more easily across short distances. Hausman (2022) links the diffusion of university innovation (post–Bayh-Dole in the US) to industry agglomeration, with faster growth in employment, wages, and corporate innovation where industries are closer to nearby universities’ research strengths; effects are stronger with proximity, density, and local skills.

Later work by Kantor and Whalley confirms that firms located near universities benefit disproportionately from academic research, particularly when their technological domains overlap. These spillovers are not primarily the result of formal technology transfer offices or patent licensing. Rather, they occur through informal channels: labor mobility, professional networks, collaborative problem-solving, and shared technical languages. Firms learn not just from published research but from proximity to people who know how to think in certain ways. This is why university quality and specialization matter more than sheer size.

Once innovation takes root, it reshapes economic geography. Firms cluster not merely to access markets or infrastructure, but to tap into dense knowledge environments. This is where the feedback loop between universities and agglomeration becomes self-reinforcing. Agglomeration economies arise from three main mechanisms: shared labor pools, input-output linkages, and knowledge spillovers. Universities contribute to all three. They stabilize local labor markets by producing a steady flow of skilled workers; they attract complementary firms seeking access to talent; and they generate continuous streams of ideas that reduce the cost of experimentation.

China’s Long Experiment with Knowledge, Place, and Power

Some policies are designed to solve a problem; a few end up reshaping history. China’s 1950s university relocation programme belongs to the latter category. Conceived during the early years of the People’s Republic as an ideological and administrative reordering of higher education, it was never meant to be a growth strategy in the modern economic sense. And yet, seven decades later, its effects are etched into the geography of Chinese industrial development.

As the paper by Jianyong Fan, Wei Tang, and Feng Zhang shows, this reshuffling of academic departments created long-lived economic spillovers that only became visible once China liberalized its economy after 1978. The relocation of universities seeded knowledge in places that would later become industrial hubs, revealing how human capital can quietly outlast ideology, policy cycles, and even economic systems.

Between 1949 and 1957, China undertook one of the largest higher-education restructurings in history. Over 70% of university departments were moved, and 33.7% were relocated across cities, often to regions with limited prior educational infrastructure. This was not incremental reform. It was a systemic rewiring.

The motivations were explicitly political and developmental:

• to dismantle Western-style comprehensive universities,

• to align higher education with Soviet planning ideals, and

• to redistribute technical training capacity across the country

The placement of departments was largely disconnected from local economic conditions at the time. Indeed, the data show no strong correlation between pre-1949 industrial strength and where departments were sent. This allows the authors to treat the relocation as a plausibly exogenous shock, an economist’s dream.

At the time, this reorganization produced no immediate economic dividends. During the planned economy era, firm behavior was tightly constrained, labour mobility was limited, and innovation had little room to translate into productivity.

The turning point arrives after 1978, when market reforms allow firms to respond to incentives and tap into local capabilities. This is where the paper’s most striking results emerge.

Using county–industry level data from 1933 to 2004, the authors show that areas receiving relocated university departments experienced significantly higher long-term industrial growth, but only after markets were liberalized. In 1995 and 2004, counties with relocated departments saw:

• Higher industrial employment

• Greater firm entry

• Higher value-added per worker

What exactly do universities do that factories do not?

The paper shows that universities matter not simply because they exist, but because of how they interact with local economies.

The mechanisms through which universities shape local economies become clearest when we look at how knowledge actually moves through space and time. One of the most important channels is knowledge spillovers. Using patent data from 1985 to 2017, the study shows that counties which received relocated university departments experienced a marked increase in innovative activity, particularly in fields closely aligned with the academic strengths of those departments. This was not just a rise in raw patent counts. It included deeper forms of innovation: more co-patenting between universities and firms, greater citation intensity, and stronger collaborative ties between academic researchers and industry. These patterns reinforce a central insight from endogenous growth theory, that ideas do not diffuse evenly or instantaneously. They travel through people, relationships, and proximity.

Beyond innovation, universities also reshaped local labour markets in lasting ways. Regions that hosted relocated departments developed a denser pool of skilled workers over time. These areas saw higher proportions of college-educated individuals, a larger technical workforce, and a stronger presence of younger, highly skilled cohorts after 1980. This generational pattern matters. It suggests that universities do not simply raise skill levels once and move on; they create self-reinforcing pipelines of talent.

At the firm level, these changes translated into measurable productivity gains. Firms operating in counties with relocated university departments recorded significantly higher total factor productivity by the mid-1990s and early 2000s. The effects were especially pronounced among small and medium-sized enterprises, precisely the firms most dependent on external knowledge, skilled labor, and informal networks. Large firms can often import expertise or internalize research capacity; smaller firms rely far more on their surrounding ecosystem.

The paper also resists romanticizing knowledge diffusion. Spillovers weaken with distance. Firms located within 20–50 km of relocated universities benefit significantly more than those farther away. Beyond 150 km, effects are minimal.

This reinforces a crucial point: knowledge is sticky. It travels through people, networks, and repeated interaction, not simply through policy declarations.

One of the most striking contributions of the study is its comparison between university relocation and large-scale industrial investment programs such as the 156 Projects and the Third Front Movement.

While those projects delivered short- to medium-term gains, their effects decayed faster. In contrast, university-driven growth proved more durable, more adaptable, and more innovation-oriented.

China’s university relocation policy was not designed as an economic growth engine. Yet it became one, quietly, gradually, and enduringly. The evidence shows that when knowledge institutions are embedded deeply enough, they outlive regimes, outlast policies, and reshape economic geography long after their architects have disappeared.

In an era obsessed with quick wins and visible infrastructure, this research offers a humbling counterpoint: the most transformative investments are often those whose returns cannot be rushed.

IISc and the Making of India’s Technology Capital

Long before Bangalore became India’s “Silicon Valley,” the city was already home to a deep institutional seed planted in 1909: the Indian Institute of Science (IISc). From the beginning, IISc was conceived not just as a teaching institution, but as a research and technical capability center, drawing talent, producing advanced knowledge, and linking with industry in areas ranging from aeronautics to materials science.

When we talk about Bangalore’s rise as the heart of India’s IT revolution and later its broader tech ecosystem, two features of IISc stand out. The first is human capital creation. IISc’s graduates and researchers didn’t simply join firms; they became founders of firms, connectors across networks, and talent magnets for the city. The second is knowledge linkage: IISc maintained ties with both government labs and emerging private tech entities, through collaboration, technology licensing, and incubator support, long before Bangalore’s tech parks sprouted.

By the 1980s and 1990s, when global demand for software services exploded, and Indian liberalisation opened new markets, Bangalore was uniquely positioned to respond. It wasn’t just state policy or cheap talent that mattered; it was the local ecosystem of skilled people, research, and experimentation that had been building quietly for decades. Firms like Infosys and Wipro found both employees and intellectual collaborators within the city’s research networks, while the presence of IISc lent credibility and depth to Bangalore’s tech profile.

This pattern echoes one of the central insights from the China university-relocation research: university presence does not immediately transform an economy, but it reserves human and knowledge capacity that can be activated when markets and institutions allow. In China’s case, the long-dormant effects of relocated departments became visible only after market reforms in the late 1970s; in Bangalore’s case, the human capital and research infrastructure cultivated through IISc became catalytic once the IT industry found a global opportunity in the 1980s–90s. In both instances, universities created latent capabilities clusters of skilled people, networks of ideas, and institutional legitimacy that local economies could draw on when conditions aligned.

Where the China example highlights how latent academic capacity can later become industrial strength, Bangalore’s story shows a complementary pathway: a research university that remained active and increasingly integrated with industry effectively grew with the local economy. In both narratives, the key is not just that universities produce graduates, but that they help build endogenous innovation ecosystems, places where human capital, knowledge creation, and firm dynamism become mutually reinforcing rather than isolated elements.

The story of Wipro’s emergence as a major player in India’s information technology sector is often told through business histories and market milestones, but the Quartz India article underlines a less obvious but economically profound driver of that success: the presence of IISc in Bangalore. Long before Wipro became a global tech services brand, the company was a diversified industrial firm best known for products like vegetable oils. Its leap into computing in the late 1970s and early 1980s was not the result of a ready-made internal capability, but of an intentional collaboration with IISc’s Digital Systems Lab and its School of Automation. This partnership, formalized through a ₹14 lakh consultancy contract in 1979, became the technical and institutional foundation for Wipro’s first minicomputer systems.

At a moment when multinational corporations such as IBM were retreating from the Indian market, local firms like Wipro saw both a market gap and an opportunity to harness domestic expertise. IISc’s researchers and facilities provided the specialized knowledge in microprocessors, systems architecture, and hardware development that Wipro lacked but urgently needed. The collaboration was not superficial; Wipro’s early R&D team worked within IISc’s Digital Systems Lab, debating design choices, experimenting with Intel’s new microprocessor systems, and essentially co-developing the technical blueprint for what became the Wipro Series-86 minicomputer in 1981.

The impacts of this collaboration were structural rather than transactional. Wipro’s initial revenues from the computer systems business, about ₹2 crore in the first year, may seem modest by today’s standards, but the stronger effect was the creation of an indigenous IT capability that anchored the company’s future strategic trajectory. Beyond hardware, the collaboration produced human capital: several of Wipro’s earliest IT employees were trained at IISc, embedding research skills, rigorous engineering practices, and an innovation mindset into Wipro’s culture. That long relationship has endured; even four decades later, Wipro continues to recruit from IISc and co-invests in research initiatives such as the Wipro-IISc Research Innovation Network (WIRIN), focusing on frontier areas like AI, autonomous systems, and robotics.

What makes this narrative compelling is its economic logic. The article illustrates how a university can act as a capacity-building hub for an entire sector long before that sector fully materializes. This echoes broader development evidence (including Chinese university reallocation policy as seen above) showing that academic institutions can create latent absorptive capacity within a region. In both cases, knowledge infrastructure does not yield immediate, visible growth on its own; rather, it primes the local economy with human capital, technology know-how, and networks that firms can harness when broader economic conditions align. In Wipro’s case, regulatory shifts, market openings, and global demand for IT services provided the context in which the latent capability seeded at IISc transformed into sustained industrial growth.

In short, IISc didn’t single-handedly create Bangalore’s IT boom. No single institution could, but it helped lay the deep groundwork that made Bangalore’s later tech explosion possible, illustrating a dynamic very similar to the university-led spillovers found in the Chinese context.

When scholars look back at the economic aftershocks of the Second World War, their attention almost always turns to the United States and Western Europe. The American story, in particular, has been examined from every angle: how wartime spending helped lift the country out of the Great Depression, how factories shifted overnight into engines of mass production, and how new technologies and organizational capabilities took root. Works such as Arthur Herman’s Freedom’s Forge, Maury Klein’s A Call to Arms, and The Economic Consequences of US Mobilization for the Second World War by Alexander J Field have shown how private enterprise, government planning, and sheer industrial improvisation combined to create what Roosevelt famously called “the arsenal of democracy.” Together, this body of work offers a vivid and detailed picture of what wartime mobilization meant in the West, how it transformed industries, rewired local labour markets, and set the stage for post-war prosperity.

Yet very little of this global discussion looks eastward, toward India, a colony that did not choose to enter the war but was pulled into it through a radio announcement. India was still a poor, mostly agrarian economy, often described as industrially weak, yet the British forced it to produce everything from uniforms and boots to trucks, machine tools, radios, chemicals, and ammunition. This sudden and uneven mobilization touched hundreds of towns and districts, leaving behind changes that were subtle at the time but would quietly shape India’s economic landscape well into the twenty-first century.

A new and very insightful paper by Aneesha Parvathaneni and Dean Yang, titled War Mobilization and Economic Development: World War II and Structural Transformation in India, takes a close look at this forgotten chapter of Indian history. Their work encourages us to rethink many assumptions we hold about how the colonial government operated and how long-term economic change actually happens in India. At the heart of their study is a straightforward but powerful question: Did the huge wartime purchasing program run by the British between 1939 and 1945 leave a long-lasting mark on India’s districts, or did its effects fade away after the war ended and the country became independent?

The paper shows that this wartime mobilization, something India never chose but had to participate in, was not just a temporary spike in production. Instead, it shaped patterns of work, industry, and growth in ways that continued for decades.

A War Declared on the Radio, and the Factories That Followed

On the morning of 3 September 1939, All India Radio announced that India was at war with Germany. The nationalist leadership had not been consulted; there was no parliamentary debate; and the economic consequences of this declaration were not merely unforeseen, they were essentially unchosen. Yet from that moment onward, the British Raj undertook an extraordinary centralization of economic authority, creating a new Supply Department to coordinate procurement, and gradually transforming India into a major manufacturing base for the Allied forces, particularly after Japan’s conquest of Malaya and Burma severed supply lines in Southeast Asia.

Once India was drawn into the Second World War, the country ended up playing a far bigger role in Britain’s war effort than most people realize today. Between 1939 and 1945, India became one of the major arsenals of the British Empire, producing an enormous range of supplies such as uniforms, boots, vehicles, machine tools, steel products, communication equipment, chemicals, medicines, and countless other items needed on different battlefronts. Factories that had earlier worked on small civilian orders suddenly had to meet strict military standards and deliver large quantities on short deadlines. Towns that were known for simple crafts or small workshops were pushed into mass production almost overnight.

The scale of this transformation was extraordinary. The total value of goods the British procured from India during the war was equal to one-sixth of India’s entire national income in 1938, the year before the conflict began. For a colony with low industrial capacity and modest private investment, this level of state-directed demand was unlike anything seen before. No previous colonial policy had mobilized resources, factories, and workers at such speed or such volume.

Historians often describe this wartime procurement drive as the last major economic intervention of the British Raj, because nothing after 1945 came close in terms of size or influence. Once the war ended and India moved toward Independence, the colonial government never again attempted an economic operation of similar scale. In many ways, the war effort forced the Raj to act like a modern industrial planner for the first and only time, and this brief, intense period would leave traces that continued shaping India’s economic landscape for decades. This exercise combined elements of coercive extraction with forms of industrial capability building that had no precedent in the colonial period.

The distribution of this procurement, however, was far from uniform.

The British were acutely aware of Japanese advances along the eastern frontier and avoided deepening supply chains in Bengal and Assam. They channeled contracts toward districts where pre-war production of relevant goods, whether boots, leather goods, machine tools, textiles, or chemicals, already existed to some degree. Some districts received massive orders; others received almost none.

This geographical unevenness is central to the paper’s key argument and becomes easier to visualize through the first map the authors present:

The map reveals clusters of high exposure not only in industrial regions such as Bombay Presidency or parts of Madras Presidency but also in the United Provinces, Punjab, and segments of the Central Provinces. The northeast, by contrast, appears largely blank due to its exclusion from large-scale wartime procurement out of military caution. This spatial heterogeneity, driven not by economic choice but by wartime necessity, provides a rare quasi-experimental window into the long-run effects of a massive industrial policy shock.

To understand how World War II changed different parts of India, the researchers first needed to know where the British placed their war orders. Fortunately, there is a detailed government report written in 1947 that lists everything the British bought during the war and which Indian firms supplied these goods. This includes boots, clothes, chemicals, vehicles, tools, and many other items. The researchers combined this information with data from the 1931 Census, which showed what kinds of industries existed in each district before the war began.

Their basic idea is simple:

If a district already made products the British needed for the war, then that district likely ended up receiving more wartime orders.

So they created a measure that estimates how much wartime demand each district was “exposed” to. This is not an exact count of how much each district produced, but a good prediction based on what they were already making in 1931 and what the British needed in large amounts during the war.

To compare districts properly, the authors had to match old district boundaries with newer ones and make sure they were looking at the same areas over time. Once that work was done, they compared districts that were highly exposed to wartime production with those that were not.

They also checked whether these two sets of districts were already on different economic paths before the war. They found that before 1939, districts looked fairly similar, which means the differences we see after the war are likely due to wartime mobilization, not something that was already happening earlier.

Six Decades of Divergence: The Persistence of a Wartime Shock

The key finding of the paper is that districts more exposed to wartime procurement experienced significantly greater movement of workers out of agriculture and into the modern sector, that is, industry plus services, beginning as early as the 1951 Census, and that this divergence persisted with remarkable stability through 2011. A one-standard-deviation increase in wartime exposure predicts a 6–7 percentage point higher modern-sector employment share six decades later.

As the authors note, the effect “shows up very soon after World War II, by 1951, and persists for six decades, through 2011, with roughly the same magnitude,” underscoring the durable nature of these externally imposed structural changes.

This durability is interesting because it suggests that the wartime mobilization did not simply create a temporary boom in industrial employment that evaporated with demobilization but instead generated capabilities, complementarities, and expectations that allowed districts to sustain a different development path long after the imperial machinery had disappeared.

More surprising still is the decomposition of modern-sector growth across industry and services. Although one might expect the long-run effects to reside mainly in manufacturing, the evidence suggests that services account for the largest share of the structural transformation, particularly local and consumer-oriented services such as trade, personal services, and small-scale transport. This pattern reflects a familiar development mechanism: wartime procurement raises incomes; higher incomes create demand for non-tradable services; and this demand, in turn, supports occupational shifts out of agriculture, even if large-scale industry does not continue to expand at wartime rates.

Heavy Industry, Light Industry, and the Asymmetry of Technological Learning

One of the most illuminating contributions of the paper is the distinction between heavy-industry procurement (vehicles, explosives, chemicals, metals) and light-industry procurement (textiles, boots, simple goods). When the Shiftshare index is recalculated using only light-industry wartime orders, the long-run impacts on structural transformation largely disappear. When it is calculated using only heavy-industry orders, the effects remain large and statistically robust.

This asymmetry speaks directly to longstanding ideas in development economics about technological depth, complexity, and learning-by-doing. Heavy-industry production required Indian firms and workers to adopt new technologies, meet stringent quality standards, organize production at larger scales, and interact with the higher end of global value chains at a moment when such exposure was otherwise rare.

As the authors write, “The majority of effects on structural transformation are driven by procurement of heavy industrial goods,” suggesting that it was the specific technological content of wartime mobilization—not the aggregate volume alone that seeded durable capabilities. In the language of Hausmann and Rodrik, this was a moment of “self-discovery,” albeit coerced, wherein districts learned what kinds of goods they could produce at the frontier when compelled to do so.

This insight has profound implications for how we understand the legacy of wartime industrialization as well as contemporary policy debates in India. State-led industrial interventions can produce long-run economic benefits, but not all interventions are alike. It is the capability-intensive, technologically demanding sectors that generate the deepest and most persistent learning externalities.

The Long Shadow of War: Consumption, Urbanization, Lights from Space, and Migration

To complement the employment analysis, the authors examine a cross-section of contemporary development indicators like consumption, poverty, urbanization, night-time lights, and migration, thereby triangulating the broader welfare consequences of wartime exposure. The patterns that emerge are both consistent with the structural transformation results and evocative in their own right.

Districts with higher wartime exposure show:

• Higher per capita consumption in 2012 (an increase of roughly ₹1,693 per capita)

• Higher urbanization, with urban population shares about 8.3 percentage points higher

• Higher night-time luminosity, around 27 log points brighter in satellite imagery

• Higher in-migration rates and lower out-migration rates relative to other districts.

The night-lights result is especially striking because it captures, in a single visual sweep, the accumulated effects of a shock that occurred seventy years earlier. As the authors put it, these developmental differences are “visible from outer space,” an empirically grounded reminder that structural transformation leaves a physical, luminous trace.

The distinction between rural and urban areas also reinforces the idea that wartime mobilization ultimately altered the spatial distribution of opportunity: rural areas do not show large consumption differences, whereas urban areas do. Structural transformation, in other words, operates not just through sectors but through places, and the wartime shock appears to have nudged certain districts onto a more urbanized trajectory.

Rethinking the Indian State Through the Lens of War Mobilization

The findings complicate several entrenched narratives about Indian economic history. The colonial state is usually portrayed as laissez-faire in economic matters, oriented toward extraction rather than development, and indifferent to the industrialization of its colony except insofar as it served imperial needs. While this broad characterization remains correct, the wartime period reveals a different facet: under extreme geopolitical pressure, the colonial administration was capable of mobilizing resources, coordinating production, enforcing standards, and nurturing industrial capabilities at a scale that far exceeded its peacetime posture.

This does not rehabilitate colonial intent, but it does highlight a latent capacity for industrial coordination that independent India both inherited and reinterpreted. Indeed, it is conceivable that some post-Independence industrial investments, particularly those of the Nehruvian public sector, were placed in regions where wartime procurement had already expanded the industrial workforce or created nascent clusters, though this remains a question for future archival research.

The paper also speaks, perhaps unintentionally but powerfully, to current Indian debates on industrial policy, whether in the form of Production-Linked Incentives (PLI), strategic semiconductor fabrication, defense manufacturing, or green technologies. It demonstrates that large, coordinated, state-driven demand shocks can induce long-lasting structural change, but it also implies that such shocks must be technologically demanding and sufficiently large relative to the economy to generate lasting learning effects.

The wartime experience should not be romanticized. The procurement effort coexisted with famine, repression, and the deep inequities of colonial rule. Yet its economic consequences, stripped of normative judgment, reveal something about how quickly and durably production structures can shift when institutions coordinate, capacities are stretched, and firms are compelled to jump into higher levels of technological sophistication.

Every nation carries economic memories that are buried beneath the more dramatic political narratives of its past, and India is no exception. The wartime factories of Kanpur, Jamshedpur, Cawnpore, Coimbatore, and dozens of smaller towns rarely appear in our histories of India’s economic development, yet the evidence suggests that the shadow they cast is long and still discernible.

If one were to fly over India at night and gaze down at the patchwork of lights that mark the country’s vast and varied geography, some of the brighter clusters, those small concentrations of human activity, commerce, and aspiration, are, in a real sense, the afterglow of procurement contracts signed in Delhi during the darkest hours of the early 1940s. These lights are not the only legacy of war, nor are they the most morally salient, but they remind us that history leaves marks in unexpected places.

The deeper lesson of the paper is not that war is good for growth; it is that coordinated, capability-intensive mobilization, however motivated, can produce structural changes that outlive the institutions that initiated them. For a country still wrestling with how to industrialize, how to expand its manufacturing base, and how to translate technological ambition into broad-based prosperity, this insight is both sobering and quietly hopeful.

India once industrialized under coercion, speed, and imperial necessity. The challenge before us today is whether we can mobilize with the same clarity of purpose—but with democratic legitimacy, technological foresight, and a commitment to shared prosperity, so that the glow we see from space in the decades ahead is shaped not by the ghosts of compulsion but by the choices of a confident, self-directed nation.

Picture this: It’s 1970 in Seoul. A young Korean mother works long hours in a dusty factory, earning less than half what a family in Ghana makes. Her child dreams of a better life amid the rubble of war. Fast forward 50 years. That child lives in a world of high-speed trains and smart cities, living like someone in Spain or Italy. This isn’t a fairy tale. It’s real. It happened in Korea, Taiwan, Singapore, and Hong Kong—places we call the East Asian Miracles. These nations started poor. They ended rich. In one generation.

Economists Reda Cherif and Fuad Hasanov tell this story in their 2025 paper, Industrial Policy, Asian Miracle Style. They ask a simple question, which Robert Lucas asked in the late 80s and early 90s: Can a poor country like Ghana or Colombia grow like these tigers? Their answer is yes. But only with the right plan. Not luck. Not just hard work. A smart government that picks fights with poverty and wins. As you read on, you’ll see how regular people, bold leaders, and tough rules turned bold ambition into reality.

Think of the world in 1970. Most poor countries were low-income countries. Their people earned 10 to 20 percent of what Americans did. Over 50 years, little has changed. The figure below, from the paper, shows it clearly. Dots cluster near a straight line. Countries like Ghana stayed poor. Malaysia and Chile grew a bit into middle-income countries. But they stopped. Only 19 out of 183 nations hit high income, above half of the U.S. levels. Oil-rich spots like Saudi Arabia made it. Tiny islands with tourism did too. Europe’s edge countries, like Poland, joined the EU and caught up. But the Asian Miracles? They stood alone.

Korea began with 9 percent of U.S. income, poorer than Ghana. Taiwan matched Malaysia. Yet by 2019, they hit 90 percent of America. Hong Kong took 6 years to cross the high-income line. Singapore, 9. Korea, 25. Taiwan, 18. How? Not just factories or schools. Those help everywhere. But growth came from smarts—total factor productivity, or TFP. From 1970 to 2000, the Miracles grew 6.3 percent a year. Korea and Taiwan hit 7.2 percent. TFP added 2.1 percent. That means new ideas, better ways, jumping into tough industries. They went beyond their ‘comparative advantage’.

Robert Lucas saw it first. In Making a Miracle (1993), he said growth is like building Liberty ships in World War II. At first, each ship takes forever. Workers learn. Speed doubles. Costs halve. Then it slows. You must jump to new tasks, with steeper curves, bigger gains. Simple industries like tourism stall fast. Clean rooms, cooking food, same since Roman times, as Mary Beard notes in her books on ancient travelers. But electronics? Cars? Aerospace? Thousands of tasks. Always changing. High R&D. Endless learning and growth.

Markets don’t climb these alone. Too many roadblocks. Building factories needs clusters, suppliers nearby. Workers need exact skills. Who pays for roads or ports just for one industry? Poor firms can’t borrow big. No one knows if it’ll work. Paul Krugman (1987) called it coordination failure. A “big push” is needed, said Murphy, Shleifer, and Vishny (1989). States must lead. They beat what David Ricardo called “comparative advantage.” Don’t stick to rice or shirts. Chase big leagues—electronics, autos. That’s how you win.

In the aftermath of World War II, a wave of optimism swept across the newly independent nations of Asia, Africa, and Latin America, where leaders emerging from colonial rule envisioned a future of self-reliance and industrial might; they embraced a strategy known as import-substitution industrialization (ISI), which promised to transform their agrarian economies by producing everyday goods like steel, automobiles, machinery, and consumer products domestically rather than relying on expensive imports from wealthy Western countries. To make this vision reality, governments erected high tariff walls and import quotas to shield fledgling local industries from foreign competition, while simultaneously investing heavily in state-owned factories, offering generous subsidies, cheap credit, and tax breaks to entrepreneurs willing to set up shop at home, all in the hope of sparking a manufacturing boom that would create jobs, reduce trade deficits, and propel their countries toward modernity.

For a time, from roughly 1965 to 1980, this approach yielded tangible results, as evidenced by figure below from Cherif and Hasanov’s paper, which plots real manufacturing value-added per capita growth rates against initial income levels and reveals a striking convergence across developing economies: factories sprang up rapidly, output per worker surged at double-digit annual rates in many places, employment shifted from farms to assembly lines, and basic capabilities in heavy industry began to take root, with Asia’s emerging tigers like Korea, Taiwan, Singapore, and Hong Kong not only keeping pace but often leading the pack through early and aggressive implementation.

Then came the crash, a catastrophic debt explosion in the 1980s that shattered this fragile progress and plunged dozens of countries into their darkest economic phase. Leaders, desperate to sustain and expand their ISI projects, had been borrowing voraciously from Western commercial banks flush with petrodollars from the 1970s oil booms, using the funds to import the advanced machinery, specialized components, and technologies that local industries couldn’t yet produce; meanwhile, domestic firms, cocooned from competition, churned out low-quality goods, shoddy cars that broke down frequently, overpriced appliances with frequent defects, sold at inflated prices to captive consumers who had no better alternatives, forcing ordinary families to pay more for inferior products while governments racked up unsustainable debts to cover subsidies, budget shortfalls, and operational losses.

The tipping point arrived with a perfect storm of global shocks: as debts ballooned unchecked, rich countries like the United States and Western Europe slipped into deep recessions in the early 1980s, slashing demand for developing-world exports such as Brazilian coffee and soybeans, Chilean copper, Peruvian fishmeal, and even oil from Mexico and Nigeria; commodity prices, the lifeblood of these economies’ foreign exchange earnings, plummeted disastrously; oil, for example, tumbled from around $35 per barrel during the late-1970s OPEC spikes to as low as $10 by 1986, leaving governments without the cash to service their loans, many of which carried variable interest rates that spiked alongside U.S. Federal Reserve Chair Paul Volcker’s aggressive hikes to combat inflation.​

Governments went broke, defaults loomed, and the crisis erupted in August 1982 when Mexico declared it could no longer pay, triggering a domino effect across Latin America, Africa, and beyond, what became known as the “Lost Decade,” marked by slashed public spending, skyrocketing unemployment, malnutrition, and social unrest. Economist Anne Krueger crystallized the disillusionment in her 1990 work, famously declaring you can’t pick winners, arguing that governments lacked the foresight and discipline to select viable industries, leading to rampant waste and inefficiency; this critique resonated widely, prompting a global pivot away from state-led industrialization toward the “standard recipe” of free-market reforms, liberalizing trade, privatizing state assets, strengthening property rights, investing in education and infrastructure, and welcoming foreign direct investment, as championed by institutions like the World Bank and IMF under the banner of the Washington Consensus.​​

This new orthodoxy sounded appealing and was vigorously promoted by influential scholars: Daron Acemoglu and James Robinson in their 2012 bestseller Why Nations Fail contended that inclusive institutions such as fair rules, secure property rights, and limited corruption unleash entrepreneurship and long-term prosperity, while extractive ones stifle it, using historical cases to argue that institutional quality trumps geography or resources; similarly, Robert Barro and Jong-Wha Lee’s 2013 research, drawing on vast cross-country datasets, demonstrated that additional years of schooling dramatically boost human capital and growth rates, positioning education as a cornerstone alongside financial deepening and open trade.​

But look closer at the evidence, and table below from the paper tells a sobering truth that challenges the recipe’s sufficiency: in 1985, when Chile and Malaysia were poster children for these reforms, with superior governance scores (Chile at 3, Malaysia at 2 versus Korea’s 5, where lower numbers indicate less risk), stronger law-and-order ratings, higher private credit-to-GDP ratios (69% and 84% versus Korea’s 44%), and greater trade openness (52% and 106% versus 62%), their per capita incomes hovered around $6,000–$8,000 in PPP terms, comparable to Korea’s, yet Korea’s growth roared ahead at 8–11% annually while the others managed just 1–5%; by 2005, Chile and Malaysia had reached $14,000–$17,000, levels Korea had surpassed back in 1991, proving that even exemplary adherence to free-market prescriptions delivered solid but middling results, fixing the ills of poor governance and enabling convergence to middle-income status without igniting the sustained 7–8% miracle growth needed to reach high-income frontiers.​

Free markets excel at curbing bad rulers, reducing theft, enforcing contracts, and fostering basic stability, but they fall short on deeper “market traps” that plague development: banks shy away from funding unproven startups in cutting-edge sectors due to high risks and information gaps; innovative ideas remain siloed within firms, failing to spill over into broader ecosystems; and no single actor coordinates the massive “big push” required for complex industries like semiconductors, where agglomeration, specialized infrastructure, and skilled labor pools demand collective action, as Paul Krugman theorized in 1987, resulting in sluggish 3% average growth that peters out in middle-income traps rather than sustaining the transformative leaps of true miracles.​

Asia’s tigers, by contrast, mastered both worlds: they diligently built schools, roads, and stable macro frameworks while aiming with laser-like precision, blending the best of state intervention with market discipline in ways their peers did not. Ha-Joon Chang’s Kicking Away the Ladder (2002) exposes the hypocrisy, documenting how today’s rich nations once protected their own “infant industries” with fervor, 19th-century Britain banning Indian textiles and imposing Corn Laws, the United States maintaining tariffs as high as 50% until the 1940s, and Meiji Japan after 1868 launching state-sponsored steel mills, shipyards, and railroads before gradually opening up; Alice Amsden’s seminal Asia’s Next Giant (1989) chronicled Korea’s high-stakes bargaining with multinational firms, where subsidies came with ultimatums like “learn our technology through licensing and reverse-engineering, or lose your support,” fostering rapid capability-building; Robert Wade’s Governing the Market (1990) revealed Taiwan’s subtle “guided market” approach, with targeted credits, export-processing zones, and planning councils that steered private firms without stifling them; and Joe Studwell’s How Asia Works (2013) emphasized exports as the ultimate disciplinary “whip,” forcing firms to prove themselves in global arenas or face extinction. These scholars, drawing from decades of fieldwork, converge on a singular insight: Asia’s path was not mystical or culturally unique, but a deliberate, replicable blend of protection, ambition, and ruthless performance standards that set it decisively apart.

The Three Pillars of the East-Asian Miracle

At the core of Cherif and Hasanov’s analysis lies a distillation of the Asian Miracles’ enduring success into three interlocking principles:

1. Proactive state intervention,

2. Unrelenting export orientation, and

3. Ironclad accountability

These were not fleeting tactics but sustained strategies deployed over decades to transform resource-poor economies into global innovation powerhouses, setting them decisively apart from the failed experiments elsewhere.

The first principle demands that governments proactively identify and steer massive resources toward sophisticated, high-R&D industries far beyond a nation’s initial comparative advantage, with the explicit goal of empowering domestic firms to master and innovate in these sectors rather than merely assembling for foreigners; in the 1970s, for instance, Korea boldly targeted electronics, automobiles, and heavy machinery despite its “revealed” export strengths lying in low-skill wigs, plywood, and textiles, using the latter initially as a springboard for quick job creation and foreign exchange to fund the climb, as figure below dramatically illustrates with the explosion of U.S. patents granted per million people in Korea and Taiwan surging post-1980 while Malaysia’s line flatlines.​

This divergence stemmed from deliberate choices: by the mid-1990s, Korea poured over 2 percent of GDP into R&D, dwarfing Malaysia’s under 0.5 percent, beginning with licensing foreign technologies from firms like Philips or Motorola, then rapidly reverse-engineering and iterating to build proprietary capabilities, as exemplified by Samsung’s evolution from a trading company to a semiconductor titan that didn’t wait for multinational handouts but invested aggressively in homegrown design and fabrication. Japan had long restricted foreign direct investment to protect nascent industries, as Chalmers Johnson detailed in his 1982 classic MITI and the Japanese Miracle, and Korea followed suit initially, recognizing that multinationals share knowledge sparingly due to home-country biases and competitive frictions; thus, domestic innovation became non-negotiable, with scholars like Keun Lee (2019) demonstrating how this early leapfrogging to the technological frontier enables escape from middle-income traps, in stark contrast to Bangladesh’s entrapment in garment production or Korea’s own swift pivot away from textiles by the 1990s, ensuring total factor productivity remained dynamically high through perpetual task mastery.​

The second principle - exports, exports, exports - serves as the merciless truth that separates viable strategies from delusions, compelling firms to prove their mettle in the global arena rather than being complacent as in classic ISI regimes; while import substitution bred complacency with no incentive to scale quality or efficiency, export mandates forced Hyundai to construct massive shipyards and automobile plants capable of penetrating demanding U.S. markets, resulting in intricate supplier ecosystems in steel, engines, and electronics along the way, whereas Malaysia’s Proton car venture floundered as a subsidized domestic flop with negligible global sales, unable to attract buyers abroad.

The third principle enforces ruthless accountability through intensified domestic competition and market-based metrics like export shares, patent filings, and global market penetration, rather than mere output quotas that reward inefficiency; states nurtured clusters of rival conglomerates like Korea’s chaebols such as Samsung, Hyundai, and Daewoo, or Japan’s prewar zaibatsu, backing multiple contenders in each targeted sector to foster Schumpeterian creative destruction, where laggards withered and winners scaled, as theorized by Philippe Aghion and Peter Howitt (1992); President Park Chung-hee exemplified this by lavishly supporting frontrunners while slashing aid to underperformers, tolerating corruption as a byproduct but prioritizing tangible results over pristine governance, ensuring resources flowed to those delivering sustained productivity surges.​

Lead Agencies and the 4A Institutional Blueprint

Grand visions require operational muscle, which the East Asian Miracles supplied through specialized lead agencies - institutions like Japan’s Ministry of International Trade and Industry (MITI), Korea’s Economic Planning Board (EPB), and Taiwan’s Council for Economic Planning and Development - that functioned as elite, semi-autonomous command centers embedded deeply within industries, amassing unparalleled expertise to execute the three principles with precision and agility. Cherif and Hasanov articulate their success through a “4A” framework: Ambition, manifested in moonshot targeting of frontier sectors with top-tier staffing drawn from the best minds beyond traditional civil service ranks; Autonomy, granting these agencies independent budgets, hiring/firing powers, and freedom from bureaucratic red tape to act decisively; Accountability, balancing vertical reporting to national leaders with horizontal scrutiny via constant firm dialogues, export data audits, and performance benchmarks; and Adaptability, enabling rapid pivots, from labor-intensive textiles to capital-intensive chips, as global technologies evolved, all without the rigidity of Soviet-style five-year plans.​

A Roadmap for the Poorest Nations

For low-income countries looking enviously at these tigers, the lessons are neither elusive nor dependent on oil windfalls, geographic luck, or begging proximity to rich neighbors, but rather a clear, actionable blueprint beginning with foundational reforms like postwar land redistribution in Korea and Japan, which created efficient smallholder farms generating food surpluses and rural savings to seed initial manufacturing without food import drains. From there, harness cheap labor for export-oriented light industries, textiles, toys, simple assembly, to accumulate capital and foreign exchange for heavier bets, steadily ascending the ladder to steel, shipbuilding, and semiconductors, precisely as Joe Studwell outlines in his incisive analysis of how resource-scarce economies bootstrap via disciplined labor advantages before pivoting to complexity.​​

Cherif and Hasanov say forget the old idea of “revealed comparative advantage.” Don’t stick to farming rice or digging minerals. Pick high-tech industries instead. Focus on your own local companies, not foreign investors. Use exports as the big test. Build strong teams to guide everything. Messy governments can still win if they follow the rules.

Their studies show growth can be uneven. Small mistakes can ruin big plans. You need years of steady work. The World Bank in 1993 focused too much on factories and schools. The real key is smart policies.

Every monsoon, Gurugram drowns. Just one afternoon of rain can turn the city’s streets into ankle-deep rivers. The underpasses near Cyber Hub, supposedly designed to ease traffic, become waterlogged pits. Cars stall, commuters are stranded, and food delivery workers wade through muck. Social media overflows with pictures of cars stranded, people wading through water, and housing societies surrounded by floods.

And each year, a familiar question arises: how can India’s most prominent corporate hub, home to Fortune 500 offices, luxury apartments, and some of the country’s wealthiest households, fail at something as basic as drainage? The costs of this dysfunction are visible everywhere: residents stranded in waterlogged streets, office workers losing hours in gridlocked traffic, delivery personnel wading through filthy water, and even the municipal authorities facing repeated public criticism. The financial, social, and health toll is immense, and it affects everyone in the city, from the poorest to the richest. One would think that when so many people stand to lose, the shared interest would create enough collective pressure to force meaningful action, to finally fix the problem once and for all. Yet, year after year, the floods return, almost like clockwork, leaving a trail of frustration and helplessness.

The answer, surprisingly, lies in a book written over half a century ago by an American economist, Mancur Olson. His The Logic of Collective Action is one of those rare works that shifts how we understand politics, society, and human behavior. Olson wanted to understand a simple question: why do some groups manage to organize effectively around a shared goal, while others, sometimes much larger and more powerful, fail spectacularly? What he discovered turns much of conventional wisdom on its head. It is not a lack of knowledge, apathy, or incompetence that keeps societies from solving problems that are, on the surface, obvious. Rather, there is a structural logic embedded in how incentives operate, a logic that makes collective inaction more likely than action.

Olson’s insight is at once simple and unsettling. Just because people share a common interest does not mean they will act together to achieve it. Even when every individual stands to gain from some collective success, many will choose to do nothing, hoping others will shoulder the effort and cost. This is not laziness or moral failure—it is rational calculation. Once you see this, the annual paralysis of Gurgaon starts to make sense. The floods, the stranded commuters, the stalled construction projects—they are not anomalies or accidents. They are symptoms of a deeper, predictable pattern in human organization: the difficulty of turning shared interest into collective action.

At the heart of Olson’s theory lies the free rider problem. Public goods, he observed, are defined by two key qualities: they are non-rivalrous (one person’s use does not diminish another’s) and non-excludable (no one can be easily prevented from consuming them). Drainage systems, clean air, public parks, effective governance, and even national defense, all these, once provided, benefit everyone, whether they contribute or not. And therein lies the trap. Because no one can be excluded from enjoying the good, each individual has a rational incentive to withhold their own contribution and hope others will shoulder the burden. Why should one household in Gurgaon spend time, effort, or money lobbying the municipal council when the payoff, a fully functioning drainage system, will be shared by over a million other residents? For the individual, the logical choice often favors inaction, even when collective inaction is disastrous.

This dynamic becomes far more pronounced as groups grow larger. Olson showed that small groups often succeed where large groups fail because the costs and benefits of participation are more visible and immediate. Take, for instance, a cluster of street vendors in Bengaluru who rely on a shared water tank or storage facility. Each vendor can clearly see if another shirks responsibility, and the consequences of non-cooperation, such as loss of water access, fines, or social pressure, are felt immediately. The direct connection between contribution and reward makes cooperation easier to enforce, whether through social pressure, mutual reliance, or shared economic interest.

Olson shows that large groups usually don’t organize on their own. In big, spread-out groups, people often feel that the effort to participate is not worth it, so most stay passive and hope others will act. Smaller, close-knit groups can hold each member accountable, making everyone’s contribution count. This shows an important idea: the bigger the group, the harder it is for individuals to work together. Without ways to connect personal benefits to the group’s goals, even people with the same interests may fail to take action.

This inversion of expectation that small, cohesive groups often punch above their weight, while large, diffuse groups stumble despite their numbers, is one of Olson’s most striking insights. It explains a persistent feature of politics and governance: power often resides not with the majority, who bear the brunt of the problem, but with small, concentrated interests who are organized and motivated enough to push their agenda effectively. Olson’s framework illuminates why narrow groups often dominate policy decisions while the broader public struggles to make its voice heard.

But Olson did not leave the story in despair. If his argument ended with the inevitability of inaction among large groups, the picture would indeed be bleak: society would be condemned to perpetual inefficiency, with public goods underprovided and collective problems unsolved. Yet he asked a crucial follow-up: how do some large groups manage to organize at all, despite the overwhelming tendency toward free riding?

His answer lies in the concept of selective incentives. The idea is quite simple but powerful: groups that succeed often do so by combining the pursuit of collective goals with tangible, individual-level rewards or penalties. These incentives are selective because they are available only to those who actively participate or contribute. Trade unions do not only bargain for better wages (which benefit all workers, even non-members). They also provide strike pay, legal assistance, and social solidarity to dues-paying members, benefits that are excludable. Professional associations operate on the same principle. Membership in bodies like the Institute of Chartered Accountants of India, medical councils, or bar associations confers private advantages: certifications, access to exclusive networking opportunities, professional credibility, and even continuing education programs. While these associations may advocate for sector-wide reforms, say, new licensing standards, ethical guidelines, or better workplace safety, the decision to join and remain active is motivated as much by personal gain as by the desire to improve the system for all. Olson shows us that this mix of collective aspiration and selective reward is what allows large, otherwise diffuse groups to overcome the free rider problem.

The brilliance of Olson’s insight is that it reframes how we think about civic and political engagement. Large-scale collective action, whether in labor movements, professional networks, or even citizen advocacy groups, is rarely driven purely by abstract ideals or moral duty. Instead, it is the careful alignment of private incentives with public goals that allows mass coordination to happen. Without this alignment, even groups with enormous numbers and shared stakes can remain inert, paralyzed by the very rational self-interest that Olson argues. These insights shattered the comfortable pluralist notion that democracy naturally reflects the aggregation of citizen interests. Instead, Olson argued, politics is systematically biased toward organized minorities. Large majorities such as taxpayers, commuters, and consumers often remain underrepresented, not because they lack numbers but because they lack the organizational capacity to mobilize their diffuse interests.

The relevance for policymaking is profound. Why do so many reforms stall despite clear majority support? Why do public budgets get skewed toward subsidies, loopholes, or projects that benefit narrow constituencies? Why do long-term public goods like drainage, air quality, or bureaucratic efficiency receive less attention than short-term gains? Olson’s answer is consistent: the distribution of organizational power favors the few over the many. Air pollution in Delhi is the prime example. Every resident suffers from smog, yet lobbying for action is fragmented. Meanwhile, polluting industries or stubble-burning farmers, though smaller in number, act as concentrated blocs to resist costly reforms. The asymmetry of organization trumps the aggregate weight of interest.

While Olson paints a rather bleak picture of large-group inaction, his work is far from the final word on collective action. Elinor Ostrom, a Nobel-winning economist, offers a strikingly hopeful perspective. Studying communities around the world, like farmers managing irrigation systems, fisherfolk sharing common waters, she found that people can cooperate to manage shared resources sustainably, even without a central authority. Local communities often create their own rules, monitor compliance, and enforce penalties for rule-breaking. Trust, social norms, and regular communication make cooperation possible, and individuals see that their contributions matter because they can witness the collective benefits firsthand.

In essence, Ostrom shows that the free rider problem is not inevitable. While Olson highlighted the difficulties of large-group coordination, Ostrom demonstrated that with well-designed local rules, shared norms, and committed leadership, communities can overcome these obstacles. Technology today can amplify these mechanisms. The lesson is clear: human behavior is not fixed by self-interest alone. With the right structures, even diffuse groups can act collectively and achieve meaningful outcomes.

I will cover Ostrom’s work in a later dispatch.

For policymakers, Olson’s insight is not a reason to give up, but a guide for smarter design. Large groups rarely organize on their own, so institutions must be structured with this reality in mind. Independent regulators, citizen ombudsmen, or representative panels can serve as counterweights, ensuring that the interests of the larger, diffuse public are not ignored in favor of small, well-organized groups. Policies can also make use of selective incentives to encourage participation and cooperation: for example, offering rebates for adopting green energy, creating priority lanes for carpoolers, or publicly recognizing citizens who engage in local community projects.

Looking at Gurgaon’s flooded streets through Olson’s lens, the situation looks different. The city’s failure isn’t just about poor infrastructure; it’s also about how people organize. When large, spread-out groups of residents have a lot to lose, they often fail to act together. Meanwhile, small groups with concentrated benefits can influence the system to serve their own interests.

Olson’s provocation remains relevant six decades later: collective rationality does not emerge automatically from shared interests. It requires deliberate design, selective incentives, and sometimes the sheer stubbornness of leadership. Otherwise, the water will keep rising.

Humble Brag

Over the past six months, I’ve had the privilege of leading the capacity-building component under the NAKSHA programme — a national geospatial initiative spearheaded by the Department of Land Resources (DoLR), Ministry of Rural Development. As part of this effort, I’ve been supporting the design and implementation of the National Capacity Building Programme, which aims to equip officials from revenue and urban development departments and technical staff from 30 States/UTs with the necessary tools and training to integrate spatial thinking into land governance.

Working closely with the department’s leadership, I’ve helped steer this initiative from strategy to execution. Recently, the work under NAKSHA gained visibility when two parliamentary questions were raised in the Lok Sabha and Rajya Sabha regarding its capacity-building activities. I had the opportunity to assist the Ministry in drafting formal responses to both these questions — a small but meaningful contribution to a large and ambitious mission

Now, over to this week’s dispatch.

In January 2019, something extraordinary happened on the floodplains of the Ganga. A temporary megacity sprang up in Prayagraj to host the Kumbh Mela, a sacred Hindu gathering that attracted over 30 million pilgrims on a single day. Streets were spotless. Medical tents were staffed. Information centres buzzed with activity. Despite the logistical complexity, the event was heralded as a resounding success.

But just six years later, in 2025, tragedy struck. During another iteration of the Kumbh, a massive stampede claimed dozens of lives. Questions swirled about what went wrong. The same state machinery that had pulled off a logistical miracle had suddenly faltered. What changed?

This contrast lies at the heart of a powerful insight from a study conducted by UCL Associate Professor Prateek Raj and other co-authors. Drawing on their fieldwork and institutional analysis, the authors offer a compelling diagnosis of India’s governance paradox: the Indian state performs brilliantly in mission mode but struggles with routine service delivery. The explanation hinges on what the authors describe the Dynamic Governance Model, a framework that explains why missions succeed and why they often fail to last.

The Indian State’s Dual Personality

To understand the authors’ key argument, we must begin with a simple observation: the Indian state has an uncanny ability to shine during “missions.” Think of India’s COVID-19 vaccine rollout, or the smooth conduct of national elections, or the organization of the Kumbh. These are high-stakes, time-bound, and visible efforts. They receive political attention, media coverage, and public scrutiny.

In contrast, the same state appears indifferent or dysfunctional in its everyday duties. Health centers remain understaffed. Public schools struggle with teacher absenteeism. Municipal systems fail to provide clean water or manage waste. These failures are not due to ignorance or apathy alone. They are symptoms of a deeper structural problem: the state is not built for routine governance, the way it is mobilized for missions.

The Anatomy of a Mission: Lessons from the 2019 Kumbh

In their study of the 2019 Kumbh Mela, the researchers observed that success did not come from rigid bureaucratic structures or a top-down command system. Rather, it came from what he calls a meta-organizational system, a loosely coordinated but purpose-aligned network of actors, ranging from government departments and consultants to volunteers and religious leaders. This system operated with remarkable cohesion, despite lacking a formal hierarchy.

What made this possible? The study identifies several key enablers.

First, there was role ambiguity. In most bureaucracies, people stick to their job descriptions. But during the Kumbh, engineers stepped in to manage logistics, consultants took on operational roles, and sanitation workers went beyond their assigned duties. This fluidity allowed the system to adapt in real time. People were driven not just by compliance, but by a shared sense of mission.

Second, the system was deliberately built in slack. There were spare buses, standby staff, and buffer budgets. While traditional administrative logic treats slack as waste, Raj shows that in dynamic systems, slack is essential. It provides room to absorb shocks, handle surges, and prevent breakdowns.

Third, there was an information infrastructure that supported rapid feedback. From loudspeakers and walkie-talkies to WhatsApp groups and drone surveillance, the system was in constant communication. This allowed for quick decisions, course corrections, and decentralized responsiveness.

Together, these features enabled what the authors call dynamic governance, a model where governance adapts to context, switches modes as needed, and relies on trust, improvisation, and coordination rather than strict rules.

The Collapse of the Model: Kumbh 2025

And yet, the same model that brought success in 2019 unraveled in 2025. A tragic stampede exposed just how fragile this dynamic system can be.

In 2025, many of the enablers were missing or weakened. Political incentives had shifted. Turf wars between departments re-emerged. Slack was reduced in the name of efficiency. The hybrid information infrastructure, once a backbone, became distorted, focused more on managing VIP visits than crowd safety.

Most crucially, the culture of collaboration had eroded. The ethos of improvisation and shared mission had given way to risk aversion and blame avoidance. Role ambiguity, once a source of strength, became a liability. No one took ownership. Everyone deferred to someone else.

The study’s conclusion is sobering: dynamic governance is not self-sustaining. It requires deliberate cultivation. Without the invisible scaffolding of trust, shared purpose, and coordination, the system collapses under stress.

Routine Governance and the Limits of Missions

These insights don’t just apply to mega-events like the Kumbh. They reveal something more fundamental about how the Indian state works, and why it often doesn’t.

In mission mode, the state activates a different gear. It becomes focused, resourceful, and adaptive. But in routine mode, governance reverts to its default: fragmented, hierarchical, under-resourced, and inflexible.

This is not a uniquely Indian problem. Many developing states suffer from what Lant Pritchett calls capability traps, where institutions look functional on paper but are hollow in practice. They chase ambitious goals without building the systems to deliver them. As a result, they “fail by trying too hard too soon.”

But India’s dual personality, its bursts of brilliance amid chronic dysfunction, is especially stark. This study shows that India doesn’t lack capacity. It lacks continuity.

Can We Institutionalize Dynamic Governance?

So, what can be done? Can the magic of mission-mode be extended into the everyday?

The study doesn’t offer silver bullets, but it points to a few possibilities:

• Invest in institutional memory. Lessons from successful missions must be codified, shared, and embedded in routine systems.

• Encourage flexible staffing and decentralized problem-solving. Routine governance must allow local officials the same latitude for improvisation that missions permit.

• Build trust-based networks across departments. Collaboration cannot be episodic; it must be part of the culture.

• Redesign performance metrics. Success must be defined not by compliance and paperwork, but by responsiveness and outcomes.

In essence, we must move from a static state to an adaptive one, capable of switching gears, sustaining energy, and responding to uncertainty.

The story of the Kumbh is not just about a festival. It’s a mirror reflecting the capabilities and contradictions of the Indian state. On its brightest days, India pulls off feats that rival any modern bureaucracy in scale, ambition, and execution. But those moments are fleeting. The deeper challenge lies in building a state that doesn't just rise for the extraordinary but quietly delivers on the ordinary.

The state that shines during a mega-event must also shine when no one is watching. It must ensure that the same dedication that brings a city to life for 30 million pilgrims is applied to keeping a village school functional, a health center stocked, or a garbage truck running on time. These may not make headlines, but they define the everyday experience of governance.

If missions are fireworks, brief, bright, and spectacular, routine governance must be like streetlights: consistent, invisible, and vital. The brilliance of a mission must not be the exception but the spark that builds institutions strong enough to outlive the campaign.

The dynamic governance model offers not just a critique of our present but a blueprint for our future. They urge us to ask: what would it take to design a state that learns, adapts, and endures, not only when the world is watching, but especially when it isn’t?

The real mission, then, is not another flagship scheme. It is to make governance so capable, so embedded, and so responsive that even the routine becomes remarkable.

In governance and public policy, incentives have long been regarded as central to improving outcomes. The logic is simple: structuring the right incentives should produce better performance if individuals respond to incentives in predictable ways. This belief has driven decades of reform, from pay-for-performance schemes in education and healthcare to results-based financing in public administration.

However, real-world examples often challenge this logic. For instance, consider a ministry that introduces an incentive program to conduct training sessions on new technology for its staff. The intended outcome is to build digital capacity so employees can use new systems effectively. But if the incentive for the ministry is tied only to the number of training sessions conducted, rather than the actual uptake or application of the technology, then the focus quickly shifts. Workshops may be hastily organized, attendance recorded, and reports filed to meet targets, while participants remain poorly equipped to use the tools. Here, the incentive structure pushes administrators to prioritize quantifiable outputs over meaningful learning, ultimately undermining the very outcome it aimed to support.

This is a classic example of what Public Choice Theory describes as goal displacement, where bureaucracies pursue what is easy to measure, rather than what truly matters. The ministry's incentive is to show activity (number of trainings), not effectiveness (actual tech adoption). Ironically, the system rewards a proxy for success while leaving the intended outcome, such as staff being able to use new technology, unmet. In this way, an incentive designed to signal progress may end up institutionalizing underperformance.

Many of these incentive-driven reforms, while well-intentioned, have failed to produce the desired outcomes. Worse, they have sometimes led to unintended consequences: perverse behavior, short-termism, erosion of intrinsic motivation, and bureaucratic inertia. This raises a fundamental question: Can the very incentives meant to improve governance undermine it?

Public Choice Theory offers a compelling lens to understand this puzzle. This post explores the relationship between incentives and governance outcomes through that lens, arguing that incentives can, in certain contexts, distort public purpose and kill outcomes. However, it also explores how better-designed incentives, rooted in a deeper understanding of institutional behavior, can still play a constructive role in policymaking.

Public Choice Theory and the Incentive Dilemma

Public Choice Theory, pioneered by economists such as James Buchanan and Gordon Tullock, applies the logic of economics to political and bureaucratic behavior. Its core assumption is simple but powerful: actors within public institutions, such as politicians, bureaucrats, voters, and interest groups, act in self-interested ways, responding to incentives much like actors in markets do.

This perspective challenges the traditional, idealistic view of the public sector as a neutral guardian of the common good. Instead, it sees public institutions as arenas where individuals pursue their own goals, often at odds with collective welfare. Gordon Tullock pointed out that bureaucrats are incentivized to grow their departments—not necessarily to deliver outcomes—because increased budgets often signal performance. Similarly, William Niskanen’s Budget-Maximizing Bureaucrat model suggests that bureaucrats prefer policies that increase their discretion and budget, rather than those that maximize social welfare.

Under this framework, incentives are not inherently good or bad. Rather, they are instruments that shape behavior, and their outcomes depend on how well they align individual motivations with public goals. The dilemma arises when incentives are poorly designed, too narrow, or too detached from the complexities of policy environments. This can lead to several adverse effects:

1. Goal displacement: Individuals begin to focus on the measurable indicator rather than the intended outcome (e.g., number of schools built vs. learning outcomes).

2. Gaming and manipulation: Agents exploit loopholes or distort data to meet targets without achieving real progress.

3. Crowding out intrinsic motivation: When professionals are overly monitored or financially incentivized, their internal commitment to public service may weaken.

4. Short-termism: Incentives tied to electoral or budget cycles can prioritize visible, quick wins over sustainable, long-term reforms.

5. Policy capture: Incentives can be designed or co-opted by interest groups to serve their own ends, especially in low-accountability environments.

Thus, Public Choice Theory helps us see incentives not as neutral tools but as deeply political and context-sensitive mechanisms. Poorly designed incentives don’t just fail to deliver outcomes; they can actively subvert them.

Can Incentives Be Redesigned to Support Outcomes?

While Public Choice Theory offers a cautionary tale about the misuse of incentives, it does not call for their abandonment. Instead, it offers principles for designing better incentives—ones that are more aligned with institutional realities and human motivations.

Align incentives with multi-dimensional goals
One of the most common pitfalls in incentive design is the over-reliance on narrow, quantitative metrics, such as test scores, number of beneficiaries reached, or units delivered, as proxies for success. While these metrics are easy to track and compare, they often fail to capture the full complexity of public service delivery. Worse, they can distort behavior by encouraging teaching to the test, data manipulation, or box-ticking exercises that prioritize quantity over quality. To address this, policy designers should move toward multi-dimensional performance frameworks that combine a broader set of indicators. These might include:

• Outcome indicators that track real-world impact (e.g., improvements in learning, reductions in disease burden, or better infrastructure use);

• Process indicators that assess the integrity, inclusiveness, and transparency of implementation (e.g., how decisions were made, who was consulted, whether procedures were followed);

• Qualitative indicators, such as citizen feedback, staff satisfaction, or expert review, provide a richer understanding of whether the service is meaningful and valued by those who receive it.

• Equity indicators that reveal how fairly the benefits of a program are distributed across regions, genders, or social groups.

For example, in the education sector, measuring only test scores might reward schools that teach to the test while neglecting creativity, critical thinking, or inclusivity. A more balanced approach might combine academic results with teacher feedback, classroom observations, student well-being, and parental satisfaction.

Incentives aligned with this multi-dimensional view of performance are less prone to gaming, better reflect the real goals of public service, and support a more holistic and sustainable approach to governance. They also signal trust in the professionalism of public servants and encourage a broader commitment to public value, not just target compliance.

Protect and leverage intrinsic motivation: Public servants are not merely utility-maximizing individuals responding to external rewards; they are often motivated by a deep sense of duty, professional identity, ethical responsibility, and a commitment to public service. These intrinsic motivations are central to why many choose and remain in public sector roles, despite lower financial rewards compared to the private sector. When incentive structures ignore these internal drivers and rely solely on extrinsic rewards like bonuses or rigid performance targets, they risk eroding the very ethos that sustains high-quality public work.

Instead, policies should aim to nurture and sustain intrinsic motivation by recognizing what matters to public servants. This includes systems of meaningful recognition, such as peer acknowledgment, public appreciation, or awards for integrity and innovation. It also involves investing in professional development, enabling employees to grow their skills, expand their expertise, and stay connected to their broader mission. Most importantly, granting autonomy and discretion, allowing public servants to exercise judgment, solve problems creatively, and take ownership of their work, can significantly enhance motivation and job satisfaction.

Well-designed public institutions understand that motivation is not merely a transaction, but a relationship built on trust, respect, and shared purpose. A healthy balance between intrinsic and extrinsic incentives is essential for sustaining morale, preventing burnout, and achieving long-term governance outcomes.

Design for adaptability, not rigidity
Incentive structures must recognize that governance is not a static exercise—public challenges evolve, contexts shift, and what works today may not work tomorrow. Yet, many incentive systems are designed as fixed, one-size-fits-all contracts that lock institutions into rigid performance frameworks. These static models often become outdated, misaligned with field realities, and resistant to course correction. When that happens, public servants are left chasing outdated targets or metrics that no longer reflect the underlying policy goals.

To avoid this, incentives should be designed with built-in adaptability. This means embedding them within feedback loops that regularly collect information on what’s working, what’s not, and why. These loops can draw from performance data, frontline feedback, citizen input, and independent audits to guide periodic review and refinement.

For example, in a health insurance scheme, early data might show that incentives for hospital admissions have increased unnecessary procedures. A rigid incentive model would continue to reward this behavior, while an adaptive one would trigger a re-evaluation, shifting incentives toward preventive care or outcome-based payment models.

Adaptability also means allowing room for contextual variation. Different regions, departments, or service environments may require different incentive mixes. A remote rural district might need flexibility on timelines and targets that wouldn’t apply in an urban setting.

By designing incentives as living instruments rather than fixed levers, policymakers can ensure that systems stay responsive to change, learn from experience, and remain focused on outcomes rather than procedures. Adaptable incentives foster resilience, institutional learning, and better alignment between intent and implementation over time.

Account for the political economy and institutional context
Incentives do not operate in a vacuum. They are embedded within the political economy and institutional architecture of a society, shaped by power relations, bureaucratic norms, informal networks, and vested interests. A common mistake in policy design is assuming that technically sound incentive structures will work universally, regardless of the context in which they are deployed. In reality, even the most rational incentive system can fail if it clashes with the underlying political and institutional realities.

For instance, a reform that rewards civil servants for transparency may be undermined in an environment where powerful actors benefit from opacity and rent-seeking. Similarly, performance-based pay for teachers might fail in systems where hiring and promotions are driven by patronage rather than merit. In such contexts, frontline actors may lack the autonomy, protection, or resources needed to respond to the new incentives, regardless of how well they are designed on paper.

To be effective, incentive design must be politically aware and institutionally grounded. This involves:

• Anticipating resistance from actors whose interests may be threatened;

• Mapping stakeholder interests, including bureaucrats, unions, politicians, private contractors, and beneficiaries;

• Building coalitions of reform champions who can support, legitimize, and sustain the reform;

• Sequencing reforms to build trust and demonstrate early wins before introducing more disruptive changes.

Crucially, this also means aligning incentives with existing institutional capacities. In fragile bureaucracies with low administrative bandwidth, simple, low-stakes incentives may work better than complex, high-stakes systems that cannot be enforced.

Incentives that ignore these realities may look elegant on paper but collapse in practice. Conversely, designs that are attuned to power dynamics, institutional norms, and implementation constraints are more likely to survive political headwinds and deliver lasting outcomes. The goal is not just to design technically sound policies, but politically viable and contextually credible ones.

Embed transparency and accountability
For incentive systems to be effective and trustworthy, they must be anchored in mechanisms that promote transparency, oversight, and corrective action. Without these guardrails, even well-intentioned incentives can be gamed, captured by vested interests, or manipulated to serve private rather than public goals. In low-accountability environments, performance metrics may be inflated, data may be falsified, and outcomes may be misrepresented, all in the name of meeting targets.

To prevent this, transparency must be designed into the incentive architecture itself. This includes:

• Public reporting of performance data and outcomes in formats that are accessible and understandable to citizens, media, and civil society;

• Independent audits and third-party evaluations that verify the accuracy of reported results and assess whether incentives are achieving their intended effects;

• Grievance redressal mechanisms that allow citizens and frontline workers to flag problems, lodge complaints, and challenge distortions;

• Feedback loops that incorporate citizen perspectives, frontline realities, and external evaluations into policy adjustments.

For example, publishing real-time dashboards on school attendance or health facility performance can create pressure for accountability, especially when combined with social audits or community scorecards. Similarly, using independent agencies to validate results tied to performance-based funding can reduce the risk of internal manipulation.

Importantly, accountability is not just about punishment; it is about learning. A transparent system allows for honest reflection on what’s working, what isn’t, and why. It creates the space for adaptive management, mid-course corrections, and shared responsibility between policymakers and implementers.

By embedding transparency and accountability into incentive systems, governments not only deter abuse, they also build public trust, reinforce institutional legitimacy, and strengthen the credibility of reform efforts. In short, transparency transforms incentives from levers of control into tools for collective learning and democratic oversight.

Embrace incentives as nudges, not levers
Not all incentives need to be heavy-handed or tied to financial rewards and penalties. In many cases, small, well-designed nudges, subtle changes in how choices are presented or information is framed, can gently guide behavior in desired directions without resorting to coercive or costly measures. Drawing on the insights of behavioral economics, these nudges recognize that people do not always make decisions based on pure rational calculation. Instead, they are influenced by cognitive biases, social norms, habits, and how options are structured.

For example, changing the default setting on government forms to automatically enroll employees in pension schemes, while giving them the option to opt out, has dramatically increased savings rates in many countries. Similarly, framing information about energy usage in terms of how one’s consumption compares to neighbors can motivate households to reduce waste. In public health, simply reminding parents via text messages about immunization schedules can significantly increase compliance, without any material incentive involved.

These small design tweaks may seem minor, but they often produce large behavioral shifts. Unlike rigid performance contracts or punitive systems, nudges are less intrusive, more cost-effective, and easier to adapt across different contexts. They work with the grain of human behavior, not against it.

Of course, nudges are not a silver bullet. They must be used ethically, transparently, and in tandem with broader institutional reforms. But as a complement to traditional incentives, they offer a powerful tool for public policy, especially in settings where trust is low, administrative capacity is limited, or cultural factors resist top-down mandates.

In conclusion, the debate over incentives in governance and policymaking is not a binary one. Incentives can motivate, distort, enable, or derail, depending on how they are designed and embedded within institutions. Public Choice Theory offers a vital corrective to simplistic views of governance by showing how self-interest, institutional rules, and incentive structures interact in complex ways.

Incentives that are poorly designed, overly rigid, or politically blind can indeed kill outcomes. But incentives that are thoughtfully crafted, context-aware, and responsive to real-world behaviors can help align individual motivation with public purpose. The task, then, is not to abandon incentives, but to treat them as instruments of design, tools that must be wielded with judgment, humility, and a deep understanding of how governance works.

In this video, Buchanan and Tullock discuss their book The Calculus of Consent to mark its 25th anniversary in 1987.

A few years ago, during a policy consultation with a state government, we were discussing an ambitious reform proposal. As the conversation gained momentum, a senior policy researcher remarked, almost dismissively, that the state should abandon initiatives for which it lacks the necessary capacity.

The statement wasn’t surprising. Many of us had heard similar sentiments before, and there are numerous examples in the Indian policy-making and implementation ecosystem that seem to justify such a proclamation. But it stuck with me. What exactly did he mean by “state capacity”? Capacity to plan? To implement? To coordinate? To persuade? To manage the expectations of the stakeholders?

That moment set me off on a journey to understand the idea of state capacity more rigorously. I began reading everything I could, from historical institutionalist theories to contemporary studies in public administration. I discovered that while the term was everywhere in governance literature, it was rarely defined with precision. More importantly, most discussions treated capacity as a static trait, something a state either had or lacked, rather than as something dynamic, distributed, and context-dependent.

That’s why I found Azad Singh Bali and M. Ramesh’s paper Policy Capacity: A Policy Design Perspective so compelling. It doesn’t just dissect the idea of capacity, it reframes it. Rather than asking whether a state is strong or weak, the paper asks a sharper question: what kind of capacity is needed to design and deliver effective policy, and where should it reside?

In this post, I walk through the key ideas in the paper and reflect on how this framework offers a more practical, design-oriented approach to understanding what governments need to solve problems.

From State Capacity to Policy Capacity: Reframing the Debate

Let’s begin with the conceptual shift this paper makes.

Scholars have defined state capacity in many different ways, depending on what aspect of governance they focus on. For instance, Rothstein (2011) argues that a capable state upholds the rule of law and treats citizens impartially. Fairness, in his view, is central to capacity. On the other hand, Weiss (1998) and Hobson (1997) emphasize the importance of a strong, autonomous bureaucracy that can plan effectively and act independently of narrow political pressures. Other definitions focus more on legitimacy and coordination. Painter and Pierre (2005), for example, describe state capacity as the ability of governments to make “intelligent collective choices”, that is, decisions that are coherent and supported across different parts of the system.

Each of these perspectives sheds light on an important piece of the puzzle. But as Bali and Ramesh point out, these definitions often treat capacity as a single, overall trait as if a government either has it or doesn’t.

What they propose instead is to shift the focus from broad institutional traits to the specific work of policy design. Rather than ask “Is this a capable state?”, they ask: Can this government design and deliver good policy?

This is what they call policy capacity - the ability of a government (at various levels) to make informed, context-sensitive, and problem-oriented decisions about how policies are crafted and implemented.

This more focused lens helps in two important ways:

1. It moves the conversation from abstract discussions about “strong states” to the concrete and practical challenge of solving real-world problems.

2. It allows us to break down capacity into smaller, more manageable parts like the ability to analyze problems, coordinate between actors, manage budgets, or build political support and to locate these capacities at the level of individuals, departments, or systems.

In short, while traditional views of state capacity highlight important features like rules, institutions, or legitimacy, Bali and Ramesh make the case that we should pay closer attention to how governments design policy because that’s where capacity is most visible, and most essential.

Capacity for What? Start with the Problem

Too often, capacity-building efforts begin with a shopping list: train more staff, install better IT systems, write more manuals, create new departments. But these efforts are rarely anchored in the specific problems a policy is meant to solve.

That’s why this paper insists on flipping the question: instead of asking “what capacity do we need?”, ask “what problem are we trying to solve and what kind of capacity does that require?”

This is where policy design enters the picture. Bali and Ramesh draw on policy design literature to show that crafting effective policy requires attention to a specific set of design attributes. These attributes act like quality checks. If they’re not met, the policy is likely to underperform, no matter how “strong” the state is in general.

They explain:

On one hand, there is a need for increased capabilities for governments to maneuver a rapidly changing policy terrain. The rise of plurilateral, collaborative, and ‘meta governance’ styles is fundamentally changing the extent to which, and how, governments interact with stakeholders in the provision of public goods and services. The contemporary policy environment is characterized by increased economic and political uncertainty; continual technology disruptions that fundamentally change how a problem is conceptualized (let alone addressed); and the growing complexity of modern problems underscore the importance of creating agile, informed, and resilient designs. Adding to this are growing expectations of citizens in increasingly contestable societies that their governments will play a larger role in the provision and financing of positional goods, and help manage risks that were previously addressed largely by the market. However, on the other hand, many country-specific studies have documented, and many governments have acknowledged, the gradual decline in capacity over the years. As early as the mid-1990s, scholars attributed this decline to the gradual ‘hollowing out’ of government, fueled by a series of public management reforms which emphasized fragmentation, contracting out, and the unfettered pursuit of economic efficiency in the delivery of public services. This seeming contradiction, of the need for greater policy capacity and the gradual decline in skills and competencies, can be addressed by applying the design orientation or problem-solving approach central to the policy sciences to policy capacity. As opposed to platitudes on the need for capacity building, it would require responding to ‘what skills and competencies are required to be able to address, or at least mitigate, the given problem at hand’. This not only requires being able to conceptualize the problem in its constituent elements, generate plausible solutions, ensure that they are politically viable, and can be implemented, but will require designers to be cognizant of the design attributes that increase the effectiveness of a given policy or program.

Here are the nine attributes they identify, each of which corresponds to common real-world challenges:

Coordination

Emphasizes the need to see the bigger picture—to ensure that new policies align with and complement existing ones. Without this, policies often overlap or conflict. For example, in Ghana, residents once held up to seven different national ID cards, creating inefficiency and confusion. In India, the RSBY health insurance scheme was run by the Ministry of Labour without proper coordination with the Ministry of Health, leading to duplication and gaps. Good coordination means policy changes should be made with awareness of how they affect other programs and actors, avoiding unintended disruptions.

Coherence

This attribute is about making sure that different policies work well together and support a common goal. It means that government actions across various departments should be aligned, not pulling in different directions. For example, if multiple programs are aimed at the same group, like farmers or low-income families. They should be based on similar ideas and not confuse or contradict each other.

Over time, small changes to policies—like adding new rules or tweaking old ones can unintentionally weaken this alignment. These changes might shift the focus, create mixed messages, or change incentives in ways that reduce the overall effectiveness.

Just like coordination, maintaining coherence requires looking at the bigger picture: Are the policies targeting the same beneficiaries in consistent ways? Do they reflect the same goals and principles? And are both the substance of the policies and the processes used to implement them working in harmony?

Goodness of Fit

This attribute means that a policy should be well-suited to the specific context in which it is being implemented. This includes the nature of the sector, like health, education, or energy, as well as how decisions are usually made and what the political environment looks like.

In simple terms, a policy design that works well in one place might not work the same way elsewhere. For example, a highly centralized program might work in a country with strong top-down governance, but not in one where power is more decentralized.

So, a good policy needs to “fit” its surroundings; it should match the way things are governed, the capacity of institutions involved, and the political realities on the ground. If it doesn’t, even a well-intentioned policy can fail.

Consistency

This attribute means that all the tools and actions used in a policy should support the same goal and not work against each other. Problems arise when some parts of a policy give positive signals (incentives) while others send mixed or negative signals (disincentives), making it harder to achieve the intended outcome.

Think of it like trying to drive a car while pressing the accelerator and the brake at the same time; progress becomes difficult or confused.

Scholars like Howlett and Rayner explain that coherence is about aligning the goals of a policy, while consistency is about aligning the methods or tools used to reach those goals. When policy tools are inconsistent, it can cause policies to slowly shift or “drift” away from their original purpose, or get layered with new, sometimes conflicting elements. This weakens the effectiveness of the policy over time. Research also shows that sudden and major changes in how a policy is carried out can negatively affect results.

Degrees of Freedom

The Degrees of Freedom attribute refers to how much flexibility policy designers have to make changes in the future. In reality, most policies aren’t created from scratch; they build on past decisions. Over time, adjustments like adding new rules or modifying old ones can limit future options. That’s why it’s important to design policies in a way that leaves room for future updates, allowing both small tweaks and bigger shifts as situations change.

Complementarity

This attribute is about making sure that different parts of a policy work well together and strengthen each other. It’s not enough for policies to simply avoid conflict—they should create positive effects when combined.

For example, one tool might improve service delivery, while another boosts accountability. If used together in the right way, they can reinforce each other and make the policy more effective.

So, when designing policies, it’s important not only to ensure coordination and coherence but also to actively look for synergies—ways in which different tools can support and enhance one another. Ignoring this can lead to wasted resources or unintended negative effects.

Targeting

The targeting attribute is about how well a policy reaches the people it's meant to help, without including those who shouldn’t receive the benefits.

There are two common problems:

• Exclusion errors: when people who should get the benefits are left out.

• Inclusion errors: when people who aren’t eligible receive the benefits.

From a policy design point of view, the goal isn’t just to avoid one type of error, but to reduce both as much as possible. Good design means thinking carefully about how these mistakes can happen and putting systems in place to minimize them, like better data, clear eligibility rules, and strong delivery mechanisms.

Reversibility

The reversibility attribute is about how easily a policy or its parts can be changed or rolled back in the future. While similar to flexibility (or degrees of freedom), reversibility focuses on how big the change is and whether it’s politically possible.

In large countries like India or Brazil, reversing a policy that affects millions, like subsidies, can be very difficult, both financially and politically. Once people start benefiting, it creates pressure to keep the program going, even if it’s no longer effective.

That’s why good policy design should plan ahead—for example, by including a sunset clause (an expiry date) or setting clear limits on cost or scale to make future changes easier.

Contingency

The contingency attribute is about planning for future risks and unexpected costs that may arise from a policy, known as contingent liabilities. These are costs the government might have to cover later, like pension shortfalls, healthcare demands, or even bank bailouts.

The challenge is that many governments underestimate these risks, and once the policy is in place, it becomes hard to add safeguards later because beneficiaries may resist changes.

Good policy design doesn't have to eliminate all future costs, but it should include ways to manage them from the start, like setting aside reserves, adding spending limits, or building in flexible rules, so the government isn’t caught off guard.

Mapping Capacity: A Multi-Level Framework

Now that we know what good design looks like, the question becomes: who needs to be capable of doing what, and at what level?

This is where one of the paper’s most useful contributions comes in: a three-by-three matrix that organizes policy capacity by type (analytical, operational, political) and level (individual, organizational, systemic).

Let’s walk through it:

This framework helps explain why policy failure is often not just about bad ideas or poor execution, but mismatched or missing capacities. For example:

• A good policy can fail if there’s not enough political buy-in (low political capacity).

• A creative policy design might be unworkable if procurement systems are broken (weak operational capacity).

• A well-funded reform might collapse if bureaucrats don’t know how to adapt it to local realities (poor analytical capacity).

The real insight here is that capacities must match the demands of policy design. If a policy requires coordination across sectors, but the system lacks inter-agency collaboration mechanisms, the design will break down. If a reform depends on citizen trust, but the state lacks legitimacy, the best-designed policy will still stumble.

A Case in Point: Health Policy Design

One of the biggest challenges in health policy, especially in developing countries, is how to ensure that people have access to quality healthcare without falling into financial hardship. In much of Asia, healthcare costs are still largely paid out of pocket by households. To tackle this, many governments have introduced reforms aimed at achieving universal health coverage (UHC). These reforms typically involve two key strategies: increasing public financing (such as through government-funded insurance) and expanding public provision (often by contracting private healthcare providers).

While these reforms are well-intentioned, evidence shows that they often fall short. Why? Simply pouring more money into the system or adding new delivery mechanisms does not automatically solve the deeper, structural problems that exist within the healthcare sector. These include things like:

• Information asymmetry (where patients don’t have enough knowledge to make informed choices),

• Moral hazard (where insured individuals or providers overuse services),

• Weak accountability, and

• Fragmented delivery systems.

Bali and Ramesh argue that unless these underlying governance issues are addressed through careful design, health reforms will not lead to meaningful improvements in access, quality, or outcomes. They emphasize the importance of designing appropriate governance arrangements that manage the complex web of relationships between key actors—patients (users), healthcare providers, insurers or payers, and the government (the state).

To guide this process, the authors apply their policy design framework to the health sector. Using the table below, they map out the different tasks involved in designing health policies and how these link to specific types of capacity.

1. Analytical Design Tasks: Understanding the Problem and Finding the Right Tools

Designing effective health policy starts with understanding what exactly is going wrong. This involves:

• Breaking the problem into parts, such as identifying the specific causes of high costs or poor access,

• Diagnosing issues like perverse incentives among providers (e.g., overprescribing, unnecessary tests),

• Generating evidence to identify what types of policy tools (health financing instruments like DRGs, i.e., Diagnosis-Related Groups or capitation) could address these problems, and

• Designing tools that encourage providers and payers to behave in ways that improve efficiency and fairness.

This level of design work requires deep analytical capacity, knowledge of economics, public health, and policy tools, as well as the ability to interpret and localize best practices for the national context.

2. Operational Design Tasks: Making It Work on the Ground

Even the best-designed policy ideas won’t succeed without solid planning and execution. This part of the design process involves:

• Ensure that the prerequisites for each tool are in place. For example, if the policy depends on subsidizing premiums for low-income families, accurate income data must be available.

• Coordinating between different government departments, like Health and Finance, to make sure that the policy is coherent and fits with existing programs.

• Establishing accountability systems, such as contracts that define standards for private providers, or penalties for non-compliance.

Operational capacity is often where health reforms falter, particularly when public systems lack procurement skills, reliable data systems, or inter-agency cooperation.

3. Political Design Tasks: Managing Stakeholders and Building Support

Health reforms are politically sensitive. They affect millions of people and often challenge entrenched interests, such as private hospitals, pharmaceutical companies, or existing bureaucracies. Therefore, a key part of design is political:

• Identifying potential resistance from key stakeholders (e.g., providers who might lose income),

• Negotiating trade-offs and building coalitions to support the reform,

• Communicating the purpose and benefits of the policy to the public and key actors to generate buy-in,

• Navigating the reform through the political system, including legislative or cabinet approval.

This requires political capacity, the ability to read the landscape, engage with interest groups, and adjust the design in politically realistic ways.

4. Skills and Competencies: Who Needs What Capacity?

In the table below, the authors further break down what types of skills and capacities are needed at different levels of the system, individual, agency, and system-wide.

• At the individual level, health policy designers need technical training in areas like health economics and insurance design, along with a practical understanding of tools like DRGs or capitation.

• At the agency level, departments need access to expertise in managing large-scale programs, inter-departmental collaboration, procurement, contracting, and monitoring.

At the system level, the broader environment must support high-quality data, trust in public institutions, and a general level of legitimacy so that regulations are followed and reforms are accepted.

What this case study shows is that no amount of funding or political will alone can fix health systems. Without the right capacities in the right places, policies will either fail to launch or stumble during implementation. At the same time, even modest reforms can have a significant impact if they are well-designed and supported by the necessary skills and systems.

In the end, the authors argue that health reforms, like all complex policies, must be treated as design problems that require strategic investments in capacity. This means thinking carefully not just about what to do, but how to do it, who will do it, and whether they are equipped to succeed.

If policy is a craft, then this paper is a guide to the tools that make it work. It brings conceptual clarity, practical insights, and a powerful reframing: capacity is not a generic virtue. It is a design imperative.

Governments that succeed in the future will be those that can design wisely, implement realistically, and adapt politically. That means not just hiring smarter people or buying better software, but investing in the distributed, layered capacities that allow policy systems to learn, adapt, and hold together under pressure.

For policymakers, scholars, and practitioners alike, Bali and Ramesh offer more than a framework. They offer a way to think, and that, too, is a form of capacity.

The Paradox of the Ambitious State

India’s policymaking energy is boundless. From Swachh Bharat Mission and Ayushman Bharat to Make in India and Digital India, we are constantly launching, relaunching, and refining public schemes. On paper, they aim to transform lives. On the ground, they often stagger under their own weight—misaligned with administrative capacity, underfunded, and poorly coordinated.

But this is not a uniquely Indian predicament. Across democracies, governments increasingly suffer from what Xavier Fernández-I-Marín and colleagues call the burden-capacity gap—the widening distance between the complexity of policy portfolios and the constrained capacity of bureaucracies to implement them. Their paper, titled Bureaucratic Quality and the Gap between Implementation Burden and Administrative Capacities, published in the American Political Science Review, puts forward a powerful idea: that the quality of bureaucracies must be judged not in isolation, but in how well the policy process is integrated vertically.

This concept—Vertical Policy Integration (VPI)—could well be the most underappreciated determinant of state effectiveness.

The authors add:

How can we explain this variation in the relationship between the number of policies requiring implementation and the administrative capacities available to do so? We argue that the answer to this question lies in the coupling of sectoral bureaucracies in charge of policy formulation and those in charge of policy implementation. We consider these patterns of vertical policy-process integration (VPI) an important yet relatively overlooked feature of bureaucratic quality. We expect that (1) the more the policymaking level is involved in policy implementation (top-down integration) and (2) the easier the policy-implementing level finds it to feed its concerns into policymaking (bottom up integration), the smaller the burden-capacity gap.

What Is VPI and Why Does It Matter?

Vertical Policy Integration (VPI) refers to the strength of connection between policy formulation and implementation. It is not about passing more laws or hiring more officials—it’s about whether different levels and arms of the state are communicating, accountable, and learning from one another.

The paper defines VPI through two dimensions:

Top-Down Integration

This refers to the extent to which policymakers—typically national ministries or apex bureaucrats—remain responsible for execution. Do they provide implementation budgets? Do they bear political or administrative consequences when programs fail? Do they directly oversee the organizational setup?

Top-down integration disciplines the policymaker: if you must pay for and deliver what you create, you’re less likely to over-legislate and more likely to align policies with capacity.

Bottom-Up Integration

This is about listening. Can implementers—local officials, frontline workers—feed their insights and operational constraints back into the policy process? Are they formally consulted? Are their experiences systematically evaluated and used to adapt?

Bottom-up integration ensures that policies reflect field realities and evolve over time. Without it, bureaucracies cannot learn, and policies fail silently.

Together, these two forms of integration—responsibility from the top, feedback from below—define how coherent, responsive, and effective a government truly is.

Without VPI, bureaucracies risk becoming Weberian in form but Kafkaesque in function—endlessly churning policies they lack the resources to implement. As the authors note, policymakers have stronger incentives to create policies than to staff and support their implementation. This leads to policy overload, administrative fatigue, and declining public trust.

By contrast, VPI aligns incentives, strengthens feedback loops, and disciplines policymaking. Governments with high VPI produce fewer but better-supported policies—and have a far lower risk of implementation failure.

Diagnosing the Gap: A Structural Problem

The authors build a compelling empirical case. Studying 21 democracies over four decades in six major policy areas, they show that:

• The number of policies (especially in social and environmental sectors) has increased dramatically,

• While administrative capacity has remained stagnant or declined,

• And this disconnect is most acute where VPI is weak.

While diagnosing the problem, the authors argue:

Power-seeking politicians have strong incentives to demonstrate their responsiveness to societal demands by constantly proposing new policies. The same logic, however, does not apply to the expansion of administrative capacities that are needed for properly implementing these new policies. As political responsibilities for implementation success are often unclear, electoral incentives for politicians to invest in administrative capacities for policy implementation are generally weaker than those for adopting new policies. Although such capacity expansion improves implementation effectiveness, attributing such improvements to the actions of particular political actors is in many instances more difficult for voters. While political actors have strong incentives to engage in policy production, their incentives to engage in costly improvements of administrative capacities are typically much weaker. Deviations from this pattern are largely confined to specific constellations where implementation failures have immediate individual consequences for voters, for example, in the case of policies related to service delivery. From a mere political logic, we should hence expect a toxic combination of strong policy growth and stagnating or even declining administrative capacities. Although the urgency of this problem varies across countries and sectors, the nature of the challenge essentially remains the same: if we assume that policymaking is exclusively driven by politics, we would expect an ever-growing burden-capacity gap, with more and more policies undermining rather than strengthening overall policy effectiveness in the long run.

Countries with strong VPI exhibit smaller burden-capacity gaps. Policymakers in these states are either constrained from producing unimplementable laws or are forced to fund and structure implementation effectively. They also receive better information from the ground and adapt policies accordingly. The result? Better alignment between ambition and ability.

From Reform to Rewiring: Toward a VPI-Conscious Bureaucracy

The real strength of the VPI framework is that it reframes governance reform not as a matter of more staffing or digitization, but as a matter of organizational architecture and coordination logic.

Most governance reforms focus on expanding inputs: more officials, more training, better IT systems. Important as these are, they cannot resolve the systemic problem of fragmented authority and disjointed learning.

VPI, by contrast, draws our attention to the invisible relational fabric of the state:

• Who owns implementation failures?

• Who designs the machinery for delivery?

• Who listens to the people closest to the problem?

Answering these questions calls for institutional rewiring, not just capacity building. It calls for a shift from government as a vertical hierarchy of commands to a network of mutual responsibility and learning.

Here are four ways to build a VPI-conscious bureaucracy:

1. Mandate co-design of major schemes between central, state, and local authorities to ensure joint ownership of both success and failure.

2. Embed feedback loops in program cycles through structured consultation with implementers and real-time feedback mechanisms.

3. Create institutional memory by rotating bureaucrats between policymaking and implementation roles.

4. Link funding to implementation realism—ensuring that policies come bundled with capacity assessments and resourcing plans.

The goal is to create a system where ambition is tempered by accountability, and ground-level intelligence shapes national strategy.

Conclusion: A State That Talks to Itself

Fernández-I-Marín and his colleagues offer a crucial insight: a capable state is not one that merely delivers, but one that listens—to itself.

Modern democracies are flooded with policies. Yet many of these wash out at the point of execution. Why? Because the state often behaves like a fragmented machine: the head doesn’t know what the hands are doing, and the hands have no voice to speak back.

VPI gives us a language—and a framework—to fix this. It shows that effective governance is not just about laws or funds, but about how power, responsibility, and knowledge flow across levels.

India, like many large states, has no shortage of policies. What it needs is a bureaucratic nervous system—a structure that connects brain to limb, plan to ground, ambition to action.

The future of governance may not lie in doing more. It may lie in connecting better.

When China began to reform its economy in the late 1970s, it faced a paradoxical challenge: how could a rigid, centrally planned state introduce market forces without losing control? Deng Xiaoping, the chief architect of China’s economic rise, captured the answer in a metaphor that has since become foundational to modern Chinese governance:

We must cross the river by feeling the stones.

This wasn’t just an ordinary phrase. It was a radically different way of thinking about reform—not as a single leap into the unknown, but as a cautious, incremental journey guided by trial and error. It implied that certainty would emerge not from ideology, but from practice. That progress would come not from blueprint-perfect models but messy, grounded experimentation. Policies would not be imposed from above with absolute confidence. Instead, they would be tried out in a few places, studied, adjusted, and only then scaled across the country.

This is the crux of China’s policy-making approach over the past four decades. It stands in stark contrast to how many other governments—especially in the West—tend to operate, swinging between sweeping reforms and abrupt reversals depending on political winds.

Raghuram Rajan explains this here-

But Deng’s metaphor was only the beginning. To truly understand how China institutionalized this philosophy at scale, we must turn to the work of political scientist Yuen Yuen Ang.

In her influential book How China Escaped the Poverty Trap, Ang introduces a compelling concept: “directed improvisation.” It’s an oxymoron at first glance—how can something be both directed and improvised?

But that’s precisely the genius of China’s system.

In Ang’s framework, China’s rise wasn’t due to top-down planning alone, nor was it a case of spontaneous bottom-up innovation. It was a blend—a dynamic process where the central government set strategic goals (the “direction”), while local governments were empowered to experiment with how best to meet those goals (the “improvisation”).

Local officials, operating under the watchful eye of the Party, were given room to pilot new ideas, adapt policies to local conditions, and even challenge conventional practices—so long as they stayed aligned with overarching national priorities. Successes were emulated. Failures were quietly abandoned. In this system, learning through doing was not just tolerated—it was institutionalized.

This is how Ang unpacks the directed improvisation:

What Deng called feeling the stones became a full-fledged governance technology—a way for a massive, diverse, and authoritarian state to adapt flexibly to complex challenges. It allowed China to sidestep the ideological rigidity that had plagued earlier socialist experiments and instead develop a governance model that rewarded learning, adaptation, and pragmatic problem-solving.

The result? A country that could launch a special economic zone in Shenzhen, test it, study it, and—once the results were in—replicate the model across the country. A country where carbon trading, rural insurance, property tax reform, and e-governance were all first incubated in select cities before going national.

But here’s the big question: does this system work? Is policy learning in China as effective as it looks from the outside? Are these stones truly helping China cross the river—or are they sometimes just slippery illusions?

To answer this, we turn to the groundbreaking work of Shaoda Wang and David Y. Yang. Their paper, “Policy Experimentation in China: The Political Economy of Policy Learning,” unpacks this experimentation machine using an enormous dataset of over 19,000 government documents and 652 policy trials conducted between 1980 and 2020.

What they find is fascinating—and a bit unsettling.

Their research shows that while China's policy experimentation is indeed vast and unique in scale, it is also deeply entangled in political incentives, elite bias, and flawed inference. As we’ll see in the sections ahead, the process of “feeling the stones” is not always as objective or effective as it may seem.

But before we dive into their findings, let’s remember what makes China’s model so distinctive: a willingness to pilot, learn, adapt, and scale—all within the boundaries of a single-party state. That model, for all its imperfections, remains one of the most important governance innovations of the modern world.

And this paper gives us a rare empirical window into how it works.

What Does Policy Experimentation in China Entail?

Between 1980 and 2020, the Chinese government conducted an extraordinary 652 distinct policy experiments. These initiatives were not fringe tests or bureaucratic footnotes. They involved major reforms—property taxation, carbon emissions trading, rural insurance, and fiscal decentralization—policies with broad national implications.

This massive experimentation infrastructure was identified and analyzed by Wang and Yang through a meticulous review of over 19,000 official government documents. Among the 652 experiments, approximately 42% eventually became national policy. The remainder faded from view—discontinued without fanfare, formal evaluation, or public discussion. Policy failures in China, as the authors note, are often absorbed quietly into the system, with lessons implicitly learned but rarely broadcast.

What sets this paper apart is not only the scope of the dataset but also the clarity with which it surfaces three core features of China’s experimentation regime—features that reveal both its strengths and limitations.

Finding 1: Elite Selection — Experiments Are Not Representative

The first major insight concerns who gets to experiment. In theory, if the central government’s goal is to understand how a policy will perform when implemented nationally, one would expect it to test the policy in a diverse and representative set of regions, encompassing rich and poor, urban and rural, coastal and inland jurisdictions.

In practice, the vast majority of experiments (87.7%) were launched in economically advanced regions. On average, experimentation sites were 44.2% richer (measured by local fiscal revenue) than non-experimentation sites. Even policies designed for poor or rural constituencies—such as agricultural insurance or poverty alleviation—were disproportionately piloted in more prosperous areas.

This skew is not entirely surprising. Wealthier jurisdictions tend to have more capable bureaucracies, stronger infrastructure, and greater implementation capacity. But Wang and Yang also point to deeper political dynamics. Local governments often compete to be chosen as pilot sites, while central ministries are incentivized to select jurisdictions that will ensure the policy performs well, what we might call "policy safe havens."

The result is a significant selection bias that undermines generalizability. Policies trialed in elite conditions often cannot be seamlessly scaled to average or under-resourced regions. The lessons learned, while useful in part, may offer a distorted view of a policy’s true effects.

Finding 2: Policy Experiments Look Great — But Only on Stage

Let’s imagine you're a local government official in China. One day, the central government selects your city to test a new policy—say, a program to improve rural healthcare or streamline business licenses. You know something very important: if this policy succeeds and gets rolled out nationally, your name will be associated with a success story. And in China’s political system, that could mean a big promotion.

So, what do you do?

You pull out all the stops. You divert more funding toward the program. You assign your best staff to oversee it. You make sure every step is monitored, reported, and tweaked to look good. You’re not just implementing a policy—you’re staging a performance. The goal is to make the policy look as effective as possible during the trial.

This is exactly what Wang and Yang found. During policy experiments, local governments spend significantly more money in the area being tested—about 1.3% more of their total budget. That’s a lot when we’re talking about municipal budgets. And this effort is especially strong in places where local leaders are young and ambitious—those who have more to gain from a successful experiment.

But here’s the catch: this extra effort disappears once the policy goes national.

When the same policy is implemented in other cities or counties—places that weren’t part of the experiment—there’s no special attention, no budget bump, and no incentive for local leaders to go the extra mile. The policy becomes just another item on a long list of government tasks. And as a result, it doesn’t perform nearly as well.

The researchers call this the "voltage drop" problem: the idea that energy and effectiveness often drop off sharply when a policy moves from pilot to scale. In this case, they found that over 70% of policies performed worse after being scaled up compared to how they performed during the trial phase.

In other words, many of these impressive policy experiments were not sustainable models—they were exceptional efforts made under exceptional circumstances.

This finding raises a big question: can we trust the results of policy trials if they’re driven by temporary incentives? If a policy only works because people are trying to impress their bosses, is it a good policy?

Wang and Yang suggest that unless the central government recognizes and adjusts for this performance inflation, it risks learning the wrong lessons—mistaking effort for effectiveness, and confusing showmanship for sustainability.

Finding 3: The Central Government Sometimes Misreads the Results

So far, we’ve seen that policy experiments in China often take place in rich areas and that local officials pull out all the stops to make those experiments look like a success. But that leads to an important question:

Is the central government aware of these distortions when it evaluates whether a policy should go national? Does it carefully separate real, policy-driven success from results that were just due to lucky timing or political theatre?

According to Wang and Yang, the answer is: not always.

They uncover strong evidence that the central government sometimes mistakes noise for signal—that is, it confuses outside factors that have nothing to do with the policy itself for proof that the policy is working.

Let’s break this down with two examples they explore in detail:

Example 1: The Land Revenue Windfall Illusion

In China, one of the main ways local governments make money is by selling land to developers. When the real estate market is booming and interest rates are low, these land sales shoot up, and cities suddenly find themselves flush with cash. This means they can build more infrastructure, provide better public services, and hit their economic targets more easily.

Now, imagine one of these cities is also a site for a policy experiment.

If the experiment coincides with a big land revenue windfall, completely unrelated to the policy, it may look like the policy is doing great. But in reality, it’s the extra money from land sales that’s lifting performance, not the policy itself.

Wang and Yang show that the central government often doesn’t account for this. When evaluating the experiment, it sees good results and assumes it must be because the policy worked. As a result, policies that happen to be tested in cities with lucky fiscal windfalls are more likely to be adopted nationwide, even though the windfall had nothing to do with the reform being tested.

Example 2: Leadership Turnover and Performance Bumps

Here’s another subtle problem.

Let’s say a policy experiment begins under one local leader, but halfway through, a new Party Secretary takes over—someone who’s younger, more ambitious, and eager to climb the political ladder. This new leader pushes hard to make the experiment succeed, not necessarily because the policy is amazing, but because they want to be noticed by Beijing.

Wang and Yang find that these mid-experiment leadership changes—especially when they bring in officials with stronger promotion incentives—often lead to spikes in local performance. Again, the policy looks successful, but the driving force is the new official’s energy and ambition, not the policy’s actual design.

And again, the central government often fails to filter this out. It may treat the experiment as a success and scale up the policy, without realizing that in most places, where there’s no leadership change or no extra ambition, the policy may fall flat.

These findings point to a critical challenge in policy learning: just because something works in one place doesn’t mean it will work everywhere. And if decision-makers don’t fully understand why it worked in the first place, they risk drawing the wrong lessons.

Wang and Yang argue that the Chinese central government sometimes lacks the analytical sophistication to properly interpret the results of its experiments. It doesn’t always separate the effects of the policy from the effects of temporary, local factors like windfalls or political turnover.

Drawing on a conceptual framework inspired by Al-Ubaydli, List, and Suskind, the authors argue that effective policy learning depends on three elements:

1. The policy itself

2. The site where it is tested

3. The conditions under which it is implemented

When the second and third elements are biased or non-representative, conclusions about the first are inevitably flawed.

Deng Xiaoping’s dictum—crossing the river by feeling the stones—remains one of the most evocative metaphors in modern political thought. It captures a state’s commitment to pragmatic, iterative reform, grounded in empirical learning rather than ideological rigidity.

Yet as Wang and Yang demonstrate, the stones underfoot may not always be stable. They may be carefully chosen, polished for display, or arranged to impress rather than to instruct. The riverbed itself may shift depending on who is doing the crossing, and under what incentives.

Still, China’s experimentation regime remains one of the boldest governance innovations of the 20th and 21st centuries. It institutionalizes learning at scale, embraces policy variation, and recognizes uncertainty as an inevitable feature of development. For scholars and practitioners alike, this paper offers a valuable empirical anatomy of that system—its ambition, its mechanics, and its limitations.

In short, even a powerful and centralized state like China can struggle with the same challenge that plagues governments everywhere: mistaking short-term success under special conditions for long-term, scalable policy effectiveness.

Learn more about the paper here:

Designing policy isn’t just about aiming high—it’s about landing right. Because good intentions without good design can still miss the mark. The aim of a policy may be noble—protect workers, improve wages, boost productivity—but the outcomes can be completely different, especially in economies where informality is high. This paradox is especially clear in India’s labor market, where a law meant to protect workers ended up pushing more businesses and workers into the informal sector.

In 2003, the state government of Andhra Pradesh took what seemed like a bold and progressive step. It introduced a reform aimed at improving workers' lives by banning the use of contract labor in a company’s core operations. The intention was to encourage businesses to give workers regular jobs with better pay, job security, and benefits like health insurance and pensions. On the face of it, this seemed like a win for workers’ rights.

But nearly two decades later, we know good intentions aren’t always enough.

A new research study by economists Ritam Chaurey, Gaurav Chiplunkar, and Vidhya Soundararajan—titled “Employment Protection Legislation and Informality: Theory and Evidence from India”—tells a more complex story. The reform may have helped some workers, but it also pushed many others deeper into the shadows of India’s already massive informal economy.

The authors make a compelling case that without considering how businesses react, even well-meaning labor laws can produce the opposite of what policymakers intended.

The Informal Reality

To understand why this happened, we need to first look at the nature of work in India. Most Indians—about 85%—work in what’s known as the informal sector. These are jobs in small shops, workshops, or home-based businesses that don’t follow labor regulations. They don’t offer job security, regular pay, or benefits.

Now, within the formal sector—companies that are registered and regulated—a lot of employers also rely on contract workers. These workers are not on the company's permanent payroll. They’re hired through middlemen or for short periods, and companies use them to avoid the costs of labor laws, such as paying for severance or social security.

So, when the Andhra Pradesh government introduced a law banning contract labor in core business operations, they were trying to reduce this workaround. They wanted companies to stop treating contract workers as second-class employees and offer them the same protections as regular staff.

But as the researchers show, what followed wasn’t a straightforward transition from insecure to secure jobs.

The Reform and Its Effects

To see what changed after the 2003 reform, the researchers dug into multiple large-scale national datasets that track firms and workers across India. They used a "before-and-after" approach, comparing what happened in Andhra Pradesh after the reform with what happened in other Indian states that didn’t change their labor laws.

Here’s what they found:

The good news - Formal companies hired fewer contract workers.

The law did have its intended effect within formal companies. Firms in Andhra Pradesh cut down their use of contract labor by nearly 50%. Many of them increased the number of full-time, regular workers they hired instead. The share of contract workers in the workforce dropped, and payroll jobs went up.

The bad news - More firms and workers moved into the informal economy.

But that wasn’t the whole story.

At the same time, more businesses chose to remain informal—they avoided registering altogether to stay out of reach of labor laws. Workers in Andhra Pradesh became more likely to be hired casually or without contracts, often earning lower wages than before.

In short, while the reform reduced one type of informal work, it led to an increase in another. Many businesses, faced with higher costs of complying with the law, simply found it easier to stay outside the formal system altogether.

The Hidden Costs: When Formality Becomes a Burden

On paper, the Andhra Pradesh reform seemed like a step forward for workers’ rights. By banning the use of contract workers in a firm’s core operations, it aimed to push companies toward hiring regular, payroll employees who would enjoy greater protections and stability.

But what this approach failed to consider is how businesses actually make decisions—especially in economies where informality is not a fringe but the norm. The reform treated formality as a moral imperative rather than an economic choice. The reality? For many businesses, especially small and mid-sized ones, formality is expensive and risky.

To explain why the reform had such unintended consequences, the researchers built a theoretical model—a kind of economic simulation that helps track how firms respond to different policy environments. It’s based on real-world data and behavior, and its findings are both sobering and illuminating.

Here’s what the model shows:

1. Rising Costs, Falling Compliance

Once the Andhra Pradesh government imposed a blanket restriction on contract labor in core activities, formal firms had fewer options. Without access to flexible, short-term labor, they had to either:

• Hire regular workers, which meant higher wages and compliance with a host of laws (e.g., Provident Fund, ESI, termination protections); or

• Cut back their operations, exit the formal sector, or avoid registering altogether.

For firms already on the margins—low in capital or struggling with demand—this new constraint tipped the balance. Formality became not just harder but untenable.

2. Exit from the Formal Sector

The model predicted, and real data confirmed, that more firms chose to go informal after the reform. The authors used data from India’s Economic Census (1998 and 2005) and found a striking shift: the likelihood of a large firm (with 50 to 100 workers) being unregistered rose by over 50% in Andhra Pradesh relative to other states.

This is significant. When large firms—those typically expected to have the resources and incentives to be formal—choose to remain outside the legal fold, it suggests that something deeper is broken in the incentive structure.

3. Lower Productivity, Shrinking Wages

What happens when more firms go informal?

Productivity drops. Informal firms tend to be smaller, use less capital, and are less likely to adopt modern technologies. They also lack access to formal credit, infrastructure, and skilled labor. As more firms move into this less productive space, the entire economic engine slows down.

This was visible in the wage data, too. Using the National Sample Survey, the researchers found that average daily wages in Andhra Pradesh declined by 4.6% after the reform, relative to other states. Workers were also:

• 14% more likely to be on informal contracts;

• 9% more likely to work in an unregistered firm;

• 3.7% more likely to be casual workers (i.e., daily wagers or piece-rate workers).

In other words, instead of moving workers into better jobs, the policy pushed them into worse ones.

4. General Equilibrium Effects: A Domino Impact

Most policy evaluations look at direct impacts: fewer contract workers, more payroll workers—check. But the deeper insights in this paper come from looking at general equilibrium effects—how one change ripples through the entire system.

Here’s how it played out:

• As more firms left the formal sector, the overall demand for labor fell.

• With lower demand, workers across the board—formal or informal—had weaker bargaining power.

• Wages dropped, even for those who remained in regular jobs.

• Consumers had less to spend, further slowing down business growth.

The model predicted a 2% decline in total factor productivity (a measure of how efficiently inputs like labor and capital are used) and a 9% drop in real income, a direct hit to household welfare. These are not abstract figures. They represent real losses—reduced purchasing power, fewer opportunities, and stalled economic mobility for millions.

At its core, the Andhra Pradesh reform mistook symptoms for causes. It treated the use of contract labor as the problem, rather than understanding why firms preferred it in the first place. In doing so, it treated a structural issue as a moral one and used a stick where a smarter carrot might have worked better.

As the researchers show, if the goal was to improve job quality and worker protection, a better strategy could have been to reduce the cost of hiring formal workers—simplifying compliance, reducing red tape, and making it easier for businesses to follow the rules. The consequences of not doing so are now painfully clear: what was meant to be a step toward formality ended up reinforcing informality, hurting the very people it aimed to help.

What Could Have Been Done Differently?

When the Andhra Pradesh government banned contract labor in core business activities in 2003, they were trying to do the right thing—stop companies from exploiting temporary workers and push them to offer regular, secure jobs instead. But as we’ve seen, the policy didn’t fully deliver on that promise. Instead of creating more good jobs, it made running a formal business harder. Many companies, especially small and mid-sized ones, responded by pulling out of the formal system entirely.

So, what could have been done differently?

That’s the question the researchers behind this study explored. And their answer is simple but powerful: instead of punishing companies for doing the wrong thing, why not make it easier for them to do the right thing?

To understand this better, let’s look at two different approaches they tested using data and simulations.

Option 1: The Path Andhra Pradesh Took – Ban Contract Workers

In this approach, the government made it harder and more expensive for companies to hire contract workers. If you were running a business and still used contract labor in your main operations, you could face penalties or even legal action.

At first glance, this seemed to work. Many companies cut back on contract workers and hired more full-time employees instead. But there was a catch. Because it became harder and more expensive to stay formal and follow the new rules, many businesses decided to avoid the system altogether. They stayed small, off-the-books, and unregistered—where the government couldn’t see or regulate them.

As a result:

• Fewer new businesses joined the formal economy.

• Wages dropped for many workers because there were fewer good jobs available.

• Overall, people had less money in their pockets, and the economy slowed down.

So even though the reform reduced the use of contract labor in some companies, it ended up hurting workers in other ways—through lower pay and more informal jobs.

Option 2: A Different Approach – Make Formal Hiring Easier

Now imagine if, instead of making it harder to use contract workers, the government made it easier to hire full-time employees. What if rules were simpler, paperwork was less, and the cost of following labor laws went down?

That’s the second option the researchers explored. And the results were much more encouraging.

In this scenario:

• Companies still reduced their use of contract workers—but did so willingly.

• More businesses stayed formal, and some informal ones even registered.

• Workers had access to better jobs and higher wages.

• The economy became more productive, and overall incomes went up.

In short, this approach didn’t scare businesses away. It gave them a reason to stay and grow within the formal system.

The big lesson from these two scenarios is clear: if you want companies to do the right thing, make it easier—not harder—for them to follow the rules.

In places like India, where the informal economy is huge and many businesses are small, labor laws need to be not just protective—but practical. If the rules are too complicated, costly, or rigid, companies won’t follow them. They’ll take the easier route and operate informally, leaving workers with no job security, no benefits, and no protections.

That’s why instead of just banning contract labor, the focus should be on:

• Simplify labor laws so they’re easier to understand and follow.

• Reducing the burden of compliance, especially for small businesses.

• Offering incentives for companies that hire workers on a regular payroll.

• Using technology to cut down paperwork and reduce middlemen.

Final Thoughts

Regulation alone doesn’t change behavior. Businesses adapt to the rules in ways that minimize their costs—even if that means avoiding the rules altogether. When enforcement is weak and the informal economy is dominant, simply punishing non-compliance won’t work. It might make things worse.

Instead, governments need to create pathways that make it easier for businesses to follow the rules. This could mean reducing red tape, offering incentives for formal hiring, or improving access to credit for small firms. Crucially, labor law reforms must account for how policies interact with informality.

At its heart, this story is about the limits of top-down policymaking in complex economies. Labor reforms are not just about laws—they’re about incentives, institutions, and everyday realities on the ground.

The Andhra Pradesh reform began with the noble goal of protecting workers. But in trying to fix one problem, it ended up deepening another. The result was a rise in low-quality jobs, a shrinking formal sector, and a missed opportunity to improve worker welfare.

If there’s a silver lining, it’s this: We now know better. With rigorous research and thoughtful design, future labor reforms in India and beyond can aim not just to punish non-compliance but to build a system where doing the right thing is also the easier thing.