Arvind Subramanian, Devesh Kapur, Ashutosh Varshney, Rohit Lamba - India's Development Odyssey

ELI5 / TLDR

Arvind Subramanian and Devesh Kapur have written an 800-page book — A Sixth of Humanity: India’s Development Odyssey — argued out at George Washington University with Ashutosh Varshney and Rohit Lamba as discussants. The central frame is “precociousness”: India did everything in the wrong order. Democracy before development, services before manufacturing, welfare before public goods, higher education before primary education, laws before the capacity to enforce them. The trade kept the country together but left it poorer than it should be. The book treats the first 30 years (1950-80) as a scarcity regime rather than import substitution, characterises the post-1991 era as East Asian growth without East Asian policies, and reframes the Modi years as a continuation of welfareism that began under UPA. The discussants push back on what’s missing — civilisational thought, the role of personalities, why one-third of India (the south and west) actually grew at Chinese rates while the Hindi heartland stagnated.

The Full Story

The book in one frame: precocious sequencing

Kapur opens with a thesis that runs through the entire 800-page volume: India’s development story is a story of doing things out of order. He calls it precociousness — unusual sequencing.

Democracy before the development. Services before manufacturing. Globalization of high-skill talent via migration before low-skilled exports through trade. Welfare before providing public goods. State capitalism before building public infrastructure. Physical capital before human capital. Higher education before primary education. Enacting laws galore without putting together implementation capabilities.

The list is the book. Each item is a chapter, each chapter is a counter-narrative against either the standard left or standard right reading of what India did. The argument is also unusually data-driven — Kapur and Subramanian read roughly 3,000 official government reports and built original time series from them. Three kinds of comparison run through: across time, across Indian states, and across countries (China most often, but also Korea, Brazil, Mexico, Turkey).

Democracy first — the choice nobody else made

The graph that opens Kapur’s case shows four models. The Western model (US, UK) — democracy grows as per capita income grows. The East Asian model (Korea) — development first, democracy at middle income. The Chinese model — rapid growth, no democracy. India — full franchise at a per capita income lower than almost anyone has ever attempted, men and women at the same time, in a society that was at independence “much more patriarchical” than it is now.

The pay-off Kapur claims is order. Nation-building everywhere is violent — the US Civil War killed 2% of the population, which scaled to India would be 30 million dead. India built its state with mass violence well below the comparator group. Most countries built nationhood around a single language, a single religion, a single ethnicity. India used democracy itself as the instrument — compromise, accommodation, the daily messiness of representative politics — and held a maddening diversity together. The same shows up in inflation: hyperinflation is a measure of economic disorder, and India avoided every bout except the 1973-74 oil shock while Turkey, Brazil, Mexico were repeatedly blown out.

But democracy has costs. Kapur calls them the Kamadhenu problem — after the Hindu cow-goddess who gives milk to everyone. The Indian state hands subsidies to rich farmers, big industrialists, and small voters alike, and gives all of them tax breaks. The chart he shows: India had the highest fiscal deficit of any major emerging economy across decades when high growth should have collapsed deficits, and the picture gets worse once you add state-owned enterprise losses. The second cost is the perpetual election machine. Pre-1967 about a quarter of the population voted in some election each year; now it’s more than half. The election-free interval at the centre has shrunk from five years to about 1.2 years once you count national, state, panchayat, and urban bodies. Politicians’ time horizons collapse to match.

The first 30 years were not import substitution

Subramanian takes over and pushes harder on the economic story. The standard reading — Amartya Sen, Sukhamoy Chakravarty, Pranab Bardhan — is that the first three decades were import substitution and growth wasn’t that bad compared with the previous 30 years. Subramanian’s contention:

No, it is not import substitution. Because when you do import substitution, you protect domestic firms against foreign competition and therefore you want to promote domestic entrepreneurship in industry. But what we turned around and did had this horrendous system of industrial licensing domestically. So we thwarted the domestic private sector instead of the aim of import substitution being to promote both the domestic private and public sector. So it created in fact a scarcity regime, not an import substitution regime.

The data: in 1950, 79% of countries had higher per capita GDP than India. By 1980, India had fallen behind essentially all comparators. China was a political and economic mess for that whole period — a black-market exchange rate premium far above India’s, the Cultural Revolution — and yet by 1980 China had pulled level with India and was about to pull decisively ahead.

Concrete temples that didn’t get built

Sunil Khilnani, in The Idea of India, called India’s public sector projects “the temples of modern India.” Subramanian inverts the line: India didn’t fall in love enough with concrete. The book’s data shows that in the first 30 years there was one bridge over the Ganga and one bridge over the Brahmaputra. State capitalism crowded out public infrastructure. The book aggregates returns from every central public sector enterprise from 1970 onward and the headline state ones — central PSEs returned well below any reasonable opportunity cost, state PSEs much worse. Subramanian’s coined refrain:

Anything that the centre does badly in India, the states do much worse.

The same story for public sector banks: total commercial lending in India ran 8-25% of GDP across 50 years, against 100-200% in China and Korea. The opportunity cost of running PSEs at sub-economic returns: roughly 1.5-2.5% of GDP foregone, every year, for fifty years. Enough to have doubled central infrastructure investment annually.

The 1991 trick — East Asian outcomes without East Asian policies

The line that should travel furthest from this talk:

India achieved in those 20 years post-1991 East Asian growth and East Asian trade in terms of outcomes, but it did it without East Asian policies and without East Asian structural transformation.

Subramanian sharpens this for the Washington audience. The World Bank has been doing a mea culpa on its old East Asian Miracle report — admitting industrial policy mattered more than the 1990s consensus allowed. He’s not buying. India achieved East Asian growth on plain-vanilla Washington Consensus policies. No mercantilism. No serious industrial policy. What India didn’t get was the East Asian structural transformation — too much agriculture for too long, too little formal manufacturing, too much high-skill services too early.

The phrasing: prolonged ruralisation, premature de-industrialisation, precocious serviceification.

The international counterpart shows up in two charts. On the migration side, Indian-Americans earn 32% more than the average American — the highest of any ethnic group — meaning India over-exploited globalisation through its high-skill diaspora. On the trade side, India and China have similar labour forces but China captures 40-45% of global low-skilled exports while India has 3-4%. The country sent its talent abroad and shortchanged its low-skilled workers at home.

One-third of India did grow like China

The factoid that should puncture every wistful Lutyens dinner conversation about Chinese envy:

If you start the clock in 1980 and run it for 40 years, one-third of India — the southern states, western states, Maharashtra, and Haryana — grew almost as rapidly as China and for as long.

Karnataka grew per capita income 7.4x between 1980 and 2020, Kerala 7.1x, Tamil Nadu 6.9x, undivided Andhra 6.8x. China was 7.3x. The right question is no longer “why didn’t India grow like China?” It’s: “why didn’t the Hindi heartland — the other two-thirds — grow like the southern third?”

Varshney’s pushback: the looming question

Ashutosh Varshney calls the book “a truly veritable intellectual achievement” and predicts it will replace Drèze and Sen as the canonical text on Indian development. He’s generous on the multidisciplinarity — economics and politics genuinely fused rather than economics with politics as a residual — and on the prose, which he calls felicitous and epigrammatic, a rare thing in academic writing.

His larger contribution is to add a fourth project to the Indian state’s founding agenda. The standard three (he wrote them as the opening lines of his last book) are national unity, dignity for those at the bottom, and the elimination of mass poverty. Kapur and Subramanian add a fourth: order. Partition killed half a million to a million people and displaced 15 million across the border. The first job of the state, Hobbesian and Huntingtonian, is security. Development cannot proceed in conditions of mass insecurity, and Indian political economy has tended to forget this.

Then Varshney sharpens his blade. If democracy is a tax on development, why does the south grow at Chinese rates?

South and west marching ahead almost like China, and north and east lagging behind. Is the more relevant explanatory variable, by any chance, not democracy but democracy plus something? What might the plus be?

His hypothesis: ideology. Democracy can coexist with very different ideological orientations among political elites and the policy machine. The southern states had serious social reform movements before universal franchise (Kerala’s land reforms; the anti-Brahmin movements that swept Madras Presidency). The northern states didn’t. So when democracy arrived in the north, the burden of social transformation fell entirely on the state, which made capturing the state a high-stakes game and produced the conflictual, low-equilibrium politics that characterises UP and Bihar today.

He also corrects a small empirical claim: Sri Lanka had universal franchise from 1931, Costa Rica a year before India. India was not the first universal-franchise democracy in the global south. It was the first large one that lasted.

Lamba’s pushback: where is the civilisational layer?

Rohit Lamba calls the book one of the three he’d put in the canon for any student trying to understand India — alongside Khilnani’s Idea of India and Ramachandra Guha’s India After Gandhi. His critique is theoretical and gets uncomfortable in interesting ways.

First, the precocious frame may do too much. Precocious democracy, precocious services, precocious globalisation, precocious redistribution, precocious wealth concentration — the same hammer hits every nail. Is precociousness a cause, a culture, a choice, or a constraint? Each version implies different stories. The book sometimes blurs them.

Second, failures are well theorised; successes are not. We have the Kamadhenu problem, the capacity-by-design pathology, the precocious sequencing trap — all explanations for what didn’t work. But why did services succeed? Why did the south grow at Chinese rates? The book has anecdotes but not a theory.

Third — and this is where he gets provocative — the book never engages the civilisational question that animates daily life for most Indians. Lamba names it carefully:

What is remarkable to a median Indian, and definitely a median Hindu Indian, is the idea that there is only one large community-civilisation that survives the spread of global Abrahamism. It’s very central actually to the median Indian, like the identity of a lot of Indians. This informs a lot of their daily living. And this somehow doesn’t enter academic discourse. I don’t know why.

He links this to Gandhi: India achieved independence with relatively little violence partly because the independence movement itself was non-violent, and the non-violence was civilisational and spiritual rather than tactical. The pluralism the book celebrates may have civilisational rather than purely democratic roots. Lamba’s complaint is that Indian academic discourse — including this book — pays a kind of fixed cost of disclaimers about Hindu majoritarianism before it can think about any of this, and the disclaimers crowd out the substantive question.

Authors respond on regional divergence

Kapur picks up the divergence puzzle. Why did the south grow and the north didn’t? His compressed answer: the south established order better than the north because pre-independence social movements (anti-Brahmin movements, land reform agitation) did some of the work of social transformation before the state arrived. In the north, the state had to do it all, which meant capturing the state became existentially valuable, which produced low-trust, conflictual politics.

Subramanian adds: the divergence really begins in the 80s and 90s, after fiscal and policy decentralisation gave states agency. And the striking common thread among the high-growth states is globalisation. Tamil Nadu does manufacturing exports like Gujarat. Karnataka does services exports via Bangalore. Haryana does both via Manesar (Maruti) and Gurgaon. And Kerala — the most unique development model in the world — does neither manufacturing nor services nor agriculture but globalises through labour migration to the Gulf.

The Modi era as continuation

Asked whether Hindu nationalism changes the development trajectory, Subramanian’s answer is unsentimental: the welfareist turn started under UPA II (2009-14) with MGNREGA and the rights-based legislation. Modi continued it but changed the form — from the rights-based welfareism of the UPA to what they call the “new welfareism” of public provision of essentially private goods (toilets, water, gas connections, Aadhaar) leveraged through technology, culminating in cash transfers. The animating spirit is the same. The signature achievements (GST, public infrastructure) had antecedents that were just lacklustre earlier and got turbocharged after 2014.

Kapur is willing to say plainly:

We believe that the Indian nation state would have had much more difficulty in surviving as a cohesive entity without democracy. Look at the Soviet Union, look at Yugoslavia, look at Sudan, the civil wars in Nigeria, Ethiopia, Indonesia, Sri Lanka. Look at all the constituents of British India from Myanmar to Pakistan — everywhere there’s been disorder. So we are very clear. Of course democracy has costs. Start with electoral finance — it is massively expensive and a big source of corruption. But that is the trade-off. India created a multi-ethnic state which few others have managed.

Subramanian adds the orthogonal point: industrial licensing had nothing to do with democracy. Import substitution had nothing to do with democracy. The choice to over-invest in services rather than manufacturing had nothing directly to do with democracy. Some of India’s worst self-inflicted policy choices were independent of the regime type.

Federalism and the looming fight

The closing exchange is on delimitation — the constitutional exercise that will redraw Lok Sabha seats based on population, possibly tilting representation toward the high-fertility, high-population, slow-growing northern states. Kapur and Subramanian flag two structural problems behind any specific delimitation outcome.

First, every federal system struggles when performance diverges across entities. Indian fiscal transfers from rich to poor states have increased over time, and the southern states (and increasingly the western states) feel they are being penalised for both their better fertility performance and their better growth performance. The richest state’s per-capita GDP is now 6x the poorest, up from 2x. A federation built for transfers under shared poverty is now managing extreme divergence.

Second, the breakdown of trust between the centre and the states under the current government — “the unilateralism, the lack of discussion, the lack of process, the lack of inclusion” — is turning cooperative federalism into uncooperative federalism. The mechanics of any specific reform matter less than the trust deficit underneath.

Kapur’s closing observation that almost nobody is thinking about: delimitation within states will probably matter more than delimitation across states. Urban India is severely under-represented and that’s why “urban India is treated like crap — all subsidies, everything goes to rural India.” If delimitation finally rebalances within-state representation, urban interests get a voice they’ve never had.

Key Takeaways

The book’s organising frame is “precociousness” — India did everything in unusual sequence: democracy before development, services before manufacturing, welfare before public goods, laws before implementation capacity.
The first 30 years (1950-80) were not import substitution but a scarcity regime — domestic industrial licensing thwarted the very entrepreneurs that import substitution was supposed to protect.
By 1980, India had fallen behind essentially every comparator country in per-capita GDP terms; in 1950, only 21% of countries were poorer than India.
Public sector enterprises returned well below opportunity cost for fifty years — central PSEs cost India roughly 1.5-2.5% of GDP every year. Enough to have doubled annual infrastructure investment.
“Anything the centre does badly in India, the states do much worse” — courts, police, public service commissions, finance commissions, state PSEs.
Post-1991 India achieved East Asian growth and trade outcomes on plain Washington Consensus policies — without East Asian industrial policy and without East Asian structural transformation.
The Indian diaspora over-exploited globalisation through high-skill migration (Indian-Americans earn 32% more than average Americans) while India captured only 3-4% of global low-skilled exports versus China’s 40-45%.
One-third of India — the south, plus Maharashtra and Haryana — grew at Chinese rates from 1980-2020. The puzzle is the other two-thirds, mostly the Hindi heartland.
The election-free interval at the centre has collapsed from five years to about 1.2 years once you count all electoral levels. Politician time horizons follow.
Modi-era welfareism is a continuation of UPA welfareism with the form changed from rights-based to technology-leveraged provision of essentially private goods (toilets, gas, cash transfers).
Some of India’s worst policy choices — industrial licensing, services-over-manufacturing — were orthogonal to democracy. Democracy is not the only thing to blame, and not the right thing to blame for those.
Varshney’s challenge: if democracy is a tax on development, why did the south grow at Chinese rates? Hypothesis — democracy plus prior social reform movements (which the north didn’t have) produced both order and growth.
Lamba’s challenge: the book theorises failure well but success poorly, uses “precociousness” as too universal a hammer, and avoids the civilisational layer that animates median Indian identity.
The federalism crisis ahead has two roots — divergent state performance (fertility and growth) creating perverse transfer incentives, and the centre’s breakdown of trust with the states. Delimitation is a symptom.
The under-discussed delimitation effect is within states — urban India will finally get representation proportional to its population.
The five Indian growth states (TN, Karnataka, Haryana, Maharashtra, Kerala) all share one trait — high engagement with global flows, whether through manufacturing exports, services exports, or labour migration.

Claude’s Take

This is the kind of panel that justifies the form. Two heavyweight authors who have written the definitive book of their generation on Indian political economy, two discussants of the calibre to actually push back, and a chair who lets the argument breathe. The book itself sounds like the rare thing — a 800-page work that is data-driven without being dry, multidisciplinary without being mushy, and willing to take positions that annoy both ideological camps simultaneously.

The single most provocative claim is the one about 1980-2020. One-third of India grew at Chinese rates for forty years. The conventional Lutyens lament — why can’t we be like China — is empirically wrong about a third of the country, and the right question is internal not external. That alone reorients how you should think about Indian growth: not as a story of China envy but as a story of why some Indian states converged on East Asia and the rest didn’t.

The democracy-as-instrument argument is also stronger than the usual “we’re proud of our democracy” version. Kapur and Subramanian aren’t romantics. They cost it out — the Kamadhenu problem, the perpetual election machine, the impossibility of land reform — and still conclude that the alternative (a Soviet-style or Yugoslav-style breakup of a maddeningly diverse subcontinent) was much worse. That’s a defensible adult position rather than a flag-waving one.

The discussants earn their fees. Varshney’s “democracy plus” formulation lands hardest — democracy alone doesn’t explain why one third of India grew like China and the other two-thirds didn’t, but democracy combined with prior social reform movements (which the south had and the north didn’t) might. Lamba’s complaint that the book lacks a theory of success is also fair — it’s much easier to explain pathology than to explain Bangalore, Tamil Nadu manufacturing, or Kerala’s labour-export model. Lamba’s final provocation — that academic discourse on India pays a fixed cost of disclaimers about Hindu majoritarianism before it can engage civilisational thought — is the kind of thing that probably needed a non-Indian-American audience to be said cleanly. Whether you agree or not, it’s a real question and the book apparently doesn’t take it on.

The federalism point at the end is the one with operational weight for the next decade. India is staring down a structural problem: a federation built for shared poverty is now managing 6x divergence in per-capita income, fiscal transfers from rich to poor states are growing not shrinking, and the southern states are being asked to subsidise the same northern states that may soon get more parliamentary seats too. The trust deficit between Delhi and the state capitals is real and the mechanics of any specific reform won’t fix it. This is going to be the dominant political-economy question of the 2030s and the book seems to flag it without resolving it — which is honest, given that nobody has resolved it.

Score: 9/10. Substantive, intellectually serious, two of India’s best public intellectuals being genuinely challenged by two of India’s best political theorists. The book itself, when it lands, is going to be the new reference text. One point off only because the panel format inevitably compresses arguments — you’d need the 800 pages to fully judge the precociousness frame, and the discussants’ best critiques (Lamba on civilisation, Varshney on democracy-plus) deserve longer engagement than 90 minutes allowed.