The Subansiri River has a beautiful name. In Assamese, Subansiri translates roughly to “golden river” and for decades, Indian energy planners saw it exactly that way. The river, which originates in Tibet, cuts through Arunachal Pradesh before joining the Brahmaputra in Assam, and its gradient made it perfect for hydroelectric power. In 2003, the government cleared India’s most ambitious hydel project on its lower reach: the Subansiri Lower Hydroelectric Project, a 2,000 MW dam at a place called Gerukamukh, right on the Arunachal Pradesh-Assam border.

Twenty-two years later, the dam is finally generating electricity. And the states it was built to power are refusing to buy it.

Subscribe now

From ₹6,000 Crore to ₹26,000 Crore: A Timeline of Delays

To understand why, you need to follow the money and the clock.

When the government gave the project environmental clearance in August 2003, the cost estimate was ₹6,285 crore at 2002 price levels. The construction contract was awarded in December 2003 to a consortium of Russian and Indian firms. Work formally began in January 2005, with a target completion date of 2010. The plan was to build a 116-metre-high concrete gravity dam, the largest in all of Northeast India and eight turbines of 250 MW each, capable of generating 7,421 million units of electricity every year.

None of that went to plan.

Construction faced early trouble from land acquisition delays, difficult monsoon seasons that left workers with only half a year to build, and unexpected geological conditions that caused landslides and slowed tunnel excavation. The river was successfully diverted in April 2007. The foundation was ready for construction in 2008. And then, in December 2011, everything stopped.

Protests had been brewing for years in Assam. Farmers, fishermen, and environmental groups downstream feared that the dam would trigger catastrophic floods, the Subansiri valley had been devastated before, including by a massive earthquake in 1950 and that daily water releases to generate peak-hour power would destroy the river’s ecology.

In December 2011, protesters physically blocked construction equipment from accessing the site. Work was suspended entirely. It did not resume until September 2019, a gap of eight full years.

During those eight years, the project sat incomplete. But loans don’t pause. Interest accumulated. Contractor escalation charges mounted. The cost estimate, which had already risen to ₹10,660 crore by 2010, ballooned further. By the time the project returned to active construction, the total projected cost had crossed ₹26,000 crore - more than four times the original estimate.

That cost feeds directly into the electricity tariff. See, when NHPC (the government company that builds and operates these dams) sets a price for its electricity, it calculates what it needs to recover its total investment over the project’s lifetime, plus a regulated return. The more the project costs, the higher the tariff.

Subansiri’s tariff, once projected at under ₹2 per unit, now stands at over ₹7 per unit. The national average tariff for NHPC’s hydel power was ₹3.15 per unit in 2023-24.

Buying Subansiri power would cost a state more than double that.

What “Buying Power” Even Means

Now, a brief explanation of how this power market works, because it matters.

When the central government builds a hydroelectric project, it allocates shares of that power to specific states, called beneficiary states. These come in two kinds. The first is a firm allocation - a fixed percentage of the plant’s output that states are obligated to purchase. The second is an unallocated central share - a pool that states can optionally draw from, typically offered when other states don’t want their share.

Subansiri’s power is allocated to 16 beneficiary states across India. Assam, being the downstream state, has a firm allocation of 78 MW and receives 7.5 MW as free power from Arunachal Pradesh’s 12% free-power entitlement as the host state. But it has also been offered an additional 49.65 MW from the unallocated central share. Meghalaya has a firm allocation of 19 MW and has been offered an additional 36.75 MW from the same pool.

When states refuse power, that electricity doesn’t disappear. It flows into India’s power exchange system - specifically the Indian Energy Exchange (IEX), a real-time market where surplus electricity is bought and sold on hourly slots, like a commodity. Prices on the exchange fluctuate with demand, often settling between ₹3-5 per unit. Selling ₹7/unit power at ₹3-4/unit means selling at a loss - a loss absorbed by NHPC, a public sector company, and ultimately by the taxpayer.

Share

The Pattern Behind the Numbers

The protest movement that halted Subansiri for eight years was trying to protect communities downstream. But the eight-year suspension did not stop the dam. The dam was always going to be built. What the suspension did was increase the cost of construction by roughly ₹20,000 crore, raise the electricity tariff to a level that makes the power commercially unviable, and deny the host state and downstream communities fifteen years of free power, royalties, and development funds.

The Subansiri Lower project finally has three of its eight units operational. Unit 1 completed its mechanical run on October 24, 2025. Unit 2 was formally commissioned on December 23, 2025. The remaining five units are expected to come online in phases through 2026-27. When fully operational, it will be India’s single largest hydroelectric plant.

Subansiri is not an isolated story. India’s hydroelectric pipeline has a recurring script: optimistic cost estimates to clear approvals, environmental and social clearances obtained before detailed assessments are complete, construction halted by litigation and local resistance, interest accumulating on debt for years with no revenue, and electricity priced out of the market it was meant to serve.

Leave a comment

India’s Supreme Court is currently deciding one of the most consequential yet trivial questions in the country’s economic history. Nine judges are sitting together, the largest possible bench for any case, to answer this: what exactly is an ‘industry’ under Indian law?

You might think this sounds like a technicality. It isn’t.

A Word Worth Billions

The Industrial Disputes Act, 1947 (IDA) has long been the backbone of India’s labour law framework. It governs how workers can be hired, fired, and heard. It defines who can strike, who is entitled to retrenchment compensation, and what mechanisms exist to resolve conflicts between employers and employees. Every single one of these protections hinges on one threshold question: does your workplace qualify as an ‘industry’?

Section 2(j) of the IDA offers a definition, but it is famously elastic. It says an “industry” includes “any business, trade, undertaking, manufacture or calling of employers” and also any “calling, service, employment, handicraft, or industrial occupation or avocation of workmen.” That’s a lot of words doing very little work. The legal chaos that followed has been decades in the making.

Share

The 1978 Earthquake

The story really begins with a water board in Bangalore. In the early 1970s, the Bangalore Water Supply and Sewerage Board fined several of its employees for misconduct. The employees challenged this before a Labour Court. The Board pushed back, arguing it was a statutory body performing essential civic duties, providing water to citizens, which made it, essentially, an arm of the sovereign state and therefore exempt from labour law.

A seven-judge Supreme Court bench disagreed. In 1978, Justice V.R. Krishna Iyer delivered a landmark judgment that changed the landscape entirely. He laid down what became known as the “Triple Test” - to qualify as an industry:

1. An activity must be systematic

2. It must produce goods or services

3. And it must involve an employer-employee relationship.

The motive, profit or public welfare, didn’t matter. The nature of the activity did.

Under this test, the court ruled that hospitals, educational institutions, clubs, charitable organisations, and even government welfare departments were all ‘industries.’ Millions of workers who had previously existed outside the protection of labour law were suddenly brought in.

Justice Iyer himself was candid about the urgency.

Parliament did eventually act.

The Industrial Disputes (Amendment) Act, 1982 tried to narrow the definition and carve out exceptions for several categories. But here’s the remarkable part: the 1982 amendment never came into force. It was passed but never notified. It sat on the shelf for over four decades, effectively becoming a legislative ghost, a law that exists but doesn’t apply.

Subscribe now

A Definition Passed Between Benches

That left the 1978 ruling as the operating law, generating a steady stream of litigation. Courts kept wrestling with edge cases. Is a social forestry department an ‘industry?’ Is a university? A government hospital? Different benches gave different answers, and the contradictions piled up.

In 2002, the State of Uttar Pradesh formally challenged the 1978 ruling before the Supreme Court, in the case now known as State of UP v. Jai Bir Singh.

A five-judge bench in 2005 heard arguments and concluded the matter was too large for them, the uncertainty was too deep, the implications too wide. They referred it to a larger bench. In 2017, a seven-judge bench led by then CJI T.S. Thakur agreed it needed nine judges, citing the ‘serious and wide-ranging implications.’ The case was referred upward again.

This March, the nine-judge bench, led by Chief Justice Surya Kant finally began hearing arguments. The verdict has since been reserved.

A Socialist-Era Word in a 21st-Century Economy

The arguments before the bench have been revealing. Attorney General R. Venkataramani, appearing for the Union of India, conceded that the Triple Test was ‘logically correct’ in the abstract.

His concern was with what he called its ‘indiscriminate application’ - the way it had expanded the definition far beyond what labour law was originally designed to cover. A modern constitutional state, he argued, performs welfare activities that are neither commercial nor sovereign in the traditional sense. Treating all of them as ‘industries’ creates distortions.

Justice B.V. Nagarathna put it plainly from the bench: the 1978 judgment, she observed, was a product of the 1970s socialist era.

That observation carries weight. India’s economy in 1978 was heavily state-dominated. The boundaries between the government as employer and the government as sovereign were blurry. Today, with a gig economy, with technology platforms, with blended public-private entities delivering everything from healthcare to infrastructure, the question of what counts as an “industry” has become far more complex and far more consequential.

Leave a comment

Why This Matters Right Now

India is currently running over 14 Production-Linked Incentive (PLI) schemes across sectors, committing upwards of ₹2 lakh crore to push manufacturing. The government talks about industrial corridors, smart cities, and becoming a global factory. The Prime Minister has framed industrialisation as a civilisational project.

And yet, the legal term on which all industrial labour relations rest, the word industry itself remains judicially unsettled.

The Industrial Relations Code, 2020, one of the four new labour codes India passed , does contain a new definition of “industry,” and came into effect in November 2025. But the nine-judge bench has made clear it is not adjudicating that code’s definition. Its task is to decide whether the 1978 ruling was correct as an interpretation of the original IDA.

That distinction matters because the labour codes, despite being enacted, remain largely unimplemented across most states. Also the judgment is not directly about the new code, but it will inevitably shape how the new code is understood, challenged, and applied.

Here is a fact that tends to get overlooked in conversations about India’s ambition to become a global technology and manufacturing power: the country that is most often cited as the model to beat, China, spent roughly three decades importing knowledge before it started exporting it.

That isn’t a metaphor. It is a description of deliberate policy.

Subscribe now

The playbook China ran

Between 2010 and 2019, China nearly doubled the number of international students studying within its borders rising from 265,090 in 2010 to 492,185 by 2018. The table below traces that arc.

Table 1 - China’s international student intake, 2010–2023.

At the same time, China was running an elaborate parallel programme, sponsoring Chinese nationals to study STEM abroad, and then building incentive structures to pull them back. In the four decades since reform and opening up, Beijing sent millions of graduate students and research scholars overseas to pursue education in STEM fields, while building a sprawling ecosystem of programmes designed to ensure the scientific know-how they acquired abroad was absorbed to advance China’s commercial and defence sectors.

The results compounded. By 2023, 57.9% of newly appointed academics in the Chinese Academy of Sciences were overseas returnees, and individuals with overseas degrees headed over 70% of key national projects and accounted for about 40% of national science and technology award winners. By 2015, foreign firms had established more than 2,400 R&D laboratories in China, and universities and enterprises were jointly running over 88,000 R&D projects worth CNY 66.65 billion.

Now read the more striking number.

The number of American students in China fell from about 20,000 in 2018 to just 700 in 2023.

The West didn’t leave because China failed. It left because China no longer needed the same kind of hand-holding. The country had absorbed enough capability to stand on its own, and geopolitics did the rest.

This is the arc India should be reading very carefully.

Where India actually stands

India has the third-largest higher education system in the world by raw student count - over 43 million enrolled. But the number that matters is the Gross Enrolment Ratio, or GER - the share of the eligible college-age population that actually attends higher education.

Table 2 - GER in higher education, selected comparators.

The numbers are stark. India’s GER stood at 19.4% in 2010 and had risen only to 28.4% by 2021–22. The NEP 2020 target is 50% by 2035. For comparison, the US GER sits at 88.2% and China’s at 60.8% as of 2024. In other words, India’s enrolment rate today is roughly where China was a decade ago and China has since blown past the very target India is still trying to reach.

That gap has a very direct commercial meaning. Tens of millions of aspirational students cannot access the quality of education they want domestically. The demand is real and unmet.

This is simultaneously a policy failure and an extraordinary commercial opportunity.

Share

What India has started and where momentum has slowed

The government recognised the problem and took a meaningful step in 2023. The UGC’s Foreign Higher Education Institutions (FHEI) Regulations finally gave overseas universities something they had always wanted but never had in India: full autonomy over admissions, fees, and faculty recruitment, plus the right to repatriate surplus funds to the home institution. GIFT City in Gujarat, already operating as a financial special economic zone, extended its regulatory light-touch to education, offering 100% tax exemption, no exchange control, and the flexibility of an education SEZ.

Table 3 - Foreign university campuses in India, operational and confirmed (as of April 2026).

But here’s the thing: these numbers, impressive as they sound, are thin. Thirteen of the 17 universities currently committed are concentrated in just four metros. From China’s own experience, the most successful version of this model involved branch campuses - NYU Shanghai, Nottingham Ningbo, XJTU-Liverpool, that became embedded knowledge nodes, not just diploma mills. India has nothing of that scale or depth yet.

The regulatory framework exists. The commercial case is overwhelming. The momentum has, for now, stalled at the proof-of-concept stage.

Why this is bigger than just seats

The argument for inviting foreign universities into India is not primarily about enrolment numbers. It is about knowledge infrastructure.

Research on academic and ethnic networks shows that when countries have a strong presence of foreign-trained researchers, technologies diffuse faster into the domestic economy, even technologies developed by others, because the knowledge is already physically present in people, labs, and relationships. China ran this at scale. India’s IITs and IISc do world-class research, but they remain islands. The connective tissue - the joint labs, the embedded foreign faculty, the research collaboration that makes knowledge sticky, is largely absent.

A broad literature on what nations need to trigger a reverse brain drain points to the same answer: when real opportunity exists within a context of coherent internal policies and investment in science and technology, returning to the home country becomes an attractive option for emigrants. India has the diaspora. It does not yet have the institutional pull.

Foreign universities in India are not a substitute for building great Indian universities. But they are, historically, one of the fastest ways to accelerate the process. China understood that proximity to excellence is not passive, it is how capability transfers.

The commercial case is self-evident, if anyone is paying attention

For international universities, the math is simple. A foreign degree earned in India costs 60–70% less than the equivalent programme at the home campus, without visa complexity or family separation. With Canada slashing international study permits, Australia tightening enrolment caps, and US visa unpredictability rising, Indian students who were once a captive export market are now a captive domestic market waiting to be served.

The question is whether India will move fast enough - on approvals, on infrastructure outside GIFT City, on faculty visa norms, on research collaboration frameworks to capture that.

China, by 2019, had built enough capability that the West began to see it as a competitor rather than a student. India is still in the phase where the openness is entirely in its interest. That window does not stay open forever. The pattern, as China showed, is that countries eventually close it themselves, once they no longer need it.

India is not yet at that juncture. Whether it uses this period well is a different question entirely.

Leave a comment

Let’s start with a question that sounds simple but has divided economists for decades: if you want people in poor countries to live longer, eat better, and send their kids to school - what actually works?

The instinctive answer is money. Richer countries have better hospitals, cleaner water, and more food. But the more pointed question is: how does a country get richer, and does the path to wealth matter as much as the wealth itself? Can targeted programs, government interventions designed to fix specific problems like malnutrition or child mortality substitute for broad economic growth? Or does everything else depend on growth first?

Subscribe now

A new paper by Lant Pritchett and Addison Lewis, both economists at the University of Oxford, has taken a firm position on this. And their answer has reignited a debate that sits at the heart of development economics - the branch of economics that studies how poor countries become less poor.

A Field and Its Fault Lines

Development economics isn’t just one school of thought. It’s a running argument between at least two broad camps, and understanding that argument is essential context for what Pritchett and Lewis are doing.

One camp - call it the growth school - holds that sustained economic expansion is the primary engine of human welfare. When GDP per capita rises, that is, when the average income of a country’s citizens increases, almost everything else follows: people eat better, live longer, and educate their children more. The policy implication is that governments should focus on the conditions that enable growth: property rights, trade, macroeconomic stability, investment.

The other camp - increasingly influential since the 1990s - argues that growth is necessary but not sufficient. Even in growing economies, specific deprivations persist: certain communities stay malnourished, certain diseases remain untreated, certain children never finish school. This school favours targeted interventions, precisely designed programs that address individual problems directly, without waiting for the economy to grow its way out of them.

Much of the momentum behind this second camp came from the spread of randomised controlled trials, or RCTs - a method borrowed from medicine where you test an intervention on one group and compare outcomes against a control group that didn’t receive it.

Organisations like J-PAL (the Abdul Latif Jameel Poverty Action Lab, based at MIT) became globally influential by running and aggregating these experiments. The 2019 Nobel Prize in Economics went to the economists who pioneered this approach: Abhijit Banerjee, Esther Duflo, and Michael Kremer.

In 2021, J-PAL’s directors published a piece making their position explicit: for millions of people living in poverty, growth is not enough. Targeted programs are equally important. It’s this claim that Pritchett and Lewis are directly contesting.

Share

What the Paper Actually Measures

The question Pritchett and Lewis ask is: how much of the variation in human wellbeing across countries can be explained by income alone?

To answer it, they build what they call a basics index - a composite measure of the physical conditions of human life. Not happiness or subjective wellbeing, but concrete, measurable things: child mortality, access to clean water, sanitation, electricity, nutrition, and schooling. The data comes from the Legatum Prosperity Index, which tracks 294 indicators across 167 countries.

They then test their findings four different ways - including generating 100,000 randomly constructed indexes, each with different indicators and weights chosen at random. The idea is to show that the result doesn’t depend on how exactly you build the measure.

Across every version, the finding is the same. GDP per capita - average income, adjusted for purchasing power across countries, explains roughly 81 to 85% of the difference in basics between countries. The relationship is also non-linear in a specific direction: the gains from growth are largest for the poorest countries.

The paper also notes two harder-edged findings. There is no wealthy country in the world with poor basics. And there is no poor country that has achieved wealthy-country levels of basics.

Growth does not guarantee good outcomes - Equatorial Guinea, which has significant oil revenues but weak human development, illustrates that clearly. But no country has escaped the constraint from the other direction either.

The Twist: Growth’s Critics Confirm the Same Thing

What makes the paper’s case particularly pointed is that it also tests measures built by organisations that explicitly set out to track development independently of income - specifically to push back against the idea that GDP tells the whole story.

The Social Progress Initiative’s Basic Human Needs index, the Oxford Poverty and Human Development Initiative’s Multidimensional Poverty Index, and the World Bank’s income poverty measure all yield the same result: GDP per capita explains over 80% of the variation in each of them.

The paper’s argument is that treating targeted programs as equally important to growth, at the level of aggregate outcomes, across countries, over decades is not what the data supports. It does not dismiss targeted interventions. It contests their relative weight.

Why This Debate Matters in India

India sits precisely in the part of the income curve where the paper says growth matters most. Over the past three decades, improvements in child mortality, electrification, and school enrolment have tracked closely with economic expansion. At the same time, India runs some of the world’s largest targeted programs MGNREGA, PM-KISAN, Ayushman Bharat, the National Food Security Act representing enormous fiscal commitments.

The question of whether these programs complement growth or substitute for it is not settled. And Pritchett and Lewis haven’t settled it either - their paper measures correlations across countries, not causal mechanisms within them. What drives the 15 to 20% of variation that income doesn’t explain is itself a rich research question.

What the paper does is sharpen the terms of the debate.

If growth explains 80% of the difference between countries on the basics of human life, what exactly is the claim being made when someone says growth is “not enough”? Not enough compared to what? For whom? On what timeline?

Those questions don’t have easy answers. But they’re probably the right ones to ask.

Leave a comment

Here’s a fact that might surprise you: the oil supply disruption unfolding right now in the Persian Gulf is, by one key measure, the largest in recorded history. Larger than the 1973 Arab embargo. Larger than the 1990 Gulf War. Larger than the 2003 Iraq War.

And yet, oil markets have not collapsed the global economy. At least not yet.

Subscribe now

That tension, between a record-breaking shock and a muted response is what Goldman Sachs’s Top of Mind Issue 147 tries to explain. And the answer tells you something important about how the world has quietly rewired itself over the last two decades.

A Shock Bigger Than All Before It

To understand how we measure oil shocks, you need to know the metric analysts use: the peak hit to supply, expressed as a percentage of total global oil production. In other words, how much oil did the world suddenly lose?

By that measure, the 2026 Iran War is historic. At roughly 16–17% of global supply, the disruption to Persian Gulf oil flows, caused by Iran’s effective blockade of the Strait of Hormuz has exceeded every previous episode.

This one blows past all of them.

The reason is geography. The Strait of Hormuz is the world’s most critical oil chokepoint, a narrow corridor through which roughly 20% of globally traded oil flows every single day. When that corridor gets disrupted, the numbers move fast.

So Why Aren’t Prices Screaming?

Here’s the thing: if you looked only at the supply shock, you’d expect oil prices to be in the stratosphere. But the price response, while significant, has been far smaller than what followed the 1980 Iran-Iraq War or the 2003 Iraq War. Prices have risen, but they haven’t spiralled.

There are a few reasons for this.

First, the nature of this shock is different. This is primarily an export shock, a disruption to flows out of a region rather than a full-scale destruction of production capacity. When Kuwait’s fields were burning in 1990 or Iranian fields were taking missile fire in 1980, the physical infrastructure was damaged. Here, the oil is largely still in the ground; it’s the shipping corridor that’s blocked. Markets are pricing in the hope that a political resolution could restore flows relatively quickly.

Second, the global oil market in 2026 is structurally different from what it was in 1973 or 1980. Strategic petroleum reserves, emergency stockpiles held by the US, IEA members, China, and others are far larger and better coordinated. The US itself is now a major oil exporter, something that was unthinkable in the OPEC era. Spare production capacity in Saudi Arabia and the UAE gives OPEC+ a buffer it can deploy. None of that existed in the 1970s.

Third and this is the deeper story, the world simply doesn’t need as much oil to produce the same amount of economic output as it once did.

Using Less to Make More

Since 2000, every major economy has been getting more efficient at using energy. This concept, how much energy it takes to produce a unit of GDP, is called energy intensity, and its decline is one of the most underreported economic shifts of the last generation.

The numbers are striking. Between 2000 and 2022, the US reduced its energy intensity by roughly 30%. The European Union did even better, cutting it by nearly 35%. East Asia and the Pacific, which industrialised rapidly during this period, still managed meaningful improvements. The world average fell by close to 25%.

India, often assumed to be a laggard on efficiency given its rapid industrialisation, has in fact reduced its energy intensity by 36% between 2000 and 2023 and the pace has been accelerating, with the rate of improvement nearly tripling from 0.61% per year in the 1990s to 1.58% per year in the 2010s.

In plain terms: the global economy today produces roughly the same unit of output while burning a quarter less energy than it would have in 2000.

This change didn’t happen accidentally. It reflects a combination of industrial efficiency gains, the shift of economies toward services, the proliferation of more fuel-efficient vehicles and appliances, and increasingly, the substitution of renewables for fossil fuels in electricity generation. Every kilowatt-hour generated by a solar panel is one that doesn’t need an oil-fired turbine.

The implication is significant. When oil prices spike, the damage they cause to economic growth is now smaller in absolute terms, because oil is simply a smaller share of production costs across most industries.

Share

What This Means for Your Economy

Goldman Sachs estimates that a sustained 10% rise in oil prices now lowers global GDP by approximately 0.1% and raises global headline inflation, that is, the consumer price index including food and energy by roughly 0.2%. These are not trivial numbers, but they are considerably smaller than equivalent estimates from the 1970s and 1980s.

The distribution of pain is uneven, though. Canada and Latin America actually benefit from higher oil prices, they’re net exporters. China and Japan, which import heavily, see modest GDP hits. The region that takes the biggest blow is Central and Eastern Europe, which remains deeply exposed to energy price swings and sees headline inflation rise by nearly 0.45% for every 10% jump in oil.

Given that pre-conflict forecasts pegged global GDP growth at 2.9% for 2026, even a modest sustained shock has real implications for where we end up by year-end. India’s inflation calculus is also complicated by the fact that domestic fuel prices are periodically administered, meaning that global price spikes don’t always pass through immediately, but when they do, the hit to household budgets can be sudden.

The Question That Remains Open

So here is where we are: the largest oil supply shock in recorded history has struck a world that is measurably more resilient to such shocks than it has ever been. The damage is real but, so far, contained. Markets are watching the Strait of Hormuz with one eye and the diplomatic back-channels with the other.

The more interesting long-run question isn’t whether this shock derails 2026 growth , it probably won’t, unless the disruption becomes protracted. The more interesting question is what this episode accelerates. Energy security concerns have historically been among the most powerful drivers of the clean energy transition. Every oil shock reminds governments, businesses, and households that dependence on a single chokepoint carries real risks.

The world got more energy-efficient partly by choice, and partly because high prices kept forcing adaptation. If prices stay elevated long enough, the next chapter of that adaptation may arrive faster than anyone expected.

Leave a comment

Here’s a fact that might surprise you: India has the world’s largest road network - over 6.6 million kilometres of it. The government has poured money into building highways at a pace few countries have matched, scaling public infrastructure capital outlay from ₹2.5 lakh crore in FY14 to ₹12.2 lakh crore in the FY27 budget estimate. That’s a jump from 1.7% of GDP to 3.3%.

But here’s the thing: building roads is only half the problem. The other half is figuring out how to pay for the next batch without drowning in debt. And that’s where a financial instrument most people have never heard of - the Infrastructure Investment Trust, or InvIT is helping India finance its highways.

Share

What’s an InvIT, really?

Think of an InvIT as a mutual fund, but instead of holding stocks or bonds, it holds toll roads, highway stretches, and other income-generating infrastructure assets. A sponsor, say, NHAI or a private road developer transfers operational road assets into a trust. The trust issues units to investors. Toll revenues and annuity payments flow through the trust to unit holders as regular distributions.

The sponsor gets upfront capital. Investors get a steady yield. And the road stays publicly owned, this isn’t privatisation. Ownership doesn’t change hands. Only the right to operate and collect revenue transfers for a fixed concession period, typically 15 to 30 years

Now, this matters because InvITs solve a problem that India’s older monetisation tools couldn’t. Before InvITs, the government’s primary instrument for recycling capital from existing roads was the Toll-Operate-Transfer (TOT) model - essentially a one-shot auction where a private bidder pays a lump sum upfront for the right to toll a highway bundle. TOT worked. NHAI awarded 10 bundles covering 2,689 km and raised about ₹49,000 crore between FY19 and FY25.

But TOT is a one-time transaction. Once you sell the concession, the capital recycling stops. There’s no mechanism for adding more assets, no platform for ongoing portfolio management, no secondary market for investors to trade in and out. InvITs, by contrast, are platforms. Assets can be injected over time. Units can be listed and traded. New investors can enter through follow-on offerings. It’s the difference between selling your house and renting out rooms while keeping the deed.

From a trickle to a flood

India’s first InvIT, the IRB InvIT Fund listed in 2017, raising about ₹4,745 crore. Early investors included the Government of Singapore and Deutsche Global Infrastructure Fund. It was an experiment. Could toll road cash flows really attract institutional capital through a trust structure?

The answer, evidently, was yes. By FY25, 26 InvITs had registered with SEBI, of which 15 were road-focused. Cumulative funds mobilised surged from ₹0.11 lakh crore in FY20 to ₹1.38 lakh crore in FY25 - a compound annual growth rate of roughly 65%. Road InvIT assets under management alone hit ₹2.46 lakh crore, growing at a 42% CAGR over the same period. Roads now account for 39% of total InvIT AUM, up from 18.5% in FY21.

The government’s own vehicle, the National Highways Infra Trust (NHIT), set up by NHAI in 2020 completed four monetisation rounds by FY25, transferring 26 highway projects spanning 2,345 km and raising over ₹44,000 crore. Its investor roster reads like a who’s who of global patient capital: the Canada Pension Plan Investment Board, Ontario Teachers’ Pension Plan, and Singapore’s GIC, among others.

See, this is the part that often gets missed in policy discussions about infrastructure financing.

The constraint isn’t just how much money India can spend on roads. It’s how fast capital locked in existing highways can be freed up and recycled into new ones.

NHAI facilitated ₹72,000 crore of debt repayment between FY23 and FY25, backed by ₹1.4 lakh crore monetised since 2020. That’s not new borrowing. That’s the same capital going around twice.

The penetration gap

And yet, for all this growth, the penetration remains startlingly low. As of FY25, only about 15,700 km of tollable, revenue-generating road assets sat inside InvIT structures, out of roughly 325,000 km of national and state highways. That’s a penetration rate of 4.8%.

The government clearly sees the runway. Under NMP 2.0, launched in February 2026, the highway monetisation target is ₹4.14 lakh crore over FY26–30, with ₹3.35 lakh crore planned through InvITs and TOT alone, covering an additional 19,200 km. Crisil Intelligence projects road InvIT AUM could reach ₹5.45 lakh crore by FY30, more than doubling from today’s levels.

The investor mix is also evolving. Sovereign wealth funds, pension funds, and insurance companies have anchored the institutional side.

NHAI’s upcoming Raajmarg Infra Investment Trust (RIIT) - the first government-backed InvIT open to public and retail subscription, rated AAA by CARE could change that equation. It holds five toll projects across Tamil Nadu, Karnataka, Andhra Pradesh, and Jharkhand, with an average tolling history of over 15 years.

Subscribe now

Beyond the balance sheet

What makes InvITs genuinely interesting isn’t just the financial engineering. It’s what happens after the assets enter the trust. Because InvITs are regulated platforms with mandatory disclosures, professional investment managers, and distribution obligations, they impose a discipline on road asset management that standalone concessions rarely do.

The FICCI-Crisil report highlights how leading InvITs are adopting lifecycle costing, evaluating total maintenance expenditure over an asset’s full service life rather than minimising year-to-year costs. NHAI’s Drone Analytics Monitoring System (DAMS), operational since October 2024, uses AI-enabled drone surveys across 53,000+ km in 24 states to detect pavement distress before it escalates.

The US Federal Highway Administration estimates that every dollar spent on preventive maintenance saves $6 to $10 in avoided rehabilitation costs. For InvITs holding assets over 20-year horizons, that arithmetic is transformative.

The deeper question, though, is whether India can convert this financial innovation into a genuine infrastructure flywheel, where monetisation proceeds fund new construction, which matures into new monetisable assets, which attract fresh capital, and so on. The plumbing is now in place. The capital is willing. Whether the pace of new road awards and project execution can keep feeding the machine is what will determine if road InvITs remain a clever financing trick or become the backbone of India’s infrastructure story for the next decade.

Leave a comment

Somewhere in a warehouse in Bhiwandi, a truck is being loaded with 40,000 units of a soap bar whose formula hasn’t meaningfully changed since the 1880s. By next week, those bars will sit on shelves in Nagpur, Patna, and Guwahati. Consumers will buy them on autopilot without once asking whether that bar is quietly connected to a forest being cleared in Borneo, or to a supply chain that India does not control.

This is not just a story about soap. It is about how the things we buy every day have become one of the largest and least discussed drivers of environmental pressure on the planet. And why fixing it may also quietly fix something else: India’s dependence on imports it can do without.

The Weight of What We Consume

Every smartphone, every packet of biscuits, every soap bar requires material - mined, grown, or refined from the earth. The sum total of this extraction is what economists call the global material footprint: a measure of how hard humanity is pressing on the planet’s resources to sustain its lifestyle.

The numbers are uncomfortable. According to the United Nations, the global material footprint rose from 43 billion metric tonnes in 1990 to 92 billion metric tonnes in 2017, and is projected to reach 190 billion metric tonnes by 2060. The planet’s material consumption is growing faster than both population and economic output. The UN Environment Programme confirmed in 2024 that global material use continues to grow by more than 23% per year.

The culprit is not any single industry. It is the entire architecture of modern consumption - fast-moving consumer goods sitting squarely at its centre.

Share

People Want to Do Better. They Just Don’t.

When you ask consumers whether they care about sustainability, they overwhelmingly say yes. A 2023 Deloitte survey found over 64% of Asia Pacific consumers believe climate change is a genuine emergency. A PwC study found consumers globally are willing to pay nearly 10% more for sustainably sourced goods. In India, a 2024 Ipsos report found 66% of consumers believe the environment is at risk and needs action.

And yet sustainable products remain a small fraction of what actually sells.

This gap has a name: the intent-action gap. It isn’t hypocrisy - it’s a structural problem. Sustainable options are often harder to find, poorly labelled, or undermined by greenwashing that erodes consumer trust. A BCG survey identified eight distinct barriers and the persistent fear that eco-friendly products simply don’t work as well.

Businesses have developed two broad responses.

1. The first is choice influencing: using marketing, labelling, and storytelling to make sustainable options more visible. Think certification logos, eye-level shelf placement, supply-chain transparency campaigns.

2. The second is choice editing: removing the unsustainable option altogether. India’s phasing out of single-use plastics is a familiar example: consumers didn’t choose to stop using plastic straws; the choice was made upstream.

Both approaches have limits. But the more interesting question is: what happens when a scientific innovation makes the sustainable option genuinely equivalent - in price, performance, and experience to the conventional one? Then neither influencing nor editing is necessary. The market can simply shift.

That is exactly what the soap bar case demonstrates.

The Brick Wall in Your Bathroom

Think of a soap bar as a brick wall. Some bricks hold the structure together; others do the cleaning. Right now, both are made from the same material: palm oil.

In chemistry, the structural component is called Total Fatty Matter (TFM) - specifically long-chain fatty acid salts like sodium stearate and sodium palmitate, which give the bar its rigidity, much like iron rods inside concrete. The active cleaning molecules - shorter-chain compounds like sodium laurate - dissolve in water, trap dirt in microscopic clusters called micelles, and generate lather.

The structural molecules? They get washed down the drain. They don’t clean your skin. They don’t create foam. They just hold the shape and then leave.

A recent research paper published in the Indian Public Policy Review makes a significant point: you don’t need palm-oil-derived molecules to hold a bar together. Alternate structurants - zeolite (a volcanic mineral used in water purifiers), kaolin clay (found in many face masks), and starch provide the same rigidity at a fraction of the environmental cost. Bars built this way clean equally well, lather equally richly, and are actually gentler on skin, since high TFM correlates with greater skin irritation in laboratory tests.

India’s Hidden Import Problem

Here is where this becomes about more than climate.

India imports a significant share of its palm oil - the country is one of the world’s largest palm oil importers, sourcing primarily from Indonesia and Malaysia, which together control 85% of global production. Every spike in global palm oil prices, every export restriction in Jakarta or Kuala Lumpur, every geopolitical disruption in the shipping lanes of the Malacca Strait translates directly into input cost shocks for India’s FMCG sector.

The soap bar is a small but telling example of a broader pattern: India using imported commodities to perform functions that domestic, scientifically viable alternatives could handle. Zeolite, starch, and kaolin clay are not exotic materials. India produces all three. Substituting them for the structuring fraction of palm oil in soap bars would reduce import dependence for that specific use - quietly, without any visible change to the consumer experience.

This is the logic of what might be called micro-substitution - finding the small, specific nodes in industrial supply chains where an imported input is performing a function that a domestic alternative can replicate.

Soap bars are one such node. Scaled across FMCG categories, the aggregate effect on India’s import bill and supply chain resilience is non-trivial.

Subscribe now

The Standard Standing in the Way

India’s soap quality standard written nearly 50 years ago grades soaps by TFM content: Grade 1 requires at least 76% TFM. Higher TFM, better soap. That is the signal the system sends and consumers have internalised it.

A survey found 80% of respondents actively avoid Grade 2 and Grade 3 soaps.

Except BIS itself acknowledged in the 1990s that TFM quantity doesn’t determine quality. The type of fatty matter matters more than the amount.

Meanwhile, the US, EU, Australia, and Canada prescribe no TFM minimum at all - regulating safety and performance outcomes instead, leaving manufacturers free to innovate.

The paper recommends BIS revisit IS 2888 and shift to a performance-based framework. In practice, a low-TFM soap bar could then legally be sold as Grade 1 - provided it cleans effectively and is safe to use.

This is a modest regulatory change. Its downstream effects would not be. Palm oil demand from soap-making alone currently stands at 10 million tonnes per year, expected to nearly double in a decade.

A reformulated Indian market - driven by updated standards could put a measurable dent in that figure, reduce deforestation pressure, and shrink a specific import dependency, all without asking consumers to sacrifice a single unit of lather.

The intent-action gap in sustainable consumption is real. But sometimes, the gap isn’t in the consumer. It’s in a standard nobody has updated since 1974.

You’ve probably been hearing a lot about India’s consumption boom lately. The 2025 Union Budget raised income tax exemption limits. GST was rationalised. The narrative writes itself: a government pivoting toward the common man, loosening the purse strings before a growth push.

But here’s the thing: Neelkanth Mishra, Chief Economist at Axis Bank and a member of the Prime Minister’s Economic Advisory Council, recently argued the opposite. None of it, he said, was a consumption stimulus.

That’s a striking claim. So let’s unpack it and then ask an even more uncomfortable question.

The GST Reform That Was Always Coming

To understand the GST rationalisation, you need a bit of history.

When India launched GST in 2017, states were nervous. They were being asked to give up their own tax powers and trust a new, untested system. To get them on board, the central government promised: if your revenues fall short for five years, we’ll make up the difference from a special fund. That fund was the compensation cess - an extra charge, on top of GST, levied on luxury goods like cars, tobacco, and aerated drinks.

The cess was always a temporary fix with an expiry date. By March 2025, the constitutional obligation to compensate states was winding down, and the cess was going away. Full stop.

Now, here’s the link to the rate cuts. The 28% GST slab - the highest bracket, which applied to luxury and sin goods had the compensation cess sitting on top of it. Once that cess disappeared, policymakers had a natural opening to look at the 28% slab itself and ask: are all these goods actually luxury items that deserve the highest rate? Cars for a large family, is that really a luxury? Food processing equipment? Several such goods got moved to lower slabs.

But notice what didn’t happen. The government didn’t spend new money. It didn’t borrow more. It didn’t create a new fiscal hole to fund consumer relief. The cess revenue had always been earmarked for states and when that obligation ended, the money simply stopped flowing through that channel.

The rate rationalisation was funded by the fiscal space that was already opening up. In

Mishra’s words: the government had the balance sheet, it used it - but it was moving money from one pocket to another, not creating new purchasing power in the economy.

The Income Tax Cut That Was Actually a Thank-You Note

The income tax slab revision is trickier, because this one was deliberate. The government raised the basic exemption limit, and salaried taxpayers in the ₹8–12 lakh range got meaningful relief. That sounds like stimulus.

Here’s the cleanest way to think about whether it actually was. A genuine stimulus means the government puts more money into people’s hands without taking it back from somewhere else. The total demand in the economy goes up. But look at what the budget actually did: the fiscal deficit - the gap between what the government spends and what it collects continued to shrink, from around 5.1% of GDP in FY24 to a target of 4.9% in FY25. Tax revenues as a share of GDP still rose.

In other words, whatever income tax relief went to one group of taxpayers was quietly offset by higher collections elsewhere - through GST, corporate tax, and other direct taxes and by keeping a lid on expenditure. The net injection into the economy was close to zero.

Think of it like this: a restaurant gives you a free dessert but quietly charges you ₹100 more for the main course. You feel like you got something. The bill is the same.

Mishra describes the income tax cut as a reward for formal, compliant taxpayers rather than a macroeconomic lever. And the data bears this out. India’s fiscal consolidation over the last four years has been one of the sharpest among major economies - the deficit has compressed from 9.2% of GDP in FY21 to under 5% today, even as the US, Europe, and China have all loosened their purse strings significantly. China’s central government debt-to-GDP, for instance, went from 40% to 90% in just six years.

India moved in the opposite direction. The consumption boost, if it exists at all, is not the government’s doing.

Share

But What If That’s the Point?

Now, here’s where it gets genuinely interesting.

A January 2026 working paper from the Barcelona School of Economics authored by economists at the Bank of England, Oxford, CREI, and the University of St. Gallen offers a framework that makes India’s pattern look less like neglect and more like strategy.

The paper studies what it calls an unbalanced industrial policy - a policy mix where a government aggressively subsidises its high-tech, tradable sectors (manufacturing, tech services, semiconductors) while simultaneously suppressing domestic consumption. Think of it as flooding the world with your goods while ensuring your own people don’t buy them all up. The result is trade surpluses, capital outflows, and over time technological dominance, because manufacturing and high-tech services are where innovation lives.

Countries that ran this playbook consistently (the paper points to East Asian economies, China most explicitly) saw manufacturing employment shares rise while the rest of the world deindustrialised. The paper finds a correlation of 0.52 between industrial policy intensity and manufacturing employment growth across 24 countries from 2002 to 2019. Countries on the receiving end, running persistent trade deficits saw total factor productivity (a measure of how efficiently an economy converts inputs into output) decline by roughly 0.1% per year for every 1% of GDP in sustained deficit.

The mechanism is elegant and a little unsettling. Surplus countries export goods and receive foreign financial assets in return. That capital flows into deficit countries, where it inflates non-tradable sectors - real estate, retail, construction. Innovation migrates away from manufacturing. Productivity slows. Meanwhile, the surplus country keeps building things, keeps innovating, and quietly pulls ahead.

India on the Surplus Side

India is not running large trade surpluses, its current account deficit narrowed to roughly 1% of GDP in FY24, but the direction of policy is pointing somewhere specific.

The Production Linked Incentive schemes across fourteen sectors. The semiconductor mission. Infrastructure investment in logistics, power, and ports. The falling cost of capital Mishra himself highlights - AAA NBFCs borrowing at 7.5% for fifteen years, power plants at 7.9% for twenty - all of this is capital being directed deliberately toward the tradable, investable, productive economy.

And fiscal consolidation keeping the deficit tight, not expanding household entitlements, not running a consumption-led budget is, in the BSE paper’s framework, itself a form of industrial policy. A government that refuses to expand its debt supply suppresses domestic consumption by limiting the liquid assets households can hold. It channels the economy’s savings toward investment rather than spending. It keeps the real exchange rate competitive, making exports cheaper.

In other words, what looks from the outside like a government that “forgot” to stimulate consumption might actually be a government that read the playbook.

The Open Question

There’s a catch, and it’s significant.

The BSE paper also warns that this strategy imposes real costs on households who can’t afford to spend, on workers whose wages don’t translate into consumption, on a domestic market that stays smaller than it could be. Household financial savings in India have already fallen to roughly 5.1% of GDP, a multi-decade low. At some point, suppressed consumption may become suppressed growth.

And there’s a harder geopolitical dimension. This approach works when you’re the surplus country exporting to a willing-deficit country. India’s bet is that it can run this playbook while global supply chains are restructuring away from China, that the window will stay open long enough to build the manufacturing base that justifies the squeeze.

Whether it does is the question Mishra didn’t quite answer. What we know is this: the reforms that looked like they were about your wallet probably weren’t. The real question is whether the thing they’re building instead is worth the wait.

You’ve probably heard that India is among the fastest-growing major economies in the world. Official figures show the economy growing at roughly 6–7% a year for most of the past two decades. That’s the number quoted in budget speeches, IMF reports, and newspaper headlines.

But in March 2026, a paper by economists Abhishek Anand, Josh Felman, and former Chief Economic Adviser Arvind Subramanian dropped a striking claim: India has been miscalculating its GDP for twenty years. Not by a small margin, but by as much as 1.5–2 percentage points a year. The Government of India’s current Chief Economic Adviser, V. Anantha Nageswaran, swiftly pushed back. The debate that followed is worth understanding - because how we measure growth determines how we respond to it.

What Even Is GDP, and How Do You Measure It?

Let’s start at the beginning. GDP is the total value of all goods and services produced in a country in a year. Think of it as a single number that tries to capture the size of an entire economy. If a factory makes steel, a restaurant serves food, and a software firm ships code - all of that goes into GDP.

Now, measuring GDP sounds simple in theory. In practice, it’s enormously hard - especially for a country like India, where a huge chunk of economic activity happens outside the formal system. The neighbourhood kirana store, the auto-rickshaw driver, the roadside welder - these are part of what economists call the informal sector, and they collectively account for roughly 44% of India’s economic output. No one files annual returns for them. No one sends their accounts to a government database. So statisticians have to estimate.

That’s where the disagreement begins.

The Subramanian Argument: The Numbers Don’t Match the Story

The Subramanian paper’s core argument is this: the methodology India adopted in January 2015 made two fundamental errors, and both pushed measured GDP higher than actual growth.

The first error was using the formal sector as a stand-in for the informal sector. Because direct data on small, unregistered businesses is hard to collect, India’s statisticians assumed that informal businesses were growing at roughly the same pace as large, registered companies.

This assumption held up reasonably well - until three major shocks hit.

• Demonetisation in 2016, when the government suddenly withdrew 86% of currency notes from circulation, hammered cash-dependent small businesses.

• The GST rollout in 2017, India’s nationwide goods and services tax, added compliance burdens that large firms absorbed but small ones struggled with.

• Then came Covid in 2020. Each shock hit the informal sector far harder than the formal sector.

But the methodology kept assuming they were in step. The result, the paper argues, was that informal sector output was significantly overstated.

The second error involved deflators and this one requires a small explanation. When economists measure “real” GDP, they strip out the effect of price rises, so that growth reflects actual increases in output, not just inflation. To do this, they divide nominal output (the raw rupee figures) by a price index - a number that tracks how much prices have changed. India largely used the Wholesale Price Index (WPI) for this task. The WPI heavily tracks commodity prices, particularly oil. When oil prices fall, the WPI falls and if you deflate nominal output by a falling index, the resulting “real” GDP looks bigger than it actually is.

The paper argues this happened systematically after 2011: WPI growth averaged 2.2 percentage points less than consumer price inflation over the period, causing real GDP to be mechanically inflated.

The evidence they cite is striking. Between 2005 and 2011, India’s GDP growth tracked closely with indicators like bank credit, exports, electricity consumption, and corporate sales. After 2011, GDP kept showing 6–7% growth, but all those other indicators told a different story. Real credit growth fell from 15.6% to 5.6% per year. Exports fell from 13.9% to 5.4%. Industrial production collapsed from 16.1% to 2.9%. The paper’s conclusion: the true growth rate between 2012 and 2023 was probably 4–4.4% a year, not the official 5.9%.

The Government’s Reply: The Model Has Holes Too

CEA Nageswaran and MoSPI Secretary Saurabh Garg responded directly, and their critique is methodological rather than political.

Their first point: the WPI is not as inappropriate as the paper suggests. International standards actually recommend using a Producer Price Index (PPI) - essentially a measure of the prices that producers charge for their output. India doesn’t have a full PPI yet, but the WPI is conceptually close to one. The CPI, which the paper recommends as a substitute, measures prices that consumers pay which is a different thing altogether. Sectors like steel, cement, IT services, and chemicals are produced largely for industrial use, not household consumption. Deflating them using a consumer price index, the response argues, introduces its own distortions.

Their second point targets the informal sector calculations. The survey data the paper uses to estimate informal sector performance covers only half the informal economy - construction is excluded entirely. The paper then applies the weak performance of the surveyed sectors to the entire informal economy. That’s a large leap. Housing services - a substantial chunk of informal sector output are also missing from the paper’s survey data.

Third, the response raises a structural point the paper doesn’t address: India’s economy changed fundamentally after 2015. Global Capability Centres (offshore offices of multinational companies, employing hundreds of thousands of Indian professionals), the rapid expansion of fintech and digital payments, and deliberate policy-driven formalisation of the economy, none of these show up neatly in electricity consumption or bank credit. Those indicators, the response argues, are less reliable proxies for a modernising economy than they once were.

Finally and this is the response’s sharpest point, the government completed a major GDP methodology revision in February 2026, developed through what even the Subramanian paper calls “commendable consultations.” That revision incorporated new survey data and updated methods. But the actual adjustment it produced was far more modest than the paper’s estimates imply. If the true overestimation were really 22% of GDP accumulated over a decade, the official revision would have looked dramatically different. It didn’t.

Share

Why Does Any of This Matter?

Here’s the thing: this isn’t just a statistics dispute between economists. If GDP growth was genuinely weaker than reported, then the policy responses to that weakness - interest rate decisions, fiscal stimulus, reform urgency were calibrated to a false picture. The Subramanian paper puts it plainly: if the numbers say the economy is strong when it is actually weak, businesses may over-invest, households may over-spend, and the central bank may keep money too tight.

On the other hand, if the methodology was broadly sound and the revisions were warranted but modest, then the alarm is overstated and the bigger story is simply that measuring a $4 trillion economy with a 44% informal sector is genuinely, unavoidably hard.

India may still be among the top ten fastest-growing major economies by almost any measure. But the distance between the official number and the real one, whatever it turns out to be is itself a kind of measurement. It tells us how much of our economy we can actually see.

Here’s a fact that might stop you mid-scroll: the global economy could grow strongly on paper this decade and most people could be worse off at the same time.

That’s not a political talking point. It’s a structural feature of how AI-driven growth may actually work according to Citrini Research’s latest piece The 2028 Global Intelligence crisis.

It brought forth an interesting concept - Ghost GDP. It refers to output - goods, services, code, analysis, decisions that gets counted in the national accounts but never circulates through households. No wages. No salaries. No spending. Output that a machine generates, profits that flow to shareholders, and a number in a government report that tells almost no one how their life actually changed.

Understanding Ghost GDP is not a fringe concern. It’s an interesting lens for making sense of the AI economy and for understanding why it will land very differently in Mumbai than in Manhattan.

Subscribe now

What GDP Actually Measures And What It Misses

GDP (Gross Domestic Product) is a measure of total economic output. Every good produced, every service rendered, counts toward it. The assumption baked into the metric since its invention in the 1930s is that output and income are two sides of the same coin. A company produces more, it hires more, it pays wages, workers spend those wages, and the whole thing loops. Output creates income creates demand.

AI disrupts that loop.

See, when a company replaces 500 analysts with a single AI system, output doesn’t fall - it might actually rise, because the AI works faster and doesn’t take holidays. But the wages of those 500 analysts disappear. The GDP number stays healthy. The households behind it don’t.

This is Ghost GDP: production that exists in the aggregate statistics but does not re-enter the consumer economy as purchasing power.

The Asymmetry Hiding in the Numbers

Here’s where it gets interesting and where it matters greatly for countries like India.

Not all economies are equally exposed to this dynamic. A 2025 World Economic Forum report and IMF research converge on a striking finding: around 60% of workers in advanced economies hold roles that face significant AI exposure. For emerging market economies, that number drops to roughly 40%. For low-income countries, it drops further still to 26%.

The intuitive explanation is economic structure. Advanced economies are dominated by services - finance, law, consulting, software, media which are precisely the white-collar sectors AI is most aggressively entering. Emerging and low-income economies still carry larger shares of employment in agriculture, manufacturing, construction, and trade. Jobs that, for now, require physical presence.

But here’s the thing: there’s a second, less obvious reason why poorer countries are structurally insulated from Ghost GDP and it’s one that economists have been slow to foreground.

The Labour Cost Shield

Deploying AI at scale is not cheap. A GPU cluster capable of replacing a meaningful portion of a firm’s white-collar workforce requires significant upfront capital, cloud infrastructure, and technical capacity to integrate and maintain. In the United States or Germany, that investment makes obvious financial sense: a capable AI agent costs a fraction of what a $180,000 knowledge worker does.

In a low-income or emerging economy, the same calculation doesn’t hold.

When the labour that AI would replace costs a tenth as much, or a twentieth, the return on AI deployment collapses. The capital cost of automation remains relatively flat globally (GPU prices don’t care where you are), but the labour cost it competes against varies enormously. In countries where a skilled analyst earns ₹6–8 lakh a year, replacing them with AI infrastructure that costs several times more is not a rational business decision. At least not yet.

This means emerging economies are, paradoxically, protected from Ghost GDP by the very income inequality that we might otherwise be trying to solve. Low wages become a buffer. The machine cannot price out what is already cheap.

The IMF data captures this starkly: India’s AI exposure sits at 26% or 40% (depends on the defination you want to agree with) for key occupational categories versus 60% for the advanced economy average. The gap is not just about the jobs people do - it’s about whether replacing those jobs with AI makes economic sense in the first place.

Share

India’s Double-Edged Position

India, though, sits in an uncomfortable middle. Its IT services sector - TCS, Infosys, Wipro and hundreds of smaller firms - exports over $200 billion in technology services annually. That sector is not low-wage by Indian standards. And crucially, it competes in markets where the labour cost comparison does favour AI. An Indian developer billing at $40 an hour was competitive against an American one at $150. Against an AI agent at $2, the arithmetic is brutal.

The Ghost GDP risk for India is therefore not primarily domestic - it’s in the export services model that has anchored the current account for two decades. The rupee, the external balance, and the employment of millions of educated graduates are all downstream of whether global firms keep outsourcing or start automating.

The broader emerging economy workforce - the construction worker, the trader, the field technician is insulated. The IT engineer earning in dollars is not. India’s AI asymmetry is internal, not just external.

The Question Worth Sitting With

Ghost GDP may ultimately resolve as a transitional problem. New jobs may emerge. Policy may intervene. Technology may unlock gains that eventually reach households through different channels.

But transitions have durations. And during the transition, the countries least likely to generate Ghost GDP are not the most technologically sophisticated - they are the ones where labour is still cheap enough to remain competitive against capital.

That’s an uncomfortable kind of advantage to hold. It raises a question that no one quite wants to ask out loud: is low income, for now, a form of economic protection?