India's Middle Class Hits a Breaking Point

ELI5/TLDR

India’s middle class is much smaller than anyone admits — only 40 million income tax-paying households — and it’s getting squeezed from three sides at once. White-collar jobs have stopped being created, social media is selling everyone a billionaire lifestyle they can’t afford, and a slick digital banking system has made it absurdly easy to borrow money to fake that lifestyle. Saurabh Mukherjea and Nandita Rajhans argue in their new book Breakpoint that this is the worst moment for the middle class since 1991, and that the era of the “get a good job” mindset is basically over.

The Full Story

The middle class is a rounding error

Most reports on India talk about the middle class as a rising tidal wave of hundreds of millions. Mukherjea says that’s mostly wishful accounting. If you actually look at the income tax data, only 40 million households earn between 5 lakh and 1 crore rupees a year. That’s it. In a country of 1.5 billion people, the entire middle class fits into the population of a mid-sized European country.

Why draw the ceiling at 1 crore? Because India is absurdly unequal. Mumbai has 200-250 dollar billionaires — more than any city except three or four global money centers. If you want the middle class to actually capture the middle half of all income earned, you can’t stop at 30 lakh. You have to go all the way up to a crore.

“If you work for an investment bank and you’re asked how big the Indian middle class is, it’s almost like asking a barber, ‘Do I need a haircut?’”

The 40 million number matters because these households are force multipliers. Each one supports roughly five other jobs — the cook, the driver, the security guard, the bus conductor, the small-town courier. Out of the 350 million Indians working outside agriculture without a salary, roughly 200 million earn their income because the taxpayer spends.

Three forces colliding

The book’s central argument is that India has weathered crises before — the 2000 dot-com bust, the 2008 financial crisis — but this one is different because three independent forces are landing on the same small group at the same time.

Force one: white-collar job creation has flatlined. IT services used to absorb a million or 1.5 million graduates a year. That engine sputtered around 2016-17 and has been dead for the last three years. Meanwhile India now churns out 8 million graduates annually — 25% more than a decade ago. Supply keeps climbing, demand has stopped. Real wages at the Nifty 50 companies have actually fallen over the last ten years, after rising in the decade before. The Naukri.com jobs index in 2025 is flat against 2019.

Force two: social media as an aspiration factory. India has a unique combination that nowhere else has: mobile data costs 1/40th to 1/50th of what it costs in Europe or America, the median citizen was a teenager when all this started, and the society is one of the most unequal on earth. Put those together and what you get is hundreds of millions of young people with unlimited access to curated glimpses of a billionaire lifestyle they have no hope of touching. Mukherjea calls this “the industrialization of aspirations.”

Force three: the JAM ecosystem made borrowing frictionless. Jan Dhan (bank account), Aadhaar (unique ID), Mobile (phone). This trio was built to deliver subsidies and democratize banking. It worked beautifully. It worked so beautifully that now anyone can take out 30, 40, 70 simultaneous loans. The authors say they met a single borrower with 700 open loans. A system designed as a social safety net has become a conveyor belt for debt.

The hollowing of middle-skill work

Nandita walks through the academic spine of this. In 2003, David Autor (with Levy and Murnane) published the foundational paper showing that automation eats the middle of the skills distribution first — the supervisors, the clerks, the people whose job is to follow a rulebook. Daniel Susskind later picked this up in A World Without Work and showed it had played out across the Western economies.

When the authors went looking for the Indian version, they found it. An economist named Sudipa Sarkar had traced NSSO labor data from the 1980s to 2017 and shown the same hollowing happening here. It hit in three waves.

Wave one, roughly twenty years ago: factories automated. Indian labor costs are 3x Bangladesh’s, so anyone running a factory in India has strong reasons to swap workers for machines. Nandita describes visiting a furniture factory outside Pune doing 100 crores of annual revenue with three employees. One feeds plywood into the machine, one packs the output, one loads the trucks.

Wave two, the last ten years: the automation crept upward. Secretaries, clerks, back-office staff — all replaced by software. Nominal wage growth for the 5 lakh to 1 crore band over the last decade has been 0.4% CAGR. Essentially zero. Cost of living meanwhile doubles every ten years.

Wave three, right now: AI is coming for the cognitive jobs. Nandita admits, with the kind of honesty that makes an interviewer wince, that half her own workload is now handled by AI and she no longer hires interns.

“The kind of work that someone I would have hired as an intern would be doing, that’s what Claude is doing. So I do not need an intern today.”

The ATM room and the call center

One scene in the book illustrates all of this. Four years ago Mukherjea visited a company his firm had invested in. The company loaded cash into ATMs. In a dimly lit room in suburban Mumbai, about 100 operators sat in front of screens watching live feeds from roughly 30,000 ATMs. AI had been trained to flag odd behavior — someone trying to withdraw cash wearing a helmet, for instance. The operator would speak through the ATM’s speaker and tell the person to remove it. If they didn’t, the alarm went off and local police were dispatched.

Those 100 people were doing the work of 60,000 security guards. Each of those guards used to earn 4-5 lakh a year. As a shareholder, Mukherjea was thrilled. Two minutes later, he did the math on what happens to 60,000 families and felt sick. He relates that he also refused to believe, in January 2022, that call centers would soon be run by bots. Four years later, talking to a bot when you call Amazon is routine.

A socialist education system in a capitalist economy

The supply side of this mess is education. India’s premier engineering entrance exams shifted from written essays to multiple-choice around the turn of the millennium. That change rewarded rote memorization and punished the messy, creative thinking that AI cannot replicate. Twenty years later, the graduates produced by that system are exactly the wrong graduates for a labor market being reshaped by AI.

The numbers from the Azim Premji University’s 2026 State of Working India report are grim. Of every 100 graduates in their year of graduation, only three land a white-collar salaried job. Forty percent of graduates under 25 are unemployed. One in three graduates across all ages has no work. Graduate unemployment is nine times the rate for people with no formal education. Mukherjea points out this is worse than the 1980s — pre-liberalization India.

“We’ve got a socialist education system in a capitalist economy, and that exacerbates the problem.”

Even the IITs are cracking. Placement rates at IIT Bombay dropped from 78% to 75% in a year. The minimum salary offered fell from 6 lakh to 4 lakh.

Meanwhile there is actual demand that isn’t being met. India has roughly 150,000 VLSI chip designers. It needs three times that. A semiconductor designer with five years of experience in Bangalore or Hyderabad can earn a crore a year — about 30x the per capita income. The country’s colleges are not producing them. They’re producing generic IT coders whose jobs AI is eating.

The debt spiral

India used to be a savings culture. That’s over. Household non-mortgage debt as a share of GDP is now higher than in the US or China. Forty percent of annual income goes to debt servicing. Sixty-seven percent of Indian families have an active personal loan. The authors estimate 5-10% of Indians are in an outright debt trap — borrowing to service old debt, falling further each cycle.

What are people borrowing for? Not houses. Vacations. Five-star resorts. Maldives trips. The imitation game that social media started. When your real wages are flat and your cost of living is doubling every decade, and your Instagram feed is telling you the good life involves Bali, the math doesn’t work. You fill the gap with credit.

Why Saurabh is (cautiously) optimistic

At the policy level, Mukherjea argues for patience. He lived through the mirror image of this story as a teenager. His family moved to suburban London in the ’90s, and he watched the back-office jobs of British insurance and banking migrate to Bangalore, Gurgaon, and Pune. He later participated in that arbitrage himself — he came to Mumbai and hired IIT and IIM graduates to do work that used to happen in London. At the time, no one in Britain wrote about it. No local politician mentioned it. Even in retrospect, it wasn’t clear what anyone could have done.

Between 2015 and 2025 he watched that same London suburb collapse. The high street is destroyed. The shopping mall has fallen apart. Crime has moved in. The area voted for Brexit. Mukherjea sees MAGA America as the same phenomenon. A society that couldn’t respond to disruption and reached for a populist answer.

“I’ve seen the town where I grew up in my teenage years destroyed by a similar disruptive economic phenomenon, and British policymakers didn’t do anything about it real-time.”

That memory is why he’s willing to extend the Indian government more benefit of the doubt than it perhaps deserves. He sees GST rationalization, trade deals, labor codes, tax cuts — all pointing in the right direction, even if slow.

The deeper optimism is cultural. India is young (median age 28), high-tech, and has no real social safety net, which Mukherjea frames as a feature rather than a bug. It forces people to adapt. His closing argument is almost a reversal of conventional wisdom: a great first job out of university is now a curse, because it locks you into an office-worker mindset in a world that no longer needs office workers. The future belongs to people who stop expecting jobs and start creating them.

Key Takeaways

The Indian middle class is 40 million households, not hundreds of millions. Defined as income tax payers earning 5 lakh to 1 crore a year. Investment bank reports inflate this for obvious reasons.
The ceiling has to be 1 crore because of inequality. Mumbai alone has 200+ dollar billionaires. To capture 50% of national income you need to stretch the band upward until it includes people who don’t feel middle-class.
Each middle-class taxpayer sustains about 5 other jobs. Directly (domestic workers) and indirectly (security guards, pilots, bus conductors). 40 million taxpayers → ~200 million dependent livelihoods.
IT services stopped creating jobs around 2016-17. The engine that absorbed Indian graduates for 15 years is effectively dead. India now adds 8 million graduates a year into a flat job market.
Real wages at India’s 50 largest listed companies have fallen over the last decade. They rose from 2007-2016. The breakpoint is somewhere in the late teens. No one has cleanly explained why that specific year.
Autor-Levy-Murnane (2003) established that automation eats middle-skill supervisory jobs first. Sudipa Sarkar’s research using NSSO data shows India followed the same pattern despite being at a different development stage.
India’s labor cost is 3x Bangladesh’s. This makes it structurally cheaper for Indian manufacturers to buy machines than hire people — the opposite of what a developing country usually wants.
The JAM trinity (Jan Dhan, Aadhaar, Mobile) is a two-edged sword. Built for financial inclusion and subsidy delivery, it has become the plumbing for mass household indebtedness. Some individuals carry 700 simultaneous loans.
67% of Indian families have a personal loan. 40% of annual income goes to debt service. 5-10% of Indians are in an outright debt trap.
Only 3 out of 100 Indian graduates get a white-collar salaried job in their year of graduation. Only 7 get any kind of job. (Azim Premji University, State of Working India 2026.)
The MCQ format killed creative thinking. When entrance exams shifted from essay-based to multiple-choice around 2000, Indian education became optimized for memorization — exactly the skill AI commoditizes.
India has 150,000 chip designers and needs 500,000. A 5-year VLSI engineer in Bangalore earns 30x per capita income. The education system isn’t producing them.
Brexit and MAGA are templates for what happens if India doesn’t respond. Mukherjea watched his London suburb collapse after back-office jobs migrated. Populism filled the vacuum.
The “good first job” is now a curse. It trains an office-worker mindset in a world with fewer offices. The authors’ prescription: raise your kids to expect to be entrepreneurs, not employees.
“Plenty of work, few jobs” is the mental model. Demand exists — in semiconductors, in manufacturing, in trades — but it doesn’t arrive as neat salaried positions.

Claude’s Take

This is one of those rare conversations where a financial-markets guy accidentally wanders into sociology and actually has useful things to say. Mukherjea’s framework — three forces colliding that would each be survivable alone — is tidy but not glib. He backs it with real data: NASSCOM job growth trends, Naukri’s index, Nifty 50 wage bills, NSSO, PLFS, Azim Premji University’s latest report. When he doesn’t know something — like why the structural break is specifically the late 2010s — he says so. That kind of epistemic honesty is rare in Indian economic commentary, which typically oscillates between boosterism and doom.

The book’s real contribution is the definitional move: redefining the middle class upward to 1 crore per year because of how lopsided Indian income distribution is. It sounds unintuitive, almost offensive — how is a crore-a-year household “middle”? — but the logic holds if you insist (as they do) that the middle class should account for the middle half of income earned, not just the middle half of people. Once you grant that axiom, the rest follows.

Where I’d push back: the debt numbers need more care. The claim that 53% of personal loans are taken “below the age of 20” is almost certainly a transcription error or a slip — it should probably be below 30. And the framing that India is “the most indebted” ex-mortgages is technically true but somewhat misleading, because mortgages are where a rich-country middle class parks most of its debt. The comparison would be fairer if both sides were ex-everything-except-consumer-loans. Small point, but the book apparently does the heavy lifting here; the podcast just gestures at the numbers.

The MAGA/Brexit analogy is the most provocative thing in the episode, and it’s doing a lot of work. Mukherjea is essentially saying: this is what happens to a society that lets its middle class get eaten without responding. It’s a useful warning, though India’s political economy is different enough that the parallel probably breaks somewhere. Still, the directional point — populism is the price of policy inaction — is hard to argue with.

An 8 out of 10. Dense, data-driven, and refreshingly free of cheerleading. The book is probably worth reading for anyone interested in where India is actually going, rather than where the consulting-firm reports say it’s going.