Jobless Growth: When the Economy Booms and the Jobs Don't Come

Every quarter the headline arrives like weather: the economy grew, faster than expected, faster than most of the world. And every quarter a second conversation happens quietly in family WhatsApp groups and hostel canteens and the queue outside a walk-in interview, where a cousin with a degree has been looking for eleven months. That gap has a name. Jobless growth is what economists call it when a country’s output rises year after year while the number of decent jobs barely moves. India is the world’s most-watched case of it, and the argument over whether the label fits is worth having honestly, because the answer decides what we do next.

What jobless growth actually means

Strip away the jargon and the idea is simple. GDP is a measure of production — the total value of everything a country makes and sells in a year. Employment is a measure of how many people are needed to make it. For most of the industrial era those two lines moved together, roughly. More output meant more factories, more shifts, more hands. Jobless growth is the condition where the first line climbs and the second one flattens: the economy expands, but the expansion does not require many additional workers to happen.

It is not the same thing as a recession, and that is exactly what makes it hard to talk about. There is no crash to point at. The stock indices are fine. Corporate profits are healthy. The airports are being rebuilt. By every number that gets printed in bold on the front page, things are going well — which means anyone who says otherwise sounds like a pessimist arguing with arithmetic.

The more precise version of the complaint is not that no jobs are being created. Jobs are created constantly in a country adding millions of working-age people every year. The complaint is about composition and quality: whether the jobs being created are the kind that pay a wage you can build a life on, offer some security, and use the education people were told to acquire. By many estimates, a large share of India’s recent employment growth has been in self-employment, unpaid family work, and a return to agriculture — categories that count as “employed” in the statistics and feel like nothing of the sort in the household.

Why the growth stopped needing people

There is no single villain here, which is part of why the problem is so stubborn. Several things are happening at once, and each of them separately would be manageable.

Growth is concentrated where few people work. Look at what actually drives India’s headline expansion — finance, telecom, software services, refining, e-commerce logistics at the platform layer, capital-intensive infrastructure. These sectors produce enormous value per employee. A refinery worth thousands of crores may employ a few thousand people. A software firm can double revenue on a modest increase in headcount. When output growth is concentrated in high-productivity, capital-intensive sectors, the arithmetic follows: GDP rises steeply while the employment line hardly registers it.

Routine work is being automated. Every layer of predictable, rule-following work — the reconciliation, the data entry, the first-line support call, the basic test script — is being handed to software. This has been true for two decades and is accelerating now for reasons anyone reading about AI already knows. The pattern visible in India’s IT layoffs is the sharpest version: an industry that was the country’s great absorber of educated young people announcing record revenue and flat or falling headcount in the same breath. Whether the technology ultimately creates as much work as it destroys is a genuinely open question, and I have argued elsewhere about will AI create jobs in the terms it deserves. But the timing matters even if the optimists are right. A job destroyed this year and a job created in 2035 are not the same job, and they are not for the same person.

The manufacturing stage got skipped. This is the deepest structural fact and the least discussed at dinner tables. Every country that moved large numbers of people out of poverty in the last century did it the same way: pulled workers off farms into labour-intensive factories — textiles, footwear, assembly, toys — where a person with limited formal education could earn a wage far above what the land gave them. India, by most accounts, went from agriculture largely to services without building that middle. Manufacturing’s share of output has been widely reported as stubbornly flat for decades. The ladder that other countries climbed has a missing rung, and the people who were supposed to step on it are standing where they were.

People are leaving the farm faster than anything absorbs them. Agriculture still supports a very large share of India’s workforce while producing a much smaller share of its output — which is another way of saying it holds enormous hidden underemployment. As people leave, they need somewhere to go. Construction takes some, informally and precariously. Gig platforms take some. And when the cities offer nothing, some go back, which is why analysts point to periods where the share of workers in agriculture has actually risen — a reversal that should alarm anyone who understands what it implies.

A country can post excellent growth numbers for a decade and still leave most of its people exactly where they started, because GDP counts what is produced, not who ends up holding it.

The honest objection

I want to be fair to the other side, because the phrase “jobless growth” is used too loosely and sometimes lazily.

The government and a number of serious economists dispute the diagnosis outright. They point to survey data showing employment rising substantially in recent years and unemployment rates falling, and they argue that critics are reading old data, using the wrong survey, or defining “job” so narrowly that only a salaried office post counts. There is a real methodological argument here about how India measures employment, what counts as work in an economy where most work is informal, and whether household surveys or firm-level data tell the truer story. Anyone who tells you the evidence is unambiguous in either direction is selling something.

The critics reply that the same data showing more people employed also shows more of them in self-employment and unpaid family labour, with real wages growing slowly or not at all — that the count went up while the quality went down. They point to the collapse in campus placements, to graduates taking work far below their qualification, to the astonishing ratios of applicants to posts whenever a government job is advertised.

Notice that both sides can be right at once. Employment can rise while good employment does not. That is not a contradiction; it is the precise shape of the problem. And it is why the argument about the label matters less than the argument about wages, security, and whether a young person with a degree can find work that uses it.

What GDP was never designed to tell you

Here is the thing I keep coming back to, and it is not really an economic point but a moral one.

GDP is a production statistic. It was built to answer one question — how much is being made — and it answers that question well. It was never built to answer the questions people actually care about: who received it, whether the receiving was fair, whether life improved. A country can double its output while the gains flow overwhelmingly to owners of capital, and the number will look magnificent throughout. Nothing in the measurement will register the difference.

This is why the growth headline and the household experience can diverge so completely that each side thinks the other is lying. Both are describing something real. The economy really is producing more. Most people’s position really has not changed much — which is exactly what you feel when you read about the cost of living in India and recognise your own arithmetic in it. Rent, fees, and medical bills rise on a schedule your salary does not follow.

The question is never whether a technology or an economy produces value. It is who gets to keep it — and that is a question about power, not productivity.

Which brings me to what I think is the actual mechanism underneath all of this. Productivity gains do not distribute themselves. Whether a machine that does the work of ten people means nine unemployed workers or ten people working shorter weeks for the same pay is not determined by the machine. It is determined by who owns it, who has bargaining power, and what rules the state writes. This is the same dynamic I trace in detail in how technology gets captured: every genuine advance arrives as a surplus, and the surplus goes wherever the existing arrangement of power sends it. Jobless growth is not a mysterious ailment of modern economies. It is what capture looks like when you plot it as a time series.

Why the framing decides the policy

If you believe the problem is that Indians lack skills, you fund more skilling programmes and wait. If you believe the problem is that growth is structurally concentrated in sectors that do not need many workers, you do different things entirely — you make labour-intensive manufacturing viable, you take the informal sector’s conditions seriously instead of treating it as a waiting room, you strengthen the bargaining position of the people doing the work, and you stop treating a rising output number as evidence that the arrangement is working.

The demographic clock makes this urgent rather than academic. India’s young population is described as a dividend, but a dividend is only paid if the work exists. A large cohort of educated young people who cannot find work commensurate with their education is not a dividend. It is a bill coming due.

So when the next quarterly number arrives and the anchors call it a milestone, the number is probably accurate. Just ask the second question immediately after: grew for whom? That question is not cynicism. It is the only way to tell the difference between a country getting richer and a country where the rich are getting richer, because the headline figure, by design, cannot tell them apart.