Bigger Is Not the Same as Better

What "growth" actually means, whether we even need it, and where India's version is heading.

Everyone in that China thread was arguing about degrees. I want to argue about something underneath it, because I think the degree panic is a symptom, not the disease.

Here's the question that's been sitting in my head: when we say a country is "growing," what is actually growing? Because a lot of what we call growth is not something you build. It's something that happens to you while you do nothing. Bodies pile up. Prices drift upward. The number on the chart goes up. And we clap.

But there are two completely different things hiding inside that one word, and almost nobody separates them. Once you do, the whole China-vs-India conversation reorganizes itself.

1. There are two kinds of growth, and only one of them is real

Think of an economy as inputs × how well you use them.

You can make the output bigger two ways:

• Extensive growth, throw in more stuff. More workers, more concrete, more capital, more hours. The pie gets bigger because you added more ingredients. This is the automatic kind. A country with a rising population and a savings rate grows like this in its sleep.
• Intensive growth, get more out of the same stuff. Same worker, more output per hour. Same rupee of capital, more value. This is productivity, what economists call Total Factor Productivity (TFP). This is the hard kind. Nobody does it by accident.

graph TD
    A[Economic Output] --> B[EXTENSIVE GROWTH<br/>add more inputs]
    A --> C[INTENSIVE GROWTH<br/>use inputs better]
    B --> B1[More people]
    B --> B2[More capital / concrete]
    B --> B3[More working hours]
    C --> C1[Higher productivity per worker]
    C --> C2[Better technology / TFP]
    C --> C3[Human capital: skill, not just headcount]
    B1 --> D{Diminishing<br/>returns}
    B2 --> D
    B3 --> D
    D --> E[Hits a ceiling.<br/>This is the middle-income trap.]
    C1 --> F[Compounds forever.<br/>This is how you actually get rich.]
    C2 --> F
    C3 --> F
                    

This is basically the Solow insight, stripped of the math: you can grow for a while just by accumulating, more people, more machines, but accumulation has diminishing returns. The tenth factory in a city adds less than the first. Past a point, the only thing that keeps living standards rising is productivity, squeezing more out of each unit. And productivity comes from exactly two places: technology and human capital. Machines that do more, and people who can do more.

So when someone asks me "aren't we all growing anyway, just by existing?", yeah. That's the trap. You are growing. Extensively. Automatically. And that specific kind of growth is the kind that runs out.

2. The automatic growth: India is getting bigger while doing nothing

Here's the part people find counterintuitive.

India's fertility rate has already fallen below replacement, around 1.9 to 2.0 children per woman, under the 2.1 you need just to hold steady (SRS 2024; NFHS-6). By the logic of biology, we should be flattening out.

We're not. India still adds roughly 12.6 million people a year, about a Bhopal every three weeks, and the population won't even peak until around 2062, somewhere near 1.7 billion (UN projections, 2026). How? Demographic momentum. So many people are already in their childbearing years that even below-replacement fertility keeps the total climbing for another generation. It's a train that keeps rolling after you cut the engine.

graph LR
    A[Fertility below<br/>replacement<br/>~1.9 per woman] -->|but| B[Huge cohort already<br/>of childbearing age]
    B --> C[Population still<br/>grows +12.6M/yr]
    C --> D[Peaks ~2062<br/>~1.7 billion]
    D --> E[Then declines]
    style C fill:#2d3748,color:#fff
    style A fill:#4a5568,color:#fff
                    

This is growth by doing nothing. The headcount rises on pure inertia. And the annual growth rate tells the honest story, it's collapsed from a peak of 2.41% in 1961 to about 0.86% in 2026. The number keeps getting bigger, but the engine behind it is already sputtering.

The "demographic dividend" everyone celebrates is exactly this: right now about 66% of Indians are working-age (15 to 59), nearly two-thirds are under 35, and only ~10% are 60+ (SRS). On paper, this is a once-in-a-civilization tailwind.

Except a dividend is a payout on an investment. And a bulge of working-age people is only a dividend if you invest in them. If you don't, it's not a dividend. It's just a very large number of people who need jobs.

3. The bill has a due date, and it's closer than people think

Here's what the demographic-dividend cheerleaders skip.

The window is finite and already closing. The working-age share of the population begins declining after 2030 (State of Working India 2026), and the absolute working-age count is projected to peak around 2041 at roughly 1.01 billion, then fall. The dividend phase effectively shuts by 2055 (UNFPA, Economic Survey). After that, the dependency ratio flips and the country starts aging, fast.

And it's not uniform. This is one of the correlations that actually matters:

• The South is already aging. Kerala, Tamil Nadu, fertility well below replacement, median age past 33, 60+ share heading toward 23 to 25% by 2036.
• The North is still young. Bihar's median age is around 22, fertility near 2.9.

So India isn't one demographic story. It's a young Bihar subsidizing an old Kerala, inside one border, with all the political friction that implies. The elderly share nationally hits ~20% by 2050 (LASI). We are, in the words of one ORF paper I found brutally accurate, at risk of "growing old before we grow rich."

And the leading edge of the collapse is already visible in a number almost nobody quotes: school enrolment in India fell by 13.4 million between 2019 and 2025. The pipeline is thinning now.

timeline
    title India's Dividend Clock
    2026 : ~66% working-age : Fertility already below replacement : Window wide open
    2030 : Working-age SHARE starts declining : (SWI 2026)
    2041 : Working-age COUNT peaks (~1.01B) : Last of the youth bulge enters workforce
    2050 : Elderly hit ~20% : "Old before rich" risk
    2055 : Dividend phase ends : Dependency ratio flips
    2062 : Population peaks (~1.7B) : Then decline begins
                    

Roughly 15 to 20 years. That's the whole runway to convert a billion-person headcount into something durable. That's the actual clock China is racing too, their classrooms literally start emptying by 2040, which is why they're moving with the violence they are. They're not being visionary for fun. They're beating a demographic countdown.

4. The GDP illusion: 7.6% growth that isn't compounding where it needs to

Now the uncomfortable part, and the reason I don't trust the "India is the fastest-growing major economy" headline as much as I'm supposed to.

India is growing fast on paper, real GDP grew about 7.6% in FY26 (MOSPI). Fastest in the G20. Genuinely good. But look at what's driving it and what it's built on:

• TFP, the productivity engine, is decelerating. RBI data has India's TFP growth averaging 2.1% a year across 2000 to 2019, but slowing to 1.4% in 2015 to 2019. The intensive engine is losing power even as the extensive one (capital + bodies) revs.
• Manufacturing is stuck at ~14% of GDP. The "Make in India" target was 25%. Missed, badly. We skipped the factory stage that lifted every East Asian economy.
• ~45% of the workforce is still in agriculture, which produces only ~16% of output. That gap is the productivity problem in one line: nearly half the country is trapped in the least productive sector.
• R&D spend is ~0.65% of GDP, and declining. China is around 2.4%. You cannot buy a productivity future at 0.65%.
• Per capita just crossed $3,000. The upper-middle-income ceiling is ~$13,845 (World Bank). We are not close.

The World Bank's own country memorandum says it plainly: India's growth over the last two decades came from TFP and capital accumulation, but the labour contribution has been limited, low participation, especially women, and within-firm productivity for small and medium firms stayed flat from 1995 to 2018. The big firms got more productive by absorbing resources. The small ones just didn't move.

This is the profile of extensive growth wearing an intensive costume. We're growing because we're adding people and capital and formalizing the economy (UPI, Aadhaar helped, real gains there), not primarily because the average Indian worker is getting dramatically more productive each year.

And here's the graveyard nobody wants to visit. The middle-income trap is a real, well-populated place:

graph TD
    A[Country hits<br/>middle income] --> B{Does it shift from<br/>extensive to intensive<br/>growth?}
    B -->|YES: raise TFP,<br/>build human capital| C[ESCAPED<br/>South Korea, Taiwan,<br/>Japan, Singapore]
    B -->|NO: keep adding<br/>bodies + capital| D[TRAPPED FOR DECADES<br/>Brazil, Mexico, Malaysia,<br/>Turkiye, South Africa]
    D --> E[Stuck at middle income<br/>30-40 years, still counting]
    C --> F[Rich, high-productivity<br/>economy]
    style C fill:#22543d,color:#fff
    style D fill:#742a2a,color:#fff
                    

Brazil has been stuck at upper-middle income since roughly 1989. Mexico since 1990. Malaysia since 1992. South Africa since 1988. These aren't poor, lazy countries, they grew, extensively, and then hit the ceiling and stopped. The thing that separated the escapees (the East Asian tigers) from the trapped wasn't population or resources. It was the ability to convert a young workforce into a productive one through education and technology before the demographic window closed. That's the correlation that runs through the whole 20th century.

India is standing at exactly that fork right now, with the clock from Section 3 ticking.

5. The conversion machine is broken, on both ends

The dividend only pays out if two conversions happen: bodies to capability, and capability to jobs. India is failing at both, and this is the part people usually get half-right.

5a. The supply side: an education system that trains for the wrong economy

The core problem is a philosophy. Indian education, from primary school to university, is built like an assembly line for memory, not capability. Success is defined as your ability to store, recall, and reproduce a fixed body of data under exam conditions. That is the entire game.

Three things fall out of that:

• The curriculum update loop is 5 to 10 years. Syllabi crawl through slow academic councils. By the time a new one ships, the tools it teaches have already been outmoded. You're training students on last decade's software for next decade's jobs.
• Departments are hyper-siloed. Theory is walled off from practice, software from hardware, engineering from design. The real world is multidisciplinary; the degree is not.
• The output is certificates, not competence. National employability sits at just 56.35% (India Skills Report 2026), meaning ~44% of graduates aren't job-ready by the report's own test. And the vocational track that's supposed to feed manufacturing is the weakest: ITI at 46%, polytechnic at 33%. The good numbers cluster at the very top, CS/IT engineers around 80%, a thin spike of world-class talent sitting on a base the system barely equips.

The global backdrop makes the mismatch brutal. The WEF Future of Jobs 2025 projects that by 2030, 92 million jobs will be displaced and 170 million created (net +78M), 22% of all jobs structurally transformed, and ~39% of current skills obsolete within five years. 63% of employers already name the skills gap as their single biggest barrier. The degrees we mass-produce are training people for the 92 million that are leaving, not the 170 million arriving.

5b. The demand side: even the job-ready can't find jobs

Here's the correction to my own argument, and it matters. Employability is a supply-side number, it tells you if you're ready. It says nothing about whether a job actually exists. And on that second question, the data is worse.

State of Working India 2026 (Azim Premji University): graduate unemployment is nearly 40% for 15 to 25-year-olds and ~20% for 25 to 29-year-olds. Of young male graduates, fewer than 7% land a permanent salaried job within a year of graduating; only 3.7% get white-collar work. India's youth (15 to 29) number 367 million, a third of the working-age population, and 263 million of them are already out of education, i.e., the potential workforce right now.

This is jobless growth: the economy expands, but it produces fewer jobs per unit of growth than it used to. So graduates end up as delivery riders, telecallers, gig workers, or they migrate, youth make up ~40% of informal migrant workers, Bihar and UP shipping people to Delhi, Haryana, Punjab.

graph TD
    START[367M youth 15-29<br/>the demographic dividend] --> SUP{Supply side:<br/>are they skilled?}
    START --> DEM{Demand side:<br/>do jobs exist?}
    SUP -->|No: rote pipeline,<br/>obsolete curricula| F1[~44% not job-ready<br/>ITI 46%, polytechnic 33%]
    DEM -->|No: jobless growth| F2[~40% grad unemployment 15-25<br/>under 7% get permanent salaried work]
    F1 --> COLLIDE[Dividend becomes liability]
    F2 --> COLLIDE
    style COLLIDE fill:#742a2a,color:#fff
                    

So it's a double bind. The unskilled can't compete, and the skilled can't find enough good jobs. This is why "just copy China and fix the syllabus" is only half an answer. Skilling raises the supply of capability; it doesn't create the demand for it. You need both, better training and an economy generating jobs worth training for. Fix one without the other and you get either idle skilled people or busy unskilled ones. India is currently managing to do both wrong at once.

6. So, do we even need growth?

Fair question, and I want to be honest instead of tribal about it, because the answer genuinely changes depending on where you're standing.

The case against chasing growth is real and I take it seriously. Above a certain income, GDP and human wellbeing stop tracking each other, this is the Easterlin paradox: richer people report being happier at any given moment, but a country getting richer over decades doesn't get proportionally happier. Nobel economists like Sen and Stiglitz have said for years that GDP was never a measure of a good life; it counts traffic jams and cigarette sales as "growth." There's a stronger version too, the social-cost-of-growth idea: past a point, the costs growth generates (congestion, pollution, stress, ecological damage) start offsetting its benefits, and growth becomes what post-growth economists call "uneconomic." Add planetary boundaries, the world economy is ~5x bigger than 50 years ago and roughly 60% of ecosystems have degraded, and for rich countries, more GDP is often just more crowding and more emissions with no wellbeing to show for it.

But here's the distinction that resolves it, and it's the same distinction from Section 1. That critique applies to quantity growth in places that already have enough. India is not there. At $3,000 a head, growth still buys the things that are non-negotiable, nutrition, hospitals that work, electricity, a job that isn't subsistence farming. Below the threshold, growth is welfare. Above it, growth is mostly noise.

So the honest answer isn't "growth good" or "growth bad." It's:

The tragedy would be spending the entire dividend window chasing headline GDP and headcount, the automatic stuff, while the thing that actually compounds (an educated, employed, productive population) never gets built. You'd end up bigger, hotter, more crowded, and still poor per person. Growth without development.

7. The visible cost: what growth-without-development does to a place

The social-cost-of-growth idea from the last section isn't abstract in India. You can watch it happen from a window. Its clearest form is the slow erosion of the tier-2 city.

Cities like Bhopal, Pune, and Chandigarh used to be a genuinely rare equilibrium: modern amenities and connectivity paired with livability, clean air, green cover, forested roads, low congestion, low stress. That combination is the whole reason people chose to stay in them instead of chasing a metro. Over the past ~3 years, that balance has been fracturing fast.

The mechanism is unplanned, metric-chasing commercial development. Green belts get paved. Open space gets built over. Roads choke. The city starts mimicking the worst, most congested traits of a tier-1 metro, without generating the economic throughput that was supposed to be the trade-off. You take on all the costs of density and none of its payoff.

And here's why this belongs in a report about growth: all of that construction, all those cars, all that concrete registers as economic activity, as growth. GDP counts the cement and the fuel and the new commercial floor space as positives. It does not subtract the lost green cover, the added commute, the worsened air, the vanished calm. This is exactly the measurement failure Sen and Stiglitz kept pointing at. The living index falls while the growth number rises. It is extensive growth in physical form, more inputs, more activity, less actual welfare per person.

I'll be straight that this part is observational, I'm describing what's visibly happening in these cities, not citing a green-cover dataset. But the pattern is consistent, and the connection to everything above is not a coincidence: a bureaucracy that can't update a university syllabus in under a decade is the same kind of bureaucracy that can't plan a city with foresight. Same static inertia, same habit of not adapting until the crisis has already landed. The classroom and the city are two outputs of one operating system.

8. What China's degree cull actually is (and what it isn't)

Now the China thing makes sense.

Scrapping 12,200 programs and adding 10,200 in AI, robotics, semiconductors, that's not really about degrees. It's a forced, violent switch from extensive to intensive growth. For 30 years China grew extensively, more workers off the farm, more factories, more concrete, the greatest extensive-growth run in history. That engine is now out of road: their working-age population is already shrinking, median age ~40, real-estate model collapsed. They can't add more bodies. They literally do not have them.

So the only lever left is productivity per person. And productivity per person is human capital plus technology. Which is exactly what they're forcing the education system to manufacture, on a deadline, because their own demographic clock (classrooms empty by 2040) is even harsher than India's.

I'm not saying copy it wholesale. Centrally deleting entire fields of human knowledge because a ministry decided AI ate them is the kind of move that looks decisive now and stupid in twenty years when it turns out you needed those "obsolete" people. Killing the humanities to mass-produce chip designers is its own failure mode. But the underlying logic, that a country running out of extensive growth has to consciously engineer the intensive kind, fast, through its people, that logic is correct. And it's the logic India hasn't internalized yet, because we're still drunk on the automatic growth.

9. Where this actually heads: three futures

Projecting forward, India lands in one of three places, and which one depends almost entirely on the conversion rate, both conversions, over the next 15 years.

graph TD
    START[India 2026<br/>66% working-age<br/>window open ~15-20 yrs] --> Q{Convert bodies to capability<br/>AND capability to jobs?}
    Q -->|Both work| A[PRODUCTIVITY DIVIDEND<br/>TFP-led growth<br/>escape the trap<br/>Viksit Bharat path]
    Q -->|One works, one fails| B[MIDDLE-INCOME TRAP<br/>grow, then stall<br/>old before rich<br/>Brazil / Malaysia path]
    Q -->|Both fail| C[SQUANDERED DIVIDEND<br/>hundreds of millions<br/>under-skilled AND jobless<br/>demographic liability]
    style A fill:#22543d,color:#fff
    style B fill:#744210,color:#fff
    style C fill:#742a2a,color:#fff
                    

The numbers that decide it are already visible, and they're not reassuring, a thin, sharp spike of world-class talent (the people staffing global research and the AI talent pool) sitting on a vast base the system barely equips, in an economy that isn't generating enough good jobs even for the ones who are ready. The correlation that scares me: a youth bulge with low conversion doesn't produce a dividend, it produces instability. That's the mechanism behind China's 16% youth unemployment, and India's own version, 40% for young graduates, is already here.

Scenario C isn't a doomer fantasy. It's the default, the outcome you get by doing nothing, because doing nothing still produces the bodies. It takes deliberate work to not land there.

10. A proposal on the table: "sprint governance", and where it breaks

If the root failure is a system that refuses to adapt until the crisis has already hit, then the natural question is: what would an adaptive one look like? One idea I keep coming back to is treating governance more like engineering. I want to put it down properly, and then hold it to the same honesty as everything else in this report, including where it falls apart.

The proposal, in three mechanisms:

graph TD
    A[Identify a specific bottleneck:<br/>urban or educational] --> B[Assign a cross-functional<br/>expert team, direct mandate]
    B --> C[Strict 90-day execution sprint<br/>non-extendable]
    C --> D[Deploy in a transparent<br/>sandbox / micro-zone first]
    D --> E{Open review:<br/>peers + citizens + data}
    E -- Fails the metrics --> F[Instant rollback /<br/>leadership rotation]
    E -- Meets benchmarks --> G[Scale it permanently]
    F --> B
                    

1. Time-bound sprints over five-year plans, defined problems get non-extendable 90-day windows and small cross-functional teams with real operating authority, instead of open-ended mandates lost in paperwork.
2. Sandbox before rollout, model or pilot a change in a micro-zone, expose the data openly, let independent panels and citizens tear at it, and kill it instantly if it degrades the living index.
3. KPI leaderboards with rotation, judge officials on verifiable outcomes (district employment, green-cover retention, resource efficiency), and rotate out teams that keep missing.

What it gets right, and this is not sci-fi. Injecting speed and measurable accountability into a bureaucracy without abandoning democracy has real precedent: the UK's "delivery units" under Blair, digital service teams like the UK's GDS and the US's USDS, Estonia's near-fully-digital state, outcome-based budgeting. The core diagnosis is sound, five-year horizons lose relevance before they finish in an exponential-tech era, and transparent, measurable outcomes genuinely do improve accountability.

Where it breaks, and I'd rather say this than sell you the clean version:

• Goodhart's Law. "When a measure becomes a target, it ceases to be a good measure." Leaderboards get gamed. Reward graduate placement and gig work gets logged as placement. Reward green cover and someone plants a fast-growing monoculture that photographs well and dies in three years. The metric climbs; the reality doesn't move.
• The 90-day horizon fights the things that matter most. Restoring an ecosystem, rebuilding an education pipeline, growing genuine skill, none of these resolve in a sprint. Sprint-thinking systematically biases toward what's measurable-fast over what's valuable-slow, and some of the most important public goods emit no 90-day signal at all.
• The technocracy-vs-democracy tension, the big one. China's speed comes precisely from not being a democracy. Engineers-as-ministers move fast because there's no one to answer to and no election to lose. You cannot cleanly bolt that velocity onto a democratic body, because the friction you want to delete, deliberation, dissent, the right to say no, is partly the point of democracy. "Accountable to a leaderboard" is not the same as "accountable to citizens," and it can quietly replace it. Rule by KPI becomes rule by whoever sets the KPI.
• Who picks the metrics and the teams? The framework assumes a neutral, competent selector exists. In practice that's the same political system you were trying to route around.

So here's where I land. The instincts are right and worth fighting for, faster feedback loops, radical transparency, real consequences for failure, technical people in technical roles. But "replace democracy with a hackathon" throws away the one thing (legitimacy, consent, protection from a bad metric-setter) that separates an accountable state from an efficient authoritarian one. The upgrade India actually needs isn't less democracy. It's a democracy that's faster, more measurable, and more honest about outcomes, accountability sped up, not accountability deleted.

11. The line I'd actually end on

Strip all of it down and here's what I believe.

We keep measuring the wrong growth. We celebrate the automatic kind, a bigger population, a bigger GDP number, a bigger everything, because it requires nothing of us and it feels like progress. But that growth is momentum. It's the train still rolling after the engine's cut. It runs out, and it runs out on a schedule we can already read: working-age share sliding after 2030, count peaking 2041, window closed by 2055.

The growth that would actually save the country is the kind we're worst at, and we're failing at both halves of it. On one side, the education machine isn't educating anyone, it's a memorization factory: mass-producing credentials, not capability. It certifies that you sat through something, not that you can do anything. On the other side, even the young people who are ready walk into an economy that isn't making enough jobs to hold them, 40% of young graduates unemployed, under 7% in permanent work. Unskilled and jobless, at the same time, at the scale of hundreds of millions.

And you can already see the bill coming due in the small things, in a tier-2 city losing its green roads to unplanned concrete, growth registering as activity while the actual quality of life drains out of it.

China looked at that same wall and did something brutal and probably half-wrong to climb it. India is still standing at the bottom, admiring how tall it's gotten just by breathing.

Getting bigger was never the achievement. We were always going to get bigger. The question was only ever whether we'd get better, and that's the one thing that was never going to happen on its own.

Sources

• China degree reforms (12,200 cut / 10,200 added, 2021 to 2025, >30% of programs; 12.7M grads 2026; youth unemployment 16.9%; demographic-collapse framing): Forbes, Bloomberg, SCMP, WION, June 2026; China Policy (Substack).
• India employability (56.35% in 2026, up from 46.2% in 2022; ITI 45.95%, polytechnic 32.92%; CS/IT ~80%): India Skills Report 2026 (ETS/CII/AICTE/AIU/Taggd), via Careers360, Nov 2025.
• Graduate unemployment / jobless growth (~40% for 15 to 25, ~20% for 25 to 29; <7% of male graduates in permanent salaried work within a year; 3.7% white-collar; youth 15 to 29 = 367M; working-age share declines after 2030): State of Working India 2026, Azim Premji University; The Wire, Mar 2026.
• Global job displacement (92M displaced / 170M created / net +78M by 2030; 22% of jobs transformed; ~39% of skills obsolete in 5 yrs; 63% of employers cite skills gap): WEF Future of Jobs Report 2025.
• Demographics (TFR ~1.9 to 2.0; working-age 66.4%; count peaks ~2041 at ~1.01B; dividend closes ~2055; population peaks ~2062; growth rate 0.86%; median age ~28 to 29 vs China ~40; school enrolment -13.4M 2019 to 2025; elderly ~20% by 2050): SRS 2024, NFHS-6, UNFPA, Economic Survey; Business Standard, Al Jazeera, Daily Pioneer, InsightsonIndia, 2026.
• Growth economics (GDP +7.6% FY26; per capita >$3,000; TFP 2.1% to 1.4%; manufacturing ~14%; agri 45% of workforce / 16% of GVA; R&D 0.65%; SME within-firm productivity flat 1995 to 2018; middle-income-trap set): MOSPI, OECD India Snapshot 2026, RBI, World Bank India Country Economic Memorandum, The Wire, GDPIndex.
• "Old before rich" / productivity-dividend framing: Observer Research Foundation, Mar 2026.
• Do-we-need-growth (Easterlin paradox; social-cost-of-growth; Sen/Stiglitz on GDP; planetary boundaries; post-growth): Lancet Planetary Health 2024, Club of Rome / Earth4All, Georgetown Journal of International Affairs.
• Governance precedents (delivery units, GDS/USDS, Estonia digital state, outcome-based budgeting; Goodhart's Law): general public-administration literature.

Growth-composition and demographic figures are projections and estimates; treat the exact years as ranges, not precision. The tier-2 urban decay in Section 7 is observational, not from a dataset. The argument doesn't depend on any single number, it depends on the direction all of them point.