What the Coverage Misses.

• Manufacturing has been shrinking as a share of the economy

  Manufacturing value added fell from 15.9 per cent of GDP in 2000 to 13.5 per cent in 2025, against China’s 24.7 per cent. The factory-of-the-world framing is a forecast the data has not yet begun to confirm. What it does not prove: a falling share is not a falling level, and if services are absorbing the value instead, that may be the better trade. What it does undercut is the labour-absorption case, which needs factories rather than value.

  What would have to changeShare follows power reliability, land and logistics cost, and scale in a small number of sectors, and it follows them years after the policy that addresses them. The eleven-point gap to China is both the size of the opening and an honest measure of the distance. Watch the input costs, not the announcements.

  Against the peer setVietnam runs at 24.5% of GDP and Mexico at 20.0% — the China+1 comparison is not India’s to inherit by default.

• Research spending has fallen as a share of the economy

  Gross R&D expenditure was 0.76 per cent of GDP in 2000 and 0.65 per cent by 2020. The software and engineering-services story told abroad does not appear in the research-intensity data, and an allocator assuming one implies the other is assuming something the series does not support. What it does not prove: the series ends in 2020, and R&D performed in India but booked to a foreign parent is under-captured by construction, so the true figure is understated by an unknown margin.

  What would have to changeEvery economy that lifted research intensity did it through sustained corporate spending, not budget allocations. The number worth tracking is whether private research spend compounds faster than GDP across a full cycle; and, given the measurement gap above, whether the statistical treatment of captive R&D centres is fixed at all.

• Remittances are larger than foreign direct investment

  Personal remittances received reached roughly 151 billion US dollars in 2025, from 13 billion in 2000, the largest inflow of its kind in the world, and a current-account stabiliser that almost never appears in an India allocation memo. What it does not prove: remittances are consumption-linked and not investable. They track Gulf construction cycles and global labour demand, so they are procyclical to things that have nothing to do with India.

  What would have to changeThe number starts mattering to an allocator only if any of it is intermediated into domestic capital formation rather than spent. That conversion, through deposits, insurance and pensions, is measurable, and it is the thing to track rather than the headline inflow.

  Against the peer setThe next-largest EM recipient in this set, Mexico, receives $64bn; the scale of India’s inflow is unmatched, and so is its exposure to the labour markets behind it.

• The economy is under-levered, not over-levered

  Domestic credit to the private sector is about 44 per cent of GDP, up from 28 per cent in 2000, against 194 per cent in China and 201 per cent in the United States. India runs at roughly a fifth of China’s credit depth. The emerging-market risk framing an allocator usually arrives with assumes the opposite problem. What it does not prove: low credit penetration is also a constraint. It is part of why the investment cycle leans so heavily on public capital expenditure, and the ratio says nothing about asset quality, which is a separate question with its own paper.

  What would have to changeHeadroom is not the same as growth. Credit deepens if underwriting quality holds while it extends, which is the precise thing that failed in the previous cycle. The test is whether the ratio can rise without the bad-loan ratio rising with it: an empirical question, checkable each quarter, not a projection.

  Against the peer setBrazil stands at 75% of GDP and Indonesia at 36%: the EM peer set is itself split, and ‘under-levered’ only means something against a stated comparator.

• Barely a third of the country is counted as urban

  About 36 per cent in 2025, against 28 per cent in 2000. China stood at 36 per cent in 2000 and reached 66 per cent by 2025. India today sits roughly where China sat a quarter of a century ago. Consumption models built on a developed-market urbanisation curve tend to run decades ahead of where the series actually is. What it does not prove: India’s official definition of “urban” is unusually restrictive, so the real figure is very probably higher. That is a measurement problem before it is an economic one, and exactly the kind of definitional break this practice records rather than smooths over.

  What would have to changeThe upside here is arithmetic (a low base on a long curve), but it is gated by municipal finance and land title, neither of which moves on a national policy cycle. The leading indicator is not migration; it is whether cities can borrow against their own revenue.

  Against the peer setIndonesia is at 59% and Vietnam at 39%: on this series India trails not just China but most of its own peer set.

• Services exports have grown more than twentyfold

  From roughly 17 billion US dollars in 2000 to 412 billion in 2025. The measured comparative advantage is in services, not in the manufacturing story told abroad. What it does not prove: these are current dollars, not inflation-adjusted, and concentration in a few sectors and a few client geographies is a real exposure rather than a footnote.

  What would have to changeThe first twentyfold came from adding people. The next leg has to come from price per hour rather than headcount, which is a different bet and a more fragile one. It competes directly with automation of the same tasks. Track realised billing rates, not export totals.

• Nine in ten adults now hold a bank account

  Up from roughly a third in 2011, one of the fastest expansions of financial access recorded anywhere. What it does not prove: an account is not credit and it is not savings. Dormancy and actual usage are separate measurements, and the headline number answers neither.

  What would have to changeAn account becomes economically useful at the point it produces a credit history. That is the unfinished half, and it is what would turn financial inclusion into a financial system, visible in lending to first-time borrowers rather than in account counts.

Source: World Bank Open Data: NV.IND.MANF.ZS, GB.XPD.RSDV.GD.ZS, BX.TRF.PWKR.CD.DT, FS.AST.PRVT.GD.ZS, SP.URB.TOTL.IN.ZS, BX.GSR.NFSV.CD, FX.OWN.TOTL.ZS. Retrieved 5 September 2026. Reported annual values; account ownership is surveyed periodically rather than annually. Peer-set values from the same source, retrieved 5 September 2026.