Data note · 21 July 2026 · revised 22 July 2026

Maharashtra's FDI: stock-market money, or money in the ground?

Maharashtra books about a third of India's foreign direct investment. A reasonable suspicion is that much of it is foreign money buying Indian shares rather than building Indian factories. The suspicion is half right — but not in the way it is usually framed, and the precise answer turns out to be unobtainable from official data for a reason worth stating plainly.

Revision · 22 July 2026 An earlier version framed everything that is not manufacturing FDI as "not money in the ground". That was too harsh in one specific way: 16.13% of Maharashtra's FDI is construction-infrastructure — ports, power, transmission — which is genuine physical capital formation that the implementation (IEM) cross-check cannot see, because IEM counts manufacturing only. The bottom line below has been narrowed from "real industrial weight is a third of the FDI share" to "real manufacturing weight", and the infrastructure blind spot is now stated. The attribution and services findings are unchanged. Corrections are displayed, never silent.
The short answer
0%
of the FDI figure is stock-market money — by definition, not by estimate
3.2×
more FDI share than implemented-industry share: the real distortion

Portfolio money is definitionally excluded

RBI's Master Direction on Foreign Investment draws a bright line at 10% of post-issue paid-up equity on a fully diluted basis. At or above it, an inflow is FDI. Below it, the identical rupee is Foreign Portfolio Investment and sits in a different account entirely. DPIIT's own factsheet demonstrates the separation on its face: it carries FPI as a distinct column — "Investment by FPI's Foreign Portfolio investor Fund (net)" — sitting outside the Total FDI Inflow column.

So sub-10% buying of listed Indian shares never enters Maharashtra's ₹8,59,196 crore (USD 107,094 mn), cumulative October 2019 – March 2026, 31.35% of the national total.

Worth noticing The two channels have moved in opposite directions. Maharashtra's FDI kept rising while net FPI equity flows turned sharply negative: +₹2,08,212 cr (FY2023-24) → −₹1,27,041 cr (FY2024-25) → −₹1,80,832 cr (FY2025-26). Foreign portfolio money has been leaving Indian equities over precisely the period Maharashtra's FDI grew.

The real distortion is geographic, not portfolio

DPIIT attributes an inflow to the state of the Indian company that receives the remittance, not the site of the project it funds. DPIIT's own phrasing for the Maharashtra deal list is "Remittance-wise – through Indian companies".

Negative finding DPIIT publishes no explicit caveat anywhere stating that state attribution follows the registered or reporting office rather than the project location. The consequence is left entirely implicit in the phrase "reported at Maharashtra state". This is arguably the single most consequential undocumented assumption in the dataset.

DPIIT's own top-five Maharashtra recipients show the artefact plainly:

#RecipientActivityInvestorAmount
1Reliance Retail VenturesStorage & warehousingSaudi PIF₹9,555 cr
2Reliance Retail VenturesStorage & warehousingSLP Rainbow, Singapore₹8,813 cr
3Ambuja CementsClinker & cement manufactureHarmonia, Mauritius₹8,341 cr
4Reliance Retail VenturesStorage & warehousingQatar Holding₹8,278 cr
5HDFC CredilaOther credit grantingKopvoorn BV, Netherlands₹7,642 cr

Three of the top four are stakes in a nationwide retail and warehousing group, booked 100% to Maharashtra because the recipient is Mumbai-registered — the warehouses they finance sit in every state. Number three is a cement company whose plants are mostly in Himachal, Gujarat, Rajasthan and Chhattisgarh. Number five is a pure financial transaction with no physical project anywhere.

A hypothesis the data killed

Going in, the obvious suspect was a large unallocated bucket quietly absorbing inflows. That was true historically and is now obsolete. In the legacy RBI regional-office series (April–June 2019), "Region not indicated" was 15.96% of all FDI equity inflow — larger than the entire Mumbai office at 9.60%. In the current state-wise series, "State Not Indicated" is 0.01%, and cross-checking the state table's total against DPIIT's annual series (~USD 341.6bn vs ~342.7bn) shows it now captures essentially 100% of national inflow.

The problem is not missing data. It is complete data assigned to the wrong geography.

What Maharashtra's FDI is actually in

DPIIT Table 6.3(i), FDI Synopsis on State – Maharashtra, October 2019 – December 2024:

Services, software and trading together are 42.5%, against 4.33% for automobiles — the only unambiguous manufacturing line in the published table. But note the second row: 16.13% is construction-infrastructure — ports, power, transmission — which is physical capital formation, just not the kind an Industrial Entrepreneur Memorandum records. So the mix splits three ways, not two: ~43% clearly non-physical (services/software/trading), ~16% physical infrastructure that the IEM cross-check below is blind to, ~4% manufacturing, and 37% unpublished (DPIIT prints only the top five sectors).

For context, construction-infrastructure is 6.9% of national FDI over the same window (real-estate development is a separate 0.6%), so Maharashtra's 16.13% is about 2.3× the national infrastructure rate — this is disproportionately a Maharashtra phenomenon.

The finding that actually answers the question

DPIIT's FDI equity inflow explicitly includes the purchase of existing shares. The factsheet says so verbatim, identically, under the country, sector and state tables:

"%age worked out in USD terms & FDI inflow received through Government Route + Automatic Route + acquisition of existing shares only."

Buying shares from a selling shareholder transfers ownership and creates zero new capital formation — no plant, no job, no machine. DPIIT has booked even a USD 3.1bn pure share swap as equity inflow.

The gap DPIIT publishes no greenfield-versus-brownfield split at all — not nationally, not by state. It is therefore impossible from official Indian data to state what fraction of Maharashtra's FDI built anything. That is the honest answer, and it is a gap in the source rather than in the analysis.

The cross-check: what the ground says

Where official FDI data is silent, implementation data is not. Setting DPIIT's FDI share against each state's share of industrial investment actually implemented (IEM Part B, five years) gives the best available proxy. Both series are shares of their own national total, so they sit on one scale.

Share of FDI (FY26) Share of implemented industrial investment (5yr) Ratio = FDI share ÷ implementation share
Maharashtra
3.2×
Karnataka
12.7×
Gujarat
0.4×
Tamil Nadu
4.4×
Andhra Pradesh
0.2×
Odisha
0.03×

Maharashtra takes 41% of tracked FDI but shows 12.7% of implemented industrial investment, and converts just 34.9% of its own stated investment intent into implementation — the weakest of the top four states.

Gujarat is the mirror image: 35% of implemented industrial investment on 13% of the FDI. Odisha is the extreme — 18% of implementation on 0.2% of FDI, an industrial economy essentially invisible to anyone reading FDI tables. Karnataka's 12.7× is the cleanest proof that this is a registered-office and services effect: Bengaluru's software FDI is real money that is not plant and machinery.

What the ratio cannot see IEM Part B counts manufacturing investment only. Maharashtra's 16% construction-infrastructure FDI builds real assets that can never appear in an IEM, so the ratio overstates the genuinely "not-in-ground" share. Read it as a divergence detector — foreign money booked here vs factories built here — not as a verdict on whether the money is productive. The fuller all-state version, on a cumulative-FDI basis, puts Maharashtra at 2.5× rather than the single-year 3.2× shown above.

Bottom line

Limits

Sources — DPIIT Quarterly Fact Sheet on FDI Inflow, April 2000 to March 2026 (Sections D/E/F, Annexure-C) · DPIIT Table No. 6.3(i), FDI Synopsis on State: Maharashtra, as on 31.12.2024 · DPIIT Table No. 16, RBI regional-office series, April–June 2019 · RBI Master Direction on Foreign Investment in India · NSDL FPI Net Investment Details (financial-year series) · DPIIT Annual Report IEM Part B, state-wise implemented investment.
Machine-readable data: maharashtra_fdi_composition_2026-07-21.json · state_iem_implementation_2026-07-19.json. Written version: MAHARASHTRA_FDI.md. Corrections are displayed, never silent.