# Builders, Not Investors

Author: Pramod Prasanth
Published: 2026-08-19
Canonical: https://pramod.ch/essays/builders-not-investors
Description: Keralam's diaspora has been sending capital home for forty years. Very little of it has been structured to build firms. The question is what would have to change. From a series asking what Keralam 2040 would have to be.

---

A house in central Keralam goes up one floor at a time. The ground floor first, then years, then the first floor, and rebar left standing out of the roof slab where a second floor will go. The paint is fresh because it was done for a visit. Each floor is a stretch of years worked in Dubai or Doha or Riyadh by people who are home for two weeks twice a year and spend part of that fortnight deciding what gets built next. Keralam's diaspora has been sending capital home for forty years. It has financed those houses, and gold, deposits, consumption, education and security. Very little of it has been deliberately structured to become patient capital for firms.

This essay asks what it would take for a fraction of that capital, and the larger pool of dormant capital sitting inside Keralam itself, to become firms instead of houses.

This series has argued that Keralam's half-finished miracle has talent, institutional, cluster, and research gaps. The capital question is the fifth, and it threads through all four. None of the earlier arguments produces firms by itself.

> Keralam's diaspora has been sending capital home for forty years. Very little of it has been deliberately structured to become patient capital for firms.

### The investor and the builder

There is no shortage of "investment" in Keralam. Investment summits announce thousands of crores. MoUs accumulate. Foundation stones are laid. What happens after that is harder to read than either the state or its critics allow. The Invest Kerala Global Summit of February 2025 collected 499 letters of intent worth roughly ₹1.8 lakh crore. By November, a hundred of those projects, worth ₹35,111 crore, had got under way. By the first anniversary in February 2026, the state counted 115 projects under construction, worth ₹54,908 crore, out of 449 expressions of interest.[^1]

In November the state reported a conversion rate of 36.23%, and the industries minister called a hundred projects taking off "a record in the history of investment summits in the country". The hundred are real. The percentage is a hundred divided by 276, which is the count of projects that had already obtained land. Against the 499 letters of intent the summit collected, the same hundred projects come to 20%. The February count runs the same way: 115 is 40% of the 285 projects with land, and a quarter of the 449.

A conversion rate counts announcements that turned into construction. It says nothing about whether what gets built compounds: whether the firm keeps its intellectual property here, hires here at the top of its wage scale, and is still here in fifteen years. Keralam is not unique in measuring the tractable thing. What is different here is the kind of capital that turns up.

Most of the capital that announces itself in Keralam is calibrated for a different kind of place. It seeks upfront subsidies and easy land. It assumes a workforce that is cheap, abundant, and weakly organised. It treats environmental clearances as a friction to be managed. It carries an embedded expectation that exit will be easy: if Keralam becomes inconvenient, the same money can move to a state with better terms. That capital, when it looks at Keralam, sees difficulty. Land is dense. Labour is organised. Voice on the shop floor is loud. Environmental concerns are politically serious. The state is treated as a counterparty, not as a service provider.

From the capital's perspective, this reads as risk. From Keralam's perspective, these are its core social achievements. They are not bugs in the investment environment. They are the human development indicators the first essay in this series called a half-finished miracle, and the productive economy that fits them has not yet been built.

The mismatch is not moral. It is structural. Capital optimised for one operating environment does not become useful when transplanted into another. A different kind of capital is required. Call the first kind the investor and the second kind the builder.

The investor is interested in arbitrage. The builder is interested in compounding. The investor has exit optionality as a design feature. The builder is committed enough that leaving would destroy what was made. The investor centralises IP and high-value functions elsewhere; Keralam gets the execution slice. The builder retains IP and core capability locally because that is where the work is. The investor measures success in quarters. The builder measures in decades.

A firm building medical devices for a global market needs a workforce that operates expensive equipment carefully, a regulatory environment that takes safety seriously, a city its engineers want to live in, and a research base it can recruit from. Keralam's social contract, organised labour, environmental sensitivity, higher wages, public investment in health and education, is the operating context that firm needs. The same social contract is exactly what footloose capital reads as obstacle.

> Organised labour, environmental expectations, and higher wages look like risk to capital optimised for low-cost arbitrage. They look like features to a firm building for the long term.

*Figure: Both percentages are true. They answer different questions. One asks how many projects that had already cleared the hardest hurdle went on to start work, which is a fair measure of administrative follow-through. The other asks how much of what was announced became something real, which is the question an investment summit invites when it publishes a headline number.*

The third essay in this series argued that Keralam needs an institution like Singapore's Economic Development Board: a body that knows what it wants and can hold a thirty-year industrial bet. The other half of that story is what kind of capital responds to such an institution. The EDB matched the firms it brought in, so that one plant's output fed the next, and went out to sell the idea to foreign companies. The institutional architecture and the capital strategy were two sides of one decision.[^2]

### The dormant capital problem

Keralam is not capital-poor. It looks capital-poor only because most of the capital is dormant.

The World Gold Council's head for India has said that gold use per person is seen to be higher in Keralam.[^3] Non-resident deposits in Keralam's banks stood at ₹3.24 lakh crore in March 2026.[^4] Remittances reached 23.2% of net state domestic product in 2023, up from 13.5% five years earlier, which is 1.7 times everything the state government collects in revenue in a year.[^5]

The problem is not the absence of capital. The problem is the form it takes. Almost none of it is structured to back firm-building. Household fixed deposits sit in bank accounts. Gold is the most patient capital of all, and the most immobilised. Real estate is a savings instrument and a status signal but not a productive investment. Diaspora remittances go above all to renovating houses, paying off bank loans and education, and what is saved sits as cash at home, a bank deposit, a plot of land or a gold chain. The Kerala Migration Survey has no line for money put into a business.

> Capital is not patient because of its origin. It is patient because of the institutions that hold it.

What I have seen in Basel is the answer to this in another form. Swiss household capital is famously cautious. The private investor base is conservative. Yet Switzerland built one of the world's most R&D-intensive economies: Basel's life sciences cluster, ETH Zurich's spin-off pipeline, the watch industry's hundred-year horizons.[^6]

Private banks holding cautious money well is the easy explanation, and on its own it is not enough. Swiss patience rests on a stack of things working at once: family ownership that does not answer to a quarterly market, pension and insurance funds with liabilities measured in decades, universities that generate opportunities a firm can buy into, a skilled labour market deep enough to staff a fifteen-year programme, corporate reinvestment out of retained earnings rather than fresh capital, legal and regulatory predictability, and research-to-industry links dense enough that the distance from laboratory to production line is short. No single one of those is the mechanism. The combination is, and it took decades to assemble.

Israel ran a different version of the same move with Yozma in 1993. The government put up $100 million: $20 million into a fund of its own, and $80 million to take a 40% stake alongside private domestic and foreign managers in new Israeli venture funds. The mechanism that mattered was the exit. Each private partner held an option to buy the government's 40% back over five years, and most of them did. The state supplied the credibility needed to raise a first fund and then removed itself. By 1996 the programme had seeded eleven funds holding $263 million, which went on to make 217 investments and exit 122 of them.[^7] Israel had no inherently patient capital. It built the institutional layer that turned cautious capital into firm-building capital.

The Keralam question is the same. The capital exists. The institutional infrastructure that would redirect it from real estate into firms does not.

A Keralam Builders Fund is one shape of the missing institution: a state-anchored, diaspora co-invested vehicle with a fifteen-to-twenty year horizon and a mandate to back qualifying firms rather than real estate. Yozma is the template.

The design has to withstand a minister who wants a favour and a manager who wants an easy year. The state anchors the fund and does not choose the firms. Independent professional managers make the investment decisions and are paid on what the portfolio returns. Diaspora and institutional investors enter on a defined co-investment structure, with an option to buy out the state's stake on fixed terms after a fixed period, which is the Yozma mechanism and the reason it worked. The mandate is confined to Keralam-domiciled firms that keep core intellectual property, research and high-value employment in the state. Reporting is public and annual, on portfolio returns and on capability retained locally, not on announcements made or money moved. That last one is the discipline the rest depends on, because a fund measured on how much money it moves will move money whether or not the firms deserve it.

Differentiated tax treatment is a second lever, and it needs stating as a proposal rather than as a complaint about parity. Non-resident capital placed in an operating company and non-resident capital placed in land are not taxed identically today; treatment turns on the vehicle, the investor's residence status, the asset, the character of the income and the structure of the transaction.[^8] What does not exist is a deliberate wedge between the two, wide enough and stable enough that a diaspora investor choosing between a flat in Kochi and a stake in a Kochi firm can see it from the outside and price it over a decade. Building that wedge is the ask. The wider category of redirected domestic savings, household deposits flowing into Keralam-domiciled patient-capital instruments rather than into central government borrowing, is the harder political move but the largest pool.

### The other half of the pair

Singapore's Economic Development Board is the institution Keralam's industrial strategy has learned from. It is only half of what Singapore built.

The EDB, founded in 1961, went out and recruited firms. In its first years it also lent to industry, until a new development bank, the Development Bank of Singapore (DBS), took that work over in 1968. In 1974 Singapore incorporated Temasek and transferred to it an initial portfolio of thirty-five companies the Ministry of Finance had been holding directly, so that the ministry could go back to writing policy while a separate body owned those assets and was judged on making them pay. In 1981 it created GIC to invest the accumulated reserves abroad.[^9] One body to attract firms, one to own industrial assets and make them pay, one to manage the surplus. Keralam's conversation has been almost entirely about the first.

This matters now because of who is holding the long money. Sovereign wealth funds were party to $46 billion of AI venture transactions between January and September 2025, close to half of all such investment in the period, and they entered that market as serious players for the first time.[^10] This is capital with a mandate to think in decades, actively looking for industrial exposure, and a great deal of it sits in the Gulf, where most of Keralam's emigrants already work.

The question that poses for Keralam is concrete. What would KINFRA, or whatever succeeds it, have to become before a GIC or a Gulf sovereign fund could treat it as a partner for a decade rather than a counterparty for one transaction?

Three things would have to be true of it. A balance sheet that survives the diligence such an investor's own auditors would run, which means published accounts, assets carried at defensible valuations, and a record of returns rather than a record of disbursements. A governance structure that outlives an election, because a fund committing for ten years is underwriting the institution and not the minister who signed. And a mandate to hold equity, take losses on individual bets, and be judged on the portfolio, which is close to the opposite of how a state agency is normally assessed.

Keralam has already built one institution that meets all three tests. Cochin International Airport Limited dates from 1994. By March 2001 it had about ₹79 crore of paid-up equity, in shares of ₹10. About two-fifths came from the state and its own enterprises, and about a quarter from the companies that would serve the airport, Air India, BPCL and two banks among them. Most of the rest came from non-resident directors and their families, and from some ten thousand small shareholders, 3,880 of them non-resident Indians spread across thirty countries, who between them put in about ₹10 crore.[^11] It paid a dividend every year from 2003-04 to 2019-20, and resumed after the pandemic. It publishes its accounts, its governance has outlasted every change of government since its founding, and the state answers for what it returns rather than for what it announced.

Acquiring the site took about 1,300 acres and the resettlement of 872 families, who received market-value compensation, six cents of land each, and, for one member of each family that lost both house and land, an airport job or a livelihood linked to it.[^11] The dividend record begins after Keralam paid that bill.

An airport is still not a firm. It has a captive catchment, a regulated tariff, and land whose value it partly creates, which puts it closer to the real estate this essay has been arguing against than to a company that has to win a customer every quarter. Cochin shows that the state can anchor a Keralam-domiciled vehicle on a twenty-year horizon and bring Keralites at home and abroad in alongside it, and that it can run one to a standard an outside investor would recognise. It does not show that the same money will back a company that might fail.

None of that requires Delhi's permission. All of it requires the state to decide that an industrial-capital institution is something it wants to own and be measured against, rather than a scheme it wants to announce.

### The returnee paradox

The temptation, when arguing for builders, is to lean on the returnee story: the Keralam-origin engineer who left for Silicon Valley, made money, and is now building something back home. It is a flattering narrative. It is also misleading as the default model.

The empirical pattern crosses over. An Observer Research Foundation study of 596 Indian high-tech startups founded between 2016 and 2023 gave the pattern its name, the Indian returnee paradox. Returnee-founded firms do better through the middle of the distribution: median capital raised of $441,000 against $113,000 for domestic-founded firms, and median revenue of $92,500 against $43,700. Domestic founders own the top of it outright. Their mean revenue was $5.39 million against the returnees' $0.68 million, and of the firms in the sample that reached a billion-dollar valuation, three were founded by domestic teams and none by returnees.[^12] Returnees cluster in the knowledge-intensive sectors where a foreign track record reads as credibility, and in none of those sectors do they outnumber domestic founders.

Returnees are not underperforming. They produce a dependable middle and seldom an outlier. The founders in this sample who built something large were domestic, and the authors suggest that local founders, embedded in local markets, institutions and ways of getting things done, may now be better placed to reach scale. They call their own analysis correlational. My own reading is that they grew up inside the discipline of scarcity rather than carrying the abundance assumptions of foreign capital. One of the study's authors is AnnaLee Saxenian, whose earlier work on brain circulation made the case that returning migrants drive technology development in their home countries. The paper calls its India finding a contradiction of that earlier work, and then narrows it to the top of the distribution, where the large outcomes now go to domestic founders.

For Keralam specifically, this matters. The patient builder here is primarily the Keralam-grown founder who never left, or who left for a defined window and came back without that becoming the founding identity. Returnees and incoming foreigners are useful: they bring networks, technical depth, and capital channels. They are auxiliary infrastructure, not the engine.

> A successful Keralam 2040 is not one where the diaspora came back to save it. It is one where fewer people felt they had to leave in the first place.

This reframes the diaspora capital argument. Diaspora capital is one channel, not the centre. The centre is dormant domestic capital, redeployed through institutions Keralam has not built.

### The AI volatility objection

The fairest objection to a fifteen-year capital horizon in 2026 is that the underlying technology now moves faster than the horizon. AI capability stacks turn over fast. A firm-building bet made today could be obsolete before the capital has finished vesting. Why fund patience in a market that no longer rewards it?

The objection does not survive contact with the firms it describes. Anything that needs a workforce, a regulatory pathway, embedded enterprise customers and a production line takes years to compound, and the AI tooling inside that work updates on its own clock without changing the clock the work runs on. A medical-device firm using AI for diagnostic imaging still sits inside a multi-year regulatory pathway and a longer adoption curve with payers. A precision-manufacturing firm using AI for process optimisation still needs the manufacturing line. The capital backing the deepest AI-era work runs on horizons longer than the AI cycle, not shorter, which is why sovereign funds are already underwriting large AI data-centre buildouts.

For Keralam the implication is narrow and useful. Patient capital and AI-enabled work are not in tension. The firms that can hold a fifteen-year horizon are the same firms that can put AI to work inside it.

### Five policy principles for a builder-friendly Keralam

The policy moves do not need to be elaborate. They need to be coherent. Their structure matters.

First, incentives that vest over time rather than upfront. A subsidy paid before a firm has built anything attracts capital that came for the subsidy. A benefit that grows as the firm hits milestones (local headcount, local R&D spend, IP retention, years of continuous operation) attracts capital that intends to compound. Singapore's Pioneer Certificate Incentive and Development and Expansion Incentive are written the conditional way. Each runs for five years; extension depends on the firm committing to further expansion; the firm reports its progress and is judged on it; and if it breaches the terms, Singapore can revoke the incentive and recover the tax relief already granted.[^13] The award requires the company to build what it said it would build. Keralam's incentives mostly do not work this way.

Second, benefits linked to local wage bill and R&D spend, not only capital expenditure. Capex is easy to measure and easy to game; a firm can build a shell facility, claim incentives, and keep all high-value functions elsewhere. Local wage bill captures whether the firm is actually building a team in Keralam. Local R&D spend captures whether the firm is doing intellectual work here or only execution. Singapore's assessment criteria start from the employment created, weighed by skills and seniority rather than headcount alone, then total business expenditure and the commitment to grow technology and knowhow locally. Fixed asset investment in plant and buildings is required in addition, and only of manufacturing projects.[^13]

Third, state procurement as the first customer. Keralam's government is itself a large buyer of health services, public infrastructure, education materials and government IT. Directing a meaningful share of public procurement to qualifying locally-built products (medical devices from a Keralam device cluster, marine engineering from the Kochi precision cluster, applied-AI tools for Keralam's public service network) de-risks the first-customer problem that hobbles most early-stage Keralam firms. It is the most under-used industrial-policy lever the state owns.

A preference regime is also the easiest of these five to corrupt, and saying so strengthens the standard rather than weakening it. Preference without published qualification criteria, measurable performance standards and genuine competition among qualifying suppliers stops being industrial policy and becomes protection for whoever is best connected. The discipline that makes it work is the same one that makes the fund work: the criteria are published, the performance is measured against them, and a supplier that fails them loses the preference.

Fourth, IP retention requirements for research-linked support. Firms that draw on the labs the previous essay described (Sree Chitra, CMFRI, the proposed concentrated research base) should accept a condition of access. Intellectual property belongs to a legal entity rather than to a place, so the condition has to be written that way: ownership, or exclusive commercial rights, held by a Keralam-domiciled entity for a defined period, with stated exceptions where genuine multi-party research makes shared or split ownership unavoidable. This prevents the joint-degree pattern transposed onto industrial research, where Keralam provides the input and the intellectual property is captured elsewhere.

The harder question sits underneath, and this essay does not settle it. A domicile requirement tells a firm where to register something. It does not make Keralam the place a firm would choose to hold its most valuable asset anyway, and a firm that resents the condition will find a way around it. What would make a Keralam-domiciled entity the natural home for the IP, rather than the compulsory one, is a question about courts, tax, talent depth and the credibility of the state as a long-term counterparty. Better to answer that before the condition is drafted.

Fifth, administrative predictability over speed. Builders do not need the fastest clearances in India. They need predictable ones: timelines they can plan against, decisions that hold across electoral cycles, and bureaucratic interlocutors with the authority to deliver on what they say. Unpredictability is more damaging to a fifteen-year bet than slowness, because the bet cannot be reset every time the political wind shifts.

> Time-vested incentives attract capital willing to commit. Upfront subsidies attract capital looking for a subsidy. The difference is in who shows up.

Each principle is mechanical, not aspirational. None requires the social contract to change. All require institutional discipline of the kind the third essay argued for, applied here to the incentive and procurement architecture rather than to the cluster mandate.

### The female talent dimension

The capital and institutional story has a talent dimension that connects directly back to the second essay in this series. A builder operating in Keralam cannot afford to ignore half the talent pool. The female labour force participation gap is a competitive constraint on every firm that builds here.

The firm that designs for female participation (care infrastructure, transit safety, dormitory and shift design, predictable working environments, leadership pipelines that do not bottleneck) recruits from a fundamentally larger pool than the firm that does not. In a state where resident women outperform resident men in tertiary attainment, the firm building on a male-only operating model is competing for the smaller half of the educated workforce.[^14] The case is a talent case, not a diversity case. It belongs in the builder argument as part of the operating-context calculation.

### Cosmopolitan flow as background infrastructure

The cleanest signal that a state is producing the right kind of capital and the right kind of firms is what its talent flow looks like at steady state. Singapore, Zurich, and Basel all carry continuous in/out movement: foreign professionals on multi-year assignments, native talent rotating through tours of duty abroad, both treated as background infrastructure rather than as policy events.

Keralam's current diaspora discourse is structured around departure as loss and return as redemption. A 2040 Keralam that has built the talent base, the institutional architecture, and the capital infrastructure should be able to drop that framing. Native outflow becomes a normal life choice when the underlying economy is strong; the pathology is not the going but the not-coming-back, and the not-coming-back is what an economy with insufficient firm-building produces. Native return becomes a normal commute, not a rehabilitation programme. Foreign inflow (researchers, engineers, founders from Korea, Brazil, Switzerland, Vietnam, the Gulf) becomes the diversity signal that the firms and the research base both need. A later essay in this series argues that the city itself has to be built to absorb that inflow.

> Keralam 2040 succeeds when leaving is a choice, not a structural necessity, and when arriving from outside is a choice, not an exception.

#### What Keralam can do now

Within the state's current authority, the moves are real. Design the time-vested incentive structure described above within the existing industrial-policy framework. Launch a Keralam Builders Fund with state anchor investment and diaspora co-investment, structured for a fifteen-to-twenty year horizon, mandated to back firms rather than real estate. Implement state procurement preferences for locally-built qualifying products and services. Create a returnee founder programme: not a generic subsidy track, but a dedicated bureaucratic interlocutor with the authority to make procurement, research-access, and IP commitments. Reform IP retention requirements for state-research-linked firms.

#### What requires federal change

Some of the harder moves need Delhi. The income-tax code has no deliberate wedge between diaspora capital placed in a firm and diaspora capital placed in land, and only the centre can create one.[^8] High-skill visas, which any serious builder will need, are a central matter. As I read them, central R&D incentives favour large firms in established clusters over smaller firms in newer places, which works against Keralam. And the room a state has to depart from central industrial policy is limited. Keralam needs movement on all four, and it controls none of them.

---

Getting the capital right is not the last decision. A state with patient money still has to say what it is for, and then build the institutions, the training and the place that one answer needs. That choice, and what it obliges, is the question the next essay opens.

The houses will still be there. They are not a mistake, and nobody who built one owes anybody an apology. A family with no institution willing to turn its savings into a firm put those savings into the safest available thing, and concrete is safe. Forty years passed and nobody built the alternative.

---

### Sources

[^1]: Invest Kerala Global Summit outcomes, reported from Kerala Industries Department figures: ["Kerala hits century in turning investment summit projects into reality"](https://www.outlookindia.com/announcements/news-media-wire/kerala-hits-century-in-turning-investment-summit-projects-into-reality) (100 projects worth ₹35,111.75 crore taken off from 499 letters of intent, 276 of which had obtained land, a conversion rate the state put at 36.23%, announced by Industries Minister P. Rajeeve on 12 November 2025) and [Information and Public Relations Department, Government of Kerala, "Invest Kerala Global Summit 1st Anniversary"](https://prd.kerala.gov.in/ml/node/338849) (449 expressions of interest; 115 projects in which construction had started, with an investment of ₹54,908.22 crore), with the government's table as on 18 February 2026 reprinted in ["Invest Kerala Global Summit Delivers: ₹54,908 Crore Projects Take Off in One Year"](https://www.newsexperts.in/invest-kerala-global-summit-delivers-%E2%82%B954908-crore-projects-take-off-in-one-year/) (285 projects with land, 115 of them under construction). See also ["Work begins on 86 projects worth ₹31,429 cr"](https://www.onmanorama.com/news/business/2025/07/04/kerala-investment-projects-progress-invest-kerala-global-summit.html) for the mid-2025 position, and the [Kerala State Planning Board, Economic Review 2025 (PDF)](https://spb.kerala.gov.in/sites/default/files/2026-01/ER%202025%20Volume1%20Eng%20final.pdf) for the industrial-policy context. Totals drift slightly between releases, which is why the figures here are given as reported rather than reconciled.

[^2]: [Economic Development Board Singapore](https://www.edb.gov.sg/): founding mandate. On matching firms for integration at Jurong Island: [Lim Hng Kiang, keynote, 25 September 2009](https://www.meti.gov.sg/newsroom/mr-lim-hng-kiang-at-the-jurong-island-reclamation-completion-ceremony--25-sep-09/). The institutional argument this section picks up is made at length in the third essay in this series.

[^3]: Kerala's gold consumption: [World Gold Council, "Jewellery demand and trade", India gold market series](https://www.gold.org/goldhub/research/jewellery-demand-and-trade-india-gold-market-series/17661), which records South India as roughly 40% of national jewellery demand and attributes the southern skew partly to Gulf remittance income; and ["Kerala leads as southern zone remains India's largest gold consumer"](https://www.theweek.in/news/biz-tech/2024/05/31/kerala-leads-as-southern-zone-remains-india-s-largest-gold-consumer-experts.html), which reports the World Gold Council's India chief executive saying that gold use per person is seen to be higher in Keralam.

[^4]: Non-resident deposits held by banks in Keralam: ₹3.24 lakh crore as at 31 March 2026, with Federal Bank holding ₹92,326.10 crore, State Bank of India ₹84,966.67 crore and South Indian Bank ₹31,947.18 crore, and most of the foreign-currency component originating in the Gulf states. Reported in ["Gulf Keralites drive surge in FCNR(B) deposits as NRI savings cross ₹3.24 lakh crore"](https://english.hindusthansamachar.in/Encyc/2026/7/18/Gulf-Keralites-remittances.php). The figure is bank-reported and consistent with the trajectory described a year earlier in ["Kerala sees record-breaking inflow from NRIs"](https://www.onmanorama.com/news/business/2025/05/29/nri-remittance-record-kerala-deposits.html), when the total was approaching ₹3 lakh crore.

[^5]: S. Irudaya Rajan, [Kerala Migration Survey 2023 (PDF)](https://iimad.org/wp-content/uploads/2024/06/KMS-2023-Report.pdf) (released June 2024; the published PDF is marked draft), International Institute of Migration and Development with the Gulati Institute of Finance and Taxation ([report to the Chief Minister, PDF](https://www.gift.res.in/wp-content/uploads/2024/07/GIFT-IIMAD-Report-to-CM.pdf)): remittances at 23.2% of net state domestic product in 2023 against 13.5% in 2018, a total of ₹2,16,893 crore against ₹85,092 crore, and 1.7 times the state's own revenue receipts.

[^6]: On the Swiss institutions that hold conservative capital on long horizons, [Basel Area](https://www.baselarea.swiss/) for the life sciences cluster, [ETH Zurich spin-off statistics](https://ethz.ch/en/news-and-events/eth-news/news/2024/01/press-release-eth-spin-offs-an-annual-record-set-in-2023.html), and the [Swiss Bankers Association](https://www.swissbanking.ch/en/) on wealth-management and fiduciary practice. The claim in this section is about institutional form rather than any single statistic.

[^7]: Yozma: the Israeli government's $100 million commitment in 1993, split between $20 million in a fund of its own and $80 million taken as 40% stakes alongside private domestic and foreign managers, with each private partner holding an option to buy the state's share back over five years; eleven funds seeded by 1996 holding $263 million, 217 investments and 122 exits. See [Yozma programme structure and outcomes](https://en.wikipedia.org/wiki/Yozma), drawing on Gil Avnimelech, ["VC Policy: Yozma Program 15-Years Perspective"](https://www.researchgate.net/publication/228921726_VC_Policy_Yozma_Program_15-Years_perspective) (2009), and [OECD, "Venture Capital Policies in Israel" (PDF)](https://www.oecd.org/content/dam/oecd/en/publications/reports/2003/01/venture-capital-policies-in-israel_g17a1564/585780028400.pdf) (2003).

[^8]: On the tax treatment of non-resident investment, [Income Tax Department, Government of India](https://incometaxindia.gov.in/). Indian law does not treat non-resident capital in operating companies and non-resident capital in real estate identically; the applicable treatment depends on the investment vehicle, the investor's residence status, the class of asset, whether the income is characterised as capital gain, rent, dividend or business income, and how the transaction is structured. The essay therefore argues for a deliberate and legible differential rather than asserting that current treatment is uniform. Establishing the precise effective-rate difference for any given pair of investments is a question for a tax adviser and not something this essay claims to have done. On visa provisions and on the latitude states hold to depart from central industrial-policy norms, [PRS Legislative Research](https://prsindia.org/) and the [Department for Promotion of Industry and Internal Trade](https://dpiit.gov.in/). These are general pointers rather than specific findings, and the claims they support are correspondingly general.

[^9]: [Temasek, "The roots of Temasek"](https://www.temasek.com.sg/en/who-we-are/the-roots-of-temasek): incorporated in 1974 under the Singapore Companies Act to own and commercially manage an initial portfolio of 35 companies previously held directly by the Minister for Finance, freeing the ministry to concentrate on policy and regulation. GIC was established separately in 1981 to manage the government's accumulated reserves, which are invested outside Singapore. Both are wholly owned by the Minister for Finance and run to distinct mandates; Temasek, not GIC, is the domestic industrial counterpart to the EDB. On the EDB's industrial lending passing to DBS in 1968: [DBS, "A bank is born" (S Dhanabalan)](https://www.dbs.com/media/features/a-bank-is-born.page).

[^10]: EY, ["Sovereign wealth funds power the VC investment wave in GenAI"](https://www.ey.com/en_ie/insights/ai/sovereign-funds-drive-genai-vc-investment-surge), 12 December 2025: "Between January and September 2025, SWFs were involved in AI venture transactions totalling $46 billion, almost half the total for the year to date." Total venture investment in generative AI reached $87 billion over the first eleven months of 2025, against $52.5 billion for all of 2024.

[^11]: Cochin International Airport Limited, incorporated March 1994 as India's first greenfield airport built with private participation and the origin of the Indian airport public-private-partnership model. Paid-up equity of ₹78.93 crore at 31 March 2001: Government of Kerala and state public-sector undertakings ₹32.45 crore, airport service providers ₹21.25 crore, NRI directors and relatives ₹14.14 crore, and 6,160 resident and 3,880 non-resident small shareholders ₹4.48 crore and ₹5.40 crore, the non-resident holders spread across thirty countries (Varkkey and Raghuram, from CIAL records). Shares of ₹10 par value. About 19,000 investors from 28 countries, and at 31 March 2025 the Government of Kerala held 33.38 per cent and government companies 8.46 per cent (CIAL RTI information manual). Dividends paid every year from FY2003-04 to FY2019-20, none for FY2020-21 or FY2021-22, and resumed afterwards: [CIAL, "CIAL Hands Over Rs. 79.82 Crore Gross Dividend to State Government"](https://cial.aero/news-Updates/DividentHandover) (50 per cent for FY2024-25, approved at the 27 September 2025 AGM). Net profit ₹156 crore on turnover ₹553.42 crore in FY2017-18. Site acquisition covered about 1,300 acres and resettled 872 families under a rehabilitation package of market-value compensation, six cents of land per family, and direct or indirect employment for one member of each family that lost both house and land. [Cochin International Airport Limited](https://www.cial.aero/); Biju Varkey and G. Raghuram, ["Public Private Partnership in Airport Development: Governance and Risk Management Implications from Cochin International Airport Ltd" (PDF)](https://www.iima.ac.in/sites/default/files/rnpfiles/2001-10-05BijuVarkey-GRaghuram.pdf), Indian Institute of Management Ahmedabad.

[^12]: Vivek Wadhwa, AnnaLee Saxenian, D P K Muthukumaraswamy and M H Bala Subrahmanya, ["The Indian Returnee Paradox: Why the Diaspora Is No Longer Driving Startup Success"](https://www.orfonline.org/research/the-indian-returnee-paradox-why-the-diaspora-is-no-longer-driving-startup-success), Observer Research Foundation, 9 February 2026. 596 high-tech startups founded in India between 2016 and 2023, with an analytical sample of 521 firms having complete founder histories. Returnee-founded firms lead on medians (capital raised $441,000 against $113,000; annual revenue $92,500 against $43,700); domestic-founded firms lead on means (revenue $5.39 million against $0.68 million) and hold all three of the sample's billion-dollar valuations. The paper records that "in no sector do returnee-founded companies outnumber domestic ones". The section title in this essay takes the paper's own term.

[^13]: [Economic Development Board Singapore, "Pioneer Certificate Incentive and Development and Expansion Incentive" (PDF)](https://www.edb.gov.sg/content/dam/edb-zh/downloads/brochures/PC%20and%20DEI%20Brochure.pdf), version 1 January 2022. A corporate tax exemption or a concessionary rate of 5% or 10% on qualifying income; an incentive period "limited to five years", extension "subject to the company's commitment to undertake further expansion plans"; qualification on "the employment created (including skills, expertise and seniority), total business expenditure which generates spin-off to the economy, as well as commitment to growing the capabilities (e.g. technology, skillsets, knowhow) in Singapore", with fixed asset investment required additionally of manufacturing projects; award "subject to the company implementing its plans"; and, under the Economic Expansion Incentives (Relief from Income Tax) Act, an obligation to "submit regular progress reports to the EDB for the evaluation of performance", with breach exposing the company to "the potential revocation of the incentive and recovery of any associated benefits".

[^14]: Ministry of Statistics and Programme Implementation, [Periodic Labour Force Survey Annual Report 2023-24](https://mospi.gov.in/publication/annual-report-periodic-labour-force-survey-plfs-2023-24), Appendix A, Table (8), persons aged 15 and above, rural and urban: in Kerala, 14.2 per cent of women are graduates and 5.4 per cent hold a postgraduate degree or above, 19.6 per cent in all, against 11.5 and 3.3 per cent of men, 14.8 per cent.