Applying land value capture to public investments

Donald Shoup, the famous urban planner, posed this great question that goes to the heart of infrastructure finance: “How do you finance a project that has a return much higher than its cost, but nobody wants to pay the upfront cost?”

The answer lies in land value capture (LVC), a topic that I have blogged on numerous occasions (see here and here). Specifically, two instruments, betterment levy on existing developed properties, and impact fee on those which are undergoing development, have the potential to raise significant revenues. 

In theory, LVC spans the spectrum from appropriating directly through ownership of a part of the developed land (through the likes of land pooling) and indirectly through one-time and/or recurring charge (betterment levy or impact fees). 

The former, the best example is Gujarat’s historical town planning schemes (TPS), is perhaps the most ideal approach, insofar as it cleanly appropriates a share of the land itself (typically 40-60% of the total land). The problem is that it is an engagement-intensive activity requiring both high credibility and state capability, both deficient across state and local governments. The best illustration of this is Andhra Pradesh’s struggling experiment to develop its capital at Amaravati by pooling land through a TPS. 

TPS entails that the state takes over the lands, develops infrastructure, and leaves developed plots for the landowners, all this within some defined period of time. The combination of the political economy of real estate and weak state capability creates a time inconsistency problem that erodes the credibility associated with such schemes. 

It is also for this reason that, despite numerous efforts over the last three decades, there are still very few instances of genuinely successful slum housing redevelopment schemes in Indian cities. And there are numerous examples of badly delayed and poorly executed such projects. 

In the circumstances, the best bet with LVC in the context of countries like India is to collect a share of the incremental value as a charge on some urban planning instrument. It is the instrument with the fewest implementation difficulties. 

Consider the numerous ongoing infrastructure projects across the country, ranging from national highways and expressways to metro railway projects, airports and ports, data centres and industrial clusters, and generally any large investment that leads to the development of an area. In all these cases, there’s an immediate spike in property prices, which is almost completely appropriated privately. If some share of the property valuation increase can be appropriated through the LVC instruments, then it can be securitised to fund the projects themselves, at least a significant share of the cost. 

Illustrative examples will include the nodes in the Delhi-Mumbai industrial corridor, stations in the Mumbai-Ahmedabad high-speed rail and the Regional Rapid Transit System(RRTS), vicinity of Navi Mumbai Airport, the corridor abutting Bharatmala pariyojana, the vicinity of several new airports like Jewar and Bhogapuram, and so on. None of these has any specific betterment levy or impact fee charged on properties in their influence zones. These are all massive missed opportunities and large revenues foregone. 

The formally reported claims of LVC revenues from all these are either in the form of unlocking value through sales or lease of government properties, or, in a few cases, they include the sale of lands acquired through TPS. 

In every one of these projects, the intent of capturing land-value gain from the project announcement onwards exists somewhere on paper. In practice, though, the only meaningful capture has come from in-kind land pooling, lease revenue/premia from sale of publicly-owned land, and a few development/impact-fee tweaks. The second one (lease revenue or premiums from sales of public lands) is pure monetisation and should not be confused with LVC. 

A true betterment levy or impact fee, in the form of a charge on existing private property in the influence zone, payable from the first transaction after project announcement, has effectively not been deployed at scale in any of these projects. Therefore, the aggregated, publicly disclosed revenue from "pure" LVC instruments (betterment levy + impact fees, excluding land sales/leases/concession fees/toll/UDF) is very small, even negligible, relative to project cost in every case here. The genuine “capture” that has occurred is overwhelmingly through public-land monetisation or in-kind retention of pooled land, where the former leaves the increment on third-party private property untouched.

Perhaps the only true significant examples in India are the impact fees in a corridor of 1 km abutting the Hyderabad Outer Ring Road (ORR) (though its success in terms of realisation is a matter of debate), and the 1% metro cess on property purchases in Mumbai to fund the metro railway project. A promising effort with the Navi Mumbai Airport Notified Area (NAINA) was that of the 50% betterment levy on the increased property value notified in 2013, which was immediately reduced to a notional 0.05% following political opposition. 

The political economy is an important consideration. Every such big public investment immediately invites large speculative purchases, especially by politically and bureaucratically connected individuals. They are an important sink for the large volumes of black money that slosh around the local economy. Vast sums are invested in the name of third parties (benamis) who front the local connected and influential. Throwing sand in the wheels of this speculative activity, much less appropriating a part of the windfall increments as public revenues, will be resisted to the hilt by the aggrieved interests. 

It is therefore important to be careful while designing the LVC instrument. It should depend on the nature of the infrastructure project. The objective should be to capture the increment in land value that the incumbent land owner enjoys from the public investment by charging the first transaction after the project announcement (which triggers the spike). Ideally, it should also try to avoid penalising the buyer, who has internalised and paid for the land increment. However, if the increment is ongoing, like with a corridor that undergoes continuous development, there’s a case for also capturing the increment flows over time through an appropriate instrument. 

It is also important to differentiate between brownfield (betterment levy) and greenfield (impact fees) areas. 

Accordingly, the most appropriate betterment levy on existing (developed or brownfield) property owners (who are not likely to transact) comes from a cess on their property taxes. In theory, if the property taxes are indexed to the market value, there would be no need for a separate cess. However, this is never the case since the property prices are generally linked to the guidance or ready reckoner rates fixed by the local authorities, which are always lower than the market value, and the wedge widens after local property booms. 

For properties undergoing development, the ideal option would be to charge from those selling the property. However, this is complicated and can be captured, and that too only partially (given the aforesaid wedge between market value and guidance value) from capital gains taxes. Administering this is infeasible (imagine two kinds of capital gains tax for land, where there’s land value increment from public investment and the rest). In the circumstances, for such properties, the options are to levy a charge on either the registration fees, the change of land use fees, the layout development fees or the building permission fees. The assumption would be that the charge shapes market expectations and cascades across the property market and is internalised by everyone in the property transactions chain. 

It may be useful to keep in mind some principles while designing the LVC instrument. In theory, it can be charged on the first transaction for a property in the influence zone after the project announcement as a significant windfall fee, or on all transactions for a period (that reflects the time for the project to realise its benefits). A simple and practical strategy would be to impose a one-time cess (on the property value) on the first transaction on the property. A flat levy/fee, while administratively appealing, does not address the diversity in the types of development projects and their widely varying land value impacts. 

An analysis of the projects mentioned above points to a few problems, even where some LVC framework exists. For one, the first transaction after the project announcement rule is not legally embedded. Most Indian betterment-levy provisions trigger at project completion (e.g., Section 37 of the DD Act 1957; Section 66 of the Bombay/Gujarat TP Act 1954/1976). By then, the original landowner has typically already transacted, and the increment has been captured privately. This is not only economically inefficient but also invites public discontent and political opposition. 

Policy makers create confusion and skirt accountability by conflating land monetisation with value capture, as is the case with CIDCO, YEIDA, or DSIRDA. Alternatively, user-side instruments like User Development Fees (UDF) in airports, and tolls and concession fees on highways, also get lumped as LVC. 

As the aforesaid discussion shows, designing an efficient (in terms of capturing value from the incumbent land owner at the time of project announcement) and least distortionary (in terms of not adding unreasonably to property development costs) LVC instrument requires careful thought and design. 

As an illustration, I asked Claude to estimate the likely LVC revenues from the development of the newly announced data centre projects in the Visakhapatnam area in Andhra Pradesh, and other investments in the state.

It estimates a realisation of Rs 700-1500 Cr annually from the application of LVC instruments to the Google AI hub catchment, Bhogapuram airport zone, defence clusters, and metro corridors. Even half of this would be greater than the combined property tax revenue collection of all the state’s 117 municipalities. 

As an implementation strategy, it may be useful to align incentives and design a revenue-sharing mechanism among all public stakeholders, especially the local government. So, for example, instead of all the LVC revenues from a national highway or other central government project going to the project development entity, a part should be shared with the local government. This creates the required local incentives to enable effective adoption of any LVC scheme. 

In terms of policy mandates, it is time to mandate that all large public investment projects be accompanied by the adoption of LVC instruments by the local governments. This would entail the following: (a) notification of a LVC influence zone around the project, (b) notification of the appropriate LVC instrument and the LVC rate, (c) enabling a statutory framework and a physical mechanism for its collection, and (d) a transparent and public accounting of the realisations. 

As a first step, all states should be encouraged to adopt an LVC policy, borrowing from MoHUA’s LVC policy. It may be useful for the Government of India to build on this policy and formulate a model document on both betterment levy and impact fees that can be issued as administrative orders by the state governments. To incentivise this, the Ministry of Finance should consider making this a mandatory requirement to access the 50-year interest-free capex loans being given to state governments under the Special Assistance Scheme for Capital Investment (SASCI).

The model documents and guidance on each of the betterment levy and impact fee instrument options may be useful to ensure that LVC instruments are adopted in letter and spirit. The documents should contain the detailed implementation design of the instrument, and the same should be incorporated into the financial closure, state support, and all other agreements associated with the project, and should be institutionally coded into the system along with the project announcement. All project appraisals should have tightly defined and enforceable requirements on LVC. 

Finally, the central government's support for any project should be made contingent on the realisation of the LVC proceeds in both letter and spirit. This should become an integral part of the financing culture of all large projects in India. This is one of the very few big bang low-hanging fruits in the mobilisation of public revenues, especially for fiscally strapped urban local bodies.

Weekend reading links

1. Samsung's spectacular turnaround, from being in the doldrums as late as in 2024.

This month its market value, which has soared by 400% in the past year, hit $1trn for the first time, propelled by furious spending on artificial-intelligence infrastructure. In the first quarter of 2026 its operating profit rose to 57trn won ($38bn), more than eight times as much as a year before. Analysts expect profits to keep rising at a blistering pace, thanks in particular to the seemingly insatiable demand for its advanced memory chips... Semiconductors accounted for 61% of sales and 94% of operating profits in the first quarter. It is one of just three firms capable of making at scale the memory chips needed for ai, alongside SK Hynix, a South Korean rival, and Micron, an American one. The number of memory chips Samsung sold in the first quarter was up by about 20% on the preceding three months, but the average selling price rose by 90%.

2. The Economist argues that the calm in global oil markets despite a supply shock of some 14 mb per day can be traced to the 4 mbpd of increased exports by the US and the 4.5 mbpd of reduction in Chinese imports, coupled with rationing across countries. 

3. As the rise of AI threatens white-collar jobs and increases the returns to capital, The Economist proposes some measures to redistribute the gains. 

If employment falls, income that once went to workers is likely to show up as high profits in AI firms, chipmakers, data centres or elsewhere in the supply chain. Clever tax reforms, such as levies on corporate profits that are above a normal return on capital, on land and on natural resources, could capture these rents. The case for inheritance taxes to prevent the entrenchment of a capital-owning elite looks even stronger than before. At the same time governments could help workers adjust. Public wage-insurance, which smooths out falls in income after job losses, can help workers find better opportunities (and so can eventually pay for itself). Denmark’s active labour-market policies, in which the state helps people find and train for new occupations, have been proved to cut spells in unemployment... 

A last set of radical ideas, such as the partial nationalisation of ai firms. This week a South Korean presidential adviser floated a citizens’ “dividend” from AI businesses, sending the local stockmarket down by 5%, before backtracking. In America politicians murmur about giving citizens shares in AI companies via “Trump accounts”. In economic terms there is little difference between a well-designed tax system and a government stake in the private sector—and countries without AI giants will have to rely on taxes rather than seizing shares in foreign companies. But America may find that some public ownership is the best way to make the social upside from the technology transparent.

4. The changing face of Reliance Industries.

5. America's remarkable productivity growth miracle since the pandemic (it predates the AI boom).

Now with AI coming of age, the productivity spurt is likely to continue. 

6. Egypt may well be the leader in land monetisation to promote economic growth in any substantial form.  (HT: Adam Tooze)

Since 2015, Egypt has increasingly contributed public land as equity, while foreign investors provide capital, development expertise, and project execution. Once a project is completed, revenues are shared according to pre-agreed division... in 2023, Egypt appointed the bank’s International Finance Corporation as its advisor for the asset monetization program, leveraging its experience supporting emerging markets. While land monetization has been tried elsewhere, Egypt’s projects are among the largest... For the Ras el-Hekma development on the country’s North Coast, Egypt contributed approximately 40,600 acres of state-owned land along the Mediterranean. The UAE (via its ADQ sovereign wealth fund) committed roughly $35 billion, the largest foreign direct investment in Egyptian history. Egypt received immediate foreign currency inflowsfor the land, a 35 percent stake in the project, and long-term profit participation... A similar project, also on the North Coast, is Alam el-Aroum/Samla near Marsa Matrouh. The Qatar Investment Authority-linked Qatari Diar is investing almost $30 billion, which includes a $3.5 billion upfront land payment for some 20 million square meters and $26 billion in development investments. A revenue share for Egypt (15 percent after cost recovery) is part of the deal.

Another arrangement is in place for Egypt’s New Administrative Capital (NAC). About thirty miles east of Cairo, the NAC is designed as the government seat and a commercial hub; reports estimate total development costs of up to $58 billion, including infrastructure and governmental, commercial, and residential districts. Foreign direct investment plays a role in specific sub‑components like the Central Business District (CBD) and future free-trade‑zone ventures. Chinese banks led by the Industrial and Commercial Bank of China provided 85 percent of funding for twenty towers in the CBD. The China State Construction Engineering Corporation developed the CBD; Gulf investors (such as the United Arab Emirates’ DP World) developed commercial parcels. The state monetized land incrementally for the NAC, and parcel sales financed development, without increasing Egypt’s debt.

All this appears very impressive. While the article paints a picture of success, it would be interesting to peel layers and scrutinise this. 

7. This is anecdotal, but tells a lot about why India lags in manufacturing.

8. Spain is undertaking an ambitious experiment in immigration.

Since 2022, Spain’s foreign-born population has surged by an annual average of 665,000, the equivalent of adding a city the size of Málaga each year. Last year the country accounted for roughly one-third of the total increase in the EU’s immigrant population, according to the Rockwool Foundation, a Berlin think-tank. Supporters say the influx has given Spain’s ageing society a much-needed burst of economic vigour. Critics call it a poorly planned strategy that is straining the country’s infrastructure and creating new social tensions... In less than a quarter of a century, Spain’s foreign-born population has gone from one in 20 residents to almost one in five, a higher proportion than even the US... Last month the Spanish government’s most contentious immigration move to date took effect — a sweeping amnesty giving at least half a million people the chance to gain residency and work permits and move out of the shadow economy... applicants must prove he was in Spain before January 1 this year and has been there for five consecutive months.

The country is already experiencing an acute housing shortage, has among the highest youth unemployment rates at over 10%, and the anti-immigrant Vox party is running third. The final outcome on the rapid rise in immigration is yet to be known. 

9. Ruchir Sharma points to an area where China trails badly behind the US, the negligible role of the renminbi as an international currency. 

With a 17 per cent share of global GDP, but only 2 per cent of central bank reserves, China is trailing 30 to 40 years behind previous superpowers at a similar stage of their ascents... Britain at its peak accounted for 40 per cent of trade, but 60 per cent of trade payments were in sterling. China by contrast has a leading 15 per cent share of global trade, but only 2 per cent of trade bills are invoiced in renminbi...

China will remain an incomplete superpower until it can match this financial firepower. For decades, it has kept its financial system more tightly sealed than any other major nation. It now ranks in the bottom fifth of nations by international investment position, which captures the level of foreign ownership in the domestic market. Foreigners own less than 5 per cent of the stocks and bonds in China, one-fifth the level in the US. Its home market is something of a local prison. Beijing has generated economic growth with heavy infusions of government money, corralled at home by capital controls. Its money supply has multiplied sixfold since 1980 to 230 per cent of GDP, among the highest in the world. This liquidity sloshes around inside the walled economy, much of it in the domestic debt market, battered lately by a property bust. Beijing is wary of easing controls, lest it unleash capital flight.

10. The US has been the biggest oil export beneficiary of the Iran war.

11. The booming market for whey protein.

Prices for whey protein isolate have soared fivefold to €28,000 a tonne since 2023, outstripping cheese and butter prices by more than four times as producers struggle to keep pace with a booming protein market... Thirty years ago whey was primarily used in animal feed or spread on farmland as a fertiliser. Whey is the liquid separated from curds during the cheesemaking process. Now, dairy groups upgrade the liquid by filtration to produce whey protein isolate, a valuable ingredient for use in sports supplements as well as groceries such as yoghurt, bread and fizzy drinks, as weight-loss drugs and the protein megatrend propel demand. “This is reshaping the economics of dairy,” said Jose Saiz, analyst at Expana, a commodity market information service. “It used to be a product with no value . . . now cheese could become the byproduct of whey production.”

12. The UK's experience of small altnets driving broadband expansion and lowering prices has come with excesses. 

Over the past decade, more than £31bn has been raised to roll out full fibre broadband across the UK, with private equity giants, including Macquarie and KKR, backing upstart challengers — or “altnets” — in the sector. Yet despite the initial investor optimism that the “altnets” could snatch customers away from industry leaders, the firms have been dogged by high build costs, lower-than-expected customer uptake and a sharp response from the country’s largest provider — BT’s Openreach — to deploy its own fibre infrastructure. The destruction of shareholder value has been brutal. At the last count, the “altnets” posted losses of more than £1.5bn in 2024, according to Enders Analysis. After accumulating some £9bn of net debt as of 2025, some companies have already been placed into administration, while others have fallen into the hands of lenders... 

Prior to the “altnet boom”, which accelerated from 2021 onwards, only 24 per cent of Britons had access to full fibre internet, according to Ofcom. Fast forward five years, that number is now more than 78 per cent, with altnets serving almost 20mn homes with full fibre, according to Assembly Research. By next year, Ofcom estimates 95 per cent of UK homes will have full fibre, putting it in line with Europe’s leaders, including Spain and Luxembourg, where regulators encouraged full fibre rollout earlier than the UK. The “altnets” have also forced BT’s telecoms infrastructure provider Openreach and Virgin Media O2 to expand their own fibre networks, giving consumers a choice between providers who are now forced to think faster and harder about how to up their game. This added competition has meant prices have fallen, with the average monthly real-terms list price of UK full fibre broadband falling from £62.38 in September 2021 to £43.46 in September 2025, a drop of 30 per cent... altnets are taking close to 1mn [customer] lines from BT annually.

13. Alan Beattie argues that EU has initiatied several trade measures against China, though their implementation has been weak, primarily due to internal opposition within the bloc. 

One of the recent instruments with a bit more bite is the Foreign Subsidies Regulation (FSR), launched in 2023 and designed to level the playing field against state-backed Chinese companies bidding for contracts or producing and selling in the EU. It’s notable that it gives a lot of investigatory and decision-making powers to the Commission, specifically to the internal market and competition directorates, which are used to having autonomous powers. By starting investigations into their operations, the FSR has managed to get some Chinese companies to pull out of public procurement bids. But when the EU tries to use its internal market powers to investigate supposedly subsidised Chinese businesses trading in the single market, it becomes clear just how strongly Beijing is willing to resist. Last week, in an investigation dating to 2024 into the Chinese cargo scanner company Nuctech, Beijing cited new supply chain security laws to forbid Chinese companies to comply with requests for information, saying Brussels’ extraterritorial reach was illegitimate.

14. Arvind Subramanian and Devesh Kapur on the contrasting tales of India and China in monetisation of lands by urban local bodies. 

China’s land revenues increased from less than 1 per cent of GDP to more than 10 per cent at its peak. In contrast, India’s revenues have stagnated at about 1 per cent of GDP through the entire growth phase. Put differently, the Chinese government’s collections from land revenue for every urban resident that was available for spending were about 15 times more than India’s in 1999; at its peak in 2020, this multiple increased to 225.

15. As AI capex surges, there are growing doubts about whether it will generate the returns required to justify it.

For each of these hyperscalers, I collected the consensus estimates of analysts for the capital expenditures and revenues between 2025 and 2030. In these five years, capital investments are expected to rise by 20 per cent a year, a growth rate never seen before in this industry. Meanwhile, revenues are expected to grow 15 per cent annually. If we make the heroic assumption that there are no costs, then the additional revenue is the profit these companies are expected to make from their additional investments in AI data centres. Yet, even under these extremely optimistic assumptions, I calculate the implied return on investment is highly negative for all of them except Amazon. These numbers show that if the hyperscalers continue on the current trajectory, the AI boom will become a story of one of the largest destructions of shareholder value in history... If the hyperscalers want to generate, say, a 10 per cent return on investment, they would have to find an additional $2tn to $5tn in revenue a year. A tall order for a group of companies that currently generates revenues of just $1.5tn per year.

See this comparison with the technology, media and telecom (TMT) bubbles of the late nineties.

In 2025, US businesses invested almost $1.5tn in IT equipment and software. At the peak of the TMT bubble, it was $466bn or $829bn when adjusted for inflation. Indeed, the US economy is growing solely because of the tech boom. I calculate that over the past four quarters, 93 per cent of US GDP growth was explained by tech investments. Even at the peak of the TMT bubble, it barely reached 60 per cent.

16. Tim Harford points to evidence that retailers jack up prices in response to shocks much faster than they bring prices down, and that retailers make their money not so much during the upcycle than in the downward phase.

Johannes Brinkmann and Nikhil Datta of the University of Warwick recently published an analysis of the impact on petrol and diesel prices of the oil price shock in 2022, following Russia’s onslaught in Ukraine. They found that in the UK, retailer margins compressed: the wholesale price of diesel rose by 39 pence per litre, while retail prices only rose 16 pence. This is the opposite behaviour to that predicted by the greedflation hypothesis. A natural explanation of this price compression is that retailers feel under more intense scrutiny when prices are rising. Brinkmann and Datta show that searches on the petrolprices.com website increased dramatically when prices did — and that areas where such searches were more common were also areas where the price compression was more intense. 

Brinkmann and Datta’s analysis is merely the latest in a long tradition of research describing “rocket and feather” pricing at the pump — capturing the idea that pump prices neither faithfully track the ups and downs of the crude oil market, nor exaggerate them — instead, they shoot up like a rocket but drift down again like a feather. What is more, the quick surge upward reaches prices less lofty than one would expect; it is during the slow descent that retailers make their money. Fifteen years ago, Matthew S Lewis and Howard Marvel noted that customers spent more effort searching when prices were rising, even though there was little benefit to that search, since most retailers were charging similar prices. When pump prices were falling, there was more variability from forecourt to forecourt and a higher return to shopping around, but most customers did not bother, feeling content that prices were moving in the right direction.

17.  Preference shares are a complicated instrument.

The “preference” investors receive — usually a slightly higher dividend — comes at the expense of voting rights. Preference shares therefore arguably resemble the worst of both worlds of debt and equity. Like bonds, they do not offer any influence over the company’s decision-making. Like ordinary shares, they do not come with a contractual claim to annual payouts as dividends are subject to management discretion. Such non-voting shares have been a prominent feature in corporate Germany. Four of the 40 blue-chips in the Dax have used them to establish substantive two-tier share structures for years: Volkswagen, Porsche and BMW, the three auto giants, and glue and detergent maker Henkel. All are dominated by controlling families who hold a tight grip over the firm. 

Two further Dax companies — Merck and Fresenius — are listed as “partnerships limited by shares”, a German legal structure known by its acronym KGaA, that makes some shareholders more equal than others through other means. Deutsche Bank listed its asset management arm DWS as a KGaA in 2018, too, warning at the time that this structure could dent its valuation. Yet the club of German blue-chips with differential voting rights will soon become smaller after BMW shareholders voted to abolish preference shares last week at its annual meeting. The group’s 54.7mn preference shares, representing 10 per cent of BMW’s equity, will soon be swapped into ordinary shares with voting rights... In the US, dual-class share structures have become increasingly popular as fast-growing tech groups want to tap public markets while keeping outsized voting power for insiders. Elon Musk’s SpaceX even wants to grant its CEO 10 times as much voting power as external investors.

Sample this about SpaceX  

Elon Musk's special class of shares currently gives him control of 85 per cent of the voting power at SpaceX.

Some thoughts on sustaining high growth rates in India

The flight of foreign portfolio investors since October 2024, coupled with the declining net FDI, has sparked a debate on what should be done to attract and retain foreign capital to India. This has assumed greater significance given the stagnation in domestic savings at about 30% of GDP.

It is an accounting reality that if the economy has to sustain high growth rates, it must have the capital to support the high investment rates required. Domestic savings must therefore be supplemented with foreign capital. But how much can foreign capital contribute?

A recent op-ed argued that to achieve the Vikasit Bharat 2047 goal, which necessitates a 9% annual growth over the next two decades, India must strive to attract foreign capital in the range of 15% of GDP.

This may be an opportunity to step back and assess the realistic envelope of foreign capital that India can target. As an analytical framing, this would entail examining both the demand and supply sides for the different types of foreign capital available, India’s track record in attracting them, and then making a judgment.

What are good comparators of countries having successfully attracted significant foreign capital? What is the envelope of foreign capital that is looking to invest in markets like India? What’s the level of capital that India can absorb without engendering too many distortions? Based on all these, what is the realistic foreign capital target for India? 

India’s net FDI inflows from foreign investors as a % of GDP have shown a continuous decline since 2020. 

Taken together, net FPI and FDI as a share of GDP have never exceeded 4%, have been declining since the pandemic, and have been negative over the last two years. (All graphics below generated using ClaudeAI). 

The highest FDI share of GDP was 3.6% in 2008, which has been more of an outlier since it has struggled to cross 2% in recent years. 

One important thing to keep in mind is that none of the bigger economies have managed to sustain FDI beyond 3% of GDP for long periods. India’s sustained 10-year average has been 1.5% of GDP. Even sustaining it at 2% will be a challenge, leave aside 3% and above. 

It is to be noted that the northeast Asian economies did not grow by importing foreign capital. South Korea grew at 8-9% for three decades with sustained FDI of about 0.7% of GDP; Taiwan, with 0.8%; Japan, with 0.2%; China, in its 1990-2010 peak-growth period, averaged ~3%. All four built investment rates of 35-50% of GDP through extraordinarily high domestic savings. The important takeaway is that, for sustained growth in a large economy, foreign capital has historically been a marginal supplement, and high growth has been delivered by domestic savings. Another important factor, which we shall discuss in detail later, has been their high investment efficiency.

Bringing together all types of foreign capital, the graphic below shows a realistic envelope of about 4.2% of GDP. Even this may be bordering on the optimistic, given India’s track record and global economic headwinds. 

In fact, reversing the recent trend of declining net FDI and FPI, and reaching even a 2% of GDP target for foreign capital would also be challenging. In the circumstances, a 6.2% to 6.5% target for GDP growth rate on a sustainable basis would be an achievement. Anything beyond that will require substantial breakthroughs. 

As a final word on foreign capital, it can be said that if India is to sustain high growth rates, it will surely have to attract foreign capital. But that will remain a marginal contributor. Instead, the heavy lifting will have to come by way of domestic capital, and especially by improving the efficiency of its utilisation. India must significantly increase domestic savings and also improve its capital allocation efficiency.

Since increasing domestic savings is itself a measure of broadbased economic growth and therefore endogenous (a point I have raised on several blog posts), capital allocation efficiency for any given investment rate becomes an important lever for sustaining high economic growth. The Incremental Capital Output Ratio (ICOR) measures the investment required to generate an additional unit of output, with a lower number indicating higher capital allocation efficiency. 

India's ICOR has been in the 4 to 5 range since the late nineties, and has never gone below 4 on a sustained five-year period. It briefly fell below 4 during the high-growth phase of mid-2000s. Further, in recent years, it has risen to 5-5.3. 

How does this compare with the East Asian economies in their high growth phases? 

Each of these countries went through three to four distinct ICOR phases as it transitioned from labour-intensive take-off, through capital deepening, to maturity (or crisis). The lowest ICORs occurred during the take-off phase in Taiwan, Korea and Japan — and during China's pre-2007 reform years. The northeast Asian growth miracles (Korea, Taiwan, and Japan in the 1960s-80s) sustained ICORs of 2.6 to 3.7 through their fastest decades. China repeated the feat in the 1980s and 1990s with ICORs of 3.6-4.1, but its capital efficiency has deteriorated sharply since 2007, with ICOR roughly tripling from 3 to nearly 8. Investment efficiency deteriorated in every country either in the run-up to the 1997 Asian Financial Crisis or, in China's case, after the 2008 global stimulus. 

The scatter plot below maps every economy-period in the dataset onto a single chart. The the dashed green diagonals are ICOR isoclines - every point on a given diagonal has the same ICOR. The main finding is important - the Northeast Asian growth miracles operated in a narrow corridor of moderate investment rates and very high efficiency. India operates in a corridor of similar investment rates but lower efficiency. China sits at the extreme right - very high investment rates, with efficiency that was once strong but has badly deteriorated. 

India has never matched the East Asian miracle benchmark of ~3, and capital efficiency may well be the single largest constraint on its high growth ambitions. To put this in perspective, at India’s current ICOR of ~5, achieving 8% growth requires a 40% investment rate, well above India’s current ~33%, whereas at ICOR 3.5, the same growth would need only 28%. 

In other words, lowering the ICOR by one full point reduces the investment requirement for any growth target by 7-10 percentage points of GDP, a larger lever than raising domestic savings, and far larger than attracting foreign capital. Therefore, the path to faster Indian growth runs through pulling the cluster of Indian dots upward and leftward, i.e., raising growth without raising the investment rate. Unfortunately, it appears to be going in the opposite direction for now. 

So what drives lower ICOR?

In simple terms, it is about growth, which comes with lower investment and higher efficiency of capital conversion. 

For a start, on the inputs side, services and labour-intensive manufacturing have a lower ICOR than capital-intensive manufacturing and real estate-driven development. Also, higher human resource quality brings greater bang for buck from an incremental unit of investment. On the efficiency side, higher capacity utilisation, expeditious project completion, and factor market reforms - land acquisition, labour mobility, and credit allocation - contribute to lowering ICOR. Finally, institutional quality, involving predictable regulation, contract enforcement, and competitive product markets, all raise the marginal productivity of capital. 

These are also the main areas of reform that have been discussed ad infinitum. 

An important observation here is that we cannot overlook the reality that while it is essential to move up the value chain to high value manufacturing, it is virtually impossible to generate high growth from it without far higher investment rates. It underscores the criticality of labour intensive sectors like textiles and footwear in the sustainable economic growth of a country, especially a large one like India. This is also underscored by the importance of labour-intensive manufacturing in the high growth periods of the northeast Asian economies.

It is also important to note that China’s ICOR tripled in 15 years (from 3 to 9) as it shifted from labour-intensive manufacturing to capital-intensive infrastructure, property, and state-directed lending. Growth fell from 10% to 5% during the same period, exactly what the ICOR deterioration predicts.

The relevance of capital allocation efficiency also highlights the importance of human resource quality. This is even more so of if the services sector is to increase its role in sustaining high growth rates. 

Update on the AI spending boom

I blogged here on the emerging AI economy. This is an update on the trends in AI spending.

As the hyperscalers - Alphabet, Amazon, Meta, and Microsoft - ramp up their data centre investments to $720 bn in 2026, their age of bountiful cash flows is over. Three of them are expected to have at least a quarter of negative cash flows, and Alphabet will only just scrape above. 

At around 40% of their revenues this year, the cloud giants’ capital expenditures will surpass those of the oil industry during the shale boom in the 2010s and the telecoms industry during the dotcom bubble in the 1990s… Nowadays investors judge the success of these firms on the basis of concentrated revenue contracts stretching far into the future, rather than dispersed sales received today. Mostly these contracts involve selling computing capacity to model-makers like OpenAI and Anthropic, which are themselves incinerating vast piles of cash. Total future revenue agreements have risen to $2trn, from $730bn last year, at Amazon, Google, Microsoft and Oracle (Meta is a buyer, rather than a seller, of computing capacity)…

Since the start of last year the big five have raised $260bn from bond markets, a quarter of all such borrowing by listed American non-financial firms… Nearly a third of the haul from selling bonds this year is in currencies other than the dollar… Much larger obligations lurk off-balance-sheet. The biggest are $820bn of future payments to lease data centres yet to be built, up from $270bn a year ago. Commitments to spend money on other things, like packing their data centres with chips, have risen as fast. Amazon, Google, Meta and Oracle now disclose $680bn of such obligations. Other bills are tied to special-purpose vehicles: separate entities with their own balance-sheets. Last year one assembled to build Meta’s new data-centre in Louisiana issued the biggest single corporate bond in history.

This is an interesting summary of the dynamic driving this investment boom

The five firms (hyperscalers plus Oracle) have assumed the role of central planners, attempting to make the complex chain of returns on investment work across the AI economy: data centres are useless if businesses don’t find models worth paying for, which only happens if model-makers can raise enough capital to make them. In the process, the hyperscalers have sacrificed their own returns… Big tech has also liberally lent its creditworthiness across capital markets. Many firms that contract with the giants can take those contracts to the bank (literally) and raise more debt.

There are perhaps three overlapping risks - circular financing, vendor financing, and off-balance-sheet leverage - which are not a set of isolated deals but a single integrated capital structure spanning seven participant types. 

Claude generated this comprehensive diagram that captures the linkages across these seven layers.

A chipmaker (and hyperscaler) writes an equity cheque to an AI lab (purple, up the left side), the lab converts it into a purchase commitment back to that same chipmaker and the cloud landlords (orange, down the right side), the commitment becomes backlog that supports the vendors' valuations and their bond issuance (blue), and financiers fund the whole build through SPVs and private credit (amber). Money does not exit the system; it rotates.

Three properties convert a set of bilateral deals into a self-reinforcing system. First, circularity inflates apparent demand - taking equity from vendors and using it to support further borrowing has made the boom dependent on convoluted financial engineering (echoing the 1990s, when telecom-equipment makers like Lucent and Nortel advanced money to customers to buy equipment, only to face write-offs when bankruptcies hit). Second, a single weak node is load-bearing - if OpenAI cannot meet commitments, Oracle’s revenue, CoreWeave’s backlog, Nvidia’s and AMD’s sales, and the SPV lenders’ collateral all reprice together. Third, the risk has been pushed to where it is least visible (into private credit and SPVs) while equity markets have already capitalised the optimistic case. If demand does not materialise, no amount of financing ingenuity can address it, and given the amounts involved it is certain to have very large financial-market and economy-wide impacts, cascading through the balance sheets of corporates, financiers and households. 

The risks are concentrated and largely contingent — they crystallise only if AI demand disappoints, but if it does, they crystallise simultaneously across all seven categories because they share the same underlying bet. Everything rests on AI demand growing fast enough to justify >$1tn in lab commitments and ~$725bn in annual hyperscaler capex. If the revenue curve stays ahead of the cost curve, the circularity is just efficient capital allocation. If it doesn't, the same circularity becomes the risk transmission mechanism.

All this circularity is also generating its set of accounting distortions. Robin Wigglesworth points to the curious spike in the “other income” line of the hyperscaler's quarterly account statements, attributable to the changes in valuations of their investments in AI companies. 

Alphabet, for example, booked $37.7bn of “other income” in just the first three months of the year, accounting for over half the company’s net income over the period. Amazon reported “other income” of nearly $16bn in the first quarter, up from $2.7bn in the same period last year. That was nearly half its overall net income for the three months. Microsoft reported “only” $942mn of other income in the first three months of the year, but this line item has now made $7.2bn over the past nine months… what constitutes “other income” in this case: the ebb and (mostly) flow in the valuations of their sizeable private investments in companies like OpenAI and Anthropic…

Not only have private investments and increasingly engorged funding rounds become a meaningful driver of the hyperscalers’ aggregate earnings, but the money the hyperscalers have pumped into the likes of Anthropic and OpenAI has allowed the AI companies to sign huge computing deals with Alphabet’s Google Cloud, Microsoft’s Azure and Amazon Web Services… OpenAI and Anthropic now make up about half of the entire cloud computing order books at Oracle, Alphabet, Amazon and Microsoft.

A vendor invests equity in its own customer; the customer uses that equity to pre-commit purchases from the vendor; the vendor books the purchase commitment as backlog; the backlog supports the vendor’s own valuation and borrowing. 

Even though AI expenditures are building the plumbing for the next-generation economy, something is unsettling about the scale and pace of spending. More worrying is the concentration and circularity of the transactions. Is AI spending going to be the mother of all circular trade bubbles?