Can the demerger save Vedanta? | Uncovering the truth about sovereign loans

ELI5/TLDR

Two stories. First, Vedanta has finally been chopped into five listed pieces — aluminium, oil and gas, power, iron and steel, and a residual entity that holds the crown jewel Hindustan Zinc. The split is less about strategy and more about a London-based parent suffocating under its own debt. Second, when one country lends money to another, the money itself is rarely the point. Soft power, export contracts, resource security, and recycling trade surpluses do most of the talking. China didn’t invent the playbook — Japan ran it for 30 years — but China runs it bigger and with sharper contracts.

The Full Story

The Vedanta surgery

Anil Agarwal spent two decades stitching Vedanta together. He has spent the last two and a half years taking it apart. On May 1, Vedanta Limited formally split into five entities. By end of June, four of them will trade on Indian exchanges as separate listed companies. The fifth, the residual Vedanta Limited, keeps the stake in Hindustan Zinc and a few smaller bits.

The group is enormous. Roughly half of India’s primary aluminium. The largest integrated zinc and silver business through Hindustan Zinc. One of the largest private upstream oil and gas players through Cairn India. Copper, iron ore, steel, power. All ultimately controlled by a London-based holding company, Vedanta Resources PLC, listed in 2003 because Indian capital markets back then weren’t deep enough to fund a multi-commodity resources business.

The London listing made sense for two decades. Then it didn’t. The stock underperformed BHP and Rio Tinto. Compliance was heavy. The 2018 Tuticorin protest deaths added reputational damage. Agarwal took the parent private in October 2018 for roughly a billion dollars, which only added more debt on top of debt already inherited from the $9 billion Cairn India acquisition. By October 2022, Moody’s started downgrading Vedanta Resources, calling its capital structure “unsustainable.” The London parent had no operating cash flow of its own — it serviced its debt entirely from dividends and brand fees paid up by its Indian subsidiaries. Bond investors couldn’t see how much of the group’s cash was actually available to repay the parent. By September 2023, the bonds had been downgraded to CA3 — junk territory.

So Vedanta announced the demerger.

How the surgery was performed

Three decisions matter.

Debt push-down. The London parent owed $5.5 billion net of cash. That debt has been distributed across the five new entities, roughly in proportion to their cash flow capacity. Aluminium, the cash cow with 38% operating margins, took the largest piece — $3.5 billion. Power took on debt matched against its 15-25 year power purchase contracts. The residual Vedanta Limited absorbed about a billion. Iron and steel got $200 million.

Dividend policy. Until now, Vedanta Limited had to pay out at least 30% of attributable profits and pass through all of Hindustan Zinc’s dividends to its own shareholders within six months. The lion’s share went to the London parent. Under the new policy, the board decides what to pay and when. The cash can stay and be deployed more flexibly.

Brand fee. Every Vedanta entity pays the London parent for the right to use the trademark. The rate has crept up every time the parent has been in trouble — 0.75% in 2017, 2% in 2021, and now roughly 3% for most entities (0.75% for copper because its margins are too thin). On group revenue of around Rs 1.74 lakh crore in FY26, a 3% fee yields about Rs 5,200 crore. That covers the parent’s annual interest bill with room to spare.

The architecture solves the same problem from two sides. The $4.8 billion parent debt is now sitting on five separate balance sheets. The cash flowing up to London is doing so at a higher rate but through cleaner channels.

What the operating businesses look like

Two pillars carry the tent. Aluminium operating profit grew 43% to over Rs 25,000 crore, with cost per ton at $1,752 — a five-year low, the payoff from years of acquiring captive coal and bauxite mines. Hindustan Zinc came in around Rs 22,000 crore, margins of 56%, cost per ton at a five-year low of $959. These are unusually good industrial numbers.

The rest carries shadows. Vedanta Power had a fatal accident — at least 13 dead, more than 20 injured — and one of its two main plants is shuttered indefinitely. Vedanta Oil and Gas has declining reserves. Iron and steel has a smaller asset base than the market has tested. Next quarter, each will face sector-specific analysts for the first time without the conglomerate’s averaging effect to hide behind.

Why states lend to other states

In 2007, Sri Lanka wanted a port at Hambantota. India said no (commercially unviable). The US said no. China said yes — $307 million at 6.3% fixed, which was actually a normal commercial rate against Sri Lankan treasury bills paying 12-14%. A decade later, the port became the world’s most famous case study in “debt trap diplomacy.”

Except the framing is wrong. Sovereign loans are almost never just commerce, but they are also almost never just traps. They are bundles — development assistance, export contracts, foreign policy positioning, sometimes monetary strategy — all moving inside a single agreement.

Four main motives drive state lending. Soft power — the Marshall Plan rebuilt Europe and kept it out of the Soviet sphere. Export promotion — finance a foreign government on the condition they hire your engineering firms. Resource and logistic security — Angola financed its post-civil-war reconstruction through Chinese loans collateralised against future oil exports, routed through a Chinese-controlled bank account. Macro-financial recycling — Ben Bernanke’s “global savings glut.” When you run massive trade surpluses, the savings have to go somewhere; lending them out is one way.

The instruments match the motive. Grants for humanitarian work. Concessional ODA loans where at least 25% of the value is effectively a gift (Japan’s loan to the Mumbai-Ahmedabad bullet train at 0.1% over 50 years with a 15-year grace period is the textbook case). Policy bank loans at near-commercial rates — China Development Bank, Japan’s JBIC, Germany’s KfW, India’s Exim Bank. Tied export credit, where India’s Exim lines require 75% procurement from Indian firms (Japan requires only 30%). Central bank swap lines, which are emergency liquidity dressed up as monetary cooperation.

China is doing what Japan already did

Japan ran this exact playbook from the 1970s through the early 1990s and was the world’s largest aid donor by 1989. Indonesia, the Philippines, and Vietnam absorbed enormous amounts of Japanese capital, with Japanese firms building the region’s modern infrastructure.

China arrived with the same model in the 2000s but with three things Japan didn’t have at the same scale — a 45% savings rate generating endless surpluses, idle steel and cement factories after the 2008 stimulus, and a political system that could coordinate policy banks, state firms, and sovereign capital in one push. Chinese overseas lending peaked at $136 billion in 2016 and has been falling since. Many flagship projects — Hambantota, Kenya’s railway, Zambia’s sovereign debt — turned out to be commercially weak. Chinese state lenders are now busier refinancing old loans than writing new ones. AidData calls Beijing the world’s largest official debt collector — 128 rescue operations across 22 distressed borrowers worth roughly $240 billion by end of 2021.

What does set Chinese lending apart is contract design. A 2021 study of 100 contracts across 24 countries found three unusual features — confidentiality clauses preventing borrowers from disclosing terms (sometimes even the existence of the loan), escrow accounts where the borrower’s commodity revenues route through Chinese banks before reaching the borrower’s treasury, and clauses designed to keep Chinese debt out of collective restructurings. None of this proves the debt trap thesis, but it does mean Chinese lenders have built a system optimised for seniority — getting paid first when things go wrong.

What Hambantota actually was

Sri Lanka’s 2022 default was not primarily caused by Chinese debt. Of the $37 billion owed externally, roughly 36-40% sat with Western bondholders at coupons of 5-8%, only 19-20% with China. The trigger was the bond maturities, not the Chinese loans — Sri Lanka had issued $12.5 billion of bonds in the 2010s when credit looked cheap, and Covid killed the tourism dollars needed to roll them over.

The port wasn’t seized. In 2017, Sri Lanka leased it for 99 years to a Chinese state shipping operator for $1.12 billion in cash, which it then used to service its other foreign debt. The original Chinese loans are still on Sri Lanka’s books. What changed hands was operational control of a failing asset.

The Chinese contracts didn’t cause the default. They did make fixing it slower, harder, and more expensive. Restructuring deals with Western bondholders closed in December 2024 and with Japan in March 2025. Talks with China over the remaining $4.75 billion are still unresolved.

Key Takeaways

Vedanta’s demerger is structural surgery dictated by the London parent’s debt, not strategic clarity. The $5.5 billion has been distributed across five balance sheets, the brand fee bumped to 3%, and the dividend policy loosened — all so cash can keep flowing up to London without choking the operating subsidiaries.
Aluminium and Hindustan Zinc do the heavy lifting. Margins of 38% and 56%, cost positions at five-year lows. The other three entities head into independent life with real shadows — a fatal accident at Vedanta Power, declining reserves at Oil and Gas, an untested asset base at Iron and Steel.
The Q1 earnings calls of the five separate entities will be the first time their managements face sector-specific analysts without the conglomerate’s averaging effect.
Sovereign lending is rarely just lending. Four motives — soft power, export promotion, resource security, surplus recycling — and the instrument matches the motive.
China is running the Japanese playbook from the 1970s-90s, only bigger. The novelty is contract design — confidentiality, escrow accounts, anti-restructuring clauses — built for seniority when things go wrong.
Hambantota was not a seizure. The port was leased for $1.12 billion in cash that Sri Lanka used to service Western bondholders, who were the bigger creditor anyway.

Claude’s Take

The Vedanta segment is the better of the two stories — it’s specific, it’s recent, and it forces you to read the demerger as a financing manoeuvre rather than a strategic pivot. The aluminium business is genuinely a high-quality asset. The residual Vedanta Limited is essentially a holding-company wrapper around Hindustan Zinc with new flexibility on dividends. The other three entities are where the discount lives. Worth watching how the market prices the four new tickers once they list — the conglomerate discount theory implies the sum-of-parts should re-rate, but only if the smaller pieces don’t disappoint on their first independent earnings calls.

The sovereign loans segment is well-trodden territory dressed up as a fresh take. The framework — four motives, five instruments — is useful pedagogy but not new. The genuine insight is the contract design point — Japan never used confidentiality clauses or escrow accounts, so the architecture really does distinguish Chinese lending from the historical model even if the strategic intent rhymes. The Sri Lanka revisionism (Western bondholders, not China, triggered the default) is also worth internalising — it’s a cleaner correction to the popular narrative than most takes manage.

Score 7. Solid Daily Brief, the Vedanta half does real work.

Can the demerger save Vedanta? | Uncovering the truth about sovereign loans | The Daily Brief #464