The Improbable Equilibrium · Chapter 5

The Conduit Island

A note on this draft. This chapter rests on the one source document in the Mauritius archive that carries the status [COMPLETE] — the offshore-finance record covering 1992 to 2024. That makes it the most source-confident chapter in the book, and the prose is allowed to run accordingly. But "complete" is not "certain": the underlying document still flags a series of figures as unverified — the island's share of foreign investment into India, the scale of African capital routed through it, the sector's headcount and its share of national output. Where the record marks a number as contested or estimated, this chapter carries the hedge forward rather than rounding it into fact. It does not resolve what the source leaves open.

The Third Pillar

There is a kind of money that has no nationality until it needs one — born in one country but reluctant to be taxed there, destined for another but reluctant to pay full freight to enter. What it requires is a way station: a jurisdiction stable enough to be trusted, lawful enough to be respected, accommodating enough to ask few questions. For the better part of three decades, a great deal of that money found its way station on an island of roughly 1.3 million people in the Indian Ocean that most of the world struggles to locate on a map.

Mauritius did not stumble into the role; it chose it, deliberately, as state policy, because a government that had looked hard at the island's prospects did not like what it saw. By the early 1990s the post-independence economy rested on two exposed legs: sugar, a monocrop at the mercy of preferential European prices that would not last; and the textile export-processing zone, which had delivered near-full employment but depended on trade preferences — above all the Multi-Fibre Arrangement, due to lapse in 2005 and take the island's wage advantage with it. A small island living on borrowed trade terms needed a third leg, one that could not be undercut by a cheaper factory in Asia, and the leg it chose was finance: a purpose-made offshore architecture to attract foreign capital seeking a low-tax, treaty-sheltered home. The instrument was the Offshore Business Activities Act of 1992, passed under the first administration of Sir Anerood Jugnauth, and officials called it the "third pillar" of diversification, alongside sugar and tourism. What it created was a factory for companies — corporate shells that existed on the island in a legal sense without necessarily doing much there in a physical one.

What made the strategy plausible, where it would have failed almost anywhere else, was a combination of assets few jurisdictions could assemble at once: a working common-law judiciary inherited from the colonial period, in which a shareholders' agreement could be enforced and an arbitral award executed; political stability resting on credible, regularly contested multi-party elections; a bilingual legal and professional class, fluent in English and French, able to service sophisticated cross-border transactions; and, almost by accident of history, a tax treaty with India, signed in 1983 for ordinary reasons, that was about to become one of the most valuable pieces of paper in the developing world. The strangeness of the bargain is worth pausing on: a nation builds its institutions over generations in order to govern itself, and Mauritius built them and then rented them out. The very things that made it a functioning plural democracy, the subject of this book's earlier chapters, were the things that made it a trustworthy address for other people's capital. The "Mauritius miracle" and the "Mauritius conduit" were two names for the same machine.

Anatomy of an Exemption

The 1992 Act did not create one kind of offshore company. It created two, and the distinction was the quiet engineering at the heart of the edifice. The Category 1 Global Business Company — GBC1 in the trade's shorthand — was on paper a real tax resident of Mauritius: it had to keep at least two resident directors, a local bank account and one board meeting a year on the island, and in exchange — on obtaining a Global Business Licence from what became, in 2001, the Financial Services Commission — it paid a headline fifteen per cent tax that a deemed-foreign-tax-credit mechanism reduced to an effective rate of around three. More valuable than the rate was the privilege residence carried: a GBC1 could draw on the island's treaty network, above all the treaty with India. The Category 2 company was the simpler animal — explicitly not resident, paying no local tax, barred from the treaty network, needing no resident directors and almost no presence. It was the closest Mauritius came to the classic "brass-plate" company, a name on a door with little behind it, and it drew little regulatory attention until international pressure eventually forced its abolition.

This bifurcation was the system's cleverest feature and, eventually, its most exposed. By splitting its companies into a "substantive" tier and a "non-resident" tier, Mauritius could present a face of respectability while the practical bar for clearing the GBC1 substance test stayed low: its "central management and control" requirement could be met with nominee directors, a near-empty account, and a board meeting attended for a few hours once a year. To the island's regulators this met the norms of the time; to critics at the Tax Justice Network and, later, the Organisation for Economic Co-operation and Development, it was substance in name only. Both descriptions were accurate; they measured against different rulers.

The treaty that turned all of this from a diversification strategy into a phenomenon was older than the architecture around it. The Double Tax Avoidance Agreement between India and Mauritius, signed on 24 August 1983, was a conventional treaty between two Commonwealth states, meant to stop the same income being taxed twice; nobody negotiating it then was contemplating waves of foreign capital pouring into Indian equities, because India's equity markets were closed to foreign institutional money. Its most consequential clause was, at signing, of almost no commercial interest: Article 13 provided that gains from selling shares could be taxed only in the seller's country of residence. For a Mauritius-incorporated GBC1 holding Indian shares, that country was Mauritius — and Mauritius did not tax capital gains. The effective tax on the gain was zero.

The dormant provision came alive in the early 1990s, when India, under the structural-adjustment programme that followed its 1991 balance-of-payments crisis, liberalised its capital account and admitted registered Foreign Institutional Investors into its equity markets. A foreign fund that wanted Indian shares now had a powerful reason to route through Mauritius: incorporate a vehicle, secure a Tax Residency Certificate, register with India's market regulator — and, on exit, treat the gain as Mauritius-resident and therefore untaxed. A treaty written to prevent double taxation had become an instrument for almost no taxation. The flows were enormous, even if their dimensions resist confirmation. India's own data showed Mauritius as the single largest source of cumulative foreign direct investment into the country — by figures the underlying record flags as unverified, on the order of thirty-four to thirty-seven per cent of cumulative inflows from 2000, with the broader, likewise unconfirmed claim that the island accounted for as much as forty to forty-five per cent at the corridor's height; market participants put the Mauritius share of foreign-institutional equity flows, by estimates the source again marks uncertain, at thirty to forty-five per cent of registered assets at peak. The numbers were always slippery; the direction was not. For a long stretch, an extraordinary proportion of the foreign money entering the world's largest democracy passed first through one small island.

The Round Trip

There was a shadow over those numbers, and it had a name: the trade called it round-tripping, the sector's most corrosive — and least measurable — legitimacy problem. Indian money owned by Indian residents would be sent abroad through legal or quasi-legal channels, dressed as a foreign holding structure owned by Indians through a Mauritius intermediary, folded into a GBC1, then sent back into India wearing the costume of foreign investment. On sale, the zero-capital-gains benefit applied; the promoter recovered capital plus gain, and India collected no tax on money that, in substance, had never really left Indian hands. The "foreign" investor was, at the bottom of the chain, often an Indian one.

This was not a fringe suspicion. India's revenue authorities, the Reserve Bank of India, academic economists and the International Monetary Fund all flagged it as a serious distortion. Everyone could see the shape of the problem; what no one could do was measure it — because the reason measurement was impossible went to the heart of what the island was selling. The value of a Mauritius structure, for the client who most wanted round-tripping, lay in its opacity: before the later reforms, the island's registers did not require real-time disclosure of the ultimate Indian owner, and India could see the funds on its own registers but could not pierce the corporate veil without slow, case-by-case treaty requests unsuited to systemic monitoring. The opacity that made round-tripping attractive was the same opacity that made it unquantifiable. Estimates of the share of "Mauritius FDI" into India that was really Indian capital completing a round trip ran, by some accounts, anywhere from a tenth to perhaps two-fifths — figures layered on a thing built specifically not to be counted.

By the 2010s the case against the arrangement was being assembled more systematically. The OECD's work on treaty shopping, which crystallised in Action 6 of its Base Erosion and Profit Shifting project, held that a treaty's benefits should go only to entities with a genuine economic connection to the treaty partner, not to shells interposed to harvest its advantages — and against that standard a GBC1 with nominee directors, a near-empty account and an annual board meeting looked like exactly the vehicle the rule was written to stop. The island's central instrument and the world's emerging tax norm were on a collision course.

The Managed Withdrawal

India had raised the capital-gains problem with Mauritius, on and off, since at least 2006. The negotiations were delicate on the Mauritian side: the offshore sector had become a meaningful slice of the island's economy and public revenue, and closing the exemption threatened demand for the GBC1 — Mauritius was being asked, in effect, to consent to the slow demolition of its own most valuable product.

The revised protocol was signed on 10 May 2016, during the Modi government's first term, and took effect from 1 April 2017 — less an amputation than a phased withdrawal. Capital gains on shares acquired on or after 1 April 2017 would now be taxed in India; a transitional concession taxed gains on shares bought after that date but sold before 31 March 2019 at half the Indian rate; and, crucially, everything held before 1 April 2017 was fully grandfathered, exempt no matter when sold. That grandfathering mattered: positions accumulated over decades kept their protection, making the change a slow fade rather than a cliff. The protocol also wrote a qualitative anti-abuse standard into the treaty for the first time — a limitation-of-benefits provision entitling a GBC1 to benefits only if it could show it had not been established primarily to obtain them, a screen where there had been none.

What made the withdrawal hard to escape was that India closed the side door at the same time as the front, renegotiating its treaty with Singapore on equivalent terms. Singapore had been the second great conduit for equity investment into India; revising both treaties in the same month foreclosed the obvious detour and made the change stick. The market felt it: new GBC1 incorporations aimed at Indian equities fell sharply from 2017, assets under management in India-focused funds declined, and the island's regulator acknowledged the contraction even as it pointed to a compensating trend it hoped would take up the slack — a growing book of Africa-focused mandates. The conduit was not closing. It was being redirected.

Building the African Gateway

The turn toward Africa was partly improvisation and partly foresight. Mauritius had not waited for the India corridor to falter before hedging; from the late 1990s it had been signing investment-promotion-and-protection agreements and tax treaties with African states, building a second business treaty by treaty. By 2020 it had concluded tax treaties with roughly twenty-two African countries — a figure the underlying record flags as unverified — among them South Africa, Mozambique, Rwanda, Senegal, Lesotho, Botswana and Zimbabwe, plus a web of investment-protection agreements reaching states without full treaties. When the India door swung shut, the African scaffolding was already standing.

The pitch rhymed with the old India pitch but differed in mechanics. Where the India treaty had offered a capital-gains exemption on a large, liquid stock market, the African treaties mostly offered reduced withholding-tax rates on the dividends, interest and royalties flowing out of African operating businesses and infrastructure projects up to a Mauritius holding vehicle. Some of the most respectable money in the development world used precisely this structure — development finance institutions including the British financier formerly known as the CDC, the Dutch FMO, the French Proparco and the African Development Bank itself channelled co-investment vehicles through Mauritius GBC1s, drawn not only by the tax saving but by the law: the Mauritian system of English common law supplemented by a commercial court offered contractual enforceability and arbitration access superior, for complex multi-party deals, to those of many destination jurisdictions. The cumulative scale was large but, as ever with this island, hard to pin down; international bodies recorded Mauritius as a major source of investment into sub-Saharan Africa while noting that — exactly as with India — a portion of "Mauritius investment" was really third-country capital merely passing through. Estimates of cumulative African investment structured through Mauritius companies ran upward of ninety-six billion dollars by one figure cited in regional research, though the underlying record cautions that methodological differences make even that number difficult to verify.

And with the African business came an African version of the old Indian argument, sharpened by circumstance. Bodies including the African Tax Administration Forum, Oxfam and the regional arm of the Tax Justice Network argued that the treaties Mauritius signed with African states were lopsided bargains, handing treasuries mostly downside — forgone withholding-tax revenue — while the upside, investment facilitation, could often be had by other means; by several published estimates, Africa was losing hundreds of millions of dollars a year through its treaty relationships with the island. The figure was disputed: the Mauritian side, through its financial-services promotion body, countered that the net investment effect more than offset the taxes given up. There was a quieter structural criticism too, that several treaties had been signed under unequal bargaining power, with some African states lacking the capacity to model what a treaty would cost — and Rwanda had begun a treaty review that took in its Mauritius agreement. The island's second act arrived already trailing the legitimacy question that had clouded its first.

The Grey List

The reckoning, when it came, came not from India or Africa but from the global machinery built to police dirty money, and for a period it threatened the sector with something close to extinction. The instrument was the Financial Action Task Force, the inter-governmental body that sets and grades the world's anti-money-laundering and counter-terrorist-financing standards, and its 2018 mutual evaluation of Mauritius produced findings mixed in a telling way. On paper — in its laws and institutions — the island scored adequately; the architecture was broadly compliant. The failure was in effectiveness: whether the machinery actually worked to detect, investigate and prosecute financial crime. The island had built the engine and not run it hard enough. The deficiencies cut close to the bone: inadequate supervision of the lawyers, accountants and corporate-service providers who administered the global-business structures; a financial-intelligence unit under-resourced for the suspicious-transaction reports it received; beneficial-ownership data not reliably available to law enforcement in good time — the opacity problem again, in the language of crime-fighting; and weak oversight of the non-profit sector for terrorist-financing risk. The island that sold itself on credibility had been found wanting on the one credential that, by the late 2010s, mattered most.

In October 2020 the Task Force placed Mauritius on its list of jurisdictions under increased monitoring — the grey list — with the public announcement following in February 2021. The worst consequence came not from the Task Force but from Europe: under its anti-money-laundering directive, the European Commission designated Mauritius a high-risk third country, effective March 2021, obliging European financial institutions to apply enhanced due diligence to any dealing with a Mauritius-connected counterparty. For an economy whose proposition rested on frictionless access to the global financial system, the friction was dangerous. Correspondent-banking relationships came under scrutiny, fund managers fielded due-diligence questions the moment the label landed, and existing clients began asking how hard it would be to move to Luxembourg, Ireland or the Cayman Islands. The grey-listing had laid bare the sector's deepest vulnerability: the whole thing rested on a reputation that could erode, and erosion, once visible, could become a run.

What followed was, by the Task Force's own later assessment, an unusually fast and thorough recovery. The government then led by Pravind Jugnauth treated the listing as the emergency it was: the financial-intelligence unit was restructured, the Commission launched intensive on-site supervision of the management companies that service the sector, and the law was tightened, extending beneficial-ownership obligations to the non-financial businesses and professions and giving the intelligence unit faster access to ownership information. At its June 2023 plenary the grading body removed Mauritius from the list — roughly two and a half years from listing to clearance, quick by the standards of these things — and the European Union lifted its own high-risk designation shortly after. The episode left two marks: a lesson in the fragility of small-state reputation, lost in a single evaluation and recovered only through political will and administrative capacity that small states find hardest to sustain; and a structural cleanup, as the crisis accelerated the abolition of the old brass-plate tier, completed in 2022, with remaining companies converted into a new "authorised company" category that kept non-resident status but carried heavier compliance. The gap closest to a pure shell was finally closed — it had taken an external shock to do what the island had not done for itself.

Two Stories About the Same Island

Step back from the chronology and the same island supports two accounts that cannot be reconciled — not because one is honest and the other dishonest, but because both rest on real evidence and answer different questions. In the first, the story of the development-finance gateway, Mauritius performs an intermediation service with no easy substitute: investors wanting exposure to emerging markets with thin financial depth and unreliable contract enforcement needed a holding vehicle in which a shareholders' agreement would hold, and the treaty network cut the withholding-tax leakage that might otherwise have made the return unattractive. The crucial move is the counterfactual — the alternative to a Mauritius structure was very often not the same investment through a domestic vehicle but no investment at all. This is, in essence, the case the island's promotion bodies made, and the logic behind the phrase the government preferred: not an offshore haven but an international financial centre, a small state monetising hard-won credibility. In the second, the story of the tax drain, the proponents are named and specific — the Tax Justice Network and its regional arm, the African Tax Administration Forum, the development charity Oxfam, and the scholars of illicit financial flows the archive cites. In their telling, Mauritius is a conduit that systematically strips revenue from countries with far greater fiscal need than the island itself while delivering next to nothing in genuine activity in return, a model engineered to exploit beneficial-ownership opacity and the arbitrage of treaties never designed for their actual use. The leaked-records investigations that put the island's name in headlines drew on precisely this framing: the conduit not as service but as siphon.

Both accounts are partly correct, and the truth is a matter of proportion the system was built to keep hidden. Some flows were genuine foreign capital seeking emerging-market exposure through a legally efficient route; some were domestic capital round-tripping for tax avoidance; some were corporate profit-shifting to low-substance subsidiaries. The sector contained all three in a mix no one can state with confidence, because the feature that made it attractive to the second and third kinds — opacity — is the feature that prevents the mix from being measured. The reforms of the late 2010s and early 2020s have steadily closed the space the harmful uses lived in, and what survives should look more like the first account than the second. Whether enough survives to sustain the sector at its historical scale is the question the next part turns on.

The Floor Beneath the Model

Every earlier threat had been particular. The 2016 revision attacked one mechanism — the capital-gains exemption on Indian shares; the grey-listing attacked one failing — weak supervision and opaque ownership. Each was survivable because each left intact the deepest thing the island was selling: a tax rate lower than the places its clients came from. The threat that emerged in 2021 was different in kind, because it went after the rate itself. Where the base-erosion project launched in 2013 had targeted specific abuses, the agreement reached through the OECD and G20's inclusive framework in October 2021 reached far broader: a two-pillar settlement, the second pillar of which established a fifteen per cent global minimum effective tax rate for large multinational enterprises. The home jurisdiction of a qualifying multinational can impose a "top-up" tax whenever a subsidiary is taxed below the floor anywhere in the world — so a GBC1 earning income at the roughly three per cent effective rate the deemed-tax-credit mechanism had always delivered would simply trigger a top-up charge in the parent's home country, clawing back the saving the structure existed to provide. The rate differential that had been the foundation of the proposition since 1992 would, for income-generating structures at least, cease to exist.

Mauritius joined the inclusive framework in 2019 and has committed to implementing the rules. One carve-out matters most: the minimum-tax rules exclude a slice of income attributable to genuine substance — real payroll and tangible assets. For a Mauritian business with actual employees doing actual work, this offers partial shelter; for a nominee-director GBC1 with little real employment, nothing. The regime does precisely what the island's critics demanded and its defenders claimed to embody: it rewards substance and punishes shells. The strategy the regulator and the economic-development board have set out in response is the only one available — shifting the value proposition from tax efficiency to genuine financial-service capability, the fund administration and cross-border legal work where the island's professional class, courts and regulatory infrastructure offer an advantage independent of any tax break, captured in the rebranding from "offshore centre" to "international financial centre." The difficulty is that building such a proposition takes human capital that accrues over decades, against formidable competitors with longer records and, in the European cases, single-market passporting rights that Mauritius can never obtain.

The stakes are not abstract. The sector employs, by figures the underlying record flags as estimates, somewhere in the region of eleven to thirteen thousand people directly — a substantial share of formal employment on an island of 1.3 million — and the financial and insurance activities of which it forms part contributed, again by unverified estimate, something like ten to twelve per cent of national output in the years before the pandemic. And the government that inherited the problem was not the one that built the architecture or fought off the grey-listing: the decisive 2024 election, the subject of an earlier chapter, returned Navin Ramgoolam to power at the head of the Alliance du Changement, and with power came an offshore sector caught mid-transition — past the grey list, past the abolition of the brass-plate company, but not yet through implementation of the global minimum tax that threatens its core. Whether the reformed sector is viable at scale is now their question, and the corpus this chapter rests on is candid that the answer is not yet known.

The credibility the island spent thirty years renting out was always, in the end, its only durable asset — and credibility, as the grey-listing showed, is far easier to lose than to rebuild. That logic — leverage the position, monetise the credibility, convert smallness into outsized importance — was never confined to the ledgers of finance; it was the island's whole theory of how to survive in a world built for larger states. And in the same decade the conduit faced its reckoning, Mauritius was applying that identical logic on a different terrain and to far higher stakes: not the routing of capital but the recovery of stolen territory. The map itself was on the table.