Rupee at 95: Corporate Arbitrage Impact

Rupee at 95: How Corporate Arbitrage Is Reviving the RBI's Oldest Weakness

Indian corporates are exploiting the onshore-NDF gap again, pushing the rupee to a two-month low near 95.50 and blunting the Reserve Bank of India's defence at the very moment New Delhi can least afford it.

On July 6, 2026, the rupee closed at 95.4725 per dollar, capping a 0.86% weekly slide — its worst run in nearly two months — despite pre-market intervention by the Reserve Bank of India and Brent crude sitting quietly around $70. The proximate cause is not oil, or the Federal Reserve, or foreign investors dumping Indian bonds. It is a legal, documented arbitrage between Mumbai's onshore forward market and London's offshore non-deliverable forwards — one that Indian exporters and importers, sitting on genuine trade paperwork, have re-learned to run. The RBI can sell dollars all day; it cannot close a 4-to-6 paisa spread that pays corporates to buy them.

That is the story here. Not a currency crisis — the World Bank's April 2026 India Development Update still credits India with roughly $690 billion in reserves — but a slow leak in the plumbing of a managed float, reopened by the very participants the central bank has spent five years trying to bring onshore. It matters because the current account deficit is widening, the balance of payments gap has, according to Nomura estimates cited by the BBC, crossed $70 billion, and the political ceiling on rupee weakness — the 100-per-dollar psychological line — is closer than it has ever been.

What corporates are actually doing

The mechanics are unglamorous. An Indian importer with a genuine dollar payable — a shipment of crude, a machine order, a chip contract — is allowed under RBI rules to hedge that exposure. In early July, the one-month USD/INR contract in the offshore NDF market has been quoting 4 to 6 paisa above the onshore forward rate, according to Reuters reporting carried by The Hindu BusinessLine. The corporate sells dollars forward at the richer offshore rate and buys them at the cheaper onshore rate against the same underlying trade document. Bankers describe it as "free money." One FX salesperson at a foreign bank told Reuters his desk had executed around 10 such transactions in recent days.

The onshore leg forces the bank to buy dollars in the interbank market to hedge itself. That is the dollar demand pressuring the currency. Rinse, repeat, and the RBI's intervention gets absorbed by counterparty flows rather than sentiment. It is a mechanical, not directional, source of pressure — which is precisely why it is so hard for a central bank to fight with words.

None of this is illegal. It is a rational response to a well-known price gap the RBI itself has spent years trying to shrink. As an IMF working paper on India's FX markets documented in 2024, offshore rupee transactions already account for more than 60% of total OTC volumes, driven by NDFs whose volume has "nearly tripled since 2016." The onshore-NDF spread, once 17 forward points in 2010–2014, had compressed to about 5 in 2020–2023. The current 4-to-6 paisa gap is therefore not extreme by historical standards — it is close to average — but it has become newly tradable because directional pressure on the rupee has turned one-way and the RBI is defending a level rather than smoothing volatility.

The historical parallel: 2013 and 2020, but with better plumbing

To grasp why the RBI is treating this seriously, look at the two episodes it is trying not to repeat. During the 2013 taper tantrum and the March 2020 COVID shock, the IMF's Schmittmann and Chua found that Indian rupee NDFs "were pricing much more depreciation than onshore markets" and that offshore quotes led onshore prices during stress. The rupee moved because London decided where it should trade before Mumbai opened. Governor Urjit Patel and, later, Shaktikanta Das treated this as a strategic vulnerability: an Indian currency being priced by a market they could not touch.

The response was to bring the NDF onshore. Since June 2020, banks operating out of Gujarat International Finance Tec-City (GIFT) have been allowed to transact non-deliverable derivative contracts, a change enshrined in the RBI's current Master Direction on Risk Management and Inter-Bank Dealings. The result, per the IMF, was tighter spreads and better transmission of onshore policy signals into offshore quotes. That worked — until this year, when a chronically weak rupee turned the residual spread into a directional trade.

The July 2026 episode is therefore not a re-run of 2013. Reserves are three times what they were then; the BBC's coverage of Modi's austerity appeal notes India's $690 billion buffer is enough for 11 months of imports, against three weeks in the 1991 balance-of-payments crisis. But it is a reminder that as India's FX market has matured, the arbitrage channel has become the dominant transmission mechanism from global dollar strength to the domestic currency. The Economist noted in April that the rupee had lost around a tenth of its value against the dollar in the fiscal year to March 2026, a slide it attributes less to the Iran war than to a persistent inability to draw in foreign investors.

Why the RBI's March cap is bending, not breaking

To slow exactly this dynamic, the Reserve Bank in late March 2026 capped banks' net open position on the rupee in the onshore market at $100 million per bank. The rule, described in The Hindu BusinessLine's account, forces every client-driven trade to be squared off almost immediately with an offsetting interbank transaction. That is why bankers describe trade sizes as "modest" and why the arbitrage spread has not closed: banks cannot warehouse the risk long enough to bid the offshore rate down.

The prudential framework sits inside a broader web. The RBI's Master Direction sets the Net Overnight Open Position Limit at up to 25% of a bank's Tier I and Tier II capital and requires exchange-traded currency derivative positions to sit within it. A separate RBI notification governing exporter and importer forward cover requires the authorised dealer bank to verify an underlying trade exposure; contracts booked in excess of 75% of the eligible past-performance limit are on a deliverable basis and cannot be cancelled. That is the paperwork wall that keeps this a corporate — not a speculator — trade.

The result is a policy stalemate. The RBI has succeeded in preventing speculative NDF flows from swamping the rupee, as they did in 2013 and 2020. It has not been able to stop legitimate hedgers from routing their real dollar demand through the fatter leg of the arbitrage. Every corporate that does this converts a natural hedge into a marginal dollar buyer — and the aggregate flow, though bounded by the $100 million cap, is enough to overwhelm the RBI's routine spot sales.

Who wins, who loses

The immediate winner is the mid- to large-cap Indian corporate treasurer with both import and export books. Large diversified groups — Reliance, Tata, Adani-linked entities, IT majors with dollar receivables against rupee costs — are precisely the kind of counterparties with the trade documentation, bank limits and legal appetite to run this trade at scale. Their treasuries have effectively been handed an income line by the spread. Foreign banks with GIFT City units — HSBC, Standard Chartered, Deutsche, Citi — capture the flow. Domestic private banks with strong corporate franchises book the fees.

The losers are more diffuse. First, the RBI itself: every dollar sold in defence of the rupee that gets absorbed by an arbitrage-linked interbank hedge is a dollar spent without buying sentiment. FX reserves have already fallen by $38 billion since the Iran war began, one of the sharpest declines in the region, per the BBC. Second, the government of Narendra Modi, which the Economist reported in April faces a political ceiling on rupee weakness that a slide toward 100 would decisively breach. Third, unhedged small importers and consumers, who pay the higher import bill without access to offshore markets.

There is a subtler loser: the credibility of the RBI's own liberalisation. Governor Sanjay Malhotra, who took over in December 2024 and has cut the repo rate by 125 basis points to 5.5% while praising a "Goldilocks" economy in an FT interview, has tied his tenure to a growth-supportive stance. Every rupee wobble narrows the space for another cut. The IMF's 2025 Article IV report already noted that RBI had judged it prudent to keep the policy rate at 5.5% given a neutral stance — the arbitrage-driven dollar demand is precisely the sort of "excessive volatility" that could force Malhotra to hold rather than cut in August.

Diplomat View

The rupee's problem in July 2026 is not the arbitrage — it is what the arbitrage reveals. India has built a modern FX market with world-class plumbing, and that plumbing now transmits every incremental dollar of stress instantly into the currency. The $100 million net open position cap is a good tourniquet but a bad long-term policy: it distorts price discovery without addressing the underlying imbalance between dollar demand (crude, gold, fertiliser, capex imports) and dollar supply (weak FDI, sluggish exports, foreign portfolio outflows).

The forecast: absent a positive shock — a US-India trade deal, a decisive de-escalation of the Iran conflict pushing Brent below $65, or a foreign inflow surprise from bond-index rebalancing — the rupee grinds toward 96 through August, and the RBI defends 97 hard. A move through 100 is off the table this quarter only because reserves are sufficient to make defence credible. The forecast changes if oil breaks above $85, if FPI outflows accelerate past $10 billion in a month, or if Malhotra signals a rate cut without a matching macroprudential tightening. Watch the MPC statement in early August: a hold that explicitly cites "external sector volatility" is the tell that the arbitrage has won the tactical round.

What to watch next:

• Early August 2026 MPC decision — a hold with hawkish language on the rupee signals the arbitrage has forced the RBI's hand.
• Weekly RBI Statistical Supplement releases — reserve drawdown of more than $5 billion in a single week would confirm intervention is escalating.
• Any tightening of the $100 million NOP cap or new curbs on GIFT City NDF participation — the RBI's move from macroprudential to administrative controls would mark defeat for the 2020 liberalisation experiment.

The Bottom Line

The rupee is not being sold by speculators; it is being sold, one documented trade at a time, by Indian companies capturing a legal spread the RBI cannot close without unwinding a decade of market reform. Until New Delhi resolves the underlying dollar-demand imbalance — through oil diversification, capital-account inflows, or a politically painful adjustment in gold and travel imports — every incremental episode of global dollar strength will land on the rupee through this same channel. The $100 million cap buys time. It does not buy a floor.