By V Shunmugam, Market economist and Partner at MCQube
Currencies don’t usually free-fall in slow motion. They do so in bursts—when macro-anxiety meets microstructure fragility. Over the last fortnight, the rupee has been that textbook case: a sharp slide driven by war-risk premia in oil, worsening risk appetite, and the sudden realisation that India’s balance of payments is not a comfort blanket when capital is rushing out. By March 30, the rupee printed fresh record lows around Rs 95 per dollar intraday, closing near Rs 94.83—an approximately 11% fiscal-year drop, its worst annual decline in well over a decade.
This time, the damage wasn’t just a number on a screen. Volatility in foreign exchange doesn’t stay contained; it becomes a product—packaged into spreads, warehoused on dealer balance sheets, and amplified through arbitrage channels never designed to carry this kind of load. When geopolitical risk spiked and foreign portfolio outflows accelerated, the rupee didn’t merely weaken; the market’s plumbing began to creak. The most revealing signal wasn’t the spot print—it was the dislocation between onshore and offshore pricing and the frantic scramble to monetise it. Settlement data from the Clearing Corporation of India shows that on March 30 alone, the daily gross USD/INR position across all OTC segments touched roughly $302 billion, with forwards accounting for about $220 billion across over 54,000 deals compared with an average $200 billion and 30,000-35,000 deals. They aren’t normal numbers; they are stress numbers.
Here lies the uncomfortable truth at the heart of this episode—dealer-driven FX markets do not merely absorb volatility—they can profit from it. The mechanics are worth understanding because they explain both how the crisis deepened and why the RBI had to act so forcefully.
In a calm market, the margin between where a bank buys and sells dollars is thin and competitive. When importers panic-buy and exporters hold back, banks widen spreads and clip a larger margin on every trade. Beyond client flow, treasury desks in stress conditions often take directional bets, carefully positioned but real, tucked inside what gets reported as “risk management”.
The bigger game, though, is the arbitrage between onshore deliverable pricing and offshore non-deliverable forwards (NDFs). NDFs exist precisely because India’s rupee has historically had capital account restrictions that prevent offshore players from holding actual rupees. Instead, global banks, macro hedge funds, and prop trading firms trade a cash-settled proxy—punting on where the rupee will fix. In normal times, the onshore and NDF markets stay broadly aligned. In stress conditions, they diverge—and that gap becomes a conveyor belt for speculation. By late March, banks had reportedly built arbitrage positions of $25-35 billion, with some estimates placing pure arb structures at a further $10-18 billion. As offshore INR weakened, domestic dealers hedged by selling rupees onshore, adding fuel to the very fire the RBI was trying to douse.
This is the paradox of arbitrage in an asymmetric system. In theory, arb aligns prices. In practice—when one side is a 24/5 offshore market running on global risk appetite, and the other is a domestic market where a central bank is trying to smooth disorderly moves—“alignment” becomes a one-way accelerator. The onshore-NDF spread, normally a handful of paise, reportedly blew out dramatically, allowing corporates to buy dollars onshore and sell them offshore—a sign that the transmission mechanism between markets has broken down.
On March 27, the RBI instructed authorised dealers to limit their net open position in INR (NOP INR) to $100 million at the close of each business day, effective April 10. This overrides the earlier regime, where banks could set limits linked to their capital base—giving dealers significant runway to warehouse INR risk.
The logic is elegant—if banks cannot hold large onshore INR positions, they cannot run the arbitrage trades that transmit offshore bearishness into domestic selling pressure. Positions that were “economically” hedged via NDF legs offshore but still exerted real pressure on the onshore market now have to be unwound. A sharp rupee bounce was reported immediately after the announcement, as traders who had been caught on the wrong side were forced to buy back dollars locally.
But the central bank is under no illusions that this is cost-free. Forced unwinds can be disorderly; banks may book losses; and tighter position limits may make dealers more cautious about quoting tight spreads to real-economy customers. The policy choice is deliberate: accept pain in the dealer P&L to reduce currency volatility. That trade-off is sometimes the right one to make—and this appears to be one of those moments.
Consider bank earnings data. ICICI Bank’s FY25 profit on exchange and derivative transactions rose roughly a third year-on-year. HDFC Bank’s net trading and mark-to-market gain for the December 2025 quarter exploded from `0.7 billion to `9.3 billion in a single year. SBI’s fee and exchange-linked income trended higher as market activity surged. These are not incidental but structural realities—the system’s most powerful intermediaries are financially incentivised by the very conditions the central bank is trying to stifle. The RBI’s position cap, among other things, is an attempt to break that alignment.
The NOP cap can be the current bandage. The surgery has to focus on building onshore FX market that is transparent and liquid enough to make offshore NDFs redundant as a price signal.
India already attempted this. Exchange-traded currency derivatives were launched in 2008 explicitly to counter NDF markets and bring currency access to SMEs and retail participants. The architecture worked—but was gradually hollowed out. With underlying exposure now mandatory for participation, volumes have collapsed, and the segment has been reduced from a price-discovery platform to a hedging utility for the fews.
The answer may not require reinvention. Allowing speculative participation within defined position limits, expanding the instrument set, and leaving product innovation to the market infrastructure institution ecosystem while regulators focus on steering rather than prescribing a reasonable starting point. Deliverability, however, would need to stay off the table—a deliverable rupee market carries dollarisation risks difficult to justify.
The broader question, perhaps, is whether the conditions now exist for the rupee to begin discovering its price at home—visibly, in a central limit order book—rather than continuing to look offshore for that signal. The NOP cap may have quietly opened that conversation.