NextFin News - India's foreign exchange reserves dropped by $14.88 billion in the week ended September 18, the steepest weekly fall in almost two years, as the Reserve Bank of India sold dollars to defend a rupee pushed toward record lows by a fresh oil-price shock and the Federal Reserve's first rate hike since 2023. The decline, reported in the central bank's weekly data on Friday, pulled the reserves pile down to $765.9 billion from $780.78 billion a week earlier — erasing nearly $20 billion in just two weeks from the all-time high of $785.7 billion touched in early September.
The scale of the drawdown reframes a question that has shadowed Indian markets all year: how long can the RBI lean against the rupee's depreciation without burning through the very buffer it is trying to showcase? The answer so far is that the buffer is large enough to absorb the hit — but the cost of defending the currency is now showing up in the headline reserve number, not just in the fine print of the central bank's forward book.
The Numbers: A Two-Week Reversal From Record High
The sequence tells the story. In the week ended September 4, India's reserves jumped a record $44.9 billion to $785.7 billion, lifting the country past Russia into fourth place globally among reserve holders. A week later, they fell $4.92 billion to $780.78 billion. In the week ended September 18, the decline accelerated to $14.88 billion, leaving reserves at $765.9 billion.
Foreign currency assets — the largest component and the part the RBI actually uses to intervene — bore almost the entire hit. FCAs fell $14.82 billion to $630.98 billion, accounting for 99.5% of the total weekly decline. Gold reserves, by contrast, edged up $68 million to $111.29 billion, and special drawing rights slipped $106 million to $18.74 billion. The near-total concentration of the drop in FCAs is the fingerprint of intervention: when the central bank sells dollars to buy rupees, foreign currency assets shrink almost one-for-one.
Even after the fall, the reserves remain $89.67 billion above the end-March 2026 level. But the direction has turned, and it has turned quickly. The two-week net decline of roughly $19.8 billion from the September peak is the largest reserve drawdown since the autumn of 2024.
The currency market context explains the timing. On September 17, the rupee slipped past 96 per dollar for the first time in more than a month, closing at 96.08 — a session in which four traders said state-run banks were spotted offering dollars, most likely on behalf of the central bank. The trigger was the Federal Reserve's 25-basis-point rate increase on September 16, its first hike since 2023, which lifted the dollar and pushed U.S. Treasury yields higher. The rupee has been one of Asia's weakest currencies through 2026, down around 6% against the dollar for the year, and the RBI's playbook has been consistent: intervene when the pair approaches its record high rather than defend a fixed line. That record stands at 96.844, set on May 20.
The Mechanism: How Defending the Rupee Shows Up in the Data
The transmission channel from currency defence to reserve drawdown is mechanical, but it is worth tracing because it is often misunderstood. When the RBI sells dollars in the spot market, it credits rupees to the buyer's account and debits its foreign currency assets. The rupee money supply expands; the foreign reserve stock contracts. The weekly statistical supplement captures this with a one-week lag, which is why the intervention visible in the September 17 session appears in the week-ended-September-18 print.
There is a second channel, less visible but equally important: forward intervention. The central bank has also been conducting dollar-rupee sell/buy swaps — selling dollars spot and agreeing to buy them back later at a fixed forward rate. These swaps drain rupee liquidity from the banking system without immediately shrinking the headline reserve number, because the future repurchase obligation is parked in the forward book rather than netted against the stockpile. That is why the central bank's own September 11 release carried a caution worth repeating: the frequent interventions supporting the rupee have likely offset some of the overseas dollar inflows that otherwise would have swelled the reserves, and those inflows are recorded as a future liability.
Put differently, the headline reserve number is not a clean measure of firepower. A portion of the roughly $136 billion the RBI mobilised this year through foreign-currency deposit schemes and concessional swaps sits on both sides of the balance sheet — an asset in the spot reserves and a matching liability in the forward book. When the central bank chooses to bring those forward sales onto the spot ledger to defend the currency, the headline number falls faster than the underlying position deteriorates.
Why September Was Different: Three Shocks Converged
Three forces came together in the second half of September to produce the largest weekly reserve drawdown in almost two years.
First, oil. Brent crude reached $106.52 a barrel on September 24, up 3.3% in a day and 22% over the month, as uncertainty over US-Iran tensions and attacks on Saudi oil terminals revived fears of disruption in the Strait of Hormuz. India imports nearly 89% of its crude, so every $10 move in Brent translates into billions of dollars of additional import bills and a structural rise in dollar demand from refiners and airlines.
Second, the Fed. The September 16 rate hike — widely expected but consequential nonetheless — widened the US-India yield spread and strengthened the dollar broadly. Emerging-market currencies came under pressure, and the rupee's breach of 96 was the local expression of that move.
Third, the calendar itself. The RBI's special deposit mobilisation drive, which had pulled in roughly $57 billion before the facility was closed early, was winding down. With that one-off inflow behind it, the reserves became more sensitive to the underlying current-account and portfolio-flow pressures that had been masked during the fundraising window.
The combination matters because each shock reinforces the others. Higher oil prices widen the trade deficit, which weakens the rupee; a weaker rupee raises the local-currency cost of oil imports, feeding inflation; and a Fed that is hiking rather than cutting reduces the room for the RBI to defend the currency with interest-rate policy alone. The result is a central bank that must choose between two imperfect tools: burning reserves or tightening financial conditions into a slowing economy.
Cyclical Shock or Structural Erosion? The Call
This is the judgment the market needs to make, and the answer is not either/or. The reserve drawdown itself is cyclical — a mean-reverting response to a commodity-price spike and a dollar-strength episode. India's reserves have fallen by more than $10 billion in a single week before, in January 2026 ($9.89 billion) and March 2026 ($11.68 billion), and on both occasions the pile rebuilt as oil eased and flows stabilised. A cyclical drawdown does not, by itself, signal a balance-of-payments problem.
But layered on top of the cycle is a structural concern that will not self-correct. India's external position remains structurally sensitive to oil: the country has never solved the problem of financing a large energy import bill with a currency that foreign investors do not naturally want to hold. Foreign institutional investors have been net sellers of Indian equities for much of 2026, and the rupee has been the weakest major Asian currency for most of the year despite repeated intervention. That combination — persistent portfolio outflows plus a currency that needs defending even when reserves are near record highs — is the structural leg of the story.
The distinction has a concrete implication. If this were purely cyclical, the RBI could afford to let the rupee absorb some of the pressure and wait for oil to fall, confident that the reserves would rebuild. If the structural leg dominates, every defence of the currency is a transfer of resilience from the balance sheet to the exchange rate — and the exchange rate, as 2026 has shown, keeps finding new lows anyway. The rupee touched 96.844 in May; it is testing the mid-96s again in September. The reserves are being spent to slow the pace of depreciation, not to reverse it.
The Second-Order Cost: What the Market Is Not Pricing
The first-order effect of the intervention is obvious: the rupee stabilises, at least temporarily. The second-order effects are where the real story lies.
The first is the forward-book liability. Every dollar the RBI sells today, whether spot or via swap, is a dollar it may need to buy back later — potentially at a higher rupee price. If the rupee continues to depreciate, the central bank's own balance sheet absorbs a valuation loss, which shows up as a reduction in its surplus transfer to the government. That is a fiscal cost of currency defence that does not appear in the weekly reserve print.
The second is the inflation pass-through. A rupee that is only slowly depreciating, rather than absorbing the oil shock in one move, spreads the import-price increase over a longer period. That keeps core inflation stickier for longer and narrows the RBI's room to cut rates even as growth moderates. The market is currently debating whether the RBI's next move is a hike — some economists see an October move as increasingly likely, while others expect the central bank to wait until December — but the currency is already doing some of the tightening for them.
The third is the signal to speculators. A central bank that intervenes reliably near a visible level — 96, in the current episode — teaches the market where the floor is. That can deter one-way bets in the short run. But it can also concentrate selling pressure just above the perceived defence line, so that when the central bank eventually steps back, the move is sharper. The RBI's stated approach of curbing "undue volatility" rather than targeting a level is designed to avoid exactly this trap; the question is whether the market believes it.
The Counter-Thesis: Why the Buffer Still Holds
The bear case is not without an answer. The strongest argument for the RBI's approach is simply the size of the cushion. Governor Sanjay Malhotra, delivering the monetary policy address in August, said:
"India's foreign exchange reserves continue to be adequate in terms of the standard metrics of reserve adequacy with import cover of over 10 months and external debt cover of 90.8%."
An import cover of more than 10 months is comfortably above the conventional adequacy benchmark of three months, and external-debt cover near 91% leaves little room for a liquidity crisis of the kind that has felled other emerging markets.
The RBI also has more than spot sales in its toolkit. It can tighten liquidity through swaps, raise the cost of shorting the rupee, and lean on state banks to manage flows. And the reserve pile, even after the September drawdown, is still far larger than it was at the start of the year — the $765.9 billion level is roughly $214 billion above the roughly $552 billion recorded in early January.
But even the optimists acknowledge a ceiling. Tanay Dalal, an economist at Axis Bank, put it plainly after the previous week's data:
"Should the RBI elect to bring forward existing dollar sales in the upto 12-month horizon, the headline reserves will likely stabilize around $750 billion."
That $750 billion figure is not a forecast of where reserves will sit comfortably; it is an estimate of where the headline number lands once the central bank recognises its forward obligations on the spot ledger. It is a useful anchor: the market should think of India's usable buffer as closer to $750 billion than to the $785 billion peak.
What Comes Next: Three Scenarios
Base case — managed drift. Oil stabilises in the $100–$105 range, the Fed signals that the September hike is a one-off, and the RBI continues to intervene selectively near 96–96.5. Reserves grind lower toward the $750 billion anchor over the next two to three months, then stabilise. The rupee depreciates gradually, in the 3–5% annualised range, without a disorderly break.
Upside case — oil-driven relief. US-Iran tensions ease, Brent falls back toward $90, and the trade deficit narrows. Portfolio flows return as the growth differential reasserts itself. In this scenario the reserves rebuild toward $780 billion and the rupee recovers toward the low 94s. This is the cyclical mean-reversion trade, and it is the one the RBI is implicitly betting on.
Downside case — a test of the buffer. Oil stays above $110 for an extended period, the Fed delivers more than one hike, and the rupee breaks decisively through 97. In that world the RBI faces a genuine trade-off: spend reserves faster to defend 97, or let the currency move and risk an inflation spike. A break of 97 with reserves falling another $20 billion in a month would be the signal that the structural leg has overtaken the cyclical one.
The Watch List
Three signals will tell the story before the next weekly reserve print. First, Brent crude: sustained trading above $110 a barrel would keep the import bill under pressure. Second, the USD/INR level: a decisive break above 97 would test the RBI's tolerance. Third, the forward-book disclosure in the RBI's monthly balance sheet, which will show whether the central bank is accumulating future dollar obligations faster than it is rebuilding spot reserves.
The falsifying signal for the base case is specific: if Brent holds above $110 for four consecutive weeks while the rupee trades above 97 and reserves fall another $20 billion, the "managed drift" thesis is wrong and the structural-depreciation case takes over.
The central judgment: India's reserves are large enough to survive this episode without a crisis, but they are being used to buy time, not to change direction. The rupee's trend is set by oil and the dollar, and no amount of intervention reverses either. The $14.88 billion drawdown is not a warning of imminent trouble — it is the invoice for a defence that slows the fall without stopping it.
Explore more exclusive insights at nextfin.ai.
