In the third week of June 2026, an Indian bank did something that would have looked strange in almost any other year. AU Small Finance Bank advertised a five year dollar deposit at 7.1 per cent. A few days later, larger banks began publishing comparable rates of their own. For context, the United States government, which can print dollars and has the deepest sovereign bond market in the world, currently pays just over four per cent on a five year Treasury note. AAA rated dollar corporate paper of similar maturity pays barely five. An Indian bank, in other words, was offering to pay nearly three percentage points more than the US Treasury for the same currency and the same maturity. On the face of it, this is the kind of arbitrage that global capital markets would have devoured and closed in seconds. It did not close because the offer is ring fenced. It is available only to the diaspora, and only because the Reserve Bank of India has resurrected its most powerful, and most expensive, dollar mobilisation tool.
To understand why India is willing to pay well over the odds for what look like plain vanilla deposits is to read the pressure fractures in the rupee, in the central bank's balance sheet, and in a global monetary system that is quietly fissuring.
What the RBI actually did
FCNR(B) stands for Foreign Currency Non Resident (Bank). These are term deposits that NRIs hold in foreign currency at Indian banks. The depositor puts in dollars, earns interest in dollars, and gets dollars back at maturity, with no rupee conversion and no exchange rate risk to the depositor. The critical "(B)" means the bank, not the depositor, carries the currency risk on the underlying funds. Hedging that risk in the forward market normally costs the bank around three to three and a half per cent per year, which is why FCNR(B) rates have historically been low.
What the RBI announced on 8 June 2026 changes that arithmetic. The central bank opened a special US dollar to rupee swap facility, available for fresh FCNR(B) deposits mobilised between 8 June and 30 September 2026, with banks able to access the swap window itself until 16 October. Under the scheme, banks can swap their dollar deposits with the RBI at the prevailing reference rate and unwind the swap at the same rate when the deposit matures. The mechanism effectively eliminates the bank's hedging burden. The RBI has also exempted these deposits from the cash reserve ratio and the statutory liquidity ratio. In effect, the central bank has stepped into the market as the counterparty of first resort, allowing banks to offer 7.1 per cent on dollars while still booking a viable rupee cost of funds.
The window comes with conditions designed to lock the funds in. The deposit must have a maturity of three to five years, with a one year minimum lock in before any premature withdrawal is permitted. The swaps that banks undertake with the RBI cannot be cancelled. The RBI is asking the banks, and through them their NRI customers, to commit for several years rather than fleetingly. The reason for that condition becomes obvious once you look at the historical context.
The 2013 precedent
The first thing to understand about the current intervention is that it is not an innovation. It is a revival. The closest historical parallel is the FCNR(B) swap window of September 2013, which became one of the defining episodes of Raghuram Rajan's early tenure as Governor.
The summer of 2013 produced what financial markets called the taper tantrum. In May, the chairman of the US Federal Reserve had suggested that the Fed might begin reducing its bond purchases. The remark was offered almost in passing. The market reaction was not. The Indian rupee fell from around 54 to a dollar in May to almost 68 by late August, a depreciation of about a quarter in three months. India's reserves were draining. The current account deficit was becoming difficult to finance.
Rajan took over on 4 September 2013. Within days, he announced a special FCNR(B) swap window allowing banks to raise three to five year dollar deposits and swap them with the RBI at a concessional fixed rate of 3.5 per cent, well below the prevailing market premium of around seven to eight. The window was open for less than three months. It raised over thirty four billion dollars from non resident Indians, an extraordinary sum at a time when total NRI deposits stood at around seventy billion. The rupee stabilised within weeks. The reserves began to rebuild. The crisis passed. The Rajan window entered the central banking canon as a high stakes, high payoff tactical instrument.
It is worth noting one thing about that earlier window that rarely appears in the celebratory tellings. When the 2013 deposits matured in 2016, the RBI had to manage a large outflow that put fresh pressure on the rupee at a time it could ill afford it. A short term defensive measure had built a refinancing cliff three years out. The lesson, applicable to the current window, is that this kind of intervention buys time. It does not solve underlying imbalances. The bill comes due, in a different form, when the deposits mature.
A decade later, the instrument has been pulled from its box. The mechanism is the same. The maturities are the same. The CRR and SLR exemption is the same. Even the calendar is similar, with a window of just under four months. The differences are in the surrounding context.
Why now
The pressure that drove the RBI to reach for this lever has been building for months. India's foreign exchange reserves peaked at around 728 billion dollars in late February 2026, just before the Iran war broke out. By May, they had fallen to roughly 682 billion dollars, a decline of about forty six billion in three months, as the RBI sold dollars to slow the rupee's fall. The rupee itself has weakened by about seven per cent so far this year and now trades around 96 to a dollar, an all time low. Crude oil has remained stubbornly above one hundred dollars a barrel. Foreign portfolio investors have been net sellers through most of the war.
Underneath these headline numbers is a quieter problem. NRI deposit inflows, which had been running at over seven billion dollars in the financial year ended March 2025, collapsed to under a billion in the year just ended. The diaspora, historically a reliable source of dollar funding for the Indian economy in moments of stress, was not turning up. The reason was simple. US dollar interest rates remain elevated, with short dated Treasury bills offering more than four per cent. FCNR(B) deposits, weighed down by the hedging cost banks were absorbing, were offering rates that simply could not compete on a risk adjusted basis.
The RBI's intervention is a direct answer to that problem. Industry estimates suggest that the window could attract between fifty and seventy billion dollars over its four month life. That would more than offset the reserves lost since the war began and meaningfully change the external account picture by year end. The FCNR(B) window is the heaviest piece of artillery in a coordinated defensive package that also includes the repo rate held at 5.25 per cent, the gold import duty raised from six to fifteen per cent, restrictions on rupee forwards, continuous foreign exchange intervention, and even the Prime Minister's public appeal to Indians to avoid overseas travel and destination weddings.
The economics of the swap
It is worth pausing on what the RBI is actually giving up by running this scheme, because nothing in central banking is free.
Consider a stylised example. A bank raises one billion dollars through a three year FCNR(B) deposit at seven per cent, owing the depositor roughly 1.225 billion dollars at maturity after compounding. Under the RBI's swap, the bank immediately sells those dollars spot at, say, 96 rupees to the dollar, receiving 96 billion rupees that it can lend out. Simultaneously, it contracts to buy dollars back from the RBI three years later. The market forward premium for INR/USD currently sits around five and a half per cent per annum, reflecting the interest rate differential between the two currencies. The RBI's concessional rate is materially below that, reportedly in the range of one and a half to two per cent.
The mathematics matters. At a market forward rate, the bank would have to buy back dollars at approximately 96 multiplied by 1.055 cubed, or roughly 112.7 rupees to the dollar. Under the RBI's concessional rate, the swap rate works out to roughly 100 rupees to the dollar. The implied all in rupee cost of these funds is approximately 8.6 per cent, against domestic three year deposit rates currently around seven and a half to eight. The seven per cent dollar offer becomes tenable for banks, particularly given the CRR and SLR exemption that frees up the full deposit base for lending.
The subsidy embedded in the scheme is the gap between the market forward and the RBI's concessional rate, multiplied by the dollar amount. For every billion dollars mobilised, the RBI is implicitly transferring roughly 120 million dollars of present value to banks and, through them, to the diaspora. If the window attracts fifty billion dollars, the implicit balance sheet cost to the RBI is around six billion dollars. If the rupee depreciates sharply beyond the contracted forward over the deposit life, that cost rises further. It is, in economic substance, a cross border fiscal payment dressed in the garb of a central bank swap. Whether the trade is wise depends on what you think the alternative is. If the alternative is a disorderly rupee fall toward a hundred or beyond, the cost looks reasonable. If the alternative is an orderly depreciation the economy would have absorbed without intervention, the policy is expensive.
The two handed approach
There is a deeper analytical point hiding inside the FCNR(B) episode. India is currently doing two things simultaneously that, on a casual reading, look contradictory.
On one hand, the Reserve Bank has been quietly diversifying its reserves away from dollar assets and toward gold, with gold's share of total reserves nearly tripling over five years and the bulk of physical gold now held within India rather than in foreign vaults. The direction of this multi year drift is unmistakable. India is reducing its structural exposure to any single sovereign issuer of reserve assets.
On the other hand, the same Reserve Bank has just launched a programme explicitly designed to attract more dollars into the Indian banking system, even at the cost of accepting substantial subsidies on its own books. The policy is an all in wager on the dollar's continued primacy in the moment, crafted to attract precisely the currency India's longer term framing says it wants to bypass.
Both moves are real. Both are strategic. They look contradictory only if you assume the goal is to take a static position on the dollar. The goal is different. It is to maximise optionality. India wants less dollar exposure in its central bank reserves so that it cannot be politically squeezed in a crisis. But India also wants access to dollar funding when it needs it, particularly during balance of payments stress, and the diaspora is the most natural and politically uncomplicated source. Building gold reserves is a long horizon defensive move. Tapping the diaspora is a short horizon offensive one. They are not in conflict. They are two tools for two different purposes, and a serious central bank uses both. The country is acting like an institutional investor with a strategic allocation and tactical overlays, rather than like a state with a single foreign exchange ideology. There is something quietly impressive about this. It is the kind of policy posture that is easier to write about than to execute.
The mathematics for the NRI
A five year FCNR(B) deposit at seven per cent is a dollar denominated instrument paid by an Indian bank. The interest is exempt from Indian income tax under Section 10(15)(iv)(fa). There is no rupee conversion, which means no exchange rate risk to the depositor. The relevant comparison is therefore not the rupee fixed deposit rate, but dollar interest rates available where the NRI is resident, with the wrinkle that NRIs in the United States, the United Kingdom, Canada, and several other jurisdictions remain liable to tax on worldwide income at home.
For an NRI in the Gulf, where personal income tax is essentially nil, the entire seven per cent is retained. For an NRI in Singapore at a 22 per cent marginal rate, the post tax yield falls to around 5.5 per cent, still meaningfully ahead of US Treasuries. For an NRI in the United States, federal and state taxes can compress the post tax yield to around 4 to 4.5 per cent, at which point the comparison with a tax efficient Treasury becomes much closer. The RBI has engineered a domestic dollar interest rate high enough to overcome the inertia of NRIs in low tax jurisdictions, while remaining within the range Indian banks can absorb given the central bank's hedging subsidy.
The risks
Three things could go wrong, and they are worth flagging.
The first is the combined question of whether the rates are competitive enough across all NRI geographies and what the RBI is taking on in exchange. The 2013 window worked partly because US dollar interest rates were near zero. Today they are not, and the post tax differential in the largest NRI markets is meaningful but not enormous. At the same time, the central bank is taking on currency risk that the market would price at three to four per cent per year without charging the banks anywhere near that rate. If the rupee depreciates sharply beyond the contracted forward, the RBI realises a real economic loss at maturity. In a severe stress scenario, the window accelerates rather than arrests eventual reserve depletion.
The second is the refinancing cliff that the 2013 window's history warns about. The deposits mobilised now will mature in 2029 to 2031. If the global rate environment is hostile then, India may find itself defending the rupee again to roll over the deposits at unfavourable terms. Today's defence is tomorrow's vulnerability, on a predictable schedule.
The third is a quieter risk worth flagging. The US Treasury publishes a semi annual currency report that flags trading partners for exchange rate practices it deems inappropriate. Offering a central bank subsidised dollar yield to attract inflows that depress the rupee's natural depreciation path could, in a different political environment, attract attention. India is on relatively good terms with Washington at present, helped by the recent reduction in US tariffs from fifty to ten per cent. But the structure of the policy is the kind of thing that, in another era, would be characterised as currency manipulation.
None of these risks is fatal. They are the trade offs that any meaningful central bank intervention involves. But they deserve to be acknowledged rather than glossed over.
What to watch
The window closes on 30 September 2026 for new deposits, with banks able to access the RBI swap until 16 October. Three measurable indicators over the next four months will determine whether the policy is judged a success.
The first is the headline mobilisation number. The 2013 window raised thirty four billion dollars. The current optimistic estimates are in the fifty to seventy billion range. If actual mobilisation comes in below thirty billion, the policy will have helped at the margin but not transformed the picture. Above fifty, it will have meaningfully rebuilt reserves and given the rupee a foundation to stabilise on.
The second is what happens to the rupee. The defence of the currency is the entire point of the exercise. If the rupee can hold around 95 to 96 against the dollar through the autumn and avoid breaching one hundred, the policy will have done its job even if the deposit number disappoints. If the rupee continues falling despite the inflows, the FCNR(B) window is buying time but not solving the underlying problem.
The third is the forward premium spread. If the market three year premium widens further as US rates stay higher for longer, the RBI's concessional rate becomes an even larger subsidy, raising the question of whether the central bank has the stamina to keep the window open at that price.
The longer view
The FCNR(B) swap window is both a brilliant anachronism and a signal of strain. It reveals that India's external position, for all the improvements in macro fundamentals of the last decade, remains structurally dollar dependent at moments of pressure. The 2013 playbook is being dusted off not because the economy is identical, but because the toolkit for a liquidity hungry emerging market in a dollar world remains surprisingly small. The Rajan window of 2013 is remembered as a success because the rupee stabilised, reserves rebuilt, and the crisis passed. The current window may end up being remembered the same way. Or it may end up as one of several measures that, collectively, helped India navigate a difficult year without quite turning the tide. Either outcome is consistent with the policy being sensible. What would be inconsistent is the central bank doing nothing while the diaspora's dollars sat idle in foreign banks at lower yields. That is the option the RBI explicitly chose not to take.
India has a diaspora of over thirty million people, scattered across the world, holding meaningful savings in currencies the Indian economy needs. The 2026 revival of the Rajan playbook is, in part, an act of institutional memory. It is the kind of move that a central bank with continuity of thinking can make, drawing on a tested template rather than improvising under pressure. By October, we will have the first read on whether the bet is paying off. The numbers themselves will be public. The question for now is simply whether the diaspora shows up at the scale the central bank is paying it to show up.