$56.85 Billion in 67 Days: Inside RBI’s Unusual Dollar Deposit Gamble

By P. SESH KUMAR

RBI’s special FCNR(B) deposit window attracted $56.85 billion in just 67 days, but questions remain over the currency risk, eventual cost and abrupt early closure.

New Delhi, August 17, 2026 — In June 2026, the Reserve Bank of India (RBI) made an unusual offer. It told Indian banks: bring in dollars from Indians living abroad, and we will take care of the one expense that has always made such deposits unattractive to you.

Banks responded by raising the interest they paid on dollar deposits from about four per cent to as much as seven. Money poured in—USD 56.85 billion in sixty-seven days, most of it in deposits.

On 14 August the RBI shut the offer a month early, nine days after its Governor had said publicly that no such closure was being considered.

What happened, in plain words

Some government announcements arrive looking like a gift and leave looking like good sense. This was one of them.

Let us rewind to the first week of June 2026. The rupee was sliding. It had started the year at about 89.5 to the dollar and had slipped past 94.

India’s stock of foreign currency–the national reserve of dollars, euros and gold that the RBI keeps for emergencies–had peaked at USD 728 billion in late February and had been worn down to about USD 667 billion by the end of May, because the RBI kept selling dollars to slow the rupee’s fall.

War in West Asia had pushed crude oil above USD 100 a barrel, which makes everything we import more expensive.

And foreign investors who buy and sell Indian shares had been pulling out steadily–a net USD 9.6 billion in the April-to-June quarter alone.

None of this was a crisis. India was not going to run out of dollars. Reserves of USD 680-odd billion cover the country’s immediate foreign bills many times over. What India had was something subtler: a confidence problem, and a central bank that was tired of spending its dollar reserves week after week to hold the rupee steady.

There was one tap that had run dry. Indians living abroad can keep dollar deposits in Indian banks. In the year to March 2025, such deposits brought in USD 7.08 billion. In the year to March 2026, they brought in USD 946 million–a fall of 87 per cent.  Nobody abroad was interested, because the interest rate on offer was a limp three or four per cent.

So on 5 June, 2026 the RBI announced a scheme, and on 8 June it put out the rules. If a bank raised fresh dollar deposits from Indians abroad, for three to five years, the RBI would absorb the one cost that had been keeping the interest rate low.

The offer was open for deposits taken between 8 June and 30 September 2026.  Two similar offers were made for banks and government companies borrowing dollars abroad.

The response was extraordinary. Forty days in, USD 20.72 billion had come.  By 13 August, 2026 the figure was USD 56.85 billion–of which USD 52.30 billion was deposits from Indians abroad, and the rest borrowing by banks and public sector companies.

In one week alone, the country’s reserves jumped by USD 14.1 billion and crossed USD 707 billion for the first time.

And then, on Friday 14 August 2026,  the RBI shut the window. Deposits would be accepted only until 31 August, not 30 September 2026. Its stated reason was the “encouraging response”.

The awkward part: on 5 August 2026, at a press conference, RBI Governor Sanjay Malhotra had been asked directly whether the scheme might be closed early. He said flows were healthy and that no such proposal was being considered.  Nine days later it was closed. Nine days is not a long life for a public assurance.

Orbital Mirrors and the Sunlight Startup: Hype vs Reality

Four words, explained once

Before going further, let us decipher four bits of vocabulary–because everything that follows depends on them, and each is simpler than it sounds.

A deposit is a loan you make to a bank. You give the bank money, the bank pays you interest, and it returns your money on an agreed date. That is all it is. When an Indian abroad puts dollars into an Indian bank, he is lending dollars to that bank.

A hedge is protection against a price moving against you. If a bank takes in dollars today and must return dollars in four years, but spends the money in the meantime lending rupees to Indian businesses, it has a problem: if the rupee falls in those four years, it will need more rupees to buy back the dollars it owes.

Buying protection against that costs money, roughly 3 per cent a year in the summer of 2026. That cost is why banks could only offer 3 or 4 per cent to depositors –the rest was being eaten by protection.

A swap, in the form used here, is a two-part promise. The bank hands its dollars to the RBI today and receives rupees. Years later, the two reverse the trade. The crucial detail is the price of that reversal. Under this scheme, the RBI Bank promised to hand the dollars back at exactly the same exchange rate–which means that if the rupee falls in the meantime, the RBI swallows the loss, not the bank.  The banks’ protection expense simply vanished. That is the whole scheme in one sentence.

Reserves are the national savings account in foreign currency. When the RBI buys dollars, reserves rise. When it sells them to prop up the rupee, they fall. Reserves are the reason nobody panics.

What was actually being sold–and why it is not a bond

A great deal of the coverage called these “NRI bonds”. They are not bonds, and the difference matters enormously.

The instrument is called an FCNR(B) deposit–Foreign Currency Non-Resident (Bank). It has existed since 1993. An Indian living abroad, or a person of Indian origin, puts dollars (or pounds, euros, yen and a few others) into an Indian bank.

The deposit stays in that currency. He gets interest in that currency. He takes his money out in that currency, and he can take all of it out of India whenever it matures.

The interest is not taxed in India. Most importantly, he takes no currency risk at all: he put in dollars and he gets back dollars, whatever happens to the rupee.

Normally these deposits run from one to five years; under the 2026 scheme, only three-to-five-year money qualified, with the depositor unable to withdraw for the first year.

The RBI then made the deal irresistible from the banks’ side. It removed the currency risk through the swap. It exempted these deposits from the rules that normally force a bank to park a slice of every deposit with the RBI or in government bonds, which meant the bank could lend every last dollar of it.

And it allowed banks to lend money against these deposits.  The ceiling on interest was generous: for three-to-five-year money, banks could pay up to about three percentage points above the international reference rate.

The banks moved within forty-eight hours. AU Small Finance Bank took its top dollar rate from 5.15 per cent to 7.10. HDFC Bank, ICICI Bank and Axis Bank went to around six per cent.  For a doctor in Dubai or an engineer in Singapore earning four per cent on safe dollar savings, six or seven per cent from a large Indian bank was the best offer available anywhere in the world.

Now, why does the word “bond” matter? Because India has genuinely sold bonds to its diaspora before, three times, and each time it did so openly.

The India Development Bonds of 1991 raised USD 1.6 billion. The Resurgent India Bonds of August 1998, sold within weeks of the nuclear tests and the sanctions that followed, raised USD 4.2 billion.

The India Millennium Deposits of 2000 raised USD 5.5 billion. About USD 11.3 billion in all, issued by the State Bank of India, with the Government of India openly standing behind the currency risk.  Everyone who bought one knew he was lending to India in its hour of need, and everyone in India knew the government was borrowing.

In 2026 the government is carrying exactly the same risk. But it is carrying it inside a RBI ledger that nobody outside can read, behind something that looks to the depositor like a slightly generous bank fixed deposit.

One commentator has built his critique of the scheme on precisely this contrast, and although it rests on a single writer’s analysis, the underlying point is a matter of record: 1998 was honest about who was taking the risk, 2026 is not.

India’s Goldilocks Era Coming to an End? RBI Sees Inflation Risks

The case for the scheme–put as strongly as it deserves

It would be easy and lazy to attack the scheme without first admitting how sensible it was.

The RBI’s alternative was to keep selling dollars from the national reserve to hold the rupee up. That is the worst way to spend money in a falling market. It drains the reserve, it shows traders exactly where the central bank will fight, and it invites them to keep attacking until the money runs out.

The swap window did the opposite: instead of selling dollars out of the reserve, it brought dollars in. And the dollars it brought in were tied up for three to five years, unlike the money of a foreign share investor who can leave on a Thursday afternoon.

Raghuram Rajan, who ran the same scheme in 2013, later called it the “least bad” choice available on his first day as Governor–and was honest enough to add that nobody would ever know for certain how much of the calm that followed was actually caused by it.  That is the right register for this kind of policy: not triumph, but the choice of the least damaging option in a bad week.

It is also not obviously expensive. If the cost of the scheme is roughly 3 per cent a year, then India effectively borrowed medium-term dollars at a price no worse than what a nervous emerging economy would have paid to issue a government bond in the same weeks–and without the drama of a public bond sale and the ratings commentary that comes with it.

And the strongest defence of the early closure is this: a central bank that stops a costly scheme the moment it has done its job is behaving better than one that lets the meter run to the end of the calendar out of pride.

On that reading, calling the closure “puzzling” mistakes good management for confusion.  Most market economists took exactly that benign view–the money simply arrived faster than anyone expected.

Autopilot Parliament: Advocacy for Executive Discretion Disturbing

Where it went wrong

All of that conceded, there is a real flaw, and it is not the one the newspapers found.

The mistake was not stopping. The mistake was building something that could only be stopped by looking at a calendar.

We may think of it as a shop offering a discount. The RBI fixed the discount–a very deep one–and then opened the doors, with no limit on how many customers could walk in, no cap for any single bank, no plan for shrinking the discount as the crowd grew, no auction to see who would accept the smallest subsidy.

Once the interest rate on offer jumped from 4 per cent to 7, the world’s Indians were never going to be the limiting factor. Demand was effectively bottomless.

So when the central bank had taken in more than it wanted, the only lever it had left was to slam the door. And slamming doors is expensive for a central bank, though the cost is not paid in dollars.

It is paid in credibility. The Governor’s August 5 answer was not an offhand remark; it was a direct reply to a direct question about this exact possibility. Reversing it nine days later, without publishing a target that had been met or a cost that had grown uncomfortable, teaches the market to treat everything the RBI Bank says about its schemes as provisional. And it created a stampede: banks that had planned to gather deposits calmly over six weeks are now telephoning customers to squeeze it all into a fortnight, with the total possibly reaching USD 60–70 billion by the end of August.

Then there is the second problem: nobody has told us the price.

Here is the arithmetic in plain terms. The RBI has taken on the currency risk for USD 52.3 billion. The protection it is giving away free is worth roughly 3 per cent a year, which on that sum is about USD 1.8 billion a year, or something like USD 5–9 billion over the life of the deposits.

On top of that sits the real question–what the rupee will be worth in 2029, 2030 and 2031, when the RBI  must hand those dollars back at today’s rate. If the rupee behaves as it usually has, the loss on that promise could run into the tens of billions of dollars. One independent estimate puts it at USD 8–30 billion, with a working middle figure of USD 15–20 billion.

I would not defend those numbers to the last decimal, and that is exactly the point: the RBI has published nothing that would let anybody test them.

Why should a college student care? Because of where that bill eventually lands.

The RBI earns money, and every year it hands its surplus to the Government of India, where it becomes part of the money the government spends on everything from roads to scholarships.

A large loss on this swap ledger, years from now, means a smaller cheque to the government, which means either less spending or more borrowing.

It will never appear in a Budget document as the cost of the 2026 scheme. No Member of Parliament will ever vote on it. It will simply be there, quietly, in a smaller number.

(This is the first of the two-part series. Views expressed are the author’s own.)

Why the RBI Is Standing Pat While the Fed Signals Higher-for-Longer Rates

Follow The Raisina Hills on WhatsApp, Instagram, YouTube, Facebook, and LinkedIn

About The Author