India’s $56 Billion Dollar Deposit Drive: The Cost May Come Due in 2031
RBI headquarters in Mumbai. (Image RBI on X)
By P. SESH KUMAR
The RBI’s dollar deposit window attracted billions in foreign currency, but questions remain over the true size of fresh savings, the cost of the currency protection and the repayment wall looming in 2029-31.
New Delhi, August 18, 2026 — India’s latest dollar-deposit drive may have delivered an impressive headline number, but the bigger question is what that money will ultimately cost. With billions of dollars mobilised in a matter of weeks, the scheme has bought the RBI time and strengthened its foreign-exchange position. Yet the liabilities, rupee liquidity created in the process and the concentration of repayments in 2029-31 raise questions that the headline inflow does not answer. The real test, therefore, is not how much money came in, but how much of it represents genuine new savings, what the RBI paid to attract it and whether India is once again merely buying time.
There is a third and quieter consequence. When a bank hands the RBI dollars, it receives rupees– newly created rupees. At about ₹95 to the dollar, USD 52.3 billion means roughly ₹5 lakh crore of new money pushed into a banking system that already had more cash than it could use. (That conversion is my own arithmetic, not an official figure, but nobody disputes the multiplication.) The RBI now has to mop that money back up, which is a further cost, and that one has not been disclosed either. Bank of Baroda’s chief economist said as much, bluntly, on the day the window closed.
We have done this before: 2013, and the bill in 2016
In September 2013 the rupee had crashed to 68.85 and the world’s money was fleeing emerging markets. The RBI offered banks almost the same deal–cheap protection on dollar deposits of at least three years, at a fixed 3.5 per cent when the market price was above six. In three months the twin windows drew USD 34 billion, of which USD 26–27 billion was deposits.
Two lessons are usually taken from that episode. Only one of them is the important one.
The lesson everybody repeats is that repayment turned out to be easy. It did. When the deposits matured in late 2016, the RBI had already arranged protection for about 80 per cent of the amount, sat on USD 365 billion of reserves, and had told everyone months in advance what it planned to do. More than nine-tenths of the money went out without a ripple in the currency market–the stock of these deposits fell from USD 44.11 billion in September 2016 to USD 20.85 billion in December–and almost nobody noticed, because the country was busy standing in queues for demonetisation.
The lesson almost nobody repeats is about what the money actually was. A large part of the 2013 inflow was borrowed money going round in a circle. A foreign bank would suggest to a wealthy Indian abroad that he open one of these deposits, then immediately lend him most of the value of that deposit against it, sometimes repeatedly. On paper, dollars flowed into India. In reality, a modest amount of genuine savings had been inflated several times over by borrowing. An economist warned at the time that such money can leave as suddenly as it arrives, and by 2016 bankers confirmed that much of the “repayment” was simply those loans unwinding.
The same machinery is running again, and possibly on a bigger scale: the rules expressly permit banks to lend against these deposits, and there are reports–from one legal commentary quoting a newspaper clipping, which is as thin as it sounds–of nine times leverage being offered to wealthy non-resident clients through an offshore branch in Gujarat. We may however, treat the specific multiple with suspicion. But let us not doubt the mechanism, which wealth managers have been advertising openly. It means the headline of USD 52.3 billion and the amount of real new savings India has attracted are two very different numbers.
So did it work? It depends on what you measure
This is where the auditor’s instinct differs from the headline writer’s. “USD 56.85 billion mobilised” is not a result. It is a count of money that came through a door, reported by the very people who benefit from the number being large.
A serious assessment would ask, first, how much of it is new. We need to subtract the borrowed round-tripping. Also subtract the money that shifted from one kind of Indian bank account to another. Further, subtract the dollars the diaspora would have sent anyway. The April-to-June figures already offer a warning: deposits from Indians abroad brought a net USD 2.8 billion that quarter, which is less than the USDC3.6 billion of a year earlier–though the window only opened on 8 June, so the real test comes later.
It would ask, second, what each dollar cost: the free protection, plus the expense of mopping up the new rupees, plus the eventual loss when the dollars are handed back–divided by the dollars that genuinely stayed. And it would compare that with the cost of simply selling reserves instead, which is the honest alternative.
Third, it would check the arithmetic against the story. Reserves went from about USD 667 billion at end-May to USD 707 billion on 7 August–roughly USD 40 billion–while USD 52 billion supposedly came in. The gap is probably explained by the RBI continuing to sell dollars, by the changing value of gold and other currencies, and by timing. That is my inference from published numbers, not an official reconciliation. But a central bank should close that gap in its own words rather than leave citizens to guess.
Fourth, it would look at the rupee, honestly. The rupee closed at 95.42 on the day the window shut–essentially unchanged. Fifty-two billion dollars did not turn the currency around. The most that can be claimed is that it prevented something worse, and that is a claim about something nobody can observe.
Fifth, it would look at when the money leaves. This is three-to-five-year money, which means a wall of repayments falling between mid-2029 and mid-2031, at a time when the free protection that created it no longer exists. Whether that wall is renewed or repaid is the actual test of whether 2026 was clever, and the answer arrives long after the people who designed the scheme have moved on.
Sixth, it would notice what kind of money this is. These deposits are debt: money India owes to foreigners, in foreign currency, on a fixed date. That is precisely the category India spent the thirty-five years after 1991 learning to be careful about. Fifty-two billion dollars of it in eleven weeks is not a small change in the shape of the country’s obligations.
And finally, it would ask the uncomfortable question about what this money is doing. One well-known investor’s charge is that the scheme does nothing about India’s real problem and merely provides the dollars that foreign investors are using to leave. It is one man’s opinion and it is arguable. But it points at the right thing. Fifty-two billion dollars of borrowed diaspora savings is not the same event as fifty-two billion dollars of foreign investment in factories. In April-to-June, direct investment brought in a net USD 7.8 billion, better than the USD 4.8 billion of a year before, but modest beside the USD 9.4 billion that foreign companies took out. Deposits buy time. Investment builds capacity. We cannot argue when anyone says India keeps buying time.
India’s Goldilocks Era Coming to an End? RBI Sees Inflation Risks
If you are an Indian abroad, should you have put money in?
For a large group of people, yes–with three warnings that matter more than the interest rate.
The attraction is real and easy to state. Six or seven per cent, in dollars, from one of India’s biggest banks, with no exchange-rate risk to you, no Indian tax on the interest, and the freedom to take the whole thing out of the country when it matures. Compared with an American government bond paying a little over four, or a dollar deposit in a Western bank, that is a genuinely better deal for the risk involved. The absence of currency risk is the single most valuable feature and the one most people undervalue–it is what separates this from a rupee deposit, where a high interest rate is quietly eaten by the rupee’s decline.
The first warning is tax at home. India not taxing the interest does not mean anybody else won’t. An Indian living in the United States pays American federal and state tax on this interest as ordinary income and must report the account to the US authorities. A gross yield of six and a half per cent becomes something nearer four after tax–at which point the advantage over an American government bond has largely disappeared. The chief executive of HDFC Bank said publicly that tax rules make the scheme unattractive to many Indians outside West Asia and Singapore, and that inflows might therefore fall short of the USD 60–80 billion being talked about. That report reaches me second-hand, so we may treat the figure loosely–but the tax logic is beyond doubt, and it explains why the Gulf, where there is no personal income tax, supplied so much of this money.
The second warning is you cannot get out easily. The money is locked for a year. After that, early withdrawal is at the bank’s discretion, and since the bank cannot cancel its own arrangement with the RBI, it has every reason to say no. If there is any chance of needing the money, arrange an overdraft facility against the deposit at the time of opening it, rather than discovering the problem later. And do not comfort yourself with deposit insurance: it covers ₹5 lakh, which is loose change at the sums involved here.
The third warning is borrowed money. The structure being marketed to wealthy clients–borrow dollars cheaply abroad, put them in at seven per cent, pledge the deposit as security, repeat–turns a conservative deposit into a leveraged bet. It works beautifully until the lender raises its rate, or demands more security, or a regulator changes the pledging rules. If you want a safe dollar return, take the deposit and borrow nothing against it. If you want the leveraged bet, at least be clear that is what you have bought.
Best case, worst case
For India, the best case is genuinely good. Crude oil softens, the trade gap stays modest, and foreign money starts flowing back in; one rating agency has already projected that India’s overall external accounts swing into surplus this financial year. The rupee steadies, imported inflation eases, and by 2029-31, when the RBI must return those dollars, Indian inflation has come down enough that the exchange rate has not moved much–so the loss on the promise is small. By then, real foreign investment is arriving. The scheme is remembered as a cheap, clever bridge across a bad patch.
The worst case is just as easy to imagine. The rupee keeps sliding at its usual pace. In 2029-31 the RBI hands back dollars having received rupees at 95, and books a loss running into the tens of billions of dollars–which shrinks the cheque it writes to the government at a moment when the Budget has come to rely on that cheque. A good part of the USD 52 billion turns out to have been borrowed money, which leaves at the first opportunity, forcing either a fresh subsidised window– the sixth since 1991, and each one teaches the diaspora to sit and wait for the next–or another drawdown of reserves into a nervous market. The rupees created along the way prove hard to mop up, at a cost never separately disclosed. The medicine that bought time becomes the reason time runs out.
For the depositor the risks are gentler but not absent. The best case is that he has locked in seven per cent in dollars just as American interest rates start falling, which is an excellent trade. The worst case is not that an Indian bank fails; it is the combination of tax at home, money locked up for years, and–in the genuinely unlikely tail–a future change in India’s rules about taking money out or pledging deposits, something India has not attempted since 1991 but has never promised never to do. For the depositor who borrowed to invest, the worst case is a demand for more security in a currency he does not control.
Why the RBI Is Standing Pat While the Fed Signals Higher-for-Longer Rates
What to learn, and what to do next
Five things follow, and not one of them needs a new law.
Publish the price. The RBI should tell the country, in its annual report and in a standing note, how large this swap book is, what the free protection is worth, what the position is worth at each reporting date, and what mopping up the rupees has cost. A bill of this size, incurred without Parliament voting on it, should not be something outsiders have to estimate from press reports. This single change would improve every future debate on the subject.
Build the brake before you open the tap. Any future window should have a ceiling on the total, a share for each bank, and either an auction–so that banks bid for the subsidy and the cheapest bid wins–or a published formula for shrinking the discount as reserves fill up. A deep discount with no limit and only a calendar to stop it will always end in an abrupt Friday evening announcement.
Spread out the repayment dates. A single wall of maturities in 2029-31 is a wound India has inflicted on itself for no reason. Encourage a spread of tenures, and announce in advance the terms on which deposits may be rolled over. A wall becomes a slope.
Count the borrowed money. If banks may lend against these deposits, the RBI should collect and publish how much of the inflow was leveraged, and cap the multiple. Without that number, “USD 56.85 billion” tells us very little–and the central bank is choosing not to find out something it could easily find out.
Finally, and least comfortably: treat the illness, not the fever. Every one of these windows –1991, 1998, 2000, 2013, 2026–has been a response to the same underlying condition. India needs foreign money; it attracts too much of the kind that can leave on a Tuesday and too little of the kind that builds a factory and stays twenty years. Net foreign investment of USD 7.8 billion in a quarter, against USD 9.4 billion taken out by foreign companies, is the real problem statement. A diaspora deposit window is a good tourniquet. It is not surgery. Thirty-five years of reaching for the same tourniquet should prompt a different conversation about the wound.
Was the scheme thought through? The part that brought money in was thought through brilliantly. The part that had to switch it off was not thought through at all–which is the signature of a policy designed by people who were confident about demand and untroubled by any duty to explain the cost. Fifty-six billion dollars arrived in sixty-seven days. India will find out what it paid for them in 2031, and by then almost nobody will connect the two.
(This is the second of the two-part series. Views expressed in the article are author’s own.)
$56.85 Billion in 67 Days: Inside RBI’s Unusual Dollar Deposit Gamble
Follow The Raisina Hills on WhatsApp, Instagram, YouTube, Facebook, and LinkedIn
About The Author
