Diaspora Bonds Explained: India's $11.3B Playbook, Israel, Nigeria

On 5 August 1998, State Bank of India opened the books on a five-year US dollar bond paying 7.75 percent, tax-free in India, and open only to people of Indian origin. It closed within weeks having raised about $4.2 billion. Two years later SBI did it again and pulled in $5.5 billion. Few governments have ever borrowed from their own emigrants at that scale, and the way India did it still shapes how India reaches for diaspora money in a crisis, including the deposit window the Reserve Bank opened this June.

This piece looks at what a diaspora bond actually is, why India's issues worked when others flopped, what Israel and Nigeria did differently, and why India has since switched from bonds to bank deposits as its emergency lever.

What a diaspora bond is, and what makes it different

A diaspora bond is a debt instrument a country (or a state-owned entity) sells to its own emigrants and their descendants. On paper it looks like any other sovereign or quasi-sovereign bond: fixed coupon, fixed tenor, repayment at maturity. The difference sits in who buys it and why.

In their influential paper on the subject, World Bank economist Dilip Ratha and Suhas Ketkar argue that emigrants price country risk differently from a fund manager in London. They have family in the country, they may plan to retire there, and they often see repayment in local currency as a feature rather than a risk because they expect to spend that money at home anyway. That gives the issuer two things institutional markets rarely offer: a lower coupon than its credit rating would demand, and access to money when global markets are closed to it.

The authors call the first effect a "patriotic discount". The second is arguably more valuable. India's 1998 and 2000 issues both came at moments when raising dollars through normal channels was expensive or impractical.

Diaspora bonds are also different from NRI bank deposits, which India relies on far more heavily today. A deposit sits on a bank's balance sheet and can be withdrawn under the account's rules. A bond is a fixed obligation. If you want the deposit side of the story, our guide to NRE, NRO and FCNR accounts covers how those work.

India's three issues, side by side

India has tapped its diaspora directly three times, always through SBI rather than the sovereign itself, and always for five years. The Ketkar and Ratha paper puts the proceeds at $1.6 billion for India Development Bonds in 1991, $4.2 billion for Resurgent India Bonds in 1998 and $5.5 billion for India Millennium Deposits in 2000, a combined $11.3 billion.

The RIB terms, summarised by chartered accountant Rashmin Sanghvi at the time, show how the product was built for a retail investor abroad. The minimum ticket was $2,000, £1,000 or DM 3,000. Interest could be paid half-yearly or rolled up. And the interest was exempt from Indian income tax, with wealth tax and gift tax exemptions on top. For a Gulf-based engineer who could not easily buy Indian government paper any other way, that was a clean, tax-efficient dollar asset with SBI's name on it.

The IMD followed in late 2000 with an 8.5 percent dollar coupon, which Business Standard notes was well above the NRI deposit returns available then. Redemption fell on 29 December 2005. RBI estimated the outflow at $7.3 billion and sold dollars to SBI to meet it, at Rs 45.34 to the dollar. (Business Standard quotes the amount raised as both $5.5 billion and $5.2 billion; the Ratha paper uses $5.5 billion.)

Why India's bonds worked

The Ratha and Ketkar analysis points to a handful of design choices that are easy to describe and hard to copy.

• A trusted issuer. SBI, not the government directly, sold the paper. Investors saw it as a "rupee-rich" bank that could always meet obligations. That perception mattered more than any rating.
• Distribution where the diaspora lives. SBI used its own US network and outsourced Gulf marketing to Citibank and HSBC. The bonds were sold where people banked.
• Simple terms. Fixed rate, bullet maturity, non-negotiable, small minimums. No one needed a broker to understand it.
• Know-your-customer at the point of sale. Restricting the bond to people of Indian origin made it a community product and kept regulators comfortable.

The trade-off is cost. These were not cheap loans. A 7.75 percent dollar coupon in 1998, tax-free, was generous by the standards of investment-grade issuers. India paid for speed and certainty, and the currency risk landed with SBI and the government. Under the IMD agreement, exchange losses from rupee depreciation were shared between the two.

A diaspora bond is cheapest for the country that needs it least, and most expensive for the one that needs it now.

Israel: seventy-five years of the same product

Israel is the model every finance ministry cites, and the scale is hard to match. The Development Corporation for Israel, which sells Israel Bonds in the US, says it has generated more than $55 billion in worldwide investments since 1951, including a record $5.7 billion in the two years after the 7 October 2023 attacks and $1 billion in the first 30 days alone.

What makes the Israeli programme different from India's is permanence. India issued three times, each in a squeeze. Israel issues every year, with a menu of fixed and floating instruments from one to 20 years and denominations from $100 to $1 million. The Ratha paper notes DCI ran roughly 200 staff in the US and registered its offerings with the Securities and Exchange Commission, which let it market openly to retail buyers.

The patriotic discount was real but has faded. Until the late 1980s, 10-to-15-year Israel bonds paid around 4 percent fixed while US 10-year Treasury yields averaged about 6.8 percent over the same period, according to Ketkar and Ratha. Today the bonds price much closer to market.

Worked example: what the patriotic discount actually costs a buyer

Take the two programmes at their historical terms and put $10,000 into each.

Israel, pre-1990 style. A 10-year bond at about 4 percent pays roughly $400 a year, or $4,000 over the life of the bond in simple interest. A Treasury at the 6.8 percent average Ketkar and Ratha cite would have paid about $680 a year, or $6,800. The gap of roughly $2,800, around 28 percent of the principal, is the gift the buyer made to Israel's treasury in exchange for the sense of contributing.

India, RIB 1998. $10,000 in the dollar tranche at 7.75 percent, with interest compounded half-yearly for five years, grows to about $14,625. Because that interest was exempt from Indian income tax, a non-resident buyer kept all of it on the Indian side (tax in the country of residence was a separate question). There was no discount here; if anything, India paid a premium.

The lesson is that "diaspora bond" covers two very different deals. One is a charitable loan with a coupon attached. The other is a well-priced crisis instrument where the diaspora's real contribution is speed and willingness to buy when no one else will. Any reader weighing an instrument like this today should take the tax and currency questions to a qualified adviser; our explainer on NRI tax on money sent to India covers the basics.

Nigeria and Ethiopia: one clean execution, one cautionary tale

Nigeria's 2017 issue is the best recent example from Africa. The Debt Management Office sold a $300 million five-year bond at 5.625 percent, registered with both the US SEC and the UK Listing Authority and sold through private banks and wealth managers. Orders peaked near 190 percent of the offer before settling at about 130 percent. The bond matured and was redeemed in June 2022, which the DMO used as evidence of its reliability.

The registration point is the one most governments miss. Ethiopia's state utility sold bonds for the Grand Ethiopian Renaissance Dam to Ethiopians in the US without registering them. In 2016 the SEC found it had raised about $5.8 million from more than 3,100 US residents between 2011 and 2014 in violation of securities law, and the utility agreed to pay back nearly $6.5 million including interest. An earlier Ethiopian diaspora bond had already sold slowly; a World Bank blog pinned that on doubts about the utility's ability to repay, low trust in the guarantor and political risk.

Put the two side by side and the pattern is obvious. Patriotism raises interest. Registration, a credible issuer and distribution through institutions people already use turn interest into money.

Why India switched from bonds to deposits

India has not issued a diaspora bond since 2000. When the rupee came under pressure in 2013, RBI chose a different tool: it offered banks a concessional dollar swap for fresh FCNR(B) deposits of three years or more. According to Business Today, that 2013 scheme brought in about $34 billion in total, including roughly $26 billion of FCNR(B) deposits.

The logic is practical. A deposit scheme runs through every bank's existing NRI desk, needs no prospectus abroad and avoids foreign securities registration entirely. The bank carries the deposit; RBI absorbs the hedging cost through the swap. The NRI gets a foreign currency deposit with no rupee risk, which is closer to what many buyers of RIBs and IMDs wanted anyway. A 2022 industry proposal for "India Millennial Bonds" never went anywhere, partly because, as one economist told Business Standard, such a move could signal that something was wrong.

RBI reached for the same tool again on 8 June 2026, after heavy foreign portfolio outflows tied to the West Asia conflict pushed the rupee lower. The new window covers fresh FCNR(B) deposits of three to five years mobilised between 8 June and 30 September, with RBI bearing the full hedging cost. Early take-up has been far stronger than in 2013. RBI data reported by Business Today showed $40.8 billion raised under the scheme by 31 July, of which $36.7 billion was FCNR(B) deposits, already above the entire 2013 FCNR haul.

For background on how these flows fit India's wider relationship with its emigrants, see our piece on the global Indian diaspora economy and our guide to NRI investment routes into India.

What to watch next

• 30 September 2026: the cut-off for fresh FCNR(B) deposits to qualify for RBI's swap. The final tally will show whether this window becomes the largest diaspora mobilisation India has run.
• 16 October 2026: the date to which RBI said the swap facility itself would stay open for banks.
• 2029 to 2031: the three-to-five-year deposits raised this year fall due. India has been here before: the IMD redemption in December 2005 required RBI to supply about $7.3 billion. A deposit wall many times that size will need the same planning, and readers should watch how RBI phases it.
• Whether anyone in Africa or South Asia issues a registered, retail diaspora bond in 2026-27. Nigeria's 2017 template is there to copy. The FEMA rules for NRIs and host-country securities law will decide where Indian-origin investors can even buy one.

The deeper point is that diaspora money is not patient by default. It is patient when the issuer is trusted, the terms are simple and the product is sold where the diaspora already banks. India learned that in 1998. The 2026 numbers suggest it has not forgotten.

About this researchMuffinPay InsightsIndependent analysis of stablecoins, CBDCs, cross-border payments and digital public infrastructure. Every figure is linked to its source. Read our editorial standards →