Purpose-Bound Money Explained: Singapore's PBM and Smart Subsidies

In November 2022, visitors to the Singapore FinTech Festival could scan a QR code and receive a voucher worth 20 digital Singapore dollars, spendable at selected food outlets. When they bought lunch, the hawker did not receive a voucher. The stall received plain tokenised money, almost instantly, with no reconciliation file to send to anyone. That small trial, run by DBS with the government's Open Government Products team, is the clearest demonstration of what the Monetary Authority of Singapore calls purpose-bound money: rules that ride on top of money, and fall away once the rules are met.

It sounds like a technical footnote. It is actually a new answer to one of the oldest questions in public finance: how do you give someone money that can only be spent on what it was meant for, without building a separate bureaucracy for every scheme?

What purpose-bound money is, and what it is not

The phrase "programmable money" causes confusion because it implies the money itself changes. PBM is designed to avoid that. In the MAS model, there are two layers:

• The money underneath. A digital form of an ordinary currency: tokenised bank deposits, a regulated stablecoin, or, in trials, wholesale central bank money. It stays fully fungible.
• The conditions on top. A wrapper that says who can receive it, where it can be spent, and until when. Validity periods and permitted merchant types were the examples MAS gave when it published its PBM whitepaper in June 2023, as reported by Mondo Visione.

When the conditions are satisfied, for example when a recipient pays an approved merchant, the wrapper is removed and the merchant receives unrestricted money. That is the key design choice. A merchant does not want a pile of tokens that only work in one scheme. A merchant wants money.

This separation matters for anyone comparing PBM with central bank digital currencies. PBM is not a CBDC and does not require one. It is a protocol that can sit on top of several kinds of digital money, which is why MAS later built it into infrastructure that also handles tokenised deposits and regulated stablecoins. Our explainer on stablecoins, CBDCs and tokenised deposits sets out those layers.

How Singapore built it, step by step

Singapore's route to PBM ran through Project Orchid, MAS's exploration of a digital Singapore dollar.

1. November 2022: four live trials. Announced at the FinTech Festival, the trials covered disbursements by DBS and Open Government Products, commercial vouchers by Temasek, FAZZ and Grab, government payouts to unbanked recipients by OCBC with the Central Provident Fund Board, and SkillsFuture grants released by UOB to training providers once eligibility conditions were met, according to Tech Wire Asia.
2. June 2023: the PBM whitepaper. Developed with the IMF, Banca d'Italia and the Bank of Korea plus industry participants, it set out a lifecycle from issuance to redemption, an interface to the digital currencies underneath, and open-source code. Pilots tested escrow for online commerce with Amazon, FAZZ and Grab, and programmable rewards with DBS, Grab, FAZZ, NETS and UOB, as Forkast also covered.
3. November 2023: the Orchid Blueprint. MAS described four building blocks for a digital money system: a settlement ledger, a tokenisation bridge to existing account systems, a programmability protocol using PBM, and a name service to replace wallet addresses with readable identifiers, Electronic Payments International reported.
4. October 2025: BLOOM. MAS launched an initiative to extend settlement using tokenised bank liabilities and regulated stablecoins across G10 and Asian currencies, with programmable compliance controls and preset rules for agent-initiated payments. The consortium includes Circle, DBS, OCBC, UOB, Partior, Stripe, Coinbase, Ant International and StraitsX, according to Charltons Quantum.

The direction is telling. PBM started as a tool for government vouchers and ended up as a general-purpose control layer for wholesale payments and even AI agents. The government use case was the proving ground.

Purpose-binding already exists. What changes?

Restricted money is not new. Food stamps, school vouchers, gift cards and India's fertiliser subsidies all bind money to a purpose. India even has a digital version: NPCI's e-RUPI, a prepaid voucher sent by SMS or QR code for a specific purpose and redeemed at participating merchants, built with the Department of Financial Services and the National Health Authority. We cover it in detail in our guide to e-RUPI digital vouchers.

What PBM changes is the cost of each new scheme. Open Government Products, which ran Singapore's RedeemSG voucher system for campaigns including CDC vouchers, has written that the old model had three problems: high implementation cost per scheme, slow and costly merchant settlement, and heavy contracting effort, as its engineering blog explained. Under RedeemSG, merchants could be paid the next day. In the PBM trial, settlement was near-instant, against one to two days for conventional voucher reimbursements, according to The Edge Singapore, which also noted the pilot could later extend to Singapore's 28,000 coffee shops, hawker centres and restaurants.

The gift PPI rules come from the RBI's master direction on prepaid payment instruments, which caps gift instruments at ₹10,000, bars reloading, and allows no cash-out except back to the source account.

Worked example: a uniform subsidy, two ways

Consider a hypothetical state scheme that gives ₹1,500 for school uniforms to each of 2,00,000 students. The budget is ₹30 crore.

As a cash DBT: ₹30 crore is credited to parents' accounts in one day. The state has no idea how much buys uniforms. Some will go to rent or debt repayment, which may be a perfectly reasonable household choice.

As a PBM voucher: The same ₹30 crore is issued as vouchers valid at, say, 400 registered tailors and uniform shops, expiring after 90 days. Three numbers now decide whether the scheme works:

• Redemption rate. Unspent vouchers return to the treasury. If redemption is 85%, ₹25.5 crore reaches shops and ₹4.5 crore lapses. Is that a saving or a failure? It depends on whether students simply could not find an approved shop.
• Merchant coverage. With 400 shops and 2,00,000 students, each shop serves 500 students on average. In a rural block with one approved tailor, the restriction becomes a queue.
• Merchant cash flow. A tailor who fills 500 orders at ₹1,500 has ₹7.5 lakh tied up. Near-instant settlement instead of a reimbursement claim is the difference between buying cloth on credit and buying it outright.

India's existing voucher data shows why the first number is not academic. NPCI's published e-RUPI statistics, which give vouchers created and redeemed by issuing bank, show about 2.94 lakh vouchers created and about 45,000 redeemed across six banks, roughly 15% by count. The page does not give a period or values, so it is not a precise redemption rate, but it is a warning that issuing a voucher and getting it spent are different achievements.

Every condition you code into money is a new way for the money not to arrive.

The risks are mostly about people, not code

• Exclusion by design. A restriction that excludes the wrong merchant, or expires before a harvest-season shopping trip, removes value from exactly the households a scheme targets.
• Fungibility. If a family already spends more than the voucher on food, binding the voucher to food changes nothing; it simply frees cash for something else. Restrictions only bind when the subsidy exceeds normal spending.
• Paternalism. Households usually know their own priorities better than a scheme designer does. PBM makes it cheap to override that judgement, which is a political temptation as much as a technical feature.
• Privacy. A programmable voucher can, by design, record where every rupee or dollar was spent. Who sees that trail is a governance question we take up in the CBDC privacy debate.
• Merchant capture. Approved-merchant lists create rents. Whoever controls onboarding controls a guaranteed customer base.

The strongest case for PBM is narrow and real: business-to-government flows such as grants released to a training provider once a course is completed, escrow in procurement, and subsidies to merchants that are now settled through slow claims. The weakest case is restricting what poor households can buy.

What this means for India

India already has most of the pieces: a voucher instrument in e-RUPI, a prepaid instrument framework at the RBI, and a retail e-rupee pilot in which programmability has been part of the RBI's testing agenda. What it lacks is a single protocol that lets a ministry define conditions once and have them work across banks, wallets and merchants. Our analysis of programmable e-rupee use cases looks at where that is heading. The DBT system shows the stakes: India moves tens of lakh crore through cash-style transfers. Even a small shift toward conditional instruments would be the largest PBM deployment anywhere.

What to watch next

• Singapore FinTech Festival, November 2026. MAS announced both the 2022 trials and the 2023 Orchid Blueprint at the festival; expect updates on BLOOM trials and any government payout using PBM.
• BLOOM trial results. Watch whether programmable compliance controls go live in cross-border corporate payments during 2026-27, which would be PBM's first large commercial use.
• India's Union Budget, expected 1 February 2027. Any announcement on conditional or voucher-based delivery for health, fertiliser or education subsidies would point toward purpose-bound designs.
• RBI e-rupee pilot updates. Programmability features tested in the retail pilot are the most likely route for PBM-style instruments in India.

This article explains payment designs and public policy. It is not legal or financial advice.

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 →