The state as the largest payer and the largest collector
Public sector payment flows cover every disbursement a government body makes and every collection made on its behalf. A government pays salaries, pensions, benefits, student grants, and farm subsidies. It collects taxes, fines, customs duties, school and hospital fees, and passport fees. No merchant in the world operates at that scale. These flows run on the same cards and credit transfers as commerce. The rules that govern them, however, come from statutes and regulations, which set the permitted channel, the applicable fee, and the form the proof of payment takes.
The state’s position in the payment chain differs from a merchant’s. A merchant chooses its provider, negotiates its fees, and designs its checkout flow. A government agency rarely has any of those three freedoms. The law dictates the channel a transaction must use. The price of the service is an administered tariff. The counterparty is a taxpayer or a benefit recipient, who can neither go elsewhere nor walk away from the transaction. Merchant acquiring methods, built on choosing a provider and negotiating fees, therefore do not apply to these flows.
| Card type | Direction | Main constraint | What breaks in practice |
|---|---|---|---|
| G2P, government to individuals | Public sector salaries, pensions, social assistance, allowances | Reaching recipients, often unbanked, on a fixed date and at scale | The last mile: no account, the cash-out point is too far away, identity cannot be verified |
| P2G, individuals to government | Taxes, fines, school meals, university fees, hospital bills, licenses | Identifying the debt unambiguously and proving payment | Reconciliation: the payer pays the right amount against the wrong debt, or pays twice |
| B2G, businesses to government | VAT, social security contributions, customs duties, fees | Matching the tax return and the certified invoice | A gap between the payment flow and the tax flow, which are checked separately |
| G2B, government to businesses | Payments on public contracts, subsidies, tax refunds | Statutory payment term and mandated invoicing channel | An invoice sent outside the official channel does not start the clock, and the supplier does not know it |
pagoPA: the mandatory public hub model
pagoPA is the Italian public platform through which all payments to public administrations flow. It routes payment orders and reconciles settlements, but it is not a scheme, a wallet, or a payment provider. PagoPA S.p.A., an Italian state-owned company, has run it since 2016. The platform holds no funds. It shows citizens no price. It imposes three things: the payment notice format, the debt identifier, and the debt’s life cycle. Of all comparable public systems, it goes furthest in placing the state between creditor agencies and the payments industry, because routing through this single hub is mandatory.
The obligation to use the platform is set by statute. Article 65(2) of Legislative Decree 217/2017, as amended by Article 24 of Decree-Law 76/2020, set the effective date at February 28, 2021. Since then, no payment service provider authorized in Italy may execute a transaction whose payee is a public administration outside pagoPA. The ban applies to the provider itself, including a foreign provider operating under the EU freedom to provide services. A license obtained in another member state does not exempt it.
The business model pairs an infrastructure monopoly with retail competition. The platform itself is a monopoly, but the market for services built on top of it stays open. Each member provider sets its own fee to the payer. That fee is displayed before the payer confirms the transaction. The payer picks a channel from all the members, so the display puts fees in direct competition. The state pooled the protocol and left the commercial margin to private firms. Every member can access the debt. Profitability on Italian public flows therefore depends on the fee a provider can sustain against competing channels.
Paying millions of recipients who have no bank account
A social payment is a benefit paid by a government agency to an eligible individual. The state knows who the recipients are, but it does not necessarily have an up-to-date directory of their bank details. Some recipients have no account, live far from a branch, change phone numbers, or cannot provide proof of address. Yet the payment must go out on a fixed date to every recipient, and each transaction must leave an audit trail. The channels used around the world fall into dedicated payment instruments, mobile money wallets, instant rails addressed by identity, and over-the-counter payouts.
| Channel | Real-world examples | What it solves | What it costs |
|---|---|---|---|
| Dedicated government card | Meeza (Egypt, 2019), e-zwich (Ghana, 2008), Qi Card (Iraq, 2007), Korti Milli (Tajikistan, 2017), Direct Express (US, 2008) | A state-issued instrument that works whether or not the recipient is banked | A terminal base to deploy and an acceptance network to build from scratch |
| Mobile money | M-PESA (Kenya), MTN MoMo, Orange Money, Wave, Wizall Money (WAEMU) | Immediate rural reach through agent networks, with no bank branch needed | Cash-out still carries a fee, and agent liquidity determines whether the payment gets through |
| Instant rail addressed by identity | APBS on NACH (India), PromptPay (Thailand), Pix (Brazil), SIPS (Somalia, 2025) | 24/7 payments addressed by national ID rather than by account number | Requires a queryable identity registry and reliable ID-to-account mapping |
| Counter and cash | Post offices, bank branches, authorized service points | Works without a digital ID or a phone | Cash-in-transit costs, lines, and risk of diversion at the service point |
A dedicated government card is a payment instrument issued by the state or under its authority, whether or not the recipient is banked. Governments use one when they also want to build a domestic card scheme, as Egypt did. The Central Bank of Egypt made Meeza the channel for public sector salaries, pensions, and subsidies. The card base passed 43.5 million cards in June 2025 (Central Bank of Egypt, 2025). Ghana chose a different technology for the same job. e-zwich, run by GhIPSS since 2008, authenticates by fingerprint and works offline, with no bank account required. It carries public sector salaries, national service allowances, and social programs. Iraq and Tajikistan run similar systems, the Qi Card and Korti Milli.
In the US, the same need produced two distinct instruments, which are often confused. The first, Direct Express, is a Bureau of the Fiscal Service program launched in 2008. The prepaid card has about 3.4 million cardholders (US Treasury, 2025). The federal government uses it to pay Social Security retirement benefits, disability benefits, and veterans’ pensions to recipients without a bank account. The second, EBT, carries the QUEST acceptance mark. It is run by the states and delivers SNAP food assistance under USDA oversight. A US grocery retailer must be certified for EBT separately, with its own product eligibility and routing rules.
Mobile money is a payment service built on an e-money account that the user accesses from a phone. It serves areas with no bank branches and no card acceptance. AfricaNenda published its SIIPS 2025 report on November 13, 2025, with the World Bank and UNECA. It counts 36 live instant payment systems in 31 African countries, which processed 64 billion transactions worth nearly $2 trillion in 2024. It names government payments as one of the two levers for reaching scale. The GSMA, for its part, counts 347 million active accounts out of about 1.2 billion registered (SOTIR 2026). The gap between the two figures suggests that growth lies in reactivating existing accounts rather than opening new ones.
Wizall Money is an e-money institution licensed by the BCEAO, the West African central bank, in Senegal, Côte d’Ivoire, Burkina Faso, and Mali. It focuses on bulk payments by companies and governments, such as salaries, allowances, and social assistance, rather than person-to-person transfers. Recipients can collect these payments without a bank account, a subscription, or a specific mobile operator. A simple SMS code is one of the accepted methods. The Senegalese government used the wallet to deliver its emergency cash transfers. Unlike Orange Money, its business model does not depend on a mobile network operator.
National ID as a payment address
Identity-based addressing means designating the recipient of a credit transfer by national ID rather than by bank details. In India, social payments are addressed to an Aadhaar number, the 12-digit identifier issued by UIDAI. The rail is the Aadhaar Payment Bridge System (APBS), which runs on NPCI’s National Automated Clearing House (NACH). This lets the state pay hundreds of millions of people without maintaining their bank details. The link between identifier and account sits in an NPCI registry, which banks populate by declaring the mapping for their customers. No other country runs a subsidy program at this scale on identity-based addressing.
In August 2026, India’s official direct benefit transfer dashboard showed a cumulative ₹52.65 lakh crore paid out. It lists 320 schemes across 56 ministries and estimated savings of ₹5.14 lakh crore (dbtbharat.gov.in). The savings figure is a government estimate of leakage avoided by removing ghost and duplicate recipients. It is the program’s main political argument. It remains an estimate, not a direct measurement of the sums that would otherwise have been diverted.
Several other countries use a government identifier as a payment address. In Thailand, National ITMX has operated PromptPay since 2017 under a mandate from the Bank of Thailand. It accepts a mobile number, a national ID number, or a corporate tax ID as an alias. In Singapore, PayNow accepts the NRIC for individuals and the UEN for entities. In Brazil, a Pix key can be a CPF or a CNPJ, the tax IDs of individuals and companies. In all four cases, the state turned an administrative identifier into a payment address, so it can route a public payment without first collecting bank details.
This form of addressing works only if an identity registry can be queried online. The SIIPS 2025 report makes the link explicit. Without verifiable identity, a large share of users stay at the lowest tier of regulatory limits, which keeps government payments from reaching scale. Digital social payments therefore require an identification system first. Both the routing of the payment and the limit that applies to the recipient depend on it.
Collecting taxes: standardized notices and QR codes
Public collection is the receipt, by a government agency or a third party acting on its behalf, of amounts owed as a tax, a fine, or a fee. It applies to an individual debt whose amount and debtor are set before payment. Two requirements set it apart from retail payment acceptance. First, the debt must be identified unambiguously, because a taxpayer settles a specific liability, not a shopping cart. Second, the payment must produce legally binding proof, because paying a tax starts legal deadlines running and extinguishes a legal obligation. The countries that got public collection right standardized the payment notice first, and only then opened collection channels.
A national QR standard is a two-dimensional code specification published by a public authority and made mandatory for acceptance. In several countries it took hold by regulatory mandate rather than organic adoption. The central bank publishes the specification and requires acceptance, which ends the patchwork of proprietary codes: each merchant displays a single code. Indonesia is the most advanced example. QRIS, run since 2019 by Bank Indonesia with the Indonesian Payment System Association (ASPI), replaced the codes each wallet used to impose on its merchants. It had 50.50 million users and 32.71 million enrolled merchants (Bank Indonesia, 2024 data).
| Standard | Operator | Since | How it was imposed |
|---|---|---|---|
| QRIS | Bank Indonesia with ASPI | 2019 | Central bank mandate to end wallet fragmentation |
| Thai QR Payment | Bank of Thailand / National ITMX | 2018 | EMVCo standard on top of PromptPay; the basis of Thailand’s cross-border links |
| DuitNow QR | PayNet | 2019 | Banks and wallets required to accept the same code |
| QR Ph | Bangko Sentral ng Pilipinas with PPMI | 2019 | Mandatory replacement of proprietary QR codes; Paleng-QR Ph program in municipal markets and local transport |
| SGQR | MAS / IMDA through the Singapore Payments Council | 2018 | A single label covering domestic and international schemes; a display standard, not a clearing standard |
| TANQR | Bank of Tanzania | 2022 | Mandatory link to TIPS, which drove QR adoption |
The QR code handles how the debt is displayed at the moment of payment. Bill presentment networks handle how it is delivered to the payer in the first place. Italy carries its tax notices, vehicle taxes, and business invoices over CBILL, an interbank service run by CBI S.c.p.a. since 2014. Italian banks offer pagoPA payments through that channel, in their apps and at their ATMs. India built the equivalent with Bharat Connect, formerly the Bharat Bill Payment System, created in 2017 and run by NPCI Bharat BillPay Ltd. A biller connects to the network once. Its bills then become payable from any connected app. Value flowing through the network rose from ₹0.96 lakh crore in 2021 to ₹14.8 lakh crore in 2025.
Small economies sometimes build their payment infrastructure in a single step. Timor-Leste launched P24 in 2024. The country’s first electronic interbank system is supplied by SIBS and operated by the Banco Central de Timor-Leste. The interoperable ATM and POS switch is used in particular for paying taxes and customs duties. The country went straight from a fully fragmented landscape to a national rail, with no intermediate stage. Public collection was the first funded use case.
Mandatory e-invoicing brings the tax authority into the chain as soon as the invoice is issued, before payment. Kenya introduced it by public notice on June 8, 2026. Starting with the 2026 financial year, all reported income and expenses must be backed by an e-invoice issued and transmitted through eTIMS/TIMS. Malaysia has been phasing in MyInvois since August 1, 2024. Guidelines issued on December 7, 2025, raised the exemption threshold to RM 1 million. Singapore takes a different approach. InvoiceNow, a national network built on Peppol, transmits invoice data to IRAS, the Singapore tax authority.
Selling to government: e-invoicing, payment terms, and penalties
A public contract is an agreement under which a public body buys works, supplies, or services from a business. Payment follows a timetable set by law, not by the contract. That rule sets it apart from an ordinary commercial relationship. The public buyer pays, often late, on that statutory timetable. The supplier’s risk is therefore payment delay rather than default. Its working capital needs are calculated from the applicable law, not from its standard terms of sale. Two parameters drive them: the mandated invoicing channel and the statutory payment period.
The invoicing channel determines when the payment clock starts. Directive 2014/55/EU requires EU contracting authorities to receive and process e-invoices that comply with the European standard. That standard comes in two syntaxes, UBL and CII, while the network that carries the invoice is chosen separately from this semantic standard. Peppol, launched as an EU pilot project in 2008 and governed since 2012 by OpenPeppol AISBL, a Brussels-based association, runs on a four-corner model, like the card schemes. Italy is the exception: it mandates its own state channel, the Sistema di Interscambio. An invoice that fully complies with the semantic standard still has no effect if it travels over a channel the receiving government does not accept.
| Jurisdiction | Mandated payment term | Late-payment penalty | Reference |
|---|---|---|---|
| European Union | 30 calendar days; 60 days in exceptional cases | Interest at the reference rate plus at least 8 percentage points, plus a flat recovery fee of at least €40 | Directive 2011/7/EU, Articles 4 and 6 |
| US (federal agencies) | 30 days after receipt of a proper invoice or acceptance; 7 days for meat, 10 days for dairy and perishable products | Interest penalty paid automatically, with no claim from the supplier | FAR 52.232-25; 5 CFR Part 1315 |
| United Kingdom | No statutory cap on agreed terms: the law works through the cost of paying late | Interest at 8 percentage points above the Bank of England base rate; fixed compensation of £40, £70, or £100 depending on the amount | Late Payment of Commercial Debts (Interest) Act 1998 |
| India (micro and small enterprises) | 45 days maximum | Compound interest with monthly rests, at three times the policy rate notified by the RBI | MSMED Act 2006, Sections 15 and 16 |
The main difference between these regimes is who has to trigger the penalty. The US federal regime pays it automatically, with no action from the supplier, so an agency’s late payment hits its budget directly. The European regime creates a right that the creditor must exercise, and few suppliers do so against a public customer whose next contract they hope to win. The two texts say much the same thing. The US penalty is actually paid; the European one is rarely claimed.
- Get the buyer’s routing identifier before signing, not when the first invoice goes out: a public body often has several, one per department.
- Distinguish the semantic standard from the transport network: the first describes what the invoice contains, the second how it travels. The two are chosen separately.
- Model the actual payment delay, not the statutory one: the law sets a ceiling, but the buyer’s behavior sets the cash position. Measure the two separately.
- Check whether the penalty is automatic or must be claimed: it completely changes the economic value of the right.
- Never bridge a public buyer’s late payment with invoice discounting without checking that the contract allows it: most public procurement regimes strictly limit the assignment of receivables and factoring.
The state as rail builder: domestic schemes and subsidies
A domestic card scheme is a national card issuing and acceptance network. Its rules, its brand, and the processing of its transactions belong to an entity based in the country. Public flows serve as the seed for this kind of infrastructure. Launching one runs into a chicken-and-egg problem. Merchants do not equip their stores without cardholders, and cardholders do not use a card that is not accepted. Public procurement and social benefits break the deadlock by providing an issuing base that does not depend on any commercial decision. Several countries have followed this sequence, as the timeline below shows.
The Turkish case shows what this method delivers when it is sustained over time: the scheme, launched in 2016, stayed small until its 2024 turning point. TROY had 90 million cards at the end of 2025, up 80% year over year, with TRY 4.8 trillion in volume. Its market share by value reached 25.3%, up from 18.3% a year earlier (BKM, releases from January 2025 and January 23, 2026). About 25 lira of every 100 paid by card in Turkey now run over the domestic scheme. That growth comes from the mandate placed on the public sector, not from cardholder choice. It will last only as long as the mandate does.
Governments also intervene on the price of payment services, through price caps or direct subsidies. Three examples show the range. In Thailand, PromptPay is free below a threshold by government decision, a measure that wiped out most paid person-to-person transfers in the country. In Indonesia, regulation caps BI-FAST at Rp 2,500 per transaction to bring down interbank transfer pricing. In Pakistan, the state subsidizes acceptance directly, on the terms described below.
Pakistan’s acceptance subsidy runs on published terms. The state pays financial institutions 0.5% of the value of each Raast QR merchant payment, capped at PKR 100. The amount is split equally between the merchant’s bank and the customer’s bank. It covers transactions made between September 1, 2025, and June 30, 2026, from a budget of PKR 3.5 billion. Banks file claims quarterly with Raast Payments Pakistan, certified by internal audit. From January to March 2026, Raast processed 664 million person-to-person transactions, against 55.9 million merchant payments. The gap shows that the rail has been adopted for domestic transfers far more than for merchant payments, even though merchant payments are what the subsidy targets.
Programmable public money: promise and track record
Programmability is the ability to attach conditions of use to a unit of money, checked at the moment it is spent. The argument comes up in almost every central bank digital currency (CBDC) project: a programmable subsidy could be spent only on its intended purpose. A food voucher would become impossible to counterfeit, a fuel subsidy impossible to resell, and a farm payment traceable all the way to the input supplier. Operational results fall short of that promise. So far, programmable uses in production are limited to a handful of territories and a handful of aid programs.
India is an instructive comparison, because the country already has a free, universal retail payment rail. The Reserve Bank of India opened its e₹ (Digital Rupee) pilot in 2022 as a closed user group, with 13 banks and 26 cities. Outstanding value stood at ₹771.7 crore on March 31, 2026, down 24% year over year (RBI, 2025–26 annual report). The decline is the key fact. With UPI available, a retail CBDC has no economic case. The RBI is pursuing two tracks. The first is targeted programmability, used for subsidies in Gujarat, Puducherry, and Chandigarh. The second is wholesale cross-border use.
| Project | Operator | License type | Stated public sector rationale |
|---|---|---|---|
| e₹ (Digital Rupee) | Reserve Bank of India | Pilot since 2022 | Programmable subsidies; retail balances down 24% year over year |
| Digital tenge | National Bank of Kazakhstan | Live | Traceable public spending: procurement, the National Fund, tax administration, subsidies |
| eNaira | Central Bank of Nigeria | Live since 2021, little adoption | Announced shift toward government-to-person payments and cross-border use |
| JAM-DEX | Bank of Jamaica | Live since 2022 | First CBDC to be granted legal tender status |
| DCash | Eastern Caribbean Central Bank | Discontinued on January 12, 2024 | First retail CBDC in a monetary union; relaunch suspended in February 2026 in favor of a regional instant payment rail |
Kazakhstan aimed its project at public spending rather than retail payments. The digital tenge is designed to make public funds traceable. It covers public procurement, management of National Fund assets, tax administration, subsidies, and infrastructure financing. On July 18, 2026, it was recognized as a form of the national currency, with the National Bank given exclusive rights over its issuance and circulation. Among the major CBDC projects, it is the only one where the public sector is the primary rationale rather than one use case among others.
Nigeria took the opposite path. The eNaira, Africa’s first retail CBDC, launched in 2021, and adoption has remained very limited. Almost none of the wallets opened have ever been used, yet the service has never been shut down. The Nigeria Payments System Vision 2028, published by the Central Bank of Nigeria on June 5, 2026, acknowledges the slow uptake. It announces a shift toward government-to-person payments and cross-border use. Failed adoption is therefore not the same as a shutdown. The eNaira is still technically live.
Operating on public flows: what to check
For a payment service provider, operating on public flows means collecting payments on behalf of a government agency or making its disbursements. The work follows specifications published by the public authority, and contracts are awarded through a procedure in which three criteria carry the most weight. The first is compliance with the protocol the agency imposes. The second is the ability to absorb mass peaks on fixed dates. The third is an accurate reading of the legal text that governs the flow. The quality of the technical integration matters less to the outcome than these three.
- Identify the legal text, not the contact person. A mandatory public platform is set by decree, not in a tender. In Italy, the ban on providers dates from February 28, 2021, and applies even to firms operating under the EU freedom to provide services.
- Check whether the debt has a standardized identifier and build all reconciliation on it. Italy’s IUV is the archetype; the principle applies wherever the state has standardized its payment notice.
- Keep collection and certification apart. Collecting for the state and certifying the receipt fall under separate regimes, sometimes three stacked ones (accepting, certifying, collecting). A project that covers only two is not compliant.
- Size for the social payday peak, not the monthly average. A single payday concentrates most of the volume, and cash-out agent liquidity determines success as much as the rail does.
- Treat ID-to-account mapping as a source of disputes, not a technical detail. That is where payments to the wrong account originate, and it is what regulators have started to address.
- Get the invoice routing identifier before the first delivery on a public contract, and confirm the expected syntax. An invalid invoice starts no clock.
- Model the end of administered prices and subsidies. Mandatory free service, per-transaction regulatory caps, acceptance subsidies: all of them can be revised by circular, often at short notice.
- Do not confuse failed adoption with a shutdown for public digital currencies. Several projects written off as dead are still live, and at least one is explicitly repositioning toward social payments.
One thread runs through every system in this guide. The public sector is the only player that can impose an acceptance standard on an entire country with a single decision. At the same time, it supplies the volume that makes the standard viable. QRIS in Indonesia, PromptPay in Thailand, pagoPA in Italy, Meeza in Egypt, and TROY in Turkey all spread by regulatory mandate, not by gradual commercial adoption. In these markets, regulatory mandate is the normal way payment infrastructure spreads, not a transitional phase before competition takes over.