Five markets, five central banks, no regional license
East Africa is the region where retail payments grew up furthest from the banks. The everyday instrument is an e-money wallet tied to a mobile phone line, not a bank account or a card. Customers load and cash out that wallet through a network of neighborhood agents who swap banknotes for e-money. The wallet handles person-to-person transfers, merchant payments, and bill payments. Bank rails are the exception rather than the rule, and they were built after the wallet, often to compete with it. The region's technical and regulatory stack, banking side included, is best read starting from the wallet.
Licensing is national in every market in the region, with no mutual recognition among them. There is no regional license of the passport kind, even though five of these markets belong to the East African Community. An e-money issuer licensed by the Central Bank of Kenya therefore has no right at all to operate in Tanzania or Uganda. Operating there takes a separate application, filed with a separate authority, under a separate law. Regional pooling stops at two cross-border large-value rails, EAPS and REPSS.
| Market | Authority | RTGS | Retail clearing | Retail instant rail | Dominant rail in practice |
|---|---|---|---|---|---|
| Kenya | Central Bank of Kenya (CBK) | KEPSS (since July 29, 2005) | NACH, Nairobi Automated Clearing House (checks, EFT) | PesaLink (IPSL, 2017); a public Kenya Fast Payment System is still only announced | M-Pesa (Safaricom), outside the bank rails |
| Tanzania | Bank of Tanzania (BoT) | TISS | TACH, Tanzania Automated Clearing House | TIPS (BoT, 2022), mandatory membership | TIPS, which links wallets and banks |
| Uganda | Bank of Uganda (BoU) | UNISS (2005) | ACH in five currencies (UGX, USD, EUR, GBP, KES) | No unified public switch documented here | MTN MoMo and Airtel Money (duopoly) |
| Rwanda | National Bank of Rwanda (BNR) | RIPPS (2011): RTGS + ACH + central securities depository | RIPPS (same platform) | eKash / RNDPS 2.0 (RSwitch, on Mojaloop, 2025) | eKash, at near-zero public pricing |
| Ethiopia | National Bank of Ethiopia (NBE) | EATS (2011), 10 hours a day, 6 days a week | EthSwitch for ATM and card interoperability | EthioPay (EthSwitch, December 2025) | telebirr (Ethio Telecom) and EthSwitch |
| Somalia | Central Bank of Somalia | National Payment System (launched July 2021) | – | SIPS + Somali Payment Switch (January 2025) | Mobile wallets, which predate the banking infrastructure |
Kenya: M-Pesa, national infrastructure owned by a telecom operator
M-Pesa is an e-money service launched in 2007 by Safaricom plc, the Kenyan telecom operator in the Vodacom/Vodafone group. The platform now sits within the M-Pesa Africa joint venture, and M-Pesa remains Africa's largest mobile money service. No other payment system in the world has been studied as closely. Its interface is a USSD menu that works on a basic handset with no internet connection. What sets it apart is institutional rather than technical. In Kenya, the dominant retail payment rail is privately owned by a listed company, not a piece of public infrastructure.
M-Pesa changed its pricing logic in fiscal 2026, and volume has largely decoupled from revenue. 17.1 billion “Kadogo” micro-transactions are free, and they account for 58% of activity (Safaricom FY26). Most of the volume therefore no longer earns any per-transaction revenue. Revenue is shifting to services built on the wallet, chiefly credit, merchant services, and co-badged cards. The gap is widening between consumer pricing, often zero, and merchant pricing, which now carries the network's economics. Any negotiation over merchant pricing has to take this shift into account. M-Pesa brings in KES 182.7 billion in revenue, or 45.6% of Safaricom group revenue. Payments are no longer a sideline for the telecom operator: they generate nearly half its revenue.
A credit ecosystem has grown up around the wallet, and it explains how deeply Kenyans use it. Two products drive it. M-Shwari is a savings and microloan account linked to the wallet, run with NCBA. Fuliza is an instant overdraft that covers a payment when the wallet balance falls short. Both products directly lift the real payment success rate, since a customer with an approved overdraft completes the payment where a customer with a zero balance fails. The conversion gap is measurable on an average-sized cart.
PesaLink, Kenswitch, and KEPSS in Kenya: the banks fight back, and the choice is still open
Kenya's legal framework rests on the National Payment System Act of 2011, implemented through regulations issued in 2014. The regulations govern the authorization of payment service providers and the oversight of payment systems. The CBK's National Payments Strategy 2022–2025 sets the policy direction (Central Bank of Kenya, National Payments System page, accessed in 2026). This framework came after mobile money: the CBK let M-Pesa launch in 2007 under a letter of no objection, subject to its oversight. The Kenyan wallet has therefore never operated outside the central bank's view, but its formal authorization regime dates only from 2011–2014.
| Rail | Operator | Since | What to know |
|---|---|---|---|
| KEPSS (Kenya Electronic Payment and Settlement System) | Central Bank of Kenya | July 29, 2005 | The RTGS: continuous, final gross settlement in central bank money. A systemically important system. It is also Kenya's access point to the regional EAPS and REPSS rails, which are hosted on it |
| NACH (Nairobi Automated Clearing House) | Central Bank of Kenya, together with the Kenya Bankers Association | – | Retail clearing with net settlement: checks and electronic transfers. Limit of KES 1 million per check since October 2009; cycle cut from T+3 to T+1 in 2013. ⚠️ Same name as India's NACH, run by NPCI, but entirely unrelated |
| PesaLink | Integrated Payment Services Limited (IPSL), a subsidiary of the Kenya Bankers Association | 2017 | Instant interbank credit transfers, more than 80 institutions connected (IPSL/KBA, 2025–2026). Volumes up 78% in 2023 over 2022 (KBA 2023 annual report); no cross-checked 2025 figure. Linked to PAPSS since February 2026 |
| Kenswitch | Kenswitch Limited (bank consortium, backed by the CBK) | 2002 | The domestic card and ATM switch: 26 member banks, more than 2,500 ATMs, about 40,000 POS terminals, and 50,000 connected agents (Kenswitch, 2026). It supports the bank agent sharing authorized by the CBK |
| Kenya Fast Payment System (FPS) | Central Bank of Kenya | announced | National instant payment rail announced under the National Payments Strategy. Status: announced, not live. The choice is still open (see below) |
Kenyan wallets have been interoperable since April 10, 2018, when Safaricom and Airtel opened direct wallet-to-wallet transfers. T-Kash (Telkom Kenya) joined in October 2018. Technical connectivity alone did not generate off-network transfers. The CBK then had to *require alignment of on-net and off-net pricing*. As long as sending off-network costs more than sending on-network, the largest player keeps its network effect, however well the systems are connected. This pricing mechanism governs the whole region, including the fee schedules of the providers that plug into it.
Two kinds of players handle acceptance in Kenya. Large banks are active, notably Equity Bank, KCB, Co-operative Bank, and NCBA, alongside regional acceptance providers such as Pesapal, Cellulant, and DPO Pay. DPO Pay, formerly DPO Group, now operates under the brand of Network International, whose Africa headquarters is in Nairobi. Network International's ownership of DPO is established; the more remote ownership chains reported in the press are not. Due diligence should therefore focus on the contracting entity named in the contract, not the brand shown on the website.
Tanzania: the central bank runs the rail, and membership is mandatory
Tanzania built retail interoperability in two stages: first through bilateral agreements, then through a central bank switch. The first stage was market-driven. With the backing of the Bank of Tanzania, Airtel and Tigo signed an interoperability agreement in September 2014 and launched commercially in February 2015, followed by Tigo and Zantel in December. Common business rules were adopted in October 2014, covering participation criteria, clearing and settlement principles, and dispute resolution. The arrangement was the *world's first market-led wallet-to-wallet interoperability*. In 2022, the central bank replaced this bilateral mesh with a switch it runs itself.
| System | Operator | Since | Role and published figures |
|---|---|---|---|
| TISS (Tanzania Interbank Settlement System) | Bank of Tanzania | – | RTGS: urgent interbank transfers and settlement of net obligations from check/EFT clearing, card switches, and the money market. 33 participating banks and institutions; the beneficiary's bank must credit the account within two hours of the instruction (Bank of Tanzania). Still runs on SWIFT MT messages |
| TACH (Tanzania Automated Clearing House) | Bank of Tanzania | – | Low-value retail clearing; its net balances settle in TISS (Bank of Tanzania, Payment System page) |
| TIPS (Tanzania Instant Payment System) | Bank of Tanzania, direct operator | 2022 | 651 million transactions worth TZS 54,950 billion in 2025, vs. 453 million and TZS 29,820 billion in 2024, up 84% in value year over year (BoT, National Payment Systems Annual Report 2025). 45 providers connected at end-2024; about 1.5 million transactions a day, with capacity sized at three times that |
| TANQR | Bank of Tanzania | 2022 | National merchant QR standard built on TIPS: one QR code accepted by all. Airtel Money, AzamPesa, Mixx by Yas, TTCL Pesa, HaloPesa (Viettel), and M-Pesa are the six integrated e-money issuers (BoT, 2025 report) |
| UmojaSwitch | UmojaSwitch Co. Ltd (cooperative of Tanzanian banks) | 2006 | Cooperative switch pooling ATMs, mobile banking, and agency banking for banks too small to build their own: more than 260 ATMs in Tanzania and Zanzibar, Visa and UnionPay acceptance (UmojaSwitch, 2026). Does not replace the central bank rails |
| SADC-RTGS | South African Reserve Bank, mandated by the SADC Committee of Central Bank Governors | 2013 | Tanzania is one of the 16 participating states. Settlement in rand only for now, a key constraint for any regional flow through southern Africa (SARB, 2026) |
Uganda: a telecom duopoly on a minimal central bank foundation
Uganda's retail market is overwhelmingly dominated by two telecom wallets, MTN MoMo and Airtel Money, alongside deliberately lean public infrastructure. The Bank of Uganda runs two systems designated as systemically important, and only two: UNISS, the national RTGS, and the ACH. Uganda has no equivalent of the public retail switch that Tanzania mandated. Interoperability therefore still depends largely on agreements between private players and on aggregators.
- UNISS (Uganda National Interbank Settlement System), Bank of Uganda, since 2005: 27 registered participants (Bank of Uganda, national payment systems oversight framework, 2025). Participation is open to non-banks, which matters a great deal in a market where mobile money carries retail volume. Migration to CMA Small Systems' RTS/X platform is under way.
- ACH (Bank of Uganda): multilateral net clearing of checks and electronic transfers in five currencies (Ugandan shilling, US dollar, euro, pound sterling, and Kenyan shilling). Domestic multicurrency clearing is the key point: routine regional flows, especially to Kenya, avoid going through a correspondent bank.
- Cross-border rails: Uganda participates in EAPS (East African Community, local currencies) and is one of the eight countries actually connected to REPSS (COMESA), along with Mauritius, the DR Congo, Malawi, Eswatini, Zambia, Rwanda, and Kenya (Central Bank of Kenya, 2026).
- Non-telecom competition: Wave Mobile Money is licensed as an e-money issuer in Uganda, among other markets (Wave, EME/EMI licenses). It is the main price disruptor in African mobile money, and worth watching for the pressure it puts on fee schedules.
- Online acceptance: DPO Pay (Network International) covers Uganda as well as Kenya, Tanzania, and Rwanda; Chipper Cash (Critical Ideas, Inc.) also operates there, with a cross-border person-to-person transfer model.
MoMo's revenue mix is shifting away from person-to-person transfers. At MTN, “advanced services” (credit, insurance, investment, and merchant payments) made up 34.1% of MoMo revenue in 2025, up from 29.9% in 2024. That year-over-year increase measures the shift: growth now comes from these services, not from P2P transfers. The operator earns direct revenue from merchant acceptance and now has its own reason to grow it, which strengthens the merchant's hand in negotiations. Five years earlier, the reverse was true.
Ethiopia: 120 million people, a long-closed market, and a stack playing catch-up
Ethiopia has some 120 million people and a recent payments stack that market research still covers only thinly. Its development was held back by a foreign exchange and market access regime historically closed to foreign players. In relative terms, its growth now ranks among the fastest on the continent. Three building blocks frame the system: telebirr for wallets, EthSwitch for interoperability, and EATS for settlement.
Agents, float, and cash-in/cash-out: the rail's physical infrastructure
The agent network is the set of physical outlets where e-money is exchanged for banknotes. Shopkeepers, kiosks, and gas stations handle the conversion both ways, acting on an issuer's behalf. None of the systems described above would work without this network. It is the region's most expensive and least visible infrastructure. Where payments happen, and where they fail, shows up on the agent map more than in any API specification.
Two published figures give a sense of the network's scale. On the Kenyan banking side, Kenswitch connects more than 2,500 ATMs, about 40,000 POS terminals, and 50,000 agents for 26 member banks (Kenswitch, 2026). The CBK has allowed banks to share agents, which extended coverage without multiplying each bank's costs. On the telecom side, MTN Group reports 1.4 million active agents and 2.1 million active merchants across its footprint. Those figures come from MTN Group's 2025 results, published March 16, 2026. The gap in scale between the banking and telecom networks shapes the entire market in the region.
Two signals help gauge how mature a market in the region is. The first is a falling share of cash-out, which shows that e-money is circulating within the ecosystem instead of leaving it after every payment. Merchant acceptance becomes more valuable as that share shrinks. The second is free micro-transactions. The 17.1 billion free “Kadogo” micropayments account for 58% of M-Pesa activity (Safaricom FY26). That pricing choice reflects a decision to stop charging a toll on small amounts in order to capture usage. Once a market reaches this stage, merchant acceptance becomes profitable.
Cross-border: EAPS, REPSS, PAPSS, and the limits of each
Four rails carry the cross-border payments reachable from East Africa. They differ in the amounts they handle, their settlement currencies, and the customers they serve. Mixing them up is the costliest mistake in a regional architecture. It triggers no technical error message: the cost shows up as longer settlement times and cash tied up.
| Rail | Scope | Settlement currency | What it does, and what it doesn't |
|---|---|---|---|
| EAPS (East African Payment System) | Four countries: Kenya, Rwanda, Tanzania, Uganda (CBK, 2026) | Local currencies of the East African Community | Large-value, by linking national RTGS systems. Open 08:30–16:00 EAT, Monday to Friday. ⚠️ Burundi is not connected, despite presentations that line up all five EAC central banks. Complements the retail rails, never competes with them |
| REPSS (Regional Payment and Settlement System) | Eight countries connected: Mauritius, DR Congo, Malawi, Eswatini, Uganda, Zambia, Rwanda, Kenya (CBK, 2026) | US dollar and euro | COMESA rail, live since October 2012, accessed through each country's own RTGS (KEPSS in Kenya). Only eight of the 21 COMESA member states are connected, which is why it should not be overrated |
| PAPSS (Pan-African Payment and Settlement System) | 28 African states, more than 190 banks and fintechs, 16 switches since the BEAC (Bank of Central African States) joined on July 9, 2026 | Local African currencies, daily netting, with net balances settled in hard currency through Afreximbank | The only pan-African rail that bypasses correspondent banks outside Africa: in effect, the infrastructure of the AfCFTA (African Continental Free Trade Area). Link to PesaLink announced in February 2026. ⚠️ Values processed are not published: coverage is growing fast, but actual usage cannot be verified |
| SADC-RTGS (formerly SIRESS) | 16 SADC states, including Tanzania | Rand only for now | The backbone of large-value payments in southern Africa. Relevant only to Tanzania within East Africa. A Dar es Salaam ↔ Johannesburg flow does not travel the same pipes as a Dar es Salaam ↔ Nairobi flow |
| Private hubs (Onafriq, M-Pesa Global) | Onafriq: 43 countries, 2,000 corridors claimed. M-Pesa Global: Safaricom in Kenya, Vodacom in Tanzania, DRC, Mozambique, Lesotho | Depends on the corridor | Wallet-to-wallet retail, where the public rails don't reach. A single connection to Onafriq replaces dozens of bilateral agreements. ⚠️ Reach figures are self-reported and unaudited; Onafriq's stated headquarters is in London, not Mauritius or South Africa |
Somalia followed the opposite path from the other markets in the region. The Central Bank of Somalia launched the Somali Instant Payment System (SIPS) and the Somali Payment Switch in January 2025. They build on the National Payment System, launched in July 2021, and carry 24/7 transfers between banks, mobile wallets, and terminals for P2P, P2M, and G2P payments. The rollout took place in a post-conflict setting, in a country where mobile money long predated the banking infrastructure. The central bank estimates that more than two-thirds of payments already go through digital channels.
Accepting payments in the region: a launch checklist that holds up
The points below cover the decisions to make before technical integration, in the order they come up. Here, the country is the unit of decision, not the region. Start by qualifying the country's dominant rail and who owns and operates it; the choice of a payment service provider (PSP) comes only after that.
- Qualify the country's model. Rail owned by a telecom operator (Kenya, Uganda): the cost and availability of collection are negotiated commercially, and the operator is also a potential competitor for the service. Rail run by the central bank (Tanzania, Rwanda): access depends on a license, pricing is administered, and counterparties are members by default.
- Check local establishment requirements. No East African license works as a regional passport. An e-money issuer must be licensed country by country (CBK, BoT, BoU, BNR, NBE), and Ethiopia's regime has historically been closed to foreign players, opening gradually since M-Pesa Ethiopia arrived in 2023.
- Choose between a direct connection and an aggregator. Direct gives the best unit cost and the slowest time to go live; an aggregator gives the reverse, whether it is DPO Pay (Network International), Cellulant, Pesapal, Flutterwave, or Onafriq for wallet interconnection. In a first market, an aggregator is almost always the right choice, as long as the exit has been planned.
- Design reconciliation before integration. The callback is not proof of payment. Reconciliation that holds up combines four elements: status queries to the operator, duplicate detection on the merchant side, fuzzy matching for hand-typed references, and an exception queue worked by a human operator. An integration that skips this passes user acceptance testing, where the data entry is clean, and fails in production, where it is not.
- Wire up refunds on day one. The outbound flow has its own permissions, limits, and timing, and it does not mirror collection. A platform that collects but cannot refund builds up operational debt that customer service ends up paying off.
- Plan for repatriation and FX. A merchant account balance is held in local currency. Converting it and moving it out of the country fall under the national exchange control regime, not the payment provider, and that is where the delays nobody budgeted for tend to hide.
- Instrument non-technical failure reasons. Agent out of float, insufficient balance, PIN entered too late, wrong reference: product changes and the acceptance network fix these, not a retry. Without that breakdown, the conversion rate stays unexplained.
- Monitor tax and regulatory risk. Transaction taxes are a recurring political risk for the sector across the continent, and Kenya's open choice between PesaLink and a public Fast Payment System shows that the dominant rail itself can change. Avoid long exclusivity on any single domestic rail.