Measuring cash share against the right denominator
Cash share is the weight of cash in a market's payments, measured against a reference total. Central banks, statistics agencies, and payment providers all publish figures under that name, but they do not use the same denominator. The first difference is the unit counted. Transaction counts and transaction values never give the same result, because cash is used mostly for small amounts. The second difference is the channel: a measure may cover in-person points of sale, remote commerce, or bill payments. A comparison between two markets is valid only if both use the same unit and the same channel.
| Market | Published measure | Value | Source and year |
|---|---|---|---|
| Euro area | Cash share by number of point-of-sale payments | 52% in 2024, down from 59% in 2022 and 72% in 2019 | ECB, SPACE 2024 |
| Euro area | Cash share by value at the point of sale | 39% in 2024, down from 42% in 2022 | ECB, SPACE 2024 |
| Malta / Slovenia | Cash share by number, point of sale | 67% and 64% | ECB, SPACE 2024 |
| Netherlands / Finland | Cash share by number, point of sale | 22% and 27% | ECB, SPACE 2024 |
| United States | Cash share by number of consumer payments | 14%, behind credit (35%) and debit (30%) | Federal Reserve, Diary 2025 (2024 data) |
| Japan | Cashless payment ratio (without cash), domestic indicator | 58.0%, or 46.3% on the international comparison indicator | METI, March 31, 2026 |
| Brazil | Boleto share by number of retail payment transactions | 7.6% in the second half of 2025 | Banco Central do Brasil |
| Global | Cash share of in-store payment value | 14% in 2025 | Worldpay, Global Payments Report 2026 |
Cash share of payments and the value of banknotes in circulation are two different quantities, and they can move in opposite directions. India's currency in circulation reached ₹41.68 trillion at the end of fiscal 2025–2026, up 11.9% year over year (Reserve Bank of India). That was the largest increase in absolute terms since the year after demonetization. Over the same period, UPI kept breaking volume records, so both series grew together.
Brazil is moving the other way. Cash withdrawals fell to 1.1 billion transactions in the second half of 2025, down 13.8% year over year (Banco Central do Brasil). Pix captured 54.7% of retail payment transactions over the same period. The drop in withdrawals has tracked the rollout of a free public payment rail to the entire adult population, something no other market has replicated so far.
OXXO, Rapipago, Abitab: Latin America's payment counters
A cash collection network is a set of physical retail locations that collect payment, on a merchant's behalf, for an order placed elsewhere. The model is used from Mexico to Uruguay under several names: cash voucher, pago en efectivo, or CIP code. The customer orders online, receives a numeric reference, then goes to a counter to pay in cash. The collection network takes the funds on the merchant's behalf and passes them on at its settlement cycle. Payment therefore becomes asynchronous, with hours or days between the order and the confirmation.
| Network | Operator | Country | Since | What to know |
|---|---|---|---|---|
| OXXO Pay and Paynet | FEMSA (OXXO Pay); Paynet (multi-retailer network) | Mexico | 2015 | The must-have rail for Mexican e-commerce. Selling in Mexico without it cuts off a significant share of the market. |
| Rapipago and Pago Fácil | Grupo Gire (Rapipago); Prisma / Multipago (Pago Fácil) | Argentina | 1990 | Bill payment and cash voucher outlets, still essential for the unbanked population. |
| Abitab | Abitab S.A. | Uruguay | 1993 | About 240 locations in Montevideo and 240 more in over 100 towns (Abitab, 2026). Grew out of a network of lottery agents. |
| Redpagos | Redpagos | Uruguay | 2001 | The second national network, formed from a merger of currency exchange houses. The Abitab / Redpagos duopoly shapes all cash acceptance in the country. |
| PagoEfectivo | Orbis Ventures S.A.C. (Paysafe group) | Peru | – | Generates a CIP code payable at a bank branch or a bodega. Also available in Argentina and Ecuador. |
| PayNearMe | PayNearMe, Inc. | United States | 2009 | The North American equivalent: paying an online bill in cash at a partner retailer. |
Pricing on these networks combines a percentage and a fixed fee per reference, with a minimum fee and amount limits. Conekta publishes an e-commerce rate for cash payments of 2.6% + MXN 3.00 before tax, with a minimum fee of MXN 5.40. Financial services and insurance pay a flat MXN 16 before tax. The maximum per reference is MXN 35,000 and the minimum MXN 10, across a network of more than 19,000 payment locations (Conekta, published pricing, accessed in 2026). Customers sometimes pay a fee of their own, from MXN 10 to MXN 13 depending on the location.
These networks also act as bank tellers. Spin by OXXO, launched in 2021, links a digital wallet to the convenience store network, turning each store into a place to deposit and withdraw cash. The offering puts the retailer in direct competition with banks for unbanked customers. The e-money is issued by an electronic payment funds institution separate from FEMSA and OXXO: Compropago S.A. de C.V. The commercial contract and compliance due diligence apply to that entity.
Brazil's boleto: a negotiable instrument, not a transfer
The Boleto Bancário is a Brazilian barcoded payment slip that a creditor issues and a debtor pays at a bank counter, at an ATM, or by debit. Because it is a negotiable instrument, it supports three things the Pix instant payment system does not: protest, enforcement, and assignment. That legal difference explains why the boleto has held its ground against Pix in B2B and collections, even though consumers have largely abandoned it for purchases. The rail is operated by Nuclea, formerly CIP, under the supervision of the Banco Central do Brasil. Febraban, the Brazilian banking federation, sets the format but does not operate it.
Mandatory registration changed how collection works. A registered boleto can be paid at any bank, even after the due date and even through a banking correspondent. Before 2018, a slip issued by one bank could be refused at another. The central registry also tells the issuer what was paid, where, and when. Reconciliation used to rely on bank files in inconsistent formats.
Pix Saque and Pix Troco let customers withdraw cash or get cashback at participating retailers in exchange for a Pix payment. The Banco Central has thus turned every participating store into a cash distribution point, with no ATM required. The merchant hands its surplus cash to the customer and receives the equivalent in its account. Volume remains modest: 8.5 million transactions in the second half of 2025, up 20.9%. Brazil treats access to cash as a service delivered by the merchant network.
Japan's konbini: 56,000 counters that never close
A konbini is a Japanese convenience store that is always open and serves as a payment counter in addition to its retail business. Customers pay electricity bills, local taxes, online orders, insurance premiums, or installments on deferred purchases there. The service is called 収納代行 (shūnō daikō), meaning collection on behalf of third parties. No Western network comes close to this density of payment locations, so cash acceptance in Japan is built around convenience store chains.
| Issue | What applies | Impact on the merchant |
|---|---|---|
| Limit per slip | ¥300,000 including tax, a limit set by the collection agents | A larger amount cannot be collected on a single barcode; it has to be split |
| Who pays the fee | The biller, as a flat fee per payment collected | The unit cost is fixed: heavy on small amounts, light on large ones |
| Cost trend | Across-the-board increase of several tens of yen from 2022, after the association asked for a fee review in October 2020 | A contract signed before 2020 is no longer at market rates; the review clause is up for negotiation |
| Confirmation | Delayed: the biller learns of the payment from the collection agent's file | Any logic that releases orders immediately must be triggered by the notification, not the due date |
| Payment method at the counter | Overwhelmingly cash | A customer with no cash has to withdraw it before paying; the in-store ATM is part of the flow |
Konbini also serve as payment points for Japanese consumer credit. NP後払い (Net Protections, 2002) introduced pay-by-invoice for e-commerce 15 years before Klarna: the seller is guaranteed payment, and the buyer pays on receipt, at a konbini or by bank transfer. Paidy (2014, a PayPal subsidiary since 2021) rolls a month's purchases into a single bill, which can also be paid at the counter. Japanese buy now, pay later therefore relies on the convenience store network for collection, not on a card-linked credit line.
Mobile money agent networks
A mobile money agent is an independent merchant appointed by an operator to convert cash into e-money and back. The agent network is the service's only physical point of contact, and wherever there is no agent, there is no access to cash. The business depends on holding two inventories at once: banknotes in the till and an e-money balance in the account. Every transaction draws down one and adds to the other, so agents rebalance both several times a day.
As the network gets denser, each agent earns less, and the figures published in Kenya document the trend. Safaricom had 333,011 M-PESA agents at the end of March 2026, up from 298,890 a year earlier and about 173,000 in 2020. Commissions paid came to KES 37.38 billion in fiscal 2026, compared with KES 37.82 billion two years earlier. A flat commission pool spread across more and more agents drives average income down to KES 112,244 a year, from 144,355 in 2024 (Business Daily Africa, based on Safaricom data).
| System | Operator | Footprint | Published figures |
|---|---|---|---|
| M-PESA | Safaricom plc / M-Pesa Africa | Kenya, plus Tanzania, DRC, Mozambique, and Lesotho through Vodacom | KES 41,680 billion and 46.41 billion transactions in the fiscal year ended March 31, 2026; 40 million monthly active customers; 3.1 million merchants (Safaricom, 2026) |
| MTN MoMo | MTN Group Fintech | 13+ markets | 69.5 million active users, 23.3 billion transactions, $500.3 billion in value, 1.4 million active agents (MTN Group, 2025 results published March 16, 2026) |
| Orange Money | Orange Middle East and Africa | West Africa, Central Africa, the Maghreb | Run country by country through e-money subsidiaries licensed by central banks, separate from the telecom subsidiaries |
| Wave Mobile Money | Wave Mobile Money Inc. | Senegal, Côte d'Ivoire, Mali, Burkina Faso, Uganda | More than 20 million monthly active users and 150,000 agents in mid-2025; 1% fee on transfers, free deposits and withdrawals (company figures, unaudited) |
| MTN Mobile Money Ghana | Mobile Money Limited, licensed as an EMI by the Bank of Ghana | Ghana | GHS 4,100 billion in transactions and a float of GHS 38.4 billion in 2025 (MTN Ghana, March 2026) |
| telebirr | Ethio Telecom | Ethiopia | More than 54.8 million registered users in July 2025 and ETB 2,380 billion over the fiscal year (Ethio Telecom, 2025) |
| bKash / Nagad | bKash Limited (BRAC Bank group); Nagad Limited | Bangladesh | More than 82 million verified users at bKash (2025); Nagad operates under postal law, not a mobile financial services license |
| GCash | G-Xchange, Inc. (Mynt) | Philippines | 81 million active users and 2.5 million merchants in January 2025; an e-money issuer without a banking license |
| Easypaisa / JazzCash | Easypaisa Digital Bank; Mobilink Microfinance Bank | Pakistan | About 18 million and 21 million monthly active users (company statements, 2025); both operate under a microfinance bank license |
Mobile money's success depends on conditions in the host market. Vodacom shut down M-Pesa in South Africa in June 2016, with about 76,000 active users against an initial target of 10 million. South Africans were already banked and carried payment cards. The service offered them account access that other instruments already provided, unlike the markets where it took off.
Cash on delivery
Cash on delivery (cash on delivery, or COD; contre-remboursement in France, ramburs in Romania) means paying for an order at the moment it is handed over, directly to the carrier. The merchant ships the goods before collecting anything. The courier collects the cash and turns it over to its employer, which passes it on to the merchant on a contractual schedule. The transaction therefore involves three successive transfers of risk: at shipment, at collection, and at payout. This payment method dominates e-commerce across much of North Africa, the Middle East, and South Asia.
The largest cost in cash on delivery is return to origin, ahead of the carrier's fee. Indian industry estimates put cash on delivery at 60% to 65% of e-commerce orders (ET Prime Research, 2024). The return-to-origin rate reportedly reaches 25% to 30%, compared with 2% to 3% for prepaid orders. Each return means two trips, warehouse handling, and restocking, all costs incurred with no sale booked.
| Lever | How it works | Key limitation |
|---|---|---|
| Explicit cash-on-delivery fee | Charge for the payment method, not the product | Seen as a penalty; set it based on the cost of returns, not the carrier's fee |
| Prepayment discount | Shift the incentive toward cards, bank transfers, or wallets | An immediate, certain cost against a delayed, probable gain |
| Phone number verification at checkout | Weed out bogus orders before shipping | Adds friction to every order, including legitimate ones |
| Card or QR payment on delivery | The courier carries a terminal or a payment code | The customer stays in control until the doorstep; no need to make change, and returns remain possible |
| Pickup point collection with payment on site | A customer who makes the trip is committed | Sharply reduces returns to origin, but shrinks the delivery area |
| Risk scoring by address and order history | Offer cash on delivery only to trusted customers | Needs enough volume to be statistically sound |
The real cost of cash for merchants
The cost of accepting cash covers everything a merchant spends to take in, store, and deposit notes and coins. Unlike cards, none of these costs comes as a per-transaction fee billed by a provider, which is why cash is often seen as free. The cost items are checkout time, end-of-day cash-up, cash-in-transit, coin supply, and cash discrepancies. The Deutsche Bundesbank and the EHI Retail Institute put a figure on the total for German retail, timing 3,125 real transactions with a stopwatch.
| Cost item | Annual total | Per transaction | As % of cash sales |
|---|---|---|---|
| Checkout time for cash payments | €1,881.9M | 0,121 € | 0,896 % |
| Back office: counting, cash-up, equipment, insurance | €1,314.8M | 0,084 € | 0,626 % |
| Cash pickup and change supply | €578.0M | 0,037 € | 0,275 % |
| Total, cash | €3,774.7M | 0,242 € | 1,797 % |
| For comparison: domestic debit card with PIN | €675.3M | 0,33 € | 0,67 % |
| For comparison: credit card with signature | €319.2M | 1,036 € | 1,817 % |
Several costs are left out of that figure, and they weigh heavily in some store formats. Employee theft and robbery hit isolated stores with large cash holdings and long opening hours hardest. Cash discrepancies eat up management time without ever showing up as a payment cost. Cash-in-transit contracts are often billed per pickup, which penalizes frequent, small collections. Idle cash is one last cost: notes sitting in the drawer and the safe earn no interest and cannot fund inventory.
- Measure before deciding: timing a sample of cash transactions and a register close-out gives a more reliable internal cost than any industry average.
- Cut handling, not cash: cash recyclers at the register, smart safes with provisional credit, automated counting. The savings come in the back office, the largest cost after checkout time.
- Renegotiate cash-in-transit at the right frequency: adding safe capacity and spacing out pickups costs less than scheduling more of them.
- Put deposit fees where they belong: negotiate the cash deposit fee as you would a card fee, and compare it with the cost of accepting other payment methods.
- Track the average ticket by payment method: it is the variable that determines the cost ranking, and the one an aggregate ratio always hides.
Accepting cash: legal obligations and physical access
Legal tender is the status the law gives to notes and coins, which makes them valid for discharging a debt. Whether a merchant must accept them is a separate question, and countries answer it differently. In the US, no federal rule requires a private business to take cash; the obligation comes from states and cities. In the euro area, the Court of Justice of the European Union takes the opposite view: legal tender in principle carries an obligation to accept notes and coins at face value. A company rolling out a cashless concept across several countries therefore has to check the law market by market.
| Market | Rule | Since | Reach |
|---|---|---|---|
| Norway | Businesses must offer cash payment up to NOK 20,000 on premises that sell to consumers, under the Financial Contracts Act | October 1, 2024; fines possible since May 1, 2025 | Physical stores open to the public |
| Denmark | Ban on accepting DKK 20,000 or more in cash in a single transaction, on anti-money laundering grounds | In force | All businesses |
| Euro area | Council position of December 19, 2025, on the regulation on the legal tender of euro banknotes and coins, codifying the CJEU ruling (joined cases C-422/19 and C-423/19) | Position adopted, text not final | Mandatory acceptance at face value, with power to discharge debts; member states must guarantee access to cash |
| US (federal) | No obligation for private businesses to accept cash | – | Yields to any state or municipal rule |
| US (states) | Cash acceptance required: Massachusetts, New Jersey, Rhode Island, Colorado, Connecticut, Delaware, Oregon, Tennessee, Montana, New York | Varies; New York State's Protection of Cash Payments law took effect March 20, 2026 | In-person sales; online, mail, and phone orders are generally exempt |
| US (cities) | Philadelphia, San Francisco, Detroit, Washington | In force in Washington since January 1, 2025 | Retail and restaurants |
An obligation to accept cash assumes consumers can still get hold of it, which makes access to withdrawals the other half of the question. The UK monitors that access closely. The LINK network counted 42,403 ATMs, 33,710 of them free to use, and 1,272 million withdrawals in 2025. As of May 2026, 3,756 machines were designated “protected” and 282 banking hubs had been recommended. In the euro area, 57% of consumers say they can withdraw cash at a store checkout, and 87% say it is easy to get to an ATM or a bank, down from 89% in 2022 (ECB, SPACE 2024).
Cash is still the only payment method that works without a telecommunications network. That property is cited for business continuity, and some markets have built it into their payment architecture. Norway's BankAxept scheme works offline and serves as a designated cash distribution channel at the grocery retailer NorgesGruppen. A continuity plan that relies only on card fallback mode therefore overlooks the one rail that stays available during a network outage.
Accepting cash: what to check before you sell
The cash collection rails covered in this guide fall into four families, each with its own mechanism and its own constraints for merchants. None substitutes for another; each one carries local collection in the markets where it dominates. Together, these four families cover most situations merchants face outside high card-penetration markets.
- Find the dominant rail, not just an available one. A gateway that offers cards in Mexico without OXXO Pay covers only part of the market, and the gap shows up in the conversion rate.
- Check who issues and who collects. The brand owner is almost never the regulated entity: Compropago for Spin by OXXO, Mobile Money Limited for MTN MoMo in Ghana, and a separate e-money subsidiary in each country for Orange Money.
- Model the cash flow lag. Payment at the counter, notification from the collection agent, payout of funds: three different timestamps, and only one of them is contractually binding.
- Track every unpaid reference. Created, expired, paid, paid late: without those four states, the measured conversion rate is wrong and the reconciliation does not balance.
- Treat agent liquidity as a service risk. A dense network is not necessarily a liquid one, and an agent who runs out of cash turns customers away without leaving a trace in the system logs.
- Calculate what cash costs you before refusing it. Checkout time and back-office work cost more than cash-in-transit, and the average ticket decides the final ranking.