Loading field map
Every payment is a promise that must be settled. The core tension in payments technology is how to move value between two parties when they do not trust each other, when they are not in the same place, and when the transfer must be final. Over the past century, six distinct frameworks have emerged to solve different versions of this problem. They did not replace one another in a clean sequence. Instead, each new framework addressed a gap—speed, access, convenience, or cost—that earlier systems left open, and the older frameworks adapted rather than disappeared. Today all six coexist, layered on top of one another, each handling a different slice of the world's payment volume.
The first framework solved the problem of moving money between banks. Before electronic systems, banks settled accounts by physically transporting gold or cash. The Federal Reserve launched Fedwire in 1915 as a telegraph-based system for transferring funds between member banks, settling in central-bank money—the ultimate trusted asset. The Automated Clearing House (ACH) network, formalized in the 1970s, added batch processing for recurring payments such as payroll and bills. SWIFT, founded in 1973, standardized the messaging that lets banks communicate payment instructions across borders. These systems share a hierarchical trust model: each bank trusts the central bank or clearing house, and the end customer trusts their own bank. Settlement is final but slow—ACH batches take one to three days, and international wire transfers can take several days. This framework remains the foundational plumbing for all other payment frameworks. Every card transaction, every PayPal transfer, and every instant payment ultimately settles through a bank-account-to-bank-account transfer.
Account-to-account systems required both parties to have bank accounts and to share bank identifiers. In the 1950s, merchants wanted a way to accept payment from any customer without checking their bank details. The card-network framework, launched by BankAmericard (later Visa) in 1958 and Mastercard in 1966, created a two-sided market: consumers get a card, merchants get a terminal, and the network connects them. The trust model shifted from hierarchical bank trust to brand trust—the Visa or Mastercard logo guaranteed that the merchant would be paid. Settlement still happens through bank accounts, but the card network adds an extra layer: the issuing bank pays the acquiring bank, and the network sets interchange fees that allocate costs between them. This framework dramatically expanded payment access for consumers and merchants, but it introduced a new coordination problem: the network had to balance the interests of issuers, acquirers, merchants, and cardholders. The card-network framework coexists with account-to-account systems by sitting on top of them for settlement while adding brand-based authorization and consumer credit.
As card networks grew, a specialized function emerged: merchant acquiring. An acquirer is a bank or licensed entity that underwrites the risk of a merchant, handles settlement routing, and manages chargebacks. In the early card era, acquirers were the only way a merchant could accept cards. They performed the full set of merchant-facing tasks: application review, terminal provisioning, transaction processing, and dispute resolution. Over time, this framework narrowed. The rise of third-party processors and payment facilitators absorbed many of the merchant-facing functions, leaving acquirers primarily as risk-bearing back-end entities that hold the merchant's settlement account and assume liability. The narrowing of merchant acquiring is a case of absorption: later frameworks took over the customer-facing parts while the acquirer retained the regulatory and financial risk. Today, a merchant may never interact directly with its acquirer; the acquirer's name appears only on the settlement statement.
The card-network and merchant-acquiring frameworks required merchants to apply for a merchant account, integrate with a processor, and pass underwriting checks. For small businesses and online startups, this friction was prohibitive. PayPal, launched in 1998, introduced a different model: the payment service provider (PSP) aggregates many merchants under a single merchant account. The PSP becomes the merchant of record, handling onboarding, transaction processing, fraud screening, and settlement. The merchant never needs a direct relationship with an acquirer or card network. This framework competed directly with merchant acquiring by simplifying onboarding to a few clicks and by offering a unified API that abstracted away the complexity of multiple payment methods. PSPs also shifted the value proposition from pure transaction processing to data services—fraud analytics, subscription billing, and multi-currency management. The card networks initially resisted this aggregation model but later embraced it through programs like Visa's Merchant Aggregator Program. Today, PSPs and payment facilitators coexist with merchant acquirers: the acquirer still underwrites the aggregated risk, but the PSP owns the merchant relationship.
Account-to-account and card-network frameworks both assumed that the user had a bank account. In many parts of the world, especially sub-Saharan Africa, that assumption did not hold. M-Pesa, launched in 2007 by Safaricom in Kenya, solved a different coordination problem: how to transfer value using only a mobile phone and a network of local agents. The trust model was neither bank trust nor brand trust but telco trust—users trusted the mobile network operator to keep a ledger of balances and to allow cash-in and cash-out at agent kiosks. M-Pesa did not require a bank account, a card, or internet access. It expanded financial inclusion on a scale that earlier frameworks had not reached. Meanwhile, mobile wallets such as Apple Pay (2014) addressed a different gap: convenience. They digitized the card credential, replacing the physical plastic with a smartphone, but still relied on the card-network framework for authorization and settlement. The mobile money framework is thus internally diverse: one branch (M-Pesa) bypasses banks entirely, while another branch (Apple Pay, Google Pay) sits on top of the card networks. Both branches share a common insight: the phone becomes the trusted device for initiating payment, displacing the card or the bank branch.
Account-to-account systems settled slowly, and card networks settled in one to three days. By the late 2000s, consumers and businesses expected real-time movement of funds. The United Kingdom launched the Faster Payment System in 2008, and similar systems followed in India (IMPS, 2010), the Eurozone (TARGET Instant Payment Settlement, 2018), and the United States (FedNow, 2023). These fast rails are a modernization of the account-to-account framework: they use the same bank-account infrastructure but add 24/7 processing, immediate finality, and standardized messaging. The trust model returns to central-bank settlement, but with a crucial difference: settlement happens in seconds rather than days. Fast rails challenge the card-network framework on speed and cost—a real-time bank transfer can be cheaper than a card transaction—but they lack the card network's consumer protections, rewards, and global reach. They also compete with mobile wallets by offering a direct bank-to-bank alternative. Today, fast rails coexist with card networks and mobile money, each serving different use cases: instant person-to-person transfers, bill payments, and merchant transactions where speed matters more than credit.
No single framework has dominated because each solves a different trust and settlement problem. Account-to-account systems provide the ultimate settlement layer in central-bank money, but they are slow and require bank access. Card networks add brand trust and consumer credit but introduce interchange costs and settlement delays. Merchant acquiring handles risk but has narrowed to a back-end function. PSPs and payment facilitators lower onboarding friction but aggregate risk under a single merchant account. Mobile money reaches the unbanked but depends on agent networks or device ecosystems. Fast rails offer speed but lack the features of card networks. The frameworks agree on one thing: settlement must eventually happen in bank money. They disagree on who should intermediate the transaction, how fast settlement should occur, and what protections to offer. The result is a layered system where a single payment may touch three or four frameworks—a card transaction authorized by a network, processed by a PSP, settled through an acquirer, and finally cleared over an ACH or fast rail. Understanding the payments subfield means understanding how these layers interact, not which one will win.