Which Payment Architecture Fits Which Corridor?

The subject here is the choice of infrastructure that connects two or more domestic payment systems across a border. The practical question is not which model sounds most modern, but which architecture addresses a corridor’s actual bottleneck—too many intermediaries, too few connections, incompatible data, difficult settlement, weak access, or slow beneficiary-leg decisions. Comparing those choices helps distinguish a faster route for payment instructions from a system that can reliably turn an incoming payment into usable funds.

There is no universal winner. A corridor linking two highly integrated systems may benefit from a direct connection. A region with many participating systems may need a hub. A group of jurisdictions willing to share governance and settlement rules may build a common platform. A large-value institutional corridor may continue to rely on correspondent banking while upgrading its standards and operating arrangements. Each choice moves complexity somewhere; none makes the receiving bank’s responsibilities disappear.

Four ways to connect payment systems

A single access point: simple entry, narrow reach

The earliest cross-border retail arrangements often used a single access point. One provider connected to the other system, and customers reached that foreign system indirectly through that provider. The model is relatively simple to establish because few participants have to coordinate. It can work for a defined corridor or a limited use case, particularly when a provider already has the necessary relationship and access.

Its weakness is that the apparent simplicity belongs to the infrastructure, not necessarily to the customer’s full payment. The instruction can still start a long chain after the front-end provider receives it. Indirect access may mean more handoffs, more opportunities for errors, and repeated checks. The BIS review describes the model as difficult to scale, limited in access, slow, and not cost-effective for small transactions. A single access point may solve a connection problem for one provider while leaving other banks and payment-service providers outside the arrangement.

This model is therefore not a failed version of every other model. It is a choice for a narrow connection, with a narrow capacity to create competition or broad interoperability. If a corridor needs only one established channel and the volume is sufficient to support it, that trade-off may be acceptable. If the goal is to make many providers and end-users interoperable, the same design becomes a bottleneck.

Bilateral links: directness with a scaling limit

A bilateral link connects two domestic payment systems directly. Participants in one system can reach participants in the other without routing through an entity that acts as a traditional correspondent. The arrangement has an important coordination advantage: only two systems need to agree on the technical and operational connection. Incompatibilities can be addressed directly, and the route can reduce reliance on intermediaries.

The India–Singapore link between the Unified Payments Interface and PayNow illustrates the promise. The systems use APIs so participating customers can make instant, low-cost transfers without joining the foreign fast-payment system. Direct interlinking can improve the user experience and reduce unnecessary steps, but it does not eliminate all dependencies. Many linked fast-payment systems still rely on traditional banks for currency conversion. They also need to manage real-time operation, liquidity, credit, security, and the risks that follow from increased use.

The deeper limit is scale. If every domestic system needs a separate connection to every other system, the number of relationships grows rapidly. Each new link brings technical integration, legal agreement, operational testing, governance, and dispute-resolution work.