Correspondent Banking Is Quietly Dying. Here Is What Replaces It.

The number of active correspondent banking relationships has fallen by roughly a quarter in the last decade. The decline has been gradual enough that most people inside the industry have stopped noticing it. They should start noticing again, because the replacement architecture is finally visible.

Why the old model is shrinking

Correspondent banking is the system by which a bank in one country settles payments in another country by holding accounts at a partner bank in that country. It works. It has worked for decades. It is also expensive, slow, opaque, and increasingly hard to justify from a compliance standpoint.

The driver of the decline has been the cost of compliance. Every correspondent relationship requires AML monitoring, sanctions screening, and counterparty due diligence. As regulatory expectations have risen, the marginal relationship has become uneconomic. Banks have responded by closing tail relationships, which has the side effect of cutting off entire corridors, particularly in emerging markets.

What is replacing it

Three pieces of infrastructure are converging into a credible replacement, and they are arriving faster than most observers expected.

First, real-time gross settlement systems are now live in more than seventy markets. Their domestic adoption is mature. Their cross-border interlinkages are finally being built. The BIS-led nexus project and similar initiatives are not theoretical anymore.

Second, regulated stablecoins are functioning as a settlement layer between regulated counterparties. The volumes are not small. Whatever your views on the consumer use case, the institutional settlement use case is real and growing.

Third, payment network operators are building their own cross-border products that bypass the correspondent network entirely. These are still smaller than the correspondent rails by volume, but they are growing faster.

What this means for banks

For banks that earn meaningful revenue from correspondent banking, the strategic choice is no longer abstract. The corridor revenue you have today is unlikely to be there in five years in its current form. The question is whether your institution wants to be the provider of the replacement rails or a customer of someone else's.

The banks that have decided to be providers are investing heavily in their own real-time interlinkages and in selective stablecoin partnerships. The banks that have decided to be customers are quietly winding down their international correspondent networks and outsourcing the function.

There is no third option that involves doing nothing.

What this means for everyone else

For corporates and fintechs sitting downstream of these flows, the practical implication is that cross-border payments are about to get materially cheaper, faster, and more transparent over the next three to five years. The corridors that have been stuck at five-day settlement and opaque fees for decades are going to look very different.

The interesting question is what new products that enables. My view is that real-time, low-cost cross-border settlement will unlock a wave of treasury and working capital products that have not been economically viable until now. That is where the next decade of fintech innovation will come from.