The correspondent banking model that has enabled cross-border payments for the past 50 years is built on trust; trust that each party in a complex chain will act in good faith and fulfill its obligations.
However, the absence of purpose-built technology and the dependence on manual processes at originating institutions and correspondent banks have prevented effective and efficient counterparty engagement and risk management.
Want proof? The United Nations Office on Drugs and Crime estimates that between 2% and 5% of the global GDP, or up to $2 trillion, is laundered through banks each year. Also, according to Accuity Research, de-risking has driven a 25% reduction in global correspondent banking relationships. Ask any central banker and the message is the same: cross-border payments through financial institutions are high risk, slow, expensive, opaque, and not inclusive. Fear and friction define this business.
The demand for cross-border payments has surged in recent years, highlighting the urgent need for new solutions that can address counterparty risk within the correspondent banking system. Otherwise, digital innovators and fintechs will continue to marginalize banks for essential services (deposit taking), but disintermediate them on the originating side, taking their customers and profits. And the “real risk” at banks and payment systems from these players doesn’t go away.
Where It Breaks Down
In the current model, a respondent bank—typically a smaller or regional financial institution—approaches a larger correspondent bank and requests access to its global network to facilitate international payments on behalf of its customers.
“The bigger bank says, ‘Great, let us see your policies and procedures regarding KYC, AML, pick the abbreviation,” said Gary Palmer, President and CEO of Payall. “’OK, we’ve approved you, here are commercial terms, funding policies and our message format. Every six months up to two years, we’re going to audit your records to see if you’ve followed your own policies.’ And it’s off to the races.”
Typically, the correspondent bank submits its policies and procedures for clearing payments and managing counterparty risk to its regulator, and they periodically audit these to ensure compliance.
Additionally, the central bank or relevant regulator requires a business plan from the originating bank as well and examines their adherence to procedures.
“Here’s where it breaks down,” Palmer said. “Those policies and procedures have been layered on top of each other—layer after layer—for 50 years. I learned this in 2017, when I met with dozens of originating institutions, correspondent banks, central banks, and regulators, and discovered there had never been software built as a part for any of the parties to digitize this business.”
“No core system, no digital bank platform and no risk or compliance engine has ever been built to address the problem set that each one of those institutions faces for cross-border payments,” he said. “They built core bank systems to do deposit accounts, car loans, maybe savings accounts or checking accounts, but never cross-border payments. So, the only way for a bank—whether it’s an originating bank, a correspondent bank, or an intermediate bank—to operate is by manual workflows.”
Nostro and Vostro
This model poses substantial risks and costs for financial institutions as well as frustrations and high costs to users. For example, a bank in Brazil may have customers who collectively make $4 billion in payments to recipients in the United States every year, with an average transaction size of $252,000. This process is facilitated by a U.S. correspondent bank.
In this scenario, the Brazilian bank might advertise on their website that it can make bank transfers to the U.S.
“There’s a concept called nostro and vostro where you’ve got banks that have pots of cash with one another,” said Hugh Thomas, Lead Commercial & Enterprise Payments Analyst at Javelin Strategy & Research. “The nostro is mine that sits with you, and vostro is yours that sits with me. They just sort of net and pool at the end of every day and figure out, ‘OK, you’ve got this much more vostro with me and I’ve got this much more nostro with you as a consequence of us having done these transactions.’ Those are, in many cases, manual processes.”
The nostro account created by the Brazilian bank may hold significant amounts of USD. These dollars sit on the bank’s balance sheet and are subject to market volatility until a customer decides to send their $252,000 payment to a U.S. recipient.
“Guess what happens next?” Palmer said. “A payment instruction goes to the backroom of the Brazilian bank where their customer made the international transfer. A human being looks at it and says, ‘Does my customer have enough money in their account? Do I have enough money in my nostro account? Does my correspondent support this type of payment? Oh, and the customer is in pharmaceutical or they’re in gaming or they’re in furniture construction—will my correspondent bank support the commercial activity or the source of funds?”
Adding to the complexity, Brazil, like many countries, has specific requirements set by its central bank for these payments. For example, if a transaction exceeds a certain threshold, then workers at originating institutions must collect additional data, such as a bill of lading or an invoice, which must then be reported to the regulator.
These country-specific nuances introduce additional manual checks to an already intensive process.
“Based on the size of the payment, what does my regulator say I need in terms of data documents and artefacts to substantiate the payment?” Palmer said. “Did my correspondent bank mandate any particular rules based on the size of the payment? What are my internal risk and compliance rules? This results in millions of manual touchpoints at originating institutions and correspondent banks every day.”
The Essence of Moving Trillions
While the answer to most of these questions may be affirmative, this time-consuming process must be repeated with each transaction. Moreover, these lengthy manual interactions increase the risk of errors or manipulation.
For example, a bank employee must not only review whether all the required documentation has been submitted but also ensure the documents haven’t been altered or forged. They may then verify whether the payment involves a sanctioned vessel, port, company, product, or currency.
As a result, a single payment could require reviewing hundreds of pages.
“Now the bank says, ‘What do we do with all this stuff? There’s no way to tie all this human-collected data, rule execution and decision-making with a payment and store it at the bank’s core.” Palmer said. “So, they file all this stuff away in paper folders, and it sits until either the Brazilian regulator comes in to ask for it, or the American correspondent bank or their regulator, as part of an audit or examination, asks for it. And how effective is it to manually audit 0.000001% of all transactions?”
“You’re trusting the foreign bank – in the case of the correspondent bank – to execute these things because you can only audit a tiny number of transactions and you’re auditing them two months, six months, two years after the fact,” he said. “This has been the essence of moving $160 trillion last year through the banks around the world.”
Digitizing the Process
Though this model has dominated for decades, several forces are driving a shift to a new paradigm. First, the consumer experience with digital payment solutions has raised expectations across the board. When users can send peer-to-peer domestic payments in seconds, they expect the same speed and transparency in cross-border payments.
Along with these heightened expectations, cross-border payments demand has surged in recent years.
However, with no global standardization of payments regulations likely anytime soon, institutions must implement systems that enable counterparty verification without intensive manual checks. These platforms can digitize the due diligence process and radically improve the efficiency of onboarding counterparties.
“Imagine a series of AI agents surveilling that counterparty over time,” Palmer said. “Have there been changes in the ultimate beneficial owner? Have there been any citations, letters, or interventions from a regulator of the counterparty? Have there been any issues from a financial reporting perspective, or irregularities, or terms of profitability?”
“New software digitizes just about everything: originating bank applications, including document collection and diligence, the boarding process, the examination and evaluation of everything from the Wolfsberg questionnaire all the way to their critical infrastructure,” he said. “It digitizes the relationship—the CRM of the counterparty as well as enables 100% real-time see-through to how an originating institution executed their risk and compliance rules, replacing slow, costly, and human-centric periodic audits of a small percentage of transactions with perpetual visibility and digital decisioning of every transaction. This capability has never existed before and replaces trust with data, removing fear and friction at correspondent banks.”
