The Statistic Banks Stopped Talking About

In 2017, the Committee on Payments and Market Infrastructures published data showing that roughly 40 percent of cross-border payments took three to five business days to reach their destination. SWIFT launched gpi in 2018 explicitly to address this. correspondent banking networks were restructured. Billions were invested in payment infrastructure upgrades.

The number has not moved enough.

In some corridors — particularly those involving African or Southeast Asian financial systems — 40 percent still feels optimistic. A corporate treasury team in London moving funds to a supplier in Lagos or Jakarta will tell you privately that three-to-five days is a good outcome, not a worst case. The infrastructure investments of the past seven years improved the best-case path. They did not fix the structural problem.

The structural problem is not the rails.

What the Real Bottleneck Actually Is

Every major correspondent banking chain — and most are three to five institutions deep — has three failure points:

  1. The sending bank's compliance screening. A wire enters the SWIFT network only after AML and sanctions screening clears. Depending on the bank's risk model and the transaction's characteristics, this can take minutes or trigger a manual review that sits in a queue for twenty-four hours.
  1. The correspondent's compliance re-screening. Most large US and European correspondents re-screen every transaction they receive before passing it on — even when the sending bank has already cleared it. This is not redundancy for operational reasons. It is redundancy driven by the bank's own risk appetite and regulatory obligation. The correspondent has no visibility into the sending bank's screening methodology. It cannot rely on the first clearance. So it runs its own.
  1. The receiving bank's settlement queue. Even after funds arrive, the receiving bank may hold them in a settlement queue pending internal reconciliation — particularly in markets where nostro account reconciliation is manual or near-real-time nostro visibility is not available.

The network speed is not the problem. SWIFT gpi has demonstrably reduced end-to-end times for transactions on the gpi track — in corridors where both institutions are participants. The problem is the compliance layer that sits on top of the network, and the structural absence of pre-validation before a payment enters the chain.

This is the distinction that most cross-border payment solutions for banks fail to address. They sell faster rails. The bottleneck is not the rail.

What "Fixing" Cross-Border Actually Means

There are three things that actually reduce cross-border payment latency and failure rates at a bank. They are not glamorous. They are operationally demanding. And they are what separates banks with genuinely competitive cross-border capabilities from those with a better website.

Data Enrichment Before Payment Entry

The single highest-impact change a bank can make to its cross-border payment operation is requiring rich payment data at submission — not inferring it later.

When a corporate customer submits a payment with no Legal Entity Identifier for the beneficiary, no purpose code, and unstructured remittance information ("invoice #47291, attention Accounts"), every downstream institution in the chain must handle that ambiguity. The correspondent cannot efficiently route without a purpose code. The receiving bank's compliance system cannot auto-clear without a counterparty identifier. The nostro reconciliation team cannot match the payment without manual intervention.

ISO 20022's pmtInf/chargeBearer and related fields exist specifically to carry this data. Most banks collect it — or could collect it — but do not make it a requirement at payment entry. Requiring it changes what the payment looks like downstream. Correspondents with structured data in the message can apply automated compliance screening at a confidence level that triggers far fewer manual reviews. Banks with automated compliance screening clear payments faster.

The upstream investment — making rich data a submission requirement for institutional clients — is the downstream operational gain in screening speed.

Pre-Validation: Stopping Problems at the Door

The second major improvement is pre-validation: checking payment data against known reference data before the payment enters the correspondent chain.

This means validating beneficiary bank BIC codes against the SWIFT Refinitiv directory, checking counterparty LEIs against the GLEIF database, running beneficiary names against sanctions lists at submission rather than at screening, and flagging corridor-specific risk signals (high-value transactions to high-risk jurisdictions, payments to entities with recent regulatory actions) before the transaction enters the chain.

Pre-validation catches data quality issues that would otherwise surface as errors at a correspondent or receiving bank — causing delays, manual investigation, or return flows. It also catches sanctions and AML flags before the payment has consumed any network resources or created any nostro exposure.

The banks that have implemented this well run a pre-validation layer between their payment origination systems and their SWIFT gateway. For cross-border payment volumes above 500 transactions per day, the return on that investment is measurable within weeks: lower exception rates, faster settlement, fewer correspondent queries.

Smart Routing: Choosing the Right Chain

Not all correspondent paths are equal. The optimal routing for a payment depends on the sending and receiving currency, the destination country, the transaction size, and the compliance risk profile of the counterparties in the chain.

Most banks use static routing tables — a fixed correspondent chain for a given currency corridor that was set up when the relationship was established and reviewed infrequently since. Static routing does not account for:

Smart routing — routing that dynamically selects the correspondent chain based on real-time corridor intelligence — is the third lever. Banks that have built this layer (typically through a payment hub or orchestration platform that sits above the SWIFT gateway) materially outperform banks on static routing on both speed and cost.

The combination of data enrichment, pre-validation, and smart routing does not require a new payment network. It requires better use of the infrastructure that exists.

ISO 20022: Enabler, Not Silver Bullet

ISO 20022 gets described in banking marketing as the solution to cross-border payment friction. The description is not wrong — the standard does enable richer data, faster reconciliation, and better compliance screening — but it is also incomplete.

ISO 20022 solves the data standard problem. It does not solve the compliance screening problem, the correspondent banking relationship problem, or the nostro reconciliation problem. These are separate issues that happen to intersect at the payment message.

The banks that have made the most of ISO 20022 are the ones that used it as an opportunity to redesign their payment data quality workflows — making rich data collection a client-facing requirement, not a post-submission enrichment step. SWIFT's own gpi tracking is built on ISO 20022 message formats. The more structured data in the message, the more SWIFT gpi can do with it.

But the message format change is a precondition, not a destination. Banks that treat ISO 20022 migration as a compliance exercise — translating their existing flat-data SWIFT MT messages to ISO 20022 XML without redesigning their data collection and validation processes — will meet the regulatory requirement and not capture the operational benefit. The standard is only as useful as the data discipline that feeds it.

The Corridor-Specific Strategy Banks Miss

Cross-border payment performance varies significantly by corridor, and a bank that treats its global cross-border book as a single product will consistently underperform in every corridor it operates.

The specific corridor dynamics that matter:

The banks with competitive cross-border payment solutions manage their corridors as distinct products, each with its own routing logic, correspondent relationships, and compliance profile. This is not how most banks organise their payments function. It is why most banks' cross-border payment performance trails their best-in-class peers by a wide margin.

Cross-Border Payment Compliance: The Structural Challenge

Compliance in cross-border payments is often framed as a regulatory burden to be minimised. That framing is accurate in the sense that compliance creates friction, cost, and latency. It is inaccurate in the sense that the solution is to reduce compliance investment.

The banks that handle cross-border payment compliance well have one structural characteristic in common: they have invested in risk differentiation rather than risk uniformity.

A uniform compliance model applies the same screening rules, the same risk thresholds, and the same escalation triggers to every payment, regardless of the counterparty relationship history, the corridor risk profile, or the transaction frequency pattern. This maximises compliance coverage and minimises compliance cost per transaction. It also maximises false positive rates in low-risk corridors and under-screens high-risk corridors where the baseline rules were calibrated for a different risk environment.

A differentiated compliance model — one that uses transaction history, counterparty relationship data, corridor-specific risk signals, and behavioural analytics to calibrate screening intensity — produces materially lower false positive rates for well-understood relationships and more targeted escalation for genuinely anomalous transactions.

The operational investment in building that differentiated model is real. The data integration required is non-trivial. But the output — faster clearance for legitimate payments, more accurate risk signal, lower manual review costs — is worth it.

SWIFT's KYC (Know Your Customer) Registry is one piece of infrastructure that supports differentiated compliance. Banks that use the registry to reduce redundant counterparty screening in established correspondent relationships — rather than running full screening on every transaction — have measurably lower per-transaction compliance costs and faster settlement in those corridors.

What Spectral Partners Can Do

The cross-border payment problems I have described — data quality at entry, pre-validation, smart routing, compliance differentiation — are addressable. They are addressable with specific process changes, specific technology choices, and specific corridor investment decisions.

What they require is an honest assessment of where the friction actually is. Not a vendor presentation on faster rails. Not a consultant's slide deck with best-practice frameworks. An honest map of your payment chains, your compliance false positive rates, your correspondent routing logic, and the actual settlement times you are achieving versus what the network can deliver.

We offer that assessment as a structured engagement: a cross-border payments diagnostic that maps your current state, identifies where the latency is actually being created, and produces a prioritised remediation roadmap with cost and timeline estimates. The output is not a strategy document. It is a set of specific, sequenced recommendations that your engineering and operations teams can act on.

The banks that have done this work consistently find the same thing: the problem was never the network. It was the three inches upstream.

Want a cross-border payments diagnostic for your institution? Talk to Amara.

Amara leads Cross-Border Payments research at Aicura Consulting, with a focus on correspondent banking infrastructure, cross-border payment compliance, and institutional payment flows across African and emerging market corridors.