Why Adding More Bank Relationships Doesn't Solve Settlement Speed

Bank relationships are the connections a business maintains with banks to route and settle payments, but adding more of them rarely fixes slow cross-border settlement. When cross-border payments crawl, the instinctive fix is to add another bank. One more correspondent, one more nostro account, one more mutually beneficial relationship — and surely settlement times will improve. It's a strategy most treasury and payment teams, PSPs, EMIs, payment companies, and other businesses handling cross-border payments have tried at least once. It's also a strategy that recent economic research suggests is largely misdirected.

The problem isn't the number of bank relationships a business holds. It's the payment network structure those relationships sit inside — and the trade-offs that structure imposes on settlement speed, liquidity costs, and counterparty risks. For teams trying to improve operational efficiency and financial stability, that distinction matters: inefficient settlement ties up liquidity, slows customer payouts, and increases exposure to counterparties. This article examines how bank relationships affect settlement speed, why the limits are structural, what research says about speed and systemic risk, where adding multiple banking partners backfires, and which alternatives — including FinchTrade's approach to settlement optimization — can improve payment infrastructure more effectively.

Key Point Summary

• What the economic research actually says about settlement speed
• The paradox: faster settlement can amplify crisis severity and impact financial stability
• Why more relationship banking doesn't fix the problem
• What actually improves settlement speed

What the economic research actually says about settlement speed

In 2025, a working paper by Agostino Capponi (Columbia University) and Jin-Wook Chang, published in the Finance and Economics Discussion Series of the Board of Governors of the Federal Reserve System, tackled this question head-on. The paper investigates how settlement speed affects financial stability in payment networks, accounting for netting benefits, liquidity costs, and counterparty risks. Capponi's related work on financial networks has circulated through the National Bureau of Economic Research, and this line of economics research has become required reading for anyone designing payment infrastructure.

The headline finding is uncomfortable for the "faster is always better" school of finance: faster settlements have ambiguous effects on systemic risk and social welfare. The optimal settlement speed is not a universal constant — it depends on network topology and on the trade-off between netting efficiency and liquidity costs on one side, and counterparty default probability on the other.

Here's the intuition. When settlement is slower, obligations between members of a payment network accumulate and can be netted against each other. If Bank A owes Bank B and Bank B owes Bank A, only the difference needs to move. These netting benefits reduce the gross liquidity each participant must hold. Faster settlement strips those benefits away: every payment must be funded in full, immediately, which raises liquidity costs across the network — especially for smaller participants and the local businesses that depend on them for services.

On the other side of the ledger sits counterparty risk. The longer an obligation remains unsettled, the longer each party is exposed to the possibility that the counterparty defaults before the money arrives. Slower settlement widens that exposure window; faster settlement narrows it, lowering counterparty default probability at any given moment.

The net welfare effect depends on which force dominates — and that, in turn, depends on liquidity conditions and network structure. There are threshold points where the calculus flips entirely.

The paradox: faster settlement can amplify crisis severity and impact financial stability

The most striking result in the Capponi–Chang paper concerns financial crises. Under deteriorating liquidity shifts — the exact conditions in which everyone suddenly wants faster settlement — accelerating settlement can paradoxically increase systemic risk. The mechanism: faster settlement reduces crisis likelihood, because exposures are open for shorter periods, but it amplifies crisis severity when a crisis does occur.

Why? Because when settlement is fast and netting is minimal, every participant must produce liquidity on demand. In stressed conditions, defaulting agents transmit their failure through the network almost instantly — there is no buffer of pending obligations to net down, no time for other members to arrange funding, and the outcome is a sharper, deeper cascade. The paper argues explicitly against a one-size-fits-all approach to settlement speed design, a conclusion with obvious implications for regulators — and for the governors of central banks weighing real-time settlement mandates — but also for any business making infrastructure decisions.

The lesson for financial stability, per Agostino Capponi and Chang, is that speed is a design parameter, not a virtue. And that reframing is exactly why piling on bank relationships fails as a strategy.

Why more relationship banking doesn't fix the problem

Relationship banking has genuine value. A long-standing relationship with one bank can mean better credit terms, smoother compliance reviews, priority support, and access to a broader range of products offered. Economics has studied this for decades — the relationship banking literature (much of it from business school research on how banks price loans to repeat customers) shows measurable benefits: information advantages, cheaper credit, and confidence on both sides of the table.

But none of those benefits changes the physics of settlement. Adding bank relationships to speed up payments fails for four structural reasons.

1. You're adding nodes, not changing topology. Settlement speed is a property of the payment network structure, not of any individual bank. If your payments to Lagos still route through two correspondents, a foreign exchange conversion, and a cut-off-time queue, opening an account at a third bank in Frankfurt changes nothing about that path. Network topology determines how value flows; a new node on the periphery doesn't restructure the core.

2. Fragmentation destroys netting benefits. This is the direct application of the research above. Spread your flows across five banks instead of one, and obligations that could have been netted within a single relationship now settle gross across five. Your aggregate liquidity requirement rises. Each account needs its own buffer, its own funding positions, its own minimum balances. The liquidity costs of fragmentation are real money — capital parked as prefunding instead of working as investments in the business.

3. Each relationship multiplies counterparty and operational exposure. Every additional bank is another counterparty whose security, solvency, and operational reliability you now depend on. Financial crises have repeatedly shown that counterparty risks correlate exactly when you need diversification most. More relationships also mean more onboarding cycles, more compliance overhead for whoever is managing the accounts, more reconciliation for accounting teams, and more places where a payment can silently stall. What began as a treasury advantage becomes a management burden — a standing concern for every board that has watched banking-partner sprawl accumulate.

4. The bottleneck is usually the rail, not the bank. Correspondent banking settles on infrastructure with cut-off times, batch windows, and time-zone gaps. A payment from a European PSP to an African partner can involve several intermediaries, each applying its own controls. No bilateral relationship, however strong, removes hops that are baked into the rail itself. Technology at the endpoint can't fix latency in the middle.

The example plays out constantly in practice: a payments company involved in emerging-market corridors adds its fourth and fifth banking partner, expecting faster settlement. Twelve months later, settlement times are unchanged, liquidity is scattered across five accounts, and the finance team's focus has shifted from growth to reconciliation. The customer experience hasn't improved at all.

What actually improves settlement speed

If more banks aren't the answer, what is? The research points to the real levers.

Change the rail, not just the counterparty. Where the strategy aims at genuinely faster settlement, the structural fix is moving value on rails that settle in near real time. Stablecoin-based settlement, for instance, collapses the multi-hop correspondent chain into a single transfer that clears in minutes, around the clock — no cut-off times, no batch windows. That is a topology change, not a node addition.

Concentrate flows where netting works. Rather than fragmenting across many banks, routing flows through a single liquidity provision point restores netting benefits. Trading counterparties that can net your buys and sells, your inflows and outflows, reduce the gross liquidity you must hold — the same access to markets with a fraction of the prefunded capital.

Match settlement speed to liquidity conditions. The Capponi–Chang framework implies firms should treat speed as adjustable. In calm conditions, faster settlement is cheap insurance against counterparty exposure. When liquidity conditions arise that are stressed, the ability to net and sequence payments is worth more. Infrastructure should support both.

Pick partners that compress the chain. The right partner isn't the sixth bank — it's the counterparty that combines FX conversion, liquidity, and settlement in one hop, with the security and compliance posture your regulators expect.

Conclusion

At FinchTrade, our bankers and settlement specialists built precisely the problem we were built to solve. As a Swiss-regulated OTC desk and cross-border settlement partner, we give PSPs, EMIs, and payment companies a single point of access to deep digital-asset and FX liquidity — with T+0 settlement, no prefunding requirements, and netting across flows. Instead of adding another bank relationship and hoping the network improves, our clients change the network itself: fiat and stablecoin legs settle through one counterparty, in one hop, at optimal settlement speed for their corridors.

The research is clear that settlement speed is a structural property, not something you can buy by collecting banks. If your settlement times are the bottleneck, the answer is a better-designed page in your payment stack — not a longer list of relationships. Talk to FinchTrade about restructuring your settlement flows around speed that's actually engineered in.

For requesting more information about how we can help reach out to us. We're here to help and answer any questions you may have.

Contact us!