Blog Post
Reducing Breakage in Correspondent Banking (Nostro) Reconciliations
"Breakage" is the silent killer of profitability in international banking. It refers to any transaction that fails to process automatically (Straight-Through Processing, or STP) and falls into an exceptions queue. In the high-speed world of correspondent banking, where margins on payments are razor-thin, manual repairs can cost more than the revenue generated by the transaction itself. In 2026, with the move to ISO 20022 offering richer data but also more complexity, managing breakage in Nostro accounts is critical. This guide explores the root causes and how to fix them.
Why Does Breakage Happen?
Even with modern systems, data mismatches persist. The most common culprits include:
- 1. Truncated References: A payment sent with a 16-digit reference might be truncated to 12 digits by an intermediary bank's legacy system. A standard "exact match" algorithm will fail this.
- 2. The "Principal vs. Fees" Trap: You send $1,000. The correspondent takes a $20 lifting fee and credits $980. Your ledger expects $1,000. This $20 difference causes a break.
- 3. FX Variances: When settling across currencies (e.g., sending GBP to settle a USD invoice), slight differences in the applied exchange rate vs. the booked rate create "dust" that clogs the reconciliation engine.
The Cost of Inaction
A single manual investigation takes an average of 15-20 minutes. If your bank processes 10,000 cross-border payments a day and has a 5% breakage rate, that is 500 repairs daily. This translates to full-time employees doing nothing but fixing data—a massive operational risk and overhead.
See how this impacts operational risk in our Top 7 Risk Tools guide.
How to Automate Repairs
The solution lies in "Intelligent Exception Management" tools like Reconwizz.
1. Smart Tolerance Logic
Instead of demanding an exact match on amount, the system can be configured with tolerances. For example, "Match if Reference is identical AND Amount is within $25." This automatically clears the fee-related breaks, posting the difference to a "Bank Fees" expense account automatically.
2. "Fuzzy" String Matching
To handle truncated references, the software strips out non-numeric characters and looks for substrings. If it finds "123456" inside a string like "REF: 000123456-ABC," it suggests a match with a high confidence score.
3. Learning from History
AI-driven tools learn from manual actions. If a human operator consistently matches payments from "Bank of America New York" with a $15 deduction, the system learns this pattern and starts auto-matching them in the future.
Conclusion: Chasing 100% STP
While 100% Straight-Through Processing is the holy grail, getting from 90% to 98% can save millions. By implementing intelligent matching rules that anticipate fees and formatting errors, you free your operations team to focus on true exceptions—like fraud or liquidity issues—rather than fixing typos. Whether managing Card Settlements or Nostro accounts, the principle remains: automate the predictable, investigate the anomaly.