MiFID Duplicate Trades: When Transaction IDs Create Silent Errors

Silent Errors in Your MiFID Transaction Reports

Duplicate trades can occur in MiFID transaction reporting even when Transaction IDs are unique. This guide explains common root causes (rebooking, system faults, replayed messages and amendment handling) and practical controls to detect duplicates without false positives such as block fills.

Detecting Duplicate Trades in Front Office Systems: Risks and Mitigation

Front Office systems are complex environments that can occasionally produce unexpected outcomes. One such risk is the creation of duplicate trades, which may go undetected without appropriate controls in place.

Causes of Duplicate Trades

Most duplicate trades are identified by filtering controls that flag repeated Transaction IDs. However, complications arise when a Front Office system or associated process generates a unique Transaction ID for what is effectively the same trade. This can occur for several reasons, including:

• User input error, such as the manual rebooking of a trade
• System faults, including issues in booking engines or replayed messages
• Coding errors in bespoke middleware or trade extraction processes
• Incorrect handling of trade amendments, where a system erroneously creates a new transaction rather than using a cancel and rebook process (i.e., issuing a CANC followed by a NEWT)

ESMA and the FCA place responsibility on investment firms to ensure their transaction reporting is complete and accurate. This is reflected in MiFIR RTS 22, Article 15.

Challenges in Identifying Duplicates

Not all trades that appear similar are duplicates. A common example is a block fill order, where a single client order is executed in multiple parts. These trades may share identical characteristics such as:

• Buyer and Seller
• Price and Quantity
• Trade date and time (to the second)

However, each execution is a legitimate market transaction and should be reported separately under MiFID II requirements.

This makes it difficult to rely solely on simple matching logic, as it increases the risk of false positives, where trades are flagged as duplicates despite being genuine.

Detection Techniques

To reduce false positives and improve accuracy:

1. Rule based filtering should exclude known trade patterns such as block fills or strategy trades. 2. Some firms employ historical pattern matching, which compares each new transaction with previously reported trades to detect potential duplicates caused by errors in trade amendments or resubmissions. 3. Quality assurance sampling is another effective approach. This involves selecting a sample of trade data and applying a series of logic and consistency checks to flag anomalies. Although less comprehensive than full data scans, this method is often sufficient to detect underlying issues while requiring fewer resources.

Resolving and Preventing Duplicates

Once a duplicate trade has been confirmed, the next step is to identify its root cause. This could involve reviewing audit trails, system logs, or manual inputs. Firms should then:

• Adjust system logic or booking processes to prevent similar issues in the future
• Implement preventative controls that check for duplicate trade characteristics within a given time frame
• Strengthen reconciliation processes between Front Office systems, regulatory reporting engines and other downstream platforms

These actions align with best practice guidance outlined in the FCA's Market Watch publications and more recently in Discussion Paper DP24/2, which highlight the need for stronger governance and control over transaction reporting processes.

Conclusion

Duplicate trades represent a significant compliance and operational risk, especially in the context of transaction reporting regimes such as MiFID II. Firms must adopt a multi-layered approach to detection, incorporating rule based filters, historic trade comparisons and quality assurance sampling. This not only meets regulatory expectations but also reinforces the integrity of trade data throughout the organisation.