A failed status is useful for routing an exception. But it still leaves the reviewer with the real work: reconstructing what happened.
What was expected to settle? What did the source actually report? What remained unsettled? What happened on the cash and security legs? What actions followed? Who owns the exception, and what evidence supports closure?
That is more useful than reducing the review to failed or not failed.
What a status cannot tell you
Settlement systems can report different states and produce different evidence objects. Status advice and settlement confirmation, cancellation and reversal are not interchangeable. Source-native reason codes can also point to very different operational conditions.
The useful habit is simple: do not flatten them into one generic lifecycle. Keep the source-native state, reason, reference and timestamp from the relevant market or account-servicer workflow.
For example, the current online ISO 15022 User Handbook distinguishes PEND and PENF as different source-native concepts. It also lists distinct reason codes for conditions involving securities availability, money, holds, linked instructions, investigation and partial processing.
Treat those codes as reported evidence, not conclusions. They can tell you what the source said; they do not, on their own, establish root cause, blame or liability.
One version detail matters here: the online 2026 MT User Handbook was published on 17 July 2026, while Standards Release 2026 goes live on 14 November 2026. If an article quotes current handbook details, refresh the relevant implementation and release context before publication.
The failed-trade evidence chain
Start with two questions: what should have happened, and what does the source say was observed?
At minimum, a reviewer should be able to answer:
- Which source-native instruction or trade references identify the event?
- When was it meant to settle?
- What quantity and, for an against-payment trade, what amount and currency were expected?
- What source-native status and reason did the source report, and when?
- Did an effective settlement event occur later?
- If settlement was partial, what settled and what remained?
- Which evidence source supports each answer?
If settlement later succeeds, do not overwrite the earlier failing observation. If the intended settlement time and the effective settlement time are both available, keep both. The gap between them is part of the operational history.
The same goes for partial settlement. Source documentation can represent settled and remaining quantities or amounts. A partial confirmation can be useful evidence without implying that the whole exception is resolved.
A reason code is evidence, not blame
Keep the source-native reason code beside its original source and context.
A team may map native reasons into simpler internal categories such as SECURITIES_AVAILABILITY, CASH_OR_LIQUIDITY, TIMING_OR_CUTOFF, HOLD_OR_RELEASE, LINKAGE_OR_DEPENDENCY, or INVESTIGATION_OR_UNKNOWN.
That mapping is reviewer normalization, not a market standard. Keep the native code and narrative visible beside it.
Why this matters: a reported condition such as insufficient securities can help explain the observed settlement state. It does not establish why the shortage occurred, whether a control failed, or who is liable.
Cash break and position break are separate evidence questions
A failed settlement can leave two separate questions on the desk: what happened to the cash leg, and what happened to the security or position leg.
For cash, relevant evidence may include the expected settlement amount and currency, a source-reported settled or remaining amount, value-date information, and a cash-account report, statement or debit/credit notification reference when available.
For the security leg, the reviewer may need the expected quantity, the effectively settled quantity, the remaining quantity, and a holdings statement or transaction statement with its as-of or period context.
These are operational evidence questions. They do not decide journal treatment, valuation or NAV.
That is why a useful reviewer note is not NAV impact = X. It is closer to professional review required — position evidence unresolved or valuation/NAV question remains for the responsible specialist.
External action is part of the record
The status is only part of the story. What happened next belongs in the record too.
Depending on the source and market, that history may include an enquiry or investigation status, a hold or release, a partial settlement, a cancellation request, a cancellation of a prior confirmation, a reversal, a modification, or a new instruction linked to a prior instruction.
When available, keep the source-native action, timestamp and reference. That is more useful than forcing every event into one normalized label.
An action record shows that the source reported an action or status. It does not, by itself, show that the action was timely, effective, compliant or legally determinative.
One leg cleared does not automatically close the exception
A common shortcut is to see one successful event and call the whole exception finished.
If the security leg is evidenced as complete but cash evidence is still unresolved, the record still has an open cash question. If cash is evidenced but the remaining security quantity is not, the position question remains open.
Partial settlement, cancellation and reversal need the same care. They are events in the history, not reasons to erase what came before.
If later evidence changes a prior closure, keep the close event and record why and when the record was reopened, along with the evidence that triggered it.
The Failed-Trade Exception Record
Treasury Desk author synthesis — reviewer-facing Failed-Trade Exception Record
This is not an industry standard, accounting template, audit workpaper, legal requirement, or CSD/CCP/custodian schema. It is simply a way for a reviewer to bring the evidence into one place without replacing the underlying records.
| Field | Reviewer question |
|---|---|
| Trade / settlement references | Which source-native trade, order, instruction or message references identify the event and its linkages? |
| Expected state and time | What was expected to settle, when, and from which source? |
| Actual source-native state | What status did the source report, at what time, and under which source/version context? |
| Reason / cause evidence | What native reason code or narrative was reported? Is any normalized cause clearly labelled as synthesis? |
| Cash impact evidence | What cash amount/currency/value-date evidence exists, and what remains unresolved? |
| Position / security evidence | What quantity was expected, settled and remaining, and which statement or transaction evidence supports it? |
| External action history | What enquiry, hold, release, partial, cancellation, reversal or other source-supported action followed? |
| Reviewer impact | Is professional review required, and which unresolved question must be handed off? |
| Exception owner | Which functional role owns the exception, and when was it assigned or escalated? |
| Evidence links / references | Which controlled source references allow a reviewer to reconstruct the record? |
| Open / closed / reopened state | What is the reviewer state now, and what event changed it? |
| Closure evidence | What source-native outcome, leg evidence, action history, handoff and closure reference support the bounded closure decision? |
| Remaining unresolved items | Which question is still open or explicitly outside this record’s scope? |
For this reviewer artifact, a simple state model can help: OPEN → READY_FOR_REVIEW → CLOSED, with REOPENED as an explicit post-close state.
That state model is Treasury Desk author synthesis, not a universal settlement standard.
For this artifact, CLOSED means the bounded operational record is complete for the review job. The outcome is evidenced; the material cash and security questions are reconciled or explicitly handed off; relevant actions are linked; professional-review questions are cleared or transferred to the responsible function; and the closure owner, time, reason and evidence references are recorded.
That is intentionally narrower than legal finality, accounting correctness, audit clearance or compliance closure.
Wholly synthetic example — every value below is synthetic
A synthetic record might show:
SYNTHETIC exception_record_id:SYN-FT-001SYNTHETIC instrument_ref:SYNTHETIC-SEC-001SYNTHETIC expected_settlement_at:2026-08-18T10:00:00ZSYNTHETIC expected_quantity:1,000 SYNTHETIC_UNITSSYNTHETIC expected_settlement_amount:EUR 100,000SYNTHETIC source-native status example:PENFSYNTHETIC source-native reason example:LACKSYNTHETIC partial confirmation:600 units settled; 400 units remainingSYNTHETIC cash evidence state:PARTIALSYNTHETIC security-leg evidence state:PARTIALSYNTHETIC external action:account-servicer enquiry recordedSYNTHETIC professional review required:trueSYNTHETIC open questions:cash evidence unresolved; position evidence unresolvedSYNTHETIC closure state:OPEN
The example does not reproduce a standard market workflow and does not represent a real institution, account, fund, customer or trade. The source-native code examples are illustrative and remain source-specific.
What this does not prove
A reviewer-facing exception record does not establish:
- one universal lifecycle or taxonomy for failed trades;
- ultimate root cause, negligence, blame or legal liability;
- legal finality from a technical settlement timestamp or status;
- the correct accounting entry, valuation treatment, cutoff decision or NAV impact;
- audit assurance, compliance closure or control effectiveness;
- that one settled leg resolves the other leg or the whole exception;
- that a cancellation or reversal erases the original event history;
- that the proposed record or closure test is an industry requirement;
- that Treasury Desk executes, repairs, cancels, reverses, settles or moves assets or cash.
So the practical question is narrower: can another reviewer reconstruct what was expected, what was observed, what was done, who owned it, and why the record was closed — without rebuilding the story from scratch?
What part of that story does your team most often have to reconstruct manually? Public or synthetic examples only.
Sources and notes
Reader-facing primary sources used for the source-native terms and evidence concepts above. Retrieved or re-opened on 2026-08-19.
- Swift — Standards Releases.
- ISO 20022 Registration Authority — Securities Settlement message definitions.
- ISO 15022 / Swift UHB — MT 548 Settlement Status and Processing Advice, status code field, and reason-code field.
- ISO 15022 / Swift UHB — MT 545 Receive Against Payment Confirmation, including date/time, quantity, amount, and partial-settlement indicator fields.
- ISO 15022 / Swift UHB — MT 535 Statement of Holdings and MT 536 Statement of Transactions.
- ISO 20022 Registration Authority — Bank-to-Customer Cash Management message catalogue.
Source-native code examples remain source-specific. The 2026 online MT User Handbook publication date and the Standards Release 2026 live date are distinct, as noted above.