The Submitter and the Swaps Trader
You are the compliance officer responsible for a bank's submissions to a major interest-rate benchmark. The bank is one of 12 contributors. On Wednesday a senior swaps trader emails the submitter: "Your 3-month has been 1bp high all week - we're paying fixed in size, take a look." The submitter's written procedure requires submissions to be based on a hierarchy: executed transactions, then committed quotes, then expert judgement, and requires that judgement to be documented. A review finds the submitter's spreadsheet excludes a €150 million deposit transaction executed at 09:05 because it was booked late, and that the submission was 0.8bp above the transaction-based rate on three of the last five days. The trader's swap position is €400 million notional paying fixed. The benchmark is designated as critical under the applicable benchmark regulation.
1. Characterise the trader's email and set out the immediate compliance actions.
Read the worked answer
The email is an attempt by a person with a direct financial interest to influence a benchmark submission. The trader is paying fixed on €400m, so a lower fixing benefits his position; the phrase "take a look" is a request to move the submission in that direction, not a neutral market comment. Under benchmark regulation and market-abuse rules, influencing or attempting to influence a benchmark can constitute manipulation even where the submitter does not comply, and the submitter's independence is compromised by receiving the request.
Immediate actions: escalate to the MLRO and the benchmark oversight committee; preserve the email, the submitter's records and the trader's position and trade history; restrict the trader's ability to contact the submitter and review whether his positions were established or changed around submission dates; and consider whether a suspicious transaction and order report should be made. The submitter should be interviewed separately and reminded of the duty to report attempts at influence. The bank should also check whether the trader has contacted other contributors, since collusive influence would be more serious than a single email.
2. Analyse the submission error against the hierarchy, identify the control failures, and say what should be done about past submissions.
Read the worked answer
The €150m deposit executed at 09:05 is an executed transaction and sits at the top of the hierarchy. It should have been included; excluding it pushed the submission about 0.8bp above the transaction-based rate. If the submitter chose to override the transaction data using expert judgement, the procedure required that judgement to be documented, and there is no evidence it was.
The control failures are: no reconciliation of late-booked transactions before the cut-off; no four-eyes review of the spreadsheet; no automated link between the transaction system and the submission; and no challenge when the submission diverged from the transaction-based rate on three of five days. Controls should include a pre-submission reconciliation, a documented exception log, independent review and retention of all inputs.
On past submissions, the bank should assess materiality, correct any error that the administrator's rules permit, and notify the benchmark administrator and, where required, the regulator. It should not quietly change historical records; the audit trail of the error and the correction is itself evidence of a functioning control environment.
3. What governance changes would you recommend for the submission function, and how would you test whether they work?
Read the worked answer
The submission function should be independent of the trading desk: it should report through operations or risk, not through the swaps business, and the submitter's remuneration and appraisal should not depend on trading results. There should be a conflicts register covering the submitter's own positions and those of close colleagues, a prohibition on trading desks contacting the submitter about submissions, and a clear escalation route for attempts at influence.
Testing should be evidence-based rather than policy-based: sample submissions against transaction data and check the documented rationale; reconcile late bookings to the submission file; review communications between the desk and the submitter for attempts to influence; and test whether the four-eyes review actually challenges outliers. The bank should also run a scenario in which a senior trader makes a request like this one and confirm that the submitter reports it. The oversight committee should receive the results quarterly, with actions tracked to closure, and the board should be told that a critical benchmark is involved, since the reputational and regulatory consequences extend well beyond the trading desk.