Regulatory Report Quality Score Contradiction
At Meridian Bank, the Liquidity Coverage Ratio (LCR) report is assembled monthly from three source systems. The data governance dashboard shows 98.2% quality across 46 controls, yet the regulator's last examination found 12 errors, including a $41 million misclassification of high-quality liquid assets. Post-mortem shows the controls run against the reporting layer after manual adjustments; 4.1% of report lines are manually overridden each month. A vendor upgrade to the treasury system changed two field definitions nine months ago, and no one updated the business glossary. Three stewards share responsibility for LCR data: one in treasury, one in risk, one in finance. Each says remediation belongs to another group. The next regulatory submission is in six weeks, and the steering committee wants a remediation plan with a fixed $250,000 budget.
1. Diagnose the governance failure that allowed a 98.2% quality score to coexist with 12 regulatory errors. Identify which stewardship accountability was missing, not which system failed.
Read the worked answer
The failure is control placement and stewardship accountability, not system reliability. The 98.2% score measures the reporting layer after manual adjustments, so it certifies the output the team already touched rather than the source data. The 12 errors-including the $41 million misclassification-entered upstream or through overrides. Missing accountability: no steward owned the two treasury field definitions changed by the vendor upgrade, and no steward owned the manual override log. Three stewards sharing LCR responsibility without decision rights created diffusion. The quality dashboard should have reported source-system control results and override rates separately. The governance failure is that quality was measured where it could be polished, not where it was created, and no one was accountable for definition change management.
2. Allocate the $250,000 budget across no more than three remediation actions. Justify the order and state what you deliberately defer.
Read the worked answer
Allocate: $120K to source-system controls for the two changed treasury fields, including automated reconciliation and a definition-change log; $80K to override analytics and removal-classify the 4.1% manual adjustments, eliminate the top recurring causes, and require steward approval for new overrides; $50K to stewardship operating cadence-a single LCR data owner, RACI, and monthly quality review with risk and finance. Defer enterprise-wide glossary tooling and a new dashboard; those are lower-yield in six weeks. Reasoning: the regulator's errors came from definitions and overrides, so fix those first. The override analytics is the fastest way to find the $41 million class of error. Do not spend on more reporting-layer controls-that repeats the existing mistake.
3. Define decision rights for the two changed treasury field definitions and explain how you would prevent a repeat at the next vendor upgrade.
Read the worked answer
Decision rights: the treasury data owner decides which definition is authoritative for LCR; the risk steward approves regulatory interpretation; finance owns report-level mapping; IT implements but does not decide. For a vendor upgrade, require a change-impact assessment before deployment: vendor release notes mapped to business glossary terms, steward sign-off on changed fields, and a regression test against regulatory report lines. Add a clause to the vendor contract requiring 60-day notice of field-definition changes. The CDO should also add a 'definition change' control to the DQ framework, so a field change cannot pass without a named steward's approval and a glossary update. This prevents the nine-month silent drift that occurred here.