Case Study 13.1: Maya's Regulatory Reporting Crisis — Verdant Bank's XBRL Failure

The Situation

Organization: Verdant Bank (fictional UK challenger bank) Maya's challenge: A XBRL taxonomy update failure that invalidated Verdant's PRA regulatory return submission Timeline: Q1 2024 Regulatory backdrop: PRA/Bank of England reporting requirements; imminent new Bank of England data standards

Background

By Q1 2024, Verdant Bank had grown significantly — 340,000 customers, £1.8 billion in assets. The regulatory reporting function had not kept pace. Verdant's regulatory reporting process:

  • Core banking balance sheet extracted manually via CSV from Finacle
  • Capital calculations performed in Excel by one analyst (Kofi)
  • XBRL conversion handled by an external XBRL bureau service — a small specialist firm hired in 2021
  • Submission via the Bank of England's BEEDS portal

The process had worked — barely. Q3 2023's submission had been filed with 6 hours to spare. The team knew it was fragile.

The Failure

On February 14, 2024, Kofi submitted Verdant's Q4 2023 PRA return to the BEEDS portal. The portal accepted the file. Two days later, Maya received a message from the BEEDS portal: submission rejected — taxonomy validation failure.

The rejection notice: the instance document referenced taxonomy version pra-taxonomy-2022-03-31 — an outdated version. The Bank of England had published taxonomy version pra-taxonomy-2023-09-30 in September 2023, effective for Q4 2023 reporting. The external XBRL bureau had not updated its taxonomy configuration for Verdant's template set.

The filing deadline for the PRA return was February 20, 2024 — six days away.

The Cascade of Problems

Maya assembled an emergency response team: Kofi, the CTO, and the external XBRL bureau. What they discovered was worse than expected:

Problem 1: Taxonomy update not communicated. The Bank of England had notified regulated firms of the taxonomy update in September 2023 — through a notice on its website. Verdant had no process for monitoring Bank of England website notices. The XBRL bureau had received the notification but had assumed Verdant would inform them if updates were needed. Neither party had acted.

Problem 2: New template structure. The 2023 taxonomy update had added a new section to the capital adequacy return — capturing data on the bank's exposure to crypto-assets (a new PRA requirement under PS7/23). Verdant had no crypto-asset exposures, but the template section still needed to be completed (with zeros where applicable). The 2022 template did not include this section, so the prior-year data was structurally incompatible.

Problem 3: Calculation mapping errors discovered during rework. As the team worked to regenerate the submission under the 2023 taxonomy, they discovered that three data points in the capital calculation had been mapped to the wrong XBRL concepts. These errors had existed in prior submissions — the wrong data had been submitted correctly formatted. The values were immaterial (all three errors totaled a £280,000 reclassification within the capital deductions schedule), but they represented an accuracy failure.

Problem 4: Excel formula error. While reviewing the capital calculation workbook under time pressure, Kofi identified an Excel error — a cell reference had broken during a prior copy-paste operation. The formula that should have summed deferred tax asset (DTA) deductions from cells D45:D52 was instead referencing D45:D51 — omitting one row. The DTA deduction was understated by £340,000, slightly overstating CET1.

The Response

Maya convened a daily crisis meeting. The immediate priority: file a corrected submission before the February 20 deadline.

Day 1–2 (February 14–15): Taxonomy update. The XBRL bureau updated its template configuration to the 2023 taxonomy. The new crypto-asset section was added with zero values (confirmed by the CFO that Verdant held no crypto-assets).

Day 3 (February 16): Excel audit. Kofi audited every formula in the capital calculation workbook. The DTA formula error was found and corrected. Three other formula references were identified as potentially fragile (relative references in merged cells) — noted for remediation.

Day 4 (February 17): Reclassification of mapping errors. The three XBRL concept mapping errors were corrected — the £280,000 reclassification was confirmed as immaterial by the CFO and documented in the submission notes.

Day 5 (February 18): Internal sign-off. Maya reviewed the corrected submission. The CRO signed off on the capital calculation. Submitted 10:42 AM.

Day 6 (February 19): Portal confirmation — submission accepted. Filed one day before the February 20 deadline.

The Regulator's Response

Three weeks later, Maya received a follow-up from the PRA Supervision team. The supervisor noted:

  1. The initial taxonomy failure had been logged — this was Verdant's first submission failure in three years of reporting, noted as a one-time event.
  2. The corrected submission had been filed on time — no deadline breach.
  3. The three XBRL concept mapping errors (prior submissions) were noted. Given immateriality, no restatement was required. However, the PRA expected Verdant to document a root cause analysis and implement controls to prevent recurrence.
  4. A recommendation (not enforcement): Verdant should implement a taxonomy monitoring process to ensure awareness of annual updates.

Maya's Remediation Program

Following the crisis, Maya implemented a six-month remediation:

Taxonomy monitoring: A calendar alert for Bank of England and EBA taxonomy publication dates. The external XBRL bureau was contracted to provide a formal notification service — they would alert Verdant of any taxonomy update within 5 business days of publication, with an impact assessment.

Submission verification process: A checklist introduced — before any submission: (1) confirm taxonomy version is current; (2) confirm template completeness against the current template specification; (3) confirm all concepts map to the correct XBRL element (a mapping inventory maintained by Kofi).

Capital calculation spreadsheet remediation: The Excel workbook was reformatted — all input data moved to a separate "inputs" tab; all calculations moved to a "calculations" tab with named ranges replacing all cell references; a separate "review" tab listing every material formula for auditor review. An independent analyst (not Kofi) reviewed all 47 calculation formulas.

Longer-term: Maya began scoping a move to a specialist regulatory reporting platform — one of the key findings from the crisis was that Verdant's scale had outgrown a manual Excel + external bureau approach. An RFP was issued to three vendors (Wolters Kluwer FRR, Vermeg, and a boutique UK specialist) in Q3 2024.

Discussion Questions

1. The taxonomy update failure resulted from a communication gap — neither Verdant nor the XBRL bureau monitored Bank of England taxonomy publications. Under BCBS 239 Principle 1 (governance), where does accountability for this failure sit? At the board? Senior management? The reporting analyst? The external vendor?

2. The three XBRL concept mapping errors had existed in prior submissions. The PRA concluded these were immaterial and required no restatement. At what level of materiality would a prior-period restatement be required, and what would that process look like? Who — internally and externally — would be involved?

3. Maya's remediation included converting cell references in Excel formulas to named ranges. Explain how named ranges reduce formula error risk compared to cell references like "D45:D51." What residual risks remain in the Excel-based approach even after this improvement?

4. The external XBRL bureau was hired to reduce internal complexity — Verdant outsourced XBRL generation to specialists. But the taxonomy failure was a failure of that outsourced service. How should Verdant structure its third-party risk management for an outsourced regulatory reporting service? What contractual provisions and monitoring controls should be in place?

5. Maya's longer-term response was to scope a specialist regulatory reporting platform. What criteria should the RFP evaluation use to compare the three vendors? Beyond functionality, what factors are particularly important for a challenger bank of Verdant's size?