Case Study 16.1: Cornerstone's ICAAP Under Scrutiny — The PRA's Deep Dive

Context

Cornerstone Financial Group is a fictional composite mid-sized UK bank with approximately £18 billion in total assets. It operates primarily in UK corporate lending, commercial real estate finance, and residential mortgages, with a smaller consumer banking division. It is rated a Category 2 firm by the PRA, which means it is subject to more intensive supervisory engagement than smaller banks but is not in the systemically important Category 1 tier.

The PRA's supervisory relationship with Cornerstone is managed by a dedicated supervisory team, and the annual cycle includes a formal review of Cornerstone's ICAAP submission. The year in question — the year of the ICAAP deep dive that this case study examines — was Cornerstone's third ICAAP under the post-CRD IV framework, submitted in Q4 of the relevant year.

Key people in this case: - Oliver Hartley: Cornerstone's Chief Risk Officer, responsible for overall ICAAP ownership - Fernanda Câmara: Head of Capital Adequacy, author of the ICAAP - James Ng: PRA Lead Supervisor for Cornerstone - Dr. Mei Xiu: PRA Senior Specialist, Stress Testing and Capital Analysis - Rafael Torres: Retained by Cornerstone as an external stress testing advisor six months before submission

Cornerstone's submitted CET1 ratio: 11.8% Cornerstone's Pillar 1 minimum: 4.5% Cornerstone's Combined Buffer Requirement (CCB + CCyB): 3.0% Cornerstone's Pillar 2A (previous cycle): 2.2% Cornerstone's internal Capital Planning Buffer (as submitted): 2.1% Total capital target (as submitted): 11.8% (and management considered this comfortable)

The Submission

Fernanda Câmara and her team spent six months producing the ICAAP. It was a 180-page document — professionally structured, thoroughly referenced, with appendices containing model documentation, board minutes from the two ICAAP governance sessions, and a schedule of management actions.

The stress test scenarios were: - Adverse: UK GDP -2.8% cumulative over 3 years; unemployment peak 7.5%; residential house prices -15%; commercial real estate -20% - Severely Adverse: UK GDP -5.2% cumulative; unemployment peak 9.8%; residential house prices -25%; commercial real estate -32% - Reverse Stress Test: UK GDP -14% over 2 years; unemployment 18%; residential house prices -55%

Under the Adverse scenario, the modeled CET1 fell from 11.8% to 9.3%. Under the Severely Adverse scenario, the modeled CET1 fell from 11.8% to 7.1%. The reverse stress test found that the scenarios above drove CET1 to 4.5%.

The ICAAP concluded that Cornerstone's Capital Planning Buffer of 2.1% was appropriate and that the institution was well-capitalized.

Oliver Hartley signed off the document. The board approved it at a 90-minute session in which two non-executive directors asked questions about the management actions section and the risk appetite framework.

The document was submitted to the PRA on the deadline date.

The PRA Deep Dive

Three weeks after submission, James Ng contacted Fernanda to schedule what he called a "capital adequacy deep dive" — a series of structured analytical sessions between Cornerstone's Capital Adequacy team and PRA specialists. This was not unusual; the PRA conducts such sessions annually for Category 2 firms. But the intensity of this cycle's engagement, Fernanda would later reflect, was a signal.

Dr. Mei Xiu led the PRA's analytical review. Her team had dedicated ten working days to reviewing Cornerstone's ICAAP before the sessions began. They brought three primary concerns to the table.

PRA Finding 1: Scenario Severity Insufficient for the Portfolio

Dr. Xiu's team had examined Cornerstone's portfolio composition in detail. Two concentrations stood out:

First, Cornerstone's commercial real estate book represented 27% of total lending — a concentration significantly above the UK banking sector average of approximately 15%. The PRA's analysis indicated that Cornerstone's CRE portfolio was skewed toward secondary office and retail assets — sectors that had already seen structural demand shifts (remote working, online retail) that made them more vulnerable to cyclical downturns than prime assets.

Second, Cornerstone had built a £2.1 billion leveraged lending book over the previous three years, providing term loans to private equity-backed companies. Approximately 60% of this book was to businesses with interest coverage ratios below 2.0x at origination — a metric the PRA flagged as indicating limited capacity to absorb income shocks in stress.

Dr. Xiu's challenge was direct: "Your severely adverse scenario specifies commercial real estate falls of -32%. Our analysis of UK CRE cycles — 1989–1993, 2007–2009 — shows peak-to-trough falls of 44–49% in secondary office and retail. For a portfolio concentrated in secondary assets, -32% is not a severely adverse scenario. It is closer to an adverse scenario calibrated to prime assets."

On the leveraged lending book, the PRA noted that the ICAAP's corporate loss model used a single corporate PD sensitivity applied uniformly across the corporate book — making no distinction between investment-grade corporates, sub-investment-grade leveraged companies, and mid-market companies. The leveraged lending book's higher sensitivity to earnings shocks was not captured.

Cornerstone's response: Fernanda acknowledged that the CRE scenario calibration had used aggregate UK commercial real estate indices rather than segment-specific indices for secondary office and retail. She committed to recalibrating using sub-sector specific data. Rafael Torres was asked to produce revised scenario parameters.

After recalibration, the Severely Adverse scenario CRE shock was revised to -43% (from -32%), and a separate leveraged lending module was built with PD sensitivities calibrated specifically to highly leveraged borrowers' income-shock vulnerability.

PRA Finding 2: The Reverse Stress Test Was a Box-Tick Exercise

Dr. Xiu was more pointed about the reverse stress test: "A GDP contraction of -14% over two years with unemployment at 18% is not a plausible scenario. The worst post-war UK annual GDP contraction was approximately -9.9% in 2020, driven by a once-in-a-century pandemic shock. A stress of that severity is not informative about your actual fragility because it is not a scenario that could realistically occur through normal macro-financial channels."

The PRA's issue was not that the scenario was too severe in absolute terms. It was that the scenario was mechanically severe — generated by running numbers backward from the capital floor through a linear model — rather than narratively constructed from a genuinely plausible failure mechanism specific to Cornerstone.

"What we want to understand," Dr. Xiu explained, "is what realistic scenario combination — what plausible sequence of events — could cause Cornerstone to become non-viable. Given your concentrations, one candidate is a severe dislocation in secondary commercial real estate, combined with a leveraged lending default wave, in conditions where you cannot raise new funding. Is that the scenario? Has the board had a genuine conversation about it?"

The board minutes from the ICAAP governance sessions showed no record of a substantive reverse stress test discussion. The two questions the non-executive directors had asked were about management actions, not about the reverse stress test results.

This finding was particularly uncomfortable. It exposed a gap not just in analytical process but in governance quality — the board had effectively rubber-stamped the reverse stress test without engaging with it.

Cornerstone's response: Rafael Torres led a redesigned reverse stress test process. The approach shifted from threshold mapping (backward-engineering from the capital floor) to narrative construction: the team identified Cornerstone-specific failure mechanisms first, then quantified them.

Three narratives were developed:

Narrative 1 — The CRE Correction: A 48% fall in secondary UK commercial real estate, driven by a combination of structural demand destruction (remote working becoming permanent) and a sharp interest rate rise (making leveraged CRE investment unviable). This scenario stressed Cornerstone's CRE book severely — but was explicitly linked to observable pre-conditions (vacancy rate increases, transaction volume collapse) that would give early warning.

Narrative 2 — The Leveraged Lending Cascade: A recession in which private equity sponsors failed to support portfolio companies, triggering defaults cascading through Cornerstone's leveraged lending book in a concentrated 18-month window. The narrative specified that the catalyst was a sharp rate rise (increasing borrowing costs for already-stretched borrowers) followed by a consumer spending shock (reducing revenues for consumer-facing PE portfolio companies).

Narrative 3 — The Funding Freeze: An idiosyncratic reputational shock (a specific hypothetical — a large unexpected credit impairment announcement in the CRE book) triggering deposit outflows and loss of wholesale funding access, combining with market-wide risk aversion in a period of elevated geopolitical uncertainty. This narrative was linked to ILAAP findings rather than ICAAP capital floors alone.

Each narrative was presented to the board in a dedicated 45-minute session. Non-executive directors were asked specifically: "Do you recognize this as a plausible failure mechanism for this institution? What early warning indicators would you want management to monitor?" The session was minuted in detail.

PRA Finding 3: Management Actions Were Not Credibly Documented

The third finding was the most operationally consequential. Cornerstone's ICAAP assumed four management actions under the Severely Adverse scenario:

  1. Suspension of the ordinary dividend for three years (saving approximately £85m CET1 per year)
  2. Reduction of new lending by 30% in years 2 and 3 of the stress (reducing RWA and credit loss accretion)
  3. Sale of the consumer banking division within 18 months of the stress onset (assumed net proceeds: £250m CET1)
  4. Issuance of £400m of new CET1 capital from existing shareholders and new institutional investors within 12 months of the stress onset

The PRA's analysis of each action:

Dividend suspension (£85m/year): Accepted in principle. The board had resolution authority. However, the ICAAP did not distinguish between ordinary share dividends and Additional Tier 1 coupon payments. AT1 coupon cancellation requires separate regulatory assessment and cannot simply be assumed. The PRA required this to be split and the AT1 treatment documented separately.

Lending reduction (RWA compression): Accepted in principle. Operationally credible — the institution can choose to reduce new lending origination. However, the assumed RWA reduction pace assumed loan redemptions at their contractual maturity with no refinancing, which overstated the RWA reduction speed (some loans would roll over). Revised assumptions required.

Consumer banking division sale (£250m proceeds): Challenged on multiple grounds. The assumed proceeds were based on a multiple of earnings applicable in normal market conditions. In a severely adverse stress, financial services M&A markets contract sharply and sale proceeds for sub-scale consumer banking businesses would likely be materially lower — particularly if the buyer is aware the seller is distressed. Additionally, 18 months was considered optimistic for a full regulatory approval and sale completion process. Revised: sale proceeds reduced to £130m under stressed market conditions; completion timeline extended to 24–30 months.

New equity issuance (£400m): The PRA's position was categorical. New equity issuance as a management action under a severely adverse scenario is not credible without documented evidence of commitments from named investors who have been approached and have indicated support. "General availability of new capital from institutional markets" is not credible under systemic stress — equity markets for financial sector issuance typically close precisely during the conditions described by a severely adverse scenario. This action was rejected entirely unless Cornerstone could provide documented investor commitments. No such commitments were produced. The action was removed.

The Revised ICAAP: An £85 Million Capital Buffer Increase

The combined effect of the three revisions — more severe scenario, genuine reverse stress test, credible management actions — was material.

Under the revised Severely Adverse scenario (CRE shock -43%, leveraged lending module added, equity issuance management action removed):

  • The stressed CET1 ratio under Severely Adverse fell from 7.1% to 5.4% — still above the 4.5% regulatory minimum but with a substantially reduced buffer
  • The Capital Planning Buffer required to ensure CET1 remains above 4.5% throughout the stress horizon, given the more severe loss path, was assessed at 3.3% rather than the originally submitted 2.1%
  • This 1.2 percentage point increase in the Capital Planning Buffer, applied to Cornerstone's £7.1 billion RWA, translated to approximately £85 million in additional required capital

Cornerstone's Total Capital Requirement (Pillar 2A plus buffers plus Capital Planning Buffer) was revised upward by approximately £85 million compared to the submitted ICAAP.

Aftermath and Lessons

Oliver Hartley reflected on the process at Cornerstone's annual Risk Committee offsite six months later. He framed three lessons that he asked the team to carry forward:

Lesson 1: The scenario is the output of a genuine question, not the input to a compliance process.

The original ICAAP scenarios were designed around the question: "What numbers do we need to put in to show our capital is adequate?" The right question is: "What do we genuinely believe could go wrong, given who we are and what we do?" For Cornerstone, that question produced scenarios centered on CRE and leveraged lending — which is where the genuine risk lived. The PRA's challenge moved Cornerstone from the first question to the second, and the resulting ICAAP was more useful as an internal management tool, not just more compliant as a regulatory submission.

Lesson 2: Management actions need to be battle-tested, not assumed.

The instinct to include management actions is correct — banks do take actions in response to stress. But those actions need to be credible under the conditions of the stress, not under normal market conditions. The rejected equity issuance assumption was not dishonest; it reflected a genuine board belief that the bank could access markets. The PRA's challenge forced a harder question: under the conditions described by the scenario you've specified, what actions are genuinely available? That is the right question.

Lesson 3: Board engagement cannot be delegated.

The gap in the reverse stress test governance was ultimately a board engagement gap. Two talented non-executives had attended two ICAAP governance sessions and had not been given the opportunity to engage substantively with the reverse stress test. The redesigned process — dedicated board sessions on each reverse stress test narrative — was more demanding of board time. But it produced a board that could genuinely speak to the question: "What could cause Cornerstone to fail?" That capability, the PRA's team made clear, is the minimum expectation.

Rafael Torres, who had advised on the revised ICAAP, offered his own observation in a conversation with Maya Osei at a risk management conference months later:

"The PRA's findings were not surprising to me. I've seen this pattern in a dozen ICAAPs. The document gets written by the people who know the numbers, optimized to show adequacy, reviewed by governance at a level of abstraction that prevents real challenge, and submitted. The PRA's job is to break the pattern. What surprised me at Cornerstone was how quickly the team adapted once the pattern was named. Once Oliver reframed the question — what would actually break us? — the team produced genuinely good analysis in three months. The capability was there. It just needed the right question."

Maya filed this observation away. Verdant Bank's first ICAAP submission was eighteen months away.

Discussion Questions

  1. The PRA rejected Cornerstone's assumption of £400m in new equity issuance as a management action. Under what circumstances, if any, would a capital raise be an acceptable management action in an ICAAP stress test? What documentation would be required?

  2. Cornerstone's original CRE scenario shock of -32% was described by the PRA as "closer to an adverse scenario for prime assets." What process should Cornerstone have used to calibrate its scenario severity before submission — and what external data sources are available for UK CRE sub-sector price indices?

  3. The three reverse stress test narratives developed after PRA challenge were specifically designed around Cornerstone's concentrations. Why does the PRA view institution-specific narratives as superior to generic macro scenarios for reverse stress testing? What does this imply about the appropriate process for designing reverse stress tests?

  4. The revised ICAAP increased Cornerstone's Capital Planning Buffer by £85 million. Who bears the economic cost of this increase? How does this illustrate the relationship between the quality of internal stress testing governance and the actual capital position of the institution?

  5. Oliver Hartley's Lesson 3 — "Board engagement cannot be delegated" — has implications beyond stress testing. What aspects of board responsibility for risk management does this case illustrate? What structural changes (committee structure, board expertise requirements, dedicated session time) might a bank make to ensure genuine board engagement with stress testing?