Structuring an ICAAP Under SS31/15 That Evidences Robust Stress Testing

The quickest way to attract Pillar 2A add-ons is to produce an ICAAP where stress testing reads as a bolt-on exercise rather than a live management tool. The PRA's concern, consistently signalled through SS31/15 and subsequent supervisory communication, is not the size of the capital number but the quality of the thinking behind it. If the stress testing section cannot be traced to real board decisions, calibrated to the firm's actual vulnerabilities, and reconciled to the capital stack, supervisors will fill the gap themselves.

Start With the Risk Identification Spine

The ICAAP should open with a risk identification section that is specific to the firm's business model, not a generic taxonomy. SS31/15 expects the firm to demonstrate that it has considered all material risks, including those not fully captured under Pillar 1. The common failure is listing risks without evidencing why each has been sized the way it has.

For each material risk, show three things: the exposure measurement, the historical and forward looking evidence base, and the link to a specific stress scenario or capital charge. If a risk is identified but not stressed or capitalised, explain why, with reference to mitigants that are themselves evidenced.

Design Scenarios That Reflect the Firm, Not the Benchmark

Scenarios borrowed wholesale from the annual cyclical scenario or anchored only to the PRA's published stresses will read as unimaginative. The PRA wants to see reverse stress testing that identifies the firm's actual breaking points, and idiosyncratic scenarios that reflect concentrations, business model dependencies, and management actions that would plausibly be taken.

Good practice: run at least one severe but plausible macro scenario calibrated to the firm's geography and portfolio, one idiosyncratic scenario targeting the largest concentration or dependency, and a combined scenario. Document the narrative logic, the transmission mechanisms into P&L, RWAs, and liquidity, and the time horizon over which each shock plays out.

Make Management Actions Credible

This is where ICAAPs most often lose supervisory confidence. Management actions that assume orderly deleveraging, dividend suspension on day one, or cost cuts that have never been executed at the stated scale will be challenged. Each action should carry an owner, a trigger, a quantified effect, and evidence of feasibility, ideally cross referenced to the recovery plan.

If management actions reduce the capital requirement materially, expect supervisors to probe the counterfactual. Show the capital path both with and without actions.

Connect Stress Outputs to the Capital Stack

The reconciliation from stress testing outputs to Pillar 2A sizing is where many documents fall apart. Set out, explicitly, which risks are covered by Pillar 1, where Pillar 2A is intended to top up for underestimation or uncovered risks, and how the stress results inform the Pillar 2B buffer. Avoid double counting, and be explicit where overlaps exist.

Where the firm's internal assessment differs from the PRA methodology for a given risk, do not obscure the difference. Set out the methodology, the quantitative result, and the rationale.

Evidence Genuine Board Ownership

SS31/15 is clear that the ICAAP is a board document. Minutes, challenge logs, and iterative drafts should show substantive engagement, not a single approval meeting. The board's risk appetite should be the anchor for scenario severity, and the ICAAP should show the board accepting, modifying, or rejecting specific assumptions.

A board that has genuinely engaged will have asked uncomfortable questions. Capture them.

What Good Looks Like

A strong ICAAP under SS31/15 reads as a coherent argument: here is our business model, here are the risks it generates, here is how we have stressed them, here is what we would do, and here is the capital we hold as a result. Each link in that chain is evidenced. Weaknesses are acknowledged and being addressed on a stated timeline.

The next action: before submission, have an independent reader, ideally someone with supervisory experience, trace the argument end to end and flag where it relies on assertion rather than evidence. That is where add-ons originate.

Frequently Asked Questions

How severe should idiosyncratic scenarios be?

Severe enough to challenge the firm's capital position meaningfully. If the scenario produces a comfortable surplus, either the scenario is too mild or the firm's analysis has not captured second order effects. Reverse stress testing should identify the point of non viability and work back from there.

How much weight should we give to management actions?

More weight is defensible when actions are pre agreed, quantified, operationally tested, and consistent with the recovery plan. Less weight when they rely on market access under stress, counterparty behaviour, or execution at a scale the firm has never achieved.

Should we disclose model limitations openly?

Yes. Supervisors already form a view of model weaknesses from peer comparison and thematic work. A firm that identifies its own limitations and sets out a remediation plan is treated more credibly than one that presents models as more robust than they are.

How often should the ICAAP be refreshed between annual cycles?

Whenever there is a material change in risk profile, strategy, or external environment. A document that has not been revisited after a significant acquisition, portfolio shift, or macro event will struggle under supervisory review, regardless of its annual cycle date.