How to Prepare a Recovery Plan That Withstands PRA Scrutiny

A Recovery Plan is read by the PRA as evidence of whether your board actually understands the firm's vulnerabilities and could act decisively under stress. Most plans fail that test not because the format is wrong, but because they read as compliance documents rather than operational playbooks. The fix is substantive, not cosmetic.

Start with the firm, not the template

The weakest plans begin with the PRA's expectations and work backwards. The strongest begin with an honest internal view of how this specific firm could get into serious trouble, what would be visible first, and what the board would realistically do. Only then is that mapped against the Supervisory Statement (SS9/17) and the Rulebook.

If your plan could, with minor edits, belong to a different firm of similar size, it is not yet fit for purpose.

Get the scenarios right

The PRA expects a range of scenarios: idiosyncratic, market-wide, and combined; fast and slow moving; near-default severity. Common failings:

• Scenarios calibrated to be survivable rather than genuinely severe. The point of near-default is to test whether options still work when the firm is close to the edge.
• Scenarios that do not actually stress the firm's real vulnerabilities. A retail bank whose scenarios ignore deposit flight dynamics post-2023 will be challenged.
• Reverse stress tests that are disconnected from the main scenario suite, treated as a separate exercise.

Each scenario should produce a clear narrative: what breaks, in what order, how fast, and which recovery options are available at each point.

Make the indicators work

Recovery indicators should trigger early enough to give management genuine optionality. Three questions to pressure-test them:

1. In each scenario, when do indicators actually breach, and is that early enough to execute the options you are relying on?
2. Are thresholds calibrated against the firm's own risk appetite and ICAAP/ILAAP, or borrowed from peer benchmarks without justification?
3. Is there a clear, pre-agreed escalation path from indicator breach to board decision, with named roles and timeframes?

Indicators that only breach once the firm is already in resolution territory are a red flag the PRA will raise.

Build options the firm could actually execute

For each recovery option, you need: a credible impact estimate, execution timeline, legal and operational dependencies, and an honest assessment of what degrades the option under stress. Asset sales that assume orderly market conditions in a market-wide scenario will not pass scrutiny. Capital raises priced off current conditions will not either.

What good looks like: option-by-option playbooks with named owners, decision rights, external advisers pre-identified, communications drafts, and a clear view of which options are mutually exclusive or sequencing-dependent.

Governance is the test the PRA actually applies

The PRA's reviewers can tell within an hour whether the board owns the plan. Signs they look for:

• Board minutes showing genuine challenge during approval, not nodding-through.
• A CEO who can speak to the plan's logic without reaching for the document.
• Evidence of a recent fire drill or playbook test, with lessons incorporated.
• Integration with ICAAP, ILAAP, ORSA (for insurers), and the wind-down plan, with inconsistencies resolved rather than papered over.

If the plan sits with risk or finance and the board sees it annually, you have a governance problem, not a drafting problem.

What to do before the next submission

Run a tabletop exercise against your most severe scenario with the actual executive team who would run it. Note every point where the plan does not give them what they need. Fix those gaps before you submit. That single exercise will tell you more about plan quality than any external review.

Frequently Asked Questions

How often should the plan be fully refreshed versus updated?

Material change in business model, risk profile, or group structure triggers a full refresh. Annual updates should still involve genuine re-testing of scenarios and indicators against current conditions, not a date change and sign-off.

How detailed should recovery options be?

Detailed enough that a competent executive who was not involved in writing the plan could pick it up and begin execution. If the option says "sell non-core portfolio," that is not an option, it is an aspiration.

What is the right role for external advisers?

Useful for benchmarking, scenario calibration, and challenge. Dangerous if they write the plan. Ownership must sit inside the firm; a plan the executive team did not build, they cannot execute.

How does the Recovery Plan relate to the wind-down plan and resolution pack?

Recovery assumes the firm continues as a going concern; wind-down and resolution assume it does not. The three should share a consistent view of the firm's vulnerabilities and critical functions. Inconsistencies between them are a frequent supervisory finding.