How to Structure a Recovery Plan Under PRA SS9/17 That Evidences Credible Options

A Recovery Plan is read by supervisors as a test of two things at once: whether the firm has credible options to restore financial strength in severe stress, and whether management genuinely understands its own fragility. Get the first right and miss the second, and the plan reads as marketing. Get both right, and you have a document that strengthens the supervisory relationship rather than inviting a Section 166.

Start with the stress, not the menu

The weakest Recovery Plans begin with a catalogue of options and reverse-engineer scenarios to justify them. Strong plans do the opposite. Build the scenarios first, idiosyncratic, market-wide, and combined, with sufficient severity to approach non-viability. The PRA expects scenarios that genuinely threaten the firm. If your reverse stress test lands comfortably above the point of non-viability, the scenarios are too mild and the whole plan loses credibility.

For each scenario, model the trajectory of capital, liquidity, and profitability quarter by quarter. This is what lets you show when each recovery indicator would breach and which options remain available at that point. Options that look fine in isolation often become unexecutable mid-stress: a rights issue in a market-wide crisis, a business line sale when buyers have vanished, a dividend suspension when it has already been cut.

Calibrate indicators to trigger action, not reassurance

Indicators set too close to regulatory minima are a red flag. By the time they breach, the firm has lost optionality. Good calibration sits meaningfully above minimum requirements and above the firm's own risk appetite, giving the board real time to decide. Include capital, liquidity, market, profitability, and macroeconomic indicators, and explain the quantitative link between each indicator level and the recovery options it unlocks.

The governance around indicators matters as much as the levels. Evidence how breaches are detected, who is notified within what timeframe, what the escalation path to the board looks like, and how the decision to invoke recovery is made. Supervisors will probe this in meetings. Vague wording here undermines the entire plan.

Make each option genuinely executable

For every recovery option, document: the quantitative impact on capital or liquidity, the time to execute from decision to effect, the operational steps required, the counterparties or approvals needed, the scenarios in which it remains viable, and the dependencies on other options. A capital raise that assumes functioning primary markets must say so. An asset sale must identify the asset class, likely discount, and tax implications.

What most firms get wrong: listing twenty options to appear well-prepared. The PRA would rather see six options that are properly worked through, with playbooks, than a long list of unquantified possibilities. Include dry-run evidence where you have it, board rehearsals, management exercises, pre-drafted communications.

Tie the plan to the wider recovery and resolution architecture

SS9/17 does not sit alone. The Recovery Plan must align with the ICAAP's stress testing, the ILAAP's liquidity stresses, the Resolution Pack, and the Solvent Wind-Down analysis. Inconsistent assumptions, different stress severities, different asset valuations, different timeframes, invite immediate challenge. Run a reconciliation across these documents before submission and document the deliberate differences.

Governance evidence the board actually owns it

The board must demonstrably own the plan. That means minuted challenge, not rubber-stamping. Include evidence of board scenarios workshops, challenge on indicator calibration, and debate on option feasibility. A plan approved without visible challenge reads as a management document the board has not internalised.

Next step

Before your next submission cycle, commission an internal pre-read against the most recent PRA Recovery Plan thematic feedback. Identify the three weakest options and either strengthen them with executable playbooks or remove them. A shorter, more credible plan outperforms a longer, aspirational one every time.

Frequently Asked Questions

How many recovery options should the plan contain?

There is no fixed number, but quality beats quantity. Most credible plans contain between six and twelve material options, each quantified and scenario-tested. Firms submitting twenty or more unquantified options typically attract more supervisory challenge, not less.

How severe should the recovery scenarios be?

Severe enough to breach recovery indicators meaningfully and approach, without reaching, the point of non-viability. If your scenarios do not stress the firm to the point where recovery options are genuinely needed, they will be rejected as insufficient.

Should the plan reference resolution?

Yes. The Recovery Plan should make clear where recovery ends and resolution begins, and how the firm would transition if recovery options failed. This boundary is a frequent area of supervisory questioning.

How often should playbooks be refreshed?

At least annually, and after any material change to the business, capital structure, or market conditions. Playbooks that reference outdated entities, personnel, or market assumptions undermine the credibility of the whole plan.

What triggers the sharpest supervisory concern?

Three things consistently: options that are not executable in the stress they are meant to address, indicators calibrated too close to regulatory minima, and inconsistency with the ICAAP or ILAAP. Fix these before worrying about anything else.