Structuring a Reverse Stress Test Narrative Under PRA SS3/18

The purpose of Reverse Stress Testing is to identify the point at which your business model becomes unviable, and then work backwards to understand what combination of events could get you there. Firms that treat RST as a presentational exercise, softening scenarios to avoid difficult conversations, consistently receive the hardest supervisory feedback. The firms that land well do the opposite: they identify genuinely uncomfortable failure points, show the board has grappled with them, and evidence concrete management actions. The goal is not to look safe. It is to demonstrate you understand how you could fail and what you are doing about it.

Define business model failure precisely

SS3/18 requires you to identify the point at which the business model is no longer viable. This is not insolvency. It is the point at which the firm cannot continue to operate as it currently does: counterparties withdraw, funding evaporates, the regulator loses confidence, or the economics of core activities break down. Most weak RSTs confuse this with capital depletion. Capital is one measure of failure, not the only one.

Write the failure definition in plain terms. For a specialist lender, it may be the loss of wholesale funding access combined with a doubling of arrears. For an asset manager, it may be sustained net outflows crossing a threshold at which fee income no longer covers fixed costs. Name the specific condition. Vague definitions produce vague scenarios.

Work backwards to scenarios that actually cause failure

The analytical discipline of RST is reverse causation. Start at the failure point and ask: what would have had to happen? Build two or three scenarios that get you there, each with a different causal spine. Typical patterns include: a concentrated credit or counterparty event combined with a funding shock; a reputational event that triggers client attrition in a fee-dependent business; a technology or operational failure that impairs the firm's ability to transact.

What most firms get wrong: they take a plausible macro scenario, apply it with severity, and observe that capital remains above minimum. That is a forward stress test, not a reverse one. If no scenario in your RST actually breaks the firm, the exercise has failed its own purpose.

Calibrate severity honestly

Supervisors read a lot of these documents. They know when assumptions have been softened. If your scenario requires a 40% fall in collateral values to break the firm, say so, and compare it against historical precedent and peer disclosures. If the probability feels uncomfortably low, that is information, not a reason to redraft. A low-probability failure point is a legitimate finding. A high-probability one demands urgent management attention.

Link vulnerabilities to management actions

Each identified vulnerability needs a corresponding response: a change to limits, a diversification plan, a contingent funding arrangement, a strategic review of a concentrated exposure. The action should be specific, owned by a named executive, and tracked. This is where boards earn credibility. A narrative that identifies a vulnerability and does nothing about it invites supervisory follow-up. One that identifies the same vulnerability and attaches a dated remediation plan demonstrates governance maturity.

Show the board's engagement

Supervisors will look for evidence that the board set the scenarios, challenged management on severity, and used the output in strategic decisions. Minute the discussions. Record the counter-views. If a non-executive pushed back on an assumption, capture that. A board paper that reads as a management briefing with the board as audience is weaker than one that reads as a joint analytical exercise.

What good looks like

A strong RST narrative is uncomfortable to read. It names concentrations, dependencies, and assumptions that the firm would rather not have in writing. It is also unambiguous about what the firm is doing in response. Supervisors do not expect firms to be invulnerable. They expect firms to understand their vulnerabilities and to act on them.

Your next decision: review your most recent RST and ask whether any scenario in it actually breaks the firm. If not, the exercise needs rebuilding from the failure point backwards.

Frequently Asked Questions

How many scenarios should an RST contain?

Two to four well-constructed scenarios with distinct causal drivers are more useful than a long list of variants. Each should trace a different route to business model failure.

Should we disclose RST findings externally?

SS3/18 does not require external disclosure, and most firms do not publish RST outputs. Internal transparency with the board and the PRA is the expectation. Consider carefully before including RST specifics in Pillar 3 or investor materials.

How does RST interact with the ICAAP and recovery planning?

RST should inform both. The scenarios that break the firm are natural candidates for recovery plan triggers, and the capital implications feed the ICAAP's severe but plausible range. Inconsistency across these documents is a common supervisory finding.

What if the RST reveals a vulnerability we cannot easily fix?

Say so, and explain the mitigation path and timeline. Supervisors prefer honest acknowledgement of a structural issue with a credible plan over silence or understatement. The worst outcome is for the regulator to identify the vulnerability before you do.

How often should RST be refreshed?

Annually at minimum, with interim refreshes when the business model changes materially: a significant acquisition, a new product line, a change in funding structure, or a shift in the external environment that alters the causal pathways to failure.