System-wide stress test Pilot exercise ("system-wide stress test", SWST)

Why conduct an exercise covering the entire financial system?

Since the strengthening of prudential frameworks after 2008, the solvency, liquidity and resilience of the banking, insurance and asset-management sectors have improved considerably — resilience regularly confirmed by sector-specific tests. But the microprudential robustness of each sector does not guarantee the stability of the whole system: institutions that are individually sound may collectively generate or amplify destabilising dynamics that tests focused on a single sector cannot detect.

Two developments call for a more integrated approach. First, the global growth of non-bank financial intermediation (NBFI) has made interconnections more complex. Second, several recent episodes (the "dash for cash" in March 2020, Archegos in March 2021, the 2021-2022 energy crisis, the "gilts" crisis of September 2022) have shown that systemic liquidity stress must be addressed in its own right, and not merely as a symptom of a solvency difficulty. The exercise therefore adopts a "bottom-up" approach, at the system level, placing liquidity explicitly at the heart of the analysis. The Financial Stability Board (FSB) recommends such exercises; the French authorities have drawn methodological lessons from the "System-Wide Exploratory Scenario" conducted by the Bank of England in 2023-2024.

Design, scope and governance

The exercise brings together twenty-five voluntary institutions — five banking groups (including all G-SIBs established in France), nine insurance groups, ten asset-management companies and one clearing house — ensuring broad and representative coverage (non-participating players are taken into account in the stress test exercise on the basis of a series of models capturing their characteristics and behaviours).

It is co-designed with market participants and articulated with existing reportings. The balance sheet is run on a static basis over ten days on a "best-effort" basis, following an iterative two-round process.

The scenario is built on an agnostic narrative with a plausible level of severity, though significantly more severe than usual tests (a 1-in-500-year event over 1 year, versus 1 in 200), calibrated with historical quantile — i.e. the worst fortnight observed roughly every twenty years — with the worst shock in  day 3 “D3”.

Figures 3&4: Scenario : an agnostic narration with a plausible level of severity but significantly amplified with respect to usual stress test exercises (1 – 500 vs 1-200)

Three transmission channels assessed

The analysis is structured around three channels through which a local shock propagates and generates second-round effects.

Next steps and inclusion in the international agenda

The results of the first round are currently being consolidated and analysed. A second round begun in June 2026: it combines a qualitative questionnaire (conditions for the sales of Money-Market Funds - MMFs, credibility of re-hedging strategies, liquidity management) with adjusted constraints reflecting the observed aggregated selling pressure and market depth, in order to test system-wide convergence.

Reconciliation work is being carried out jointly with the AMF; a final summary report is expected by the end of 2026.

Beyond the French case, the exercise is part of the macroprudential NBFI agenda (FSB, G7 report on system-wide testing, Eurosystem reflections), where this type of approach is recognised as a promising tool for analysing contagion and vulnerabilities linked to interconnections.