
Field Notes
Geopolitical Stress Test Results: What the ECB Asked 110 Bank CROs to Prove
TL;DR: The ECB's 2026 geopolitical stress test reversed the usual method: 110 banks started from a CET1 loss of at least 300 basis points and had to name the geopolitical events that would cause it. Most produced meaningful scenarios, but several underplayed liquidity effects and many relied on recovery actions that other banks planned to use at the same time.
Key takeaways:
- The ECB announced the exercise on 12 December 2025 and published aggregate results on 31 July 2026, with 110 banks under direct supervision taking part.
- The 300 basis point target was a minimum, set against a historical median annual depletion of around 225 to 250 basis points in EU banking crises.
- Escalation of the war in Ukraine, China-US trade disruption and Taiwan Strait tensions were the most frequently named conflicts; about a quarter of banks named a Middle East conflict.
- The exercise does not change Pillar 2 Guidance, but the final report says qualitative deficiencies may be weighed in the governance part of the SREP and could affect Pillar 2 requirements.
On 31 July 2026 the European Central Bank published the results of a stress test that gave banks the answer and asked them for the question.[^2] Each of 110 banks had to find the geopolitical events that would wipe out at least 300 basis points of its CET1 ratio, then show how capital, liquidity and operations would respond.[^1][^2] The exercise replaced the annual stress test banks would otherwise run inside their internal capital assessment, and it sat under the first of the ECB's supervisory priorities for 2026 to 2028, resilience to geopolitical risks.[^2][^9]

The Reverse Stress Test Starts From the Loss
A reverse stress test fixes the damage first and works back to the cause. The ECB set the depletion at a minimum of 300 basis points because EU banking sectors that lost capital in a given year over the last 15 to 20 years saw a median hit of around 225 to 250 basis points.[^3] About 20 percent of banks chose a harsher target.[^3] The report warns that the figure should not be compared with the 370 basis point adverse result of the EBA's 2025 EU-wide test, which used one common scenario for 64 banks.[^3][^10]
- Which conflicts banks chose. Further escalation of the war in Ukraine, China-US trade disruptions and tensions in the Strait of Taiwan were mentioned most often.[^3] About a quarter of banks named a Middle East conflict, including closure of the Strait of Hormuz, though the exercise was launched before tensions escalated in February 2026.[^3][^4]
- How the shock travels. Most banks treated the real economy as the main transmission channel, followed by financial market impact.[^2][^3]
- Where losses land. Banks projected loan losses concentrated in sectors such as manufacturing, transport, construction and agriculture.[^2][^3]
- What breaks operationally. 86 banks named cyberattacks as relevant and 57 as the main disruption type; the press release called cyber incidents and third-party service failures the most prominent threats.[^2][^3]
- No rescue assumed. Banks could not assume fiscal or government support beyond measures already approved.[^3]

Five Places Bank Scenario Design Fell Short
The ECB's overall verdict was that banks were generally able to produce economically meaningful scenarios.[^2] Sharon Donnery, a member of the Supervisory Board, said before the results that the test was "not about passing or failing."[^6] The weaknesses are where the follow-up will sit.
- Exposure fit. In some cases there was no clear connection between a bank's simulation and its most geopolitically vulnerable portfolios, and some calibrations were rudimentary.[^3]
- Liquidity. Several banks produced only a muted liquidity response despite a large capital decline.[^2] Median liquidity coverage fell from 186 percent to 163 percent over the year, while some foreign currency ratios fell below 100 percent.[^3]
- Narrative and numbers. The press release flagged gaps between scenario narratives and their translation into solvency and liquidity impacts.[^2] Claudia Buch, chair of the Supervisory Board, added that second-round effects and interactions between risk channels were sometimes not well identified.[^4][^5]
- Mitigants. Banks estimated their actions would recover about half of the lost capital.[^4] The most common were ICT and cyber recovery at 74 percent and suspending dividends at 59 percent, with 40 percent planning portfolio sales and 22 percent new capital.[^3][^4] The ECB found some assumptions overly optimistic, such as selling loans or raising capital at ambitious prices in a stressed market.[^2]
- Herding. Many banks planned the same actions, which may work less well if all of them act at once.[^3] Buch put it plainly: "If many are looking for the same exit, the exit might be blocked."[^4]
What the record shows: the exercise surfaced weak links between scenario and balance sheet more often than weak scenarios. What we infer: the scenarios came from strategy and economics teams, while the liquidity and mitigation numbers came from treasury and capital planning, and the two were not reconciled. The ECB does not say this; it is our reading of where the listed gaps cluster.

What Follows in the Supervisory Dialogue
The press release says the results will not change Pillar 2 Guidance and may inform qualitative SREP assessments.1 The full report goes further: qualitative deficiencies may be considered in the governance part of the SREP and could therefore affect Pillar 2 requirements.2 Forvis Mazars read the immediate consequence as greater scrutiny rather than an extra capital charge, and warned that "a recovery option cannot be considered credible simply because it appears in a plan."3 Supervisors will follow up with the banks concerned through the ongoing dialogue on capital and recovery planning.1
For a CRO, the supervisory dialogue sets the deadline. The bank will be asked to defend choices it made months ago, often by people who were not in the room when the scenario was written. A file that shows only the final capital number will not survive that conversation. A file that shows why each conflict was chosen, which desks priced the losses and who signed off the recovery actions stands a better chance. What we recommend: rebuild the scenario file so a supervisor can trace each step. Tie every named conflict to the portfolios it hits. Run the liquidity effects, including foreign currency, from the same narrative as the capital loss. Discount any mitigant that depends on market conditions, since a SUERF note by ECB staff found 40 to 60 percent of banks planned actions of that kind.4 Buch's own summary is the test to meet: "Scenario planning is a governance task, not only a technical one."5 The same discipline applies outside banking; geopolitical risk reaches a corporate balance sheet through the same channels, as our note on what prices move first shows. For the quarterly process behind a scenario file, see our assessment framework. Sector coverage is on our finance page.
Fortius Intel note: The open question is whether supervisors will treat herding as a bank-level weakness or a system-level one. If every bank's recovery plan sells the same assets into the same stressed market, no single bank can fix that alone, and the next supervisory round will show whether the ECB expects banks to model their peers' behaviour or intends to model it itself.
Methodology: this note draws on ECB banking supervision press releases of 12 December 2025 and 31 July 2026, the ECB's final report, a Supervision Blog post by Claudia Buch and an interview with Sharon Donnery, the EBA's 2025 results, and analysis by Forvis Mazars, Management Solutions and SUERF. SUERF Policy Note 423 is written by ECB staff and is not treated as an independent check on ECB figures. Each figure was checked against a second source where one exists.
Footnotes
-
European Central Bank, Banking Supervision, press release on the results of the geopolitical risk reverse stress test, 31 July 2026, https://www.bankingsupervision.europa.eu/press/pr/date/2026/html/ssm.pr260731~93964644b0.en.html, cross-checked with the Banco de Espana copy, https://www.bde.es/f/webbe/GAP/Secciones/SalaPrensa/ComunicadosBCE/NotasInformativasBCE/26/presbce2026-118en.pdf (both retrieved 8 November 2026). Source for: the 31 July 2026 date; 110 participating banks; replacement of the ICAAP stress test; banks generally producing meaningful scenarios; shortfalls in granularity, narrative-to-impact consistency, mitigant realism and solvency-liquidity interactions; muted liquidity responses; cyber and third-party disruption as the most prominent threats; real economy as main channel; sectors of loan losses; overly optimistic sale and capital-raising assumptions; no change to Pillar 2 Guidance; follow-up in the capital and recovery planning dialogue. ↩ ↩2
-
European Central Bank, "Geopolitical risk reverse stress test of euro area banks: 2026 SSM thematic stress test, final results," 2026, https://www.bankingsupervision.europa.eu/ecb/pub/pdf/ssm.geopolstresstest202608.en.pdf (retrieved 8 November 2026). Source for: the 300 basis point target as a minimum and about 20 percent of banks choosing a higher one; the historical median of around 225 to 250 basis points; the warning against comparison with EU-wide tests; most frequently named conflicts; about a quarter naming a Middle East conflict including Hormuz; 86 banks naming cyberattacks as relevant and 57 as the main type; no fiscal support assumed; weak links to vulnerable portfolios and rudimentary calibration; median LCR from 186 to 163 percent and some foreign currency ratios below 100 percent; management action shares (ICT resilience 74 percent, dividends 59 percent, portfolio sales 40 percent, capital 22 percent); herding risk; deficiencies possibly affecting Pillar 2 requirements through the governance part of the SREP. ↩
-
Forvis Mazars, "ECB's geopolitical stress test results," October 2026, https://www.forvismazars.com/group/en/insights/latest-insights/ecb-s-geopolitical-stress-test-results (retrieved 8 November 2026). Source for: the immediate consequence being greater supervisory scrutiny rather than an additional capital requirement; the quote that a recovery option cannot be considered credible simply because it appears in a plan. ↩
-
SUERF Policy Note 423, "Preparing for geopolitical risk," by ECB staff writing in a personal capacity, September 2026, https://www.suerf.org/publications/suerf-policy-notes-and-briefs/preparing-for-geopolitical-risk/ (retrieved 8 November 2026). Source for: 40 to 60 percent of banks indicating actions whose executability depends on market conditions. ↩
-
Claudia Buch, ECB Supervision Blog, "Preparing for the unknown: lessons from the ECB's reverse stress test on geopolitical risk," 11 September 2026, https://www.bankingsupervision.europa.eu/press/blog/2026/html/ssm.blog20260911~97fe7661d0.en.html (retrieved 8 November 2026). Source for: the exercise being launched before tensions escalated in February 2026; second-round effects and channel interactions not always well identified; banks estimating that actions would recover about half of depleted capital; cross-check of the 74 percent figure; the quotes "If many are looking for the same exit, the exit might be blocked" and "Scenario planning is a governance task, not only a technical one." ↩
About the author
Shekhar Attri, Co-Founder & CTO. An Indian Army Special Forces veteran with 21 years of service and a gallantry medal, Shekhar's corporate security advisory work spans Singapore, India, the Philippines, and the UAE, alongside PhD research on machine intelligence under incomplete information.