Fiscal and monetary support programmes, coupled with restrictions on shareholder payouts meant that the EU banks entered the stress test with solid capital cushions which helped them sail through the prolonged recessionary scenarios, largely unscathed. The Irish banking system was projected to be the most vulnerable, while Deutsche Bank and Societe Generale had the weakest outcomes amongst the European global systemically important banks (G-SIBs).

With the next round of the Supervisory Review and Evaluation Process (SREP) only due at the end of 2022, banks should now focus their attention towards strengthening their operating model and risk governance, as these qualitative factors will form a key input to the outcome of the next SREP. Furthermore, EU banks should also look to accelerate their efforts towards the integration of climate risk into their overall risk management framework as one of the key preparatory measures ahead of the climate risk stress tests slated for 2022. The significance of this cannot be overstated given that thus far the ECB is the only regulator to explicitly state that the outcome of the on-going assessment of climate risk integration practices would have an indirect influence on the pillar 2 capital requirements in the next round of the SREP.

Challenges faced and lessons learned

In this year’s stress tests, banks faced additional challenges across data, modeling, reporting and governance. The table below summarises these challenges and some of the best practices we observed in our engagements with a number of banks.

Strong starting point helped banks sail through the prolonged recession scenario

The banks had entered the stress test with their strongest-ever starting point of 15.0% CET1 (Common Equity Tier 1 ratio) on a fully loaded (FL) basis, because of government support programmes and restrictions on shareholder payouts. Under the adverse scenario, the FL CET1 was projected to decline 485 basis points (bps) to a low of 10.2% at end-FY 23. The corresponding leverage ratio was projected to decline 124bps to 4.3%. (All mentions of CET1 and leverage ratio in this article are on an FL basis.) Earlier last week, the ECB had stated the restrictions on dividend payouts would expire September-end, an advance signal that the test outcome would be positive.

The sample of 50 banks that were stress-tested accounted for 70% of banking system assets across 15 EU and European Economic Area (EEA) countries. The tests framed harsh scenarios since the deep recession in 2020 was projected to be followed by a further three years of recession. That translated to a projected cumulative reduction of 3.6% in EU real gross domestic product (GDP), and factored in a ‘lower for longer’ interest rate environment. The deviation in the projected three-year cumulative real GDP from the baseline scenario was -12.9% – significantly wider than the -8.3% deviation in the 2018 test. The scenarios also assumed that the lending moratoriums would not be prolonged beyond their expiration dates (i.e., January 1, 2021 onwards) but that the public guarantee schemes would remain in effect through the projected horizon.

This drove the aggregate CET1 reduction across the banks to €265 billion. The cumulative losses from credit risk were projected at €308 billion (423bps of CET1), followed by market risk (€74 billion / 102bps) and operational risk losses (€49 billion / 68bps). These were only partly offset by net interest income and fee income.

The EBA concluded that banks with a high reliance on interest rate revenue and those with a high domestic focus were more negatively impacted. Pertinently, the variation in outcomes is much wider than in the previous tests, ranging from a minimum decline of 80bps to a maximum decline of 996 bps on a CET1 basis. Though two banks fell below their total SREP capital requirement, they are both currently under M&A negotiations. These cases are discussed later.

Polish banks shine, while Irish are the most vulnerable

In terms of CET1 depletion, banks from Poland were the best performers while the Danish and Irish were the most vulnerable. The story is slightly different in leverage ratio terms, with Norwegian and Polish banks being the top two, but the Irish and Italian banks being the most vulnerable. On an absolute basis, Irish banks were projected to have the lowest end-point (end-FY23) CET1 ratio (8.44%), while German banks had the lowest leverage ratio (3.62%).

Weakest performers already in talks for sale

The most vulnerable banks in the adverse scenario were Banca Monte dei Pasche di Siena (MP) and HSBC Continental Europe (HCE). While MP’s CET1 was projected to be wiped out (-0.1% at end-FY23), HCE’s was projected at 5.9%. In terms of the projected leverage ratio, MP and HCE were again the weakest. However, two other banks also breached the minimum 3.0% requirement – Nederlandse Waterschapsbank (NW, 2.5%), which was actually the most resilient but suffered from a weak starting point of 2.4%, and BNG Bank (2.8%). It is worth noting that MP is currently in discussions to be acquired by local peer Unicredit, while HSBC Holdings, as part of its group-wide restructuring initiatives, is in discussion with ‘My Money Bank’ (owned by Cerberus Capital) for the sale of HCE.

Among G-SIBs, Deutsche and Soc Gen most vulnerable while Santander most resilient

Among the European G-SIBs, Deutsche Bank (DB) and Societe Generale (SG) were the worst hit, with respective projected CET1 declines of 620bps and 562bps. While the impact on DB was more pronounced owing to low interest rates and its high cost structure, SG was also projected to have a very small positive pre-provision income which effectively led to high capital erosion.

Both banks were still comfortably above the minimum regulatory requirement albeit DB came very close to the threshold on the leverage ratio, which was projected at 3.1% at end-FY23. DB was quick to respond with a press release to remind the market of the capital build year-to-date, with a net income of €1.9 billion in 1H21, its best first-half return since 2015. On the positive side, Santander was the most resilient among G-SIBs, with a peak-to-trough decline of 324bps, but a delta decline of just 258bps at end-FY23. However, ING and Credit Agricole had the highest CET1 ratios of 11.0% and 10.6%, respectively, with Santander coming in at 9.3%.

SREP-driven Pillar 2 changes on hold until end-2022

As is usually the case, stress testing outcomes would be used as a key quantitative input by supervisors in determining Pillar 2 capital requirement for each bank as part of the SREP. In addition, qualitative inputs, especially the risk governance framework (qualitative assessment of each bank’s performance during the stress test), will also be used to determine the same.

While the previous methodology for setting Pillar 2 requirements was formula-driven, the new methodology would lead to classification of the banks into four buckets, based on two steps. The first step would be to bucket them by the level of CET1 depletion (0-3%, 3-6%, 6%-9%, >9%) and the second step would entail the application of expert judgment by the JSTs based on the idiosyncratic profile of the bank, leading to bucket-wise overlays (0-1.0%, 0.5-2.0%, 1.0-2.75%, >1.75%).

However, it is worth noting that the ongoing pandemic-driven capital relief measures, which allow the banks to operate below their Pillar 2 capital requirements, will remain in effect until end-2022 which is when the full review would take place. Until then, pillar 2 add-ons may only be imposed on a case-by-case basis against select vulnerable banks.

So, what’s next?

Now that the banks are over the line on the regular stress testing exercise, the time is ripe to prepare for the supervisory climate risk stress test slated for 2022 as part of the integration of climate risk into the broader risk management framework. A key differentiating fact of the EU regulators is that they are the only ones globally to have explicitly stated that the outcome of the assessment of climate risk integration practices may have implications for capital requirements (as part of the SREP process).

This year, both the EBA and ECB are running their own top-down pilot climate vulnerability assessment exercises. The plan for next year is to engage with banks in a bottom-up exercise. French and Dutch banks are relatively better positioned owing to their experience in similar exercises run by their local supervisors. Banks in other regions will need to learn from their experience and accelerate their planning to ensure appropriate teams with multi-disciplinary skillsets such as modeling, credit risk and environmental risk are equipped and ready to start working together when the time comes.

A good practice would be for banks to start leveraging the Network for Greening the Financial System (NGFS) scenarios to develop internal models to run long-horizon (30-years) scenario analysis across both, physical and transition risks.