Europe

Evolving reporting standards: Insight into the UAE and EU stress test methodologies

16 September 2024

5 minute read

Banks in the UAE are currently completing a stress test exercise mandated by the Central Bank of UAE (CBUAE) due by the end of September (2024). Concurrently, the European Banking Authority (EBA) is set to shift the financial landscape in the EU with the revision of its stress test methodology.

The EBA is adopting a more complex approach for the 2025 stress test with the aim of improving the quality of the exercise primarily through forecasting impacts in capital under the CRR 3 and increasing the granularity of the information to be collected.

Below we take a closer look at the nature of each approach.

What do the two frameworks have in common?

Both methodologies take inspiration from internal exercises such as the internal capital adequacy assessment process (ICAAP), internal liquidity adequacy assessment process (ILAAP), and recovery plans. They also consider the impact of risk-weighted assets (RWA) through Pillar I risks, and profit and loss (P&L) impacts being transmitted to available capital through similar items (e.g., net interest income (NII), dividends, etc). Moreover, assumptions such as static balance sheet and perfect forecast scenario are shared across both stress test frameworks.

How do the frameworks differ?

The main differences stem from the fact that EBA’s starting point (baseline scenario) will be reported under CRR3 and market risk, and NII involve more complex methodologies for forecasting losses.

Credit risk is treated similarly across both frameworks, although EBA’s methodology allows for more flexibility in forecasting portfolio degradation and provisions under each scenario. Also, the starting point differs, the EU allows for IRB models for Pillar 1 credit risk capital requirements.

Operational risk is also similar for capital requirements forecasting, although EBA includes the forecasting losses for conduct risk and other operational risk events.

Regarding other P&L items, while dividends are assumed to remain constant in CBUAE’s methodology, EBA assumes an average of the previous year. It’s worth noting that EBA’s methodology assumes little to no deviation from COREP and FINREP figures, and if so, they should be clearly described in the explanatory notes. CBUAE’s methodology makes no reference to other regulatory reports.

Below we take a closer look at the specific details around the key differences between the two methodologies:

Credit Risk

While CBUAE’s methodology implies the exclusion of debt securities qualified for HQLA, this doesn’t apply to EBA’s methodology as the only relevant scopes are accounting classification and capital asset classes.
UAE banks must consider CBUAE’s requirement on provisions (CBUAE Circular 28/2010) as a floor to their impairment values. Also, for 2023 values, if the provisions are below the floor, it is deducted from CET1 capital. EBA’s methodology has no reference to similar floors.
While CBUAE’s methodology requires banks to ensure the impact of COVID-19 is exempted from default rate estimation, this isn’t the case for EBA’s methodology. However, under EBA’s methodology, PGS and other COVID-19 specific measures are referred and should be reported. It is also assumed the impact of COVID-19 measures or adjustments performed must be mentioned in the explanatory notes if they have a material impact in the results.
CBUAE’s methodology requires the reporting of debt sustainability assessment for wholesale and retail to be used for a specific SICR trigger under the adverse scenario. This isn’t the case in EBA’s methodology, as it is based on the agreement of banks using their internal definition of SICR – although it should be described in the explanatory notes.
EBA’s methodology does not require the application of the backstop trigger of a threefold increase in lifetime PD for classification to stage 2, while CBUAE’s does.
While CBUAE’s methodology requires to report per country the countries whose exposure is above 2% of the total, EBA’s methodology only requires banks to report, at most, the 10 countries with the largest exposures (or the countries with the largest exposures up to 95%).
The approaches to FX impacts differ in the fact that while CBUAE methodology includes estimating translation reserves, EBA methodology only considers the impact of FX in risk parameters (TR, LR, PD and LGD) and resulting impact in RWA, EL and impairment. Also, FX hedges are not referred to in EBA’s credit risk methodology while they are in CBUAE.
While CBUAE still has transitional adjustments of IFRS9, these are no longer applicable in the EU.
CBUAE’s explanatory note requirements for ECL forecasting methodology used are more detailed than EBA’s methodology. However, it is expected that the ECB team following the exercise will be requesting similar information mentioned in the UAE methodology. There is not as much expectation on verifying the IFRS 9 models per se, but instead have as a principle to assess any kind of model used to estimate the starting point and point-in-time parameters.
EBA’s framework allows banks to use their own methodologies to estimate rating degradation and the applicable regulatory parameters. Estimation of RWA is similar, except for the fact EBA shall follow CRR3. CBUAE’s methodology has no mention on the possibility of using own methodologies for rating degradation.
For securitisations, EBA’s methodology requires filling a separate template under specific rules. There is no such specific template in the CBUAE methodology.

Market Risk

The CBUAE considers less counterparties to default and adds concentration risk to be measured in their counterparty credit risk estimations while EBA assumes the default of the 3 most vulnerable counterparties (based on the external PD for a 3-year horizon).
In EBA’s exercise, the methodology for CVA impact in RWA is significantly more complex than CBUAE’s, depending on if the bank has a trading exemption or has comprehensive approach, and if exposures are under standardised or basic approach. EBA’s methodology has additional constraints for CVA losses estimation depending on the type of collateral posted and requires detail on the sensitivities to several market risk factors.
EBA’s methodology allows for trading exemption banks to have an impact of zero on the full revaluation of fair value through profit or loss (FVPL) positions held with trading intent and their related hedges, while the CBUAE assumes they should always be estimated.
EBA’s methodology has more requirements and additional details to be included regarding amortised cost debt securities revaluation, although the methodologies are close – the revaluation is per risk factor, and impacts are fully recognised in the first year. However, no impact is assumed for the baseline scenario and risk factor impact is more detailed.
CBUAE’s methodology requires banks to use the 2023 year-end market risk RWA as a floor for the market risk RWA under baseline and adverse scenarios, whereas EBA’s methodology does not make mention of this.

Operational Risk

EBA’s methodology assumes constant capital requirements equal to Dec2024 restated under CRR3, while CBUAE’s assumes this to be the floor of the capital requirements to be projected for operational risk.
EBA’s methodology also adds the need to forecast conduct and other operational risk losses, which CBUAE’s does not require.

NII

While CBUAE assumes that NII should fall 10% under the adverse scenario, EBA’s methodology for NII only assumes that it cannot increase.
EBA’s methodology requires banks to report (for both assets and liabilities) the starting point reference rate and margin, the new reference rate according to the scenario and the new margin, subject to the pass-through constraints and comply with the intertemporal consistency constraint for EIR existing and maturing. The replacement of maturing positions related to both the reference rate and the margin is based on the methodologically prescribed average point of maturing. The CBUAE requires to report effective interest rate (EIR) and volumes directly into the template and consider the relevant constraints and interest rate shocks in forecasts.
EBA’s methodology implies that non-performing loans yield income (net of provisions and subject to caps on the EIR in the adverse scenario) while CBUAE’s does not.

P&L items

While dividends are constant at the 2023 value for CBUAE, EBA uses an average of the dividends paid between 2020 and 2024.
EBA’s methodology allows banks to use own methodologies to forecast non-interest income and does not suggest a preferred approach unlike the CBUAE. However, the macroeconomic scenario and some constraints should be considered (i.e., not considering FX effects and following general assumptions of the exercise such as static balance sheet).

Capital items

EBA’s methodology implies capital measures such as capital issuance post December 2024 should not impact available capital – measures taken until May 2025 should be reported but will not affect the results. CBUAE methodology allows approved new capital issuance before the end of June 2024 to be included during the forecast horizon.

Other items