30 September 2025 Pillar 3 Report |
UBS Group | Risk-weighted assets
8
Key differences between the internal model approach and the standardized approach
Internal model approach
Standardized approach
Key impact
Risk weighting
Reliance on internal ratings where each
counterparty / transaction receives a rating.
Reliance on external credit assessment institutions
where permitted in the regulatory framework.
Modelled approach produces RWA that is more risk
sensitive.
Granular risk-sensitive risk weights differentiation
via individual probability of default (PD) and LGD
for mortgages.
Less granular risk weights based on loan-to-value
(LTV)
bands for mortgages.
The Group’s residential mortgage portfolio is
focused on the Swiss market, and the Group has
robust review processes in place concerning
borrowers’ ability to repay. This results in the
Group’s residential mortgage portfolio having a low
average LTV and results in an average risk weight
of around 20% under the advanced internal
ratings-based (A-IRB) approach.
Modelled LGD captures transaction quality
features including collateralization. Under the
foundation internal ratings-based (F-IRB)
approach, the LGD values are calculated based
on the rules set by regulatory authorities. This is
applicable for banks and large corporates.
No differentiation for transaction features (except
where a claim is subordinated).
Impact relevant across all asset classes.
Credit risk mitigation recognized via risk-sensitive
LGD or exposure at default (EAD).
Limited recognition of credit risk mitigation.
Standardized approach RWA is higher than
modelled RWA for most transaction types.
Wider variety of eligible collateral.
Restricted list of eligible collateral.
Limited recognition of collateral results in higher
RWA for Lombard lending and securities financing
transactions (SFTs).
Repo value-at-risk (VaR)
permits the use of VaR
models to estimate exposure and collateral for
SFTs. Approach permits full diversification and
netting across all collateral types.
Conservative and crude regulatory haircuts with
limited risk-sensitivity.
The effects
of guarantees and credit derivatives
are considered through either adjusting PD
and / or LGD estimates. UBS applies the F-IRB
approach for guarantee recognition.
In case of eligible guarantees and credit derivatives,
substitution is applied and the risk weight
applicable to the protection provider can be
assigned to the protected portion of the underlying
exposure.
CCF
A credit conversion factor (CCF) is applied to
model expected future drawdowns over the 12-
month period, irrespective of the actual maturity
of a particular transaction. The CCF includes
downturn adjustments and is the result of
analysis of internal data and expert opinion.
Credit exposure equivalents are determined by
applying CCF to off-balance sheet items. The CCFs
vary based on product type, maturity and the
underlying contractual agreements.
Modelled CCFs can be more tailored and
differentiated.
EAD for derivatives
Internal model method (IMM) facilitates the use
of a Monte Carlo simulation to estimate
exposure.
The standardized approach for CCR (SA-CCR) is
calculated as the replacement costs plus regulatory
add-ons that take into account potential future
market moves at predetermined fixed rates.
For large,
diversified derivatives portfolios,
standardized EAD is higher than modelled EAD.
Application of multiplier on IMM exposure
estimate.
Differentiates add-ons by five exposure types and
three maturity buckets only.
Variability in holding period applied to
collateralized transactions, reflecting liquidity
risks.
Limited netting can be recognized.
The repo VaR approach is a model based on a
Monte Carlo simulation and historical calibration
to estimate exposure, computed as quantile
exposure.
The comprehensive approach considers the adjusted
exposure after applicable supervisory haircuts on
both the exposure and the collateral received to
take account of possible future fluctuations in the
value of either the exposure or the collateral.
For large, diversified SFT portfolios, standardized
EAD is higher than modelled EAD.
Maturity in risk weight
Regulatory RWA function considers maturity: the
longer the maturity, the higher the risk weight.
No differentiation for maturity of transactions,
except for interbank exposures.
Model approach produces lower RWA for high-
quality, short-term transactions.
Credit valuation
adjustment
Not applicable under the final Basel III standards.
UBS calculates the CVA risk capital requirement
using both the standardized approach (SA-CVA)
and the full basic approach (BA-CVA) in line with
the final Basel III standards.
The SA-CVA uses
sensitivities to market risk factors (e.g. interest rates
and credit spreads) and uses those sensitivities with
regulatory-prescribed risk weights and correlations
to arrive at a capital charge. The BA-CVA approach
is simpler and less risk sensitive.
Where the BA-CVA and the SA-CVA are applied
under the output floor calculation, the application
of internal ratings is not permitted.
Securitization exposures
in the banking book
The regulatory capital requirements are
calculated using a hierarchy of approaches. First,
the securitization internal ratings-based approach
(SEC-IRBA) is applied, if possible. If this approach
cannot be applied, one of the standardized
If the SEC-IRBA cannot be applied, the regulatory
capital requirements are calculated using the
following hierarchy of approaches:
the securitization
external ratings-based approach or the
securitization standardized approach (SEC-SA).
Otherwise, a 1,250% risk weight is applied as a
fallback.