31 December 2024 Pillar 3 Report |
UBS Group | Credit risk
28
Credit risk under the advanced internal ratings-based
approach
Annual |
Under the A-IRB approach, the required
capital for credit risk is
quantified through empirical models
that we have
developed
to
estimate
the
probability
of
default
(PD),
loss
given
default
(LGD),
exposure
at
default
(EAD)
and
other
parameters, subject to FINMA approval.
The table below
shows the main
features of our
key credit risk
models, including numbers of
key models used
by portfolio
and the main differences between models, as well as the
description of the main characteristics of approved models.
CRE: Main features of our key credit risk models
1
Portfolio in scope
Major asset classes
Model
approach
Number of key
models
Main drivers
Number of
years of loss
data
Probability of default
Sovereigns and central banks
Central governments and
central banks, Corporates:
other lending
Scorecard
1
Political, institutional and economic indicators including
qualitative factors
>15
Banks and other financial
institutions
Banks & Securities dealer,
Corporates: other lending
Scorecard
7
Financial data including balance sheet ratios, profit
and
loss data and qualitative factors
>15
Funds
Corporates: other lending
Scorecard
5
Financial data and ratios constructed from it (such as net
asset value, volatility of returns), qualitative factors
>15
Large corporates and
internationals
Corporates: other lending
Scorecard,
market data
3
Financial data including balance sheet ratios and profit
and loss, market data and qualitative factors
>15
Enterprises in Switzerland
Corporates: other lending,
Retail: other retail
Scorecard
2
Financial data including balance sheet ratios and profit
and loss, behavioral data and qualitative factors
>25
Commodity traders
Corporates: specialized
lending
Scorecard
2
Financial data including balance sheet ratios and profit
and loss, as well as non-financial criteria. Volume,
liquidity and duration of financed commodity
transactions
>20
Ship finance
Corporates: specialized
lending
Scorecard
1
Freight rates, ship market
values, operational expenses
and group information
>20
Owner-occupied mortgages and
other wealth-management
financing
Retail: residential
mortgages, Corporates:
other lending
Scorecard
5
Behavioral data, affordability relative to income,
property
type, loan-to-value, assets and qualitative
factors
>10
Income producing real estate
mortgages
Retail: residential
mortgages, Corporates:
specialized lending
Scorecard
3
Loan-to-value, debt-service-coverage,
financial data (for
large corporates only), behavioral data and qualitative
factors
>20
Lombard lending and
concentrated equity-based
lending (CEL)
Lombard: Retail: other
retail, CEL: Corporates:
other lending
Simulation
approach based
on historical
returns
3
Lending value ratio, collateral
asset class, historical asset
returns, counterparty factors
>10
Credit cards, consumer loans and
leases in Switzerland
Retail: qualifying revolving
retail and other retail,
Corporates: other lending
Scorecard
3
Client type and characteristics and behavioral data
>9
Other portfolios
Corporates: other lending,
Public sector entities, and
Multilateral development
banks, Corporates:
specialized lending
Scorecard,
pooled rating
approach,
rating template
6
Financial data including balance sheet ratios and profit
and loss, market data and qualitative factors.
Separate
models for Commercial Real Estate loans, Debt REITs,
Mortgage originators, Public sector entities and
Multilateral development banks / Supranationals
>15
Loss given default
Investment Bank – all
counterparties
Across the asset classes
Statistical
model
4
Counterparty and facility specific, including industry
segment, region, collateral, seniority, legal
environment,
bankruptcy procedures and macro-economic factors
>20
Swiss corporate and mortgage
lending portfolios
Corporates: other lending,
Corporates: specialized
lending, Retail: residential
mortgages
Statistical
model
4
Collateral type and client segment, loan-to-value,
time
since last valuation, location indicator
>10
Ship finance
Corporates: specialized
lending
Statistical
model
1
Loan-to-value of ship and financial collaterals
>20
International residential
mortgages and other wealth-
Retail: residential
mortgages, Retail: other
retail, Corporates: other
lending
Statistical
model
3
Loan-to-value, market value
shock
>10
Lombard lending and
concentrated equity-based
lending (CEL)
Lombard: Retail: other
retail, CEL: Corporates:
other lending
Simulation
approach based
on historical
returns
3
Loan-to-value, collateral asset class and liquidity,
historical asset returns, counterparty factors
>10
Credit cards, consumer loans and
leases in Switzerland
Retail: qualifying revolving
retail and other retail,
Corporates: other lending
Statistical
model
3
Collateral, accrued interests, client and product
characteristics, changes in original payment
plan
>9
Commercial real estate in the US
Corporates: specialized
lending
Statistical
Model
1
Loan-to-value, debt-service-coverage,
occupancy,
property type and region
>7
Exposure at default
Banking products
Across the asset classes
Statistical
model
11
Facility type and product type,
commitment type,
headroom, and client characteristics
>9
Traded products
Across the asset classes
Statistical
model
4
Product specific market drivers, e.g.
interest rates.
Separate models for OTC/ETD and SFT that generate
the
simulation of risk factors used for the credit exposure
n/a
1 Table captures the model landscape of UBS Group AG,
which also includes the models that are only applied to certain portfolios in legacy Credit Suisse infrastructure.