Barclays 2010 Annual Report Download - page 92

Each PD model outputs an estimate of default probability that is PIT,
TTC or a hybrid (e.g. a 50:50 blend). Bespoke conversion techniques,
appropriate to the portfolio in question, are then applied to convert
the model output to pure PIT and TTC PD estimates. In deriving the
appropriate conversion, industry and location of the counterparty and
an understanding of the current and long-term credit conditions are
considered. Both PIT and TTC PD estimates are recorded for each client.
Within Barclays, the calculation of internal ratings differs between
wholesale and retail customers. For wholesale portfolios, the rating system
is constructed to ensure that a client receives the same rating regardless
of the part of the business with which it is dealing. To achieve this, a model
hierarchy is adopted which requires users to adopt a specific approach to
rating each counterparty depending upon the nature of the business and
its location. A range of methods are utilised for estimating wholesale
counterparty PDs. These include bespoke grading models developed
within the Group (internal models), vendor models such as MKMV Credit
Edge and RiskCalc, and a conversion of external alphabet ratings from
either S&P, Moody’s or Fitch. Retail models, especially those used for
capital purposes, are almost exclusively built internally using Barclays
data. In many cases bureau data is used to complement internal data.
In addition, in some low data/low default environments, external
developments may also be utilised.
A key element of the Barclays wholesale framework is the PD Masterscale
(see below). This scale has been developed to distinguish meaningful
differences in the probability of default risk throughout the risk range.
In contrast to wholesale businesses, retail areas rarely bucket exposures
into generic grades for account management purposes (although they
may be used for reporting purposes). Instead, accounts are managed
at a more granular and bespoke level.
Exposure at default (EAD) represents the expected level of usage of the
credit facility should default occur. At the point of default, the customer
exposure can vary from the current position due to the combined effects
of additional drawings, repayment of principal and interest and fees. EAD
parameters are all derived from internal estimates and are determined
from internal historical behaviour. The lower bound of EAD for regulatory
capital purposes is the current balance at calculation of EAD. For derivative
instruments, exposure in the event of default is the estimated cost of
replacing contracts where counterparties have incurred obligations which
they have failed to satisfy.
Should a customer default, some part of the exposure is usually recovered.
The part that is not recovered, the actual loss, together with the economic
costs associated with the recovery process, comprise the loss given default
(LGD), which is expressed as a percentage of EAD. The Group estimates
an average LGD using historical information. The level of LGD depends
principally on: the type of collateral (if any); the seniority or subordination
of the exposure; the industry in which the customer operates (if a business);
the length of time taken for the recovery process and the timing of all
associated cash flows; and the work-out expense. The outcome is also
dependent on economic conditions that may determine, for example,
the prices that can be realised for assets, whether a business can readily
be refinanced or the availability of a repayment source for personal
customers. For the purposes of regulatory capital an adjustment is made
to the modelled LGD to account for the increased losses experienced
under downturn conditions, giving a ‘downturn LGD’.
E. Reporting
The Group dedicates considerable resources to gaining a clear and
accurate understanding of credit risk across the business and ensuring
that its balance sheet correctly reflects the value of the assets in
accordance with applicable accounting principles. This process can be
summarised in five broad stages:
– measuring exposures and concentrations;
– monitoring weaknesses in portfolios;
– identifying potential problem loans and credit risk loans
(collectively known as potential credit risk loans or PCRLs);
– raising allowances for impaired loans; and
– writing off assets when the whole or part of a debt is considered
irrecoverable.
Risk management
Credit risk management continued
Geographical analysis of loans and advances to customers
2010 2009
1
2
3
4
5
1
2
3
4
5
1UK
2Other EU
3US
4Africa
5Rest of World
Maturity analysis of loans and advances to customers
2010 2009
1
2
3
4
5
1
2
3
4
5
1On demand
2Not more than
three months
3Over three months
but not more than
one year
4Over one year but
not more than five
years
5Over five years
Loans and advances to customers by industry
2010
2009
£bn
18.2
14.8
Manufacturing
4.8
5.1
25.8
24.5
Construction and property
Government
10.7
9.2
Energy and water
19.2
17.4
Wholesale and retail
distribution and leisure
40.6
46.7
Cards, unsecured loans
and other personal lending
29.2
28.7
Business and other services
149.1168.1
Home loans
29.6
26.2
Other
93.1
87.4
Financial institutions
90 Barclays PLC Annual Report 2010 www.barclays.com/annualreport10