Barclays 2008 Annual Report Download - page 84

82 Barclays PLC Annual Report 2008 |Find out more at www.barclays.com/annualreport08
Risk management
Credit risk management
Measurement, reporting and internal ratings
The principal objective of credit risk measurement is to produce the most
accurate possible quantitative assessment of the credit risk to which the
Group is exposed, from the level of individual facilities up to the total
portfolio. The key building blocks in this quantitative assessment are:
– Probability of default (PD)
– Exposure in the event of default (EAD)
– Loss given default (LGD)
Barclays first began to use internal estimates of PD in its main businesses
in the 1990s. Internally derived estimates for PD, EAD and LGD have since
been used in our major risk decision-making processes, enabling the
application of coherent risk measurement across all credit exposures,
retail and wholesale.
With the advent of the Basel II accord on banking, Barclays has been
given permission to use internal rating models as an input to its regulatory
capital calculations. In preparation, Barclays spent considerable time
developing and upgrading a number of such models across the Group,
moving towards compliance with the Basel II advanced internal ratings
based approach. As part of this process, all Basel credit risk models have
been assessed against the Basel II minimum requirements prior to model
sign-off to ensure that they are fit to be used for regulatory purposes.
Applications of internal ratings
The three components described above – the PD, EAD and LGD – are
building blocks used in a variety of applications that measure credit risk
across the entire portfolio. These parameters can be calculated
incorporating different aspects of the credit cycle into the estimates:
– PD estimates can be calculated on a through-the-cycle (TTC) basis,
reflecting the predicted default frequency in an average 12 month
period across the credit cycle, or on a point-in-time (PIT) basis,
reflecting the predicted default frequency in the next 12 months.
– LGD and EAD estimates can be calculated as downturn measures,
reflecting behaviour observed under stressed economic conditions,
or as business-as-usual (BAU) measures, reflecting best modelled
behaviour under actual conditions.
These parameters, in suitable combination, are used in a wide range of
credit risk measurement and management and as our understanding and
experience have developed, we have extended the use and sophistication
of internal ratings into the following:
– Credit Approval: PD models are used in the approval process in both
retail and wholesale portfolios. In high-volume retail portfolios,
application and behaviour scorecards are frequently used as decision-
making tools. In wholesale and some retail mortgage portfolios, PD
models are used to direct applications to different credit sanctioning
levels, so that credit risks are reviewed at appropriate levels.
– Credit Grading: originally introduced in the early 1990s to provide a
common measure of risk across the Group using an eight point rating
scale; wholesale credit grading now employs a 21 point scale of default
probabilities.
– Risk-Reward and Pricing: PD, EAD and LGD metrics are used to assess
profitability of deals and portfolios and to allow for risk-adjusted pricing
and strategy decisions.
– Risk Appetite: measures of expected loss and the potential volatility of
loss are used in the Group’s Risk Appetite framework (see page 78).
– IAS 39: many of our collective impairment estimates incorporate the
use of our PD and LGD models, adjusted as necessary.
– Collections and Recoveries: model outputs are frequently used to
segment portfolios allowing for suitably prioritised collections and
recoveries strategies in retail portfolios.
– Economic capital (EC) allocation: most EC calculations use the same PD
and EAD inputs as the regulatory capital (RC) process. The process also
uses the same underlying LGD model outputs as the RC calculation,
but does not incorporate the same economic downturn adjustment
used in RC calculations.
– Risk management information: Group Risk and the business units
generate risk reports to inform senior management on issues such
as the business performance, Risk Appetite and consumption of EC.
Calculation of internal ratings
To calculate probability of default (PD), Barclays assesses the credit quality
of borrowers and other counterparties and assigns them an internal risk
rating. Multiple rating methodologies may be used to inform the rating
decision on individual large credits, such as internal and external models,
rating agency ratings, and for wholesale assets market information such
as credit spreads. For smaller credits, a single source may suffice such as
the result from an internal rating model. Barclays recognises the need for