Barclays 2008 Annual Report Download - page 81

1
Business review
Barclays PLC Annual Report 2008 79
Risk Methodologies
Fundamental to the delivery of the Group’s risk management objectives are
a series of risk methodologies that allow it to measure, model, price, stress,
aggregate, report and mitigate the risks that arise from its activities. Many of
the most important processes relate to the internal ratings used in granting
credit and are discussed separately on page 82. The specific methodologies
used to manage market risk, liquidity risk, capital risk and operational risk are
also discussed in their corresponding sections. At a more general level, the
Group’s approach to risk management can be illustrated through its use of
stress testing and the controls around model governance.
Stress testing
As part of the annual stress testing process, Barclays estimates the impact
of a severe economic downturn on the projected demand and supply of
capital. This process enables the Group to assess whether it could meet its
minimum regulatory capital requirements throughout a severe recession.
The Risk Appetite numbers are validated by estimating the Group
sensitivity to adverse changes in the business environment and to include
operational events that impact the Group as a whole using stress testing
and scenario analysis. For instance, changes in certain macroeconomic
variables represent environmental stresses which may reveal systemic
credit and market risk sensitivities in our retail and wholesale portfolios.
The recession scenarios considered incorporate changes in
macroeconomic variables, including:
– Weaker GDP, employment or property prices
– Lower equity prices
– Interest rate curve shifts
– Commodity price movements
Such Group-wide stress tests allow senior management to gain a better
understanding of how portfolios are likely to react to changing economic
and geopolitical conditions and how the Group can best prepare for and
react to them. The stress test simulates the balance sheet and profit and
loss effects of stresses across the Group, investigating the impact on
profits and the ability to maintain appropriate capital ratios. Insights
gained are fully integrated into the senior management process and the
Risk Appetite framework. This process of analysis and senior management
oversight also provides the basis for fulfilling the stress testing
requirements of Basel II.
Group-wide stress testing is only one of a number of stress test
analyses that are performed as part of the wider risk management
process. Specific stress test analysis is used across all risk types to gain
a better understanding of the risk profile and the potential effects of
changes in external factors. These stress tests are performed at a range
of different levels, from analysis covering specific stresses on individual
sub-portfolios (e.g. the impact of higher unemployment on the US cards
portfolio) to regularly assessed stress scenarios(such as the effect of a
sudden rise in global interest rates on Barclays Capital’s market exposures).
Model Governance
Barclays has a large number of models in place across the Group, covering
all risk types. To minimise the risk of loss through model failure, a Group
Model Risk Policy (GMRP) has been developed. This has been extensively
reviewed and enhanced during the course of 2008.
The GMRP helps reduce the potential for model failure by setting
minimum standards around the model development and implementation
process. The Policy also sets the Group governance processes for all
models, which allows model risk to be monitored, and seeks to identify
and escalate any potential problems at an early stage.
To help ensure that sufficient management time is spent on the more
material models, each model is provided with a materiality rating. GMRP
defines the materiality ranges for all model types. The materiality ranges
are based on an assessment of the impact to the Group in the event of a
model error. The materiality affects the approval and reporting level for
each model, with the most material models being approved by the
Executive Models Committee, a technical sub-committee of Group
Executive Committee. Although final level of model sign-off will vary,
depending on model materiality, the standards of model build,
implementation, monitoring and maintenance do not change with the
materiality level.
Documentation must be sufficiently detailed, to allow an expert to
understand all appropriate aspects of model development. It must include
a description of the data used for model development, the methodology
used (and the rationale for choosing such a methodology), a description
of any assumptions made, as well as details of where the model works well
and areas that are known model weaknesses.
All models are subject to a validation and independent review process
before the model can be signed-off for implementation. The model
validation exercise must demonstrate that the model is fit for purpose and
provides accurate estimates. The independent review process will also
ensure that all aspects of the model development process have been
performed in a suitable manner.
The initial sign-off process ensures that the model is technically fit for
purpose as well as ensuring that the model satisfies the business
requirements and all the relevant regulatory requirements. As detailed
above, the process for model sign-off is based on materiality, with all of a
business unit’s models at least initially being approved in business-led
committees, and Group involvement increasing as the models become
more material.
Once implemented, all models within the Group are subject to an
annual validation, to ensure that they are performing as expected, and
that assumptions used in model development are still appropriate. In line
with initial sign-off requirements, annual validations are also formally
reviewed at the appropriate technical committee.
In addition to annual validation, models are subject to quarterly
performance monitoring. Model performance monitoring ensures that
deficiencies are identified early, and that remedial action can be taken before
the deficiency becomes serious enough to affect the decision-making
process. As part of this process, model owners set performance triggers
and define appropriate actions for their models in the event of breaches.
Externally developed models are subject to the same governance
standards as internal models, and must be initially approved for use
following the validation and independent review process. External models
are also subject to the same standards for ongoing monitoring and annual
validation requirements.
Within Barclays Capital, where models are used to value positions
within the trading book the positions are subject to regular independent
price testing which covers all trading positions. Prices are compared to
direct external market data where possible. When this is not possible,
more analytic techniques are used, such as industry consensus pricing
services. These services enable Barclays to anonymously compare
structured products and model-input parameters with those of other
banks engaged in the trading of the same financial products. The
conclusions and any exceptions to this exercise are communicated to
senior levels of business and infrastructure management.