Barclays 2009 Annual Report Download - page 125

www.barclays.com/annualreport09 Barclays PLC Annual Report 2009 123
Traded market risk
Barclays policy is to concentrate trading activities in Barclays Capital. This
includes transactions where Barclays Capital acts as principal with clients
or with the market. For maximum efficiency, client and market activities
are managed together. In Barclays Capital, trading risk is measured for the
trading book, as defined for regulatory purposes, and certain banking books.
Risk Measurement and Control
The measurement techniques used to measure and control traded market
risk include Daily Value at Risk (DVaR), Expected Shortfall, average of the
three worst hypothetical losses from the DVaR simulation (3W), Global
Asset Class stress testing and Global Scenario stress testing.
DVaR is an estimate of the potential loss arising from unfavourable market
movements, if the current positions were to be held unchanged for one
business day. Barclays Capital uses the historical simulation methodology
with a two-year unweighted historical period at the 95% confidence level.
The historical simulation methodology can be split into three parts:
– Calculate hypothetical daily profit or loss for each position over the most
recent two years, using observed daily market moves.
– Sum all hypothetical profits or losses for day one across all positions, giving
one total profit or loss. Repeat for all other days in the two-year history.
– DVaR is the 95th percentile selected from the two-year history of daily
hypothetical total profit or loss.
Market volatility decreased from the extreme levels observed in the second
half of 2008, but remained above pre-crisis 2007 levels. As a consequence of
the unweighted DVaR historical simulation methodology, the extreme 2008
volatility will continue to impact DVaR until late 2010.
The DVaR model has been approved by the FSA to calculate regulatory
capital for the trading book. The approval covers general market risk in
interest rate, foreign exchange, commodities and equity products, and
issuer specific risk for the majority of single name and portfolio traded credit
products. Internally, as noted before, DVaR is calculated for the trading book
and certain banking books.
When reviewing DVaR estimates, a number of considerations should
be taken into account. These are:
– Historical simulation uses the recent past to generate possible future
market moves but the past may not be a good indicator of the future.
– The one-day time horizon does not fully capture the market risk of
positions that cannot be closed out or hedged within one day.
– Intra-day risk is not captured.
– DVaR does not indicate the potential loss beyond the 95th percentile.
DVaR is an important market risk measurement and control tool and
consequently the model is regularly assessed. The main approach employed
is the technique known as back-testing which counts the number of days
when a loss (as defined by the FSA) exceeds the corresponding DVaR
estimate, measured at the 99% confidence level.
The FSA categorises a DVaR model as green, amber or red. A green
model is consistent with a good working DVaR model and is achieved for
models that have four or less back-testing exceptions in a 12-month period.
For Barclays Capital’s trading book, green model status was maintained for
2009 and 2008.
Expected Shortfall is the average of all hypothetical losses from the
historical simulation beyond DVaR. To improve the control framework, formal
monitoring of 3W (average of the three worst observations from the DVaR
historical simulation) was started in the first half of 2009.