RBS 2011 Annual Report Download - page 232

230 RBS Group 2011
Risk management: Market risk continued
Quantitative risk appetite
The Executive Risk Forum (ERF) approves the quantitative market risk
appetite for trading and non-trading activities. The Global Head of Market
&Insurance Risk, under delegated authority from the ERF, sets and
populates a limit framework, which is cascaded down through legal entity,
division, business and desk level market risk limits.
At the Group level, the risk appetite is expressed in the form of a
combination of VaR, sensitivity and stress testing limits.
Adaily report summarises the Group’s market risk exposures against the
agreed limits. This daily report is sent to the Head of Restructuring &
Risk, Global Head of Market & Insurance Risk, business Chief Risk
Officers and appropriate business market risk managers.
Legal entities, divisions and lower levels in the business also have an
appropriate market risk framework of controls and limits in place to cover
all material market risk exposures.
The specific market risk metrics that are appropriate for controlling the
positions of a desk will be more granular than the Group level limits and
tailored to the particular business.
In line with the overall business strategy to reduce risk exposures, the
Group’s market risk limits were adjusted down during 2011.
The majority of the Group’s market risk exposure is in the GBM and Non-
Core divisions and Group Treasury. The Group is also exposed to market
risk through interest rate risk on its non-trading activities. There are
additional non-trading market risks in the retail and commercial
businesses of the Group, principally interest rate risk and foreign
exchange risk. These aspects are discussed in more detail in Balance
sheet management - Interest rate risk on pages 131 and 132 and
structural foreign currency exposures on page 133.
Risk models
VaR is a technique that produces estimates of the potential change in the
market value of a portfolio over a specified time horizon at a given
confidence level. For internal risk management purposes, the Group’s
VaR assumes a time horizon of one trading day and a confidence level of
99%. The Group's VaR model is based on a historical simulation model,
utilising data from the previous two years.
The VaR 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, equity and specified commodity products
and specific risk in interest rate and equity products.
The VaR model is an important market risk measurement and control
tool. It is used for determining a significant component of the market risk
capital and, as such, it 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 daily VaR estimate, measured at a 99% confidence level.
The FSA categorises a VaR model as green, amber or red. A green
model status is consistent with a good working model and is achieved for
models that have four or fewer back-testing exceptions in a 12-month
period. For the Group’s trading book, a green model status was
maintained throughout 2011.
The Group’s VaR should be interpreted in light of the limitations of the
methodology used, as follows:
xHistorical simulation VaR may not provide the best estimate of future
market movements. It can only provide a prediction of the future
based on events that occurred in the two-year time series. Therefore,
events that are more severe than those in the historical data series
cannot be predicted.
xThe use of a 99% confidence level does not reflect the extent of
potential losses beyond that percentile.
xThe use of a one-day time horizon will not fully capture the profit and
loss implications of positions that cannot be liquidated or hedged
within one day.
xThe Group computes the VaR of trading portfolios at the close of
business. Positions may change substantially during the course of
the trading day and, if so, intra-day profit and losses will be incurred.
These limitations mean that the Group cannot guarantee that losses will
not exceed the VaR.
The RNIV framework has been developed to quantify those market risks
not adequately captured by the market standard VaR methodology.
Where risks are not included in the model, various non-VaR controls (for
example, portfolio size limits, sensitivity limits, triggers or stress limits) are
in place.
Risk models are developed both within business units and by Group
functions. Risk models are also subject to independent review and sign-
off to the same standard as pricing models. Meetings are held with the
FSA every quarter to discuss the traded market risk, including changes in
models, management, back-testing results, risks not included in the VaR
framework and other model performance statistics.
Anumber of VaR model and methodology enhancements were
introduced during 2011. The quality of the market data time series used
in the ABS mortgage trading business was improved, moving from
interpolated weekly data to daily observed time series. This change has
improved the accuracy of the correlation between the different time series
in the daily data. Additionally, the basis modelling between cash and
derivatives has been refined by introducing additional time series for the
sub-prime and subordinated residential bonds, reducing the over-reliance
on the commercial mortgage basis which was used as a conservative
proxy.
*unaudited
Business review Risk and balance sheet management continued