Morgan Stanley 2010 Annual Report Download - page 106

The Company’s VaR model generally takes into account linear and non-linear exposures to equity and
commodity price risk, interest rate risk, credit spread risk and foreign exchange rates as well as linear exposures
to implied volatility risks. The VaR model also captures certain implied correlation risks associated with
portfolio credit derivatives as well as certain basis risks (e.g., corporate debt and related credit derivatives).
Among their benefits, VaR models permit estimation of a portfolio’s aggregate market risk exposure,
incorporating a range of varied market risks and portfolio assets. One key element of the VaR model is that it
reflects risk reduction due to portfolio diversification or hedging activities. However, VaR risk measures should
be interpreted carefully in light of the methodology’s limitations, which include the following: past changes in
market risk factors may not always yield accurate predictions of the distributions and correlations of future
market movements; changes in portfolio value in response to market movements (especially for complex
derivative portfolios) may differ from the responses calculated by a VaR model; VaR using a one-day time
horizon does not fully capture the market risk of positions that cannot be liquidated or hedged within one day; the
historical market risk factor data used for VaR estimation may provide only limited insight into losses that could
be incurred under market conditions that are unusual relative to the historical period used in estimating the VaR;
and published VaR results reflect past trading positions while future risk depends on future positions. VaR is
most appropriate as a risk measure for trading positions in liquid financial markets and will understate the risk
associated with severe events, such as periods of extreme illiquidity. The Company is aware of these and other
limitations and, therefore, uses VaR as only one component in its risk management oversight process. As
explained above, this process also incorporates stress testing and scenario analyses and extensive risk monitoring,
analysis, and control at the trading desk, division and Company levels.
The Company’s VaR model evolves over time in response to changes in the composition of trading portfolios
and to improvements in modeling techniques and systems capabilities. The Company is committed to continuous
review and enhancement of VaR methodologies and assumptions in order to capture evolving risks associated
with changes in market structure and dynamics. As part of regular process improvement, additional systematic
and name-specific risk factors may be added to improve the VaR model’s ability to more accurately estimate
risks to specific asset classes or industry sectors. Additionally, the Company continues to evaluate enhancements
to the VaR model to make it more responsive to more recent market conditions while maintaining a longer-term
perspective.
VaR for 2010. The table below presents the Company’s Trading, Non-trading and Aggregate VaR for each of
the Company’s primary market risk exposures at December 31, 2010 and December 31, 2009, incorporating
substantially all financial instruments generating market risk. A small proportion of trading positions generating
market risk is not included in VaR, and the modeling of the risk characteristics of some positions relies upon
approximations that, under certain circumstances, could produce significantly different VaR results from those
produced using more precise measures.
The counterparty portfolio, which reflects adjustments, net of hedges, relating to counterparty credit risk and
other market risks, was reclassified from Non-trading VaR into Trading VaR at January 1, 2010. This
reclassification reflects regulatory considerations surrounding the Company’s conversion to a financial holding
company and the trading book nature of the Company’s counterparty risk-hedging activities. Aggregate VaR was
not affected by this change; however, this reclassification increased Trading VaR and decreased Non-trading
VaR for the year ended December 31, 2009.
Since the reported VaR statistics are estimates based on historical position and market data, VaR should not be
viewed as predictive of the Company’s future revenues or financial performance or of its ability to monitor and
manage risk. There can be no assurance that the Company’s actual losses on a particular day will not exceed the
VaR amounts indicated below or that such losses will not occur more than five times in 100 trading days. VaR
does not predict the magnitude of losses that, should they occur, may be significantly greater than the VaR
amount.
100