JP Morgan Chase 2015 Annual Report Download - page 145

JPMorgan Chase & Co./2015 Annual Report 135
Value-at-risk
JPMorgan Chase utilizes VaR, a statistical risk measure, to
estimate the potential loss from adverse market moves in a
normal market environment. The Firm has a single VaR
framework used as a basis for calculating Risk Management
VaR and Regulatory VaR.
The framework is employed across the Firm using historical
simulation based on data for the previous 12 months. The
framework’s approach assumes that historical changes in
market values are representative of the distribution of
potential outcomes in the immediate future. The Firm
believes the use of Risk Management VaR provides a stable
measure of VaR that closely aligns to the day-to-day risk
management decisions made by the lines of business, and
provides the necessary and appropriate information needed
to respond to risk events on a daily basis.
Risk Management VaR is calculated assuming a one-day
holding period and an expected tail-loss methodology which
approximates a 95% confidence level. VaR provides a
consistent framework to measure risk profiles and levels of
diversification across product types and is used for
aggregating risks across businesses and monitoring limits.
These VaR results are reported to senior management, the
Board of Directors and regulators.
Under the Firm’s Risk Management VaR methodology,
assuming current changes in market values are consistent
with the historical changes used in the simulation, the Firm
would expect to incur VaR “band breaks,” defined as losses
greater than that predicted by VaR estimates, not more
than five times every 100 trading days. The number of VaR
band breaks observed can differ from the statistically
expected number of band breaks if the current level of
market volatility is materially different from the level of
market volatility during the 12 months of historical data
used in the VaR calculation.
Underlying the overall VaR model framework are individual
VaR models that simulate historical market returns for
individual products and/or risk factors. To capture material
market risks as part of the Firm’s risk management
framework, comprehensive VaR model calculations are
performed daily for businesses whose activities give rise to
market risk. These VaR models are granular and incorporate
numerous risk factors and inputs to simulate daily changes
in market values over the historical period; inputs are
selected based on the risk profile of each portfolio as
sensitivities and historical time series used to generate daily
market values may be different across product types or risk
management systems. The VaR model results across all
portfolios are aggregated at the Firm level.
For certain products, specific risk parameters are not
captured in VaR due to the lack of inherent liquidity and
availability of appropriate historical data for these products.
The Firm uses proxies to estimate the VaR for these and
other products when daily time series are not available. It is
likely that using an actual price-based time series for these
products, if available, would affect the VaR results
presented.
In addition, data sources used in VaR models may not be the
same as those used for financial statement valuations. In
cases where market prices are not observable, or where
proxies are used in VaR historical time series, the sources
may differ. The daily market data used in VaR models may
be different than the independent third-party data collected
for VCG price testing in VCG’s monthly valuation process
(see Valuation process in Note 3 for further information on
the Firm’s valuation process). VaR model calculations
require daily data and a consistent source for valuation and
therefore it is not practical to use the data collected in the
VCG monthly valuation process.
Since VaR is based on historical data, it is an imperfect
measure of market risk exposure and potential losses, and
it is not used to estimate the impact of stressed market
conditions or to manage any impact from potential stress
events. In addition, based on their reliance on available
historical data, limited time horizons, and other factors, VaR
measures are inherently limited in their ability to measure
certain risks and to predict losses, particularly those
associated with market illiquidity and sudden or severe
shifts in market conditions. The Firm therefore considers
other measures in addition to VaR, such as stress testing, to
capture and manage its market risk positions.
The Firm’s VaR model calculations are periodically
evaluated and enhanced in response to changes in the
composition of the Firm’s portfolios, changes in market
conditions, improvements in the Firm’s modeling techniques
and other factors. Such changes may also affect historical
comparisons of VaR results. Model changes undergo a
review and approval process by the Model Review Group
prior to implementation into the operating environment.
For further information, see Model risk on page 142.
The Firm calculates separately a daily aggregated VaR in
accordance with regulatory rules (“Regulatory VaR”), which
is used to derive the Firm’s regulatory VaR-based capital
requirements under Basel III. This Regulatory VaR model
framework currently assumes a ten business-day holding
period and an expected tail loss methodology which
approximates a 99% confidence level. Regulatory VaR is
applied to “covered” positions as defined by Basel III, which
may be different than the positions included in the Firm’s
Risk Management VaR. For example, credit derivative
hedges of accrual loans are included in the Firm’s Risk
Management VaR, while Regulatory VaR excludes these
credit derivative hedges. In addition, in contrast to the
Firm’s Risk Management VaR, Regulatory VaR currently
excludes the diversification benefit for certain VaR models.