JP Morgan Chase 2003 Annual Report Download - page 71

J.P. Morgan Chase & Co. / 2003 Annual Report 69
performed daily and disclosed at the period-end date. Second,
VAR and stress tests are tw o distinct risk measurements yielding
very different loss potentials. Thus, although the same trading
portfolios are used for both tests, VAR is based on a distribution
of one-day historical losses measured over the most recent one
year; by contrast, stress testing subjects the portfolio to more
extreme, larger moves over a longer time horizon (e.g., 2–3
w eeks). Third, as VAR and stress tests are distinct risk measure-
ments, the impact of portfolio diversification can vary greatly.
For VAR, markets can change in patterns over a one-year time
horizon, moving from highly correlated to less so; in stress test-
ing, the focus is on a single event and the associated correla-
tions in an extreme market situation. As a result, w hile VAR over
a given time horizon can be low ered by a diversification benefit
in the portfolio, this benefit w ould not necessarily manifest itself
in stress-test scenarios, which assume large, coherent moves
across all markets.
Net int erest income st ress test ing
The VAR and stress-test measures described above illustrate the
total economic sensitivity of the Firm’s balance sheet to changes
in market variables. The effect of interest rate exposure on
reported Net income is also critical. The Firm conducts simula-
tions of Net interest income for its nontrading activities under a
variety of interest rate scenarios, which are consistent w ith the
scenarios used for economic-value stress testing.
Net interest income stress tests measure the potential change in
the Firm’s NII over the next 12 months. These stress tests highlight
exposures to various interest rate–sensitive factors, such as rates
(e.g., the prime lending rate), pricing strategies on deposits and
changes in product mix. These stress tests also take into account
forecasted balance sheet changes, such as asset sales and securiti-
zations, as w ell as prepayment and reinvestment behavior.
RIFLE
In addition to VAR, JPM organ Chase employs the Risk identifica-
tion for large exposures (“ RIFLE” ) methodology as another sta-
tistical risk measure. The Firm requires that all market risk–taking
businesses self-assess their risks to unusual and specific events.
Individuals w ho manage risk positions, particularly complex posi-
tions, identify potential w orst-case losses that could arise from
an unusual or specific event, such as a potential tax change, and
estimate the probabilities of such losses. Through the Firm’s
RIFLE system, this information is then directed to the appropriate
level of management, thereby permitting the Firm to identify
further earnings vulnerabilities not adequately covered by VAR
and stress testing.
Nonst at istical risk measures
Nonstatistical risk measures other than stress testing include net
open positions, basis point values, option sensitivities, position
concentrations and position turnover. These measures provide
additional information on an exposure’s size and the direction in
w hich it is moving. Nonstatistical measures are used for moni-
toring limits, one-off approvals and tactical controls.
VAR by risk type
2003 2002 (b)
As of or for the year Average Minimum Maximum At Average Minimum Maximum At
ended December 31, (in millions) VAR VAR VAR December 31 VAR VAR VAR December 31
By risk type:
Interest rate $63.9 $ 43.1 $ 109.9 $ 83.7 $67.6 $ 50.1 $ 94.7 $ 59.6
Foreign exchange 16.8 11.0 30.2 23.5 11.6 4.4 21.2 18.4
Equities 18.2 6.7 51.6 45.6 14.4 5.4 32.7 8.4
Commodities 2.9 1.7 4.9 3.3 3.6 1.6 13.3 1.9
Hedge fund investments 4.8 3.2 8.7 5.5 3.2 2.5 3.6 3.2
Less: portfolio diversification (38.0) NM NM (58.4) (28.8) NM NM (26.9)
Total Trading VAR(a) 68.6 43.2 114.7 103.2 71.6 57.0 102.8 64.6
Nontrading activities 151.8 81.5 286.0 203.8 97.3 68.9 139.3 107.7
Less: portfolio diversification (45.5) NM NM (25.7) (48.6) NM NM (61.0)
Total VAR $174.9 $ 83.7 $ 331.4 $ 281.3 $120.3 $ 87.6 $ 160.2 $ 111.3
(a) Amounts exclude VAR related to the Firm’s private equity business. For a discussion of Private equity risk management, see page 74 of this Annual Report.
(b) Amounts have been revised to reflect the reclassification of certain mortgage banking positions from the trading portfolio to the nontrading portfolio.
NM - Because the minimum and maximum may occur on different days for different risk components, it is not meaningful to compute a portfolio diversification effect. In addition, JPMorgan Chase’s average
and period-end VARs are less than the sum of the VARs of its market risk components, due to risk offsets resulting from portfolio diversification.
The table below show s both trading and nontrading VAR by risk type, together w ith the Corporate total. Details of the VAR exposures
are discussed in the Trading Risk and Nontrading Risk sections below.