JP Morgan Chase 2010 Annual Report Download - page 174

Notes to consolidated financial statements
174 JPMorgan Chase & Co./2010 Annual Report
Derivatives
Exchange-traded derivatives valued using quoted prices are
classified within level 1 of the valuation hierarchy. However, few
classes of derivative contracts are listed on an exchange; thus, the
majority of the Firm’s derivative positions are valued using inter-
nally developed models that use as their basis readily observable
market parameters – that is, parameters that are actively quoted
and can be validated to external sources, including industry
pricing services. Depending on the types and contractual terms of
derivatives, fair value can be modeled using a series of tech-
niques, such as the Black-Scholes option pricing model, simula-
tion models or a combination of various models, which are
consistently applied. Where derivative products have been estab-
lished for some time, the Firm uses models that are widely ac-
cepted in the financial services industry. These models reflect the
contractual terms of the derivatives, including the period to
maturity, and market-based parameters such as interest rates,
volatility, and the credit quality of the counterparty. Further, many
of these models do not contain a high level of subjectivity, as the
methodologies used in the models do not require significant
judgment, and inputs to the models are readily observable from
actively quoted markets, as is the case for “plain vanilla” interest
rate swaps, option contracts and CDS. Such instruments are
generally classified within level 2 of the valuation hierarchy.
Derivatives that are valued based on models with significant
unobservable market parameters and that are normally traded
less actively, have trade activity that is one way, and/or are traded
in less-developed markets are classified within level 3 of the
valuation hierarchy. Level 3 derivatives include, for example, CDS
referenced to certain MBS, certain types of CDO transactions,
options on baskets of single-name stocks, and callable exotic
interest rate options.
Other complex products, such as those sensitive to correlation
between two or more underlying parameters, also fall within level
3 of the valuation hierarchy, and include structured credit deriva-
tives which are illiquid and non-standard in nature (e.g., synthetic
CDOs collateralized by a portfolio of credit default swaps “CDS”).
For most CDO transactions, while inputs such as CDS spreads
may be observable, the correlation between the underlying debt
instruments is unobservable. Correlation levels are modeled on a
transaction basis and calibrated to liquid benchmark tranche
indices. For all structured credit derivatives, actual transactions,
where available, are used regularly to recalibrate all unobservable
parameters.
Correlation sensitivity is also material to the overall valuation of
options on baskets of single-name stocks; the valuation of these
baskets is typically not observable due to their non-standardized
structuring. Correlation for products such as these is typically esti-
mated based on an observable basket of stocks and then adjusted
to reflect the differences between the underlying equities.
For callable exotic interest rate options, while most of the as-
sumptions in the valuation can be observed in active markets
(e.g., interest rates and volatility), the callable option transaction
flow is essentially one-way, and as such, price observability is
limited. As pricing information is limited, assumptions are based
on the dynamics of the underlying markets (e.g., the interest rate
markets) including the range and possible outcomes of the appli-
cable inputs. In addition, the models used are calibrated, as
relevant, to liquid benchmarks, and valuation is tested against
monthly independent pricing services and actual transactions.
Mortgage servicing rights and certain retained interests
in securitizations
Mortgage servicing rights (“MSRs”) and certain retained interests
from securitization activities do not trade in an active, open
market with readily observable prices. Accordingly, the Firm
estimates the fair value of MSRs and certain other retained inter-
ests in securitizations using DCF models.
• For MSRs, the Firm uses an option-adjusted spread (“OAS”)
valuation model in conjunction with the Firm’s proprietary
prepayment model to project MSR cash flows over multiple in-
terest rate scenarios; these scenarios are then discounted at
risk-adjusted rates to estimate the fair value of the MSRs. The
OAS model considers portfolio characteristics, contractually
specified servicing fees, prepayment assumptions, delinquency
rates, late charges, other ancillary revenue, costs to service
and other economic factors. The Firm reassesses and periodi-
cally adjusts the underlying inputs and assumptions used in
the OAS model to reflect market conditions and assumptions
that a market participant would consider in valuing the MSR
asset. Due to the nature of the valuation inputs, MSRs are
classified within level 3 of the valuation hierarchy.
• For certain retained interests in securitizations, the Firm esti-
mates the fair value for those retained interests by calculating
the present value of future expected cash flows using model-
ing techniques. Such models incorporate management's best
estimates of key variables, such as expected credit losses, pre-
payment speeds and the appropriate discount rates, consider-
ing the risk involved. Changes in the assumptions used may
have a significant impact on the Firm's valuation of retained
interests, and such interests are therefore typically classified
within level 3 of the valuation hierarchy.
For both MSRs and certain other retained interests in securitiza-
tions, the Firm compares its fair value estimates and assumptions
to observable market data where available and to recent market
activity and actual portfolio experience. For further discussion of
the most significant assumptions used to value retained interests
and MSRs, as well as the applicable stress tests for those assump-
tions, see Note 16 on pages 244–259, and Note 17 on pages
260–263 of this Annual Report.