HSBC 2008 Annual Report Download - page 165

163
• the elapsed time between the date to which the
market data relates and the balance sheet date;
and
• the manner in which the data was sourced.
Models provide a logical framework for the
capture and processing of necessary valuation inputs.
For fair values determined using a valuation model,
the control framework may include, as applicable,
independent development or validation of (i) the
logic within valuation models; (ii) the inputs to those
models; (iii) any adjustments required outside the
valuation models; and, (iv) where possible, model
outputs. Valuation models are subject to a process of
due diligence and calibration before becoming
operational and are calibrated against external
market data on an ongoing basis.
The results of the independent validation
process are reported to, and considered by, Valuation
Committees. Valuation Committees are composed of
valuation experts from several independent support
functions (Product Control, Market Risk
Management, Derivative Model Review Group and
Finance) in addition to senior management. The
members of each Valuation Committee consider the
appropriateness and adequacy of the fair value
adjustments and the effectiveness of valuation
models. If necessary, they may require changes to
model calibration or calibration procedures. The
Valuation Committees are overseen by the Valuation
Committee Review Group, which consists of Heads
of Global Banking and Markets’ Finance and Risk
Functions. All subjective valuation items with a
potential impact in excess of US$5 million are
reported to the Valuation Committee Review Group.
Determination of fair value
Fair values are determined according to the
following hierarchy:
• Quoted market price: financial instruments with
quoted prices for identical instruments in active
markets.
• Valuation technique using observable inputs:
financial instruments with quoted prices for
similar instruments in active markets or quoted
prices for identical or similar instruments in
inactive markets and financial instruments
valued using models where all significant
inputs are observable.
• Valuation technique with significant
unobservable inputs: financial instruments
valued using valuation techniques where one or
more significant inputs are unobservable.
The best evidence of fair value is a quoted price
in an actively traded market. In the event that the
market for a financial instrument is not active, a
valuation technique is used.
The judgement as to whether a market is
active may include, but is not restricted to, the
consideration of factors such as the magnitude and
frequency of trading activity, the availability of
prices and the size of bid/offer spreads. In inactive
markets, obtaining assurance that the transaction
price provides evidence of fair value or determining
the adjustments to transaction prices that are
necessary to measure the fair value of the instrument
requires additional work during the valuation
process.
The majority of valuation techniques employ
only observable market data, and so the reliability of
the fair value measurement is high. However, certain
financial instruments are valued on the basis of
valuation techniques that feature one or more
significant market inputs that are unobservable, and
for them, the derivation of fair value is more
judgemental. An instrument in its entirety is
classified as valued using significant unobservable
inputs if, in the opinion of management, a significant
proportion of the instrument’s balance sheet value
and/or inception profit (‘day 1 gain or loss’) is
driven by unobservable inputs. ‘Unobservable’ in
this context means that there is little or no current
market data available from which to determine the
price at which an arm’s length transaction would be
likely to occur. It generally does not mean that there
is no market data available at all upon which to base
a determination of fair value (consensus pricing data
may, for example, be used). Furthermore, in some
cases the majority of the fair value derived from a
valuation technique with significant unobservable
inputs may be attributable to observable inputs.
Consequently, the effect of uncertainty in
determining unobservable inputs will generally be
restricted to uncertainty about the overall fair value
of the financial instrument being measured. To help
in understanding the extent and the range of this
uncertainty, additional information is provided in the
section headed ‘Effect of changes in significant
unobservable assumptions to reasonably possible
alternatives’ below.
In certain circumstances, primarily where debt is
hedged with interest rate derivatives or structured
notes issued, HSBC records its own debt in issue at
fair value, based on quoted prices in an active market
for the specific instrument concerned, if available.
When quoted market prices are unavailable, the own
debt in issue is valued using valuation techniques,
the inputs for which are either based upon quoted