Barclays 2003 Annual Report Download - page 106

104
Critical Accounting Estimates
UK accounting standards require that the Group adopt the accounting
policies and estimation techniques that the Directors believe are most
appropriate in the circumstances for the purpose of giving a true and
fair view of the Group’s state of affairs, profit and cash flows. However,
different policies, estimation techniques and assumptions in critical areas
could lead to materially different results.
The following are estimates which are considered to be the most
complex and involve significant amounts of management valuation
judgements, often in areas which are inherently uncertain.
Bad and doubtful debts
The estimation of potential credit losses is inherently uncertain and
depends upon many factors, including general economic conditions,
changes in individual customer’s circumstances, structural changes
within industries that alter competitive positions, and other external
factors such as legal and regulatory requirements and other
governmental policy changes.
Specific provisions are raised when the Group considers that the
creditworthiness of a borrower has deteriorated such that the recovery
of the whole or part of an outstanding advance is in serious doubt.
For larger accounts this is usually done on an individual basis and all
relevant considerations that have a bearing on the expected future cash
flows are taken into account, for example the business prospects for the
customer, the realisable value of collateral, the Group’s position relative
to other claimants, the reliability of customer information and the likely
cost and duration of the work-out process. Subjective judgements are
made in this process that may vary from person to person and team to
team. Furthermore, judgements change with time as new information
becomes available or as workout strategies evolve, resulting in frequent
revisions to the specific provisions as individual decisions are taken,
case by case.
Within the retail and small businesses portfolios which are comprised of
large numbers of small homogeneous assets, statistical techniques are
used to raise specific provisions on a portfolio basis, based on historical
recovery rates. These statistical analyses use as primary inputs the
extent to which accounts in the portfolio are in arrears and historical
information on the eventual losses encountered from such delinquent
portfolios. There are many such models in use, each tailored to a
product, line of business or customer category. The models are updated
from time to time. However, experience suggests that the models are
reliable and stable, stemming from the very large numbers of accounts
from which the model building information is drawn. These models do
not contain judgemental inputs, but judgement and knowledge is
needed in selecting the statistical methods to use when the models are
developed or revised.
General provisions are raised to cover losses which are known from
previous historical experience to be present in loans and advances at the
balance sheet date, but which have not yet been specifically identified.
These provisions are adjusted at least half-yearly by an appropriate
charge or release of general provision based on statistical analyses,
other information about customers and judgements by management
and the Board.
In outline, the statistical analyses are performed on a portfolio basis as
follows: For larger accounts, gradings are used to rate the credit quality
of borrowers. Each grade corresponds to an expected default frequency
and is calculated by using statistical methodologies and expert
judgement. To ensure that the result is as accurate as possible, several
different sources may be used to rate a borrower (e.g. internal model,
external vendor model, ratings by credit rating agencies and the
knowledge and experience of the credit officers). The general provision
also takes into account the expected severity of loss at default, i.e. the
amount outstanding when default occurs that is not subsequently
recovered. Recovery is usually substantial and depends, for example, on
the level of security held in relation to each loan, and the bank’s position
relative to other claimants. Also taken into account is the expected
exposure at default. Both loss given default and exposure at default are
statistically derived values.
For the large numbers of retail accounts, the approach is in principle the
same as for the corporate and business accounts. However, individual
consideration of accounts is not practicable, and statistical
methodologies are used to assess the loss in portfolios of accounts.
The general provision also includes a specifically identified element to
cover country transfer risk calculated on a basis consistent with the
overall general provision calculation.
In establishing the level of the general provision, management
judgement is applied to the results of the statistical analyses. This is
applied at business level where management takes account of the quality
of the statistical analyses and the relevance of historical data used in the
analyses to individual or groups of customers, current information, and
the general economic and environmental factors mentioned above.
Further information on credit risk provisioning is set out on page 53.
Fair value of financial instruments
Some of the Bank’s financial instruments are carried at fair value,
including derivatives and debt securities held for trading purposes.
The fair value of a financial instrument is the amount at which the
instrument could be exchanged in a current transaction between willing
parties, other than in a forced or liquidation sale.
Financial instruments entered into as trading transactions, together with
any associated hedging, are measured at fair value and the resultant
profits and losses are included in dealing profits, along with interest and
dividends arising from long and short positions and funding costs
relating to trading activities. Assets and liabilities resulting from gains
and losses on derivative and foreign exchange contracts are reported
gross in other assets or liabilities, reduced by the effects of qualifying
netting agreements with counterparties.
Financial instruments are either priced with reference to a quoted
market price for that instrument or by using a valuation model. Where
the fair value is calculated using financial markets pricing models, the
methodology is to calculate the expected cash flows under the terms of
each specific contract and then discount these values back to a present
value. These models use as their basis independently sourced market
parameters including, for example, interest rate yield curves, equities
and commodities prices, option volatilities and currency rates. Most
market parameters are either directly observable or are implied from
instrument prices. However, where no observable price is available then
instrument fair value will include a provision for the uncertainty in the
market parameter based on sale price or subsequent traded levels.
Financial Review
Critical Accounting Estimates