Credit risk management


Credit risk is the risk of suffering financial loss should any of the Group’s customers, clients or market counterparties fail to fulfil their contractual obligations to the Group. Credit risk may also arise where the downgrading of an entity’s credit rating causes the fair value of the Group’s investment in that entity’s financial instruments to fall. The credit risk that the Group faces arises mainly from commercial and consumer loans and advances, including credit card lending.

The granting of credit is one of the Group’s major sources of income and as its most significant risk, the Group dedicates considerable resources to controlling it. The importance of credit risk is illustrated by noting that two-thirds of risk-based economic capital is allocated to credit risk. Credit exposures arise principally in loans and advances.

In managing credit risk, the Group applies the five-step risk management process and internal control framework described previously (page 82).

Specific credit risk management objectives are:

  • – To gain a clear and accurate understanding and assessment of credit risk across the business, from the level of individual facilities up to the total portfolio.
  • – To control and plan the taking of credit risk, ensuring it is coherently priced across the business and avoiding undesirable concentrations.
  • – To support strategic growth and decision-making based on sound credit risk management principles and a pro-active approach to identifying and measuring new risks.
  • – To ensure a robust framework for the creation, use and ongoing monitoring of the Group’s credit risk measurement models.
  • – To ensure that our balance sheet correctly reflects the value of our assets in accordance with accounting principles.

Organisation and structure

Barclays has structured the responsibilities of credit risk management so that decisions are taken as close as possible to the business, whilst ensuring robust review and challenge of performance, risk infrastructure and strategic plans.

The credit risk management teams in each business are accountable to the Business Risk Directors in those businesses who, in turn, report to the heads of their businesses and also to the Risk Director.

These credit risk management teams assist Group Risk in the formulation of Group Risk policy and its implementation across the businesses. Examples include:

  • – maximum exposure guidelines to limit the exposures to an individual customer or counterparty
  • – country risk policies to specify risk appetite by country and avoid excessive concentration of credit risk in individual countries
  • – policies to limit lending to certain industrial sectors
  • – underwriting criteria for personal loans and maximum loan-to-value ratios for home loans

Within Group Risk, the Credit Risk function provides Group-wide direction of credit risk-taking. This functional team manages the resolution of all significant credit policy issues and runs the Credit Committee, which approves major credit decisions.

The principal Committees that review credit risk management, formulate overall Group credit policy and resolve all significant credit policy issues are the Group Wholesale Credit Risk Management Committee, the Group Retail Credit Risk Management Committee, the Risk Oversight Committee and the Board Risk Committee (see page 84 for more details of this Committee). The Board Audit Committee also reviews the impairment allowance as part of financial reporting.

Measurement, reporting and internal ratings

The principal objective of credit risk measurement is to produce the most accurate possible quantitative assessment of the credit risk to which the Group is exposed, from the level of individual facilities up to the total portfolio.

The key building blocks in this quantitative assessment are:

  • – Probability of default (PD)
  • – Exposure in the event of default (EAD)
  • – Severity of loss given default (LGD)

Barclays first began to use internal estimates of PD (internal ratings) in all its main businesses in the 1990s. Internally derived estimates for PD, EAD and LGD have been used since then in all our major risk decision making processes, enabling the application of coherent risk measurement across all credit exposures, retail and wholesale.

With the advent of the Basel II accord on banking, Barclays has been given permission to use internal rating models as an input in its regulatory capital calculations. In preparation, Barclays has spent considerable time developing and upgrading a number of such models across the Group, moving towards compliance with the Basel II advanced internal ratings based approach. As part of this process, all Basel credit risk models are assessed against the Basel II minimum requirements prior to model sign-off to ensure that they are fit to be used for regulatory purposes.

Applications of internal ratings

The three components described above – the probability of default, exposure at default and loss given default – are building blocks used in a variety of applications that measure credit risk across the entire portfolio.

Two examples are Risk Tendency (RT) and Expected Loss (EL) which are statistical estimates of the average loss for the loan portfolio for a 12-month period, taking into account the portfolio’s size and risk characteristics under either current credit conditions (RT) or average credit conditions (EL). As such, RT uses a point-in-time PD while EL uses a through-the-cycle PD but the basic calculation is the same for both:

PD x EAD x LGD

Since through-the-cycle PDs provide a measure of risk that is independent of the current credit conditions for a particular customer type, they are more stable than point-in-time ratings. RT and EL provide insight into the credit quality of the portfolio and assist management in tracking risk changes as the Group’s stock of credit exposures evolves in size or risk profile in the course of business.

As our understanding and experience have developed, we have extended the use and sophistication of internal ratings. The other main business processes that use internal estimates of PD, LGD and EAD, are as follows:

  • – Credit Grading – originally introduced in the early 1990s to provide a common measure of risk across the Group using an eight point rating scale; wholesale credit grading now employs a 21 point scale (Barclays Masterscale).
  • – Credit Approval – a rating scale is used to set differentiated credit sanctioning levels based upon a PD, so that credit risks are reviewed at appropriate levels.
  • – Risk Appetite – measures of expected loss and the potential volatility of loss are used in the Group’s Risk Appetite framework (see page 86).
  • – Pricing – within the corporate mass market portfolios we first developed and used risk adjusted pricing models in the early 1990s to differentiate risk reward decisions.
  • – IAS Impairment calculations – many of our collective impairment estimates incorporate the use of our PD and LGD models.
  • – Economic capital (EC) allocation – most EC calculations use the same through-the-cycle PD and EAD inputs as the regulatory capital (RC) process. The process also uses the same underlying LGD model outputs as the RC calculation, but does not incorporate the economic downturn adjustment used in RC calculations.
  • – Risk management information – Group and the main business units have for several years received either Key Information Packs or other risk reports focused on EL and EC information to inform senior management on issues such as the business performance, Risk Appetite and consumption of EC.
Calculation of internal ratings

To calculate probability of default (PD), Barclays assesses the credit quality of borrowers and other counterparties and assigns them an internal risk rating.

Multiple rating methodologies may be used to inform the rating decision on individual large credits, such as internal and external models, rating agency ratings, and for wholesale assets market information such as credit spreads. For smaller credits, a single source may suffice such as the result from an internal rating model.

Barclays recognises the need for two different expressions of PD depending on the purpose for which it is used. For the purposes of calculating regulatory and economic capital, long-run average through-the-cycle PDs are required. However, for the purposes of pricing and risk tendency, PDs should represent the best estimate of probability of default, typically in the next 12 months, dependent on the current position in the credit cycle. Hence, point-in-time PDs are also required.

When each PD model is constructed, its output is specified as one of point-in-time (PIT) or through-the-cycle (TTC) or a hybrid, e.g. a 50:50 blend. Using this distinction between PIT and TTC, the PDs are bucketed into both PIT Default Grades (DGs) and TTC bands, adopting techniques that are relevant to the model’s initial output calibration, the industry and location of the counterparty and an understanding of the current and long-term credit conditions. Two grades are therefore recorded for each client, the DG and the TTC band. A customer may therefore be rated DG 6 reflecting sectoral performance and TTC band 8 reflecting long-term credit conditions.

This same PIT/TTC distinction is applied to agency ratings. Within Barclays, an agency alphabet rating is also expressed in terms of PIT DG and TTC band. It is therefore no longer possible to produce a static mapping of agency letter ratings to either DGs or TTC bands because they are considered a hybrid of both PIT and TTC. As such, any mappings would change over time with movements in the credit cycle.

Barclays internal rating system also differentiates between corporate and retail customers.

For corporate portfolios (primarily Barclays Capital, BCB and the commercial areas of IRCB), the rating system is constructed to ensure that each client receives the same rating independent of the part of the business with which they are dealing. To achieve this, a model hierarchy is adopted which requires users to adopt a specific approach to rating each counterparty depending upon the nature of the business and its location. A range of methods is approved for estimating counterparty PDs. These include bespoke grading models developed within the Barclays Group (Internal Models), vendor models such as MKMV Credit Edge and RiskCalc, and a conversion of external alphabet ratings from either S&P, Moody’s or Fitch.

A key element of the Barclays framework is the Masterscale. This has been developed to record differences in the probability of default risk at meaningful levels throughout the risk range (see table below).

In contrast to corporate businesses, retail areas do not bucket exposures into generic grades or bands for account management purposes (although they may be used for reporting purposes). Instead, accounts are managed based on internal, product specific segmentations of accounts, for instance, deriving from the cut-offs of the associated models. The cut-offs may be in the form of a score, a probability of default, a measure of forecast loss or a more sophisticated risk/reward based measure.

Exposure at default (EAD) represents the expected level of usage of the credit facility when default occurs. At default the customer may not have drawn the loan fully or may already have repaid some of the principal, so that exposure is typically less than the approved loan limit. When the Group evaluates loans, it takes exposure at default into consideration, using its extensive historical experience. It recognises that customers may make heavier than average usage of their facilities as they approach default. For derivative instruments, exposure in the event of default is the estimated cost of replacing contracts with a positive value should counterparties fail to perform their obligations.

When a customer defaults, some part of the amount outstanding on their loans is usually recovered. The part that is not recovered, the actual loss, together with the economic costs associated with the recovery process combine to a figure called the loss given default (LGD), which is expressed as a percentage of EAD.

Using historical information, the Group can estimate how much is likely to be lost, on average, for various types of loans. To illustrate, LGD is lower for residential mortgages than for unsecured loans because of the property pledged as collateral.

The level of LGD depends on: the type of collateral (if any); the seniority or subordination of the exposure; the industry in which the customer operates (if a business); and the jurisdiction applicable and work-out expenses. The outcome is also dependent on economic conditions that may determine, for example, the prices that can be realised for assets, whether a businesses can readily be refinanced or the availability of a repayment source for personal customers.

The ratings process

The term ‘internal ratings’ usually refers to internally calculated estimates of PD. These ratings are combined with EAD and LGD in the range of applications described previously. The ‘ratings process’ refers to the use of PD, EAD and LGD across the Group. In Barclays, the rating process is defined by each business. For central government and banks, institutions and corporate customers many of the models used in the rating process are shared across businesses as the models are customer specific. For retail exposures, the ratings models are usually unique to the business and product type e.g. mortgages, credit cards, and consumer loans.

A bespoke model has been built for PD and LGD for Sovereign ratings. For Sovereigns where there is no externally available rating we use an internally developed PD scorecard. The scorecard has been developed using historic data on Sovereigns from an external data provider covering a wide range of qualitative and quantitative information. Our LGD model is based on resolved recoveries in the public domain, with a significant element of conservatism added to compensate for the small sample size.

To construct ratings for institutions , corporates , specialised lending and purchased corporate receivables and equity exposures, we use external models, rating agencies and internally constructed models. External models employed include Moody’s Credit Edge, rating agency ratings and Moody’s RiskCalc. The applicability of each of these approaches to our customers has been validated by us to internal rating standards. The data used in validating these primary indicators are representative of the population of the bank’s actual obligors and exposures and its long-term experience.

Internally built PD models are also widely used. We employ a range of methods in the construction of these models. The basic types of PD modelling approaches used are:

  • – Structural models
  • – Expert lender
  • – Statistical

Structural models incorporate in their specification the elements of the industry accepted Merton framework to identify the distance to default for a counterparty. This relies upon the modeller having access to specific time series data or data proxies for the portfolio. Data samples used to build and validate these models are typically constructed by adding together data sets from internal default observations with externally obtained data sets from commercial providers such as rating agencies and industry gathering consortia.

The Barclays Masterscale (Wholesale)

Excel File Download table as excel file
DG/TTC Default Probability
Band >=Min Mid <Max
1 0.00% 0.010% 0.02%
2 0.02% 0.025% 0.03%
3 0.03% 0.040% 0.05%
4 0.05% 0.075% 0.10%
5 0.10% 0.125% 0.15%
6 0.15% 0.175% 0.20%
7 0.20% 0.225% 0.25%
8 0.25% 0.275% 0.30%
9 0.30% 0.350% 0.40%
10 0.40% 0.450% 0.50%
11 0.50% 0.550% 0.60%
12 0.60% 0.900% 1.20%
13 1.20% 1.375% 1.55%
14 1.55% 1.850% 2.15%
15 2.15% 2.600% 3.05%
16 3.05% 3.750% 4.45%
17 4.45% 5.400% 6.35%
18 6.35% 7.500% 8.65%
19 8.65% 10.000% 11.35%
20 11.35% 15.000% 18.65%
21 18.65% 30.000% 100.00%

Expert lender models are used for parts of the portfolio where the risk drivers are specific to a particular counterparty, but where there is insufficient data to support the construction of a statistical model. These models utilise the knowledge of credit experts that have in depth experience of the specific customer type being modelled. Where possible, the characteristics identified by the expert lenders for use in these models are linked during the modelling process to the Merton framework. This linkage ensures that the model is intuitive and that there is some economic rationale for the default process that is being captured by the model.

For any of the portfolios where we have a low number of default observations we adopt specific rules to ensure that the calibration of the model meets the Basel II and FSA criteria for conservatism. We have developed our own internal policy which describes specific criteria for the use of parametric (e.g. Pluto Tasche) and non-parametric low default portfolio calibration techniques.

Statistical models such as behavioural and application scorecards are used for our high volume portfolios such as SME . The model builds typically incorporate the use of large amounts of internal data, combined with supplemental data from external data suppliers. Where external data is sourced to validate or enhance internally-held data as part of the risk assessment process or to support model development and BAU operation, a similar approach is adopted towards ensuring data quality to that applied to the management of internal data. This entails adherence to the Group’s procurement and supplier management process, including the agreement of specifications and service level agreements.

Typically, modellers do not manipulate external data before using it as input to the model estimation or validation procedure. Changes required in the estimation and validation process are documented in the model build papers.

For all the above asset classes we use the Basel II definition of default, utilising the 90 day past due criteria as the final trigger of default.

Our retail banking operations have long and extensive experience of using credit models in assessing and managing risk in their businesses and as a result models play an integral role in retail approval and customer management (e.g. limit setting, cross-sell etc.) processes.

Models used include application and behavioural scorecards and/or PD/LGD and EAD models. These may be used in isolation, in combination to produce measures of forecast loss or as part of a suite of models that underpin risk/reward based decisions. The score cut-offs will be set at the appropriate level depending on the specific objective, such as ensuring all the accepted accounts meet the minimum required return on EC. It is Barclays philosophy to embed the Basel models as extensively as possible in the portfolio management process. This is an ongoing initiative and we expect greater convergence over time.

In line with Basel II requirements, Barclays will use all available relevant data, including data relating to other Barclays accounts and external agency data. Barclays does not use pooled data.

Most retail models within Barclays are built in-house, although occasionally external consultants will be contracted to build models on behalf of the businesses. Whilst most models are statistically derived, some expert lender models are used, particularly where data limitations preclude a more sophisticated approach. For mortgage originations Barclays use a third party scorecard (Omniscore), supported by a series of policy rules, to arrive at a lending decision.

All new models, including third party models, are measured against the required Group minimum standards as detailed in the Barclays Model Risk Policy.

For retail asset classes, Basel II specifies that the definition of default must include a trigger based on days past due, with the number of days being between 90 and 180. All Barclays advanced internal ratings-based models are compliant with this, with the majority using 180 days as the trigger. In all cases LGD models are specified so that they have a definition of default aligned to that used in the corresponding PD model.

The control mechanisms for the rating system

Each of the business risk teams is responsible for the design, oversight and performance of the individual credit rating models – PD, LGD and EAD – that comprise the credit rating system for a particular customer within each asset class. Group-wide standards in each of these areas are set by Group Risk and are governed through a series of committees with responsibility for oversight, modelling and credit measurement methodologies.

Through their day-to-day activities, key senior management in Group Credit Risk, the businesses and the business risk teams have a good understanding of the operation and design of the rating systems used. For example:

  • – The respective Business Risk Heads or equivalents are responsible for supplying a robust rating system.
  • – The Group Risk Director, Credit Risk Director and Wholesale and Retail Credit Risk Directors are required to understand the operation and design of the rating system used to assess and manage credit risk in order to carry out their responsibilities effectively. This extends to the Business CEOs, Business Risk Directors and the Commercial/Managing Directors or equivalent.

In addition, Group Model Risk Policy requires that all models be validated as part of the model build (see page 89). This is an iterative process that is carried out by the model owner. Additionally, a formal independent review is carried out after each model is built to check that it is robust, meets all internal and external standards and is documented appropriately. These reviews must be documented and conducted by personnel who are independent of those involved in the model-building process. The results of the review are required to be signed off by an appropriate authority.

In addition to the independent review, post implementation and annual reviews take place for each model. These reviews are designed to ensure compliance with policy requirements such as:

  • – integration of models into the business process
  • – compliance with the model risk policy
  • – continuation of a robust governance process around model data inputs and use of outputs

Model performance is monitored regularly; frequency of monitoring is monthly for those models that are applicable to higher volume or volatile portfolios, and quarterly for lower volume or less volatile portfolios. Model monitoring can include coverage of the following characteristics: utility, stability, efficiency, accuracy, portfolio and data.

Model owners set performance ranges and define appropriate actions for their models. As part of the regular monitoring, the performance of the models is compared with these operational ranges. If breaches occur the model owner reports these to the approval body appropriate for the materiality of the model. The model approver is responsible for ensuring completion of the defined action, which may ultimately be a complete rebuild of the model.

Risk tendency

In 2007, Risk Tendency increased 4% (£95m) to £2,355m (2006: £2,260m), significantly less than the 23% growth in the Group’s loans and advances balances. This relatively small rise in Risk Tendency reflected, in particular, the improving risk profile of the UK unsecured loan book. Other factors influencing Risk Tendency included: methodology changes in Barclaycard, UK Retail Banking and International Retail and Commercial Banking – Absa; the sale of the Monument portfolio; and a maturing credit risk profile in the international card portfolios.

UK Retail Banking Risk Tendency decreased £30m to £470m (2006: £500m). This reflected an improvement in the credit risk profile in the UK unsecured consumer lending portfolios, partially offset by the impact of methodology changes and asset growth.

Risk Tendency in Barclays Commercial Bank increased £15m to £305m (2006: £290m). This reflected some growth in loan balances offset by improvements in the credit risk profile.

Barclaycard Risk Tendency decreased £190m to £945m (2006: £1,135m). This reflected improvement in the credit risk profile of UK cards, the sale of part of the Monument portfolio and methodology changes in UK cards, partially offset by asset growth in the international portfolios.

Risk Tendency at International Retail and Commercial Banking – excluding Absa increased £145m to £220m (2006: £75m), reflecting an increase to the risk profile and balance sheet growth in Emerging Markets and Western Europe.

In International Retail and Commercial Banking – Absa, the increase of £110m in Risk Tendency to £255m (2006: £145m) included a change to the methodology following the introduction of Basel compliant, PD, EAD and LGD models. Excluding this change, Risk Tendency increased £90m, reflecting a weakening of retail credit conditions in South Africa after a series of interest rate rises in 2006 and 2007 and balance sheet growth.

Risk Tendency in Barclays Capital increased £45m to £140m (2006: £95m) primarily due to drawn leveraged loan positions. The drawn liquidity facilities on ABS CDO Super Senior positions are classified as credit risk loans and therefore no Risk Tendency is calculated on them.

Since Risk Tendency and impairment allowances are calculated for different parts of the portfolio, for different purposes and on different bases, Risk Tendency does not predict loan impairment.

Credit risk mitigation

The Group uses a wide variety of techniques to reduce credit risk on its lending. The most basic of these is performing an assessment of the ability of a borrower to service the proposed level of borrowing without distress. In addition, the Group commonly obtains security for the funds advanced, such as in the case of a retail or commercial mortgage, a reverse repurchase agreement, or a commercial loan with a floating charge over book debts and inventories. The Group ensures that the collateral held is sufficiently liquid, legally effective, enforceable and regularly valued. Various forms of collateral are held and commonly include cash in major currencies; fixed income products including government bonds; Letters of Credit; property, including residential and commercial; and other fixed assets. For further discussion concerning credit risk mitigation, see credit risk Note .

The Group actively manages its credit exposures and when weaknesses in exposures are detected – either in individual exposures or in groups of exposures – action is taken to mitigate the risks. These include steps to reduce the amounts outstanding (in discussion with the customers, clients or counterparties if appropriate), the use of credit derivatives and, sometimes, the sale of the loan assets. (Credit derivatives may also be traded for profit; details of these activities may be found on page 105 and Note to the accounts).

The Group also uses various forms of specialised legal agreements to reduce risk, including netting agreements which permit it to offset positive and negative balances with customers in certain circumstances to minimise the exposure at default, financial guarantees, and the use of covenants in commercial lending agreements.

Barclays manages the diversification of its portfolio to avoid unwanted credit risk concentrations. A concentration of credit risk exists when a number of counterparties are engaged in similar activities and have similar economic characteristics that would cause their ability to meet contractual obligations to be similarly affected by changes in economic or other conditions.

Credit risk mitigation to address concentrations takes several dimensions. Maximum exposure guidelines are in place relating to the exposures to any individual counterparty. These permit higher exposures to highly rated borrowers than to lower rated borrowers. They also distinguish between types of counterparty, for example, between sovereign governments, banks and corporations. Excesses are considered individually at the time of credit sanctioning, are reviewed regularly, and are reported to the Risk Oversight Committee and the Board Risk Committee.

Notes
a
Head office functions and other operations comprises discontinued businesses in transition.
b
Excludes ABS CDO Super Senior positions as these are classified as credit risk loans and therefore no Risk Tendency is calculated on them.

The Risk Oversight Committee has delegated and apportioned responsibility for risk management to the Retail and Wholesale Credit Risk Management Committees.

The Retail Credit Risk Management Committee (RCRMC) oversees exposures, which comprise unsecured personal lending (including small businesses), mortgages and credit cards. The RCRMC monitors the risk profile and performance of the retail portfolios by receipt of key risk measures and indicators at an individual portfolio level, ensuring mitigating actions taken to address performance are appropriate and timely. Metrics reviewed will consider portfolio composition at both an overall stock and new flow level.

The Wholesale Credit Risk Management Committee (WCRMC) oversees wholesale exposures, comprising lending to businesses, banks and other financial institutions. The WCRMC monitors exposure by country, industry sector, individual large exposures and exposures to sub-investment grade countries.

Country concentrations are addressed through the country risk policy, which specifies Risk Appetite by country and avoids excessive concentrations of credits in individual countries. Country risk grades are assigned to all countries where the Group has, or is likely to have, exposure and are reviewed regularly to ensure they remain appropriate.

Country grades, which are derived from long-term sovereign foreign currency ratings, range from 1 (lowest probability of default) to 21 (highest probability of default). A ceiling is applied where a country is graded 12 or worse so that the counterparty cannot receive a higher risk grading than the country, unless some form of protection is available in the event of a cross-border event, such as a significant portion of a counterparty’s assets or income being held or generated in hard currency.

To manage exposure to country risk, the Group uses two country limits: the Prudential Guideline and the Country Guideline. The Prudential Guideline is identified through the strict mapping of a country grade to derive a model-driven acceptable level of loss given default. The Country Guideline for all graded countries is set by the Group Credit Committee (GCC) based on the Prudential Guideline and the internal appetite for country risk. The Country Guideline may therefore be above or below the Prudential Guideline.

Country risk is managed through the application of Country Loss Given Default (CLGD). All cross-border or domestic foreign currency transactions incur CLGD from the Country Guideline agreed at GCC. The level of CLGD incurred by a counterparty transaction will largely depend on three main factors: the country severity, the product severity and counterparty grade.

CLGD is incurred in the country of direct risk, defined as where the majority of operating assets are held. This may differ from the country of incorporation. However, where transactions are secured with collateral, the country risk can be transferred from the country of the borrower to the country of the collateral provider. This is only permitted where the collateral covers the borrowing and is not expected to decrease over time.

Country Managers are in place for all countries where the Group has exposure and they, under the direction of GCC, have responsibility for allocating country risk to individual transactions. The total allocation of country limits is monitored on a daily basis by Group Credit Risk, as headed by the Group Credit Risk Director. Discretions exist to increase the Country Guideline above the level agreed by GCC where the Country Guideline is below the Prudential Guideline. All requests to increase the Country Guideline in line with individual discretions must be submitted to and applied centrally through Group Credit Risk.

A further mitigant against undesirable concentration of risk is the mandate and scale framework described on page 86. Mandate and scale limits, which can also be set at Group level to reflect overall Risk Appetite, can relate either to the stock of current exposures in the relevant portfolio or to the flow of new exposures into that portfolio. Typical limits include the caps on UK commercial investment property lending, the proportion of lending with maturity in excess of seven years and the proportion of new mortgage business that is buy-to-let.

Concentrations of credit exposure described in this credit risk management section and the following statistical section are not proportionally related to credit loss. Some segments of the Group’s portfolio have and are expected to have proportionally higher credit charges in relation to the exposure than others. Moreover, the volatility of credit loss is different in different parts of the portfolio. Thus, comparatively large credit impairment charges could arise in parts of the portfolio not mentioned here.

Securitisations

In the course of its business, Barclays undertakes securitisations of its own originated assets as well as the securitisation of third party assets via sponsored conduit vehicles and shelf programmes.

Barclays securitises its own originated assets in order to remove risk from the Group’s credit position, to obtain regulatory capital relief, and to obtain term liquidity for the Group balance sheet.

For these transactions Barclays adopts the following roles in the securitisation process:

  • – Originator of securitised assets
  • – Executor of securitisation trades including bond marketing and syndication
  • – Provider of securitisation trade servicing, including data management, investor payments and reporting

Barclays also acts as an administrator and manager of multi-seller conduits through which interests in third-party-originated assets are securitised and funded via the issuance of asset backed commercial paper.

In relation to such conduit activity, Barclays may also provide all or a portion of the backstop liquidity to the commercial paper, programme-wide credit enhancement and, as appropriate, interest rate and foreign currency hedging facilities.

RWAs reported for securitised assets as at December 2007 are calculated in line with rules set out in IPRU (BANK) as well as any individual guidance received from the FSA as at the end of this period.

As of 1st January 2008, Barclays calculates securitisation RWAs using the ratings based approach and/or the supervisory formula method as per the FSA’s revised rules, which implement the Basel Accord and Capital Requirements Directive.

Further information about securitisation activities and accounting treatment is in Note . The Group’s accounting policies, including those relevant to securitisation activities (policies 4 and 10), are on pages 165 and 168.

Barclays employs External Credit Assessment Institutions to provide ratings for its asset backed securities. Their use is dependant on the transaction or asset class involved. For existing transactions, we employ Standard & Poor’s, Moody’s and Fitch for securitisations of corporate, residential mortgage and other retail exposures and Standard & Poor’s and Moody’s only for securitisations of small and medium-sized entity and revolving retail exposures.

Risk management

Credit risk management


Monitoring of loans and advances

As the granting of credit is one of the Group’s major sources of income and its most significant risk, the Group dedicates considerable resources to gaining a clear and accurate understanding of credit risk across the business and ensuring that its balance sheet correctly reflects the value of the assets in accordance with accounting principles. This process can be broken down into the following stages:

  • – Measuring exposures and concentrations
  • – Monitoring weakness in exposures
  • – Identifying potential problem loans and credit risk loans (collectively known as potential credit risk loans or PCRLs)
  • – Raising allowances for impaired loans
  • – Writing off assets when the whole or part of a debt is considered irrecoverable
Measuring exposures and concentrations

Loans and advances to customers provide the principal source of credit risk to the Group although Barclays can also be exposed to other forms of credit risk through loans to banks, loan commitments, contingent liabilities and debt securities; see page 60). The value of outstanding loans and advances balances, their risk profile, and potential concentrations within them can therefore have a considerable influence on the level of credit risk in the Group.

As at 31st December 2007, outstanding loans and advances to customers and banks were valued at £389bn (2006: £317bn), of which £349bn (2006: £286bn) was granted to personal or corporate customers (see figure 1). Loans and advances were well distributed across the retail and wholesale portfolios.

Loans and advances were well spread across industry classifications (figure 2). Excluding Financial Services, Barclays largest sectoral exposures are to home loans, other personal and business and other services. These categories are generally comprised of small loans, have low volatility of credit risk outcomes, and are intrinsically highly diversified.

Balances are also diversified across a number of geographical regions (figure 3, based on location of customers). The majority of Barclays exposure is to the UK, which includes secured home loans exposure, followed by the United States, Africa and the rest of the European Union.

Fig. 1: Loans and advances

Excel File Download table as excel file
2007  2006 
£m  £m 
Retail businesses
Banks – –
Customers 164,062 139,350
Total retail businesses 164,062 139,350
Wholesale businesses
Banks 40,123 30,930
Customers 185,105 146,281
Total wholesale businesses 225,228 177,211
Loans and advances 389,290 316,561

Barclays risk is therefore spread across a large number of industries and customers and in the case of home loans, for example, well secured. These classifications have been prepared at the level of the borrowing entity. This means that a loan to the subsidiary of a major corporation is classified by the industry in which the subsidiary operates, even through the parent’s predominant sphere of activity may be in a different industry.

UK exposure to home loans accounts for just over 60 per cent of the Group’s total home loans exposure. The loan-to-value ratios (LTV) on the Group’s UK home loan portfolio are shown in figure 4. The valuations in the chart are those which applied at the last credit decision on each loan, i.e. when the customer last requested an increase in the limit or, if there has been no increase, at inception of the loan. Business flows (new business versus loans redeemed) have not materially changed the risk profile of the portfolio.

The impact of house price inflation will result in a reduction in LTV ratios within the mortgage book on a current valuation basis. On this basis, LTV on the residential mortgage book averaged 33% at the end of 2007 (2006: 34%). This ratio is a point-in-time analysis of the stock with LTV updated to current house prices by reference to an external price index and as a result may be influenced by external market conditions as well as changes in the stock of loans.

Barclays also actively monitors the risk profile of its loans and advances to customers, with a view to the early detection of any concentrations in higher risk segments. Figure 5 depicts Barclays wholesale loan profile by existing risk grade (see page 82 for a description of the rating system). The majority of Barclays exposure is to the higher quality names with just under 70% of exposure to customers with a DG of 10 or better. It is important to note that Barclays prices loans to risk. Thus, higher risk loans will usually have higher interest rates or fees or both. The profitability of a higher risk portfolio may, therefore, equal or exceed that of a lower-risk portfolio.

Barclays also actively monitors exposure and concentrations to sub-investment grade countries (see country risk policy, page 81). Details of the 15 largest sub-investment grade countries, by limit, are shown in figure 6.

Contractual maturity represents a further area of potential concentration. The analysis shown in figure 7 indicates that just over 40% of loans to customers have a maturity of more than five years; the majority of this segment comprises secured home loans.





Monitoring weaknesses in exposures

Barclays actively manages its credit exposures. Corporate accounts that are deemed to contain heightened levels of risk are recorded on graded early warning or watch lists comprising three categories of increasing concern. These are updated monthly and circulated to the relevant risk control points. Once listing has taken place, exposure is very carefully monitored and, where appropriate, exposure reductions are effected. Should an account become impaired, it will normally, but not necessarily, have passed through all three categories, which reflect the need for ever-increasing caution and control.

Where an obligor’s financial health gives grounds for concern, it is immediately placed into the appropriate category. All obligors, regardless of financial health, are subject to a full review of all facilities on, at least, an annual basis. More frequent interim reviews may be undertaken should circumstances dictate.

Within Local Business, accounts that are deemed to have a heightened level of risk, or that exhibit some unsatisfactory features which could affect viability in the short/medium term, are transferred to a separate ‘Caution’ stream. Accounts on the Caution stream are reviewed on at least a quarterly basis at which time consideration is given to continuing with the agreed strategy, returning the customer to a lower risk refer stream, or instigating recovery/exit action.

Within the personal portfolios, which tend to comprise homogeneous assets, statistical techniques more readily allow potential weaknesses to be monitored on a portfolio basis. This applies in parts of UK Retail Banking, Barclays Wealth, International Retail and Commercial Banking and Barclaycard. The approach is consistent with the Group’s policy of raising a collective impairment allowance as soon as objective evidence of impairment is identified.

Potential credit risk loans

If the credit quality of a loan on an early warning or watch list deteriorates to the highest category, consideration is given to including it within the Potential Problem Loan (PPL) list. PPLs are loans where payment of principal and interest is up to date but where serious doubt exists as to the ability of the borrowers to continue to comply with repayment terms in the near future.

Should further evidence of deterioration be observed, a loan may move to the Credit Risk Loan (CRL) category. Events that would trigger the transfer of a loan from the PPL to the CRL category could include a missed payment or a breach of covenant. CRLs comprise three classes of loans:

  • – ‘Impaired loans’ comprise loans where individual impairment allowance has been raised and also include loans which are fully collateralised or where indebtedness has already been written down to the expected realisable value. The impaired loan category may include loans, which, while impaired, are still performing.

CRL and PPL balances by geography



CRLs and PPLs as a percentage of Loans and Advances





Notes
a
In 2003, credit risk loans and potential problem loans were disclosed based on the location of the booking office. In 2004-2007 they were disclosed by location of customers.
b
Does not reflect the application of IAS 32, IAS 39 and IFRS 4 which became effective from 1st January 2005.
c
From 1st January 2005, the application of IAS 39 required interest to be recognised on the remaining balance of an impaired financial asset (or group of financial assets) at the effective interest rate for that asset. As a result, interest is credited to the income statement in relation to impaired loans, therefore these loans technically are not classified as ‘non-accrual’. In 2005, the Group replaced the ‘non-accrual’ category with one termed ‘impaired loans’. The SEC requires loans to be classified, where applicable, as non-accrual, accruing past due 90 days or more, ‘troubled debt restructurings’ and potential problem loans.
  • – The category ‘accruing past due 90 days or more’ comprises loans that are 90 days or more past due as to principal or interest where there is no expectation of ultimate write-off (whether in part or full) of the principal owed. Impairment allowance will be raised against these loans if the expected cash flows discounted at the effective interest rate is less than the carrying value.
  • – The category ‘impaired and restructured loans’ comprises loans not included above where, for economic or legal reasons related to the debtor’s financial difficulties, a concession has been granted to the debtor that it would not otherwise be considered. Where the concession results in the expected cash flows discounted at the effective interest rate being less than the loan’s carrying value, an impairment allowance will be raised.

The term Credit Risk Loans has replaced the term Non-Performing Loans (NPLs) as the collective term for the total of these three classes since it recognises the fact that the impaired loan category may include loans, which, while impaired, are still performing. This category includes drawn ABS CDO Super Senior positions.

Potential Credit Risk Loans (PCRLs) comprise potential problem loans (PPLs) and credit risk loans (CRLs). Figures 8 and 9 show CRL and PPL balances by geography. The amounts are shown before deduction of value of security held, impairment allowances (from 2005 onwards) and provisions or interest suspense (2004 and earlier), all of which might

reduce the impact of an eventual loss, should it occur. The significant increase to non-UK CRL and PPL balances is principally due to the inclusion of US-located ABS CDO Super Senior positions and other credit market exposures.

Figures 12 and 13 show impairment allowances as a percentage of CRLs and PCRLs. Including the drawn ABS CDO Super Senior positions, allowance coverage of CRLs and PCRLs decreased to 39.1% (31st December 2006: 65.6%) and 33.0% (31st December 2006: 57.0%), respectively. These movements reflect the fact that allowance coverage of ABS CDO Super Senior credit risk loans was low relative to allowance coverage of other credit risk loans since substantial protection against loss is also provided by subordination and hedges. On ABS CDO Super Senior exposures, the combination of subordination, hedges and write-downs provided protection against loss levels to 72% on US sub-prime collateral as at 31st December 2007.

Figures 14 and 15 show allowance coverage of CRLs and PCRLs excluding the drawn ABS CDO Super Senior positions decreased to 55.6% (31st December 2006: 65.6%) and 49.0% (31st December 2006: 57.0%), respectively. The decrease in these ratios reflected a change in the mix of CRLs and PCRLs. Unsecured retail exposures, where the recovery outlook is low, decreased as a proportion of the total as the collections and underwriting processes were improved. Secured retail and wholesale and corporate exposures, where the recovery outlook is relatively high, increased as a proportion of PCRLs.





Notes
a
In 2003, credit risk loans and potential problem loans were disclosed based on the location of the booking office. In 2004-2007 they were disclosed by location of customers.
b
Does not reflect the application of IAS 32, IAS 39 and IFRS 4 which became effective from 1st January 2005.
Allowances for impairment and other credit provisions

Barclays establishes, through charges against profit, impairment allowances and other credit provisions for the incurred loss inherent in the lending book.

Under IFRS, impairment allowances are recognised where there is objective evidence of impairment as a result of one or more loss events that have occurred after initial recognition, and where these events have had an impact on the estimated future cash flows of the financial asset or portfolio of financial assets. Impairment of loans and receivables is measured as the difference between the carrying amount and the present value of estimated future cash flows discounted at the financial asset’s original effective interest rate. If the carrying amount is less than the discounted cash flows, then no further allowance is necessary.

Impairment is measured individually for assets that are individually significant, and collectively where a portfolio comprises homogenous assets and where appropriate statistical techniques are available.

In terms of individual assessment, the trigger point for impairment is formal classification of an account as exhibiting serious financial problems and where any further deterioration is likely to lead to failure. Two key inputs to the cash flow calculation are the valuation of all security and collateral and the timing of all asset realisations, after allowing for all attendant costs. This method applies in the corporate portfolios – Barclays Commercial Bank, Barclays Capital and certain areas within International Retail and Commercial Banking and Barclaycard.

For collective assessment, the trigger point for impairment is the missing of a contractual payment. The impairment calculation is based on a roll-rate approach, where the percentage of assets that move from the initial delinquency to default are derived from statistical probabilities based on experience. Recovery amounts and contractual interest rates are calculated using a weighted average for the relevant portfolio. This method applies to parts of International Retail and Commercial Banking, Barclaycard and UK Banking and is consistent with Barclays policy of raising an allowance as soon as impairment is identified.

Unidentified impairment allowances, albeit significantly lower in amount than those reported above, are also raised to cover losses which are judged to be incurred but not yet specifically identified in customer exposures at the balance sheet date, and which, therefore, have not been specifically reported.

The incurred but not yet reported calculation is based on the asset’s probability of moving from the performing portfolio to being specifically identified as impaired within the given emergence period and then on to default within a specified period. This is calculated on the present value of estimated future cash flows discounted at the financial asset’s original effective interest rate.

The emergence periods vary across businesses and are based on actual experience and are reviewed on an annual basis. This methodology ensures that the Group only captures the loss incurred at the balance sheet date.

These impairment allowances are reviewed and adjusted at least quarterly by an appropriate charge or release of the stock of impairment allowances based on statistical analysis and management judgement.

Where appropriate, the accuracy of this analysis is periodically assessed against actual losses.

As one of the controls of ensuring that adequate impairment allowances are held, movements in impairment allowances to individual names above £10m are presented to the Group Credit Committee for agreement.

The Group Credit Risk Impairment Committee (GCRIC), on a semi-annual basis, obtains assurance on behalf of the Group that all businesses are recognising impairment in their portfolios accurately and promptly in their recommendations and in accordance with policy, accounting standards and established governance.

GCRIC exercises the authority of the Barclays Risk Director, as delegated by the Chief Executive, and is chaired by Barclays Credit Risk Director.

GCRIC reviews the movements to impairment in the businesses, including those already agreed at Group Credit Committee, Potential Credit Risk Loans and Risk Tendency.

These committees are supported by a number of Group Policies including: Group Retail Impairment and Provisioning Policy; Group Wholesale Impairment and Provisioning Policy; and, Group Model Policy.

GCRIC makes twice-yearly recommendations to the Board Audit Committee on the adequacy of Group impairment allowances. Impairment allowances are reviewed relative to the risk in the portfolio, business and economic trends, current policies and methodologies and our position against peer banks.

Fig. 16: Impairment charges for bad and doubtful debts

Excel File Download table as excel file
2007 2006 2005
£m £m £m
UK Banking 849 887 671
Barclaycard 838 1,067 753
International Retail and
Commercial Banking 252 167 33
Barclays Capital 846 42 111
Barclays Global Investors – – –
Barclays Wealth 7 2 2
Head office functions and other operations 3 (11) 1
Total impairment charges 2,795 2,154 1,571

Note
a
Does not reflect the application of IAS 32, IAS 39 and IFRS 4 which became effective from 1st January 2005.

GCRIC has delegated the detailed review of loan impairment in the businesses to the Retail and Wholesale Credit Risk Management Committees.

In 2007, total impairment charges on loans and advances and other credit provisions increased 30% (£641m) to £2,795m (2006: £2,154m) reflecting charges of £782m against ABS CDO Super Senior and other credit market positions.

Impairment charges on loans and advances and other credit provisions as a percentage of Group total loans and advances rose to 0.71% (2006: 0.65%); total loans and advances grew by 23% to £389,290m (2006: £316,561m).

Retail impairment charges on loans and advances fell 11% (£204m) to £1,605m (2006: £1,809m). Retail impairment charges as a percentage of period-end total loans and advances reduced to 0.98% (2006: 1.30%); total retail loans and advances rose by 18% to £164,062m (2006: £139,350m).

Barclaycard impairment charges improved £229m (21%) to £838 (2006: £1,067m) reflecting reduce flows into delinquency, lower levels of arrears and lower charge-offs in UK Cards. We made changes to our impairment methodologies to standardise our approach and in anticipation of Basel II. The net positive impact of these changes in methodology was offset by the increase in impairment charges in Barclaycard International and secured consumer lending.

Impairment charges in UK Retail Bank decreased by £76m (12%) to £559m (2006: £635m), reflecting lower charges in unsecured Consumer Lending and Local Business driven by improved collection processes, reduced flows into delinquency, lower trends of arrears and stable charge-offs. In UK Home Finance, asset quality remained strong and mortgage charges remained negligible. Mortgage delinquencies as a percentage of outstandings remained stable and amounts charged-off were low.

Impairment charges in International Retail and Commercial Banking – excluding Absa rose by £38m (93%) to £79m (2006: £41m) reflecting very strong balance sheet growth in 2006 and 2007 and the impact of lower releases in 2007.

Arrears in some of International Retail and Commercial Banking – Absa’s key retail portfolios deteriorated in 2007, driven by interest rate increases in 2006 and 2007 resulting in pressure on collections.

Wholesale and corporate impairment charges on loans and advances increased £436m to £701m (2006: £265m). Wholesale and corporate impairment charges as a percentage of period-end total loans and advances increased to 0.31% (2006: 0.15%); total loans and advances grew by 27% to £225,228m (2006: £177,211m).

Barclays Capital impairment charges and other credit provisions of £846m included a charge of £782m against ABS CDO Super Senior and other credit market exposure and £58m relating to drawn leveraged finance positions.

The impairment charge in Barclays Commercial Bank increased by £38m (15%) to £290m (2006: £252m), primarily due to higher gross impairment charges in Larger Business, partially offset by a lower charge in Medium Business due to a tightening of the lending criteria.

Writing-off of assets

After an advance has been identified as impaired and is subject to an impairment allowance, the stage may be reached whereby it is concluded that there is no realistic prospect of further recovery. Write-off will occur, when, and to the extent that, the whole or part of a debt is considered irrecoverable.

The timing and extent of write-offs may involve some element of subjective judgement. Nevertheless, a write-off will often be prompted by a specific event, such as the inception of insolvency proceedings or other formal recovery action, which makes it possible to establish that some or the entire advance is beyond realistic prospect of recovery. In any event, the position of impaired loans is reviewed at least quarterly to ensure that irrecoverable advances are being written off in a prompt and orderly manner and in compliance with any local regulations.

Such assets are only written off once all the necessary procedures have been completed and the amount of the loss has been determined.

Subsequent recoveries of amounts previously written off are written back and hence decrease the amount of the reported loan impairment charge in the income statement.

Total write-offs of impaired financial assets decreased by £211m to £1,963m (2006: £2,174m).

Note
a
Does not reflect the application of IAS 32, IAS 39 and IFRS 4 which became effective from 1st January 2005.