In this chapter we introduce the credit risk measurement framework in a banking operation. Credit‐risky lending is the core business of a bank and its key to profitability. It is also the biggest driver of a bank's regulatory capital requirement. As credit risk exposure is what banks do, it cannot be avoided, so it must be well managed. Strategy, decision‐making, risk–reward optimisation, diversification, and minimisation of loss are not possible without extensive and thorough credit risk management. The bank must use all of its qualitative and quantitative judgement capabilities to best assess credit risk.
This chapter is a long one and focuses on the risk management process and its principles. Regulatory capital requirements, which are driven primarily by credit risk exposure, are considered in Chapter 15.
The business of banking – lending money to and transacting with counterparties who carry default risk – creates credit risk exposure on the bank's balance sheet. This must be managed actively. In many cases, once a loan is originated, it cannot be removed from the balance sheet, so the discipline of credit risk management is essentially one of trying one's hardest to get the loan origination decision right, and avoiding concentration.
The other side of the approach to credit risk management is to sell loans where possible, to remove them from the balance sheet via securitisation, or to use credit derivatives. This topic is covered in detail in Choudhry (2007).
Banks generally operate one of two types of approval process: (i) via a credit committee; or (ii) via delegated authority from the credit committee to a business line head. The committee process is designed to ensure that there is proper scrutiny of any transaction that commits the bank's capital. The sponsor bringing the transaction to the committee is the front office business line; the committee will approve or decline based on the risk–reward profile of the transaction.
Procedure (ii) is common for high‐volume business, for which the committee process as a consequence of it being time consuming would not be practical. As we note above, there is uncertainty that the “know your risk” principle can be diluted, particularly in a competitive environment where a bank is trying to build volume. Given this uncertainty, “market share” should not be a performance indicator, or target, for a bank's business line. Rather, performance should be measured only via the amount of genuine shareholder value added that the business generates.
The point of credit risk limits is to set an upper bound to the loss that can be suffered by a bank at any one time.
The basic principles of credit limit setting are universal for every bank and follow the essential requirements of prudence and concentration. An element of diversification in the loan portfolio is necessary, although at all times the bank should practise the basic principle of “know your risk”. In other words, diversity as an end in itself is not recommended good practice; a bank should diversify only into sectors that it thoroughly understands and in which it has some competitive advantage or valuable skill base.
In standard textbooks on finance and banking, we read that it is the capital base that drives the limit‐setting process. Essentially, what this is saying in practice is that we take the amount of capital available and allocate it as per credit limit buckets for each of the businesses. Actually, the proper and intellectually robust way to do this is the other way around: the bank should determine its strategy and business model, as well as preparing budgets based on the risk exposure that it considers it has the expertise to manage. This process then drives the level of capital and regulatory capital that the bank should then set up. Once this amount is known and achieved, it can then be allocated to specific business lines as lower level credit limits by geography, industry, product, and so on.
The essential principles governing limit setting include the following:
As part of a transaction origination process, reviewers must consider what “ancillary business” can be generated from the same borrower. The bank must set a policy that dictates how much this ancillary business drives the origination process, whether the lending business can be a “loss leader” to an extent or can create sufficient shareholder value added in its own right.
The process of setting credit limits is very important to all banks – vanilla commercial banks, in particular – insofar as credit risk exposure generates the highest losses for such institutions. The process should follow prudent and robust policy and be run according to cycle‐proof principles to avoid getting overextended during a bull market, when loan origination standards are relaxed. Credit limits are set for a range of criteria, which are deliberately set as overlapping so as to ensure that all the various different categories of risk exposure are captured.
Macro‐level credit limits are set per individual obligor, originated within the business lines but approved by the Executive Credit Committee and secondarily approved by the asset and liability committee (ALCO).1 When necessary, if the size of a transaction dictates it, further approval may be needed by the Executive Management Committee (ExCo) and the Board itself. The level of capital allocation required for a particular limit application determines how far up the governance structure it needs to go. Formal limits on capital allocation are therefore set at ExCo approval level.
The limit‐setting process is designed to produce overlapping limits. Limits will be set in the following categories:
Limit excess is a serious breach of management governance and must be reported to ALCO (and, if necessary, ExCo) for corrective action. This can be effected by one or more of the following: (i) cease further business with the specific obligor; (ii) transfer some of the exposure, either by secondary market sale, securitisation, or hedging with credit derivatives; (iii) increase the limit; or (iv) transfer some capacity from another part of the business and/or another obligor.
The loan origination process differs across banks. The detail of an individual specific process is not of major interest to us. What is important is that this origination process adheres to basic principles of prudence, and that these are controlled and managed to ensure they are “through the cycle”. That is, a reduction in standards, or a relaxation of standards during a period of economic growth, is something that should require Board approval. Enlarging the balance sheet during a bull market is a risky strategy, because it is during this time that standards are lowered and low‐quality and/or underpriced assets are put on the book.
An example of this occurred at the failed UK banks Northern Rock and Bradford & Bingley, which originated large numbers of 100LTV and 125LTV mortgages, as well as more risky buy‐to‐let mortgages. The failed bank HBOS (in common with many banks at the time) operated a loan origination process for retail and corporate loans that delegated the approval decision to a black box computer model, which rated all applications in a tick box process that assigned a credit score and then approved on that basis. This is understandable for high‐volume business models, but sacrifices a large element of “know your customer” in the approval process.
The essential guidelines for a through‐the‐cycle asset origination standards process include:
Excluding the peak of an overheating economy just about to enter a recession, loan defaults typically do not occur at the start or end of a loan's term. Another exception is right at the end of a bull market, when bank loan origination standards have been lowered and asset prices (credit spreads) are at their most undervalued, when banks write much low‐quality business. Leaving that aside, the most common time of default is generally between 45 and 55 months after the loan start date. This means that default statistics lag considerably the actual state of the economy. Given historical default rates, which banks use to assist them in setting their credit limits, there is a danger that business continues to be written at lower credit standards at the time when the bank should be reigning in risky business. This is why the basic principles we summarise above should be observed at all times; they should act as a guiding light for a bank's Executive Credit Committee.
Financial information, metrics, and analysis are central to the extension of credit. Banking is a customer business, however, and ideally decision makers must balance what numbers and models indicate with their own judgements. This approach is not followed by every bank, many of which follow the model‐based approach in isolation. Ideally though, qualitative factors, despite being less tangible and harder to quantify, are crucial. There is no substitute for “know your customer”.
Traditional banking was very much an “expert system”, where bankers allocated credit in their sectors using subjective judgements. A popular framework was the “five Cs” – character, capital, capacity, collateral, and cycle – weighted as deemed appropriate. Without rigorous analytics, credit was sometimes extended to clients sponsored by the relationship banker who shouted loudest. But, again, soft factors are important and two broad areas for consideration are management and business.
Management needs technical and organisational skills to succeed. The team must understand their business, demonstrate adaptability to changing environments, and have the capacity to control risk and act decisively. Do they have the ability to execute their plan? Experience and background checks should be completed to assess past track records of key leaders.
Consistency of message is important and progress in implementing plans can be checked against past annual reports and press releases. The quality and thoroughness of financial reporting are paramount. Face‐to‐face meetings are highly desirable and an opportunity to ask probing questions. The openness of responses is a good indicator as to whether a good working relationship is possible, through both good and difficult times. Will problems be disclosed promptly, so that they can be worked out with action taken?
Corporate governance must be examined so that the relationships and responsibilities between management and directors are understood. Good corporate governance ensures that proper checks and balances and protections are in place in the interests of investors, lenders, and other stakeholders, and also against unethical and illegal activities. Policies are defined and determined in the company charter and its bylaws, along with corporate rules and regulations. Strong internal controls, authorisation and approval procedures, and the independence of the internal audit department are vital.
As a first step, a lender must understand the basics of its client's business. Beyond knowing the company's business sector, exactly what does it provide, and in so doing, how does it make money? In other words, what is the business model? If a bank cannot meet the company and attain a clear understanding of its business and external operating environment, it does not make sense to extend credit to it.
Success and long‐term viability often depend on competitive advantage or franchise value. Is the company innovative, unique, or efficient? Does it benefit from a strong brand? Does it have sufficient market share? How high are the barriers for new entrants to challenge its position?
A bank needs to develop informed views on industry sectors and regions. A company can have a strong financial profile and top market share, but a weak or volatile business environment will reduce its credit quality. The state of the economy in its markets, diversification of buyer base, regulation, labour markets, and the pace and depth of structural change are all key factors.
For many years, retail credit decisions were made by local bank relationship managers based on qualitative factors. Models used to analyse wholesale loans could not be applied to small, unrated borrowers. Data has been expensive to collect and verify, with the cost amplified given small potential profit. Often, sound retail credit decision frameworks have been developed by smaller banks with dedicated resources, local knowledge, and relationships. Part of the decision should be based only on a borrower's ability to pay, but also on their sense of obligation to pay. Borrowers may have a local business and range of other debts (credit cards, mortgages, loans) to consider.
Strictly subjective retail credit decisions allow for inconsistency, and sometimes bias on the basis of race, religion, national origin, gender, or marital status. Perceptions as to the most favourable types of employment could also lead to less rational decisions.
Ratings act as the basis for bank credit approval, pricing, monitoring, and loan loss provisioning. Whereas external ratings agencies were founded in the nineteenth century (mostly to analyse US railroads), bank internal ratings only took off in the 1990s with Basel I.
The Basel Committee defines a rating as a “summary indicator of the risk inherent in an individual credit”. Ratings “typically embody an assessment of the risk of loss due to failure by a given borrower to pay as promised”. A rating system is defined as “the conceptual methodology, management processes, and systems that play a role in the assignment of a rating”. Ratings have two dimensions: (i) borrower propensity to default; and (ii) transaction characteristics (for example, product, terms, seniority, and collateral).
Under the Basel “Standardised Approach”, banks use ratings developed by external rating agencies. The starting point for capital required against assets is 8% of the nominal, against which a risk weighting is applied according to a matrix of ratings and types (sovereign, bank, or corporate). An AAA‐rated sovereign asset with a 0% risk weight requires no capital. A $100 million A‐rated bank asset with a 20% risk weight requires 1.6% or $1.6 million of capital ($100 million × 8% × 20%).
A bank loan pricing model is based on the capital required and the target return on that capital. The model considers funding, default probability, recovery rate, and taxes as costs against the interest rate charged in calculating a net gain to weigh against the capital required. If loan rates in the marketplace are below the rate needed to meet the target return, the bank needs to decide whether it will accept a lower return to maintain market share or the client relationship.
Since Basel II, banks meeting strict criteria are permitted to use internal ratings to calculate regulatory capital for credit risk. The rationale is that “internal ratings can prove to be more sensitive to the level of risk in a bank's portfolio”. Internal ratings may incorporate supplementary customer information, which is usually out of the reach of external credit assessment institutions. Banks will (have incentives to) further refine internal credit risk management and measurement techniques.
Internal ratings must satisfy the “use test” and serve as the basis of risk, limits, pricing, provisioning, and capital management decisions, and not be simply for regulatory risk capital calculations.
Both Basel I and Basel II were designed to ensure that banks maintain an adequate capital buffer based on an expected loss methodology. The standard credit loss profile, illustrated in Figure 3.1, lies behind the capital calculation.
Figure 3.2 shows how this general credit losses distribution drives the minimum capital level. Expected losses are covered in the loan asset pricing, typically via the target rate of return (given by the cost of capital). Unexpected losses are covered by the capital reserves, shown in Figure 3.2 as economic capital. Of course, regulatory capital rules assign risk weightings based on the type of loan asset counterparty, and these risk weightings drive the regulatory capital minimum. In the Basel II regime, credit ratings are used to determine risk weightings for banks that use the standardised approach.
Where losses exceed the unexpected losses used to derive the capital calculation, the shortfall results in cessation of the bank as a going concern.
In theory (and in practice if one was putting together a bank's capital structure from scratch today), the expected loss profile effectively drives the capital structure. As Figure 3.3 shows, if the UL is 10%, then this is the minimum equity base requirement.
Critics of Basel II believed that while giving banks more responsibility to measure risk creates greater focus and better analysis, it opened up possibilities for opportunistic behaviour and regulatory arbitrage.
The dominant practice of banks is to use ratings to manage corporate credit risk as described. Ratings remain relatively constant, and are often linked to a schedule of average default probabilities. Rating mobility is a function of a bank's philosophy, which can be either Through‐the‐Cycle (TtC – more active migration) or Point‐in‐Time (PIT – less migration).
The quality of the portfolio shifts as the distribution of ratings evolves.
Retail exposures are not generally managed using ratings on an individual borrower basis. Exposures are grouped into segments with similar risk characteristics. Often, the distinction between borrower and product is limited or eliminated. Borrower characteristics (for example, population segment, income, credit history) and those of the facility (for example, product type, credit limit, collateral) are blended in formulating segments. To demonstrate homogeneity of risk, genuine segmentation requires all borrowers within a segment to be treated the same.
Retail exposures include loans (for example, personal finance, education loans, auto loans, or leasing), revolving credits (for example, overdrafts, revolving credit plans, or home equity lines), credit cards, residential mortgages, and small business facilities. Typically (and from a BIS regulatory stand point) retail exposures are managed on a product (facility) basis as opposed to an obligor basis.
There are four tests for characterisation as retail exposure:
For retail exposures, delinquent exposures are managed separately.
In specialised lending, both the source of repayment of a loan and prospects for recovery in an event of default are based on the cash flow from a project or property rather than on the ongoing, open‐ended operations of the borrower. Assets pledged as collateral serve to mitigate risk and as a secondary source of repayment. Types of specialised lending include project finance, income producing real estate, high volatility commercial real estate, object finance, and commodities finance.
Specialised lending possesses unique loss distribution and risk characteristics. Given the source of repayment, the exposures exhibit greater risk volatility, with both high default rates and high loss rates in times of distress. Banks use different internal risk rating criteria. Historical data is often not as readily available or comparable and relevant to the current special financing exposures being assessed.
A special risk management focus for specialised lending includes financial strength and flexibility, collateral control, project phase, and marketability.
As described earlier, the largest rating agencies have been performing analyses and collecting data for well over a century. Rating agencies have large teams and resources, and regular contact with the entities they rate. However, while external ratings are undoubtedly useful, they are no more than opinions, and are not a substitute for analysis and informed decision‐making. Banks with borrower relationships may have insights that lead them to assign different ratings from those of the agencies.
Rating agencies serve a wide range of constituencies that use ratings for differing purposes. As such, there are a range of opinions on their process. Some want ratings to be adjusted quickly to give signals of possible deterioration, while others want them to be more stable and uninfluenced by short‐term developments. The role of the economic cycle in ratings is often debated.
Rating agencies rely on the accuracy and completeness of information supplied by borrowers, and do not search for or expose fraud. Analysts concluding that available data raises some questions can assign more conservative ratings or fail to assign or withdraw ratings, but users cannot rely on rating agencies for more.
Some are of the opinion that ratings are biased to the upside, given that borrowers pay for their ratings.
After assigning ratings, banks estimate risk parameters for key exposures including probability of default (PD), loss given default (LGD), and exposure at default (EAD). Exposures are risk weighted by Basel‐mandated asset class (corporate, sovereign, bank, retail, and equity) to arrive at total risk weighted assets (RWA) to determine capital requirements.
Probability of default (PD) describes the likelihood of default over a particular time horizon. PD estimates are derived from internal default experience, mapping to external data, and statistical default models. Except for pooled retail exposures, PD for a particular grade must be a long‐run average of 1‐year default rates. Retail PD estimates must be derived primarily from internal data.
Expected loss given default (ELGD or LGD) is defined as the “economic” loss for non‐defaulted assets, accounting for inflows (via sale of supporting collateral, unsecured recoveries, and guarantor payments) and outflows (via additional post‐default drawdowns on the credit facility, internal administrative costs, and external legal and valuation fees) measured relative to the exposure at default.
An “economic” loss (unlike an accounting loss) takes into account all relevant factors including material discount effect, and material direct and indirect costs associated with holding and collecting the defaulted facilities.
Methods used to estimate ELGD for credit facilities fall into one of two categories. Subjective methods are primarily driven by expert judgement and used mainly on portfolios with few defaults and/or by banks in the early stage of internal model development. Objective methods largely rely on formal mathematical procedures, and can be further divided into explicit methods (i.e. market LGD approach and workout LGD approach) and implicit methods (i.e. implied market LGD approach). The decision to select one of these objective methods is largely driven by the nature of the portfolio in question, exposure type (loan vs bond), and the availability of data.
Estimates are based on historical recoveries (including collateral) in economic downturn conditions and used in calculating regulatory capital. LGD is economic rather than accounting loss, which includes direct and indirect costs discounted back to the point of default. Interpretations of key parameters differ by bank and are not always comparable. Definitions of downturn vary, with some banks using two consecutive quarters of negative GDP growth, while others emphasise product downturn rather than overall economic conditions. The relative financial condition and capabilities by bank are important, as funding levels affect discount rates, stronger banks negotiate better collateral terms, and more capable and well‐staffed teams work through defaults more quickly and efficiently. While PD is largely the same across all types of exposures to a borrower, LGD is likely to vary significantly by product. Banks are expected to be conservative, and auditors and external supervisors must be able to validate the model. Repurchase Value Estimators are used when banks need to take possession of and sell property and goods.
Expected exposure at default (EAD) is the gross exposure upon default. For fixed credit facilities (such as term loans), EAD is simply the amount outstanding (although EADs slightly above 100% are not uncommon given interest accrual). For revolving facilities (such as lines of credit, liquidity facilities, and overdrafts), EAD is the drawn amount plus an estimate of the amount of the remainder of the commitment likely to be drawn at the time of default. These estimates are often referred to as either the Credit Conversion Factor (CCF) or Loan Equivalent.
LGD and EAD for corporate, sovereign, and bank exposures are based on a BIS required period of no shorter than 7 years. Estimates for retail exposures are based on at least 5 years of data unless the bank demonstrates that recent data is a better predictor.
If a supervisor agrees that a bank's total expected loss is less than its provisions, the difference can be included in Tier 2 capital. This can occur in practice as the assumptions used in modelling the provisions required do not necessarily align with those for the capital calculations (for example, a point in time LGD is often used for provisioning, while a downturn LGD is used as a basis for capital calculations).
When exposures default, actual losses can exceed LGD estimates. At this stage, banks need to make a Best Estimate of Expected Losses (BEEL) for each defaulted asset considering the current economic climate, so as to cover the possibility of additional losses. Again, historical data from a full economic cycle should be used as a basis, with losses consistent with any provisions or charge‐offs taken.
Treatment of assets with a reduced likelihood of performing in full was a source of controversy in the financial crisis, as with hindsight it is clear that often problems should have been identified and loss provisions taken earlier. Under the IAS 39 “incurred loss” model, recognition of credit losses was delayed until there was evidence of a trigger event. While designed to limit the ability to build hidden reserves that could be used to boost earnings in future difficult periods, the model enabled earnings management, which postponed losses.
Under IFRS 9, banks are required to recognise expected credit losses at all times and to update the amount of expected credit losses recognised at each reporting date. IFRS 9 broadens the information that banks are required to consider when determining expected credit losses (for example, reasonable and supportable historical, current, and forecast information). To reduce complexity, the same impairment accounting will now be applied to all financial assets.
This forward‐looking approach is much more transparent for central banks, regulators, creditors, and shareholders. The model eliminates thresholds and triggers for loss reporting, which should reduce “cliff effects”. Disclosure is enhanced, with banks required to explain the basis for their expected credit loss calculations, to explain how they measure expected credit losses, and to explain how they assess changes in credit risk.
Even default has no simple definition. While some advocate a quantitative standard for clarity and easy understanding, most regulators believe interpreting default as only non‐payment fails to capture clear indicators of loss and recognition of increased expected losses.
The most common objective definition of default is when the borrower is:
Rating agencies use this narrow definition above.
Wider definitions include when the borrower is:
In practice, banks can only use broader definitions if information is available to them. Sometimes the definition of default is addressed in product documentation, which includes covenant breach and cross‐default. Products specifying a broad range of technical defaults mean that default may likely occur well before non‐payment. Laws vary by jurisdiction as to the lender's ability to place the borrower in default, so banks may have difficulties in being consistent.
The cash waterfall is the priority of payments by which classes of lenders receive interest and principal. Senior lenders are paid first, followed by junior (also called subordinated or mezzanine) lenders. On loan initiation or bond purchase, senior lenders receive lower returns in exchange for lower credit risk. In a default, equity holders receive what (if anything) remains after assets are distributed to debt holders.
Lenders analyse the size of the company or project assets and collateral relative to debt to assess the value of seniority. If the loan is a general obligation of a large company to finance a small project, seniority may be less important. If repayment is secured only on the asset financed and the cash flow it can generate, a higher priority of payment may be more important.
Loan covenants require borrowers to meet certain conditions over the period of a loan. Breaches of covenants may trigger default, higher pricing, penalties, or termination (“call” or “acceleration”) of the loan. Covenants can relate to reporting standards, financial performance, ownership, or business activities.
Common financial covenants include:
Covenants may be waived, and only serve to ensure early alerts and constructive dialogue when conditions become more difficult. However, covenants do cede some control to banks, and management can be forced to take actions not believed by them to be in the best interests of the company.
Currency convertibility and transfer risk is the loss that can occur if local currency cannot be converted to another currency and/or transferred abroad. The situation can arise when a country is experiencing capital outflows in a time of political and economic crisis. Governments may impose currency controls, or there may be simply no market for the currency. The risk is different from devaluation or appropriation of assets.
Risk assessment must include analysis of economic conditions, political situation, and legal framework. To the extent the country is integrated into the global economy through trade and capital markets activity, currency restrictions would be more likely to be avoided given the severe repercussions.
Risk can be mitigated by the use of cash collateral and letters of credit with banks in other jurisdictions, or increases in interest rates linked to currency disruption.
As discussed earlier, definitions of default vary. Impairment can be described generally as when an exposure is judged by management to have deteriorated so there is no longer a reasonable expectation as to the collection of the full amount as scheduled.
Banks can analyse a number of triggers for borrower deterioration to determine whether an asset is impaired:
Banks should disclose impairment triggers to supervisors.
Loans generally appear on bank balance sheets as assets using nominal principal values. Once a loan is identified as impaired, the current probability of default and loss given default is applied and discounted to establish the new value. Both the loan and capital (shareholders' equity) are marked down on the balance sheet. Impairment provisions appear on the income statement as an expense. Debate as to the optimal balance accounting for loans is ongoing, with some arguing that constant marking‐to‐market is needed.
Credit risk measurement for standard loans involves the most basic credit risk measures. Some products are undrawn, so exposure is dependent on usage. These contingent liabilities are off‐balance‐sheet, and Credit Conversion Factors (CCFs) are applied as estimates of risk. The exposure is multiplied by the CCF to assess capital required.
Banks provide liquidity facilities for clients to draw down as needed. Facilities can be committed or uncommitted. Clients issuing commercial paper (CP) (marketable notes maturing in 1 year or less) need backstop liquidity facilities to repay maturing issuance should rollover not be possible. Backstops are a rating agency requirement for the high credit ratings demanded by CP investors.
Under Basel I, the CCF for liquidity facilities under 1 year was 0%. Under Basel II, the CCFs for standard banks providing facilities under 1 year and over 1 year were 20% and 50%, respectively. In reassessing risk and increasing capital requirements under Basel III, the distinction based on term is eliminated and the CCF for all facilities is 50%. Lower CCFs for facilities that could be drawn only in the event of market disruption (not client credit deterioration) have been eliminated.
Any facility that is uncommitted and can be cancelled unconditionally and without notice, and requires the bank to proactively approve new drawdowns, has a CCF of 0%.
Guarantees of financial indebtedness are integral to world trade. These include loan guarantees, letters of credit, and banker's acceptances. Banks must categorise guarantee facilities in three ways: unutilised, utilised, and utilised with payment/obligations owing to the beneficiary. CCFs are generally 0–50%, 50–100%, and 100%, respectively.
BIS rules stated previously that bank exposure to another bank's letter of credit was subject to a “sovereign floor”, for example, the risk weighting could not be lower than that of the sovereign. This was prohibitive for importers using local banks to issue letters of credit in countries where the sovereign was unrated and external ratings were used by the other bank. The BIS has now waived the sovereign floor to allow the risk weighting to go below 100%.
The BIS also eased proposals for stricter capital measures for trade finance in certain short‐term, self‐liquidating, trade‐related contingent liability products collateralised by the underlying shipments, allowing the CCF of 20% to apply to the actual remaining maturity rather than a 1‐year floor.
Other off‐balance‐sheet banking products include Revolving Underwriting Facilities, in which a group of underwriters agrees to provide loans or buy notes in the event that a borrower is unable to issue in the capital markets.
The following terms are important in understanding credit risk measurement.
As a first step, lenders must know exactly to whom they are lending. The legal entity and type (individual, partnership, trust, or corporation) and its powers to conduct business and engage in borrowing must be fully understood. The structure can be simple or highly complex, involving organisational charts and legal shells.
Banks must analyse exposure to all entities in a legal or economic group when setting limits, and take a view on the group's industry sector to manage concentration. Given support and cross‐support arrangements, the performance of non‐borrowing entities can either improve or hinder the ability of the borrowing entities to pay. Covenants and default events can be negotiated to apply to the entire group. The bank must be confident that controls are in place so that intragroup loans and transactions are conducted at fair market terms and on an arm's length basis.
On some transactions, co‐borrowers are jointly and severally liable, meaning that if one borrower can no longer repay debt, the other is responsible for full payment.
The Basel Committee recognises that banks have complex global businesses across regions and tax regimes, necessitating multiple entities across the group. However, banks' own intragroup support arrangements and dependencies must be fully understood with risks well managed.
Credit agreements include the terms of drawing down the amounts of a facility and types of accounts (for example, revolving credit account, term loan account). A schedule of drawdown amounts and dates can be specified, particularly in project finance. Drawdowns are subject to meeting the conditions precedent, which can involve providing these documents:
Often, the provision of collateral is the precedent for another condition.
The timing and likelihood of drawdown are necessary to estimate exposure and risk.
Collateral is (an) asset(s) pledged by the borrower to the lender to secure a loan. In the event of default, and where the bank decides not to restructure the transaction (and the counterparty cannot refinance externally), the bank will take possession of the collateral in order to offset their loss exposure. It is essential that the lender “perfect a security interest” in the collateral, so that it can easily take control and sell without dispute should the borrower fail to meet the terms of the loan.
Collateral ranges from cash, securities as well as property, to the asset being financed. The lender is best protected if the collateral is marketable in all economic conditions, characterised by low price volatility and denominated in the same currency. Banks can choose to lend up to a percentage of the value of the asset to protect against a decline in value. This is called taking a “haircut”, and the concept behind fixing loan‐to‐value percentages in the residential mortgage market. Some loans are “overcollateralised”, meaning the bank receives collateral worth more than the loan. Collateral value must be monitored regularly, with additional collateral (“margin”) required if the value declines.
Taking collateral must not be seen as risk‐free lending. Banks must only lend to clients they believe have the ability to repay from their operations. Reputation risk, collateral value volatility, and the process of taking and liquidating collateral must be considered.
Guarantees take many forms and are issued by all types of entities including banks, corporations, and sovereigns, as well as individuals. Banks issue direct guarantees and indirect or counter‐guarantees (where non‐performance of a second party's guarantee is guaranteed).
Guarantees include:
Sovereign guarantees back projects deemed in the public interest, and support development and promotion of infrastructure, new industries, regions, and exports. Many sovereigns have state‐owned development and export/import banks.
Banks offset client loans against deposits to reduce risk and capital requirements through netting. This is possible when:
Netting allows a bank to do more business with its clients.
Netting is key to non‐balance‐sheet activities and businesses including securities clearing, payment systems, and derivatives.
Derivatives are financial contracts with a value derived from the performance of securities, interest rates, currencies, commodities, indices, credit, and other assets. Derivatives can have set terms (contract size, dates, underlying) and be exchange‐traded and highly liquid, or bilateral tailored contracts.
Banks hedge by taking derivative positions designed to perform in a manner opposite to their actual exposures. For example, a bank with a large credit exposure to a corporation can take a position in a credit derivative that would pay out if the corporation defaulted. “Basis risk” is the risk that the derivative does not perform in the direction expected or in the same magnitude as the position being hedged.
The most common types of derivatives are:
Derivatives do not require the actual purchase and sale of the underlying. Most positions are closed out by buying or selling prior to or at maturity, with the gain or loss exchanged in cash. Derivatives are generally margined throughout the term of the trade.
Credit risk management models serve many purposes. While generating outputs mandated by regulation, models must be built to meet the needs of the bank and its business for optimal decision‐making. Different measures provide a variety of insights in both normal and stressed conditions, and aid in balancing profitability and business objectives with risk. It is imperative that models are built on sound and reliable data.
Risk estimates (PDs, EADs, and LGDs) for capital purposes may not be the same for pricing or impairment purposes. Point‐in‐time or through‐the‐cycle views may differ. For example, one will not necessarily price a 12‐month loan on a through‐the‐cycle credit expectation.
Probability of Default is estimated from a range of sources. The simplest and most widely used throughout the world is rating agency ratings. Banks also use their own historical default databases or purchase those compiled by third parties. For some sectors, decades of default data is available. PD can also be estimated by monitoring bond and credit default prices.
Statistical methods to estimate PDs include:
Banks must make careful judgements as to how data is used. While default is rare (roughly 2% on average globally), consequences for debt portfolios are severe given small earnings margins and no upside as in equities. While modelling monthly or quarterly data from portfolio segments is common, defaults observed may not be a good indicator for forward‐looking analysis if a portfolio is growing or the market is new. The risk of PD being understated is significant.
An important concept in PD is “distance to default”. PD increases as the market value of the assets of a company decreases towards the book value of the liabilities. Issues considered are:
The default point is sometimes when the two values converge, although companies may continue to trade if the liabilities are longer term and creditors believe in the business.
Models must provide PD in both unstressed and stressed economic scenarios. Higher interest rates, which make debt more costly, can be integral to stress scenarios. “Point‐in‐time” PDs are estimated for unstressed conditions while “through‐the‐cycle” PDs estimate the trough. Both are fixed for 1 year, but PIT PDs will be volatile as the economy evolves, while TTC PDs will be more stable. Obligors must be classified as to how they are likely to respond to the economic cycle at both peaks and troughs.
The most common loss given default (LGD) measure is “Gross” (total losses are divided by EAD) because it is simple to calculate and requires the least data. Another LGD measure is “Blanco” (losses divided by the unsecured portion of a credit line), which is important when a bank has significant collateralisation. As a conservative measure, collateral is “haircut” in the calculation to allow for a fall in value, thereby decreasing LGD. Banks calculate “Downturn” LGD.
LGD can be difficult to calculate as bank recovery rates vary and workouts take different lengths of time, so peer data is not always useful. Formulas have been developed to best achieve comparability. Models “time weight” LGD, meaning historical data is not analysed simply by averaging loss severity of each default, but also considers the time periods in the economic cycle when they are likely to occur.
There are three objective LGD estimation methods:
Given the challenges involved in calibrating LGD models, banks reference external data sources, such as Pan European Credit Data Consortia (PECDC), S&P LossStat, and Paris Club restructure data.
Exposure at Default is the gross total of extended credit plus estimated additional drawings for 1 year or until maturity. The greatest analytical challenge in setting Credit Conversion Factors (or Loan Equivalents) is estimating additional drawings. Globally, unused commitments are huge and it is logical that a corporation would seek to drawn down in stress scenarios. Examples of products where modelling is needed include committed loan and liquidity facilities and credit cards.
Strong information management systems are vital in assessing EAD, as the bank must ensure that troubled entities draw only under the terms permitted by the facility and up to the limit. Collateral must be monitored, priced, and margined. The bank must deal efficiently and quickly in default situations.
Banks must have a standard and regular process for validation and review of credit risk models. Validation must assess the accuracy and consistency of ratings and risk components in an independent manner, with input from relevant departments. A fundamental role of supervisory authorities is to ensure this process is conducted in a meaningful and thorough manner, with banks making available the inputs and calculations.
Backtesting is an important part of model validation, and includes comparison of model results against actual ratings migration and loss experience by category. Benchmarking of internal estimates against external sources is another useful quantitative review, and can add objectivity. The entire process should make apparent changes in drivers, trends, and correlations. Outcomes can include revision of risk categories and adjustments to data timeframes. Strong risk aggregation capabilities are vital and deficiencies (a problem in the crisis) should be exposed to ensure business grows only as quickly as control infrastructure. Business line leaders must have a basic understanding of the models and ensure their risks are fully incorporated into the bank‐wide risk process.
Risk appetite is the quantum of risk a bank is prepared to assume in pursuit of its strategy, and is established, integrated into business plans, and monitored by the Board. The risk appetite statement sets out the risk profile by identifying risks and boundaries. The statement should be actionable and include quantitative measures. High level limit and target measures can be set against earnings at risk, probability of insolvency, and the chance of experiencing an annual loss. Best‐practice discipline dictates that the bank's Board sign off on the risk appetite statement. By being a Board‐issued document, it ensures that adherence to risk tolerance is taken seriously as a requirement of senior management.
Implementation requires proper policies, procedures, and controls. Regulators will monitor the level of Board involvement, adherence to policies, breach of limits, and policy changes. Crucially, the risk appetite statement must be used to identify excessive risk taking, which can threaten a bank.
Trading book exposures must be managed actively and held for “trading intent” or short‐term gain. Positions may be proprietary or arise from market making to serve clients. Capital should be allocated to cover losses from a very short period (10–20 days). Standardised method banks must use set parameters, but advanced banks with approved internal model banks can use their own EAD and Value‐at‐Risk. In addition to the risk of default of the issuer of the securities, there is a small credit risk in trading, as the counterparty could default before settlement with the position needing to be replaced at a worse price. Most securities trade on a delivery‐vs‐payment basis (DVP), so there is little or no settlement risk in exchanging securities and cash.
Exchange traded derivatives are margined and guaranteed by exchanges, minimising credit risk. Over‐the‐Counter (OTC) derivatives are bi‐lateral contracts with substantial Counterparty Credit Risk (CCR), which must be carefully measured and managed.
Swap counterparties agree to exchange fixed‐rate for floating‐rate payments. If one counterparty defaults, the other has a loss if the trade has a mark‐to‐market gain. Assuming no recovery, the loss is the present value of the net payments remaining (replacement cost). Non‐defaulting counterparties with a negative mark‐to‐market position are not released from the position and will likely make a single cash payment to the receiver. There may also be some transaction cost in replacing the position in the market.
Buyers of options have similar CCR risk.
Swaps, foreign exchange and interest rate forwards, options, other derivatives, and securities finance transactions (repo) are subject to fluctuations in value over the life of the contract. Besides replacement value, credit risk measurement must include Potential Future Exposure (PFE), defined as the maximum expected credit exposure. PFE is important because some transactions have longer maturities where losses may emerge over time. Also, positions with large downsides in extreme markets (for example, options sold) are more fully captured.
CCR “add‐ons” are determined by multiplying the notional principal amount by Credit Conversion Factors set by type (interest rate, commodities, credit, currency, and equities), features, position, and term.
Netting agreements allow parties to net the mark‐to‐market values of their trades so that in the event of default the credit exposure is limited to the net positive value of the total. Netting is generally effected under an International Swap Dealer Association (ISDA) Master Agreement signed between the parties, which specifies methods for calculating a single settlement amount in the termination currency. ISDA has obtained legal opinions from major jurisdictions confirming the enforceability of netting. “The ISDA” also specifies margining arrangements and collateral terms so that CCR in the normal course of business is further reduced with less capital required.
Securities finance transactions have similar netting and are generally executed under the ICMA Global Master Repo Agreement.
Bond and credit default swap (CDS) prices can be used to extract the market view on probability of default and guide the pricing of loans and securities.
Default curves can be constructed by extracting credit spreads over risk‐free (for example, Treasury) rates, using as many securities and maturity pricing observations as possible. Assumptions, requiring carefully thought‐out best estimates, must be then made as to:
The process of curve construction is also known as “bootstrapping”. Given the investor premiums described above, it must be noted that actual historical default rates are less than those implied by bond and CDS prices.
The default intensity or hazard rate is the default probability for each time period. This is used to construct cumulative probability of default rates and cumulative probability survival rates (the two are inverses) for each time period. Investors generally demand relatively higher increases in credit spreads earlier along the curve for lower rated credits, implying increasing default intensity with time.
With the rapid growth and complexity of derivatives markets since the 1980s, credit risk models have advanced in sophistication. Early systems of managing risk were closed form approximations, with limited and static credit categories, default probabilities, recovery rates, term structure, potential future exposure, and netting measures. Little focus was directed to correlation, diversification, and credit migration. With limited historical data and less computing power, more elaborate models were not possible.
Given the complexity, number of dimensions, and uncertainty of the CCR of a bank's derivatives portfolio, Monte Carlo simulations are now the norm. While it is data and IT system intensive, Monte Carlo can incorporate the multiple sources of risk, correlations, and mitigants (including netting). Modelling therefore shifts from the deterministic setting to the probabilistic setting.
Large numbers of joint scenarios are generated based on numerous risk‐based factors pertaining to market conditions, defaults, credit migration, correlations, and recovery over the term of the portfolio. This is especially necessary as credit events and in particular defaults are rare, yet have huge impacts. Besides estimating risk, the simulations provide many insights into profit maximisation and hedging.
The exposure at default value (EAD) is based on the specific asset drawn amount plus the expectation that the facility usage will increase, up to and beyond limit, as the obligor approaches default. Historical data and observation of sector in default is used to drive this. The EAD is, therefore, invariably above the drawn amount and up to the facility limit and beyond (includes some “costs” in certain cases.)
The risk weighting percentage (RW%) is driven by the credit model. In this case, the RW% for retail is based on IRB (uses firm‐specific PD and LGD), while the RW% for commercial real estate (CRE) has been calculated via the regulatory authority's “slotting” technique (whereby the regulator applies its own risk‐weight value because it does not deem the bank's internal model and/or data analytics to be robust enough for this asset class). The RW% for non‐CRE is the standardised approach (no model is needed, and it uses internal rating, which will be model derived).
Note that the 150% RW level is used if there are any non‐performing assets on the balance sheet.
Figure 3.4 is an illustration of the summary of the RWA calculation using the input parameters described here.
Now we examine the credit process of a bank, including how it organises the credit function, and approves, analyses, structures, prices, monitors, and works out credit assets.
As made clear in previous chapters, the Board of Directors has overall responsibility for all risks of the bank and sets strategy, policy, and limits. The Board must have a Risk Management subcommittee, which includes the CEO and the heads of the management level Credit, Market, Asset Liability, and Liquidity as well as the Operational Risk Management Committees. The Board subcommittee must take a coordinated and integrated approach to the range of risks. It is vital that its integrity and independence are maintained, with a system in place to report to internal and external auditors, for when and reasons why the full Board does not accept recommendations.
Given the critical importance of credit risk, the Credit Policy Committee (also called the Credit Risk Management Committee) should be chaired by the CEO and include the Chief Risk Officer, Heads of Credit, Credit Risk Management, Treasury, and the Chief Economist. This committee is responsible for implementing Board‐level credit strategy and policy as well as the following:
Credit approval:
The Board and Credit Committee policies and strategies must be communicated effectively throughout the bank and permeate to each level and function. Delegation of authority and responsibility with clear reporting lines and accountability are paramount. Credit risk functions must be staffed and resourced sufficiently.
On an operating level, banks need to have separate Banking and Credit Risk Management Departments. The Banking Department (often organised by industry specialisation) manages customer relationships and takes a commercial approach in identifying business opportunities, pitching for deals, and negotiating and closing transactions. The Credit Risk Management Department measures risk and enforces limits and standards, with constant overseeing of the entire portfolio. Subdepartments of the Credit Risk Management Department include Credit Portfolio, Credit Modelling, Monitoring and Collection, Collateral Management, and Restructuring.
While separate, all departments and functions must rely on each other for credit risk management to be effective. Credit Modelling must receive timely and accurate data to provide useful risk measures to the Credit Department and Credit Portfolio for action.
A credit score is a numerical expression of creditworthiness generated by a statistical model using pertinent data. In contrast to commercial lending where extensive analysis is performed on the borrower with judgements being made, high volume/small size transactions generally dictate that retail lending is based on automated scoring without any human intervention. This applies to approval, pricing, terms, monitoring, control, and collections. Credit scores are sometimes also used for Small to Medium Enterprise (SME) lending.
Banks must satisfy supervisors that the data is relevant to exposures, the model has a sound track record in predicting default and is regularly tested and updated, and a system is in place for governing the use of the credit scores.
Credit bureaus (also referred to as consumer reporting or credit reference agencies) collect and aggregate personal and financial information from sources including creditors, lenders, utilities, debt collection agencies, and public records. A particular focus is past borrowing and bill paying habits. Bureaus provide their clients with credit reports for credit risk assessment and scoring, or for other purposes such as offering employment or renting of property.
While formats vary, all credit reports contain the same basic information:
In the past, credit reporting involved only negative information, but now scheduled repayment of debt is used as a positive. Negative items closed out are removed from credit reports over time (there are regulations that govern these time periods, which are often country specific).
The minimum for a report is generally one undisputed account opened for 6 months or more with no indication that one of the holders is deceased.
Credit reports do not contain information on race, religion, national origin, sex and marital status; location of residence; age; income and employment history; interest rates charged on accounts; child support obligations; customer initiated report checks; history of credit counselling; or any information not proven to be predictive of future credit performance. Much of this information is stored at the credit bureaus themselves, and can be used for validating client information required by the local regulations, if required.
Data quality is fundamental to the generation of useful credit scores. Potential sources of mistakes include:
Consumers can check their credit reports and notify credit bureaus of any errors. Credit bureaus have a legal duty to respond promptly; however, there is criticism about their efficiency in that a greater burden is placed on the consumer. Credit scores are modelled by both credit bureaus and banks. The most widely used credit scoring system is FICO, pioneered in the US in the 1950s by what was then Fair, Isaac and Company.
Consumers need to know the basics and means to improve their credit scores to meet their financial goals. Steps for consumers to take with the greatest effect are:
While credit scoring methodologies vary and are not disclosed in detail, overall weighting of factors has been estimated as follows: payment history 35%, debt to credit ratio 30%, average age of accounts 15%, types of credit 10%, and enquiries 10%.
Credit scores constantly change, therefore consumers must be reminded, educated, and be provided with incentives for good personal financial management.
Credit scores have a number of benefits:
Lenders will generally not rely on the credit score only. Potential borrowers complete an application where income and employment are generally considered. While larger banks generally use “black box” approaches that allow little autonomy on decision‐making for their local managers, smaller banks may often rely on their knowledge of community, customers, businesses, and relationships in approving credit. This makes them, in theory, superior SME lenders compared to the larger banks, although in practice small banks' level of loan losses are generally not markedly superior. There is nevertheless a genuine element of positive customer service perception associated with the ability to apply local knowledge and expertise in the loan origination process.
Some lenders utilise “cut‐off scores”. Loan applicants with scores below a set level are rejected unless the bank chooses to make an exemption (referred to as overrides). Cut‐offs generally are more stringent for home loans than for high interest debt such as credit cards or personal loans.
As discussed elsewhere, the Board uses the risk appetite statement to set the quantity and type of risk the bank will tolerate within its total capacity to pursue its business objectives. Highest level limits for broad exposure groupings are set after considering the impact of the potential transaction on the regulatory and/or economic capital required to support the credit position. This includes using probabilities of default to estimate expected losses under different scenarios (from normal to recession or to major disruption and with consideration of frequency). These may be expressed in absolute amounts for the riskiest exposures or as a percentage of total exposure. An internal economic capital framework considering correlation, concentration, and large single exposures is integral to setting limits. Risk is then allocated on an operating level across business lines, products, industries, and regions, as well as individual borrowers in the form of further limits.
Limits are incorporated in Pillar II of Basel II as part of the supervisory review process. Limits must not be merely a “rubber stamp” of business requests leading to the approval of further increases when requested only. Forward thinking credit management involves consideration of reasonable exposure levels in anticipation of new business opportunities and the potential medium‐term movements of the credit cycle. For instance, when there is an expectation of a downturn in the credit cycle, risk limits may be tightened to alleviate future volatility in the credit outcome (bad debt charge or default rates).
Business lines can work to stay within limits by extending credit only to borrowers with increasingly higher internal ratings (corporates) and credit scores (retail). However, tightening lending standards cannot be relied upon solely to reduce credit risk.
Limits serve varying purposes. In some instances, limits are a firm form of policing against taking on risks not deemed tolerable. In other instances, limits are essentially a form of an early warning system, where credit officers and management are alerted to an increase in risk that merits further discussion and analysis. Sometimes a client is involved in multiple products (including trading) and the bank may not be able to ensure that the total intra‐day exposure limits are not breached, in which case limits should be set conservatively low. As credit risk is constantly changing due to portfolio effects and new data, limits need regular adjustment whether nominal exposures change or remain the same.
In extending credit to a corporation, banks will consider attributes such as market share, quality of products and services, innovation, brand, and operating efficiency. However, a company's performance and ability to meet debt obligations are strongly linked to macroeconomic and industry factors.
Key factors include the phase of the economic cycle, interest, inflation, and exchange rates. Sensitivity to these factors varies by industry. While the durable goods, auto, transportation, and informational technology industries tend to perform in line with the economy (cyclical), the household goods, food, and utilities industries are more constant (non‐cyclical). Historically, the performance of the financial and homebuilding industries has been highly interest rate sensitive.
Industry characteristics to consider include:
With a broader understanding of the industry, it is then possible to analyse a company's competitive advantages, positioning, and its prospects in the future.
Corporate financial statements include volumes of data that can be used to perform extensive credit analysis. The most widely used technique is financial ratios, with four main categories:
It is possible to generate dozens of financial ratios from a company's balance sheet and income statement. However, credit analysts do not rely on ratios as much as in the past or they do not have the same expectation, as in the past, that benchmark ratio levels must be met. Often, one ratio appearing weak is offset by a stronger one, and more sophisticated statistical modelling techniques are available now.
Ratio by industry sector and company size (for example, large corporate, middle market, SME) is used to assess norms and relative strength. Historical ratios are useful in identifying trends. Table 3.1 provides a summary of the formulas and measures for various ratios.
Table 3.1 Formulae and description for the key ratios mentioned above
Ratio | Formula | Measures |
Leverage | ||
Debt to Equity | debt/equity | Owed compared to owned |
Interest Coverage | EBIT*/interest expense | Ability to pay interest using cash flow |
Liquidity | ||
Current or Working Capital Ratio | current assets/current liabilities | Liquidity reserves |
Quick Ratio or Acid Test | (cash+marketable securities+receivables)/current liabilities | Ready liquidity (excludes inventory) |
Profitability | ||
Gross Profit Margin | (revenue‐cost of goods sold)/revenue | Ability to generate profit |
Return on Equity | net income/shareholder equity | Return on shareholder's investment |
Return on Assets | net income/average assets | Return on assets used |
Efficiency | ||
Asset Turnover | net sales/average assets | Use of assets to generate sales |
Payables Turnover | cost of sales/trade payables | Speed of paying bills |
Operating Cycle | average inventory/cost of sales per day | Time between acquisition of inventory and realisation of cash |
*EBIT: earnings before interest and taxes
Once the credit strength of a potential borrower is analysed, a bank can consider the structure and risks of the proposed facility. The bank must understand the purpose, amount, and duration of the borrowing. The loan must be for an activity within a company's business remit or an individual's financial goals. It is acceptable to finance the repayment of existing debts, providing the new loan terms can accommodate the client's needs better and can be serviced.
By understanding a borrower's requirements, the bank can offer the most suitable facilities, which include the following:
Further detail on this asset classification is given in Chapter 4.
In analysing the risk of a facility, bankers must always ask the question “Just how am I going to be repaid?” The bank must identify the primary source of repayment and consider whether there are secondary sources of repayment. For example, in a revolving line of credit to a company, the bank is looking to the successful conversion of the working capital into cash. Should that fail (decline in sales, receivables not collected), secondary sources could be other cash flow from operations, new business, fees, the sale of other assets, and divestitures.
Figure 3.5 illustrates the nature of outstanding loan balances of different facilities. As the revolver is on demand, usage is unknown and will vary. The bullet payment means principal risk remains for the life of the loan. Amortising loans generally require regular payments combining principal and interest, with the proportion for interest declining as principal is reduced.
For many loans, the bank's exit strategy is important. When a loan to help a company make an acquisition results in high debt levels, the bank needs to believe the merger will be cash generative to reduce leverage in a reasonable timeframe. The bank might also take the view that the company would have access to the bond markets as an alternative funding in the future.
To the extent that a bank's internal rating/scoring, industry, facility, source of repayment, and exit strategy analysis raises questions about the extension of credit, it may still wish to proceed by requiring credit “mitigants”.
These include covenants and guarantees (Chapter 4).
Borrowers are assigned a single internal rating/credit score, but risk varies depending on facility type and credit mitigants. Internal ratings assess PD, but guarantees and collateral mean some facilities can have much lower expected losses. Basel II includes a methodology for “notching” internal ratings upward to generate higher “facility ratings” for secured borrowing.
Risk‐adjusted return on capital (RARoC) is the profitability of capital after considering all costs. This profitability measure expresses expected profit (net of all costs including expected losses) as a percentage of economic capital, or worse case loss. RARoC is preferred to risk‐adjusted return on equity (RARoE) as it includes all capital rather than purely core Tier 1 equity. The relationship to economic capital needs to be understood, although it is regulatory capital that ultimately should drive the overall target and strategy, given that the regulatory capital level is required by legislative fiat.
The RARoC (or RARoE) model expresses facility and connection level income net of expected losses (i.e. it is risk adjusted) as a return on regulatory capital. It should be a key component of principal front book asset pricing calculators. By applying a RARoC target the bank is able to:
The formula is given by:
The element “capital benefit” is contentious and is not an input that is applied at every bank. This is certainly not agreed with by the author of this chapter. Capital benefit is included by some banks who argue that because the business line is charged the cost of capital held against it, it should also benefit from income received when the capital liabilities are “hedged”. That this leads to inconsistencies should be evident when one realises that a business line can start the year with “profit” on the books simply because of this capital benefit, before having undertaken any genuine shareholder value‐added work of any kind. Nevertheless, it is not uncommon to see this element on the numerator line. The better‐managed banks will dispense with this element, and it is a true test of a bank CFO's genuine understanding of the principles of corporate finance to witness whether this element is included in the bank's RARoC analysis.
RARoC enables a bank to compare opportunities more consistently by ensuring that relative risks are considered. Capital can be better allocated to business areas and products, right down to individual loans. RARoC is useful in evaluating ongoing performance as well as deal flow.
Banks set target or “hurdle” RARoCs for the use of capital. Opportunities that do not meet the required RARoC should be rejected or escalated to senior management for additional consideration. Business decisions to proceed can be made because a product is a “loss leader” valued by clients or a new line of business where profitability takes time to develop.
EVA (Economic Value Added) is net profit less the cost of use of economic capital (also known as NIACC – net income after capital charge). The cost is usually calculated as a weighted average cost of the bank's sources of capital. Unlike RARoC, EVA is an absolute measure. EVA is sometimes described as “excess return”. This “economic profit” is different from simple accounting profit because it considers implicit costs as well as tangible costs.
A comparison of RARoC and EVA is summarised in Table 3.2.
Table 3.2 Comparison of RARoC vs EVA
RoE: in principle the methodology is: | ||
RoE = |
– Costs are driven by the transaction | |
– A hurdle RoE % is set to determine the income required | ||
– The RoE hurdle can be adjusted as appropriate to market/sector/product | ||
– Income required to meet RoE hurdle % less costs drives the price | ||
RoE is return on capital employed, which is usually adjusted for risk purposes as RARoE | ||
EVA: in principle the methodolgy is: | – Costs are driven by the transaction | |
– Profit (EVA) is calculated as a % of costs | ||
EVA = |
Income – Costs – Cost of capital | – The EVA can be varied as appropriate to market/sector/product |
– Total of costs + EVA drives the required price | ||
EVA is the value created (i.e. economic profit) through undertaking the deal |
A conventional vanilla pricing approach for a corporate bank relationship manager would use these inputs:
Input (4) is the Treasury‐applied “term liquidity premium”, which is mainly, though not wholly, a function of the tenor of the lending and the liquidity of the asset once originated.
Figure 3.6 shows what the pricing screen might look like.
Asset Pricing Calculator | |||||||
Product type | Loan | Customer | ANO & Sons | ||||
Utilisation | 100% | ||||||
Interest rate basis | LIBOR | Asset costs illustration | |||||
Amount | £1,000,000 | Tier 1 capital | £12,640 | Cost of capital | £2,100 | ||
Term (months) | 60 | TLP bps | 236 | TLP | £8,325 | ||
PD | 0.064% | Expected loss rate | 0.06% | Expected loss | £480 | ||
LGD | 5% | Undrawn liquidity buffer | 0.00% | Liquidity buffer | £0 | ||
Total costs | |||||||
Recommended pricing | RM proposed pricing | ||||||
Margin bps | 431 | Proposed margin bps | 325 | ||||
Target margin bps | 331 | Proposed fee | £0 | ||||
Minimum margin bps | 306 | Proposed non‐util fee | £0 |
Figure 3.6 Corporate lending pricing screen
Beyond risk appetite and measurement, the Credit Policy Committee establishes and monitors the credit authorisation process. This includes:
Larger and more complicated loans are generally approved by credit committees. The client officer sponsors the loan proposal memorandum and is responsible for completeness. The committee scrutinises the proposal to ensure underwriting standards are met and uses their experience and knowledge of similar customers as well as their own facilities to evaluate risk and reward in making a decision. Smaller and less complicated loans are approved through delegation of authority.
Monitoring the credit authorisation process is an important control function and can yield business insights. This includes tracking:
Breaking data down by client, product, and facility can demonstrate where underwriting standards are falling in line with the market.
The Credit Policy Committee should ensure that minutes of Credit Committee Meetings are complete and include the justification for their decisions.
Banking and collateral are synonymous with each other, as evidenced by the numerous references in this and other chapters. Collateral is an excellent risk mitigant and is recognised in the Basel Accord's loss given default methodology. However – and while banks would almost always prefer to receive collateral – credit should only be extended when it is deemed that the borrower is likely to be able to repay without it. The liquidation of collateral can involve legal, market, and reputational risk, as well as time and resources.
The lending bank must have clear legal rights over the collateral, and must be able to liquidate or take possession in a timely manner in the event of default, insolvency, bankruptcy, or defined credit event. This is called “perfecting” an interest in the collateral. Collateral rights vary by jurisdiction, therefore business collateral transactions with multinational corporations will require focused legal advice.
The amounts and types of collateral are negotiated and depend on:
Collateral for loans is most commonly real estate (mortgages), equipment, inventory, cash, or securities. The borrower must prove it holds the deed to the collateral, and certify that it is not already pledged to another creditor (unless the bank is willing to take a second charge).
Where possible, banks should use historical data to project liquidation value. Banks consider potential links between borrower default and declines in collateral value. For example, if a company fails because of lack of sales, buyers for equipment for the manufacture of its products may be limited.
Collateral management is often the responsibility of a dedicated department whose responsibilities include:
In most jurisdictions, a lien refers to a transaction where the bank does not hold the collateral. Banks will ideally hold securities collateral in their own custody department. However, only clear terms in lending documentation ensure the ability to sell the collateral and keep the proceeds in the event of default.
Demand for securities collateral is growing because banks are required increasingly to post collateral for their own financing, derivatives, and clearing businesses. Central bank trades with banks designed to provide liquidity and effect monetary policy are collateralised. Banks seek to rehypothecate (use collateral received for its own collateral pledges) whenever possible.
When new loans are booked and the economy, industries, and companies change, the risk and profitability of the overall credit portfolio also change. Credit policy adapts to rebalance the portfolio by adjusting limits, pricing, maturities, and requirements (for example, collateral, guarantees, and covenants) for new business. However, the credit portfolio management team has the tools to control risk and optimise the profitability of the portfolio, which can then complement lending policy and be quicker and more efficient. This is particularly important should concentrations build and credit concerns develop rapidly. Furthermore, portfolio management techniques can reduce risk and maximise the lending capacity for banks to offer competitive market pricing for important clients and products as often as possible.
The effectiveness of credit portfolio management depends on accurate, uniform, aggregated, and timely data.
Strategies include:
Selling and securitising loans frees capacity and can improve profitability if returns are low. While basis risk and liquidity are factors, the global credit default swap (CDS) market has grown massively with gross notional CDS amount reported by the Depository Trust Clearing Corporation to have exceeded $25.5 trillion at year end 2010 alone (http://www.isdacdsmarketplace.com/market_statistics). However, it must be noted that the CDS market is not available for the majority of credit risk reference names.
Traditionally, loan profit and loss were attributed to the responsible lending officer from start date through to maturity. Now, many banks credit the lending officer with the EVA or excess return of the loan (spread between expected return and economic capital) at inception, and then transfer the loan and P&L to the credit portfolio team to manage. Regardless, the relationship banker must remain engaged with the loan throughout its duration in working through any issues with the client and keeping management informed.
Banks need to have a periodic, objective, and comprehensive loan review function, which includes audit and is ultimately responsible to the Board. Smaller banks may outsource loan review. Key functions include:
Behavioural scoring is an additional means of assessing credit risk made possible once a credit facility is extended. Banks can observe the patterns of activity and gain insights into how a customer manages their financial affairs. Exceeding credit limits, missed payments, and returned cheques are negative behaviours. Excessive utilisation of overdrafts or numerous credit card advances can be indicators of financial difficulties. Behavioural scores are generated by weighting recent trends, and are used in evaluating renewals or requests for new or larger borrowings.
Vintage analysis is another means of using existing business to better understand risk. Loan underwriting characteristics and standards vary over different periods of time. Similar loans from time periods are grouped together to create vintage pools, which are further segmented by borrower rating. Pools are tracked to assess how performance is linked to changes in the economy, interest rates, and other variables. Vintage analysis is particularly useful for longer term mortgages in forecasting repayment rates, delinquencies, and charge‐offs. Newer loans can be compared to earlier vintages at similar points in their lifecycle to test the impact of risk credit policy changes, for instance a decision to cut back on risk.
The definition of default is when a borrower fails to meet the conditions of a loan. Default can be a failure to make a payment as scheduled or technical, involving a breach of covenants. Default must be distinguished from the legal terms insolvency (borrower without means to pay) and bankruptcy (borrower under court supervision due to default or insolvency).
Corporations, individuals, banks, and even countries default. Difficult economic conditions, financial mismanagement, poor strategy and investment, excess leverage, and fraud all contribute to default. In almost all instances, it is preferable to work with the borrower to try to maximise value in the future than to become enmeshed in lengthy legal proceedings where lawyers and administrators rank first in order of payments.
Banks attempt first to refinance debt, which involves extending the maturity and thus reducing and rescheduling payments. If a realistic assessment of the company's finances and business outlook shows the debt burden is still too great, a debt restructuring is necessary. This will include a distressed situation if the company has defaulted on payments or moved into bankruptcy. Banks have restructuring departments to review businesses and work with the company and other creditors to create a workout plan.
Beyond refinancing, a workout can involve exchanging debt for equity and reduction of debt. Before agreeing to concessions, creditors need to agree that the new business plan will rehabilitate the company and make debt levels serviceable. Parties involved will have different incentives and interests, consequently, agreement is not straightforward.
If parties cannot agree that the company continuing as an ongoing concern is worth more than the assets, the parties should then work towards an orderly liquidation.
Consumer debt problems are often the result of loss of employment, decrease in income (less overtime, for instance), or financial mismanagement/illiteracy. The National Credit Act of 2007 established a debt counselling process for consistent debt restructuring, enforcement, and judgement.
A consumer may apply to a debt counsellor to:
Provision is made for basic living expenses before setting a new repayment schedule, but the consumer loses access to new credit.
Before applying for court permission to institute repossession, a creditor must first send the customer a written notice of options to repay or seek debt counselling.
All credit providers, credit bureaus, and debt counsellors must register with the NCR.
To be effective, debt counselling must ensure that borrowers (i) pay manageable levels of principal and interest; (ii) incur reasonable debt counselling fees; (iii) are given fair expectations of what can be achieved; and (iv) receive education and help with budgeting to avoid future problems.
Granting credit and monitoring the resulting default risk is the primary activity of banks. As credit risk is difficult to “hedge” or remove from the balance sheet (for most banks in the world anyway), it is self‐evident that the most effective credit risk management policy is to operate a sound loan origination policy. This chapter has described the basic principles of credit risk management, and like most bank risk management issues there is always an element of judgement based on knowledge and experience to apply to ensure good practice. A through‐the‐cycle approach combined with knowing one's risk, at all times, is probably the most effective ingredient of a viable credit risk management process.