Credit Appraisal, Risk Rating and Portfolio Management

Introduction and context

Commercial and retail banks transform deposits and other short-term funds into loans and advances. This intermediation creates income, but it also exposes the bank to credit risk: the possibility that a borrower or counterparty will fail to meet contractual obligations. Sound credit management therefore requires more than deciding whether an applicant appears trustworthy. It requires a disciplined process for originating, analysing, pricing, approving, documenting, monitoring and, when necessary, restructuring or recovering credit.

Credit appraisal is the investigation and evaluation of a proposed facility. Risk rating converts qualitative and quantitative evidence into a structured assessment of default risk. Portfolio management considers the combined risk of all exposures rather than treating each loan as an isolated transaction. These activities are connected: weak appraisal produces poor ratings, poor ratings produce inappropriate pricing and limits, and weak portfolio oversight can allow correlated losses to accumulate.

This chapter presents the credit process from origination through monitoring and explains the 5Cs of credit, internal ratings, probability of default (PD), loss given default (LGD), exposure at default (EAD), expected loss, concentration risk, provisioning, stress testing and portfolio limits.

Learning objectives

By the end of this chapter, students should be able to:

1. Explain the stages of the credit process and the responsibilities of key participants.
2. Apply the 5Cs framework to commercial and retail borrowers.
3. Distinguish credit scoring, internal risk ratings and regulatory classifications.
4. Calculate and interpret PD, LGD, EAD and expected loss.
5. Analyse concentration risk and design appropriate portfolio limits.
6. Explain provisioning, monitoring, early-warning systems and stress testing.
7. Evaluate governance, ethical and technological issues in credit decisions.

1. The credit process

A robust credit process normally contains the following stages:

The process should be risk-based. A small, standardised consumer loan may be approved through automated rules, whereas a large corporate facility requires detailed due diligence, industry analysis and independent credit committee review. Separation of relationship management from credit approval is an important control because the employee seeking business may otherwise have an incentive to understate risk.

Information and due diligence

Typical information includes identity and beneficial ownership documents, financial statements, bank statements, tax records, credit-bureau data, business plans, cash-flow forecasts, details of existing debt, collateral evidence and litigation searches. For a company, the analyst should investigate ownership, governance, management competence, related-party transactions, customer and supplier dependence, industry structure and regulatory exposure.

The analyst must assess information quality. Audited statements are useful but not infallible; management accounts may be more current but less independently verified. Inconsistencies between declared income, account activity and tax records are warning signs. Documentation should be proportionate to risk, but “know your customer” and anti-money-laundering requirements apply throughout the relationship.

2. The 5Cs of credit

The 5Cs provide a practical framework for organising judgement. They are not a substitute for evidence or a statistical model.

Character

Character concerns the borrower’s willingness to repay and integrity in dealing with the bank. Evidence includes repayment history, credit-bureau records, references, litigation, transparency, governance and the conduct of existing accounts. A borrower with strong income but a pattern of concealment may present greater risk than a less wealthy but transparent borrower.

Capacity

Capacity is the ability to generate sufficient cash flow to meet debt obligations. Repayment should ordinarily come from operating or personal income rather than from repeatedly selling collateral. Relevant measures include debt-service coverage, leverage, liquidity and cash-flow sensitivity.

A commonly used debt-service coverage ratio is:

CFADS is cash flow available for debt service during a stated period, and Debt Service is scheduled interest plus principal due in that period. Both quantities must use the same period and currency. A DSCR above 1 indicates cash flow exceeds scheduled service, but a bank normally requires a margin above 1 because forecasts are uncertain. Definitions differ: some institutions use EBITDA or net operating cash flow instead of CFADS, so the formula must be specified in policy.

Capital

Capital represents the borrower’s own financial commitment and ability to absorb losses. For a company, analysts examine equity, tangible net worth and leverage. For a household, they examine savings, assets and existing liabilities. Greater borrower capital generally reduces moral hazard and loss severity, although it does not eliminate default risk.

Collateral

Collateral is an asset pledged to support repayment. Its value, liquidity, legal enforceability, ownership, insurance, priority and volatility matter more than its nominal appraisal value. Collateral is a secondary source of repayment. It should not justify lending to a fundamentally unsustainable business.

A simple loan-to-value ratio is:

The collateral value should be a prudent, current and enforceable value, with suitable haircuts where prices are volatile or liquidation is costly.

Conditions

Conditions include the purpose of borrowing, economic conditions, industry trends, interest rates, exchange rates, regulation and the terms of the facility. A viable borrower can still become vulnerable when demand collapses, input costs rise or a foreign-currency loan is mismatched with domestic-currency income.

The 5Cs should be integrated. For example, strong collateral cannot compensate fully for weak capacity, and a strong historical character assessment may lose relevance after a major change in ownership or business model.

3. Credit analysis and facility structuring

Financial analysis

The analyst studies profitability, operating cash flow, liquidity, leverage, working-capital efficiency and trends. Ratios should be compared with the borrower’s history, peers, covenants and business model. A single ratio is rarely decisive.

Important measures include:

These are accounting ratios and can be distorted by valuation choices, seasonality, off-balance-sheet obligations or exceptional items. Cash-flow analysis is therefore essential. The analyst should reconcile profit with cash generated and identify whether receivables or inventory are absorbing funds.

Repayment and structure

A facility should be structured around the purpose and timing of cash flows. Working-capital borrowing may require a revolving line, while an equipment purchase may require an amortising term loan. Maturity should not exceed the realistic economic life of the asset without a clear refinancing rationale. Covenants may require minimum DSCR, maximum leverage, restrictions on dividends, insurance, information delivery or maintenance of collateral.

Pricing should reflect expected loss, funding cost, operating cost, capital usage, liquidity and a risk-adjusted return objective. A simplified conceptual relationship is:

This is not a complete profitability model: taxes, capital costs, hedging and recovery expenses may also be relevant.

4. Internal risk ratings and credit scoring

An internal rating is a bank’s assessment of the likelihood that an obligor will default over a specified horizon. Commercial ratings commonly combine financial strength, management quality, industry risk, business position, repayment performance and security. Retail lending often uses application and behavioural scorecards based on verified income, indebtedness, transaction history and bureau characteristics.

A rating must have:

• a defined meaning and time horizon;
• consistent data and documented overrides;
• independent validation;
• controls against discrimination and manipulation;
• regular review when risk conditions change.

Ratings are not the same as credit approval. A low-risk rating may support approval, but purpose, legal capacity, exposure limits and affordability remain relevant. Nor are internal grades automatically equivalent to external agency ratings or regulatory risk weights.

Probability of default

PD is the probability that an obligor defaults during a specified period, often one year. Default must be defined, for example by a material past-due status, bankruptcy or evidence that repayment is unlikely. A one-year PD and a lifetime PD answer different questions.

A basic empirical estimate is:

This estimate assumes the denominator is correctly defined, observations are comparable and defaults are measured consistently. In practice, banks adjust estimates for data limitations, portfolio composition and forward-looking economic information.

Loss given default

LGD is the proportion of exposure not recovered after default, including relevant recovery costs and the timing of recoveries. A stylised calculation is:

EAD is exposure at default. PV(Recoveries) is the present value of recoveries discounted at an appropriate rate, and Recovery Costs include legal, administrative and liquidation costs. LGD lies between 0 and 1 in ordinary applications, though realised losses can exceed original principal when costs, interest and fees are included. Collateral usually reduces LGD, but its value is uncertain and recovery may be delayed.

Exposure at default

EAD is the amount expected to be outstanding when default occurs. For a term loan, it may approximate current principal plus accrued amounts. For an undrawn commitment, EAD requires an estimate of how much the borrower will draw before default.

For a revolving commitment, a simplified expression is:

CCF is the credit conversion factor, between 0 and 1 in the usual case, estimated from historical utilisation and borrower behaviour. Contractual cancellation rights may affect the estimate, but they do not guarantee zero future drawdown.

Expected loss

For a single exposure over a defined horizon, the central relationship is:

PD is dimensionless, LGD is the fraction of loss, and EAD is a monetary amount; therefore EL is a monetary amount. The formula assumes the three parameters are defined consistently for the same exposure and horizon. If PD or LGD varies by time, a multi-period model is required rather than mechanically multiplying unrelated annual estimates.

#### Worked example

A bank has a corporate exposure with EAD of $2,000,000, a one-year PD of 3%, and LGD of 40%.

The expected loss is $24,000 for the specified horizon. It is an average statistical loss, not a prediction that exactly $24,000 will be lost. The actual outcome will usually be zero or a much larger realised loss.

Unexpected loss reflects the volatility around expected loss. Banks address it through capital, diversification, limits and risk-based pricing; provisions generally address expected credit losses under the applicable accounting framework.

5. Provisioning and problem-credit management

Provisioning recognises credit losses in financial reporting. The precise rules depend on the applicable accounting standard and jurisdiction. Under an expected-credit-loss approach, an institution uses historical experience, current conditions and reasonable, supportable forecasts. Exposures may be grouped according to changes in credit risk, with a short-horizon allowance for exposures whose risk has not increased significantly and a lifetime allowance when credit risk has increased significantly or the exposure is credit-impaired.

Provisioning is distinct from regulatory capital. Provisions reduce accounting profit and carrying value, while capital absorbs unexpected losses and supports solvency. Regulatory classifications, supervisory provisioning floors and accounting allowances may not be identical.

Early-warning indicators include arrears, excesses over limits, returned payments, declining deposits, covenant breaches, frequent requests for extensions, adverse bureau information, falling collateral values, tax arrears and management departure. An early-warning system should trigger investigation rather than automatic punishment. Remedial options include increased monitoring, additional security, revised repayment schedules, restructuring, negotiated settlement and legal recovery. Restructuring should not be used merely to conceal a borrower’s inability to pay.

6. Portfolio management and concentration risk

A loan can be acceptable individually yet dangerous in combination with other loans. Portfolio management examines risk by product, geography, industry, borrower group, currency, maturity, collateral type, rating grade and economic driver.

Concentration risk arises when exposures share a common source of loss. Examples include many loans to one corporate group, property lending in one city, exporters exposed to one foreign market or households whose repayment depends on the same employer. Correlation is crucial: diversification benefits are weakest when borrowers default under the same conditions.

A concentration report may include:

Portfolio limits may be absolute amounts, percentages of capital, percentages of total loans or risk-weighted measures. They should include escalation thresholds, approval authority, exception reporting and corrective action. Limits are not static: growth, macroeconomic conditions, capital, liquidity and risk appetite should influence them.

A simple Herfindahl–Hirschman concentration index is:

where sᵢ is the share of total exposure attributable to segment i, expressed as a decimal. If four equal segments each have a 25% share, HHI = 4 × 0.25² = 0.25. A higher HHI indicates greater concentration. HHI does not capture differences in credit quality or correlation, so it should supplement, not replace, limit analysis and stress testing.

7. Stress testing and scenario analysis

Stress testing asks how portfolio performance would change under adverse but plausible conditions. Scenarios may involve recession, unemployment, interest-rate increases, currency depreciation, property-price decline, commodity shocks, drought or operational disruption. Sensitivity analysis changes one variable; scenario analysis changes several connected variables; reverse stress testing identifies conditions that could threaten viability.

A simplified portfolio expected-loss estimate is:

Under stress, PD, LGD and EAD may all change. For example, unemployment can increase household PD, falling property prices can increase mortgage LGD, and borrowers drawing unused lines can increase EAD. The bank should assess impacts on provisions, capital, liquidity, earnings, covenant breaches and management actions.

Stress testing is limited by model risk, uncertain correlations, data gaps and the possibility that historical relationships fail during crises. Scenarios should therefore include expert judgement and multiple severities, not only a single forecast.

8. Governance, ethics and technology

Effective credit governance requires a board-approved risk appetite, clear delegated authorities, independent risk oversight, internal audit, documented policies, reliable management information and a credible three-lines model: business ownership, independent risk and compliance oversight, and internal audit assurance.

Ethical credit practice includes fair treatment, transparent pricing, privacy protection, responsible collection and avoidance of discriminatory lending. Affordability assessment is particularly important in retail banking. A technically valid model can still create unfair outcomes if its data reflects historical exclusion or if proxy variables reproduce prohibited discrimination.

Artificial intelligence and machine-learning models can improve fraud detection, document extraction, cash-flow forecasting and early-warning systems. However, they introduce risks involving explainability, data quality, drift, cybersecurity, privacy and accountability. Human review, model validation, performance monitoring, adverse-impact testing and an override framework are essential. Automation should improve consistency, not remove responsibility from the bank.

Applied case: Meridian Bank’s logistics portfolio

Background

This illustrative case concerns Meridian Bank, which has expanded lending to small and medium-sized logistics firms. The portfolio has grown rapidly, and 35% of borrowers serve construction companies. Many loans are secured by commercial vehicles and receivables.

Problem

Construction activity slows, receivables are paid later, and used-vehicle prices decline. Several borrowers request repayment extensions. The portfolio’s average rating remains acceptable because historical financial statements do not yet show full deterioration.

Analysis

The risk team identifies three linked channels. First, delayed construction payments reduce borrower cash flow and increase PD. Second, falling vehicle values and longer liquidation times increase LGD. Third, borrowers draw unused revolving lines to fund wages and fuel, increasing EAD. The common industry exposure means that apparently diversified borrower names are correlated.

The team calculates a baseline expected loss of $1.8 million. Under a moderate scenario, PD rises by 50%, LGD rises from 40% to 55%, and projected EAD rises by 10%. If all changes apply proportionally, stressed expected loss is:

This is a scenario estimate, not a forecast of realised loss. Management also tests a severe scenario involving a larger construction contraction and widespread payment delays.

Approach and solution

Meridian introduces monthly monitoring of receivable ageing, fuel costs, account turnover and vehicle values. It caps additional exposure to construction-dependent borrowers, increases review frequency for weak accounts, verifies collateral insurance and applies prudent haircuts to vehicle values. New facilities require demonstrated repayment capacity under higher interest rates. The bank offers viable firms structured amortisation but does not automatically extend loans where cash flow is unsustainable.

Outcome

The portfolio limit slows growth and encourages diversification toward healthcare and consumer distribution. Early intervention reduces forced sales of collateral, while additional provisions and capital planning prepare the bank for the stressed outcome.

Learning points

The case demonstrates that portfolio risk depends on common economic drivers, not only borrower counts. PD, LGD and EAD can deteriorate together. Forward-looking monitoring, limits and scenario analysis are therefore complements to individual loan appraisal.

Challenges and limitations

Credit decisions are affected by incomplete information, uncertain forecasts, changing borrower behaviour, model error, valuation uncertainty and legal delays. Historical default data may not represent a new product or crisis. Ratings may become procyclical if banks downgrade many borrowers simultaneously during a downturn and restrict credit precisely when businesses need it. Collateral can create false confidence, and excessive reliance on automated scores can exclude applicants with thin but not necessarily poor credit histories.

Professional judgement remains necessary, but it must be documented and reviewable. Analysts should distinguish facts, assumptions, estimates and stress outcomes; challenge optimistic forecasts; test repayment under adverse conditions; and escalate conflicts of interest. The objective is not to eliminate all defaults, which is impossible, but to make risk visible, priced, limited and actively managed.

Chapter Summary

Credit management is a continuous process beginning with origination and ending with repayment, recovery or relationship closure. The 5Cs—character, capacity, capital, collateral and conditions—organise the appraisal of willingness, ability, support and context. Internal ratings and scorecards translate evidence into consistent risk assessments, but they require defined horizons, validation and governance.

PD measures the chance of default, LGD measures the proportion lost after default, and EAD measures the exposure when default occurs. Their product gives expected loss for a defined exposure and horizon. Provisioning recognises expected credit deterioration under applicable accounting rules, while capital protects against unexpected loss. Portfolio management adds a system-wide perspective by controlling concentrations, correlations, currencies, industries and maturities. Stress testing examines how adverse conditions affect PD, LGD, EAD, provisions, capital and liquidity. Ethical governance, fair treatment, model controls and responsible technology are essential to sustainable banking.

Key Terms

Credit appraisal: Systematic evaluation of a proposed borrowing facility.

Credit risk: Risk of financial loss because an obligor fails to meet contractual obligations.

5Cs of credit: Character, capacity, capital, collateral and conditions.

Debt-service coverage ratio: Cash flow available for debt service divided by scheduled debt service.

Loan-to-value ratio: Loan amount expressed as a percentage of collateral value.

Internal rating: Bank-assigned assessment of an obligor’s credit risk.

Credit scorecard: Statistical or rules-based tool used to rank retail applicants or accounts.

Probability of default: Likelihood that an obligor defaults within a specified horizon.

Loss given default: Proportion of exposure lost after considering recoveries and costs.

Exposure at default: Amount expected to be outstanding when default occurs.

Expected loss: PD multiplied by LGD and EAD for a defined exposure and horizon.

Credit conversion factor: Estimated proportion of an undrawn commitment expected to be drawn at default.

Provision: Accounting allowance recognised for expected credit losses.

Concentration risk: Risk arising from large or correlated exposures.

Portfolio limit: Maximum permitted exposure to a borrower, segment or risk factor.

Stress testing: Analysis of portfolio effects under adverse scenarios.

Early-warning indicator: Evidence suggesting that credit quality may be deteriorating.

Covenant: Contractual condition imposed on a borrower.

Collateral: Asset pledged to support repayment.

Risk appetite: Amount and types of risk a bank is willing to accept.

Review Questions

Short Answer Questions

1. What are the main stages of the credit process?
2. Distinguish capacity from collateral in the 5Cs framework.
3. Define PD, LGD and EAD.
4. Why is concentration risk different from single-obligor risk?
5. State two purposes of early-warning systems.

Long Answer Questions

1. Explain how a bank should combine financial analysis, qualitative assessment and facility structure in a commercial credit appraisal.
2. Discuss the assumptions and limitations of EL = PD × LGD × EAD.
3. Explain the relationship between provisioning, expected loss, unexpected loss and regulatory capital.
4. Analyse how governance and separation of duties reduce credit risk.
5. Discuss the benefits and risks of using artificial intelligence in retail credit decisions.

Analytical / Application Questions

1. A borrower has EAD of $750,000, PD of 4% and LGD of 45%. Calculate expected loss and explain why the realised result may differ.
2. A bank’s portfolio is concentrated in one industry, and that industry faces a recession. Identify the likely effects on PD, LGD and EAD and recommend three management actions.
3. Design a stress test for a mortgage portfolio exposed to rising interest rates and falling property prices. Specify variables, transmission channels, outputs and possible limit actions.

Multiple Choice Questions with Answers

Question 1

Which element of the 5Cs primarily concerns the borrower’s ability to generate cash to repay debt?

A. Character

B. Capacity

C. Collateral

D. Conditions

Correct Answer: B. Capacity

Question 2

A loan has EAD of $500,000, PD of 2% and LGD of 40%. What is its expected loss?

A. $2,000

B. $4,000

C. $10,000

D. $40,000

Correct Answer: B. $4,000

Question 3

Which measure estimates the amount expected to be outstanding when a borrower defaults?

A. PD

B. LGD

C. EAD

D. DSCR

Correct Answer: C. EAD

Question 4

A facility has a drawn amount of $300,000 and an undrawn amount of $200,000. If the CCF is 50%, what is simplified EAD?

A. $200,000

B. $300,000

C. $400,000

D. $500,000

Correct Answer: C. $400,000

Question 5

What is the principal purpose of a portfolio limit?

A. To guarantee that no borrower defaults

B. To control exposure to a borrower, segment or risk factor

C. To eliminate the need for credit appraisal

D. To replace financial statements

Correct Answer: B. To control exposure to a borrower, segment or risk factor

Question 6

Which situation most clearly illustrates concentration risk?

A. One borrower makes an early repayment

B. A bank has many unrelated borrowers in different industries

C. Many borrowers depend on the same declining industry

D. A customer provides additional collateral

Correct Answer: C. Many borrowers depend on the same declining industry

Question 7

What does a stress test examine?

A. Only the most likely economic forecast

B. Portfolio effects under adverse but plausible conditions

C. Whether collateral documents are signed

D. The number of new accounts opened

Correct Answer: B. Portfolio effects under adverse but plausible conditions

Question 8

A DSCR greater than 1 generally indicates that:

A. Collateral value exceeds the loan amount

B. Cash flow available for debt service exceeds scheduled debt service

C. The borrower has no credit risk

D. The loan has zero LGD

Correct Answer: B. Cash flow available for debt service exceeds scheduled debt service

Question 9

Why should collateral normally be regarded as a secondary source of repayment?

A. Collateral can never be sold

B. Collateral replaces the need to assess cash flow

C. Realisation may be delayed, costly or uncertain

D. Collateral always increases PD

Correct Answer: C. Realisation may be delayed, costly or uncertain

Question 10

Which control best reduces the risk that sales incentives distort credit approval?

A. Removing all documentation requirements

B. Allowing relationship managers unrestricted approval authority

C. Separating relationship management from independent credit approval

D. Approving all applications automatically

Correct Answer: C. Separating relationship management from independent credit approval