Credit risk is the risk that a borrower or counterparty fails to meet its obligations — missing interest or principal payments, or failing to deliver on a contract. It is the largest risk for most banks, and it matters to bond investors, suppliers who sell on credit and anyone who lends money. Modern credit risk management breaks the risk into measurable parts, so that losses can be estimated, priced, provisioned and backed by capital.
What you will learn
- Define the components of credit risk: PD, LGD and EAD
- Calculate expected loss for a loan and a portfolio
- Understand how collateral, pricing and limits manage credit risk
- Distinguish expected and unexpected loss and the role of capital
The building blocks
| Component | Meaning | Main drivers |
|---|---|---|
| Probability of default (PD) | The likelihood a borrower defaults within a period, usually one year | Financial strength, credit history, industry, economy |
| Loss given default (LGD) | The share of the exposure lost if default occurs, after recoveries | Collateral, seniority, recovery process and costs |
| Exposure at default (EAD) | The amount owed at the time of default | Loan balance plus likely drawings on unused limits |
Expected loss
Expected loss (EL) = PD × LGD × EAD
Worked examples
Worked example: expected loss on one loan
A bank lends 1,000,000 to a company. It estimates a one-year PD of 2% and an LGD of 45% (illustrative figures).
EL = 2% × 45% × 1,000,000 = 9,000.
The bank should expect, on average, to lose 9,000 a year on loans like this — a cost to be covered in the loan’s pricing and recognised in loan-loss provisions.
Worked example: a small portfolio
| Segment | EAD | PD | LGD | Expected loss |
|---|---|---|---|---|
| Corporate loans | 5,000,000 | 1% | 40% | 20,000 |
| SME loans | 2,000,000 | 4% | 55% | 44,000 |
| Retail loans | 3,000,000 | 3% | 75% | 67,500 |
| Total | 10,000,000 | 131,500 |
The portfolio’s expected loss is 131,500, or 1.3% of exposure (illustrative figures). Retail lending carries the largest expected loss here because it is unsecured and LGD is high.
Assessing PD
- Credit analysis of financial statements, cash flow, management and industry — see credit appraisal, solvency ratios and the Altman Z-score.
- Credit scoring models for retail borrowers, increasingly using machine learning — see AI in finance.
- Credit ratings from rating agencies or internal rating grades mapped to PD estimates.
- Market signals such as bond spreads and credit default swap prices for listed borrowers.
Reducing LGD: collateral and seniority
Collateral, guarantees and seniority lower the loss if default happens.
Worked example: LGD with collateral
A 1,000,000 loan is secured on property valued at 700,000. Allowing a 30% haircut for price falls and forced sale, expected recovery is 490,000; recovery costs are 50,000 (illustrative figures).
LGD = (1,000,000 − 490,000 + 50,000) ÷ 1,000,000 = 56%.
Without collateral, LGD on such a loan might be much higher. Haircuts must be realistic: collateral values often fall precisely when many borrowers default.
Pricing for credit risk
Risk-based pricing builds expected loss into the interest rate:
Loan rate ≈ Funding cost + Expected loss rate + Operating costs + Capital charge
With a 7% funding cost, an expected loss rate of 0.9% (2% PD × 45% LGD), operating costs of 1.5% and a capital charge of 1.2%, the minimum rate is about 10.6% (illustrative figures). Lending below this rate destroys value, even if the loan is repaid.
Expected and unexpected loss
Expected loss is an average cost, covered by pricing and provisions. Unexpected loss — losses above the average in a bad year — is covered by capital. Because defaults cluster in recessions, portfolio losses can be several times the expected level. Bank regulation sets capital requirements for credit risk under the Basel framework, using standardised risk weights or approved internal models of PD, LGD and EAD — see capital adequacy and Basel.
Accounting standards also require provisions: IFRS 9 uses an expected credit loss model, recognising 12-month expected losses for performing loans and lifetime expected losses when credit risk has increased significantly.
Managing portfolio credit risk
- Concentration limits on single borrowers, groups, sectors and regions.
- Diversification across industries and borrower types.
- Covenants in loan agreements that give early warning and rights to act.
- Monitoring and early warning — overdue payments, covenant breaches, falling ratings.
- Collections and recovery processes for non-performing loans.
- Risk transfer through guarantees, credit insurance or credit derivatives.
- Stress testing of portfolios against recession scenarios — see stress testing.
Counterparty credit risk
Credit risk also arises in trading: if a derivative counterparty defaults while a contract is in the money, the gain is lost. Collateral (margin), netting agreements and central clearing reduce this risk — see derivatives.
Credit risk for businesses
Companies selling on credit face the same risk with trade receivables. Credit limits per customer, credit checks, deposits, credit insurance and close monitoring of overdue invoices apply the same principles — see working capital management.
Common mistakes
- Relying on collateral instead of assessing the borrower’s ability to repay.
- Using optimistic collateral values without haircuts.
- Pricing loans without covering expected loss and capital costs.
- Allowing concentrations in one sector or group.
- Underestimating how defaults rise together in a downturn.
Key terms
- Probability of default (PD) — the likelihood that a borrower defaults within a period.
- Loss given default (LGD) — the share of exposure lost after recoveries.
- Exposure at default (EAD) — the amount owed when default occurs.
- Expected loss — PD × LGD × EAD, the average anticipated loss.
- Unexpected loss — losses above the expected level, covered by capital.
Frequently asked questions
Why do banks hold capital if they already provision for losses?
Provisions cover expected losses; capital absorbs unexpected losses in bad years so the bank remains solvent.
Does collateral eliminate credit risk?
No. It reduces loss severity, but values can fall, recovery takes time and costs, and legal processes may be slow.
How is credit risk different from default?
Credit risk includes the risk of default and of deterioration in credit quality, which reduces the value of loans and bonds even before default.
Key takeaways
- Credit risk is broken into PD, LGD and EAD; expected loss = PD × LGD × EAD.
- Collateral and seniority reduce LGD; analysis and scoring estimate PD.
- Loan pricing must cover funding, expected loss, costs and capital.
- Provisions cover expected loss; capital covers unexpected loss; limits and diversification control concentration.
Explore this in ARBIMETRA LAB
Apply these ideas to your own data. Files are analysed in your browser and are not uploaded.
- Logistic Regression: Binary outcomes and odds ratios.