GARP FRM Part II Credit Risk Measurement and Management Study Guide

Quick answer: Study the 23 chapters in four connected blocks: credit foundations and borrower assessment; default probability and portfolio loss; derivatives and counterparty exposure; and collateral, xVA, stress testing and securitisation. For every measure, learn the object, horizon, probability type, dependence assumption and residual risk.

The GARP FRM Part II Credit Risk Measurement and Management book is not mainly a list of formulas. It tests whether you can choose the right credit object before calculating. A PD may describe an obligor, LGD a facility, EAD an exposure at default, Credit VaR a portfolio loss quantile and CVA a market-consistent expected counterparty loss. Mixing those objects produces plausible but wrong answers.

This is one printed Part II book. Complete it as a credit-risk unit, then combine it with the other Part II books and Current Issues for full-exam practice.

Block 1: Credit Foundations, Governance and Assessment

Chapters 1–8 build the operating framework.

Start with the basic distinctions. Insolvency is a financial condition, default is failure to meet a contractual obligation and bankruptcy is a legal process. Credit risk also extends beyond loans: settlement, trade credit, guarantees, derivatives and contingent commitments can create exposure.

Governance then controls how risk enters the portfolio. Learn the three lines of defense and the relationship between origination, independent assessment, approval, limits and oversight. A credit committee adds collective challenge; it does not remove delegated accountability.

The management and capital chapters connect policy with measurement:

• lending policy defines eligible borrowers, products and approval rules;
• connected and concentrated exposures must be aggregated;
• IFRS 9 staging changes the expected-loss horizon;
• expected loss belongs in pricing and provisions;
• unexpected loss drives economic capital at the selected solvency standard.

Chapters 5–7 then compare judgmental, empirical and financial models, scoring with rating, TTC with PIT, issuer with issue rating, and discrimination with calibration. CAP or accuracy ratio can show ranking power without proving that predicted PDs are calibrated.

Sovereign risk closes the block. Compare foreign-currency and local-currency default, then triangulate fiscal and institutional evidence with ratings, bond spreads and CDS. A market spread is not pure default probability because recovery, liquidity and risk premium matter.

Block 2: Default Probability, Credit VaR and Portfolio Dependence

Chapters 9–12 move from one obligor to a portfolio.

Default probabilities may come from historical rating transitions, hazard rates, market spreads or the Merton firm-value model. Keep three distinctions explicit:

1. marginal default during an interval versus cumulative default by a horizon;
2. default conditional on survival versus unconditional default;
3. real-world PD for forecasting versus risk-neutral PD for valuation.

Credit VaR then applies a loss distribution. State the horizon, confidence level and whether the reported quantity is total loss or unexpected loss above expected loss. Vasicek, CreditRisk+ and CreditMetrics are not interchangeable: they use different state variables, dependence structures and loss construction.

Portfolio risk introduces a second common mistake. Asset correlation is not observed default correlation. A one-factor model uses latent asset returns to produce conditionally independent defaults given the common factor. Default-indicator correlation follows from the joint and marginal default probabilities. More granular exposures reduce name-specific concentration, but systematic risk remains.

Block 3: Credit Derivatives and Counterparty Exposure

Chapters 13–16 shift from loan-style loss to contracts whose value changes through time.

A CDS exchanges a premium leg for credit-event protection. Standard coupons can create an upfront payment, so the coupon is not automatically the current market spread. Credit indices, CDS forwards and options, basket products and synthetic CDOs add spread and dependence effects. Base and compound correlation are model-implied quotes, not directly observed properties of default.

For derivatives, counterparty exposure is positive replacement cost—not gross notional. A loan is normally one-directional; a derivative can switch value direction as markets move. Separate:

• current replacement cost;
• potential future exposure during the contract;
• settlement exposure around payment exchange;
• margin-period exposure during default management and close-out.

Netting also needs precise language. Payment netting combines qualifying cash flows. Close-out netting converts covered transactions into one net payable or receivable after termination. Compression reduces redundant gross notional while preserving agreed market exposure. None of those results is valid without the relevant contractual and legal enforceability.

Block 4: Collateral, Clearing, Exposure, CVA and Structured Credit

Chapters 17–23 should be studied as one chain.

A Credit Support Annex defines eligible collateral, valuation, threshold, minimum transfer amount, haircuts, margin timing and disputes. Variation margin covers current mark-to-market; initial margin protects potential movement during close-out. Collateral reduces unsecured credit risk but creates funding, liquidity, operational and legal risk. Rehypothecation can lower funding cost while increasing collateral-return risk; segregation does the reverse trade-off.

Central clearing transforms the bilateral network. Novation places the CCP between members. Initial margin, default-fund contributions and the loss waterfall absorb member-default loss in a prescribed order. The CCP improves multilateral netting and standardization but concentrates model, liquidity and recovery risk.

Future-exposure metrics then feed CVA:

• EE is average positive exposure at one future date;
• PFE is a high percentile at one future date;
• EPE averages EE through time;
• effective EPE prevents exposure reductions from being recognized too early.

CVA combines exposure, marginal default probability, LGD and discounting. DVA adds own-default effects, and BCVA treats both sides consistently. Incremental CVA is the finite portfolio change from a trade; marginal CVA is a small-change or allocation concept. Wrong-way risk appears when exposure is high when counterparty credit quality is poor.

The final chapters apply these ideas to stress and structure. A CCR stress must revalue the portfolio and credit conditions without double counting market loss. Structured-credit and securitisation questions require the contractual waterfall: junior protection absorbs loss before senior claims, excess spread and overcollateralization supply enhancement, and correlated defaults can reach senior tranches after junior support is exhausted.

A Direct Revision Routine

For each chapter, write six lines:

1. Object: obligor, facility, exposure, portfolio or tranche.
2. Measure: PD, EL, UL, Credit VaR, PFE, CVA or another named output.
3. Horizon: date, interval or lifetime.
4. Probability: real-world or risk-neutral; marginal, conditional or cumulative.
5. Dependence: netting, correlation, copula or waterfall assumption.
6. Residual risk: what the model or mitigant leaves behind.

Then practise the close pairs: EL versus economic capital, TTC versus PIT, issuer versus issue rating, asset versus default correlation, structural versus reduced-form default, CDS coupon versus spread, notional versus exposure, variation versus initial margin, PFE versus EE, incremental versus marginal CVA, and stressed expected loss versus stress loss.

Move to full Part II practice when you can explain each pair without notes and connect scoring to PD, PD to Credit VaR, factor dependence to joint default, exposure profiles to CVA, and pool loss to tranche waterfall. Continue through the remaining books from the GARP FRM exam-preparation hub.