Credit Risk Capital Modelling Under Basel Internal Ratings Based Approach (IRB)

Day 1

Overview – Capital Requirements For Credit Risk

• Capital requirements for credit risk – Standardised Approach (SA) and Internal Ratings Based Approach (IRB)

• Revised SA and SA capital floor to IRB

• IRB – supervisory formula to measure credit risk capital requirement (Pillar 1 minimum 8% of risk-weighted assets) – logic and weaknesses

• Key IRB parameters and how they affect credit risk capital requirement (KIRB) – probability of default (PD), loss given default (LGD), exposure at default (EAD), effective maturity (M), correlation (ρ) and granularity (g)

• IRB – Foundation (FIRB) and Advanced (AIRB)

• Case Study 1 – Overview of banks' models for estimating PD, LGD and other IRB parameters (where applicable)

• Where credit risk arises on banking book – loans, undrawn credit facilities, credit guarantees and securitizations (counterparty credit risk grouped more with market risk)

• Why loan book creates additional credit risk capital requirements beyond Pillar 1 – stress testing (Pillar 2 Guidance), concentration / large exposures and IFRS 9 expected credit loss (ECL) provisioning

• Basel 3 through-the-cycle 12 month (TTC) PD vs IFRS 9 point-in-time (PIT) PD (12 month or lifetime) – interaction of ECL accounting under IFRS 9 and Basel 3 for IFRS 9 Stage 1, Stage 2 and Stage 3 loans

• Case Study 2 – How much extra capital for credit risk is needed to cover IFRS 9 ECL volatility?

PD Models

• Construction of ratings transition matrices, including continuous time homogeneous method

• Case Study 3 – Transition matrix PD model – illustrated with corporate loan portfolio

• Case Study 4 – Logit and probit PD models – illustrated with corporate loan portfolio

• Case Study 5 – Linear discriminant PD model – illustrated with corporate loan portfolio

• Case Study 6 – Linear regression PD model – illustrated with corporate loan portfolio

• Case Study 6 – Cox proportional hazard PD model – illustrated with retail loan portfolio

• Case Study 7 – Scoring PD model – illustrated with retail loan portfolio

Day 2

LGD Models

• Recovery rate data

• LGD cyclicality

• Taking into account collateral and guarantees received

• Case Study 8 – Regression LGD model, taking into account collateral and guarantees received and macroeconomic factors

Other IRB Parameters – Modelling / Measurement / Basel-Specified

• EAD – treatment of undrawn credit facilities

• M – effective maturity measurement, floor and cap

• ρ - Basel specification

• g – measurement

• Case Study 9 – Example calculations of EAD, M, ρ and g

Calculation Of KIRB

• Case Study 10 – Calculation of AIRB Pillar 1 credit risk capital requirement for a mixed retail and corporate loan book

PD and KIRB In Capital Stress Test

• Linking transition matrices to macroeconomic covariates

• Case Study 11 – Macroeconomically-conditioned continuous time homogeneous method transition matrices to stress PD in capital stress test. Resulting change in KIRB vs baseline

• Conducting capital stress test under IFRS 9 ECL. Dealing with loan migration between Stages 1, 2 and 3, resulting changes in ECL measurement horizon and probability-weighted averaging over multiple scenarios

• Case Study 12 – Calculating IFRS 9 ECL provisions in a capital stress test

IRB vs SA

• How different are KIRB and Pillar 1 SA credit risk capital requirement (KSA)?

• Case Study 13 – What is likely impact on credit risk capital requirement for IRB banks of revised credit risk SA and capital floor?

• Case Study 14 – Comparison of IRB vs SA for hypothetical portfolio

• Pros and cons of moving from SA to IRB

• IRB model approval process

• Case Study 15 – Attempted IRB model applications – Metro Bank and other UK specialist banks / challenger banks

• Case Study 16 – What percentage of banks’ loans are on IRB vs SA?