SAS Credit Risk Modelling - LGD
SAS Credit Risk Modelling - LGD: A to Z of LGD Modelling
Loss Given Default (LGD) is the most end-to-end driven component of credit risk modelling. Unlike PD or EAD, LGD is not a single modelling step-it is a complete lifecycle that starts with raw data and ends with deployment, monitoring, and regulatory defence.
Most LGD material available today focuses only on the final estimation technique. It assumes that default definition, recovery construction, discounting, write-offs, restructuring, governance, and deployment are already correct or not critical. In real banks, this assumption is wrong.
LGD models fail far more often because earlier steps were incomplete or poorly designed than because of weak modelling techniques.
This book exists to close that gap.
SAS Credit Risk Modelling - LGD: A to Z of LGD Modelling is written to cover every step of LGD credit risk modelling from start to finish, exactly as it is performed in real banking environments.
What "A to Z" Means in This Book
"A to Z" in this book is literal.
This book documents the entire LGD lifecycle, including steps that are rarely explained together in one place, but are routinely reviewed by validators, auditors, and regulators:
LGD foundations and regulatory context
Governance, ownership, and model lifecycle
LGD-specific data architecture and source systems
Default definition and default event construction
Recovery cashflow identification and validation
Recovery timing and present-value discounting
Write-offs, collateral haircuts, and net recovery
Restructuring and forbearance treatment
Observed LGD target construction
Observation and recovery windows
LGD-specific feature engineering
Binning and variable reduction for LGD
Stability and leakage checks
LGD modelling techniques used in banks
Full SAS implementation and reproducibility
Validation, calibration, downturn LGD, and stress testing
Deployment, monitoring, and recalibration
Integration with PD, EAD, IFRS 9, and Basel
A complete end-to-end bank-style case study
Every stage that materially affects LGD outcomes is covered.
Why This Book Is Different
This book does not isolate LGD modelling from the processes around it.
It treats LGD as a system, not a formula.
It explains not only how each step is done, but why banks do it that way-and what goes wrong when they don't.
All examples are synthetic and fictional, but the workflows, controls, and decisions reflect real banking practice.
Why This Book Uses SAS
LGD models in real banks are built and governed in SAS environments because of transparency, auditability, and regulator acceptance.
This book therefore uses clear, production-style SAS code throughout, with no black boxes and no unexplained shortcuts.
How to Read This Book
This book is designed to be read sequentially.
LGD is cumulative-each step depends on earlier decisions. Skipping chapters risks misunderstanding downstream logic, just as in real LGD projects.
Final Note
This book is written for practitioners who want to understand LGD A to Z, not just how to estimate it, but how to build, validate, deploy, monitor, and defend it in real banks.
Everything that follows is designed for that purpose.
Customer Reviews
ThriftBooks ® and the ThriftBooks ® logo are registered trademarks of Thrift Books Global, LLC
