99908 - CREDIT RISK

Learning outcomes

Credit risk is the main source of risk for a commercial bank, far more relevant then other risk types like market, interest, operational, etc.. and can have dramatic systemic effects, as an under estimation of credit risk is at the root of the financial and economic turmoil of the 2008. The course will provide a broad overview of credit risk management in (commercial) banking. First, we will review the methodologies behind the more widely used (structural and reduced form) models for individual credit risk components (EAD,PD,LGD); then we will analyse how credit risk is measured at portfolio level (expected and unexpected losses, risk aggregation and contributions) by the most popular metrics (the Moody’s KMV model, CreditMetrics™ and Credit Risk Plus™). Finally, we will address credit risk regulatory topics, both in terms of capital requirements (Basle II and III) and of accounting standards (IFRS9). The theoretical insight will be paired with practical exercises: some of these models and metrics will be applied to real credits data using the Python scientific stack.

Course contents

This extensive program bridges academic learning and professional practice. It provides the foundation for rigorous education in credit risk management and equips students with competencies demanded by today’s financial and consulting industry, both in domestic as well as international business environment.

The course aims at providing a comprehensive view on Credit Risk modeling for financial institutions. This program is designed to provide both theoretical knowledge and practical skills, ensuring students are prepared to pursue careers in financial institutions or consulting in the risk management area, with specific focus on quantitative aspects of credit risk modeling. The course aims at providing not only the theoretical knowledge, but also the instruments to be successfully introduced in the risk management world in terms of attitude, analytical skills, problem solving capacity and team work.

 

1. Aim of the course

The program articulates outcomes in terms of knowledge and skills and is totally integrated in the other topics included in the Master in Quantitative Finance program, aimed at providing quantitative instruments to be applied in finance and risk management.

Knowledge

• Mastery of credit risk fundamentals, including expected, unexpected, and extreme losses.
• In-depth understanding of regulatory frameworks: Basel accords, Single Supervisory
• Mechanism (SSM), IFRS9, EBA/ECB guidelines.
• Knowledge of core statistical and econometric approaches applied to risk analysis.
• Understanding of advanced credit risk models: Probability of Default (PD), Loss Given Default (LGD), Exposure at Default (EAD).
• Familiarity with machine learning methodologies and their role in credit risk modeling.
• Insights into ESG and novel risk integration in supervisory contexts.
• Knowledge of validation frameworks for credit risk systems, ensuring compliance with supervisory expectations.

Skills

• Ability to design, calibrate, and validate internal risk models in compliance with regulatory standards.
• Capability to integrate forward-looking macroeconomic scenarios into credit risk models.
• Skills in conducting stress tests and interpreting results for decision-making.
• Competence in building Satellite Models for proactive portfolio management.
• Collaborative experience through case studies and group projects.
  
  

2. Course Structure

The program spans for 30 hours in persons[1] lessons. Each module is progressively structured to introduce, consolidate, and advance students’ knowledge. The program is based on eight core modules which collectively cover the breadth of modern credit risk modeling. Modules combine lectures, applied case studies and group projects. In some cases, practitioners from consulting credit risk modeling teams will be invited for the lecture.

During the courses, alongside the theoretical concepts, cases and exercises will be proposed, to be performed both ad individual and group level.

Students not attending lessons are expected to study the same material discussed and suggested in this document as well as the material distributed during the lessons.

 

3. Module Descriptions Introduction to Credit Risk

This module provides the foundation for the entire program. It begins with a detailed analysis of credit risk as the possibility of financial loss due to borrower default. Students are introduced to key financial instruments that embed credit risk, the structure of lending contracts, and the concept of default as defined by regulatory bodies. The course situates credit risk within both managerial and regulatory contexts, ensuring students understand its dual purpose: value creation for banks and systemic stability for economies.

Students explore expected, unexpected, and extreme losses, learning to compute and interpret them. Special focus is placed on Basel requirements, the three-pillar framework, and IFRS9 standards. The module concludes with an introduction to risk parameters (PD, LGD, EAD) that underpin subsequent modules. Practical sessions involve team presentations on risk metrics, computation of CET1 ratios, and critical discussion of recent regulatory changes.

PD Models

The Probability of Default (PD) module combines theoretical depth with practical implementation. Students are introduced to the historical development of PD modeling, starting with statistical classification methods and progressing to advanced logistic regression and survival models. The course emphasizes three phases: estimation, integration, and calibration. Participants learn to work with diverse datasets, identify risk drivers, and segment portfolios into meaningful groups.

Practical sessions involve exercises. The role of overrides, group logic, and margins of conservatism are examined in depth.

LGD Models

Loss Given Default (LGD) models are critical for assessing recovery potential and regulatory capital requirements. This module covers both theoretical underpinnings and applied techniques. Students are introduced to LGD as the portion of exposure not recovered during the default process. The course distinguishes between performing and defaulted assets, emphasizing the importance of modular versus single-step estimation approaches.

The discussion will go through the construction of LGD models using workout data, discounted recovery cash flows, and regression-based estimation methods. Calibration is addressed through treatment of open and ceded positions, downturn adjustments, and extraordinary disposals. Exercises will be discussed during the module. Group projects focus on designing and presenting LGD models to hypothetical bank committees.

EAD Models

Exposure at Default (EAD) measures the outstanding exposure at the time of default. This module introduces students to Credit Conversion Factors (CCFs) and other key metrics used to project exposures. The regulatory definitions of on- and off-balance sheet exposures are explained, as well as managerial implications for capital planning.

Students learn to build EAD models step by step: from defining variables, through statistical selection of drivers, to calibration against observed long-run averages. Emphasis is placed on downturn adjustments and treatment of margins of conservatism. Workshops involve exercises.

Satellite Models

Satellite models link credit risk parameters (PD, LGD, EAD) with macroeconomic indicators to create forward-looking risk assessments. This module examines their role in stress testing, IFRS9 provisioning, and ICAAP exercises. Students learn to design PD satellite models, conduct univariate and multivariate analysis, and apply sensitivity testing across macroeconomic scenarios.

Case studies include the impact of GDP, unemployment, and housing price indices on credit risk. Students also learn how satellite models were applied during the COVID-19 crisis and how they are evolving to integrate climate and ESG risks. The module includes exercises.

IFRS 9 Models

This module merges provisions estimation and staging allocation, providing students with a comprehensive understanding of IFRS9 requirements. Students learn the conceptual shift from incurred to expected loss models, understanding the significance of staging allocation and lifetime expected loss. The course covers construction of Point-in-Time and Through-the-Cycle matrices, scenario conditioning, and integration with PD, LGD, and EAD models.

Practical sessions and case studies on portfolio staging. Students critically evaluate the challenges IFRS9 introduced, including sensitivity to macroeconomic forecasts and alignment with Basel capital requirements.

Machine Learning for Credit Risk

Machine learning (ML) is transforming credit risk modeling. This module introduces students to the framework for developing ML models. Topics include supervised classification and regression algorithms, ensemble methods, bias-variance trade-offs, and interpretability tools such as SHAP values. Students also explore the application of ML to transactional credit risk models, including time series-based probability of default estimations.

Students learn to critically assess ML results, balance accuracy with interpretability, and consider supervisory expectations regarding transparency.

Credit Risk Model Validation and Regulatory Review

The final module focuses on model validation, ensuring reliability and compliance with supervisory standards. Students study regulatory requirements (ECB, EBA, Basel IV), internal model investigations, and annual validation reporting. Both quantitative and qualitative validation techniques are explored, including back-testing, benchmarking, and representativeness analysis.

Workshops provide experience with validation real cases. Special attention is given to IFRS9 validation requirements and integration of stress testing.

[1] Some online lessons might be organized upon need.

Readings/Bibliography

Mandatory:

• Slides and exercises discussed during the lessons.
  

 

Recommended:

• Basel II/III/IV regulations and EBA/ECB guidelines.
• IFRS9 standards.
  
  

Additional:

• Bellini, T. IFRS 9 and CECL Credit Risk Modelling and Validation

Teaching methods

The program adopts a blended approach:

• Theoretical in-person[1] lectures provide academic foundations.
• Exercises for real life problems.
• Group projects replicate professional consulting assignments.

[1] Some online lessons might be organized upon need.

Assessment methods

The assessment is based on an oral exam, with a duration of 20-30 minutes on average and 3-5 questions or exercises proposed to the student.

The final outcome will depend on how many questions will be answered and the quality of the answers and will ai at assessing both the student’s theoretical preparation as well as the ability to solve case studies and exercises.

To solve the exercises, Excel or calculators can be used.

The attendance of in person lessons is considered a significant plus.

Students are not requested to present any material before the oral exam, even though they are strongly recommended to solve the exercises proposed in the slides discussed during the lessons.

The final outcome of the exam is the typical grade 0-30L grade.

To register for the exam, please follow the standard procedure on Almaesami.

Teaching tools

Specific MS Teams meetings can be requested by students to attend online, in case of impossibility to attend in-person lessons. The request must be delivered to the Professor at maximum the day before the lesson.

Office hours

See the website of Marco Stella