Operational Risk Modeling in Financial Services

PATRICK NAIM is the CEO of Elseware and widely recognized as an expert for operational risk modeling and quantification. Patrick has extensive experience in advising banks, insurance and energy companies for over 20 years in Continental Europe, the United Kingdom, and North America. He is also the author of Risk Quantification: Management, Diagnosis and Hedging and Bayesian Networks: a Practical Guide to Applications, both from Wiley. LAURENT CONDAMIN, PHD, is Managing Partner and Researcher at Elseware. For the past 10 years, he has been advising the largest financial institutions. His areas of expertise are operational risk modeling, stress testing, credit rating modeling, project risk analysis, insurance coverage optimization and cost-benefit analysis.

• List of Figures xiList of Tables xvForeword xixPreface xxiPart One Lessons Learned in 10 Years of PracticeChapter 1 Creation of the Method 31.1 From Artificial Intelligence to Risk Modelling 31.2 Model Losses or Risks? 5Chapter 2 Introduction to the XOI Method 72.1 A Risk Modelling Doctrine 72.2 A Knowledge Management Process 82.3 The eXposure, Occurrence, Impact (XOI) Approach 92.4 The Return of AI: Bayesian Networks for Risk Assessment 10Chapter 3 Lessons Learned in 10 Years of Practice 133.1 Risk and Control Self-Assessment 133.2 Loss Data 243.3 Quantitative Models 303.4 Scenarios Workshops 363.5 Correlations 413.6 Model Validation 47Part Two Challenges of Operational Risk MeasurementChapter 4 Definition and Scope of Operational Risk 574.1 On Risk Taxonomies 574.2 Definition of Operational Risk 68Chapter 5 The Importance of Operational Risk 715.1 The Importance of Losses 715.2 The Importance of Operational Risk Capital 745.3 Adequacy of Capital to Losses 76Chapter 6 The Need for Measurement 776.1 Regulatory Requirements 776.2 Nonregulatory Requirements 82Chapter 7 The Challenges of Measurement 937.1 Introduction 937.2 Measuring Risk or Measuring Risks? 937.3 Requirements of a Risk Measurement Method 957.4 Risk Measurement Practices 98Part Three The Practice of Operational Risk ManagementChapter 8 Risk and Control Self-Assessment 1058.1 Introduction 1058.2 Risk and Control Identification 1078.3 Risk and Control Assessment 113Chapter 9 Losses Modelling 1219.1 Loss Distribution Approach 1229.2 Loss Regression 134Chapter 10 Scenario Analysis 13710.1 Scope of Scenario Analysis 13710.2 Scenario Identification 15010.3 Scenario Assessment 163Part Four The Exposure, Occurrence, Impact MethodChapter 11 An Exposure-Based Model 17911.1 A Tsunami Is Not an Unexpectedly Big Wave 17911.2 Using Available Knowledge to Inform Risk Analysis 18011.3 Structured Scenarios Assessment 18111.4 The XOI Approach: Exposure, Occurrence, and Impact 182Chapter 12 Introduction to Bayesian Networks 18512.1 A Bit of History 18512.2 A Bit of Theory 18612.3 Influence Diagrams and Decision Theory 18712.4 Introduction to Inference in Bayesian Networks 18712.5 Introduction to Learning in Bayesian Networks 189Chapter 13 Bayesian Networks for Risk Measurement 19113.1 An Example in Car Fleet Management 191Chapter 14 The XOI Methodology 20314.1 Structure Design 20314.2 Quantification 20914.3 Simulation 214Chapter 15 A Scenario in Internal Fraud 21915.1 Introduction 21915.2 XOI Modelling 219Chapter 16 A Scenario in Cyber Risk 22716.1 Definition 22716.2 XOI Modelling 234Chapter 17 A Scenario in Conduct Risk 23917.1 Definition 23917.2 Types of Misconduct 24117.3 XOI Modelling 246Chapter 18 Aggregation of Scenarios 25518.1 Introduction 25518.2 Influence of a Scenario on an Environment Factor 25718.3 Influence of an Environment Factor on a Scenario 25818.4 Combining the Influences 26118.5 Turning the Dependencies into Correlations 262Chapter 19 Applications 26519.1 Introduction 26519.2 Regulatory Applications 26719.3 Risk Management 278Chapter 20 A Step towards “Oprisk Metrics” 28720.1 Introduction 28720.2 Building Exposure Units Tables 28820.3 Sources for Driver Quantification 28920.4 Conclusion 290Index 291