Measuring Operational and Reputational Risk A Practitioner's Approach

by Soprano, Aldo; Crielaard, Bert; Piacenza, Fabio; Ruspantini, Daniele

• ISBN13: 9780470517703
• ISBN10: 0470517700
• Edition: 1st
• Format: Hardcover
• Copyright: 2009-04-20
• Publisher: Wiley

Summary

This book shows practitioners the best models to use in a given situation, according to the type of risk an organization is facing. It is based on extensive applied research on operational risk models, testing results on Unicredit datasets. Theoretical models and available research do not include a direct testing on real databases, as banks' relevant information isn't available to the general public and academia. For operational risk this remains a key challenge in the development of models correctly interpreting the risk structure and elements. UniCredit operational risk team, formed by experienced former academics and finance practitioners, have applied and tested various models and fitting techniques.

Author Biography

ALDO SOPRANO is Managing Director, UniCredit Basel 2 project manager for Central and Eastern European countries. Previously he was Group Head of Operational Risk Management. A Graduate in Economics, he holds a Master in Finance. In his career he has also been responsible for market risk management, credit risk control, capital allocation and more recently Chief Risk Officer of UniCredit Kazakhstan. He is the author of several articles on risk management and was the Chairman of the International Institute of Finance’s Working Group on Operational Risk.

BERT CRIELAARD works in the Operational Risk Department of UniCredit (Holding) and is group-wide responsible for operational risk management in the Corporate, Private Banking and Asset Management business divisions. Previously he worked in the insurance and asset management industry in Italy and the Netherlands. He is (co-)author of articles on insurance in risk management.

FABIO PIACENZA is a senior quantitative analyst at UniCredit Group Operational Risk Management in Milan. Graduated in mathematics, he is author of several articles on operational risk related topics.

DANIELE RUSPANTINI works in UniCredit Group Milan in the Operational Risk Management team, graduated in mathematics, he is co author of articles on quantitative risk management.

Table of Contents

Foreword	p. xi
Preface	p. xiii
Acknowledgments	p. xvii
The Development of ORM in UniCredit Group	p. 1
A brief history of a fast-growing group	p. 1
Creating a new function	p. 2
Developing the new control system	p. 3
Challenges in the early stages	p. 4
Methodology to measure operational risk	p. 4
Training and internal communication focus	p. 5
International regulatory challenges	p. 6
Reputational risk management	p. 7
The Calculation Dataset	p. 9
Definitions	p. 9
Rules of thumb	p. 10
Internal loss data	p. 12
Business line mapping	p. 12
Event type classifications	p. 14
Data quality analysis	p. 18
Special cases	p. 18
Minimum loss threshold	p. 19
External data	p. 20
Public or external data sources	p. 21
Consortium data	p. 22
Scenario data	p. 23
Business environment and internal control factors	p. 23
Scenarios	p. 24
Insurance information	p. 24
Scaling data	p. 25
The Unicredit Group Operational Risk database evolution	p. 26
Final considerations	p. 26
Loss Distribution Approaches	p. 27
Calculation dataset building	p. 29
Internal calculation dataset	p. 29
External calculation dataset	p. 31
Scenario-generated calculation dataset	p. 32
Risk indicators calculation dataset	p. 32
General LDA framework	p. 32
Operational risk classes	p. 35
Identically distributed risk classes	p. 35
Inflation adjustment	p. 38
Data independence	p. 39
Parametric estimation and goodness-of-fit techniques	p. 41
Severity distributions	p. 41
Graphical methods	p. 44
Analytical methods	p. 44
Frequency distributions	p. 48
Applying extreme value theory	p. 50
g-and-h distribution theory	p. 53
Calculating operational capital at risk	p. 56
Loss severity distribution	p. 58
Loss frequency distribution	p. 61
Annual loss distribution	p. 63
Single class capital at risk	p. 66
Insurance modeling	p. 66
Appropriate haircuts reflecting the policy's declining residual term	p. 68
Payment uncertainty	p. 68
Counterparty risk	p. 71
Application of insurance	p. 72
Adjustment for risk indicators	p. 73
Operational risk classes aggregation	p. 75
Copulae functions	p. 76
Elliptical copulae	p. 78
Archimedean copulae	p. 81
Choice of copula	p. 84
Correlation coefficients	p. 84
The closed-from approximation for OpVaR	p. 86
Effect of the minimum threshold on capital at risk	p. 88
Confidence band for capital at risk	p. 89
Stress testing	p. 91
Loss data minimum threshold setting	p. 91
Empirical application on Algo OpData	p. 92
Descriptive statistics	p. 95
Autocorrelation analysis	p. 95
Capital at risk estimates using parametric models	p. 99
Capital at risk estimates using EVT	p. 116
Capital at risk estimates using the g-and-h distribution	p. 129
Capital at risk estimates considering correlation	p. 133
Regulatory capital requirement	p. 136
The consolidated capital requirement	p. 136
The individual capital requirement	p. 137
Economic capital requirement	p. 140
Integration of Operational risk in the budgeting process	p. 145
Analyzing Insurance Policies	p. 147
Insurance management and risk transfer	p. 147
Qualifying criteria in the Basel 2 capital Framework	p. 148
Rating of the insurance company	p. 149
Duration and residual term of the insurance contract	p. 150
Policy termination requisites	p. 151
Claims reimbursement uncertainty and ineffective coverage	p. 152
Conclusions	p. 152
A practical application to traditional insurance	p. 153
Insurance policies to cover financial institutions' operational risks	p. 153
Operational event types and available insurance coverage	p. 155
Managing Reputational Risk	p. 159
Introducing reputational risk	p. 159
A financial institution's reputational risk exposure	p. 160
Managing reputational risk: a matter of policy	p. 162
Reputational risk measurement	p. 164
Reputational risk as a function of share price volatility	p. 164
Measuring reputational risk using scenarios	p. 173
Scoring-card-based models for reputational risk assessment	p. 177
A recent example of reputational event	p. 178
A description of the event	p. 179
Background	p. 179
How the fake trading occurred	p. 179
The discovery and first reactions	p. 180
Measures planned and taken	p. 182
Immediate consequences for SocGen	p. 183
Reputational issues and comments	p. 183
The lessons learned - what can we do to avoid being next?	p. 186
Psychological, 'soft' factors	p. 186
Control instruments	p. 187
Managing data and signals	p. 187
Conclusions	p. 189
References	p. 193
Further reading	p. 195
Index	p. 201
Table of Contents provided by Ingram. All Rights Reserved.