Journal of Operational Risk

Marcelo Cruz

Lehman Brothers

Welcome to the first issue of the second volume of The Journal of Operational Risk.With this issue we start the second year of the Journal’s existence. Compared to a year ago, we are pleased to see an increasing number of high quality submissions.

Welcome to the first issue of the second volume of The Journal of Operational Risk.With this issue we start the second year of the Journal’s existence. Compared to a year ago, we are pleased to see an increasing number of high quality submissions. One year on from its inception and the Journal now has a healthy and desired “competition” for a slot in its issues. I am also thankful for the number of congratulatory messages on the event of our first anniversary. I can only thank you for the support and ask you for your continuing collaboration.

The Journal of Operational Risk is a vehicle for communicating results in the modeling and management of operational risk. The journal aims to be the forerunner in publishing articles in areas such as: statistical/actuarial methods and estimation issues, causal models, scenario analysis based models, Bayesian methods, uses of external data within the framework, etc. We also encourage submissions for publication on new ideas and research in subject areas such as corporate governance, business continuity plans, enterprise-wide risk, financial crime and the development of controls to avoid them, insurance, etc. This Journal also aims at being a new forum for discussions on this subject, facilitating the publication of top-quality operational risk technical papers. Research in operational risk is a field that is growing at a fast pace in both the financial industry and academia. There are currently many lines of research, most of them trying to overcome the challenges presented by the new regulatory standards created by the Basel II Accord. However, currently there is not a single forum for the debate of these ideas. The Journal of Operational Risk has been published to fulfill this much needed role.

At this moment I would like to invite you to really appreciate the high quality of the articles in this issue of The Journal of Operational Risk. We thank the authors for their trust in this young publication and appreciate all the encouraging messages sent by a number of colleagues in the industry and academia. All of this gives us the comfort that we are taking the right steps to build this new global industry forum for operational risk.

Research Papers

In this first issue of the second volume, there are three research papers in the main section. One of them shows an approach that allows the aggregation of the different types of data. The remaining two papers deal with the use of extreme value theory and heavy tails.

In the first paper, “A “toy” model for operational risk quantification using credibility theory”, Bühlmann et al apply Bayesian credibility to combine the internal losses, external data and scenario analysis into a single figure. Credibility theory seems to be emerging as a dominant thought on the aggregation of the different types of operational risk data; the focus of this paper is centered around the quantification of low frequency/high impact losses.

In the second paper, “Multivariate estimation for operational risk with judicious use of extreme value theory”, El-Gamal et al, realize that indiscriminate use of extreme value theory to all data would lead to unacceptably high capital estimates. In their view, practitioners need to use judgment in modeling the data; they present estimating capital by using copulas in a multivariate estimation of capital.

In the third paper, “Heavy-tailed distributional model for operational losses”, Giacometti et al analyze a real database received from a large European bank using statistical distributions in an actuarial type of approach. The authors claim that assuming a Poisson distribution for the frequency distribution fails as the frequency process closely follows a non-homogeneous Poisson process. They address a number of problems usually faced by the practitioner in dealing with modeling operational risk databases.

Operational Risk Forum

In this current issue, Schäl and Stummer provide a more rigorous, academic view on operational risk data modeling following Basel II standards. In the final article, Makarov shows two applications of the exact extreme value theorem.