A systemic approach to operational risk measurement in financial institutions

ABSTRACT This paper presents an implementational systemic approach framework for operational risk (SAFOR), where the operational risk (OR) management’s target is to manage and mitigate the risk-around-loss causes. The guiding method is general systems theory used for describing essential system features of a complex domain. The important implications of the general systems theory for our work are that an organization, such as a bank, should be seen as a system, interacting with other systems through its boundaries. In order for this system to function efficiently, it needs to implement control structures, for example for giving feedback. While this is not new to organizational efforts, but the feedback in a setting of OR should be considered on two levels. The first level is the most immediate with feedback given more or less automatically when something goes wrong or is clearly less optimal than desired. This is, to a large extent, already present in most well-functioning banking institutions of today. The other level, feedback through a deliberation filter (such as a formal decision), is not as well covered. This paper considers deliberation feedback through investigating models for decision making under uncertainty and looking at suitable input data from OR measurements. Our OR framework for banks is surveyed as a systemic–holistic approach. A holistic treatment of the problem emphasizes the environment and the ongoing administration of the system in the implementation. Therefore, in the SAFOR model we propose an implementational guidance system that takes the whole bank’s organization, including its environment, into account. This means that a systemic–holistic approach may avoid a large fraction of the operational mistakes and errors that have their origin in the design of the systems, eg, people, machines, plant/environment and management. Moreover, to meet the mission of the bank’s safety and security, overall systems effectiveness must result in an efficiently operating control. Our SAFOR framework not only considers and discusses the OR models, it also includes a method for decision making. We emphasize its importance because reasoning about OR is often scenario based. For each scenario, value-at-risk (VaR) and conditional value-at-risk (CVaR), or other computational models, yield OR prediction intervals. Therefore, a decision must be made that takes into consideration what sort of actions that has to be taken to reach reasonable OR levels. However, for a small number of scenarios, there has to be a discriminating principle and methods and processes for its approach.