Applying fuzzy scenarios for the measurement of operational risk

Abstract

Operational risk measurement assesses the probability to suffer financial losses in an organisation. The assessment of this risk is based primarily on the organisation’s internal data. However, other factors, such as external data and scenarios are also key elements in the assessment process. Scenarios enrich the data of operational risk events by simulating situations that still have not occurred and therefore are not part of the internal databases of an organisation but which might occur in the future or have already happened to other companies. Internal data scenarios often represent extreme risk events that increase the operational Value at Risk (OpVaR) and also the average loss. In general, OpVaR and the loss distribution are an important part of risk measurement and management. In this paper, a fuzzy method is proposed to add risk scenarios as a valuable data source to the data for operational risk measurement. We compare adding fuzzy scenarios with the possibility of adding non fuzzy or crisp scenarios. The results show that by adding fuzzy scenarios the tail of the aggregated loss distribution increases but that the effect on the expected average loss and on the OpVaR is lesser in its extent.