ABSTRACT
When measuring operational risk, having enough data is one of the main concerns. Collecting data for operational risk is difficult. Managers might be worried first about whether the reference period is homogeneous with the current period, and second, whether all operational failures have been recorded. However, there is a third issue that usually escapes from the control of operational risk managers, and that is the amount of events that has been observed. If there are too many, then the manager might focus on frequency and tend to be more concerned about reducing operational risk events, but if there are too few events, then a statistical problems arises because the sample size of the available data may not be sufficient to draw conclusions that are statistically sound. Some of these issues will be dealt with in the chapter that discusses underreporting; however, we now want to concentrate on the combination of data.
