September 2014 Q3.ii

[DC92]

Outline how ABC Life could measure its operational risk exposures.

I would have answered the question by looking at 'low frequency, high severity' and 'high frequency, low severity' operational risk events.

For the former I would have discussed using extreme value theory, and mentioned both the block maxima method as well as the threshold exceedance method, stating that a GEV distribution and GP distribution could be fitted to those methods, respectively.

For high frequency low severity operational risk events I would have suggested using non-life reserving and pricing techniques. More data available, and the data should be more stable.

Also would have mentioned that operational risks are dependent on the economic cycle (Eg fraud is more likely to occur in a recession).

My approach is quite different from the one in the Examiners Report. Would I receive marks for going with this approach?

Thanks
DC92

 

[Simon James]

Hi. You would certainly have scored some marks. The key point is that there is likely to be little data, so quantitative techniques may be impractical, so discussing Extreme Value techniques may be limited to 1/2-1 mark as EVT requires data to fit a distribution.

It's not immediately obvious to me why using non-life techniques would mean more data is available?

If you have positioned your point about the recession well to ensure it answers the question (Eg "operational risks are more likely to occur in a recession so the state of the economic cycle can be used an indicator of operational risk exposure"), then you would gain marks.