With spring in the air, the U.S. clocks have gone forward, stealing an hour from us.
Obviously it is an illusory lost hour, but it does provoke an interesting thought experiment: if you were to actually lose an hour of the working day, where could you improve efficiency to make up for the lost time? In the risk management department, it would be a tough call. Let’s take a walk in a risk manager’s shoes for a day…
The average day starts with checking the overnight results. Here, time is at a premium as any obvious data errors need to be identified, and if the error is large enough, it can prompt a complete rerun of the risk numbers. This is, in part, due to the fact that risk reports are run against all levels of the hierarchy, with each new level being treated as a new entity.
If the base risk numbers could be run at the trade or position level, and the logic required to transform these into hierarchically aggregated risk reports done at the point of enquiry, it would dramatically reduce the time taken to correct errors. If only the trades or positions affected by the error needed to be rerun, as soon as the correction was in the system the right numbers would be running through the logic, creating correct risk reports.
Once the runs have been checked, there is the process of distribution. This will vary from firm to firm, but generally involves transplanting risk results into spreadsheets or reporting databases, and emailing or releasing the numbers once this is complete. The time criticality of this process is largely determined by a few factors:
- How much do the stakeholders use the risk numbers? If risk numbers are used as a strategic tool, then they need to be available before trading starts. If they are simply a reporting base, this is not as urgent.
- How much do the stakeholders trust the risk numbers? If the models for the risk or the pricing within the risk are not fully trusted, then it is likely that even with the best intentions, the risk reports ultimately will not be used by the stakeholders.
- How well do the stakeholders understand the risk numbers? Are the numbers created to reflect the risk takers’ view, senior management’s view, or is it the risk manager’s job to translate between the two?
If the risk calculation engine could send the results immediately to the reporting database, where they could be checked see above using the same distribution tools and dashboards used by the stakeholders, then this entire step could be removed. As corrections occurred, the results would be dynamically corrected as well, with potential operational errors involved in manual report distribution disappearing, and a simple traffic light system could be employed to show when a report was considered good.
Whilst we are here, we could throw in online risk model validation like a dynamic back-testing report, so that as well reducing the time to get the reports into the stakeholders’ hands, the doubts around the validity of the models could also be mitigated. By combining dashboards between desk levels, stress testing and enterprise level risk numbers, it would be possible to illustrate how these metrics impact each other and work together, significantly increasing the understanding of risk from the multiple points of view.
Once the reports are out, the stakeholders themselves may find errors and email the risk managers, who then check those errors and make adjustments if necessary. This may be time critical, based on the importance of the numbers to the stakeholders. The bigger risk of efficiency loss is the crossing of emails, and the obvious operational risk impact that has.
If the system’s dashboards could be used to message around the acceptability or otherwise, and be fully audited, this would remove a significant headache and potential extra time loss.
It would seem that by streamlining the calculation process and increasing efficiency in the production control and result distribution process, there would be a good chance of increasing the cultural adoption of risk. This involves putting new eyes on the system and its outputs, and improving the communication between those who set the risk appetite and those who consume it.
Something for the risk department to think about as they enjoy the lighter spring evenings.