Getting to the ‘Aha’ Moment for Useful Risk Metrics – Highlights of Discussion
from CCRO’s April 2012 Members’ Meeting
by Jim Pierobon
Yeager was
quick to point out that many risk managers tend to settle on a metric “before
they even know what question it is they’re trying to answer.” CCRO members
don’t want to calculate risk metrics just for the sake of having metrics. “We
typically calculate them because somebody has a burning question they need to
answer,” Yeager said.
It’s clear from the CCRO dialogue since then that a better practice – dare we say the best practice – is to start by agreeing not only about what question needs answering but how the answer will be useful to relevant colleagues throughout the company.
So instead
of moving to define something as general as ‘earnings at risk” what would be
really useful is how sensitive corporate earnings might be to five or six variables.
“If you
get people to set the math aside for a
second and just think about what it is they want to know, you can get an
amazingly robust list that can lead to helpful thoughts about metrics,” Yeager
said. “Your goal is to get the stakeholder who’s going to be making decisions
on this metric to say, “Aha, THAT’s what I was looking for. I can completely
believe the numbers.”
“It’s not about a particular metric, it’s about the insight that you’re looking for,” Yeager explained.
Some of the April meeting attendees cited the forward price risks inherent in Financial Transmission Rights, or FTRs.
Yeager
pounced on this is a good illustration of the challenge of crafting a useful risk metric:
“Is it the
movement between now and settlement? Do you want to understand it as a full
distribution? Do you want to understand fifth percentile thirty day?” Yeager
asked. “There are all sorts of different ways that you might go after that.”
Uploaded with permission from Wikimedia Commons; CREDIT AaCBrown.
Yeager
drilled down about FTRs to make his point. “There seems to be . . . an over-reaction of FTR auction prices
to the last few days of day-ahead prices. So the question is: ‘Does your model
even pick that up? Or does it just pick up the random standard deviation sort
of behavior of the FTR market.’ “
Anderson recalled while working a an energy company before assuming the helm of the CCRO how a “huge amount of money” was spent trying to devise a global Value at Risk (VaR) across all products and books.
“The senior executive wanted just one
number,” Anderson remembered. “Not surprisingly, they never did come up with
that. It was finally determined that such a metric was potentially
'misinformation',” he said, to much laughter among the meeting’s participants.
An unidentified speaker offered: “I
recall very well when you get in that trap that you were talking about. You
have traders that are not wrong, they’re just thinking differently about risk.
The risk manager, meanwhile, is the guy with the limits. They are unable to
reach an understanding.”
This attendee found it helpful to use “sharp ratios” to come up with a potential earnings range given a VaR limit - or earnings targets. In that context, it’s helpful to ask, “what's a reasonable VaR?” The risk manager can discuss with the trading manager how much are we going to make on these positions? What (are) your earnings going to look like.”
Yeager and
Anderson aspire for working group sessions along these lines that the Group and
CCRO members can rely on when risk managers inside a company present
the metric and results “to remind everyone,” Yeager said, “why we wound up
where we did.”
