So, if dollar offset is likely to
cause failed hedges, what else is the market
using? So far, practitioners have noted regression
and less-frequently scenario analysis as the most common
alternative methodologies.
Regression, however, really
encompasses a set of widely diverging methodologies.
There’s plenty of room for maneuvering, with multiple
variables that each company can define for each set of
hedges. “Mathematicians like regression,” notes Leda
Braga, head of Cygnifi’s valuation services. “It’s more
complex but it’s also a more elegant method.”
Cygnifi identifies the following
variables that need to be defined in choosing data for
running regression analysis:
·
How far
backwards?
·
And, at what frequency?
·
Should you perform intra-period
averaging?
·
Drop extreme values?
·
Use changes in MTM or cumulative changes?
·
Or, time of data snapshot?
Other questions involve the
appropriate value of R square and what slope to use in
the regression. And those are not necessarily the only
ones. The bottom line is regression may be elegant, but
it also allows for quite a bit of “massaging” to help
get a better answer. Systems like Cygnifi, in fact, are
working toward optimizing the regression analysis to
minimize EPS volatility. (Cygnifi hopes to have an
analytical tool that would tell hedgers how to adjust
the notional amount of a hedge to achieve a more highly
correlated relationship.)
While useful, such tests may carry
some risk. The more liberal the test, the more likely it
is to “pass” a hedge into special accounting; but does
that mean the hedge is moving further away from
achieving its risk management goal? “That’s a danger,”
Ms. Braga agrees. The point made by Nick Grantley of
Shell is also relevant here: the lesser the hedge’s
borderline, the higher the chances that it will pass the
effectiveness bar. Hence, some of the hedges that have
difficulty qualifying for hedge accounting may also be
less effective in other ways.
