|
Economic
capital is considered a key component in the management
of risk, profitability, and, therefore, value. Use of
economic capital in value management allows banks to
take risk into account in a much more tailored and
individual way, helping them (in theory at least) to
eliminate the “measurement arbitrage” inherent in the
old regulatory capital approach. But despite its obvious
advantages, many are vocally critical. Economic capital,
they say, places too much emphasis on rare events,
focuses entirely on the debt holder’s perspectives,
doesn’t always give the right incentives for management
and does not, in itself, provide a sufficient
early-warning system of critical losses. Ironically, it
is the biggest players in the banking world—who have
been perceived as front-runners in applying economic
capital management approaches—that have tended to suffer
most in recent crises. What is behind this apparent
paradox? Has economic capital proved a paper tiger, too
weak to deliver on its promises—to help banking
institutions deliver value for shareholders? And what
lessons can we learn from recent events?
Where
does economic capital fall down?
It is
clear that economic-capital concepts have some inherent
weaknesses, which make them vulnerable to (conscious or
unconscious) exploitation or misuse. These include:
1. Event
risk—very rare events escape the net
Most
economic capital models are value-at-risk (VAR)-based.
While they typically operate at high confidence levels
of 99.93 percent to 99.97 percent, they do not capture
losses from very rare events, independent from the
amount of associated losses. The losses we are currently
seeing in the subprime collateralized debt obligation (CDO)
markets (which, on paper, were originally rated AAA) are
one example of such a rare event, and so was the drawing
of liquidity lines provided to structured investment
vehicles (SIVs) holding this paper. These events fell
outside of the range of probable losses which economic
capital provides for.
2. New
risks—not enough historical data
Risk
models typically rely on historical data to measure
risk. But in the case of new risks, there is no
appropriate historical data to feed the models, and
standard bank-wide models may be too coarse to capture
the true risk of a new business. Recent cases in point
include:
§
Loss
data for subprime loans mostly distributed to investors
via CDO structures—and the resulting bad quality of the
loan process.
§
Correlation data on defaults by different obligors, and
how this changes in different market environments (e.g.
in the case of an interest-rate rise). Correlations are
a key driver of probability of default (PD) and loss
given default (LGD) in loan-portfolio tranches. But
correlation is notoriously difficult to measure based on
historical time series, and investors into CDOs do not
generally have the means, material or methodology to
measure the amount of dependence risk to which they are
exposing themselves. Bank-wide measurement systems are
not sufficiently sensitive to provide a proper measure.
3. Risk
category—interdependencies may not be captured
No model
covers all risk categories to the same extent, and some
are typically not covered at all. Very often,
interdependencies between categories are not fully
captured quantitatively, mostly due to lack of
historical data. So, in many cases, risks may go
undiscovered for a long time until they suddenly strike.
This is certainly the case for SIVs and conduits, which
were treated as a much lower risk than on-balance-sheet
exposures until suddenly—through the drawing of
liquidity lines or consolidation—they turned into
on-balance-sheet exposure. At one time, it was hardly
conceivable that those liquidity lines would ever be
drawn. In our current market situation, there is hardly
any liquidity line that has not been drawn.
What
lessons can be learned?
All of
this raises big questions for managers. Economic capital
does have some serious failings. But with no better
alternative available at present, and the Pillar 2
regulators due to visit banks in the second half of 2008
to review their latest models, how can we make the best
of it? Below are some pragmatic ways to breathe life
into your economic capital approach.
§
Place
much more emphasis on scenario analysis and stress
testing. Permanently challenge—using qualitative
analysis and common sense—the output from your
statistical models, based on historical data. Adapt
models, or restrict their use in the event of
significant differences in the outcome of the analysis.
§
Holistically analyze the various business models your
bank operates:
§
What are
the main revenue drivers?
§
What are
the potential cost drivers?
§
Under
what circumstances would the business model break down
or collapse?
§
What
economic capital would the market require the bank to
hold in order to operate the business model, given full
transparency?
§
No
restriction of significant risk categories, parts of
business models, or other silos are allowed. This could
help capture more risks—in particular interdependencies
between risk drivers.
§
Integrate the perspective of other stakeholders into the
economic-capital concept:
§
The
current approach focuses on the debt-holder perspective
and the bank’s default probability.
§
A more
balanced view would include a going concern perspective
(including the likelihood of wipe-out of profits, and
loss of equity triggering a recapitalization).
§
This
approach would also capture the breakdown of single
business models—and the events we currently see in the
market.
§
Risk
measurement is just one side of the story: it has to be
related to management action. This may result in a
decision to walk away from unacceptably risky business;
or you may decide to change or add to your business
model to avoid relying on a certain source of revenue
with unacceptable risks.
§
As well
as measuring capital, you also need to manage capital
actively. You need the right instruments in place to
manage your business relative to your capital base. Your
scenario analyses must relate to, and reflect, the
complete capital picture, including not just economic
capital but regulatory, rating-agency and book capital.
This assessment will, of course, be different for every
bank, and it may be tough to apply a number to it: but
that is exactly what Pillar 2 of Basel II is calling
for, with bank-wide stress testing.
Economic
capital is still alive—just keep challenging
Despite
the fact that economic capital may not be a panacea in
bank management, KPMG member-firms still believe in its
usefulness. However, banks need to make sure they are
really exercising the freedom which the concept allows
them. Evaluating risk is not a mechanical exercise; it
is not a matter of crunching numbers through a machine
and blindly believing the results. Evaluating risk
should be an ongoing and challenging qualitative
exercise, geared to improving and calibrating your
statistical models, so that they can be truly effective
when they are used to support business decision-making.
Our
recommendation is simple: Keep challenging yourself.
Poke holes in your model. Ask yourself: What assumptions
have I made? If they’re wrong, what would happen to my
business? Focus on understanding your business model:
What makes it work? What drives profits and losses?
Devote proportionally more time to developing new
scenarios. Make sure you have a process to attach
numbers to those scenarios, relate the numbers back to
your risk-bearing capacity—and act on what you find. As
long as practitioners are exercising true economic
thinking in this way, economic capital still has a
chance to be meaningful and helpful.
This
article is an excerpt from a thought leadership document
entitled “Frontiers in Finance” (March 2008), Daniel
Sommer and Holger Spielberg, partners at KPMG Germany. |
