risk management
Meet Rex
What'sNewS Issue 1 January/February 2001
Creating the tools that will transform credit risk theory into management practice
is the responsibility of Adriaan Kukler, manager of the Credit Risk Portfolio
Modeling (CRPM) project. Structured around the implementation and refinement of
the Rabobank Economie Capital Calculation System (REX), CRPM's ultimate aim is,
according to Kukler, 'to let Rl measure the economie performance of credit assets
and actively manage the portfolio.'
Refined approach
REX to the rescue
Work in-progress
Confident outlook
Applications and anxieties
That, however, will come only in the
project's fourth and final phase. In its
initial pilot phase, which began in January
last year, the goals have been more mod
est. 'To begin with we are concerned
mainly with collecting relevant data from
credit applications in order to help us re-
fine the quantification of credit risk and
hence the parameters of the model,' Kuk
ler explains. 'The multi-phase approach
will also allow people's experience to
grow alongside the system.'
Phase two is scheduled for Quarter 4,
2000, subject to approval and coopera-
tion. 'For example, we will be looking at
new sorts of limits based on calculations
of economie capital, new sorts of ap-
provals and new criteria for credits,' says
Kukler. This phase should also see a re
finement of the risk quantification
methodology. Initially the system will only
take into account counterparty risk. Ex-
pected loss and unexpected loss figures
will be calculated from the multiplication
of the counterparty's default probability,
the exposure and the projected recovery.
In phase two counterparty risk analysis
will be supplemented with information
from credit admin systems,
further improving accuracy
of especially dynamic expo-
sures and facilities. After
back-testing, the system can
then be used by product
managers, relationship
managers and account man
agers for performance eval-
uation. The expected loss
figure calculated on a per
transaction basis can also be
used by control RI for its
loss forecasting and reserving
The quantification of portfolio concentra-
tions and active credit portfolio manage
ment are objectives which take the CRPM
project beyond Basle's requirements. 'In
the third phase, one year later, we'11 be
looking to quantify the correlation effect.
Then it will also be possible to calculate
the marginal risk contributions of a single
asset or sub-portfolio of assets in compari-
son with the existing total portfolio of the
bank.' In the final phase, it's hoped that
the mark to market of loans can be calcu
lated, enabling an accurate comparison of
the internal and external price of RI's
credit assets to be made. Effectively, REX
will teil RI whether to keep a loan or sell
it. Rolled out globally, REX will provide
an overview of all the credits in the sys
tem, opening the door to fully centralized
portfolio management. It will also estab-
lish a direct line between the origination
and distribution sides of business.
With the Basle committee still deliberating
on the exact criteria that banks should use
in making internal assessments of credit
risk, is there a danger that Rl might have
jumped the gun? 'Clearly, until the Basle
committee finalizes its BIS II recommenda-
tions, we can't be 100% certain that the
CRPM system will be compliant, but let's
just say our confidence level is around
99.99%,' says Kukler. 'Our target is more
than what BIS II requires, it's about imple-
menting best practice in the industry.'
continued from page 3
portfolio correlation and reduce capital
requirements. Effectively, introduction of
these measures means you can improve
the returns on risk at both ends of the
process: on the origination side through
the use of EVA and on the distribution
side by reducing portfolio correlations
through diversification.'
The applications of this doublé sided ap
proach to risk management are seemingly
endless - improved risk sensitivity analysis
enables sharper product pricing, a new cri
teria for limit settings and exposure assess-
ment and an EVA-based approach to rela
tionship profitability and management. It
will also provide a more accurate measure
of performance of all entities within the
bank. Understandably, however, the
prospect of EVA-based evaluation might
give rise to anxiety in some quarters - after
all this new emphasis on quality rather
than quantity, wouldn'r this suggest a
shrinkage in bonuses? Singh believes such
anxieties are groundless. 'It's a fact of hu-
man nature that once you mention per
formance, people start getting nervous,
but the current system of only maximizing
return on solvency (ROS) isn't good for
anyone in the long run. We're concerned
with developing the tools that will help
improve individual and organizational
performances by helping people make ra
tiona! decisions. Once people becorne fa-
miliar with them, approaching transac-
tions with the true risk return equation in
mind will become second nature.'
ht the meantime, however,
CRPM's high IT component
means there's still a lot of
t| work to be done on data gath-
ering ar>d integration. 'We're
ciirrently vvorking on phase
two,' says Kukler, 'involving
y the collection of outstandings
and the development of inter-
faces to all the production sys-
tems, next to the inclusion of
bank counterparties in REX.
Adriaan Kukler, ensuring Initially we'11 be targeting the
success of CRPM SKY database and the CRAM
database for the Sundial secu-
policies. ritization program. After that we'11 be go-
ing from office to office to see what more
needs to be done.'
