P.
79
2018 Pillar 3 Disclosures
Annex
The mean distribution of the Trading Book VaR by desk
for 2018 and 2017:
2018
2017
Funding Desk and DPV
874
1,024
Forex Desk
550
390
Debt Desk
1,239
780
Equity Desk
257
527
Derivatives Desk
493
260
Credit Desk
165
135
Banknotes
21
27
In addition, an analytical measurement derived from
the VaR, known as the market risk Component VaR
is calculated and reported daily, serving to establish
the contribution to the total risk of each position and
market risk factor (risk concentration), approximating
the sensitivity of the VaR to variations in the portfolio
positions.
The component VaR can be obtained at a greater level
of breakdown and reported by:
•
Product.
•
Risk level.
Parametric VaR
With the aim of increasing the control over the VaR
historical simulation model, the parametric VaR is
calculated and reported daily to provide a point of
comparison for the risk estimate.
This methodology is based on statistical hypotheses of
normality in the distribution of probability of changes
in the risk factors. Using the historical series of
market prices (provided by the Market Data Service),
we calculate (in the market risk measurement tool)
the volatility and correlation between assets, which
together with the hypothesis of the distribution of
probability of changes provide an estimate of the
potential change of a position.
Expected shortfall
Another more advanced method supplementing market
risk measurements is the Expected Shortfall. The aim in
this case is to measure the expected loss in the event
that the VaR levels were to be exceeded. It therefore
quantifies the risk within the loss zone. This is an
asymmetric measurement which, unlike the VaR, not
only takes into consideration the frequency of losses
but also their magnitude in the event that the VaR were
exceeded.
Back testing
Monitoring tests to check the goodness-of-fit of the
market risk model are carried out; for this purpose,
clean and dirty back-testing studies are performed,
which help us demonstrate the suitability of this model
in the daily activity.
Contrast statistics
With the purpose of completing the models in further
detail and more effectively and complementing back-
testing, stricter goodness-of-fit tests are performed to
help identify possible inefficiencies in their calculation.
These tests are an essential tool to manage market risk,
especially when a part of it lies on the use of models
and systems that stem from a series of hypotheses that
require practical confirmation.
The metrics used are carried out on 2 levels:
•
General metrics applicable to all methodologies of
VaR calculation
•
Specific normality metrics applicable to parametric
methodologies
Management results
On the basis of the risk tools, the management results
for the trading books are calculated on a daily basis.
The criterion followed is mark-to-market for positions
with directly observable market prices (funds, bills,
futures, options on organised markets) and mark-to-
model (theoretical valuation) with market inputs for
operations without a quoted price (deposits, OTC
derivatives, etc.).
Sensitivity measurements
Although the limits are structured with regard to the
VaR measurement, which summarises all types of risks
and portfolios in a single indicator, there is a series of
supplementary measurements for the monitoring of
A|A.I