FRANKFURT (MNI) – The following is the first part of a verbatim
text of a technical note from the European Central Bank (ECB) on the
macroeconomic scenarios and reference risk parameters following the
publication of the EU-wide stress tests’ results:
INTRODUCTION
This note presents the main technical features of the stress test
exercise that has been conducted by the CEBS and national supervisory
authorities, in cooperation with the ECB. ECB staff provided the
macroeconomic scenarios (benchmark and adverse) and the corresponding
key micro parameters (probabilities of default (PDs), loss given default
(LGDs), and haircuts for holdings of government bonds in the trading
book). The changes in these parameters under the adverse scenario
represent a substantial stress for the European banks.
The “benchmark” scenario is on average not very far from currently
available forecasts, while the adverse one, taking stock of prevailing
tail risks especially related to the sovereign debt situation is in
turn substantially below these forecasts.
In addition, the severity of the stress arises from the combination
of the increase in the haircuts and especially from the increase in the
PDs and LGDs under the adverse scenario. The reference haircuts were
computed from changes in the prices of 5-year sovereign bonds. The
impact of the increase in the haircuts on government debt in the trading
book is mitigated by the fact that banks holdings of government
securities are primarily in the banking book, and the average maturity
of these securities is only around 5 years. On the other hand, the
increase in PDs and LGDs affects all portfolios in the banking book and
is substantial. For instance, comparing the end-2009 values with those
under the adverse scenario in 2011, PDs of corporate assets double or
triple in some countries, while for the euro area they increase by over
61%, on average.
1. The Macroeconomic Scenarios
For the purpose of conducting the stress-test exercise, two
macroeconomic scenarios, covering the period 2010-11, were developed: a
benchmark scenario (see Table 1), and an adverse scenario (see Table
2), taking stock of prevailing tail risks, especially related to the
sovereign debt situation. The adverse scenario GDP, cumulated over
2010-11, is close to three percentage points lower than the benchmark
one for the European Union (EU) and for the euro area as a whole.
The benchmark scenario is mainly based on European Commission (EC)
forecast numbers that were available when work on the CEBS exercise
began in March 2010, i.e., the Autumn 2009 European Economic Forecast
(November 2009) and the EC Interim Forecast (February 2010). This was
complemented with more up-to-date information on country forecasts in
cases of significant changes. Assumptions for market interest rates as
well as for exchange rates were set in line with the methods employed by
the EC to construct their forecast.
In this scenario, the slow recovery initiated in 2010 is expected
to gain further momentum, with e.g. GDP growth for the euro area
reaching 1.5% in 2011 after 0.7% in 2010 largely in response to the
ongoing world trade pick-up. At the same time, unemployment remains high
even increasing in a number of countries, owing to the lagged effects
of the past activity slowdown. Consumer price inflation is assumed to be
contained and stable overall, as the upswing occurs in economies where
the degree of slack is substantial. There are however a number of
countries where inflation declines or increases significantly
reflecting their cyclical positions or fiscal policy measures.
The scenario involves somewhat more contained dynamics in 2010,
while by contrast it appears generally on the upside for 2011. On
balance over the two years, differences with currently available
forecasts are limited.
The adverse scenario12 has two main features, a global confidence
shock, that affects demand worldwide, and an EU-specific shock to the
yield-curve, originating from a postulated aggravation of the sovereign
debt crisis. The latter impact is differentiated across countries,
taking into account their respective situation.
The global confidence shock occurs in a context of downgraded
employment and profit expectations worldwide. It affects both private
investment and consumption, through a lasting downward shock to these
variables, cumulating overall to some 2 percentage points of GDP points
over the horizon, concentrated over the second half of 2010 and the
first quarter of 2011. The EU is directly affected by this confidence
shock and by the effect on exports of the implied lower world demand.
In addition, related to prevailing sovereign debt risks, a common
upward shift in the yield curve was applied for each country in the EU
(reaching 125 basis points for the three-month rates and 75 basis points
for the 10-year rates at end-2011), supplemented with country-specific
upward shocks to longterm government bond yields (overall amounting to
70 basis points at end-2011 for the euro area). The rise in short-term
rates reflects an assumption of tensions in the interbank market as
was seen during earlier financial turmoil episodes. The country-specific
bond yield shock in turn accounts for differentiated fiscal situations
and related market perceptions.
Accordingly, the distribution of the country-specific upward shock
to long-term interest rates across countries reflects two elements.
First, a widening of spreads in line with market developments since the
beginning of May 2010. Second, an additional widening of spreads
reflecting an average additional increase of 30 basis points. Its impact
on each countrys long-term bond yields was determined in proportion to
the volatility of 10-year sovereign bond spreads that was observed
between December 2009 and June 2010. Taken together, the
country-specific shock implies an additional average increase of 70
basis points (see Table 3). To underline the importance of the combined
shocks affecting interest rates, it is worthwhile to mention that, for
example, they result in 2011 in 10-year government bond yields of 4.7%
for Germany and 14.7% for Greece (see Table 7).
The macroeconomic effects of these assumptions were calibrated
using econometric models, also taking into account trade spillovers
across EU countries. GDP growth is particularly affected in the adverse
scenario, and is lower than in the benchmark scenario for all countries,
on average by about one percentage point in 2010 and by close to two
percentage points in 2011. The unemployment rate is higher, especially
in 2011 (e.g. by 0.6 percentage point in the euro area), while inflation
is significantly lower in 2011 (e.g. by 0.4 percentage point for the
euro area). The adverse scenario generally appears to be substantially
below available forecasts and projections, thereby corresponding to the
materialisation of downside risks to economic growth prospects.
2. Probabilities of Default and Loss given Default
Estimates of probabilities of default (PD)13 and loss given default
(LGD)14 parameters were computed at the country level for five main
portfolios (financial institutions, sovereign, corporate, consumer
credit and retail real estate). For all countries in the exercise, these
parameters were computed for both the benchmark and adverse scenarios
for 2010 to 2011.
To calculate the PDs and LGDs conditional on the different
scenarios, sector-specific regression models16 were used to link PDs and
LGDs with macroeconomic variables.
These models provide estimates of sector-specific elasticities of
PDs and LGDs with respect to changes in macroeconomic variables
conditional on shocks to the system. In the models, three propagation
channels for the shocks were identified: the demand channel; the supply
channel and the long-term borrowing costs channel. To obtain
country-specific PD and LGD parameters for 2010 and 2011 under the
benchmark scenario, these elasticities were multiplied by the projected
changes in macroeconomic variables for each country using the PD and LGD
levels that were observed at end-2009 as a starting point. Similarly, to
obtain PDs and LGDs under the adverse scenario in 2010 and 2011, the
differences between the macroeconomic variables in the benchmark and
adverse scenarios for each year were multiplied with the elasticities
implied by the sector-specific regression models. For the purposes of
using these parameters for stress-testing the balance sheets of
individual financial institutions, national supervisory authorities were
encouraged to use as a starting point their own PD and LGD levels for
2009 and to apply the changes of these parameters in 2010 and 2011 with
respect to their values in the benchmark scenario in the respective year
according to the outcomes of the ECB models. For some of the largest
banks for which a full bottom-up exercise was conducted, together with
supervisory authorities, supervisors could decide to allow these banks
to feed the common macroeconomic scenarios into the banks own internal
models for the computation of PDs and LGDs.
[TOPICS: M$$EC$,M$X$$$,M$$CR$,MT$$$$]