263SYSTEMIC RISK MANAGEMENT IN THE BANK ACTIVITY: NEW [601853]

263SYSTEMIC RISK MANAGEMENT IN THE BANK ACTIVITY: NEW
APPROCHES

Professor PhD, Marin Opri țescu, University of Craiova , [anonimizat]
Teaching Assistant PhD Student: [anonimizat], University of Craiova,
[anonimizat]

ABSTRACT: The role of the bank system as essential link the saving-investment process
makes of its stability a priority on the agenda of the public authorities. One of the major objectives
of a central bank is to prevent the systemic risk by promoting an efficient bank monitoring, which
should contribute to the achievement of the stab ility and viability of the entire financial system.
Thus, the central banks developed methods and pr ocesses for the continual supervising and
evaluation of the banks – premises of the preven tion of the apparition of a great variety of bank
crisis or other unpleasant surprises regarding the entities of the bank system.

Key words: model, systemic risk, warning in time system.

JEL codes: G21, G24, G32

General considerations

The bank activity involves risks which are manifested at the level of each bank entity, but
which can be transmitted in the entire bank system or, in the case of the international/transnational
banks, in more bank systems. The banks pursue the reaching of some objectives that are many times
divergent, in that they use specific instruments both in order to raise its quota on the market, and also to attract available capitals necessary to the performance of some speculative transactions.
On the other hand, the changing environment th e banks operate in, marked by the increase
of the volatility, the internationalisation and the liber alization of the financial markets increased the
effect of contagion, as it was proven by the propagati on of the effects of the financial crisis at the
end of the 90s in Thailand to the rest of Asia , eastern Europe and South America, on the entire
world bank system. These events determined the supervising authorities to pay an increased
attention to the financial risks and, implicitly, to the administration of the systemic risk.

The models used in the global appreciation of the bank risks

In order to prevent the systemic risk, in order to assure the stability and the viability of the entire
bank system, the monetary authorities developed systems of monitoring the activity and the results of the banks. All the bank systems have at least an au thority of regulation and supervising, which have
responsibilities, powers of regulation and implementation of the different assumed decisions. In most of
the bank systems, the regulation and supervising authority goes to the central bank. In order to be efficient, the supervising authorities must rejoice an appropriate implementation power and an adequate
degree of autonomy, most of the times they having to resist the pressures exercised by the government,
banks, stock holders, deponents or creditors.
The supervising authority normally uses a top down approach that is focused upon the
evaluation of the manner the banks id entify, quantify, administrate and control the risks, and as it is the

264case, establish a diagnostic for the observed problems. In the practice of the banking supervising,
directed towards the evaluation of th e risk and the identification of th e potential problems that can affect
the bank system, the authorities use instruments and pr ocedures of global appreciation of the bank risks,
known as warning in time systems . The most known supervising syst em of the bank rating is the
CAMEL model, used in USA. Thus, Federal Reserve Bank evaluates the banks thanks to a category of
performances included in the CAMEL system ( acronym formed of the names of the composing
elements of the process of examination of the safe ty and solidity of the banks – solvability, the quality
of the assets, the management, the incomes and the liqui dity). Each element is attributed grades from 1
to 5. The banks that have received the grades 4 or 5 are considered banks in difficulty, their activity
being considered risky and they are exposed to bankr uptcy. These are strictly supervised and forced by
the bank authorities to optimise their financial behaviour.
In the American approach of the risk positi on of a bank, the bank performance is defined by
solidity, given by an appropriate adequacy of the capital. The management of the bank and the profitableness (which is the purpose of each bank) are not objectives of the Central Bank, being
considered the most efficient means of minimizati on of the risks and covering of the exposure to the
risk.
The analysis systems of the financial indicators based upon the prudential norms suppose the
monitoring and the analysis of the main economical -financial indicators of a bank institution: the
adequacy of the capital, the liquid ity, the exposure to a single debtor etc. and in case they exceed the
established limits, the supervising authority interferes aggressively.
The comprehensive methods of evaluation of the bank risks suppose an evaluation of the risk
profile of the bank through the quan tification of all the risks that correspond to every activity treated
distinctively and the attribution of certain scor es for every activity. The cores are afterwards
aggregated so that the final score of the bank as a whole is obtained.
The statistic models have the advantage that they identify those risks that have the greatest
possibility to generate situations adverse to the bank, based on the prevision of the probability of the
future evolutions. These models remove all the di sadvantages of using the static models in an
economy characterized by dynamism. The most used models are the ones which estimate a probability of a decrease of the grade held by the bank, predict a supervising incapacity or
estimates potential losses which can induce, under extreme conditions, banking bankruptcies. The
misfit of these models stands in the fact that they suppose the existence of some considerable
historic registers, so that the solutions prove to be veridical. At the same time, the solutions of the
models depend on the right choice of the variables upon which the provision is made.
The bank rating and early warning system used by the National Bank of Romania is
CAAMPL. This was implemented in 1999 in order to promote an efficient supervising, aligned to
the international practices and standards. The fundamental objective of the supervising system is the identification initially of those banks that are c onsidered inefficient, according to the evaluation
criteria of the financial and operational aspects established by the monetary authority or manifest adverse trends, requiring a special supervising.
The components of analysis of the CAAMPL syst em are: the adequacy of the capital (C), the
quality of the assets (A), the quality of the stock holding (A), the management (M), the profitableness
(P) and the liquidity (L). In determining the composed degree of classification , from the five elements,
four are analysed in accordance to a series of indi cators, each bank being attributed a composed rating
and a final score, which reflects th e total of points given to the indicators which define the CAAMPL
components (the banks are classified into five cate gories, the superior level being represented by the
rating 1). The final grade is obtained by summing the va lue of the rating for each indicator, to which is
added the value of the rating for the components “the quality of the stock holding” and “the quality of
the management”.
From 2004 BNR has modified the modality of ev aluation of the management of the banks
by using some models of evaluation of the risks of the scorecard type which allow the calculation

265and the attribution of a grade for each component of the management system: the planning, the
monitoring, the control, the evaluation. These data-p rocessing applications have been possible as a
consequence of accessing the PHARE financing funds and the consultancy offered within these projects. From 2005 BNR has modified the approach in the evaluation of the L component, so that
it would quantify the capacity to plan, monitor and c ontrol the bank risks, to evaluate the adequacy
of the internal audit systems.
The rating system used by BNR is submitted to a continuum process of perfecting, imposed
by a multitude of factors, of which we mention: the evolution of the national economy, the evolution and the vulnerability of the bank system, the necessity of the harmonisation of the bank legislation to the international standards. For instance, upon the liquidity problems the Commercial
Bank Unirea, Eurom Bank (ex Dacia Felix) or Ba ncorex dealt with during 1998-2000, the National
Bank of Romania included two essential compone nts (the quality of the stock holding and the
management) and improved the liquidity component (2001) by determining the liquidity index as
report between the effective liquidity and the necessary one.
In order to consolidate the supervising activity which allows BNR to plan, monitor, control
the bank risks and also to evaluate the quality of the informational systems at the level of each bank,
the banks rating system CAAM PL includes a new component the market risk sensitivity (S) . This
one evaluates the market risk and the first simu lation of this component was achieved on June 30
th,
2004. According to the theoretic model, the banks were submitted to some interest rate shocks
consisting in the modification of the interest rate by four percents for the national currency and tow percents for the unique European currency and the Amer ican dollar. The scenario was chosen so that it
would generate a loss at the level of each bank. The conclusion of the simulation was that, despite the
insufficient development of the derived financial inst ruments meant to reduce the market risk, most of
the Romanian banks registered low exposures to the interest rate risk. The weakness of this simulation
was the unavailability of some data necessary to the evaluation of the sensitivity for some banks.
According to the model, the weight of the quantitative analysis was of 40%, the difference being represented by an analysis of qualitative type.
According to CAAMPL, the totally active weights of the banks classified in accordance to the
five types of rating, during 1999/2006, is presented in figure no. 1.
Fig. no.1. – The weight of banks classified in the five composite ratings in the banking
asset volume during 1999-2006
Source: National Bank of Romania, Romanian Banking System, 2006, p.88
From the presented data we the following can be observed:
• the increase of the ratings granted to the banks starting with 1999, year that marls the The weight of banks classified in the five composite ratings in the ba
asset volume
December 31st 1999 – December 31st 20060.64
78.74
7.16
6.736.720.00
70.78
19.41
7.99
1.833.38
76.39
16.01
3.79
0.439.85
66.78
19.33
3.89
0.150.00
79.39
20.05
0.46
0.100.00
86.86
12.55
0.54
0.050.00
78.21
21.38
0.39
0.022.11
73.81
23.69
0.37
00 20.00
77.37
22.27
0.340.00
74.71
24.76
0.530.00
74.12
25.34
0.540.0010.0020.0030.0040.0050.0060.0070.0080.0090.00100.00
RATING 1 RATING 2 RATING 3 RATING 4 RATINGDecember 1999
December 2000
December 2001
December 2002
December 2003
December 2004
December 2005
March 2006
June 2006
September 2006
December 2006

266beginning of the process of restructuration of the bank system;
• the polarization of the bank system on the superior floor of the bank rating system,
containing the banks with 2 and 3 rating, which denotes the consolidation of the bank system.
However, in 2007, according to a study elabor ated by BNR [Georgescu Florin, 2007] we
observe that more banks were retrograded by BN R from rating group 2 to group 3 at the end of
September 2005, mainly because of the deterioration of the profitability indicators.
According to the estimations presented by the prime-vice governor of BNR, Florin Georgescu,
six banks received weaker evaluation grades as compared to December 2006, exiting rating group 2.
Thus, the number of the banks included in this gr oup decreased from 17 to 11. In exchange the number
of the banks of the rating group 3 increased from 19 to 26, and the one of the banks from the rating
group 4, increased from 1 to 3, given the fact that the total number of the credit institutions on the market increased from 37 to 40. Presently, the rating group 2 represents the best efficiency points the
banks in the system can obtain. From the end of 2003, no bank had been included in the rating group 1.
It is also shown in the study that there is a s light degradation of the profitability rate of the
banks, against the expenses for the expansion of the networks of units and the increase of the
number of wage-earners, and also the reduction of the interest margins. Thus, BNR pursues the
financial solidity of the banks on the market and th e degree of potential contagion after a six component
rating system – CAAMPL. BNR does not make pub lic the ratings given to every bank, as the
independent financial evaluation agencies do, however it presents from time to time the number of the
banks which are included in every rating step. Acco rding to the data by the end of September 2007, no
bank on the market had received the alarm points, which correspond to rating group 5.

Fig. no. 2. – The weight of banks classified in the five composite ratings in the banking
asset volume
Source: Georgescu Florin (2007) – Bilanțul evoluției sistemului bancar la un an de la
aderarea la Uniunea European ă, Bucharest, 2007
At the same time, beside the rating syst em CAAMPL, according to a study made by BNR
[National Bank of Romania, 2007] it is described the mechanism of a early identification of the
deterioration events of the CAAMPL rating system with one year prediction horizon, and also the
results of its estimations for December 31
st, 2007.
Its main component is represented by a statistic quantification of the probability of deterioration
of the CAAMPL rating, estimated and tested using exclusively micro prudential data that cover the
December 1999 – December 2006 period. The specific methodology of estimation, testing and implementation combines elements encountered in the specialty practice to the exigencies stipulated by

267the Basel II agreement regarding the problematic of th e internal models for the credit risk, in order to
obtain some high and consistent performance in time of the process of identification of the events of
deterioration of the credit institutions in the Romanian banking sector.
The high accuracy of the statistic model is assu red by the performance of the noting function, in
the terms of the discrimination power, of the stability and of the adequate calibration of its estimations.
The value of the surface indicator of the ROC curb is high and robust, the stability tests indicating
reduced fluctuations around the level of 85%, which re presents, moreover, the appropriate value for the
entire observations sample. Plus, the verification of th e ex-ante delimitation capacity of the deterioration
events of those of maintenance and improvement of the CAAMPL rating shows a higher performance at
the level of the testing sample (January 2003 – Decem ber 2005) than that at the level of the estimation
sample (December 1999 – December 2002).
The model is very well applied especially in the case of the credit institutions with a high weight
in the Romanian bank system, the weighted average of the general accuracy rate being over 90 percent
as opposed to 80 percent in the case of the arithmetic mean in the conditions of the calibration of the
alarm limit so that the percent of the false alarms be equal to the one of the unidentified deteriorations.
The performance percents are notable in the cas e of BNR, BRD and Raiffeisen Bank, namely 100
percent for the first and the third bank and 95,89 per cent for the second one. On the other hand, in the
case of three of the banks with foreign capital and of a bank with autochthonous private capital the
scoring function has an unsatisfying performance. This result is determined by the great number of
consigned false alarms, considering the fact that the events of deterioration of the rating are integrally
signalled by the model. For these reasons, the ex-ante id entification of the events of deterioration of the
rating is made exclusively upon an alarm limit, th e success probability of the signal furnished by the
model being evaluated inclusively upon a noting scale but also through qualitative analysis.
The noting scale includes six risk classes, whic h assure a good segmenting of the probability of
deterioration in terms of the value of the obtained sc ore. The criteria used for its construction were the
homogeneity of the events from the same class and the significant delimitation of the empirical
probability of deterioration of the CAAMPL rating be tween the different risk categories. The empirical
resulted probabilities are robust estimations in the term s of the risk of deteri oration of the CAAMPL
rating and allow the gradation of the alarms gene rated through the application of the alarm limit,
especially for the values of the theoretical probability of deterioration adjacent to the level of the alarm
limit. Three categories of these have low deteriorati on risk, one has average deterioration risk and two
have high deterioration risk.
Taking into consideration the previously exposed aspects, the author of the study
[Bogdan Moinescu, 2007] considers that beside th e CAAMPL rating system, the use of the model
of prevision of the deterioration of the rating pres ented in this study can contribute to the increase of
the accuracy of the process of id entification of the banks with financial difficulties. Thus, the
central bank could administrate more efficiently th e limited resources it has in order to avoid some
eventual individual disequilibrium at the system level.
The results of the application of the system of provision of the events of deterioration of the
CAAMPL rating for December 31st 2007 indicates 18 credit institutions with low deterioration risk,
5 with average deterioration risk and 8 with high de terioration risk. The interpretation of the signals
supplied by the model based upon some qualitative information and the historic performance
registered on each bank by itself leads to the following conclusions:
− the credit institutions with important weight in the Romanian banking sector will register a
performance at least as good as in 2007 comparatively to 2006;
− there are no banks with composed 3 rating at the end of 2006 for which the model will
signal a deterioration of the performance at the end of 2007;
− the deterioration signal of the rating in the cas e of five banks represents more likely a false
alarm, even though the empirical downgrade probability afferent to these entities is above average;
− it is very likely that for three credit institu tions summing 2 percent of the aggregated asset

268of the Romanian banking system the CAAMPL system deteriorate from 2 in December 2nd 2006 to
3 in December 2007.
At the same time we must mention that the fu ture inclusion in the quantification model of
the probability of deterioration of the CAAMPL ra ting of some macroeconomic variables, as the
exchange rate and the interest rate or the in tegration of the collected information through the
activity of bank inspection in the process of interpretation of the signals supplied by the statistic model can lead to the refinement of the system of prevision of the events of deterioration of the
rating presented in this study. Also, as the great rating agencies will not e more of the credit
institutions from the Romanian bank system in report to the nine ones from the present, the
utilisation of the issued ratings, but mostly of their component regarding the appreciation of the
perspective will enlarge the sphere of informa tion available for the completion of the signals
supplied by the system of prevision of the events of deterioration of the CAAMPL rating. Plus, the
annual reanalysis of the performance of the sta tistic model – expression of the application of
requirement stipulated by paragraph 443 of the Basel II Agreement represents another component which will contribute to the maintenance of its accuracy at a high level in the future.

Modern approaches of the global risk
The supervising authorities, and also the th eoreticians pay nowadays a special attention to the
macroprudential analysis in order to evaluate the vulnerability of the bank systems to shocks. The
novelty of this recent approach, consecrated at the end of the 90’s, of the last century consists in the fact
that the systemic risk is analysed from the perspec tive of its interaction to the real economy, the focus of
the supervising activity being on the contamination risk and the mutual exposure of the banks to
macroeconomic shocks. We assist thus, to the minimi sation of the factors specific to each bank that can
have an adverse evolution and can amplify the exposure to risk.
Practically, this approach uses aggregated macro prudential quantitative indicators at the level of
the bank sector (liquidity, adequate capital, the quality of the assets) and macroeconomic indicators (the GIP level and dynamics, the evolution of the inflationist process, the policy of the incomes, etc.) which
concur to the establishment of the interaction between the real sector of the economy and the health of
the bank system. The macro prudential analysis frame is complete when in the model are used data regarding the entire financial system and there are used techniques of the stress tests type.
We consider that the efforts of applying are considerable and the success of this type of
analysis depends on the degree of integration of the financial system in every country and on the
creation of some international standards so that this demarche is unitarily implemented .
The decision of the Administration Council of BNR from October 2004 through which the
Direction of Financial Stability is created, having a role in the elaboration of some representative
financial stability indicators for the supervision of the national financial system and to assure their
international comparability, denotes the fact that the macro prudential analysis is agreeable by the
monetary authority from our count ry too. The problem will be difficult, considering the structure of
the Romanian financial system, and also the to tal liberalisation of the capital account, which will
impose the Central Bank the enforcement of the supervision process for the assurance of the
stability of the bank system.
Usually, the authorities use more warning in time systems, precisely to assure a high efficiency
of the supervision. The Committee from Basel through the New Basel II Agreement set the basis of the
consolidated supervision , considering the transnational character of the banks. Thus, the authority from
the origin country must supervise, on a consolidated basis, the banks form the host countries, which does not exclude the compulsoriness of the banks fr om the host country in respecting the prudential
norms specific to the banking market where they ope rate. The banks reciprocally supply themselves
information regarding the management and the stock holding of these credit institutions, especially as far as the liquidity, the solvability the scheme of guarantying the deposits, the limitation of the great

269exposures, the accounting procedures and the internal control mechanisms are concerned.
Therefore, at the same time with the exposur e of our country to the European Union, the
National Bank of Romania became a member of the Eu ropean System of the Central Banks, quality in
which it is represented in all the its work structures . This representation supposes the participation to the
regulation process at the level of the European Union which is developed on four work levels.
Among the main benefits of this process we can find the increase of the speed of adopting
the decisions by delegating the components of tec hnical regulation to the Specialty committees and
the possibility of reaching in time the convergence in the plan of the supervision practices at the
competent authorities from the E.U. Among the actions taken by BNR in order to reach the convergence in the plan of the practices of supervision of the banking activities we can find:
– the adapting of the reporting system of the credit institutions at the COREP requirements –
Common Reporting – standardised frame of prudential reporting in the EU – and FINREP –
Financial Reporting – standardised frame of financial reporting used by the prudential supervision
authorities from the EU – through the configuration of their reporting forms and their integration in
the reporting electronic system of BNR;
– the use of the recommendations elaborat ed by the Committee of the European Bank
Supervisors (CEBS);
– the signing of ten bilateral memorandums with supervision institutions from the original
country of the financial groups present on the Roma nian market for the flex ibility of the exchange
of information necessary in the achievement of an efficient supervision;
– the participation to the information exchange with the supervisors from South-Eastern
Europe by constructing a regional platform, as a c onsequence of the dominant role the subsidiaries
of some foreign banks play in the financial intermediation of this region;
– the promotion of a mutual supervision through the participation to twinning programs,
professional training seminaries, bilateral meeting between BNR as supervision authority from the
host country and those from the origin country, such as Banca d’Italia, Austrian Financial Market
Authority and Austrian National Bank, Hungarian Financial Supervisory Authority and Bank of
Greece.

Conclusions
The recent developments focus upon sophisticated systems that use econometric techniques for
the estimation of the bankruptcy probability or the deterioration of the rating. Based upon the
information offered by these instruments, set off, wh en it is needed, inspection actions which aim for the
specific identified aspects or are established the prio rities in the case of the general examinations which
are performed on a regulated basis. Thus, the early warning systems of the deterioration of the bank
performances allow the improvement of the efficacy of the bank inspection activity and a better administration of the limited resources the prudential control authorities dispose of.
Hence, we consider that there are some challenges in the implementation of a supervision
process based upon the mechanisms of evaluation of the risk management:
– The thoroughness of the qualitative aspects of the supervision process, through:
ƒ A greater involvement of the management of th e credit institutions in the process of risks
administration;
ƒ The evaluation of the credit institutions is focused upon the establishment of the risk
profile of the institution;;
ƒ The supervision consolidated upon a ti ght cooperation, both between the credit
institutions members of the groups, and also th e corresponding supervision authorities from other
countries;
– The implementation of the primary and sec ondary legislation approved by the end of 2006
in order to apply the stipulations of the Basel II Agreement starting, the latest, by 2008;

270– The main changes in the new approach regarding the prudential supervision are:
ƒ The diminishing of the conformity and the ach ievement of a prudential supervision based
upon the evaluation of the risks of the credit institutions;
ƒ The focusing more on the assurance of respecting the principles of corporatist governance
ƒ The evaluation of the stability, the accuracy and the efficiency of the process of
management of the risks of the credit institutions, especially regarding:
• The quality of the strategy of modernisation of the risks management;
• The efficient existence and the of the specialised and managerial
support committees;
• The efficient functioning of an adequate internal control system;
• A transparent and efficient system of managerial reporting.
As a conclusion, the Basel II Agreement is the mo st important referential frame in the micro
prudential nowadays. In the frame of the new appr oach, the supervision activity is oriented more
and more towards the analysis of the risk pr ofile of the credit institution, of the means and
instruments existing at the hand of its leaders for the efficient administration of the specific risks.
The rating and early warning systems, the stress tests and the ones of interbank contamination
represent sophisticated techniques that allow the successful achievement of the previously
mentioned objectives. These instruments use relevant information on the characteristics of a credit
institution and its counter parties, based upon whic h a synthetic measure of their performances
and/or vulnerability is supplied.

References:
1. Georgescu Florin (2007) – Bilanțul evoluției sistemului bancar la un an de la aderarea la
Uniunea European ă, Bucharest;
2. Georgescu Florin (2007) – Provocări pentru supravegherea sectorului financiar-bancar ,
Bucharest;
3. Moinescu Bogdan (2007) – S istem de previziune a evenimentelor de deteriorare a
ratingului CAAMPL , Studies Notebooks no. 23, National Bank of Romania;
4. Oprițescu Marin – coordinator (2006) – Managementul riscurilor și performan țelor
bancare , Universitaria Publishing House, Craiova;
5. xxx – (2004) ”International Convergence of Capital Measurement and Capital
Standards” , Basel Committee on Banking Supervision .
6. xxx – (2007) ”Raport asupra stabilit ății financiare” , National Bank of Romania ;
7. xxx – (2008) ”Buletin lunar nr. 1/2008” , National Bank of Romania .