Risk Adjusted Pricing Strategies for the C orporate Loans
Business - Do They Really Create Value?
Thomas J ung
K£W Bankengruppe
Palmengartenstr. 5-9
D-60325 Frankfurt
Phone: +49 621 4014524
E-Mail: mail@ thomas-jung.org
Jiirgen Strohhecker
HfB / Business School of Finance and Management
Sonnemannstrafe 9-11
D-60314 Frankfurt am Main
Germany
Phone: +49 69 154008-110
E-Mail: strohhecker@ hfb.de
Abstract
For a fairly long time, the German corporate loan business has been regarded in
publications on the subject as a value destroyer. On the basis of the bank lending
statistics of the Deutsche Bundesbank, our analysis clearly shows a cyclical as well as a
structural problem preventing the big four German banks from creating value. In order
to analyze possible strategies to solve this issue within a systemic approach, we built a
model which included almost 200 variables. Running this model until the year 2010 the
outcome is appalling: the break-even retum on equity will not be reached. We therefore
modelled in a second step the widely discussed strategy of risk adjusted pricing. The
outcome raises hope - although this strategy in itself is not sufficient to solve the
problem entirely, the retum on equity can be increased.
Key Words: Risk Adjusted Pricing, Corporate Loans Business, Value Creation
Strategies
1 Value Destruction in the German Corporate Loans Business
Regarding retum on equity (RoE), the German corporate loans business has been
destructing value for several years.’ As early as 1999, Andreas Stehman, former Vice
President at Mercer Management Consulting, pointed out:
“In the loans business with corporate clients, shareholder value is destroyed. In 1998,
Deutsche, Dresdner, Commerz, BHF, Hypo-Vereinsbank, Bankgesellschaft Berlin and
IKB have reduced shareholder value by more than three billion DM with corporate loans.
A total lending volume of 900 billion have yielded a mere 3.4 billion of net earnings after
provisions for loan losses” (Stehmann, 1999, p. 23).
Nowadays, the problem is still prevailing. Even though the number of corporate
insolvencies in Germany decreased by 3.5 % in 2005, 37,500 insolvencies are still close
to the all-time high (Creditreform (Ed.), 2005). Loan loss provisions, which are the
biggest expense item in the corporate loans business, are expected to remain at a high
level. The precarious financial situation of the corporate loans business has been a
subject of lively discussion in academic publications as well. In their provocatively
titled book ,,The corporate loans business - a value destroyer?," Rolfes, Schierenbeck
and Schiiller (2001) state:
“The classical corporate loans business has become unprofitable for most banks - due to
high credit risks and low margins. Banks, which position themselves as loan suppliers for
corporate clients, increasingly risk destroying shareholder value”.
This paper investigates the German corporate loans business of the four major private
banks in Germany (Deutsche Bank AG, Dresdner Bank AG, Commerzbank AG und
Bayrische Hypo- und Vereinsbank AG), also known as the Big Four (Deutsche
Bundesbank, 2004, p. 110), between 1999 and 2004. For scenario and policy analysis,
the time horizon is extended to 2010. In accordance with the bank lending statistics of
the Deutsche Bundesbank, corporate loans are defined as loans to German non-financial
corporations and self-employed persons (including sole proprietors) (Deutsche
Bundesbank, 2002, p. 146). The RoE before tax is used as primary measure; it is
compared to capital costs in order to measure the development of shareholder value.
RoE is computed by subtracting loan loss provisions from the net income and dividing
the difference by the regulatory capital (Duhnkrack, 2002, p. 159). Value creation
occurs when RoE is higher than the cost of capital; when RoE is lower than the capital
cost rate, value destruction is observed (Horter, 1997, p. 178). Using the Capital Asset
Pricing Model (CAPM), capital costs between 1999 and 2004 reached 10.85% .2
Figure 1 shows the RoE of the corporate clients business for the Big Four. It clearly
indicates that the threshold to value creation was never crossed. After having suffered
from a negative RoE in 2002 (-2.9%), the banks managed to reverse the trend with RoEs
of 0.4% in 2003 and 3.5% in 2004. Nevertheless, compared to the target value of
10.85 %, a substantial gap persists.
1 For the definition of “destructing value” see p. 2.
2 B =0.922, risk free interest rate of 4.07% and weighted relationship between market and book value for
the Big Four of 1.25 (based on data analysis).
rar
oO
>
&
844
wy —e—Actual RoE
i —e—RoE objective
0: besereeewe nee ceeseceew seed lence cress gM cceecsees
4 T T T ; 1
1999 2000 2001 2002 2003 2004
Figure 1: RoE of the corporate clients business for the four big German private banks
Moreover, Figure 1 indicates the existence of a structural problem besides the cyclical
problem. RoE cycles are primarily caused by the business cycles, which induce
increasing and decreasing loan loss provisions. However, even during the economically
strong years 1999 and 2000, RoE fell short of the 10.85 % target. This alludes to the
structural roots of the problem.
Faced with the unsatisfactory achievement of the RoE objective, bank managers have
been reacting with cost cutting programmes, i.e. process optimisation, layoffs and cuts
in overheads. Additionally, the lending volume - especially for financially weak
customers - was reduced in order to cut back provisions for non-performing loans.
Obviously, those measures were not sufficient to transform the corporate loans business
into a value-creating business segment. A strategy widely discussed in academic
publications on the subject and stimulated by the Basel II regulatory framework is the
strategy of risk adjusted pricing (e.g. Doring, 2001, p. 4 and Zielke, 2004). The paper
deals with the issue whether or not risk adjusted prices are the appropriate action to
eradicate the problem.
2 Dynamic Hypotheses about Bank Management's Past Behaviour
2.1 Management Reaction to Unsatisfactory Achievement of Objectives
In 1999 and 2000, Germany’s gross domestic product (GDP) was growing slightly
above average. Relatively few companies became insolvent and, as a result, loan loss
provisioning of the Big Four was tolerable (2.978 billion Euros in 1999 and 2.352
billion Euros in 2000). Although actual RoE was still below the target, behaviour over
time showed a rising trend. In expectation of further increasing rates of retum and
ongoing economic growth, the Big Four were expanding their lending volume by 4 % in
2000 compared to 1999. Rising net interest revenues from a growing lending volume
raised actual RoE, which caused expectations to mount even higher. The result is the
reinforcing feedback loop R1 shown in Figure 2.
Lending RoE objective
portfolio +
Achievement of
RoE objective
Ca
-
Expected RoE
Net interest +
revenues *
+ Actual RoE
Figure 2: Loan business growth engine
Throughout the years 2001 and 2002, economic growth was declining, becoming even
slightly negative in 2003. For many companies with tight financial resources,
deteriorating sales figures were too hard to cope with. As a result, the number of
insolvencies rose sharply by 41 % during 2003 to 39,470. Of course, this development
affected the operating profit of the Big Four. Loan loss provisions were rising even more
dramatically and RoE was crashing. In 2002, the banks were suffering from 6.119
billion Euros of loan loss provisions - more than twice as much as in the preceding year.
A major fraction of those expenses was caused by the German corporate loans business
(Deutsche Bundesbank, 2004, p. 24):
“As in the years before, in 2003 the biggest fraction of the German banks’ loan loss
provisions have probably been caused by the corporate loans business. Banks were still
facing a high number of corporate insolvencies“.
Due to the combination of dramatically increased provisions and higher levels of
administrative expenses profitability slumped. In 2002, RoE even became negative,
creating pressure to act. Largely, managers took actions in two directions in order to
increase actual RoE and close the gap to targeted RoE (break-even).
Lending RoE objective
portfolio +
Achievement of
RoE objective
ap he
Net interest Expected RoE
+
revenues +
‘Actual RoE
+ +t -minsolencies
Loan loss ne:
rovisions
+
Regulatory
capital
Figure 3: Reducing the lending volume
First, they lowered the lending volume especially for financially weak customers or
rather bad rating classes. This measure had a double impact on actual RoE (Figure 3).
On the one hand, loan loss provisions could be reduced, which simultaneously lowered
4
expenses and therefore increased the numerator of the RoE measure (balancing feedback
loop B1). On the other hand, the reduction in the volume of loans lowered the regulatory
capital required, which decreased the denominator and increased RoE (balancing
feedback loop B2).3
Second, in 2001 the Big Four initiated major cost cutting programmes to reduce
administrative overheads (balancing feedback loop B3 in Figure 4). In 2002, operating
expenses (excluding compensation and benefits) were reduced by nearly 10 % to 10.7
billion Euros, which was almost the same level as in 2000. In 2003, those expenses were
cut back again by approximately 6%. Additional cost cutting in the personnel
department caused mass layoffs. The Big Four decreased the number of employees by
10,850 (Deutsche Bundesbank, 2004, p. 21). In their 2002 annual report, Commerzbank
wrote: “The 2002 financial year was characterized by the need to lower personnel costs
substantially. For this reason, we pressed ahead with our cost-cutting policy introduced
the year before” (Commerzbank, 2002, p. 46). As a result, personnel costs of the Big
Four were significantly lower in 2002 (11.1 billion Euros) compared to 2000 (12.2
billion). In 2003, the intensity of cost cutting efforts levelled off and personnel costs
decreased only slightly to 10.957 billion Euros (Deutsche Bundesbank, 2004, p. 36).
Lending RoE objective
portfolio +
Achievement of
RoE objective
of cost-cutting
Be ¥ ‘ initiatives
Bl
Y Net interest Expected RoE
vf B3
revenues
actual RoE
Insolvencies
Loan loss Administrative
provisions Gy / expenses
+
Number and intensity
Regulatory
capital
Figure 4: Cost-cutting initiatives
While these measures led to the desired outcome in the short term, there were also side
effects to account for:
“In 2003, developments of interest-based business were characterised mainly by banks’
efforts to create the conditions for a structural improvement in their performance by
consistently adjusting their balance sheets for risky assets. However, the decline in the
volume of business, particularly in the area of riskier loans with a higher rate of interest,
depressed net interest received” (Deutsche Bundesbank, 2004, p. 16).
3 Of course the volume of loans is not only affected by management decisions. As shown by econometric
research of Deutsche Bundesbank, high liquidity and low economic growth decreased demand, too (Calza
et. al., 2001 and Calza et al., 2003).
Additionally, there is the danger of significant counter-acting consequences in the
medium and long term. Figure 4 explains the peril of tuming the reinforcing feedback
loop R1 into a vicious cycle by pushing the balancing loops too hard. Reducing the
lending volume because of unsatisfactory achievement of the RoE objective will
decrease net interest revenues. Lower net interest revenues will decrease actual RoE,
causing expectations to suffer and curtailing the lending volume even further.
2.2. Modelling Management's Reactions
This section extends and formalizes the causal loop model in order to analyse the
dynamic interaction of the various measures with the strategy of risk adjusted pricing.
The complete model has more than 200 variables. Therefore, it is impossible to describe
it in full detail in this paper. Instead, we provide an overview of the main variables and
highlight the most important dynamic interrelationships between those variables. For in-
depth inspection, we are happy to provide the Vensim model file.
The model’s core variable is the actual RoE. It is computed by dividing earnings before
tax (EBT) by the regulatory capital. EBT is the result of earnings minus total expenses.
While earnings can be equated by net interest revenues for the purpose of the model,
expenses are made up of three major components: personnel expenses, administrative
expenses and loan loss provisions.
Average life of
loans
Lending portfolio
Lending portfolio Lending portfolio
increase rate decrease rate
<Fractional lending
portfolio increase
rate>
<Write-offs of
non-performing
loans>
Average quality
of loans
Quality of credit
Credit portfolio portfolio Credit portfolio
quality increase rate quality decrease rate
<Average quality of new
loans and loan
extensions >
Figure 5: Volume and quality of loans as co-flow structure
Focusing on the income items first, one major factor determining net interest revenues is
the lending volume, which we modelled as stock (Figure 5). The lending volume
decrease rate is modelled using the fractional decrease rate structure (Sterman, 2000, p.
523). Average life of loans is set to four years using the same data from the European
Central Bank as for computing the empirical credit margins (calculated using diverse
Monthly Bulletins provided by the European Central Bank). Additionally, the lending
volume is decreased by write-offs of non-performing loans, which is done after seven
years on average. For computing the lending volume increase rate, the stock lending
volume is multiplied by the fractional loan volume increase rate, which is itself
influenced by various factors. The result is a structure that uses a normal fractional rate
plus adjustments (similar to Sterman, 2004, p. 524). If on average new loans have a
different creditworthiness than repaid loans, the quality of the bank’s credit portfolio
changes. To model quality as an attribute to the lending volume, a co-flow structure is
used (Sterman, 2000, p. 497-509).
Net interest revenues are calculated by multiplying the lending volume by the average
credit margin. The Big Four’s credit margin is influenced by both the customers’
creditworthiness and the economic situation. Between 1999 and 2004, it developed in
the opposite direction compared to most other variables, especially GDP growth (Figure
6). When the economy was growing moderately, competition was intense and interest
margins were low. In an economic downtum, competition abated and banks were able to
negotiate better terms. In 1999 and 2000, credit margins were low while GDP growth
was relatively high. In 2001, 2002 and 2003 GDP growth suffered and margins were up
to 1.86 %, 1.63 % and 1.55 % respectively. The increased credit margins may have, at
least partially, resulted from a moderate risk adjusted pricing policy (we will revisit this
point later on), besides from reduced competition. As insolvencies held steady in 2004,
competition was accelerating again and margins decreased.
3.5% 4
—@— Interest margin
—e—GDP growth rate
3.0%
2.5%
2.0%
15%
1.0%
0.5%
0.0%
-0.5% T r T T 1
1999 2000 2001 2002 2003 2004
Figure 6: Interest margin and GDP growth rate
Based on the data of 1999 to 2004, a non-linear relationship with a negative slope
between GDP growth rate and credit margin has been modelled. Figure 7 shows the
table function used in the model. As banks can change the credit rate for loans only in
the case of extensions or when they provide new loans, changes in GDP growth do not
affect margin immediately. Therefore, a smooth structure is used to represent this delay
(Sterman, 2000, p. 428).
1.9%
1.8%
B
Nu
se
1.6%
15%
14%
1.3%
Interest margin [% per year]
B
N
x
11%
1.0%
-2.0% -1.0% 0.0% 1.0% 2.0% 3.0% 4.0% 5.0%
GDP grow th rate [% per year]
Figure 7: Table function modelling the relationship between GDP growth and credit margin
On the expense-side, the dominant component is loan loss provisions. The lending
volume is multiplied with the ratio of loan loss provisions to determine the absolute
amount of loan loss provisions. A ratio of loan loss provisions of 0.0071 indicates, for
example, that per 1,000 Euros of loan volume 7.10 Euros of loan loss provisions have to
be made. The data available show that loan loss provisions are strongly correlated with
GDP growth. The correlation coefficient is -0.95. The minus sign is not surprising, since
in times of low or negative GDP growth rates, insolvencies are high, causing a soaring
number of non-performing loans. At the same time, the value of loan collateral suffers.
As aconsequence, banks have to increase loan loss provisions, which, in turn, increases
the ratio of loan loss provisions, too. Therefore, we used the table function shown in
Figure 8 to model this relationship. Again, we used a smooth structure to model the
delayed effect of GDP growth on loan loss provisions.
0,025
0,020
0,015
0,010
Ratio loan loss provisions
0,005
0,000
-1,0% 0,0% 1,0% 2,0% 3,0% 4,0% 5,0%
GDP grow th rate [% per year]
Figure 8: Relationship between “Ratio loan loss provisions” and “GDP growth rate”
Personnel expenses were determined based on the number of employees and their
average salary. As a result of similar salary bands, we did not differentiate between sales
people and credit analysts. Disaggregating the employees in those two categories would
have increased complexity without adding value. Instead, the staff was grouped into
rookies and experienced employees, as this is more important for the amount of
personnel expenses. Over time, employees usually gain experience, which results in
higher salaries. Therefore, a two echelon aging chain was modelled (Figure 9) (Sterman,
2000, p. 470-472). Experienced employees do not only eam higher salaries, they also
raise the discriminatory power when rating corporate customers. More precise credit
ratings are one major prerequisite for successful risk adjusted pricing strategies. Based
on interviews conducted at Commerzbank AG, the time to gain experience was set to
eight years. The decision mule for hiring or firing employees was modelled as a close gap
structure (Hines, 2005, p. 29).
Gap of employees “ ~ <Desired number of
employees >
Time to hire Co ‘ Ploy
Total number of Desired layoffs Time to lay off
employees
Rookie hire rate . . Layoffs of
Time to gain experienced
experience employees
<Desired layoffs > Rookie Experienced
employees , employees =
poy Assimilation ploy Experienced
rate quit rate
ie qui Personnel
Rookie quitrate expenses Labour tumover rate of
experienced employees
Labour turnover aN kie' <Average salary of
rate of rookies Meade [OORes experienced
salary> employees>
Figure 9: Model structure of the employee sector
The third component of expenses, other administrative expenses, has only a relatively
small impact on RoE. Therefore, to avoid complexity, those expenses were modelled as
a fraction of personnel expenses.
As was mentioned earlier, RoE is the core variable influencing management's actions.
Again, we used a close gap structure to model the rules guiding those actions. Based on
the actual RoE measure, management is anticipating the future RoE development and is
setting this up in contrast to the RoE objective. As Figure 10 illustrates, expected
achievement of the RoE objective is reported to management and finally influences
decisions on the number of employees, their salaries as well as the rate of new loans and
loan extensions. In all these three cases, changes in the same direction are observed
when the “reported expected achievement of the RoE objective” varies. If RoE
achievement is down, “desired number of employees”, “employee salaries” and the
“new loan and loan extension rate” drops too. In all three cases, non-linear table
functions were used to model the specific relationships.
Desired number of
Actual RoE RoE objective
employees
qd
Forecasted Expected Reported expected Empl
RoE — achievement of achievement of ROE =m vata eee
RoE objective objective salaries
Time to adjust
/~
New loans and loan
RoE forecast Time to report
extensions
Figure 10: Core close gap structure of the model
4 Dashed arrows indicate effects via cause-and-effect chains.
10
Based on the CAPM, the normal RoE objective was set to 10.85%. However, as
empirical data conceming the loans business indicated, the RoE objective is influenced
by the economic situation. When the economy is expanding and corporate clients’ credit
demand is high, the Big Four seem to reduce their RoE goals to gain market share. If
economic growth slows down, market share can hardly be expanded, making the RoE
objective more important. As a result, the banks focus on the business which is really
profitable. Therefore, the RoE objective was modelled as a variable dynamically
affected by GDP growth.
Moreover, the model focuses on the decision rule for new loans and loan extensions as
it is central to the loans business. Figure 11 shows the structure expanding Figure 5. The
“normal fractional new loans and extension rate”, which is set to 0.25 % per year, is
modified by management decisions on the supply side and customers’ decisions on the
demand side. As mentioned above, management is reluctant to increase the lending
volume, if actual RoE falls short of the objective. Additionally, there is a negative
relationship between loan loss provisions and the willingness to supply new loans and
extend existing loans. Both effects correspond to the R1 and B1 feedback loops in
Figure 3.
<E ffect of economic trend on
demand for new loans and -
loan extensions > Lending
Ss portfolio
Lending portfolio
increase rate
<E ffect of ROE goal
achievement on new loans
and loan extensions supply>
Intended fractional new
~ loans and loan Fractional
extensions rate lending portfolio
increase rate
<E ffect of loan loss Normal fractional new : : .
Pra Time to adjust lending
provisioning on new loans and loans and loan ae
; : portfolio increase rate
loan extensions supply> extensions rate
Figure 11: Decision rule for new loans and loan extensions
On the demand side of the market, requests for new loans and loan extensions will
decrease, if the economic situation is bad. When investment activities are low, corporate
clients will not ask for new loans to buy capital equipment. We modelled this effect by
using the trend function for the formulation of the economic growth trend and a linear
relationship between the GDP growth trend and the effect on demand for new and
extended loans. As can bee seen in Figure 11, the fractional loan increase rate adjusts to
the various pressures with a delay. The smooth structure uses half a year as adjustment
time.
For model testing purposes, several tests were undertaken (Sterman, 2000, p. 858-891).
First, the boundary adequacy and structure assessment tests were carried out by
discussing causal loop diagrams and stock and flow diagrams with loan business
experts. To ensure dimensional consistency, Vensim’s units check function was used.
The model equations were scanned carefully for suspect parameters. Several extreme
5 This is the reciprocal value of the average life of loans.
11
condition tests were performed without violating basic physical laws. The integration
error test revealed insignificant deviations when changing time step and integration
methods. For the behaviour reproduction test, twelve key variables were selected, where
actual data was available. The simulation output was compared to the actual value.
Figure 12 shows the result of the behaviour reproduction test for a selection of four
important variables. Naturally, the fit is not perfect. However, the model reproduces the
data available with reasonable accuracy.
240B €
Volume of loans n> Actual RoE
g Simulation results
200B €} Simulation results & 47) (Base case)
(Base case) e
of 5
Actual data———*,
160 B €t T 4
T T T T 1 T T T T T 1
1999 2000 2001 2002 2003 2004 2005 1999 2000 2001 2002 2003 2004 2005
4B€
Number of employees Loan loss provisions
Actual data
Simulation results
(Base case)
Actual data Simulation résults
(Base case)
3.5007 T T T T T 1 1B€ T T T T 1
1999 2000 2001 2002 2003 2004 2005 1999 2000 2001 2002 2003 2004 2005
Figure 12: Comparison between actual data and simulation output
Having evaluated the model carefully, it can be used with confidence for exploring
various future scenarios and policy analysis.
2.3 Base Run Simulations for Various GDP Growth Scenarios
The model provided a significant amount of feedback. For example, V ensim’s loop tool
indicated that the stock variable lending volume was part of several hundred loops.
Although most relevant variables were modelled endogenously, there was one very
important exogenous variable: GDP growth rate. Because of the importance of this
exogenous factor which cannot be influenced by bank management, we assumed three
scenarios for the future economic development. While the base case scenario supposed
that the German economy was able to grow according to its potential growth rate at
around 2 % per year, the best case scenario was more optimistic and the worst case
scenario much more pessimistic. Table 1 provides an overview of the concrete values
assumed for the three scenarios.
12
2005 2006 2007 2008 2009 2010
Best case scenario 2.75% 3.00% 3.50% 3.25% 3.13% 3.00%
Base case scenario 2.00% 2.50% 2.25% 2.00% 1.75% 1.50%
Worst case scenario 0.50% 0.50% 0.25% -0.10% 0.00% 0.25%
Table 1: GDP growth scenarios for 2005-2010
When simulating the model for the three scenarios, using the values provided by Table
1, the results show that even under very optimistic assumptions about the economic
development in Germany, RoE will not rise above cost of capital. As Figure 13
illustrates, the Big Four will continue to destroy shareholder value in the corporate loans
business.
Best case scenario
Base case scenario
% per year
N ,
¢
‘ ,
ih oe
Worst case scenario ~ =
-2 TTT Tt 7T T7171 T_T T 1
1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
Figure 13: RoE projection for the three scenarios
A closer look at some major measures collected in Table 2 reveals further insight. Loan
loss provisions, the lending volume and the number of employees fluctuate by more or
less 50 % between the best case and the worst case scenario. While loan loss provisions
in the best case scenario, for example, come to 0.949 billion Euros in 2010, they amount
to 1.464 billion Euros in the worst case scenario. Even more impressive is the relative
distance between the two scenarios for the ratio of loan loss provisions. In the worst
case scenario, the ratio is more than double the value of the best case scenario. With a
fluctuation margin of 80 %, RoE is found in a medium position.
Furthermore, although the lending volumes differ by more than 20 % between best and
base case scenario, net interest revenues are pretty much the same. Even in the worst
case scenario, the fractional decrease of net interest revenues is less than the per cent
decline in the volume of loans. Nevertheless, the scenario simulations clearly show that
a structural gap between actual RoE and target RoE persists.
13
Value for 2010
Basecase Bestcase Worstcase
eee scenario scenario scenario
Loan loss provisions (in bn €) 1.237 0.949 1.464
Net interest revenues (in bn €) 2.355 2.476 1.665
Lending volume (in bn €) 170.76 206.10 98.34
Headcount 3,724 4,960 2,330
Ratio of loan loss provisions (in %) 0.72 0.46 1.48
Return on equity (in %) 7.25 7.91 1.58
Table 2: Major measures for the three simulation scenarios
A Monte Carlo analysis was carried out to cross-check the outcome of the scenario
simulations. Instead of three deterministic scenarios for the GDP growth rate, we used
auto correlated pink noise to represent the business cycle (Sterman, 2000, S. 917-922).
In order to generate pink noise, the values for four parameters had to be set. Mean value
for GDP growth rate was used as one Monte Carlo parameter ranging from 0.5 to 3.5 %
per year with an increment value of 0.1. The standard deviation was set to 1 % per year
and auto correlation time was assumed to be 1.5 years. The noise seed constant, which is
used to change the behaviour of the random number generator, was the second Monte
Carlo parameter. It was incremented from 1 to 100, resulting in 100 simulations per
value for the mean GDP growth rate. All in all 3,100 simulations runs were carried out.
Figure 14 shows the result.
20744 r
Confidence bounds
os 50% (4:4! 75% Ifill 95% 222 100%ZZ
v4
> Per
g es UU
8
wu OT
lo}
a
a Mean value for es
4 -1044 actual RoE
<
-20 T T T
1999 2002 2005 2007 2010
Figure 14: Monte Carlo simulation for RoE
The target RoE and thus the threshold to value creation is not reached once. While there
are some constellations with bad RoE results, the upside potential is very limited.
Within the given structure, RoE values beyond 10 % per year cannot be achieved. It is
obvious that the corporate loans business suffers from structural problems. Without
improving business policies, value creation is out of range. Credit margin and, as a
consequence, net interest revenues are still too low to attain the target. The problems of
fierce competition and too low margins are still prevailing. At the beginning of 2005,
Klaus-Peter Miller, CEO of Commerzbank AG, pointed out in an interview conducted
by Financial Times Germany: “Because of the extremely low credit margins we reject
14
even clients with high creditworthiness” (Maier an Jenkins, 2005). Still, revenues are
not sufficient to raise RoE to the desired level.
2.4 Conclusions from the Base Run Simulations
As a result of the simulation analysis conducted so far, two major conclusions can be
drawn. First, the cyclical component of the problem manifesting itself primarily in the
ups and downs of RoE is caused by the business cycle. As GDP growth and insolvencies
of corporate clients are exogenous variables, which the bank’s management cannot
influence directly, management needs appropriate policies to react to the cycles. Second,
a structural problem component can be identified. In the corporate loans business, value
destruction occurs throughout a business cycle. During the economic revival phase,
competition increases and credit margins decline. Therefore, even in a prospering
economy, the Big Four do not reach the target RoE set by their shareholders. The credit
margin is too low to cover the risk of the corporate loans business.
3 Risk Adjusted Pricing as One Possible Strategy for Y ield Increase
Eamings growth is one important leverage for solving the structural problem discussed
above and thus for improving profitability of the domestic corporate loans business.
The banking market in Germany is polypolistic, as even the five largest banks together
hold only a market share of 17% (Pawlowski and Burmester, 2001, p. 346). Thus an
individual bank is a price taker and its influence on the credit margin is limited.
Consequently, a lump-sum increase of credit interest does not present an ingenious way
of boosting income.
3.1 The Concept of Risk Adjusted Pricing
In short, the basic idea of risk adjusted pricing is that financially weak borrowers pay a
higher credit spread (risk premium, credit margin) than financially strong borrowers, as
their probability of default and therefore their credit risk is comparatively higher.
The term “credit risk” comprises “default risk” as well as “spread risk”. While the first
one characterizes the risk of a total loss due to an insolvency of a debtor, the latter can
be described as the risk that the borrower s credit quality worsens over time and thus the
credit spread negotiated at the very beginning of the contract no longer compensates the
bank for the risk it carries (Schierenbeck, 2001, p. 6).
“Credit risk” further comprises “expected loss” and “unexpected loss” (Kirm&e, 2001,
p. 1018). The expected loss can be anticipated by banks with the help of statistical
methods (Broker and Lehrbass, 2001, p. 77), as it is virtually the means of the losses
assumed over a given period of time. The three components needed to calculate it, are
the probability of default (pd), the loss given default (gd) and the exposure outstanding
at the time of default (ead).
Ex post, the unexpected loss can be calculated as the difference between the expected
loss anticipated ex ante and the loss which has actually occurred in a specific period of
time (Pfingsten and Schrock, 2000, p. 12). Ex ante, the unexpected loss is calculated via
quantiles or downside-risk-measures such as the concept of value-at-risk. In a sense, the
15
unexpected loss is the real risk a bank bears when offering credits. For this reason,
regulators all over the world urge banks to charge capital in order to cover potential
losses and thus to guarantee the stability of the whole financial system.® In practice, the
unexpected loss is measured with the help of credit portfolio models such as
CreditRisk+ or CreditMetrics (Broker and Lehrhass, 2001, p. 778).
Within the theoretical framework of the current rate method (Marktzinsmethode), banks
eam an interest differential income due to the deliberate acceptance of the debtor's
credit risk (Rudolph, 2001, p. 334). The “consequent risk adjusted pricing of credits
leads to a price policy within which the single debtors bear, at a given time, the possible
occurring costs of their own defaults via the calculation of risk premiums in an ideal
case. Ultimately, the debtors compensate the bank for the credit risk induced by them.”
(Schiller and Tytko, 2001, p. 211). For our purpose, risk adjusted pricing is defined as a
state, in which the debtor pays a risk premium which equals at least the amount of his
expected loss plus an additional return for the regulatory capital, which banks are urged
to hold in order to cover potential losses (Zielke, 2004, p. 35).
The concept of risk adjusted pricing is not new. However, because of a too faint-hearted
differentiation between the debtors of the different rating classes, banks were
inadequately compensated for the risks they had taken in the past. In total, risk
premiums were too low (Kirm&e, 2002, p. 380). In this sense, Jochen Sanio, head of the
BaFin (German’s Federal Financial Supervisory Authority), criticizes that “the
competition for the cheapest conditions has taken on disastrous dimensions”. He adds:
“The price of a credit has to contain a sufficient risk premium” (Sanio, quoted by
Dahms, 2005). The criticism of many scientists veers toward the same direction. Paul,
professor for Finance and Banking at Bochum University, and Stein, Managing Director
of the Institute for Finance and Banking in Bochum, state:
,Particularly in the German corporate loans business, in the past, the problem was that
the conditions were not conforming to the risk differentiation aimed at by the individual
banks. Frequently, competitors offered unchanged conditions even for bad risks so that a
single bank was forced to join the price cutting war” (Paul and Stein, 2003, p. 49).
If banks do not pursue the concept of risk adjusted pricing, financially strong debtors
pay a risk premium, which is too high, whereas financially weak debtors pay a spread
which is too low. In fact, financially strong debtors subsidize their counterparts, a
phenomenon known as adverse selection (Schiller and Tytko, 2001, p. 216). As a
consequence, they search for cheaper alternatives, either in the capital market or at a
different bank. The financially weak debtors, then again, remain within the credit
portfolio. Thus, in the long run, the average credit quality of the credit portfolio worsens
while, at the same time, the margin is inadequately low (KirmBe, 2002, p. 380).
Applying the concept of risk adjusted pricing can lead to a loss of lending volume. The
absolute amount of the loss is primarily influenced by two variables, the first one being
the difference between Bank A’s credit margin compared to the market interest rate, the
second one being the loyalty of the customers. As borrowers disclose confidential
6 In their internal control systems, the Big Four apply the concept of economic capital, either parallel to
the concept of regulatory capital or exclusively. For the purpose of this paper both we consider that both
are of equal value so that for reasons of simplification the concept of regulatory capital is pursued.
16
information, a certain inhibition threshold exists to change a creditor (Paul, 2000,
p. 1224). Nevertheless, an average increase in margin leads in general to a loss of
lending volume and vice versa.
3.2 Modelling of Risk Adjusted Pricing
In order to simulate the strategic starting point of risk adjusted pricing, we assumed that
¢ risk has already been taken into account in the credit margins in the past, but that
the differentiation between the different classes of credit worthiness has not
proved satisfactory and that
* competitors of the considered banks do not change their pricing scheme.
Against the background of this, we considered a change in the credit margin as a change
in relation to the market credit margin at that point of time. We further supposed that the
lending margin, as calculated for the observation period, counted for a portfolio quality
of 1 and that the portfolio quality was constant in the observation period. Thus, we
modelled the change in lending margins as a shift of the lending margins of the big
banks in comparison to the market credit margin. While creditworthy corporate
customers will profit from lower interest rates, doubtful clients will have to pay higher
rates. On average, risk adjusted pricing should lead to higher interest margins.
As was mentioned earlier, applying risk adjusted pricing influences the lending volume.
We assumed a linear relationship between the interest margin relative to the market and
the demand for loans. In our model, the borrowers’ response to a change of credit
margins can be made more or less elastic by changing the “slope of loans demand
curve” (Figure 15). A slope of zero, for example, indicates that increasing interest
margins have no impact on demand at all. More realistic values for the slope parameter
are, of course, negative reducing demand for loans when interest margins increase.
Up to now, the topic of the elasticity of demand in relation to credit margin has not been
dealt with in any study.” Hence, we assumed that financially weak debtors are restricted
in their choice of lenders so that their elasticity is low. In contrast, financially strong
debtors face many alternatives, including the capital market, and thus possess a high
elasticity. In between these two groups are the debtors with an average financial
stability. As the model is not disaggregated, the parameter “slope of loans demand
curve” represents the behaviour of these three groups on average: in general, while an
increase of the credit margin leads to a loss of lending volume, a decrease leads to
growth (Maier and Jenkins, 2005). Therefore we set the slope to a modest negative
value.
7 Therefore, these values are subject to a certain degree of imprecision due to the estimation. However, as
a sensitivity analysis in 3.4 indicates, demand elasticity is not a very sensitive parameter.
17
Actual RoE
+ ei
Net interest Provisions for
revenues loan losses
+ «
Competition Interest margin
intensity -
ne ‘
Effect of interest margin on ie Z | Lending portfolio
demand for new loans and . .
Risk adjusted loan extensions Lending portiolio = =
aa i increase rate
pricing intensity rf
Slope of loans Fractional lending Average quality
demand cure portfolio increase rate of loans
- +
Average quality of new
and prolongated loans
Quality of credit
portfolio
Credit portfolio
quality increase rate
Figure 15: Modelling the effects of risk adjusted pricing
Aside from the borrower's rating, the intensity of competition determines elasticity of
customers’ demand for loans (Pothoff, 2004, p. 17). The intensity of competition is in
tum strongly influenced by the economic growth: the higher the growth of the GDP, the
higher the competition and vice versa. Accordingly, demand elasticity rises in times of
cyclical uptus, whereas it diminishes in cyclical downturns.
For the purpose of modelling the empirical values of the observation period, the shares
of the different classes of credit worthiness were held constant so that the portfolio
quality takes a normalized value of 1.
As a result of the above mentioned different elasticities of the various classes, the
portfolio quality changes over time, beginning in the year 2005, which initiates three
cause-and-effect chains. First, since the empirical credit margin we calculated is
assumed to be an average margin for all classes of creditworthiness, it increases when
the portfolio quality worsens and vice versa. Secondly, depending on the change in the
portfolio quality, the ratio of loan loss provisions alternates with a time-delay: the worse
the quality, the higher ceteris paribus the coefficient and vice versa. Last but not least,
the desired credit volume per employee varies, as credits of financially weak debtors
take a longer time to be handled with (problem loans).
Desired volume of ,
loans per employee —>> Desired number of employees
Effect of loan quality
Average quality of loans on interest margin
—> Interest margin
Effect of loan quality
on tisk provisioning —» Ratio loan loss provisions
Figure 16: Effect of changes in the average quality of loans
18
Furthermore, we modified an essential rule of behaviour observed in the past. As an
altemative rule of action, the policy makers do not reduce any longer the lending
volume when the ratio of loan loss provisioning is unacceptable and vice versa. Instead,
they alter their pricing scheme according to this ratio. Generally speaking, the big banks
lower the credit margin for financially strong customers under all economic conditions.
The degree of the reduction depends upon the economic situation: the reduction tums
out higher in a phase of an economic boom and lower during an economic downturn. By
contrast, the credit margin is raised for financially weak customers. A gain, the absolute
level depends on the business cycle. In between these two extremes, the credit margin of
customers with an average credit rating is lowered in an economic upturn, whereas it is
raised in a recession. All in all, policy makers raise credit margins in a recession and
lower it during a boom. As one result of this action, the portfolio quality improves over
time.
3.3 Simulation Findings for Different Economic Scenarios
Simulating the introduction of risk adjusted pricing in 2005 with a medium intensity
(0.15) and a medium sensitivity of customers’ demand for loans (slope -2), we obtained
the following results:
Base case scenario
12
B Risk adjusted pricing
> 8
&
x
8 4
& No risk
cE] adjusted pricing
g
<9
“1999 ' 2001 ' 2003 ' 2005 ' 2007 ' 7 2010
Best case scenario Worst case scenario
12 12
3 Risk adjusted pricin s i i
5 justed pricing py, - Risk adjusted pricing
= a
a a
= =
y No risk W
& adjusted pricing é*
: /
< < No risk
adjusted pricing
TT TT T TT T T eT 1 T TT T TT T | oe 1
1999 2001 2003 2005 2007 2010 1999 2001 2003 2005 2007 2010
Figure 17: Effect of risk adjusted pricing strategies in different economic scenarios
As Figure 17 indicates, the strategy of risk adjusted pricing achieves the best effect in
times of economic downturns. As a result of the decreasing competition, which leads at
the same time to decreasing elasticity, banks can demand credit margins, which
adequately compensate them for the risk they carry. Even in times of economic upturns,
the return on equity rises, because the portfolio quality changes for the better and thus
19
the loan loss provisions decrease. The following table summarizes the fundamental
results:
Value in year 2010
Measure Basecase Bestcase Worst case
Loan loss provisions (in bn Euro) 0.966 0.832 1.200
Net interest revenue (in bn Euro) 2.129 2.427 1.802
Lending volume (in bn Euro) 141.13 190.26 83.20
Headcount 3,676 5,268 2,484
Ratio of loan loss provisions (in %) 0.68 0.43 1.44
Return on equity (in %) 9.0 8.7 8.2
Table 3: Major measures for applying risk adjusted pricing in the three simulation scenarios
The results in the different scenarios converge noticeably, as banks are paid according to
the credit risk they carry. This becomes particularly obvious when looking at the net
interest revenue, which is raised in a recession despite a shrinking lending volume. By
contrast, the financially strong borrowers and financially stable borrowers are attracted
in an economic uptum, as they possess a high elasticity in relation to the credit margin:
the lending volume increases noticeably, even though not as strong as in the base
scenario. In addition, this effect leads to a decrease of loan loss provisions in the
medium-term.
A second Monte Carlo simulation with the same set of parameters as for the first one
(page 14), confirmed these results. As illustrated by Figure 18, risk adjusted pricing
contributes to value enhancement of the German corporate loan business. The mean
level of the RoE can be raised by approximately 1.5 percentage points. The break-even
will now be reached with a likelihood of 25 %. Furthermore, it becomes obvious that
tisk adjusted pricing hedges banks against negative returns in economic downtums in
particular.
20) Mean value for actual RoE
using risk adjusted pricing
ptt,
Mean value for actual RoE
104 without risk adjusted pricing
Confidence bounds
50% 11/4 75%Illlll 95% 100%
T T T 1
1999 2002 2005 2007 2010
Actual RoE [% per year]
-20
Figure 18: Monte Carlo simulation for RoE with risk adjusted pricing strategy
20
3.4 Sensitivity of the Demand for Loans to Interest Margin
As stated above, we applied estimated values for the elasticity of credit demand in
relation to the credit margin. In order to challenge the results, we conducted a sensitivity
analysis in the next step. For the base case scenario of economic development, risk
adjusted pricing was introduced in 2005 with a slope of the loans demand curve varying
from 0 (no impact of changes in interest margin on demand) to -6 (high impact of
changes in interest margin on demand). As Figure 19 shows, when customers are very
sensitive to changes in interest rates (-6), risk adjusted pricing causes actual RoE to peak
in 2006. The decline in the years 2007 to 2010 is due to the fact that lending volume is
decreasing, because customers refuse to obtain loans with the risk adjusted interest
rates. However, even in the case of very sensitive customers, risk adjusted pricing
improves actual RoE to 8.3% in 2010 when compared to the base case simulation,
which leads to an RoE of 7.25 % (Figure 13).
10%
3
5 9%!
2"
w 8%} 2010
[4 i
= 2009
g 7% 2008
2007
9
o% 7 ~ 2006
nT pact 4 3 2 “a ~ 2005
lemand Slope of loans demand curve No impa t
ci
On demand
Figure 19: Sensitivity analysis of customers’ reactions to interest margin changes
for economic base case
Even with a variation in economic development in the broad range of 0.5 % to 3.5 %
mean GDP growth from 2005 to 2010 and when analysing the simulated RoEs for 2010,
risk adjusted pricing policies are under all circumstances preferable to traditional pricing
(Figure 20). For the combination of high economic growth (> 2.5 % per year) and very
price sensitive customers, risk adjusted pricing has no clear advantage. However, there
is no disadvantage either. Once again, it becomes obvious that risk adjusted pricing can
protect the corporate loans business performance against economic downswings. It is
most effective when the economy grows on average by 1 to 1.5 % per year.
21
11% 5
10% +
9% +
8% +
T% 4
—e—RAP, slope -6
—a— RAP, slope -2
—e—RAP, slope 0
—m—0N0 risk adjusted pricing (RAP)
6% |
Actual RoE 2010 [% per year]
5% 4
4%
0,5 1 1,5 2 2,5 3 3,0
GDP growth rate [% per year]
Figure 20: Sensitivity of RoE in 2010 to various scenarios of economic growth, risk adjusted
pricing and customers’ price sensitivity
3.5 Conclusions from Simulation Risk Adjusted Pricing Strategies
Risk adjusted pricing may enhance the value of the German corporate loan business
even if customers’ sensitivity to changes in interest rates is much higher or lower than
originally assumed. Applying this strategy, banks are able to hedge their income against
economic downturns and to improve their results even in times of booms (if customers
are not too sensitive to changes in interest rates).
Despite this impressive achievement, we want to emphasize that this strategy alone is
not sufficient to solve the problem of the value destruction. Target RoE is achieved only
in the unrealistic case of 1.25 % steady GDP growth in the years 2005 to 2010 and only
in the presence of completely insensitive customers, who do not reduce their demand for
loans when interest rates are increased.
Additionally, several fundamental prerequisites must be met. The existence of adequate
rating systems is one critical requirement, as incorrect estimations of the credit
worthiness of the borrowers lead to an over- or underestimation of the risk. While an
overestimation causes a loss of lending volume due to comparatively high margins, an
underestimation entails margins which are too low to compensate for the risk. Thus, the
risk of each individual borrower has to be assessed correctly. Another success factor is
the sales force. The employees must be able to discuss critical and difficult aspects with
customers, but at the same time increase customers’ loyalty. Y et, this point should be
regarded as an opportunity: when arguing about weaknesses, the customer is able to
eliminate them and thus improve his financial situation. Additionally, increased
perceived pricing fairness raises the loyalty as well.
4 Transformation of the German Corporate Loan Business into a Value Creator
For a fairly long time, the German corporate loan business has been regarded in
publications on the subject as a value destroyer. The analysis of the variables and their
22
interrelations confirms this point of view. Our model further reveals that the German
corporate loan business will not reach the target RoE without a policy change.
This paper has analyzed risk adjusted pricing as one strategy to solve the structural
problem of value destruction. The outcomes support the hypothesis and document that a
policy change in the pricing scheme may raise the economic profit depending on the
state of the economic growth by up to 5 percentage points. Banks can boost value
creation if they demand fair credit margins. By applying risk adjusted pricing, a bank is
hedged against painful losses, and the strategy achieves its best results in economic
downturns - offence is once more the best defence.
In order to question the findings, we conducted two sensitivity analyses. Those analyses
have further strengthened our argument.
Despite the positive findings, we advise to restrain optimism. Although the structural
problem can be partly solved, risk adjusted pricing without supporting measures will not
lead to achieving the targeted RoE automatically. In order to solve the structural
problem as a whole, a constitutive program is needed. A new calibration of rating
systems together with training of the credit analysts enhances the discriminatory power
of credit analysis and thus leads to shrinking loan loss provisions with a certain time-
delay. Furthermore, banks should make use of the chances of credit portfolio
management: with the help of credit derivatives and asset-backed securities, financial
institutions can actively manage the risks of their credit portfolios without jeopardizing
the underlying business connection with their clients. Thus, the German corporate loan
business can be transformed into a value creator.
5 References
Broker F, Lehrbass F. 2001. Kreditportfoliomodelle in der Praxis. In Handbuch Bankcontrolling,
Schierenbeck H, Rolfes, B, Schiiller S (ed.). 2nd edition. Gabler: Wiesbaden, 773-787.
Calza A, Gartner C, Sousa J. April 2001. Modelling the demand for loans to the private sector in the Euro
area, In European Central Bank Working Paper Series, European Central Bank (ed.),
Working Paper No. 55.
Calza A, Gartner C, Sousa J. January 2003. Aggregate loans to the Euro area private sector. In European
Central Bank Working Paper Series, European Central Bank, Working Paper No. 202.
Commerzbank AG (ed.). Access 23.12.2004. Geschaftsbericht 2001, URL: https://www.commerzbank.de/
aktionaere/archive/konzem/2001/gb2001/gb_de01/gb_de01.pdf .
Commerzbank AG (ed.). Access 23.12.2004. Geschaftsbericht 2002, URL: https://www.commerzbank.de/
aktionaere/archive/konzern/2002/gb2002/interaktiv/download/commerz_ar_d_2002.pdf.
Commerzbank AG (ed.). Access. 23.12.2004. Geschéaftsbericht 2004, URL: https://www.commerzbank.de/
aktionaere/archive/konzern/2004/gb2004/interaktiv/download/commerz_gb_d_2004.pdf.
Dahms J. 28.01.2005. Sanio kritisiert deutsche Banken: Der oberste Finanzaufseher warnt Geldhauser vor
desastrésem Wetthewerb um Kredit-Kunden. In Berliner Morgenpost, URL:
http://morgenpost. berlin! .de/content/2005/01/21/wirtschaft/729842. html (28.01.2005).
Deter H, Diegelmann M (ed.). 2003. Creditor Relations - Beziehungsmanagement mit
Fremdkapitalgebern. Bankakademie-V erlag: Frankfurt a. Main.
Deter H, Diegelmann M. 2003. Creditor Relations - Instrument zur langfristigen Finanzierungssicherung.
In Creditor Relations - Beziehungsmanagement mit Fremdkapitalgebern. Deter H,
Diegelmann M (ed.). Bankakademie-V erlag: Frankfurt a. Main, 1-19.
Deutsche Bundesbank (ed.). Juni 2002. Bankenstatistik: Richtlinien und Kundensystematik - Statistische
Sonderveroffentlichung.
23
Deutsche Bundesbank (ed.). Access: 23.12.2004. Monatsbericht September 2004, URL:
http://www.bundesbank.de/download/volkswirtschaft/monatsberichte/2004/200409mb_bbk.
pdf.
Deutsche Bundesbank (ed.). Access: 06.01.2005. Bankenstatistik Dezember 2004: Statistisches Beiheft
zum Monatsbericht 1, URL: _http://www.bundesbank.de/download/volkswirtschaft/
bankenstatistik/2004/122004bankenstatistik.pdf.
Dohring J. 2002. Kreditkalkulation und Basel II. Seminar ,,Finanzdienstleistungen in der
Unternehmensfinanzierung“ an der Universitat Diisseldorf.
Duhnkrack T. 2002. Wertorientierte Steuerung des Firmenkundengeschafts. In Handbuch
Firmenkundengeschaft - Technologie, Rating und Risikosteuerung als Kernkompetenz der
Banken. Juncker K, Priewasser E. (ed.). 2nd edition. Verlag Fritz Knapp: Frankfurt a. Main,
153-162.
European Central Bank (ed.). Access: 23.11.2004. Monatsbericht Juni 2004, URL:
http://www.bundeshank.de/download/ezb/monatsherichte/2004/200406ezb_mb_gesamt.pdf.
Horter S. 1997. Shareholder Value-orientiertes Bank-Controlling. Verlag Wissenschaft & Praxis:
Sternenfels [et al.]. Doctoral Dissertation Eichstatt, Kath. Universitat.
Kirm®e S. 1998. Kreditrisikosteuerung im Firmenkunden-Portfeuille. In Das Firmenkundengeschaft - ein
Wertvernichter?. Rolfes B, Schierenbeck H, Schiller S (ed.). Fritz Knapp: Frankfurt a.
Main, 101-130.
Maier A, Jenkins P. Access: 28.01.2005. Niedrige Margen belasten Commerzbank. In Financial Times
Deutschland (ed.), URL: http://www. ftd.de/ub/fi/1106385295643.html?nv=cpm.
Oehler A (ed.). 2000. Kreditrisikomanagement: Portfoliomodelle und Derivate. Schaffer-Poeschel:
Stuttgart, 2000.
Paul S. 2000. Marketing. In Geld-, Bank- und Borsenwesen - Handbuch des Finanzsystems. von Hagen J,
von Stein JH. 40th edition. Schaffer-Poschel: Stuttgart, 1211-1293.
Paul S, Siewert KJ. 1998. Bank-Controlling I: Ertragsmanagement in Kreditinstituten, 3rd edition.
Bankakademie-V erlag: Frankfurt a. Main, 1998.
Paul S, Stein S. 2002. Basel II und die deutsche Kreditwirtschaft - ein Uberblick. In Basel II, Mittelstand
und Kreditpreise. Paul S (ed.). Bankakademie Verlag: Frankfurt a. Main, 35-70.
Pawlowski N, Burmester C. 2001. Ableitung von Standardrisikokosten auf der Basis von Expected-Loss-
Kalkulationen. In Handbuch Bankcontrolling. Schierenbeck H, Rolfes B, Schiiller S (ed.).
2nd edition. Gabler: Wiesbaden, 345-372.
Pfingsten A, Schréck G. 2000. Die Bedeutung von Krediteinstufungsmodellen im Bankwesen. In
Kreditrisikomanagement: Portfoliomodelle und Derivate. Oehler A (ed.). Schaffer-
Poeschel: Stuttgart, 1-23.
Potthoff C. 2004. Wetthewerb driickt Preise fiir Grofkredite: Deutsche Banken stéhnen tiber
Kampfkonditionen auslandischer Konkurrenten - Firmen mit guter Bonitat profitieren. In
Handelsblatt of 09.12.2004, Nr. 240, 17.
Rolfes B. 2001. Renditeanspniche fiir Gesamtbank und Geschéftshereiche. In Handbuch Bankcontrolling.
Schierenbeck H, Rolfes B, Schiiller S (ed.). 2nd edition. Gabler: Wiesbaden, 509-518.
Rolfes B, Schierenbeck H, Schiller S (ed.). 2001. Das Firmenkundengeschaft - ein Wertvernichter?. Fritz
Knapp: Frankfurt a. Main.
Rudolph B. 2001. Kalkulation von Risikokosten auf Basis des Optionspreismodells. In Handbuch
Bankcontrolling. Schierenbeck H, Rolfes B, Schiiller S (ed.). 2nd edition. Gabler:
Wiesbaden. 331-343.
Schierenbeck H. 2001. Das duale Steuerungsmodell. In Handbuch Bankcontrolling Schierenbeck H,
Rolfes B, Schiiller S (ed.). 2nd edition. Gabler: Wiesbaden, 87-101.
Schierenbeck H. 2001. Ertragsorientiertes Bankmanagement, Band 2: Risiko-Controlling- und integrierte
Rendite/Risiko-Steuerung, 7th edition. Gabler: Wiesbaden.
Schierenbeck H. 2001. Relationship- Management als zentraler Erfolgsfaktor im Firmenkundengeschaft. In
Das Firmenkundengeschaft - ein Wertvernichter? Rolfes B, Schierenbeck H, Schiller S
(ed.). Fritz Knapp: Frankfurt a. Main, 187-216.
24
Schierenbeck H, Rolfes B, Schiller S (ed.). 2001. Handbuch Bankcontrolling. 2nd editon. Gabler:
Wiesbaden.
Schiller B, Tytko D. 2001. Risikomanagement im Kreditgeschaft - Grundlagen, neuere Entwicklungen
und Anwendungsbeispiele. Schaffer-Poeschel: Stuttgart.
Stehmann A. 1999. Das Firmenkundenkreditgeschaft als Wertfalle deutscher Banken - Die isolierte
Sanierung des Firmenkundengeschéfts wird scheitern / Neue Wettbewerber kommen auf. In
Frankfurter Allgemeine Zeitung from 09.08.1999, Nr. 182, 23.
Sterman JD. 2000. Business Dynamics: Systems Thinking and Modeling for a Complex World. Mc-Graw-
Hill: New Y ork [et al.].
Suyter A (ed.). 2004. Risikomanagement: Aktuelle Entwicklungen und Auswirkungen auf Banken und
Unternehmen, Fritz Knapp: Frankfurt a. Main.
Von Hagen J, von Stein JH. 2000. Geld-, Bank- und Borsenwesen - Handbuch des Finanzsystems. 40th
edition. Schaffer-Pdschel: Stuttgart.
Zielke A. 2004. Risikoadjustiertes Pricing von Kreditrisiken. In Risikomanagement: Aktuelle
Entwicklungen und Auswirkungen auf Banken und Unternehmen. Suyter A. (ed.). Fritz
Knapp: Frankfurt a. Main, 25-44.
25
