Incorporating Sustainability Criteria into Credit Risk Management

Abstract

Does a commercial debtor's economic, environmental and social performance in terms of sustainability affect its credit risk rating? Does adding criteria aimed at assessing a lender's environmental, social or sustainability practices provide added value to traditional financial rating criteria? Many analyses have reported that a correlation exists between companies' environmental and their financial performance. We checked out the assertion that it ‘pays to be sustainable’ by analyzing the role that criteria pertaining to sustainability and environmental orientation play in the commercial credit risk management process. Our results show that sustainability criteria can be used to predict the financial performance of a debtor and improve the predictive validity of the credit rating process. We conclude that the sustainability a firm demonstrates influences its creditworthiness as part of its financial performance. Copyright © 2008 John Wiley & Sons, Ltd and ERP Environment.

Full text

Copyright © 2008 John Wiley & Sons, Ltd and ERP Environment
* Correspondence to: Olaf Weber, GOE – Gesellschaft für Organisation und Entscheidung, Sonnwendbuehel 2, Unterach 4866, Austria.
E-mail: weber@goe.ch
Business Strategy and the Environment
Bus. Strat. Env. 19, 39–50 (2010)
Published online 27 November 2008 in Wiley InterScience
(www.interscience.wiley.com) DOI: 10.1002/bse.636
Incorporating Sustainability Criteria into Credit
Risk Management
Olaf Weber1,2*, Roland W. Scholz2 and Georg Michalik2
1 GOE – Gesellschaft für Organisation und Entscheidung, Unterach, Austria
2 ETH Zurich, Institute for Environmental Decisions IED, Natural and Social Science Interface
(NSSI), Switzerland
ABSTRACT
Does a commercial debtor’s economic, environmental and social performance in terms of
sustainability affect its credit risk rating? Does adding criteria aimed at assessing a lender’s
environmental, social or sustainability practices provide added value to traditional fi nancial
rating criteria? Many analyses have reported that a correlation exists between companies’
environmental and their fi nancial performance. We checked out the assertion that it ‘pays
to be sustainable’ by analyzing the role that criteria pertaining to sustainability and envi-
ronmental orientation play in the commercial credit risk management process. Our results
show that sustainability criteria can be used to predict the fi nancial performance of a debtor
and improve the predictive validity of the credit rating process. We conclude that the
sustainability a fi rm demonstrates infl uences its creditworthiness as part of its fi nancial
performance. Copyright © 2008 John Wiley & Sons, Ltd and ERP Environment.
Received 20 September 2007; revised 16 September 2008; accepted 16 September 2008
Keywords: sustainability; banking; loan; risk; rating; decision making; credit
Introduction
THE CRISIS ABOUT ASSET BACKED SECURITIES BASED ON RISKY MORTGAGES IN 2007 SHOWED THAT PRUDENT RISK
management practices in the credit rating process are more important than ever. This is valid not only for
mortgages, but also for loans to small and medium sized enterprises (SMEs) that are diffi cult to rate with
respect to their credit standing as well. Thus we analyzed the infl uence of a commercial debtor’s economic,
environmental and social risks in terms of sustainability on its credit risk rating. Furthermore, if this is the case,
the question of whether adding criteria aimed at assessing a lender’s environmental, or sustainability, practices to
the classical credit rating criteria improves the risk rating ability should be answered.
A number of academic surveys have identifi ed a positive correlation between environmental performance and
fi nancial performance (Annandale et al., 2001; Dasgupta et al., 2002; Dowell et al., 2000; Klassen and McLaugh-
lin, 1996; Nakao et al., 2007). Studies have been done on the chain of cause and effect between environmental
performance and fi nancial performance (Bansal and Roth, 2000; Lankoski, in press; Reinhardt, 1999; Steger,
2000) as well as on factors that infl uence the strength of the correlation (Russo and Fouts, 1997). Other analyses
suggest that a positive environmental performance can be associated with neutral to positive economic (Ilnitch

40 O. Weber et al.
Copyright © 2008 John Wiley & Sons, Ltd and ERP Environment Bus. Strat. Env. 19, 39–50 (2010)
DOI: 10.1002/bse
and Schaltegger, 1995; Schaltegger and Figge, 2000) or fi nancial performance (Benson et al., 2006; Elsayed and
Paton, 2007; Kreander et al., 2005). On the other hand there are empirical studies that do not show a clear positive
relation between sustainability performance and fi nancial performance (Margolis and Walsh, 2001; Wagner et al.,
2001a). Why can these different results be found?
There are different concepts of sustainability in fi rms (Barnett and Salomon, 2006) such as the reputation of a
fi rm for other managers, the membership in a code of conduct (Cowton and Thompson, 2000), stakeholder
ratings, the environmental or social performance as part of the sustainability performance, policies and
outcomes.
Concepts of business performance range from basic fi nancial indicators such as cashfl ow rate, return on assets
(Simpson and Kohers, 2002) or investment to market indicators such as total income return or price–earnings
ratio (McGuire et al., 1988) to integrated concepts such as the sustainable value-added (Figge and Hahn, 2004).
Additional infl uences on the results of the analyses are regions, sectors, services, regulations (Bleischwitz, 2004),
market capitalization measures (Cerin and Dobers, 2001) or generally the kind of sample used for a study.
The methodical spectrum for analyses on sustainability and business performance is very broad as well. It ranges
from statistical correlative analyses using large samples (Annandale et al., 2001; Dasgupta et al., 2002; Dowell et
al., 2000; Klaassen and Botterweg, 1976; Nakao et al., 2007) to case studies analyzing cause–effect relations
(Steger, 2004, Romero Castro and Pineiro Chousa, 2006).
Environmental risks that led to credit defaults in the past were often linked to real estate collateral. If a lender
accepts real estate collateral as loan security without checking that site or building for contamination, the calculated
value of the security could be higher than the market value in case of soil contamination. As a consequence lit-
erature indicates that most banks consider environmental risks as part of the credit appraisal process (Coulson
and Dixon, 1995; Wanless, 1995; Weber et al., 2008a). Furthermore the World Bank and the International Finance
Corporation (IFC) published detailed guidelines for integrating environmental assessment into the credit risk
assessment (IFC, 1998) that found entrance to environmental assessment strategies of commercial banks and
multi-lateral development banks (Annandale et al., 2001). Thus, what are typical environmental risks in the lending
business?
They are the following (Case, 1996; Thompson and Cowton, 2004; Weber, 2005).
• Sites used as collateral that are contaminated: the contamination of a site affects the value of the collateral in a
signifi cant way, because decontamination is costly.
• Regulatory driven investments: a fi rm can be obliged to invest in environmental technologies because of regula-
tions and has fi nancial problems because of this.
• Market changes: environmental attitudes of consumers or industries may change so that some products cannot
be produced or sold anymore.
• Reputation risk: banks attract bad reputations if they are doing business with fi rms that are in trouble because
of environmental or social problems or if they fi nance projects that are seen as environmentally or socially prob-
lematic by stakeholders.
Most prior research that examined environmental credit risks focused on the security risks from contaminated
real estate collateral and on the rating phase of the credit risk management process (Billiot and Daughtrey, 2001;
Keidel, 1997; Kühne, 1999). But what are the strategies and activities in the other phases of the credit risk man-
agement process?
Credit Risk Management
We used the Brundtland defi nition of sustainable development as our basic defi nition (Brundtland, 1987) and tried
to apply this defi nition to the level of credit management in the banking business to determine how banks could
manage their sustainability risks adequately (Jennings and Zandberger, 1995; Schmidheiny and Zorraquin, 1996).
Furthermore, we broadened concepts that integrate environmental risks into credit management (Hart, 1995) and
included social and economic risks (Gladwin et al., 1995) as well.
Many factors infl uence the borrower’s capital stock, its earnings or its liquidity and must be rated by the lender.
These factors infl uence a borrower’s ability to repay the loan and as a result infl uence the bank’s credit risk. In

Incorporating Sustainability Criteria into Credit Risk Management 41
Copyright © 2008 John Wiley & Sons, Ltd and ERP Environment Bus. Strat. Env. 19, 39–50 (2010)
DOI: 10.1002/bse
credit risk management, a company’s balance sheet, its quantitative indicators and its qualitative indicators (such
as management skills) are among the indicators used to identify and manage the risk of a borrower (Caouette
et al., 1998; Fitch, 1997; Saunders, 1999; Weber, 1997b).
In our study, we concentrated on the so-called counterparty credit risks, which are mainly infl uenced by the
reputation of the debtor, the ability to repay, future earnings, the debtor’s capital and its ratio to debt and the value
of the collateral (Saunders, 1999).
Sustainability risks could signifi cantly infl uence these factors (Thompson, 1998b, Coulson and Dixon, 1995).
Thus a debtor’s earnings could be infl uenced by increased costs of prescribed investments into environmental
technologies. These investments, in turn, decrease the debtor’s liquidity and therefore the ability to repay the loan.
Moreover, the need for additional capital to invest in environmental end-of-pipe technologies results in a decrease
in the borrower’s capital–debt ratio, which in turn leads to an increase in the bank’s credit risk. In addition, the
site of a commercial borrower used as collateral can depreciate, thus increasing the credit risk for the bank. There
can also be indirect infl uences that affect a borrower’s earnings; non-adequate wages in developing countries, for
example, can result in consumer boycotts, putting a borrower’s reputation at risk.
Scientifi c analyses of the management of sustainability credit risks in banks have concentrated mostly on banks’
environmental risk management. One result of this fi nding was that in about 10% of all credit losses in German
banks environmental risks were involved (Scholz et al., 1995). In a follow up study, Weber (1997a) deduced that
the main reasons for credit losses attributable to environmental issues were a reduction in securities from con-
tamination, expenditures to resolve environmental issues forced by a regulatory body, environmental disasters and
environmentally caused market changes. These risks indicate that banks should place increasing importance on
environmental credit risk management in their corporate lending operations (Nitsche and Hope, 1996; Thompson,
1998a).
The risks mentioned above, however, interact with a debtor’s fi nancial performance. A debtor with a good capital
basis and liquidity is able to negate environmental risk by reversing the effect of environmental damage – such
as decontaminating a contaminated site – more easily than a debtor with little capital. While the former could be
a good, low risk client to a bank, the latter could be very risky.
In spite of this, most banks manage sustainability risks separately from traditional credit risks, because they do
not perceive them as being that infl uential (Weber et al., 2008a), instead of integrating them into their general
risk management (Lundgren and Catasús, 2000; Watchman, 2005).
Thus sustainability risk – the uncertainty about the future outcome of loans for a lender emerging from envi-
ronmental, economic and social sustainability risks of a fi rm being a borrower – could infl uence its economic
performance, i.e. its solvency or future earnings. There is a potential link and the challenge is to identify under
which circumstances this link exists (Berman et al., 1999; Edwards, 1998; Esty and Porter, 1998; Friedman, 1970;
Hart and Ahuja, 1996; Judge and Douglas, 1998; Klassen and McLaughlin, 1996; Louche, 2001; McGuire et al.,
1988; Munn, 1998; Pava and Krausz, 1996; Reinhardt, 1999; Repetto and Austin, 1999; Schaltegger and Figge,
2000; Schaltegger and Synnestvedt, 2002; Stigson, 2001; Wagner and Schaltegger, 2003; Wagner et al., 2001b;
Weber, 2006). In turn the economic performance of a company has a link to its creditworthiness and its credit
risk (Caouette et al., 1998; Coulson and Monks, 1999; Keidel, 1997; Saunders, 1999; Schaltegger and Thomas,
1996) as it can infl uence the solvency or future earnings. This risk is transferred to the lender who has to rate and
manage the risk. Thus, if the lender does not identify these sustainability risks, or rates these risks incorrectly, his
credit decision possibly does not refl ect the real risk of the credit.
In the following sections we describe which economic, environmental and social indicators we used to analyze
the infl uence of a fi rm’s sustainability risks on its creditworthiness.
Sustainability Risk Indicators in Credit Risk Management
It is not easy to apply economic concepts of sustainability to the level of a fi rm (Jennings and Zandberger, 1995)
and on the level of its creditworthiness. According to Welford (1995), to be sustainable, a fi rm should minimize
its use of non-renewable resources and its emissions of pollutants. This defi nition corresponds to the meaning of
strong sustainability. Moreover, the fi rm should contribute to a reduction of unemployment and use long-term
considerations and indicators of success (Callens and Tyteca, 1999).

42 O. Weber et al.
Copyright © 2008 John Wiley & Sons, Ltd and ERP Environment Bus. Strat. Env. 19, 39–50 (2010)
DOI: 10.1002/bse
Studies that look especially at the interaction between environmental risks and the credit risk of a company are
not very frequent. Though some of them analyze the interaction between environmental risks and credit risks
(Coulson and Monks, 1999; Coulson, in press; Cowton and Thompson, 2000; Thompson, 1998b; Thompson and
Cowton, 2004) or the integration of environmental aspects into the credit rating process (Weber, 2001; Weber
et al., 2008a; Weber and Sell, 2002), there are only a few studies that show whether the integration of environ-
mental risk rating into the rating process improves the process, resulting in lower credit risks (Billiot and
Daughtrey, 2001).
A specifi c challenge of sustainability is the social dimension. The social risks of a fi rm level include such con-
siderations as the handling of participation, salary fairness, job safety, health standards at the workplace and
creation of new jobs (Callens and Tyteca, 1999). They were not regarded as early as environmental aspects, but
became more and more important as an infl uence factor for the fi nancial performance of companies and thus
have to be integrated in the risk management process as well (Sharma and Ruud, 2003; von Geibler et al., 2006).
Other studies, however, could not fi nd a correlation between a fi rm’s social performance and its fi nancial perfor-
mance (Hamilton et al., 1993; Mueller, 1991).
Methods
The Rating of Commercial Loans
We classifi ed loans into two groups according to risk: non-default loans and default loans. Group 1, the non-default
loans, contained loans in which the client fulfi lled all of his contractual loan agreements. Group 2 contained those
loans that were fi nancially negative for the bank. This includes loans that required cost intensive supervision
measures or resulted in credit defaults as defi ned in the Basel II framework.
While on the one hand classifying the loans in two groups is simple and comprehensible, on the other hand
statistical methods – such as discriminant analysis – reinstate a continuum of loan risk from the two groups, using
the distances between the scores of the cases and the group centroids of the two categories as indicators of the
risk of a case (Caouette et al., 1998).
Sample and Procedure
In 40 German banks, credit offi cers chose their most recent non-default and default commercial loans to attain a
random sample of a spectrum of loans. These 40 banks are a sample of the member banks of the German Savings
Bank Association.
The participants were credit offi cers who mainly give loans to small and medium sized enterprises. They were
not specialists in rating the environmental or sustainability risks of their debtors. The credit rating process in these
cases is based on traditional scoring or rating of the creditworthiness.
At the beginning of a session, we asked the loan offi cers to classify the loans as non-default or default based on
the description above, in order to calculate the prognostic validity of the criteria. The credit offi cers rated their
loans, using the questionnaire described in the following section with scales from 1 = ‘very bad’ to 5 = ‘very good’,
in a face-to-face situation from a post hoc view in which they already knew the fi nancial outcomes of the loans.
Questionnaire
The questionnaire consisted of 91 items covering four fi elds. We used 33 criteria for the traditional credit risk
assessment (xtrad), basing it on a standardized credit risk rating system that already was used in many banks and
is under a continuous validation process. The remaining 58 items were sustainability criteria, which we extracted
from a number of studies described in the sections above. We checked the homogeneity of the items in the four
groups with a reliability analysis using Cronbach’s alpha.
As a result of the reliability analysis, we selected 85 items for the later analysis. The 85 items were those that
did not decrease the alpha of the scale. Thirty-one of the items were economic sustainability criteria (xecon), 15 were
environmental criteria (xenv) and 6 were social criteria (xsoc) (Table 1). All of the groups scored a Cronbach’s alpha

Incorporating Sustainability Criteria into Credit Risk Management 43
Copyright © 2008 John Wiley & Sons, Ltd and ERP Environment Bus. Strat. Env. 19, 39–50 (2010)
DOI: 10.1002/bse
Traditional criteria
(Cronbach’s alpha = 0.91)
Economic sustainability
criteria
(Cronbach’s alpha = 0.83)
Environmental sustainability
criteria
(Cronbach’s alpha = 0.76)
Social sustainability criteria
(Cronbach’s alpha = 0.75)
Reputation Net debt service Costs of environmental
measures
Wage policy
Legal capacity to borrow Sustained growth Emissions Health policy
Competency of
management
Quality of growth Environmental friendliness of
construction
Social security of the
employees
Follow-up regulation Sector development Consideration of nature and
landscape
Workers’ participation
Relations to the lender Integration of environmental
aspect in economic
decision making
Soil erosion Conservation of workplaces
Potential for development Robustness against crises Sealing of soil Flexibility of working
conditions and working
hours
Attainment of budget Personal resources Sewage emission
Dividend policy Community relations Sewage quality
Sector Risk of accidents Air emission
Region Job creation Noise emission
Product and market Adequacy of fi rm size Resource protection
Competition Eco-effi ciency Material use
Clients Information and
communication
Ratio of renewable and non-
renewable resources
Suppliers Material productivity Use of renewable energy
Volume of orders Spatial relation Use of water (amount)
Future margin Commuter mobility
Agency report Car fl eet
Credit limit Energy effi ciency
Account turnover Technical update of power
plants and machines
Outstanding interest and
amortization
Amount of waste
Auditing company Waste management
Management systems Toxic waste
Trustee Contaminated sites
Personal securities Technology management
Physical securities Material substitution
Liquidity ratio Longevity
Return on equity Recycling capacity
Cash fl ow ratio Redemption of used products
Debt ratio Miniaturization of products
Free cash fl ow Ecological product design
Equity-to-fi xed-assets ratio Contracting
Self-fi nancing ratio
Risk of change in interest
rates
Table 1. List of items separated into the four criterion blocks used in the questionnaire, including the Cronbach’s alpha for each
criterion

44 O. Weber et al.
Copyright © 2008 John Wiley & Sons, Ltd and ERP Environment Bus. Strat. Env. 19, 39–50 (2010)
DOI: 10.1002/bse
of 0.75 or higher, which is considered satisfactory (Cronbach, 1951). After the reliability check, we merged the
items into the four categories using the mean of the single items.
Furthermore, we used the participants’ experience with sustainability risks (experienced versus inexperienced),
the debtor’s categorization as belonging to the service sector or production sector and the type of bank (state bank
versus savings bank) as categorical control variables.
Analysis
To analyze the results of the rating process we conducted a two-step analysis (see Figure 1). In the fi rst step, we
analyzed how accurate sustainability criteria (independent variables) are at predicting credit risks using multiple
regression analysis. In a second step, we analyzed whether combining traditional and sustainability criteria as
independent variables improved the validity of discriminating non-defaults from defaults using discriminant
analysis.
Discriminant analyses can be used to predict group membership on the basis of quantitative predictor variables.
They are standard methods of credit risk prediction (Lachenbruch, 1975; Saunders, 1999). Though there are newer
methods as well (Bensic et al., 2005), discriminant analyses are still frequently used models in credit risk rating
practices (Altman, 1968).
One type of validity in discriminant analysis is predictive validity, measured by correct classifi cation of the loans.
If the integration of sustainability criteria improved the predictive validity of the discriminant analysis, there should
be a signifi cant improvement in the correctness of the classifi cations. To measure this improvement, we used the
McNemar test (Siegel and Castellan, 1987). This test is typically used in a repeated measurement situation, in
which each subject’s response is elicited twice. The McNemar test determines whether classifi cation by traditional
ratings and classifi cation by a combination of both traditional and sustainability ratings are equivalent.
Results
In different German regions, we asked all of the savings banks and state banks with a balance sheet of higher than
1 million b to participate in our study by having three representatives at a time analyzing one default and one
non-default. Out of the 40 banks, 70% participated. Thus we analyzed 180 loans by 74 loan offi cers. If a bank was
Figure 1. Two steps to analyze the predictive validity of traditional and sustainability criteria.
Economic sustainability criteria
Environmental sustainability criteria
Social sustainability criteria
Traditional rating criteria
Categorial classification of the
loans by the credit officers
Step 1: Regression analysis
Step 2: Discriminant analysis
Economic sustainability criteria
Environmental sustainability criteria
Social sustainability criteria
Traditional rating criteria
Environmental sustainability criteria

Incorporating Sustainability Criteria into Credit Risk Management 45
Copyright © 2008 John Wiley & Sons, Ltd and ERP Environment Bus. Strat. Env. 19, 39–50 (2010)
DOI: 10.1002/bse
not able to provide three representatives, we asked the remaining representatives to analyze more than two
cases.
Eight percent of the participants had had between 1 and 5 years of job experience, 34% had had between 6 and
9 years and 58% had had more than 10 years. The majority of the participants were between 30 and 39 years old.
Looking at job experience and age, we assert that the participants were very experienced in the credit rating
process.
In Table 2 we present the descriptive statistics for the ratings.
Analysis of Covariates
We checked the infl uence of covariates by using Chi square tests to evaluate whether there is a signifi cant relation
between the classifi cation of a loan as default or non-default and the covariates ‘type of bank’ (savings banks versus
state banks, Chi square = 0.280, df = 1, p = 0.596), ‘sector’ (service sector versus production sector, Chi square =
0.479, df = 1, p = 0.489) and ‘experience of the participants with sustainability ratings’ (experienced versus non-
experienced, Chi square = 0.340, df = 1, p = 0.560). We found no covariate to be signifi cantly related to the clas-
sifi cation of a loan as good or bad.
Furthermore, we analyzed whether there are signifi cant differences between the covariates and the ratings (tra-
ditional and sustainability), using three 2 × 4 ANOVAs for repeated measurement (traditional rating, economic,
environmental and social risks as the four levels of the repeated measurement). We did not fi nd that the type of
bank (df = 1, F = 2.86, p = 0.093), the sector (df = 1, F = 2.72, p = 0.102) or the experience (df = 1, F = 2.31, p =
0.131) had a signifi cant effect on the ratings. Analyses of the interactions between the covariates and the ratings
did not indicate that these interactions were signifi cant either (type of bank, df = 3, F = 0.30, p = 0.993; sector,
df = 3, F = 1.50, p = 0.219; experience, df = 3, F = 0.269, p = 0.848).
Predicting Credit Risks Using Sustainability Criteria
As mentioned above, in our fi rst step we analyzed the infl uence of sustainability risks on the credit risk of the
borrower using the three sustainability ratings as independent variables and the traditional rating as a dependent
variable in a stepwise regression model. In Table 3 we present the results of this analysis.
With R square = 0.790 (adjusted R square = 0.614) after Step 3, the independent variables – economic, envi-
ronmental and social risks – were able to predict a signifi cant amount of the variance of the credit risk operation-
alized by the traditional ratings. Economic sustainability was entered in Step 1 (R square = 0.737, pF change < 0.0001),
social sustainability was entered in Step 2 (R square = 0.771, pF change < 0.0001) and environmental sustainability
Variables Group N Mean Std deviation
Traditional criteria all 180 3.42 0.53
default 67 3.00 0.45
non-default 113 3.67 0.39
Economic sustainability all 178 3.36 0.48
default 66 3.03 0.41
non-default 112 3.56 0.41
Environmental sustainability all 122 3.20 0.62
default 43 3.18 0.66
non-default 79 3.22 0.61
Social sustainability all 159 3.14 0.65
default 56 2.71 0.56
non-default 103 3.38 0.57
Table 2. Descriptive statistics for the ratings. N for the criterion groups is different because of missing values

46 O. Weber et al.
Copyright © 2008 John Wiley & Sons, Ltd and ERP Environment Bus. Strat. Env. 19, 39–50 (2010)
DOI: 10.1002/bse
was entered in Step 3 (R square = 0.790, pF change = 0.003). Collinearity statistics did not indicate collinearity of the
variables (Belsley, 1991).
The, at fi rst view, surprising result that loans with higher ratings in economic and social sustainability (positive
beta weights), but lower ratings in environmental sustainability (negative beta weights), have higher ratings in
traditional risk rating (Table 3) demands deeper analysis. Thus, we distinguished between non-defaults and defaults
and correlated traditional ratings with environmental sustainability ratings. This correlation analysis indicated a
signifi cant positive correlation between traditional criteria and environmental criteria for the non-defaults (r =
0.316, p = 0.005), but no signifi cant correlation for the defaults (r = 0.019, p = 0.905).
Combining Traditional and Sustainability Criteria to Predict Credit Risks
To analyze whether integrating sustainability criteria, in addition to traditional criteria, could improve the prediction
of credit risks, we conducted a multiple linear discriminant analysis to determine whether four predictors – traditional
rating, economic risk, environmental risk and social risk – could predict whether a loan would be non-default or
default. We subsequently conducted a McNemar test for related samples to compare the result with the number of
correct classifi cations from using only traditional criteria.
The overall Wilks’s lambda of the discriminant function was signifi cant (lambda = 0.494, df = 4, Chi square =
81.04, p < 0.0001), indicating that overall the predictors differentiated non-defaults from defaults. In Table 4 we
present the beta weights and the signifi cance of the variables in the discriminant function.
As expected, traditional criteria showed the highest weights. The second highest weights showed economic
sustainability. Environmental criteria again – as in the regression analysis – showed a negative weight, while social
sustainability showed a positive weight.
When we tried to predict whether a loan belonged to the group of non-defaults or defaults, we were able to
classify 86.6% of the loans correctly with 85.0% correct classifi cation for the defaults and 87.3% correct classifi ca-
tion for the non-defaults. In order to take chance agreement into account, we computed a kappa coeffi cient and
obtained a value of 0.706 (a value of unity indicates perfect prediction). Finally, to assess how well the classifi ca-
tion procedure is able to make predictions in a new sample, we estimated the percentage of loans accurately clas-
sifi ed using the leave-one-out technique and found that we had correctly classifi ed 85.7% of the cases.
As a next step, we wanted to compare the results of the discriminant analysis above to the prediction of non-
defaults and defaults using only traditional criteria.
Criterion Standardized beta weights Unstandardized beta weights p
Economic sustainability 0.529 0.584 <0.0001
Environmental sustainability −0.197 −0.160 0.003
Social sustainability 0.399 0.300 <0.0001
Constant 1.017 <0.0001
Table 3. Results of the stepwise regression analysis using the traditional rating as the dependent variable after Step 3 (R square
= 0.79)
Criterion Standardized beta weights Non-standardized beta weights
Traditional 0.653 1.732
Economic sustainability 0.381 1.013
Environmental sustainability −0.435 −0.690
Social sustainability 0.258 0.445
Constant −8.509
Table 4. Standardized and non-standardized beta weights of the predictor variables of the discriminant function

Incorporating Sustainability Criteria into Credit Risk Management 47
Copyright © 2008 John Wiley & Sons, Ltd and ERP Environment Bus. Strat. Env. 19, 39–50 (2010)
DOI: 10.1002/bse
The overall Wilks’s lambda of the discriminant analysis that used only traditional criteria was signifi cant as well
(lambda = 0.625, df = 1, Chi square = 83.45, p < 0.0001). We were able to correctly classify 78.9% of the loans in
our sample, with 80.6% correct classifi cation for the defaults and 77.9% correct classifi cation for the non-defaults
(kappa = 0.564).
We conducted a McNemar test for repeated measurement, in order to analyze whether the integration of sus-
tainability criteria improved the number of correct classifi cations. This test evaluates whether the proportion of
correct classifi cations using only traditional criteria differs from the proportion of correct classifi cation using tra-
ditional and sustainability criteria.
While traditional criteria correctly classifi ed only fi ve loans, which had not been correctly classifi ed by the com-
bination of traditional and sustainability ratings, the combination of traditional and sustainability ratings correctly
classifi ed 13 loans, which could not have been correctly classifi ed by traditional criteria alone. The one-sided
McNemar test indicated a signifi cant improvement of correct classifi cations by integrating sustainability criteria
into the discriminant analysis (exact pone-tailed = 0.048, N = 119). Thus wrong predictions could be reduced by 22.7%
using sustainability criteria compared to using only traditional criteria.
Discussion
We analyzed whether it was possible to predict SMEs credit risks using sustainability criteria and whether a com-
bination of traditional and sustainability criteria would improve upon using traditional credit rating criteria alone
for the prediction of non-defaults and defaults.
Our results indicated that sustainability criteria were able to predict traditional ratings. This result agrees with
many other studies showing correlations between a fi rm’s fi nancial performance and its sustainability performance
(Cortazar et al., 1998; Dasgupta et al., 2002; Klassen and McLaughlin, 1996; Nakao et al., 2007; Russo, 2003;
Steger, 2004).
After analyzing the negative weights of environmental sustainability separately for non-defaults and defaults,
we found a signifi cant positive correlation for the fi rst and no signifi cant correlation for the latter. The reason for
this effect could be that integrating the environmental performance only had a positive effect on a fi rm’s perfor-
mance when it was possible to implement complementary assets that moderate the relationship to fi nancial per-
formance (Christmann, 2000; Coulson and Monks, 1999; Sharma and Vredenburg, 1998). Thus fi rms in the
default group were not able to make connections between their environmental performance and an ensuing fi nan-
cial benefi t. If an environmental investment had a negative impact on the fi rm’s fi nancial performance, we should
have found a negative correlation for the bad loans as well. Thus, lenders need to analyze the relation between
their debtors’ sustainability risks and fi nancial performance and must not analyze sustainability risks in an isolated
way.
The integration of sustainability risks into the rating process resulted in improved risk prediction and risk
management for lenders, because sustainability risks infl uence the risk of the loans (Coulson, 2007; Thompson,
1998a; Wagner, 2007). Some of the environmentally caused credit defaults that Scholz et al. (1995) have reported
could have been prevented if the lenders had used a rating system that consisted not only of economic and fi nan-
cial indicators, but also of sustainability indicators. However, both types of indicator should be connected in a
multivariate way to take the relation between them into account.
Conclusion
Additional research analyzing the effi ciency of incorporating sustainability criteria into the credit risk management
process would be reasonable. Our results showed an improvement in correct risk classifi cation from 78.9 to 86.6%
or a reduction of wrong predictions of 22.7%. The use of additional criteria, however, increases expenses. Thus,
further analysis of the cost–benefi t relation between using additional criteria and improving risk management
would be of value both for science and for industry.

48 O. Weber et al.
Copyright © 2008 John Wiley & Sons, Ltd and ERP Environment Bus. Strat. Env. 19, 39–50 (2010)
DOI: 10.1002/bse
References
Altman EI. 1968. Financial ratios, discriminant analysis and the prediction of corporate bankruptcy. Journal of Finance September: 589–
609.
Annandale D, Bailey J, Ouano E, Evans W, King P. 2001. The potential role of strategic environmental assessment in the activities of multi-
lateral development banks. Environmental Impact Assessment Review 21: 407–429.
Bansal P, Roth K. 2000. Why companies go green: a model of ecological responsiveness. Academy of Management Journal 43(4): 717–736.
Barnett ML, Salomon RM. 2006. Beyond dichotomy: the curvilinear relationship between responsibility and fi nancial performance. Strategic
Management Journal 27: 1101–1122.
Belsley DA. 1991. Conditioning Diagnostics: Collinearity and Weak Data in Regression. Wiley: New York.
Bensic M, Srlija N, Zekic-Susac M. 2005. Modelling small-business credit scoring by using logistic regression, neural networks and decision
tree. Intelligent Systems in Accounting, Finance and Management 13(3): 133–150.
Benson K, Brailsford TJ, Humphrey JE. 2006. Do socially responsible fund managers really invest differently? Journal of Business Ethics 65(4):
337–357.
Berman SL, Wicks AC. Kotha S, Jones TM. 1999. Does stakeholder orientation matter? The relationship between stakeholder management
models and fi rm fi nancial performance. Academy of Management Journal 42(5): 488–506.
Billiot MJ, Daughtrey ZW. 2001. Evaluating environmental liability through risk premiums charged on loans to agribusiness borrowers. Agri-
business 17(2): 273–297.
Bleischwitz R. 2004. Governance of sustainable development: co-evolution of corporate and political strategies. International Journal of Sustain-
able Development 7(1): 27–43.
Brundtland GH. 1987. Our Common Future. Oxford University Press: Oxford.
Callens I, Tyteca D. 1999. Towards indicators of sustainable development for fi rms – a productive effi ciency perspective. Ecological Economics
28: 41–53.
Caouette JB, Altman EI, Narayanan P. 1998. Managing Credit Risk: the Next Great Financial Challenge. Wiley: New York.
Case P. 1996. Land, lending and liability. Chartered Banker 2(4): 44–49.
Cerin P, Dobers P. 2001. What does the performance of the Dow Jones Sustainability Group Index tell us? Eco-Management and Auditing 8:
123–133.
Christmann P. 2000. Effects of ‘best practices’ of environmental management on cost advantage: the role of complementary assets. Academy
of Management Journal 43(4): 663–680.
Cortazar G, Schwartz ES, Salinas M. 1998. Evaluating environmental investments: a real options approach. Management Science 44(8):
1059–1070.
Coulson A. In press. How should banks govern the environment? Challenging the construction of action versus veto. Business Strategy and the
Environment. http://dx.doi.org/10.1002/bse.584. Access date 17 November 2008.
Coulson A, Dixon R. 1995. Environmental risk and management strategy: the implications for fi nancial institutions. International Journal of
Bank Marketing 13(2): 22–29.
Coulson A, Monks V. 1999. Corporate environmental performance considerations within bank lending decisions. Eco-Management and Audit-
ing 6: 1–10.
Cowton CJ, Thompson P. 2000. Do codes make a difference? The case of bank lending and the environment. Journal of Business Ethics 24:
165–178.
Cronbach LJ. 1951. Coeffi cient alpha and the internal structure of tests. Psychometrika 16: 297–334.
Dasgupta S, Laplante B, Wang H, Wheeler D. 2002. Confronting the environmental Kuznets curve. Journal of Economic Perspectives 16(1):
147–168.
Dowell G, Hart S, Yeung B. 2000. Do corporate global environmental standards create or destroy market value? Management Science 46(8):
1059–1074.
Edwards D. 1998. The Link Between Company Environmental and Financial Performance. Earthscan: London.
Elsayed K, Paton D. 2007. The impact of fi nancial performance on environmental policy: does fi rm life cycle matter? Business Strategy and the
Environment. http://dx.doi.org/10.1002/bse.608. Access date 17 November 2008.
Esty D, Porter ME. 1998. Industrial Ecology and competitiveness. Strategic implications for the fi rm. Journal of Industrial Ecology 2(1): 25–43.
Figge F, Hahn T. 2004. Sustainable Value Added – measuring corporate contributions to sustainability beyond eco-effi ciency. Ecological Eco-
nomics 48(2): 173–187.
Fitch T. 1997. Dictionary of Banking Terms (3rd edn). Barron’s: Hauppauge, NY.
Friedman M. 1970. The social responsibility of business is to increase its profi ts. The New York Times Magazine 33: 122–126.
Gladwin T, Kennelly JJ, Krause TS. 1995. Shifting paradigms for sustainable development: implications for management theory and research.
Academy of Management Review 20(4): 874–907.
Hamilton S, Jo H, Statman M. 1993. Doing well by doing good? The investment performance of socially responsible mutual funds. Financial
Analysts Journal 49(November/December): 62–66.
Hart S. 1995. A natural resource-based view on the fi rm. Academy of Management Review 20: 986–1014.
Hart S, Ahuja G. 1996. Does it pay to be green? An empirical examination of the relationship between emission reduction and fi rm perfor-
mance. Business Strategy and the Environment 5: 30–37.
Ilnitch A, Schaltegger S. 1995. Developing a green business portfolio. Long Range Planning 28(2): 29–38.

Incorporating Sustainability Criteria into Credit Risk Management 49
Copyright © 2008 John Wiley & Sons, Ltd and ERP Environment Bus. Strat. Env. 19, 39–50 (2010)
DOI: 10.1002/bse
International Finance Corporation (IFC). 1998. Environmental Assessment, Operational Policies OP 4.01. IFC: Washington, DC.
Jennings PD, Zandberger PA. 1995. Ecologically sustainable organizations: an institutional approach. Academy of Management Review 20(4):
1015–1052.
Judge WQ, Douglas TJ. 1998. Performance implications of incorporating natural environmental issues into the strategic planning process: an
empirical assessment. Journal of Management Studies 35(2): 241–262.
Keidel T. 1997. Ökologische Risiken im Kreditgeschäft [Environmental Risks in the Credit Business]. Gabler: Wiesbaden, Germany.
Klaassen LH, Botterweg TH. 1976. Project evaluation and intangible effects: a shadow project approach. In Environmental Economics (Vol. 1),
35–49, Nijkamp P (ed.). Martinus Nijhoff: The Hague.
Klassen RD, McLaughlin CP. 1996. The impact of environmental management on fi rm performance. Management Science 42(8): 1199–1214.
Kreander N, Gray RH, Power DM, Sinclair CD. 2005. Evaluating the performance of ethical and non-ethical funds: a matched pair analysis.
Journal of Business Finance and Accounting 32(7/8): 1465–1493.
Kühne G. 1999. Umweltrisiken im Firmenkundenkreditgeschäft: ein integratives Konzept für Kreditprüfung, -überwachung, und -steuerung [Environ-
mental Risks in the Commercial Lending Business: an Integrative Concept for Credit Check, Credit Monitoring, and Credit Controlling]. Deutscher
Universitätsverlag: Wiesbaden, Germany.
Lachenbruch PA. 1975. Discriminant Analysis. Hafner: New York.
Lankoski L. 2007. Corporate responsibility activities and economic performance: a theory of why and how they are connected. Business Strategy
and the Environment. http://dx.doi.org/10.1002/bse.582. Access date 17 November 2008.
Louche C. 2001. The corporate environmental performance–fi nancial performance link: implications for ethical investments. In Sustainable
Banking. The Greening of Finance, Bouma JJ, Jeucken M, Klinkers L (eds). Greenleaf: Sheffi eld; 187–200.
Lundgren M, Catasús B. 2000. The banks’ impact on the natural environment – on the space between ‘what is’ and ‘what if’. Business Strategy
and Environment 9: 186–195.
Margolis JD, Walsh JP. 2001. People and Profi ts? The Search for a Link Between a Company’s Social and Financial Performance. Erlbaum: Mahwah,
NJ.
McGuire JB, Sundgren A, Schneeweis T. 1988. Corporate social responsibility and fi rm fi nancial performance. Academy of Management Journal
31(4): 854–872.
Mueller S. 1991. The opportunity costs of discipleship: ethical mutual funds and their returns. Sociological Analysis 52(Spring): 111–124.
Munn K. 1998. Responsible Care and Related Voluntary Initiatives to Improve Enterprise Performance on Health, Safety and Environment in the
Chemical Industry, Working Paper SAP2.59/WP.109. International Labour Offi ce: Geneva.
Nakao Y, Amano A, Matsumara K, Genba K, Nakano M. 2007. Relationship between environmental performance and fi nancial performance:
an empirical analysis of Japanese corporations. Business Strategy and Environment 16(2): 106–118.
Nitsche C, Hope C. 1996. The Banking Sector and Environmental Issues – Some Empirical Evidence from Britain and Germany, Research Papers
in Management Studies 7. Judge Institute of Management Studies: Cambridge.
Pava ML, Krausz J. 1996. The association between corporate social-responsibility and fi nancial performance: the paradox of social costs. Journal
of Business Ethics 15: 321–357.
Reinhardt FL. 1999. Bringing the environment down to earth. Harvard Business Review Jul./Aug.(114): 149–157.
Repetto R, Austin D. 1999. Estimating the fi nancial effects of companies’ environmental performance and exposure. Greener Management
International Autumn(27): 97–110.
Romero Castro N, Pineiro Chousa J. 2006. An integrated framework for the fi nancial analysis of sustainability. Business Strategy and Environ-
ment 15(5): 322–333.
Russo MV. 2003. The emergence of sustainable industries: building on natural capital. Strategic Management Journal 24: 317–331.
Russo MV, Fouts PA. 1997. A resource-based perspective on corporate environmental performance and profi tability. Academy of Management
Journal 40(3): 534–559.
Saunders A. 1999. Credit Risk Measurement: New Approaches to Value at Risk and Other Paradigms. Wiley: New York.
Schaltegger S, Figge F. 2000. Environmental shareholder value: economic success with corporate environmental management. Eco-Manage-
ment and Auditing 7(1): 29–42.
Schaltegger S, Synnestvedt T. 2002. The link between ‘green’ and economic success. Environmental management as the crucial trigger between
environmental and economic performance. Journal of Environmental Management 65: 339–346.
Schaltegger S, Thomas T. 1996. Pollution added credit rating (PACT): new dimensions in emissions trading. Ecological Economics 19: 35–
53.
Schmidheiny S, Zorraquin F. 1996. Financing Change: the Financial Community, Eco-Effi ciency, and Sustainable Development. MIT Press: Cam-
bridge, MA.
Scholz RW, Weber O, Michalik G. 1995. Ökologische Risiken im Firmenkreditgeschäft [Ecological risks in commercial credit business]. In
Kreditrisiken aus Umweltrisiken [Credit Risks Caused by Environmental Risks], Overlack-Kosel D (ed.). Economica and Deutscher Sparkas-
senverlag: Bonn; 1–49.
Sharma S, Ruud A. 2003. On the path to sustainability: integrating social dimensions into the research and practice of environmental manage-
ment. Business Strategy and the Environment 12: 205–214.
Sharma S, Vredenburg H 1998. Proactive corporate environmental strategy and the development of competitively valuable organizational
capabilities. Strategic Management Journal 19: 729–753.
Siegel SN, Castellan J Jr. 1987. Nonparametric Statistics for the Behavioral Sciences (2nd edn). McGraw-Hill: New York.
Simpson WG, Kohers T. 2002. The link between corporate social and fi nancial performance: evidence from the banking industry. Journal of
Business Ethics 35: 97–109.

50 O. Weber et al.
Copyright © 2008 John Wiley & Sons, Ltd and ERP Environment Bus. Strat. Env. 19, 39–50 (2010)
DOI: 10.1002/bse
Steger U. 2000. Environmental management systems: empirical evidence and further perspectives. European Management Journal 18(1): 23–
37.
Steger U. 2004. What is the business case for corporate sustainability. Perspectives for Managers June: 2–4.
Stigson B. 2001. Making the link between environmental performance and shareholder value: the metrics of eco-effi ciency. In Sustainable
Banking. The Greening of Finance, Bouma JJ, Jeucken M, Klinkers L (eds). Greenleaf: Sheffi eld; 166–172.
Thompson P. 1998a. Assessing the environmental risk exposure of UK banks. International Journal of Bank Marketing 16(3): 129–139.
Thompson P. 1998b. Bank lending and the environment: policies and opportunities. International Journal of Bank Marketing 16(6): 243–252.
Thompson P, Cowton CJ. 2004. Bringing the environment into bank lending: implications for environmental reporting. The British Accounting
Review 36: 197–218.
von Geibler J, Liedtke C, Wallbaum H, Schaller S. 2006. Accounting for the social dimension of sustainability: experiences from the biotech-
nology industry. Business Strategy and Environment 15(5): 334–346.
Wagner M. 2007. Innovation and competitive advantages from the integration of strategic aspects with social and environmental management
in European fi rms. Business Strategy and the Environment. http://dx.doi.org/10.1002/bse.585. Access date 17 November 2008.
Wagner M, Schaltegger S. 2003. How does sustainability performance relate to business competitiveness? Greener Management International
44: 4–16.
Wagner M, Schaltegger S, Wehrmeyer W. 2001a. The relationship between the environmental and economic performance of fi rms. What does
theory propose and what does empirical evidence tell us? Greener Management International 34(Summer): 95–108.
Wagner M, Wehrmeyer W, Carlens J. 2001b. The relationship between environmental and economic performance of fi rms and the infl uence
of ISO 14001 and EMAS: an empirical analysis. Paper presented at the Eco-Management and Auditing Conference, Nijmegen.
Wanless D. 1995. The Gilbert Lecture 1995: Banking and the Environment. The Chartered Institute of Bankers: London.
Watchman P. 2005. Banking on Responsibility. Part 1 of Freshfi elds Bruckhaus Deringer Equator Principles Survey 2005: the Banks. Freshfi elds
Bruckhaus Deringer: London.
Weber O. 1997a. Credit management and sustainable industrial development. Paper presented at Science for a Sustainable Society – Integrating
Natural and Social Sciences, Roskilde.
Weber O. 1997b. Integration von Wissensmodulen bei der Kreditvergabe [Integration of Knowledge Modules into the Credit Decision Process] (Vol. 7).
Lang: Frankfurt a.M..
Weber O. 2001. Assessment of enterprises environmental risks: using photographs as perceptual stimuli. Journal of Environmental Psychology
21: 165–178.
Weber O. 2005. Sustainability benchmarking of European banks and fi nancial service organizations. Corporate Social Responsibility and Envi-
ronmental Management 12: 73–87.
Weber O. 2006. Investment and environmental management: the interaction between environmentally responsible investment and environ-
mental management practices. International Journal of Sustainable Development 9(4): 336–354.
Weber O, Fenchel M, Scholz RW. 2008a. Empirical analysis of the integration of environmental risks into the credit risk management process
of European banks. Business Strategy and the Environment 17: 149–159.
Weber O, Koellner T, Habegger D, Steffensen H, Ohnemus P. 2008b. The relation between sustainability performance and fi nancial perfor-
mance of fi rms. Progress in Industrial Ecology 5(3): 236–254.
Weber O, Sell J. 2002. Valuation of contaminated sites as collateral. Local Land and Soil News 4(2): 21.
Welford R. 1995. Environmental Strategy and Sustainable Development. The Corporate Challenge for the 21st Century. Routledge: London.