Reducing credit risk by promoting sustainability and
responsibility
Master thesis, MSc Finance
Geno Kavaldzhiev, S2724200
Faculty of Economics and Business,
1
Abstract
2
1. Introduction
Just a couple of years before the beginning of the implementation of Basel IV there seem to be many problems among banks in the European Union. Basel IV will be a framework for bank capital requirements and financial disclosure. The European Central Bank posted in the last quarter of 2014 the results of a year long stress test of the resistance among the 130 largest euro area banks. 136 billion euros are described as non-performing exposure and 25 banks as having a capital shortfall of 25 billion euros. When using a standard definition for non-performing exposure (any obligations that are 90 days overdue or that are impaired or in default), the amount of non-performing exposures reached 879 billion euros.1
According to data and analysis from the middle of the same year, revealed by Financial Times and made by a study from Fitch, European banks’ bad loans hit one trillion euros.2 Fitch surveyed one hundred banks due to be assessed by the test
mentioned above. According to that survey, 29 banks see the number of impaired loans rising by more than 20%, while one third of the banks see their bad loan volumes falling or staying the same. According to the article, the increases are partly due to the switch to a new more conservative loan classification and partly due to decline in the quality of their assets. Is it possible that socially responsible activities can partly explain the difference in these numbers? Is it possible that corporate social responsibility (CSR) can reduce the problems? Could it be a factor in the process of improving the quality of the banks’ portfolios?
What if the relations between banks and its customers can be improved by following socially responsible practices? Suppose that such actions can have a positive influence like better communication and more trust among financial institutions and their clients? What if this better influence can improve financial performance (FP). The topic of the current research paper is the impact of socially responsible behavior on the ocurrance of financial distress among the financial institutions.
Such an issue is important because, as mentioned above, there are problems in the financial industry in Europe. Social responsibility is still not considered as a
1 https://www.ecb.europa.eu/press/pr/date/2014/html/pr141026.en.html
3
necessity, as there are no widespread laws that define it as compulsory for the institutions. Nevertheless, it does have a great impact on the governance of the banks in many ways. It is considered as a way to improve relations with customers, suppliers, institutions and the government (Torres et al., 2012). It is also viewed as a motivation to focus on sustainable and long-term projects that can lead to better returns (Eisenbach, 2014). The development of the literature about the outcomes of CSR can point out if there is a strong and proven need of new regulations and laws that would determine a better definition, implementation and regulation of such socially responsible programs. The current paper tries to further develop the debate for social responsibility and its potential benefits or weaknesses.
There are many parties that could be interested in the results. On one hand, this could be the management of different financial institutions across Europe. On the other hand, policy-makers, the governments and decision makers from the European Union can use such investigation to continue the debate and maybe initiate a compulsory common minimum levels of social responsibility or a new enhanced definition for corporate social performance (CSP).
4
The objective, as mentioned above, is to explore whether further implementation and development of socially responsible strategies, policies or campaigns from the banks’ management could have a positive impact on the performance of the bank (viewed as a decrease in NPLs or loan losses). The main research question is “Is there a negative relationship between the four main scores for social responsibility from ASSET4 Datastream and the ratio of non-performing loans to total loans among the financial institutions in the European Union and the European Free Trade Association?”. The current research brings more light into the theory for corporate social performance by exploring whether social responsibility can improve the financial performance of banks in Europe by leading to a decrease in the levels of problematic loans.
The main contribution of this paper is the development of the debate on the advantages and disadvantages of CSR among financial institutions. As to my knowledge, this is the first paper to discuss the relation between corporate social responsibility and the ratio of non-performing loans to total loans among banks from the European Union and the European Free Trade Association.
The remaining of the paper will be constructed in the following way – the literature review will be presented in the next section. The third section will be dedicated to explaining the data and the methodology that are used in the research, followed by a presentation and a discussion of the results in Section 4. Finally the conclusion will be displayed in the last section, including also suggestions for future research.
2. Literature Review
2.1 Theory
5
is important for the sustainability, competitiveness, and innovation of EU enterprises and the EU economy. It brings benefits for risk management, cost savings, access to capital, customer relationships, and human resource management.”3 This definition includes the
environmental and social aspect of the CSR assessment by talking about sustainability and responsible impact on the society. It also cover the economic and corporate governance features by mentioning competitiveness and innovation. Finally, it is brought by the European Commission, which further makes it appropriate due to the fact that the current research is focusing on banks within Europe.
Attracting and retaining good and valuable borrowers is a way corporate social behavior could affect the levels of non-performing loans across banks. Any type of scandal caused by cases like financing of a project that leads to high pollution, the reports of too large amounts of bonuses for the senior management or mistakes of the banks’ employees due to low level of skills or training can redirect important customers towards the competition and instead attract lower quality borrowers. Such factors can eventually lead to an increase in the amount of non-performing loans. Even though corporate social responsibility may sound similar to corporate governance, there is a difference - CSR protects the interests of all the stakeholders of the company, while corporate governance protects mainly the interests of the shareholders.
Many papers explore the connection between corporate social behavior, on one hand, and banks’ performance or characteristics, on the other hand. In this group of papers can be mentioned the projects of Simpson and Kohers (2002), Scholtens (2009), Keffas and Olulu-Briggs (2011), Soana (2011), Carnevale, Mazzuca and Venturini (2012) Weber (2012), Wu and Shen (2013) and Eisenbach et al. (2014). Speciffically, Wu and Shen (2013) investigate the relationship between non-performing loans and social responsibility. The authors find a negative relation between non-performing loans and CSR, therefore supporting the hypotheses in the next section.
An important aspect of social responsibility regarding the operations of the banks appears to be the Equator Principles (EP). It is presented as “a risk-management framework, adopted by financial institutions, for determining, assessing and managing environmental and social risk in projects and is primarily intended to provide a minimum
6
standard for due diligence to support responsible risk decision-making.”4 The principles
apply to project-related financial products like Project Finance Advisory Services, Project Finance, Project-Related Corporate Loans and Bridge Loans. The rules and regulations of the EP are taken under consideration for loans larger than 100 million US dollars. In total there are ten principles that take part in the evaluation process that concern different aspects of the decision process. The principles affect the following characteristics – review and categorization, environmental and social assessment, applicable environmental and social standards, environmental and social management system and action plan, stakeholder engagement, grievance mechanism, independent review, covenants, independent monitoring and reporting and finally reporting and transperancy. An illustration of cases that are due to be assessed and implemented using the EP includes the development of a power plant, mine, oil and gas related projects, infrastructure or real estate development or any other type of project that can create major environmental or social risks.
Scholtens and Dam (2007) find a difference between banks that adopt the EP and those that do not - the bigger size is a characteristic of the adopters. Examples of benefits for the adoption of the principles are a reduction in risk and an improvement of reputation. Scholtens and Dam (2007) reveal that “most other financial and firm characteristics do not show significant differences.”
The effect of social responsibility may depend on the level of income. Eisenbach et al. (2014) examine the impact of a socially responsible code of conduct like the Equator Principles on the performance of banks. The results suggest positive abnormal returns (the difference between actual and expected returns) for institutions that have adopted the EP. When the focus is shifted towards developing countries, the results become the opposite – institutions that have adopted the EP underperform the rest of the companies in the project finance industry.
One big problem according to the related literature is the direction of the causality–which one (CSR or financial performance) should be the dependent variable and which one - the independent. Waddock and Graves (1997) first suggested the “virtuous circle” that describes the hesitation in choosing the main independent and the
7
dependent variable. According to Waddock and Graves (1997), there are two theories – Slack Resources and Good Management. The first theory suggests that companies with better performance would have a larger amount of resources available to invest for socially responsible causes leading to the statement that better financial performance leads to better social performance. On the other hand, the second theory proposes another view – it states that socially responsible behavior improves relations with the stakeholders (employees, government, community) leading to a better financial performance.
The management of financial institutions would be more interested to explore the Good management theory of Waddock and Graves (1997), because it explores whether corporate social responsibility leads to better firm performance. Therefore, for the practical implications of the results, the Good Management theory seems to be more important for the current research.
8
cleaning a park in the city or planting new trees or new gardens. The employees voluntarily take part in such initiatives. Finally, insider-initiated corporate philanthropy means sacrificing profits for socially responsible causes.
According to Benabou and Tirole (2010) there are two stages of development of the capitalism - “the invisible hand of the market” (managerial capitalism) and “the shareholder-value” capitalism. The “shareholder-value” capitalism has one main goal – maximize the profits of the shareholders. But there is also another type of capitalism called “Stakeholder Capitalism” where the main goal is changed. According to Freeman and Liedtka (1997) the stakeholder theory combines the interests of both shareholders and stakeholders. It is not only the return of the investors that matters, but also the interests of stakeholders like clients, suppliers and employees.
It can be noticed as if in the beginning of the development of the theory the influence of social responsibility is neglected and highly ignored. Among the first papers that mention corporate social behavior is the paper of Jensen and Meckling (1976). However, the authors state the belief that social responsibility does not affect the performance of the firm. There is also the paper of Milton Friedman (1970), supporting the same view – that the sole purpose of a company is to create value for the shareholders by creating and increasing profits. Preserving the interests of the stakeholders, according to early concepts of CSR, presented by these two papers, are not viewed as an important part of the operations and activities of the firm.
Later on, the paper of Carroll (1979) provides a summary of the different definitions of social responsibility and creates the first broad concept of CSR. Firstly, the paper separates the interpretation into four types – economic, legal, ethical and discretionary responsibilities. The model includes types of social issues like product safety, environment and discrimination. Finally, the definition includes the type of response from the organization – ranging from agreement or making actions with the purpose of correcting the issue to disagreement and rejection of the problem.
9
(2014) test whether geographical distance between companies within the United States has an impact on the CSR policy. According to Jiraporn et al. (2014), distance matters in creating a CSR program. In the context of the current research, the levels of social responsibility may differ across countries and continents with the increase of distance between the observations from the sample. The methodology of the current paper includes controls for countries due to the chance that the relationship may differ across our sample of countries within Europe.
The paper of Renneboog, Horst and Zhang (2008) observes an increasing importance of CSR in the corporate world, suggesting the need for further research in the field of corporate social behavior. The authors present a short review of laws refering to social responsibility in Europe throughout the past years, giving examples of UK (where pension funds and charities are regulated by law to share with the public their social, environmental and ethical investment policies), Germany (the Renewable Energy Act), the Netherlands (“Green Savings and Investment Plan”) and others like Belgium, Germany, Italy and Sweden.
The development of the literature is enhanced by the paper of Scholtens (2009), who suggests a framework for the assessment of corporate social responsibility that estimates the level of responsible performance among banks using publicly available information. The author creates a measure for social performance and analyzes the relationship between CSP and firm characteristics. The findings suggest that CSR has a positive relation to financial performance and bank size. Furthermore, the paper finds an increase in the number of banks reporting corporate social performance as well as an increase in the responsible financial products offered by the institutions.
10
2.2 CSR and cost of capital
It is difficult to say whether banks’ financial performance improves due to better corporate social responsibility. A borrower with better financial performance due to higher social responsibility could lead to better and more secure loan payments, leading to a decrease in non-performing loans and loan losses for the bank. Therefore, if a socially responsible bank lends money to socially responsible clients, the risk, in theory, could be lower. El Ghoul et al. and Goss and Roberts (2011) discuss the benefits of CSR and firms’ cost of capital. El Ghoul et al. (2011) examine the cost of equity capital for companies from the United States. The main findings of El Ghoul et al. (2011) suggest that ”improving responsible employee relations, environmental policies, and product strategies contributes substantially to reducing firms’ cost of equity” - firms with higher CSR performance exhibit significantly lower cost of capital. On the other hand, Goss and Roberts (2011) find slightly different results – American companies with socially responsible behavior below the average pay a larger cost for their loans, but there is no benefit for being among the top socially responsible companies. Possible explanations for these findings could be that if some banks lend to more socially responsible companies, this could decrease the amount of lenders for the irresponsible companies and could increase the price of the loan. On the other hand, lending to socially responsible companies can limit the investment opportunities for the bank and therefore decrease the profits. The paper of Nandy and Lodh (2012) suggests that environmental performance does matter in the process of getting a better corporate loan conditions.
2.3 The direction of the relationship
11
Table 1 Previous research papers
Authors Measure of CSR Measure of firm performance
Relationship Empirical results Waddock and
Graves (1997)
KLD ROE,ROA, Return on Sales
positive The results of the regression analysis suggest that the relationship is positive. Simpson and Kohers (2002) Community Reinvestment act ratings (CRA)
ROA; loan losses to total loans
positive The relationship between corporate social and financial performance in the banking industry is explored using OLS. Makni, Francoeur and Bellavance (2009) Canadian Social Investment Database
ROA, ROE, stock market returns
negative Using OLS, the test of CSR and FP among Canadian firms reveals a negative “Granger causal” relationship between the environmental dimension of CSP and the firm
performance measures. Jo and Harjoto
(2011)
KLD Tobin’s q (firm value)
positive The Heckman’s two stage model suggests a positive influence of CSR engagement on firm value.
Keffas and Olulu-Briggs (2011)
FTSE4GOOD 38 financial and economic ratios
insignificant The paper examines the differences in performance between CSR and Non-CSR banks using Data
Envelopment Analysis. Soana (2011) Ethibel ratings Return on average
equity, return on average assets, cost-to-income ratio, market to book ratio,
price-to-book ratio and price per earnings
ratio
insignificant The paper analyzes 21 international banks and 16 Italian banks. The Pearson correlations do not show statistically significant results. Carnevale, Mazzuca and Venturini (2012) Social report publications
Book value per share and earnings
per share
insignificant The sample of European banks is analyzed as panel data using cluster robust standard errors.
Wu and Shen (2013)
EIRIS ROA, ROE, NPL, net interest income
(NII) and non-interest income (NonII) positive with ROA, ROE, NII and NonII and negative with NPL
12
Tharyan and Whittaker (2014)
associated with stock market returns and CSR weaknesses are negatively associated. The authors use cluster robust standard errors. Kruger (2015) KLD Cumulative
abnormal return
negative The results of the event study show that negative CSR events has a strongly
negative impact on shareholder value and positive CSR events have a less strong negative impact on shareholder value. The table above summarizes the results of different studies exploring the relationship between social responsibility and firm performance.
There are a lot of previous papers describing the trends in the literature for social responsibility. For instance, Orlitzky, Schmidt and Rynes (2003) and Margolis, Elfenbein and Walsh (2009) find a positive relation between CSR and CFP. Contrary to the previous two projects, Kitzmueller and Shimshack (2012) point out that the literature does not fully support the hypothesis that corporate social performance reduces corporate costs, revealing once again the persistent need for further research on the topic.
3. Data and Methodology
3.1 Sample data
The definition of credit risk used in the current paper relies on the definition taken by the World Bank, which defines it both as “Risk of default: The risk that a counter party will be unable to perform as agreed” and as “Risk of loss: The risk that as a result of a counter party’s inability to perform as agreed, the lender suffers a loss.”5
There is a variation between different definitions for non-performing loans across institutions and across countries. According to Thomson Reuters Datastream, it is “the amount of loans that the bank foresees difficulty in collecting. It includes but is not restricted to: Non-accrual loans, Reduced rate loans, Renegotiated loans, Loans past
13
due 90 days or more. It excludes: Assets acquired in foreclosures, Repossessed personal property.”6 The definition of Bankscope merges impaired loans and
non-performing loans, mentioning that such definition varies across countries and even across banks.7 Though it could be noted that there is a difference between
non-performing loans and impaired loans.8 Therefore the data, including non-performing
loans, are taken from Datastream and not from Bankscope, where non-performing loans are defined as impaired loans.
There are two possible proxies for credit risk – loan losses and non-performing loans. Pesola (2011) mentions that loan losses are actual data on actual events, while non-performing loans and loan-loss provisions are just “accounting concepts”. Loan losses is an actual loss, while non-performing loans do not represent a clear amount of money lost, rather an increase in possibility that the amount of non-performing loans may become a loss. It was noticed, however, that the papers regarding bank financial distress are split between non-performing loans and loan losses. Pesola (2011), Simpson and Kohers (2002) and Jokivuolle, Pesola and Viren (2014) focus on loan losses, while Makri, Tsagkanos and Bellas (2014), Kauko (2012) and Wu and Shen (2013) target non-performing loans. The papers of Pesola (2011), Makri, Tsagkanos and Bellas (2014) and Jokivuolle, Pesola and Viren (2014) analyze the data as panel data. The papers of Beck, Jakubik and Piloiu (2013) and Jokivuolle, Pesola and Viren (2014) analyze non-performing loans (the first paper) and loan losses (the second paper) on country level, rather than firm-specific level. The current paper focuses on non-performing loans.
Another question concerns the choice of a CSR measure. Some papers, like the one from Scholtens (2009), construct their own measures for social responsibility, others take the result of different indices as proxy for it (see examples in Table 1). The four total scores for the four main categories from the ASSET4 Thomson Reuters index are used here, or, more specifically, corporate governance score, economic score, environmental
14
score and social score. The paper of Chatterji et al. (2015) compares the results of six rating agencies - KLD, Asset4, Calvert, FTSE4GOOD, DJSI and Innovest. Some of the main conclusions of the paper are that, firstly, those indicators have low correlations with each other and, secondly, they use different definitions and different aspects of social responsibility (for example, employment, employee diversity and employees’ health are considered differently among these agencies). Moreover, they measure the same aspects with different methodologies.
It appears that choosing the right rating agency is another important element in the analysis. The current paper finds the ASSET4 ESG framework as the most appropriate for research purposes.9 It includes over 750 individual data points merged
into 250 key performance indicators (KPIs). All these indicators consists of 18 categories, which are further grouped into four pillars that are integrated into a single overall score (Overall Performance). It can be assumed that the four total scores are detailed enough to capture all the aspects of corporate socially responsible behavior, considering all these detailed scores in the construction of the final four scores of ASSET4. The scores are created using a Z-score, which is further normalized to take values between 0% and 100%. The Z-score reveals the units of standard deviation of the specific company score from the mean value of all companies. The ASSET4 index is preferred over FTSE4GOOD index because FTSE4GOOD uses 300 indicators, combined into only three pillar scores leading to a final score ranging from 0 to 5.10 It is
easier to compare different companies on CSP with a score ranging from zero to one hundred than one ranging from zero to five. Another widely used CSR assessments are the MSCI ESG RATINGS. The company in question starts creating ratings for corporate social responsibility in 2010 – that is an obstacle as the current research uses data since 2005 (explained further below). KLD is a widely used measure in the related literature, but it is not available for all of the required years. KLD transforms into MSCI ESG, but there is a chance of a change in the evaluation system, which could result in data that
9http://extranet.datastream.com/data/ASSET4%20ESG/documents/ASSET4_ESG_Methdology_FAQ_061
2.pdf
15
cannot be compared between the periods of time. The paper of Cheng, Ioannou and Serafeim (2014) is an example from the literature than once more supports the decision for the use of ASSET4 scores. Furthermore, these scores are the most detailed source of information that is accessible for the purpose of this research and therefore are assumed as the most appropriate for the current paper.
These 18 categories of ASSET4 consist of measures that could be good signs and predictors of the performance of the bank. One of the main tasks of the banks is to attract and retain customers. Part of these duties include actions to engage with the best borrowers in terms of loyalty, skills and safe investments. The process of gaining the attention of such customers depends on the image of the bank and on all the news regarding topics like the vision and strategy of the bank, board structure and functions (corporate governance performance), resource and emission reduction, product innovation (environmental performance), loyalty of clients and shareholders (economic performance). Furthermore, employee’s training and development (social performance) takes an important part of the activities of the banks, because unskilled employees can eventually give a big amount of loans to low-skilled borrowers and entrepreneurs or companies with bad perspectives. All these factors take a role in the calculation of the scores and therefore are considered a good measure of social responsibility for the purpose of the current research.
16
The data are taken from Bankscope and Thomson Reuters Datastream. Bankscope is used to create the list of financial institutions across Europe. Bankscope incorporates financial information about more than 32 000 banks, including the top 8 000 European Banks. The first step of the construction of the sample is executed in Bankscope by limiting the financial institutions only to active and listed banks. The focus is on listed banks following the fact that ASSET4 relies mainly on public information. For that reason companies that are appropriate for the analysis must be publicly traded companies. The number of banks that meets the criteria is 518 banks. This figure is further reduced to names that have at least one year of records for the required information in Thomson Reuters Datastream. The last transformation results in a list of 402 companies. The sample is further reduced to include only countries part of the European Union (EU) or the European Free Trade Association. This action results in 324 banks. Finally, the sample is limited to companies with data for corporate governance score, economic score, environmental score, social score and loan losses or non-performing loans. The last correction leads the number of financial institutions in our example to 51 for the test using net loan losses and 71 for non-performing loans. In total, there are 72 banks that will take part in the analysis. The list of the banks is available in Appendix A.
Data for the country-specific variables for countries except Switzerland and Norway are taken from Eurostat, annual real GDP growth rate for Switzerland and Norway are taken from Eurostat, unemployment and general government gross debt as percentage of gross domestic product (GDP) for Switzerland are taken from the World Bank and finally general government gross debt as percentage of gross domestic product (GDP) for Norway is also taken from the World Bank. For the case of Switzerland, unemployment is the national estimate for total unemployment as percentage of total labor force and general government gross debt as percentage of GDP for Switzerland and Norway is total central government debt as percentage of GDP.
17
3.2 Methodology
3.2.1 Main variables
Four different proxies for CSR will be examined from the ASSET4 universe – corporate governance score, economic score, social score and environmental score. In the main model, financial performance will be presented by the ratio of non-performing loans to total loans (NPL).
3.2.2 Control variables
Many researchers control for firm characteristics like size, risk (leverage) and performance (profitability). Following the example of Wu and Shen (2013), Gregory, Tharyan and Whittaker (2014) and Jiraporn et al. (2014), the logarithm of total assets will be used as a control for size. Small and new banks on the market would try to gain a better market position. These institutions will apply a riskier strategy to get a larger market share and larger profits. On the other hand, huge and established financial institutions would already have a big list of clients and a good market position and would be able to put more strict rules on existing and new clients in lending opportunities. Therefore, a negative relationship is expected between size and NPL.
A second common factor is the company risk or leverage. Risk will be presented by the measure of financial leverage – the ratio of debt to total assets (Jiraporn et al. 2014). A large ratio could mean that banks have enough deposits and have to expand by finding new borrowers. A big urgency to find borrowers could lead to a decrease in the quality of the requirements for the loans and for the borrower. Thus, risk would have a positive relation with NPL.
18
from the first and the second paper both use ROA and ROE, but the authors from the last paper conclude that ROA does not have significant relationship with non-performing loans, while ROE does. Therefore ROE is believed to have better explanatory power than ROA and it will represent the proxy for profitability.
There is also a control variable presented by the ratio of loans to deposits (Wu and Shen 2013). A larger ratio of loans to deposits could suggest that the company would not take more highly profitable opportunities with high risk, but would rather focus on gaining more depositors. Therefore, the amount of risky new borrowers would decline, leading to a possible decline in NPL. The relation between the ratio of loans to deposits and NPL is expected to be negative.
19
3.2.3 Hypotheses
Following Wu and Shen (2013), the following hypotheses will be tested:
H1: There is a significantly negative relationship between the economic score and the ratio of non-performing loans to total loans.
H2: There is a significantly negative relationship between the social score and the ratio of non-performing loans to total loans.
H3: There is a significantly negative relationship between the environmental score and the ratio of non-performing loans to total loans.
H4: There is a significantly negative relationship between the corporate governance score and the ratio of non-performing loans to total loans.
3.2.4. Regression design
The current study will analyze the collected data as panel data. The panel data set is a combination between time-series data and cross-sectional data. Time-series data are data about one object (for example country or company) collected for many points of time (days, months, years). Cross-sectional data are data for many objects but collected for one specific point of time. The data represent information about different companies (cross-sections) collected for ten years – the years between 2005 and 2014.
Many of the companies in our data set do not have data for every year between 2005 and 2014. This fact could be easily explained by the fact that the names of the companies taking part in ASSET4 Datastream is being updated every year and the names of the companies change, some of the companies drop out to leave a place for others. Therefore, our set is unbalanced panel data set. The data are analyzed using panel least squares.
The main equation of the model will take the following form:
20
where t reflects the year, i - the company, j – the country and k represents the CSR proxy, CSR serves as the proxy for social responsibility, NPL is the ratio of non-performing loans to total loans, SIZE is the logarithm of total assets, RISK is the ratio of total debt to total assets, ROE is return on equity, LoanDP is the ratio of net loans to total deposits, GDP is annual real GDP growth rate, UNEMP is the annual average unemployment rate in percentage, DEBT is general government gross debt as percentage of gross domestic product (GDP), Z denotes the time fixed effects, Y – country dummies, X controls for possible nonlinear effects and υ is the error term.
In each one of the tests, the variable CSR is presented separately by each of the four proxies for social responsibility, namely economic score, social score, environmental score and corporate governance score.
There are a couple of assumptions of the least squares to be considered. There could be patterns in the residuals. The existence of autocorrelation could result in incorrect standard errors in the least squares test. Homoskedasticity is another assumption that is made for the purpose of having a least squares estimation. The presence of heteroskedasticity can lead to wrong calculations for the standard errors. An example is used from Peterson (2009) to deal with possible heteroskedasticity and possible autocorrelation between the residuals. The author suggests using cross-section fixed effects and clustering by period or period fixed effects and clustering by cross-sections. He also states that in order for the regression to be unbiased, a sufficient number of clusters needs to be present. Having only ten years in our sample, it is obvious that we cannot cluster by period. Kezdi (2003) finds that to have unbiased standard errors when using fixed effects one needs to have at least ten years of data and 50 cross-sections. The sample of the current research consists of ten years and 71 cross-sections, therefore our results should be unbiased. To achieve unbiased standard errors, the tests will include time fixed effects and clustering by banks.
21
included to account for possible heterogeneity. Different banks could have different internal rules, regulations and procedures that could influence the results. Clustering by banks should be able to capture such variation. There are also years of economic growth and years of economic decline, when the prevailing economic trend determines the levels of difficulties among the individuals and the business that leads to distress among the banks. The controls for time fixed effects are used to take care of such economic cycles. Controlling for time should be able to take in mind possible unobserved factors that remain the same during a year but change across time. There is not a separate dummy variable for the period of the financial crisis from 2007-2008, as according to the data such a variable is not necessary. All of the cases with a ratio of non-performing loans to total loans above ten comes from the years between 2009 and 2014, rather than from the years of the crisis. The tests include also country dummies, following Campbell (2007) and Jiraporn et al. (2014), to capture possible exogenous factors. Each dummy equals 1 if the company is based in the corresponding country and 0 otherwise.
Following Goss and Roberts (2011) and Barnett and Salomon (2012), dummy variables are included to account for possible nonlinear relationship between the variables. There are dummy variables for levels of CSP between 0 and 25, between 25 and 50 and between 75 and 100. If the direction of the relationship changes with the increase or decrease in the values of the CSR scores, such variables should be able to capture the difference. To deal with multicollinearity, a dummy variable for CSR score between 50 and 75 is excluded from the tests, following the fact that the current research is more interested whether extremely high or extremely low scores lead to really high or really low amount of non-performing loans.
3.3. Robustness checks
22
period of time to actually reach to the public. Secondly, following the examples of Benabou and Tirole (2010) and Waddock and Graves (1997), the regressions will be tested for causality by changing the places of the dependent variable and the main independent variable in each regression. Using the example of Cheng, Ioannou and Serafeim (2014), the tests are executed again using the average of the four scores for social responsibility. Finally, a test using loan losses as dependent variable is performed, following Pesola (2011), Simpson and Kohers (2002) and Jokivuolle, Pesola and Viren (2014), to check if the results remain the same using a different proxy for financial performance.
3.4 Basic statistics
The descriptive statistics and the correlation matrix of the data for the model using non-performing loans as percentage of total loans can be seen in Table 2 and Table 3.
Table 2 Descriptive Statistics
Mean Median Maximum Minimum Std. Dev.
Economic score 61.67 71.44 99.06 1.76 30.06
Social score 70.56 85.03 98.83 4.90 28.67
Corporate governance score 53.09 58.16 95.67 1.93 28.99
Environmental score 66.74 84.97 96.76 8.74 31.46
Government debt 74.90 71.70 177.00 23.60 32.83
GDP 0.70 1.30 6.30 -9.10 2.83
Loans to deposits 1.66 1.40 79.65 0.43 3.37
Net loan losses to net loans 0.00 0.00 0.07 -0.01 0.01 Non-performing loans to total loans 5.21 2.94 119.52 0.00 8.59 Total debt to total assets 33.48 34.79 70.18 3.29 13.90
ROE -7.32 9.48 79.14 -4298.47 198.65
Log of total assets 8.17 8.12 9.48 6.32 0.65
Unemployment 9.03 8.00 27.50 2.50 4.76
23
Some interesting numbers can be seen in the descriptive statistics in Table 2. There seem to be a couple of extreme values among the general government gross debt as percentage of GDP, the ratio of non-performing loans to total loans and the return on equity. Both the extreme values for the ratios of return on equity and government gross debt take place in Greece. Throughout the years starting from 2008, a declining negative annual real GDP growth with increasing government debt and increasing unemployment decrease the ROE from 1015% to extreme negative values ranging from 163.66% to -4298.47%. The biggest levels of the ratio of non-performing loans are mainly in countries like Greece, Ireland and Italy, following the Ireland banking crisis and the Greece government debt crisis.
The correlation matrix in Table 3 shows a high correlation between the Social score and the Environmental score, suggesting that maybe companies follow values from both types in a similar manner. All the CSR scores except the economic score have a bit high correlation with the size presented by the logarithm of total assets. It seems that size does matter for the socially responsible levels among the banks. Fortunately, it could be noticed that multicollinearity is not a problem in the final data set.
24
Table 3 Correlation matrix Economic score Social score Env. score Corp. gov. score Gov. debt GDP Loans to deposits Net loan losses to net loans Non-perf. loans to total loans Total debt ROE The log of total assets Unem- ployment Economic score 1.00 Social score 0.66 1.00 Environmental score 0.43 0.58 1.00
Corporate Governance score 0.42 0.78 0.57 1.00
Government debt -0.21 -0.03 -0.10 0.14 1.00
GDP -0.02 -0.09 -0.06 -0.17 -0.45 1.00
Loans to deposits -0.05 0.05 0.02 0.07 -0.12 -0.01 1.00
Net loan losses to net loans -0.16 -0.21 -0.06 -0.07 0.22 -0.05 -0.15 1.00
Non-performing loans to total loans -0.12 -0.11 -0.03 0.04 0.32 -0.18 -0.16 0.28 1.00
Total debt to total assets -0.07 0.12 0.06 0.13 -0.12 -0.09 0.73 -0.11 -0.16 1.00
ROE 0.07 -0.01 0.13 -0.05 -0.30 0.28 -0.08 -0.10 -0.17 -0.13 1.00
Log of total assets 0.28 0.58 0.58 0.67 -0.05 0.00 0.03 -0.19 -0.07 -0.04 0.06 1.00
Unemployment 0.02 0.07 0.11 0.13 0.47 -0.42 -0.03 0.05 0.24 0.11 -0.18 -0.14 1.00
Table 4 Distribution of observations and non-performing loans by country Observations per country
Average ratio of Non-performing loans per country Austria 19 5.42 Belgium 20 2.14 Czech Republic 3 0.58 Denmark 21 2.53 France 34 3.36 Germany 17 1.41 Greece 26 7.77 Hungary 5 8.91 Ireland 19 12.31 Italy 83 7.61 Netherlands 2 61.73 Norway 9 1.02 Poland 36 6.25 Portugal 25 3.11 Spain 52 4.54 Sweden 36 0.78 Switzerland 25 0.91 United Kingdom 49 2.6
The table above reveals the total number of observations and the average ratio of non-performing loans to total loans by country.
Fig. 1: Distribution of banks by country
27
Table 5 Distribution of observations per year Year Number of observations
2005 37 2006 42 2007 47 2008 53 2009 56 2010 59 2011 62 2012 55 2013 43 2014 27
The table presents the total number of observations by year.
The analysis of the results will continue in the next section.
4. Results
A panel unit root test rejects the null of no cointegration at the 1% significance level for the variables. A set of variables is cointegrated when a linear combination of the variables is also stationary. Stationarity could be a problem when two variables have significant influence on each other, but such a relation is due to random events and the variables are in fact unrelated to each other. The t-statistic in each of the four tests is less than -2.4 and the corresponding p-value is 0.
Redundant fixed effects test is implemented to check whether there is truly a need to control for fixed effects. The null of no fixed effects is rejected for cross-section (firm) fixed effects and both cross-section and period fixed effects at the 1% significance level with p-value of 0. The p-value for the F-test of period fixed effects only is above 0.55.
28
interval for cross-section random effects and with probability of 0.0061 at 1% significance interval for period random effects. Therefore, the equations will be estimated using fixed effects.
The main test with a dependent variable presented by the ratio of non-performing loans to total loans including country controls, control for nonlinearity and clustering by firm is presented below in Table 6. The number of observations for each of these four regressions is 481, there are 71 cross-sections and the time period is ten years. The problems of heteroscedasticity and autocorrelation are removed using cross-sectional clustering and period fixed effects.
It can be noticed that the scores for social responsibility do not receive a statistically significant estimate. All the country dummies are significant at 1% significance level, but they are not reported. The controls for possible nonlinearity in the relationship do not show significant results, as none of them is significant at the 10% significant level. The government gross debt as percentage of GDP and return on equity do show significant results. These results confirm the expectation for a positive relation between the government debt and non-performing loans. The outcome also suggests that return on equity, with a coefficient of 0, has no influence on the levels of problematic loans.
The robustness check using the lagged value of the CSR receives similar results, as can be seen in Table 7. The test using the CSR index as dependent and the NPL among the independent variables is presented in Table 8. Again, the results are insignificant and the main variable is not statistically significant.
29
our main independent variable and the dependent variable. The R2 is not more than 0.7
in any of the regressions. Finally, the tests are repeated using the ratio between net loan losses and net loans as dependent variable. The results remain the same – no significant relation. The last test is not reported as most of the coefficient estimates equal to 0. Therefore, it could be concluded that there is not a clear relationship between CSR and CFP among banks.
Unfortunately, none of the social responsibility scores in the tests are able to explain the variation in the levels of non-performing loans. Having low coefficients and low statistical significance, the results for the CSR proxies reveal that there is not a solid significant relationship.
The results could imply that non-performing loans depend on country-specific factors like levels of education, levels of income for the individuals and the families, levels of development of the laws and regulations. Such factors are not observed by the current research, but could actually cause the numbers of non-performing loans to differ across countries and across banks.
Surprisingly, there are only 13 unique observations among 383 with a ratio of net loan losses to net loans bigger than 0.01. This fact could be also another reason why a significant relation is not found here. It could partly be explained by the possibility for a bank to sell a bad loan to other financial institutions before the realization of the loss and therefore to limit the risk and the loss. Another possible factor here can be the use of proper mortgages that reduce or entirely remove the loss.
30
Table 6 Panel Least Squares
1 2 3 4 Constant 49.07*** 54.28*** 54.48*** 51.29*** (4.63) (6.18) (4.79) (4.81) Government debt 0.14*** 0.15*** 0.14*** 0.14*** (0.05) (0.06) (0.06) (0.06) GDP -0.01 0.06 0.03 0.03 (0.18) (0.19) (0.18) (0.18) Loans to deposits -0.04 -0.05* -0.04 -0.04 (0.03) (0.02) (0.03) (0.03)
Total debt to total assets -0.04 -0.04 -0.04 -0.04
(0.03) (0.03) (0.03) (0.03)
ROE 0*** 0*** 0*** 0***
(0) (0) (0) (0)
The log of total assets 0.29 -0.57 -0.5 0.11
(0.42) (0.54) (0.72) (0.51) Unemployment -0.01 0.01 0 0.01 (0.18) (0.18) (0.18) (0.17) Economic score 0.01 (0.02) Social Score 0.04 (0.05) Environmental score 0.03 (0.04)
Corporate governance score 0.01
(0.03) LOW-CSR dummy 1.83 0.94 0.75 0.35 (1.17) (2.47) (2.01) (1.46) SMALL-CSR dummy 1.10 -0.65 1.16 0.39 (0.85) (1.53) (1.25) (0.98) HIGH-CSR dummy 0.33 -0.74 -0.19 -0.77 (0.57) (1.68) (1.38) (0.73) R2 0.57 0.57 0.57 0.57 Observations 481 481 481 481
Time fixed effects YES YES YES YES
Country controls YES YES YES YES
The table above presents the results from the regression using period fixed effects, country dummies, controls for possible nonlinearity and cross-sectional clustering. Dependent variable – the ratio of non-performing loans to total loans. The numbers in brackets represent the standard error of the estimates. * denotes significance at the 10% significance level
31
Table 7 Panel least squares using the lag of the CSR proxy
1 2 3 4 Constant -4.84 -1.49 0.43 -1.96 (5.78) (7.39) (5.81) (5.74) Government debt 0.13** 0.15** 0.14** 0.14** (0.06) (0.06) (0.06) (0.06) GDP 0.01 0.12 0.1 0.08 (0.19) (0.21) (0.21) (0.2) Loans to deposits -1.25*** -1.08*** -1.12*** -1.19*** (0.32) (0.3) (0.3) (0.27)
Total debt to total assets 0.01 0.01 0.01 0.01
(0.04) (0.03) (0.04) (0.03)
ROE 0*** 0*** 0*** 0***
(0) (0) (0) (0)
The log of total assets 0.29 -0.61 -0.76 -0.06
(0.43) (0.61) (0.81) (0.5)
Unemployment 0 0.03 0.03 0.03
(0.19) (0.18) (0.18) (0.18)
Lagged economic score -0.03 (0.02)
Lagged social score 0.03
(0.04)
Lagged environmental score 0.03
(0.04)
Lagged corporate governance score -0.02
(0.03) LOW-LAG-CSR dummy 0.42 1.06 0.78 -1.21 (1.40) (2.11) (1.53) (1.25) SMALL-LAG-CSR dummy 0.48 -0.91 0.43 -0.51 (1.00) (1.27) (1.07) (1.11) HIGH-LAG-CSR dummy 1.03 -0.76 -0.34 0.38 (0.72) (1.44) (1.48) (0.78) R2 0.66 0.66 0.66 0.66 Observations 410 410 410 410
Time fixed effects Yes Yes Yes Yes
Country controls Yes Yes Yes Yes
The table above presents the results from the regression using period fixed effects, country dummies, controls for possible nonlinearity and cross-sectional clustering. Dependent variable – the ratio of non-performing loans to total loans. The numbers in brackets represent the standard error of the estimates. ** denotes significance at the 5% significance level
32
Table 8 Panel least squares, testing for reverse causality
1 2 3 4 Constant 95.87*** 152.7*** 217.22*** 179.08*** (34.15) (33.96) (31) (47.28) Government debt -0.35*** -0.3** -0.11 -0.18 (0.14) (0.13) (0.11) (0.11) GDP -2.14*** -1.24** -1.58*** -0.24* (0.83) (0.58) (0.56) (0.45) Loans to deposits -0.22 -0.06 0.15 0.46*** (0.17) (0.27) (0.19) (0.18)
Total debt to total assets -0.22 0.07 0.08 0.01
(0.19) (0.22) (0.2) (0.18)
ROE 0 0 0 0.01
(0.01) (0) (0) (0.01)
The log of total assets 21.34*** 28.41*** 34.58*** 27.12***
(3.61) (3.88) (3.64) (5.76) Unemployment -0.67 -0.48 -0.57 0.2 (0.68) (0.52) (0.58) (0.54) NPL1 -0.23 (0.19) NPL2 0.32 (0.28) NPL3 0.24 (0.27) NPL4 -0.06 (0.14) R2 0.52 0.6 0.64 0.61 Observations 481 481 481 481
Time fixed effects Yes Yes Yes Yes
Country controls Yes Yes Yes Yes
The table above presents the results from the regression testing for causality in the relationship, including period fixed effects, country dummies and cross-section clustering. Dependent variables – the scores for social responsibility – economic score, social score, environmental score and corporate governance score. The numbers in brackets represent the standard error of the estimates.
33
Table 9 Panel least squares, using the mean of the CSR scores 1 Constant 1.18 (5.53) Government debt 0.14*** (0.06) GDP 0.02 (0.19) Loans to deposits -0.05** (0.03) Total debt to total assets -0.04
(0.03)
ROE 0***
(0) The log of total assets -0.25
(0.57) Unemployment 0 (0.18) CSRMEAN -0.04 (0.03) LOW-CSR-MEAN dummy -2.86 (1.93) SMALL-CSR-MEAN dummy -1.03 (1.34) HIGH-CSR-MEAN dummy 1.11 (-.69) R2 0.57 Observations 481
Time fixed effects Yes
Country controls Yes
The table above presents the results from the regression using period fixed effects, country dummies, controls for possible nonlinearity and cross-section clustering. CSRMEAN equals the average of the four proxies for social responsibility. Dependent variable – the ratio of non-performing loans to total loans. The numbers in brackets represent the standard error of the estimates.
34
Table 10 Panel least squares, controlling for outliers
1 2 3 4 Constant -6.68 -2.88 -1.29 -3.7 (5.13) (6.75) (4.92) (5.3) Government debt 0.12** 0.14** 0.13** 0.13** (0.06) (0.06) (0.06) (0.06) GDP 0.02 0.1 0.07 0.05 (0.17) (0.18) (0.18) (0.18) Loans to deposits -0.05 -0.05** -0.04 -0.04 (0.03) (0.03) (0.03) (0.03)
Total debt to total assets -0.04 -0.03 -0.03 -0.04
(0.03) (0.03) (0.03) (0.03)
ROE 0*** 0*** 0*** 0***
(0) (0) (0) (0)
The log of total assets 0.21 -0.64 -0.61 0.06
(0.42) (0.55) (0.72) (0.51) Unemployment 0.08 0.1 0.09 0.1 (0.16) (0.16) (0.16) (0.16) CSR1 0.01 (0.02) CSR2 0.06 (0.04) CSR3 0.04 (0.04) CSR4 0 (0.02) LOW-CSR dummy 2.26* 2.16 1.29 -0.39 (1.20) (2.12) (1.86) (1.19) SMALL-CSR dummy 1.37 -0.13 1.23 -0.04 (0.87) (1.37) (1.18) (0.87) HIGH-CSR dummy 0.30 -1.47 -0.52 -0.42 (0.57) (1.46) (1.37) (0.61) R2 0.65 0.65 0.64 0.64 Observations 462 462 462 462
Time fixed effects Yes Yes Yes Yes
Country controls Yes Yes Yes Yes
The table above presents the results from the regression using period fixed effects, country dummies, controls for possible nonlinearity and cross-section clustering. Dependent variable – the ratio of non-performing loans to total loans. The numbers in brackets represent the standard error of the estimates. *denotes significance at the 10% significance level
35
The results of this research are consistent with the results of Carnevale, Mazzuca and Venturini (2011), Keffas and Olulu-Briggs (2011) and Soana (2011) who also find insignificant results when estimating bank performance and CSR. This research does not manage to confirm the results of Wu and Shen (2013) who find a negative relationship between corporate social responsibility and non-performing loans.
5. Conclusion
This paper tries to explore the possible relationship between the ratio of non-performing loans to total loans, and the levels of social responsibility, presented by the four scores of ASSET4 Datastream. The results suggest that there is not a significant relationship between the main variables. The tests are repeated using the lagged values of the CSR proxy, following the chance that social performance requires time to result in real changes. Changing the places of the main independent variable and the dependent variable tests a possible reverse causality. Using the ratio of net loan losses to net loans as a dependent variable receives similar results as the other tests. The test with the average estimate of the four scores and the test with removed outliers do not change the results either. The conclusion of these tests is that there is not a significant relationship between financial distress and corporate social performance among banks in the European Union and the European Free Trade Association.
A big limitation of this paper remains the possibility of exogenous factors. There are many factors that could affect non-performing loans, but many of them are also not observable, like borrowers’ characteristics (income or professional experience), experience of the management and the employees of the bank and many others. Furthermore, our sample is determined in Bankscope and the data is taken from Datastream. Other sources like Bloomberg, for example, may have financial data for a larger number of banks from the ASSET4 universe.
36
results and could capture a better relationship. Further analysis could be constructed to analyze whether the time of the sample data also matters – for example is there a difference in the relation exploring corporate data before the year 2000 and in the last 15 years.
37
References:
Barnett, M., Salomon, R., 2012. Does it pay to be really good? Addressing the shape of the relationship between social and financial performance. Strategic Management Journal 33, 1304-1320.
Beck, R., Jakubik, P., Piloiu, A., 2013. Non-performing loans: What matters in addition to the economic cycle?
Bénabou, R., Tirole, J., 2010. Individual and corporate social responsibility. Economica 77, 1-19.
Campbell, J., 2007. Why would corporations behave in socially responsible ways? An institutional theory of corporate social responsibility. Academy of Management Review 32, 946-967.
Carnevale, C., Mazzuca, M., Venturini, S., 2012. Corporate social reporting in European banks: the effects on a firm's market value. Corporate Social Responsibility and Environmental Management 19, 159-177.
Carroll, A., 1979. A three-dimensional conceptual model of corporate performance. Academy of Management Review 4, 497-505.
Chatterji, A., Durand, R., Levine, D., Touboul, S., 2015. Do ratings of firms converge? Implications for managers, investors and strategy researchers. Strategic Management Journal.
Cheng, B., Ioannou, I., Serafeim, G., (2014). Corporate social responsibility and access to finance. Strategic Management Journal 35, 1-23.
Eisenbach, S., Schiereck, D., Trillig, J., Flotow, P., 2014. Sustainable project finance, the adoption of the Equator Principles and shareholder value effects. Business Strategy and the Environment 23, 375-394.
El Ghoul, S., Guedhami, O., Kwok, C., Mishra, D. R., 2011. Does corporate social responsibility affect the cost of capital? Journal of Banking and Finance 35, 2388-2406.
Freeman, E., Liedtka, J., 1997. Stakeholder capitalism and the value chain. European Management Journal 15, 286-296.
38
Gregory, A., Tharyan, R., Whittaker, J., 2014. Corporate social responsibility and firm value: disaggregating the effects on cash flow, risk and growth. Journal of Business Ethics 124, 633-657.
Jensen, M., Meckling, W., 1976. Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics 3, 305-360.
Jiraporn, P., Jiraporn, N., Boeprasert, A., Chang, K., 2014. Does corporate social responsibility (CSR) improve credit ratings? Evidence from geographic identification. Financial Management 43, 505-531.
Jo, H., Harjoto, M., 2011. Corporate governance and firm value: The impact of corporate social responsibility. Journal of Business Ethics 103, 351-383.
Jokivuolle, E., Pesola, J., Viren, M., 2014. What drives loan losses in Europe? Bank of Finland Research Discussion Paper 6.
Kauko, K., 2012. External deficits and non-performing loans in the recent financial crisis. Economics Letters 115, 196-199.
Keffas, G., Olulu-Briggs, O., 2011. Corporate social responsibility: how does it affect the financial performance of banks? Empirical evidence from US, UK and Japan. Journal of Management and Corporate Governance 3, 8-26.
Kezdi, G., 2003. Robust standard error estimation in fixed-effects panel models. Budapest University of Economics.
Kitzmueller, M., Shimshack, J., 2012. Economic perspectives on corporate social responsibility. Journal of Economic Literature, 51-84.
Krüger, P., 2015. Corporate goodness and shareholder wealth. Journal of Financial Economics 115, 304-329.
Makni, R., Francoeur, C., Bellavance, F., 2009. Causality between corporate social performance and financial performance: Evidence from Canadian firms. Journal of Business Ethics 89, 409-422.
39
Margolis, J., Elfenbein, H., Walsh, J., 2009. Does it pay to be good... and does it matter? A meta-analysis of the relationship between corporate social and financial performance.
Matthews, J., Rusinko, C., 2010. Sustainability Disclosure: Increasingly Important for Banks and Commercial Lenders. Commercial Lending Review 25, 13-19.
Milton, F., 1970. The social responsibility of business is to increase its profits. New York Times Magazine 32, 122-126.
Nandy, M., Lodh, S., 2012. Do banks value the eco-friendliness of firms in their corporate lending decision? Some empirical evidence. International Review of Financial Analysis 25, 83-93.
Orlitzky, M., Schmidt, F., Rynes, S., 2003. Corporate social and financial performance: A meta-analysis. Organization Studies 24, 403-441.
Pesola, J., 2011. Joint effect of financial fragility and macroeconomic shocks on bank loan losses: Evidence from Europe. Journal of Banking and Finance 35, 3134-3144.
Petersen, M., 2009. Estimating standard errors in finance panel data sets: Comparing approaches. Review of Financial Studies 22, 435-480.
Renneboog, L., Ter Horst, J., Zhang, C., 2008. Socially responsible investments: Institutional aspects, performance, and investor behavior. Journal of Banking and Finance 32, 1723-1742.
Scholtens, B., 2009. Corporate social responsibility in the international banking industry. Journal of Business Ethics 86, 159-175.
Scholtens, B., Dam, L., 2007. Banking on the Equator. Are banks that adopted the Equator Principles different from non-adopters? World Development 35, 1307-1328.
Servaes, H., Tamayo, A., 2013. The impact of corporate social responsibility on firm value: The role of customer awareness. Management Science 59, 1045-1061.
Simpson, W., Kohers, T., 2002. The link between corporate social and financial performance: evidence from the banking industry. Journal of Business Ethics 35, 97-109.
40
Stellner, C., Klein, C., Zwergel, B., 2015. Corporate social responsibility and Eurozone corporate bonds: The moderating role of country sustainability. Journal of Banking and Finance, 59, 538-549.
Torres, A., Bijmolt, T., Tribo, J., Verhoef, P., 2012. Generating global brand equity through corporate social responsibility to key stakeholders. International Journal of Research in Marketing 29, 13-24.
Waddock, S., Graves, S., 1997. The corporate social performance-financial performance link. Strategic Management Journal 18, 303-319.
Weber, O., 2012. Environmental credit risk management in banks and financial service institutions. Business Strategy and the Environment 21, 248-263.
41
APPENDIX
Appendix A. List of financial institutions, included in the test
Name Country
Aareal Bank AG Germany
Ageas (EX-FORTIS) NV Belgium
Alior Bank SA Poland
Allied Irish Banks PLC. Ireland
Alpha Bank SA Greece
Banca Carige Italy
Banca Popolare Di Milano Italy
Banca PPO.DI Sondrio Italy
Banca PPO.Emilia Romagna Italy
Banco BPI SA Portugal
Banco de Sabadell SA Spain
Banco Espirito Santo SA Portugal
Banco Popolare Italy
Banco Popolar Espanol SA Spain
Banco Santander SA Spain
Bank Millennium SA Poland
Bank of Ireland Ireland
Bank of Piraeus SA Greece
Bank PKA.Kasa Opieki SA Poland
Bank Zachodni WBK SA Poland
Bankia SA Spain
Bankinter SA Spain
Banque Canton.VE. Switzerland
Barclays PLC. United Kingdom
BBV.Argentaria SA Spain
42
BNP Paribas France
Caixabank SA Spain
Close Brothers GP.PLC. United Kingdom
Commerzbank AG Germany
Credit Agricole SA France
Credit Suisse Group AG Switzerland
Danske Bank A/S Denmark
Deutsche Bank AG Germany
Deutsche Postbank AG Germany
Dexia Belgium
DNB ASA Norway
EFG International AG Switzerland
Erste Group Bank AG Austria
Eurobank Ergasias SA Greece
Gam Holding AG Switzerland
HSBC Holdings PLC. United Kingdom
ING Bank Slaski SA Poland
ING Groep NV Netherlands
Intesa Sanpaolo Italy
Investec PLC. United Kingdom
Julius Bar Gruppe AG Switzerland
Jyske Bank AS Denmark
KBC Groep NV Belgium
Komercni Banka AS Czech Republic
Lloyds Banking GP.PLC. United Kingdom
Mbank SA Poland
Mediobanca BC.FIN SA Italy
National Bank of Greece SA Greece
Natixis France
Nordea Bank AB Sweden
43
Permanent TSB GHG.PLC. Ireland
PKO Bank SA Poland
Raiffeisen Bank INTL.AG Austria
Royal BK.OF SCTL.GP.PLC. United Kingdom
SEB 'A' SA Sweden
Societe Generale France
STD.Chartered PLC. United Kingdom
Svenska HANDBKN.AB Sweden
Swedbank AB Sweden
Sydbank A/S Denmark
UniCredit Italy
Unione Di Banche Italian Italy
Valiant Holding AG Switzerland
44
Appendix B. Data Description
Indicator Description
Corporate Score The corporate governance pillar measures a company's systems and processes, which ensure that its board members and executives act in the best interests of its long term shareholders. It reflects a company's capacity, through its use of best management practices, to direct and control its rights and responsibilities through the creation of incentives, as well as checks and balances in order to generate long term shareholder value.
Economic Score The economic pillar measures a company's capacity to generate sustainable growth and a high return on investment through the efficient use of all its resources. It is reflection of a company's overall financial health and its ability to generate long term shareholder value through its use of best management practices. Environmental Score The environmental pillar measures a company's impact on living
and non-living natural systems, including the air, land and water, as well as complete ecosystems. It reflects how well a company uses best management practices to avoid environmental risks and capitalize on environmental opportunities in order to generate long term shareholder value.
Social Score The social pillar measures a company's capacity to generate trust and loyalty with its workforce, customers and society, through its use of best management practices. It is a reflection of the company's reputation and the health of its license to operate, which are key factors in determining its ability to generate long term shareholder value.
Non-performing Loans
The amount of loans that the bank foresees difficulty in collecting. It includes but is not restricted to:
45
Renegotiated loans,
Loans past due 90 days or more. It excludes:
Assets acquired in foreclosures, Repossessed personal property. Non-performing
loans as percentage of total loans
Non-Performing Loans / Loans-Total * 100
Total assets The sum of cash & due from banks, total investments, net loans, customer liability on acceptances (if included in total assets), investment in unconsolidated subsidiaries, real estate assets, net property, plant and equipment and other assets.
Total debt as percentage of total assets
(Short Term Debt & Current Portion of Long Term Debt + Long Term Debt) / (Total Assets - Customer Liabilities on Acceptances) * 100
Return on Equity (Net Income – Bottom Line - Preferred Dividend Requirement) / Average of Last Year's and Current Year’s Common Equity * 100 Loans - Net The total amount of money loaned to customers after deducting
reserves for loan losses.
Net Loan Losses The actual amount the bank lost during the year from uncollectable loans. It is calculated by subtracting recoveries from the amount of loans charged off.
Deposits - Total The value of money held by the bank or financial company on behalf of its customers. The item includes demand, savings, money market and certificates of deposit along with foreign office and deposit accounts. Excluded are securities sold under repurchase agreement.
Annual real GDP growth rate
