Dissertation draft [603201]

Dissertation draft

Master thesis
CSR and Financial Performance of European Banks

Student: [anonimizat]: Conf.univ.dr. Iulian Romeo Ihnatov

1. Introduction
This research represents an analysis of the CSR outcomes on the financial achievements.
A socially responsible application can change fundamentally the administration and activity of
banks, which can have a bigger impact on the financial score of both the banks and the wider
economy. If social responsibility has a posi tive impact on the financial performance of the banks,
then all the stakeholders and society at large will benefit. Therefore, it is important for all
stakeholders involved to understand the basic mechanisms that motivate firms to engage in social
activiti es.
A CSR scandal can produce a negative effect on the businesses, which are stirred by the
unfavorable market reactions. After the 2008 financial crisis, banks had to restore their
credibility and image. CSR is believed to be a method of reviving this cre dibility. Therefore, it
has become a priority for banks to be socially responsible, in order to remain competitive and to
increase the trust that the clients have in it (Hsu, 2012; Schultz, Castell ó and Morsing, 2013) .
Efforts to research the nature of the connection between CSR and FP (Financial
Performance) concluded that there is a complex connection and the research process is
theoretically and methodologically intractable. However, banks are more pressured nowadays to
improve their social conduct and b e more responsible. Its stakeholders are borrowers, owners,
employees, regulators. Therefore, the banks decisions and conduct affect a large number and
variety of people. Furthermore, questions will be answered like what resources should managers
operate t owards socially responsible activities and how should the stockholders respond to
allocations of resources towards social purposes. While these financial institutions do not
produce hazardous chemicals or unload toxic pollutants, it may have an indirect in volvement in
the environmental issues by supplying funds for the businesses involved in these types of
actions. Furthermore, another important responsibility is to provide information to clients and
reduce information asymmetry. In the end, another crucial part of banks` CSR activities, is to
bring efforts to combat corruption, which is usually met in the financial sector.

2. Literature review

2.1 Previous research on CSR and financial performance
The concept of CSR was popularized by early researchers i n order to boost organizations
to act for the clarification of several social issues. McGuire (1969) characterizes four approaches
to CSR: the traditional approach, the enlightened approach, the responsible approach and the
confused approach. The tradition al one is a neoclassical economics view and they declare that
CSR has no function in the business. The enlightened approach believes that CSR is a basic
function for the organizations. The responsible approach claims that CSR should be considered
an unbias ed endeavor, regardless of the likelihood of making a profit. The confused approach
actualizes simultaneously an ethical rationale of CSR and a belief of achievement after
implementing it. From a general finance outlook, any under or nonperformance in fina ncial
institutions, can be perceived as a corporate governance failure. Shleifer and Vishny (1997) give
the definition of corporate governance as a group of mechanisms used to facilitate firms to bring
a return on capital to the capital suppliers. If the e nvironments of corporate governance are not
optimally designed to assist the stakeholders of the institutions, it would be needed to ask if other
features of the corporate environment of the institutions are designed suitably, like the structure
of CSR.
The European Commission (2002) defines CSR as “the voluntary in tegration of social
and environmental concerns into business operations and into their interaction with
stakeholders”. Carroll (1979) characterizes CSR as including the legal, ethical, economic and
other unrestricted obligations that institutions have to society. Wood (1991) adds to the Ca rroll
model a classification of structural standards: inputs, social responsibility processes, throughputs
and outputs as a result. Vilanova, Lozano and Arenas (2008) introduce the five dimensions of
CSR: vision, community relations, workplace, accountability, and marketplace. Vision involves
CSR theoretical development regulations and organization value. Community relations include
cooperation with the stakehol ders, like customers, suppliers or partners. Workplace incorporates
human rights and work practices in the organization. Accountability is about communication
clarity and financial reporting. Marketplace includes the relationship between CSR and important
business processes, like sale and purchase.
Applying this to individual organizations, this is compatible with Freeman`s (1983 ) idea
of stakeholder theory, which implies that firms are responsible to different groups, like
employees, suppliers, customers and even society at large. Different organizations present
distinct objectives and guidelines for performance, depending on who are their stakeholders. In a
market economy, in equilibrium, the stakeholders should compel the organizations to implement
the greatest likely return to the specific provided capital. This includes returns to shareholders.
Therefore, the focus should be on the financial performance of organizations, being a quick
measurable source of returns.
Friedman (1962 ) stated that organizat ions are specialized to accomplish specific
functions, and society can benefit from this if the organizations do not comprehend

multifunctional objectives. Deriving from this, t he structural -functionalist theories introduced the
term “social” into business performance (Parson s, 1991) , which coherently spoken society into
diverse societies, to which different capacities are ascribed.
However, the topic of CSR is a controversial one, even though it has received increased
consideration in recent years. Organ izations are required to allocate additional investments in
CSR and these investments are inspected through the economic cost – benefit process, and
expected gains from CSR activities encourage CSR agreements. There are arguments that CSR
actions increase the expenses without bringing adequate offsetting benefits, damage performance
and go up against activities that maximize value. These CSR actions that bring additional costs
include organizing charitable donations, establishing proposals for community imp rovement, and
developing actions for pollution reduction.
There were several studies that examined the relationship between CSR and FP at a firm
level. The empirical results are mixed. Aupperle, Carroll and Hatfield (1985) applied surveys to
determine th e perspective of CEOs on CSR activities and announced that there is a powerful
negative connection between CSR and accounting -based performance measures. Furthermore,
Moore (2001) targets eight essential UK companies in the supermarket industry and discove rs a
negative relationship between CSR and FP. Nelling and Webb (2009) found no indication that
CSR is connected to FP. They used the KLD index as CSR measure and ROA for FP measure.
However, there are many studies that found a positive relationship betw een CSR and FP.
Cochran and Wood (1984 ) indicate that there is needed more broad measures of CSR for further
research. McGuire, Sundgren and Schneeweis (1988) found a positive relationship between CSR
and accounting -based and market -based financial perfor mance measures. They concluded that
the preceding performance is more closely related to CSR, compared to the firm`s successive
performance. Waddock and Graves (1997) had two considerable findings: the better the FP the
better future CSR performance and a better CSR performance leads to a better future FP. They
used KLD data for CSR performance and ROA, ROE and ROS for the measurement of FP.
Orlitzky, Schmidt and Rynes (2003) provided a meta -analysis, which concluded a positive
relationship between CSR and FP. They suggest that practicing socially responsibility can bring
economic benefits to the firms and contribute to maximization of wealth. Tsoutsoura (2004)
selected firms from S&P 5 00 Index for the period 1996 – 2000 and found a significant and
positive relationship between KLD data and FP, measured with ROA, ROE and ROS. Anderson
and Myers (2007) discover that investors are not suffering losses if they base their investment
decisions on criteria connected to CSR that are compatible with their social beli efs. This was
based on KLD data. Shen and Chang (2009) demonstrate that organizations with powerful CSR
climates generally perform better than organization with fragile CSR environments across
diverse financial metrics. Barnea and Rubin (2010) studied the connection between CSR
investment and firm features. The findings show that insiders may over -invest in CSR projects if
their personal gains are high and the personal costs are low. It could be perceived a green –
washing form, and a focus on the CSR approac h investment and not the essence. This brings
benefits to individuals, but it can hurt the organizations. Dhaliwal, Zhen Li, Tsang and Yang
(2011) brought the conclusion that it brings future benefits for the firms to engage in CSR

activities. They found t hat firms with an increased cost of capital in the preceding year are more
probable to make public the current CSR activities. Furthermore, firms with a better CSR
performance gain a lower cost of capital in the following year. Plus, the firms with better CSR
performance obtain encouraging analyst coverage and accomplish lower absolute predicted
errors and dispersion. El Ghoul, Guedhami, Kwok and Mishra (2011) identify that organizations
with compelling CSR have lower firm risk, lower costs of equity, and h igher assessment. Cheng,
Ioannou and Serafeim (2011) show that organizations with higher CSR present lower capital
restrictions, which indirectly brings more favorable circumstances for investment and higher
appraisal. Goss and Roberts (2011) presented the fact that non -responsible organizations pay a
larger fee to get access to financial resources than organizations implicated in social endeavors,
because CSR investments reduce the risk and gain the attention of potential investors and
lenders.
CSR action s boost the connections with stakeholders, this leading to enhanced returns. A
socially responsible firm may encounter fewer employment difficulties, fewer objections from
the community, and fewer environmental burdens from the government. Additionally, so cially
responsible firms may encounter improved relationships with investors, bankers and government
officials. Firms that care about being socially responsible are performing better in the society of
today.
The banking system has a crucial duty in the de velopment and safety of the economy.
However, there are not a lot of studies presenting the relationship between CSR and banks
financial performance. Shen and Lee (2006 ) found out that the banking system safety and
soundness over time produces fundamental advantages for society. Scholtens and Dam (2007)
analyze the how faithful is the commitment to CSR policies. The conclusion was that the banks
that embrace the Equator Principles are seeking substantial CSR procedures, but they indicate
lower ROA. In one s tudy, Ahmed, Islam and Hasan (2012) presents an insignificant, but positive
relationship between CSR and operating performance for a very small sample of banks in
Bangladesh. Wu and Shen (2013) analyzed the relationship between CSR and CFP in the global
banking sector, through the reasons that could motivate banks to adopt these policies: strategic,
altruistic and greenwashing. They concluded that there is a positive, uncertain or neutral
relationship. Their empirical results demonstrated that CSR is associ ated positively with ROA,
ROE, Net Interest Income and Non -Interest Income. However, is it negatively assoc iated with
non-performing loans. Wu and Shen (2013) state that the character of the banking endeavor,
identified by the resources usag e acquired from non-proprietary stakeholders, obligates the
feedback provision to the society more frequently than other industries. The alleviation of
intermediation carried out by banks includes supervisory authorities, academics, governments
and even society control s regarding the general activity, including the CSR ones. Banks are
apprehensive about their particularized businesses, based on the contact and trust with the public.
This consciousness defines the CSR existence in the annual financial statements. Schultz,
Caste lló and Morsing (2013) found out that after the financial crisis from 2008, which implicated
all the major world banks and crippled their reputation and credibility, CSR is recognized as a
strategy to re -establish that credibility. Dorasamy (2013) states that the involvement of banks in
CSR activities advocates the adherence of systems by potential aspirants, which generates a

positive impact on sustainable growth. Polychronidou, Ioannidou, Kipouros, Tsourgiannis, and
Simet (2014) declared that the involvement of the bank in CSR procedures demonstrate the
advantages accumulated by the bank. Carnevale, Mazzuca and Venturini (2014) used a sample of
European Banks and disclosed the positive influence of the published sustainability report on the
share price. Corne tt, Erhemjamts and Tehranian (2015) analyzed the relationship between CSR
and FP of U.S. Commercial Banks in the context of the recent financial crisis. The conclusion
was that banks are benefiting from being socially responsible, since the CSR score has a positive
and significant relationship with the FP. Another conclusion was that big banks seek socially
responsible pursuits to a greater extent than banks that are smaller. The largest banks
experienced an increase in the CSR health and a decrease in CSR burdens after 2009. Forcadell
and Aracil (2017) examined if the reputation of the European banks listed on the Dow Jones
Sustainability Index (DJSI) is improved by the CSR policies. The conclusion was that the effort
to increase the reputation through CSR policies has financial performance benefits.
2.2 Bank performance
Numerous studies analyzed bank profitability and they focused on three factors: bank
specific, bank industry specific and macroeconomic variables. Most researchers measured bank
performance using Retu rn on Equity (ROE) or Return on Assets (ROA). Even though the
literature began sooner, the most notable examples are by Molyneux and Thornton (1992) , in a
study of bank profitability from 18 European countries in the period 1986 -1989, and Demirgüç –
Kunt and Huizinga (1999) , who researched the internal and external incentives of profitability
for banks in 80 countries be tween 1988 and 1995. In the second study, the descriptive variables
involve several bank specific ratios, such as cash and securities to total assets, staff expenses to
total assets, and bank capital to total assets, bank industry specific ratios such as concentration or
the structure of ownership (government or private), and macroeconomic variables such as
inflation, money su pply growth and interest rates. They found that foreign banks are more
profitable than domestic ones in developing countries. Demirgüç -Kunt and Huizinga (1999 )
deduced that banks with high equity are the most performing, and they present a low default risk
and financing costs.
Kwan (2003) compared , between 1992 and 1999, the banks performance in seven Asian
countries. Athanasoglou, Delis and Staikouras (2006) examin ed the determinants of profitability
in seven Central and Southern European countries between 1998 and 2002, including bank
specific factors like the index that indicates the bank reform progress that is typical to transition
economies. They found a negati ve and significant relationship between this indicator and bank
profitability (ROA and ROE). In the study of Pasiouras and Kosmidou (2007) , it was found a
positive relationship between bank size and bank profit. However, Micco, Panizza and Yanez –
Pagans (20 07) discovered a not significant relationship between bank size and ROA. To
summarize these results, different measures of cost are mainly negatively correlated with profits.
Bigger banks, greater reliance on loans for revenue, greater GDP growth, higher m arket
concentration, and greater amounts of equity capital to assets have been correlated with greater
profitability. Greater provisions for loan losses, higher liquidity, and more dependence on debt
have been suggestive for lower bank profits.

Hefferna n and Fu (2010) had a study on Chinese bank performance and suggested that
net interest margin (NIM) and economic value added are more precise measures of performance
than ROA and ROE. They found that some financial and macroeconomics ratios are meaningful
to performance. Olson and Zoubi (2011) found that expense ratios, loan rations, credit risk,
capital strength, inflation, and the government ownership proportion have significant
developments on both ROA and ROE. Sufian and Habibullah (2012) investigated the effect of
globalization on the performance of banks in China, using ROA and bank specific
characteristics, macroeconomic variables, and globalization factors as the determinant variables.
They found that the highly capitalized banks are more profitabl e, and that expense choice
behavior has a negative connection with bank profitability. The advancement of regulation
application, the expansion of credit and acceptance of reasonable macroeconomic policies
increased the competition in the banking sector. Beltratti and Stulz (2012) analyzed the effect of
bank and country governance, domestic regulation, the balance sheet of the bank and the before
crisis profit on ban k performance, while examining why some banks had a progress during the
crisis. The conclu sion was that the banks had a superior performance in the countries with strict
capital adequacy requirements and supervision authorities that are independent. The banks
originated from the countries with strong supervision authorities had low market retur ns,
considering that the shareholders choose the costly method of raising new equity during the
crisis. Bold, de Haan, Hoeberichts, van Oordt and Swank (2012) analyzed the impact of the last
global financial crisis on bank performance, reaching the conclusion that bank profitability wa s
affected by the economic cycle during the recent crisis. They showed that, during complex
recessions, the ROA decreases with 0.24% if the real GDP declines by 1% at the level of
banking industry. This is related to the fact that the private sector bank l oans are significantly
dependent to the level of GDP, considering that a decrease in GDP decays the asset condition
and boosts the non -performing loans. Căpraru and Ihnatov (2014) assessed the main
determinants of bank profitability in Central and Eastern Europe, using RO AA, RO AE and NIM
as proxies. The results showed that management efficiency and capital adequacy growth affect
the bank profitability for a ll dependent variables, while inflation and credit risk do not determine
NIM, only ROAA and ROAE. Furthermore, banks with higher capital adequacy are more
profitable. Andrie ș, Căpraru, Ieșan -Munteanu and Ihnatov (2016) conducted an analysis on the
determin ants of bank profibaility on European banks. By using multiple regression, the
conclusion presents that there was a negative and significant statistically impact of the
international crisis on ROA. The bank sample wa s separated in three categories : banks w ith high
adequate capital, large banks by total assets and foreign banks. The variables heightened and
diminished the crisis effect.
While there is a vast and comprehensive literature that examines the bank specific,
industry specific and macroeconomic fa ctors that have an effect on the profitability of the banks,
the conclusions are different. This may happen because of the different countries involved in the
studies, that present distinct characteristics, but also due to the contrasting macroeconomic
circumstances and the used time period and data sets.

3. Data and methodology
The research is based on a panel of 19 countries from Europe: Austria, Belgium, Cyprus,
Czech Republic, Denmark, Finland, France, Germany, Hungary, Italy, Netherlands, Norway,
Poland, Russian Federation, Spain, Sweden, Switzerland, Turkey and United Kingdom of Great
Britain and Northern Ireland. It was chosen 66 banks present in these countries, which were all
commercial, both private and public. Individual banks may be owned by bank h olding
companies, which provide more financial services, but the rating for the annual sustainability
report is applied to the individual bank, no the bank holding company. The number of banks
from each country is presented in Appendix 1.
There are used t hree categories of variables that the literature shows it influences bank
performance: bank specific varia bles, bank industry variables and macroeconomic variables. The
bank level data are from Bankscope database, and are reported as percent, except Total Assets
which is expressed in thousand s of Euros. The banking industry and macroeconomic data are
from ECB and World Bank. The CSR data is taken from Global Reporting Database, which is an
independent organization which pioneer ed sustainability reporting. G lobal Reporting Initiative
(GRI) develops sustainability standards to worldwide businesses and governments, which are
used to communicate their impact on climate change, human rights, governance and social well –
being. They create the GRI Reporting Standard s, which are developed with true multi –
stakeholder contribution and rooted in public interest. Therefore, the organizations that are
publishing their sustainability reports are registering them with GRI. These reports receive a
rating, reflecting the exten t to which the GRI Standards have been applied.
There are used three measures of FP that are considered to capture major dimensions of
financial performance in the banking indu stry: return on assets (ROA), return on equity (ROE)
and Net Interest Income (N IM). The most widely used is Return on A ssets, which measures the
ability of the bank to acquire deposits at decent costs, invest these funds in loans and investments
that are profitable, and to perform profitable daily operations.
In order to take the 20 08 financial crisis into consideration, the data is divided into three
segments: from 2005 -2007, from 2008 -2010 and from 2011 -2016.

Table 1. Variables in the model (computation and expected result)
Notation Variable Computation Expected result
Dependent variables
ROA Return on Assets Net profit/Total assets
ROE Return on Equity Net profit/Average Common Stock Equity
NIM Net Interest Margin Difference between interest income and
interest expense/Total assets of the bank
Independent variabl es
Bank specific variables
Capital Capital adequacy Equity/Total Assets +/-
Risk Credit risk Non-performing loans/Total loans –
Size Bank size Logarithm of Total Assets +/-
Banking industry variables
Bank concentration Market
concentr ation index Herfindhal -Hirschman index +/-
Macroeconomic variables
GDP Economic growth GDP per capita growth (annual %) +
Inflation Inflation Inflation rate (annual %) +/-
CSR
CSR CSR rating Dummy variable +
External assessment External assessment Dummy variable +

3.1 Bank specific variables
For the bank specific factors, it was chosen those that are most generally encountered in
the literature: capital adequacy, credit risk and bank size.
Capital adequacy is expressed in percentag es of equity of total bank assets and it assesses
if the bank is risk averse or it gets involved in more risky projects. The banks that present a
higher level of capital adequacy are treated as to be more secure, while those that present a lower
capital ad equacy are implicated in more risky activities. High equity banks continue to be
profitable during troublesome times and can earn financing with diminished costs.
Credit risk is calculated as the ratio of non -performing loans to total loans. It indicates the
condition of loan portfolio and an increase results in a recorded profit decrease.
Bank size is determine d as logarithm of Total Assets, and there are conflicting outcomes
when it comes to bank profitability, since bigger banks present a wide area of products and
services and they can take advantage of scale economies. However, at the same time, they may
take big and unnecessary risks, which have a negative effect on bank stability.
3.2 Bank Industry variables
It was chosen another variable used in l iterature, degree of concentration in the banking
industry , measured by the HHI index . A high level of concentration can be the result of an

accord between the largest banks from the whole to put in practice higher interest rate levels,
which may have a po sitive effect on the profitability or may be the development of a competitive
banking system.
3.3 Macroeconomic variables
The important macroeconomic factors used in literature are: GDP per capita growth and
inflation . An increase in the GDP per capita growth produce s an increase of demand in loans. A
decrease in the economy reduced the demand, which generates an increase in non -performing
loans and has a negative effect on bank profits. The inflation rate presents the change in the price
levels and the inflationary environment in the economy. It is measures annually per country and
can have both a negative and positive impact on the bank profitability. When banks anticipate it,
they alter the interest rate to increase proportionally the costs and income. If it is not expected, it
may affect the economy and the profit.
3.4 CSR indicators
The CSR variable is based on the GRI rating. Global Reporting Initiative present
sustainability reports based on the GRI Standards. There are four Standards: G1 (publishe d in
2000), G2 (published in 2002), G3 (published in 2006) and G4 (published in 2013).
There is given an adherence level to these standards, which reflect the range to which the
GRI Standards have been applied to the report . This adherence levels are diff erent depending on
the GRI Standards. Therefore, it was assigned a score to this ratings, which were used as dummy
variables in the analysis.

Table 2. CSR adherence levels and score
Adherence levels Score Explication
No publication of sustainability rep ort 0 No publication of sustainability report
Undeclared 1 No explicit application level declared
C, Context Index only 2 There is GRI Content available, but it
is not in accordance to the standards

B, In accordance core

3 Contains the fundamental a spects of a
sustainability report, with
communication about the economic,
environmental and social , and
governance performance.

A, In accordance Comprehensive
4 Contains additional information about
the governance, ethics and integrity,
and the strategy and analysis of the
organization.

Another indicator is used for External A ssurance , which is mea sured through a dummy
variable, which is 1 if the report is externally assured and 0 if it is not.

3.5 Descriptive statistics

The average Return on As sets for the entire sample is 1.2% during the period 2005 – 2016.
Between the years 2005 – 2007 , the average of ROA was 1.7%, but it decreased to 1.3% during
the financial crisis. There is not a big difference between the average of ROA and the median,
therefore there are not significant differences between the banks in the sample. The average
Return on Equity for the entire sample is 8.9% , with the median being 10.7% . The average ROE
decreased visibly from 2008, considering that it was 17.205% between the years 2005 – 2007. It
could mean that the companies were not that successful in generating cash internally during the
financial crisis. The average Net Interest Margin is 2.369%. This variable did not present high
differences between the years. However, th e average NIM is not very high.

The Capital Adequacy is 11.809% on average, with the highest level at 28.7% and the lowest
at 5.13 %. The average ratio of non -performing loans to total loans is 4.52%, the highest being
recorded as 36.279. This are not ver y high ratios, demonstrating that the banks presented in the
sample do not attain very high quality portfolios. The banks are not on average in a very
competitive market.

The average economic growth per capita is low, being 0.987%, with big differences b etween
the countries.

Table 3. Descriptive statistics (%, except the variables marked with *)
Average Median Maximum Minimum Std. dev. s
Dependent variables
ROA 1.214 1.100 7.480 -4.800 1.028
ROE 8.936 10.700 39.920 -170.090 14.462
NIM 2.369 2.100 6.400 0.710 1.115
Bank specific variables
Capital 11.809 11.800 28.700 5.130 3.720
Risk 4.520 2.900 36.279 0.000 5.349
Size* 18.330 18.147 21.844 14.449 1.762
Banking industry variables
Bank concentration * 68.231 66.800 100.000 22.534 19.276
Macroeconomic variables
GDP 0.987 1.100 8.720 -8.706 2.872
Inflation 2.523 1.900 15.524 -2.097 2.865

For the CSR variable, most of the banks did not publish a sustainability during the period
2005 – 2007. During the years 2008 – 2010, there was noticed an increased in the published
reports, with most published reports presenting the highest adherence level. There was a visible
difference in the period 2011 – 2016, with 35% of the banks having no published report.
However, the quality of the published reports decreased, most of them having a B or “In
accordance -Core ” adherence level.

Figure 1. CSR score on average during the period 2005 -2007

Between the years 2005 – 2007, 75% of the banks published no reports, while 2%
published a report with no application level declared. 9% of the published a C rating report,
while 3% published a B rating report and 11% published an A rating report.

Figure 2. CSR score o n average during the period 2008 -2010 75%2%9%3%11%CSR Ratings 2005 -2007
62%
7%7%11%13%CSR Ratings 2008 -2010

Between the years 2008 -2010, 65% of the ba nks published no reports, while 7 %
published a report with no application level declared. 7 % of the published a C rating report,
while 11 % pu blished a B rating report and 13 % published an A rating report.

Figure 3 . CSR score o n average during the per iod 2011 -2016

Between the years 2011 -2016, 35% of the banks published no reports, while 17%
published a report with no application level declared. 7% of the published a C rating report,
while 27% published a B rating report and 16% published an A rating report.

3.6 Methodology

There are two central hypotheses:

– The CSR reporting of a European bank can predict the financial performance of the bank;
– The more socially responsible the bank is, the higher the financial performance;
The main research q uestion is:

What is the strength of the link between the level of CSR reporting and the financial
performance of the selected European Banks?

35%
17% 5%27%16%CSR Ratings 2011 -2016

There are three regressions equations estimated with the FP measures as the dependent
variable and the CSR me asure as the independent variable, plus a set of control variables.
Ordinary least squares regression is used to estimate the regression parameters and standard
regressions diagnostics are used to evaluate the efficiency of the results. The estimated equat ion
is the following one:
Y = α + X1β1+ X2β2 + X3β3 +X4 β4 + ε (1)
Where:

Y – Dependent variables (ROA, ROE, NIM);
X1 – Vector of CSR indices;
X2 – Vector of bank -specific factors;
X3 – Vector of bank industry determinants;
X4 – Vector of macroeconomics determinants;
ε – Error term;
β1 – Matrix of variable coefficients;

3 Results
2005 -2007

The first regression was done for the years 2005 -2007. In the beginning, it was used ROA as
a dependent variable and only one indicator from each category: CSR, ROA from the previous
year, logarithm of total assets, bank concentration and GDP per capita growth. It can be observed
that the coefficient, ROA from the previous year, L ogarithm of TA, bank concentration and GDP
per capita growth are statistically significant, since the p -values are lower than 0.05. The CSR
variable is not statistically significant. The R -squared is 0.327, therefore the model explains
32.74% of the variab ility. The model is:

ROA =0.036456 -0.386531 ROA( -1) -0.000287 CSR -0.001292 LOG_TA – 0.007285
BANK_CONCENTRATION + 0.084609 GDP_CAP_GROWTH

Dependent Variable: ROA
Method: Least Squares
Date: 05/20/18 Time: 13:55
Sample (adjusted): 2 261
Included observations: 218 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 0.036456 0.008456 4.311183 0.0000
ROA( -1) 0.386531 0.059755 6.468562 0.0000
CSR 0.000287 0.000811 0.353747 0.7239
LOG_TA -0.001292 0.000394 -3.282583 0.0012
BANK_CONC -0.007285 0.002987 -2.439303 0.0155
GDP_CAP_GROWTH 0.084609 0.034168 2.476275 0.0141
R-squared 0.327403 Mean dependent var 0.017369
Adjusted R -squared 0.311540 S.D. dependent var 0.010351
S.E. o f regression 0.008589 Akaike info criterion -6.649618
Sum squared resid 0.015638 Schwarz criterion -6.556467
Log likelihood 730.8084 Hannan -Quinn criter. -6.611993
F-statistic 20.63928 Durbin -Watson stat 1.975516
Prob(F -statistic) 0.000 000

Then, it was used ROE as a dependent variable and only one indicator from each category: CSR,
ROE from the previous year, logarithm of total assets, bank concentration and GDP per capita
growth. It can be observed that the ROE from the previous year and GDP per capita growth are
statistically significant, since the p -values are lower than 0.05. The coefficient, CSR variable,
logarithm of TA and bank concentration are not statistically significant. The R -squared is 0.0803,
therefore the model explains only 8% of the variability. The model is:

ROE =0.029116 + 0.225689 ROE( -1) -0.002989 CSR + 0.005172 LOG_TA + 0.000632
BANK_CONCENTRATION + 0.005657 GDP_CAP_GROWTH

Dependent Variable: ROE

Method: Least Squares
Date: 05/20/18 Time: 14:09
Sample (adjusted): 2 261
Included observations: 250 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 0.029116 0.068127 0.427370 0.6695
ROE( -1) 0.225689 0.062065 3.636329 0.0003
CSR -0.002989 0.006413 -0.466163 0.6415
LOG_TA 0.005172 0.003293 1.570900 0.1175
BANK_CONC 0.000632 0.025181 0.025118 0.9800
GDP_CAP_GR 0.005657 0.002849 1.985450 0.0482
R-squared 0.080332 Mean dependent var 0.179456
Adjusted R -squared 0.061486 S.D. dependent var 0.080911
S.E. of regression 0.078384 Akaike info criterion -2.230682
Sum squared resid 1.499155 Schwarz criterion -2.146167
Log likelihood 284.8353 Hannan -Quinn criter. -2.196668
F-statistic 4.262628 Durbin -Watson sta t 2.021830
Prob(F -statistic) 0.000982

It was more significant the equation with ROA as dependent variable than the ROE one.
However, the CSR variable did not present significance for any of the equations. Therefore, with
these variables, CSR did not have significance over the financial performance, for 2005 -2007.
In the end, it was introduced all the values into equation. For the ROA as a dependent
variable, the ROA from previous year, Loan to Assets ratio and NIM are statistically signi ficant,
the p value being lower than 0.05. The coefficient, CSR variables, logarithm of Total Assets,
Loans to Total Deposits ratio, bank concentration, GDP per capita growth and Rule of Law are
not statistically significant. However, the R squared value i s pretty high, being 0.625. Therefore,
the model explains 62.5% of the variability.

Dependent Variable: ROA
Method: Least Squares
Date: 05/20/18 Time: 14:31
Sample (adjusted): 41 225
Included observations: 48 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 0.006611 0.017689 0.373752 0.7108
ROA( -1) -0.267484 0.101875 -2.625616 0.0126
CSR -0.001214 0.001174 -1.034052 0.3080
STATUS 0.000302 0.000640 0.472258 0.6396
EXTERNAL_ASSURANCE 9.15E -05 0.002810 0.032543 0.9742
LOG_TA -0.000421 0.000695 -0.606256 0.5481
LOAN_TA 0.013838 0.007119 1.943937 0.0598
LTD -0.000474 0.001448 -0.327116 0.7455

BANK_CONC 0.010059 0.007019 1.433222 0.1604
NIM 0.182622 0.099418 1.836908 0.0745
GDP_CAP_GR 8.41E-05 0.000486 0.172942 0.8637
RULE_LAW 0.000263 0.001932 0.136214 0.8924
R-squared 0.652155 Mean dependent var 0.014638
Adjusted R -squared 0.545868 S.D. dependent var 0.004342
S.E. of regression 0.002926 Akaike info criterion -8.617852
Sum squared resid 0.000308 Schwarz criterion -8.150052
Log likelihood 218.8285 Hannan -Quinn criter. -8.441070
F-statistic 6.135839 Durbin -Watson stat 1.949506
Prob(F -statistic) 0.000015

For the ROE as a dependent variable, no variable was statistically significant. The R squared
value is also low, being 0.1957. Therefore, the model explains 19.57% of the variability.

Dependent Variable: ROE
Method: Least Squares
Date: 05/20/18 Time: 14:41
Sample (adjusted): 41 225
Included observations: 55 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 0.129618 0.242230 0.535102 0.5953
ROE( -1) 0.158407 0.151792 1.043582 0.3025
CSR -0.013758 0.010631 -1.294149 0.2025
STATUS 0.001250 0.009832 0.127168 0.8994
EXTERNAL_ASSURANCE -0.000568 0.041104 -0.013808 0.9890
LOG_TA 0.000758 0.009832 0.077082 0.9389
LOAN_TA 0.054465 0.114816 0.474364 0.6376
LTD -0.015592 0.022523 -0.692254 0.4925
BANK_CONC 0.0461 95 0.108600 0.425364 0.6727
NIM 0.612681 1.539816 0.397892 0.6927

GDP_CAP_GROWTH -0.199608 0.772810 -0.258289 0.7974
RULE_LAW -0.019509 0.032335 -0.603334 0.5495
R-squared 0.195709 Mean dependent var 0.191527
Adjusted R -squared -0.0100 40 S.D. dependent var 0.050183
S.E. of regression 0.050435 Akaike info criterion -2.946049
Sum squared resid 0.109377 Schwarz criterion -2.508086
Log likelihood 93.01636 Hannan -Quinn criter. -2.776685
F-statistic 0.951203 Durbin -Watson stat 2.502085
Prob(F -statistic) 0.503176
Therefore, for the years 2005 -2007, the CSR variable did not present any significant
influence over the financial performance.

2008 -2010
Then, it was done the regressions for the years 2008 -2010. In the beginning, it was used
ROA as a dependent variable and only one indicator from each category: CSR, ROA from the
previous year, logarithm of total assets, bank concentration and GDP per capita growth. It can be
observed that the coefficient, ROA f rom the previous year, Logarithm of TA and GDP per capita
growth are statistically significant, since the p -values are lower than 0.05 or 0.1. The CSR
variable and bank concentration are not statistically significant. The R -squared is 0.451,
therefore the model explains 45% of the variability. The model is:

ROA =0.027086 +0.536368 ROA( -1) -0.000531 CSR -0.002300 LOG_TA – 0.003041
BANK_CONCENTRATION + 0.026232 GDP_CAP_GROWTH

Dependent Variable: ROA
Method: Least Squares
Date: 05/20/18 Time: 1 4:51
Sample (adjusted): 2 198
Included observations: 167 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 0.027086 0.007787 3.478244 0.0006
ROA( -1) 0.536368 0.062189 8.624839 0.0000
CSR -0.000531 0.000497 -1.069436 0.2865
LOG_TA -0.002300 0.000862 -2.669228 0.0084
BANK_CONC -0.003041 0.003206 -0.948632 0.3442
GDP_CAP_GROWTH 0.026232 0.016025 1.636972 0.1036
R-squared 0.451690 Mean dependent var 0.013269
Adjusted R -squared 0.43466 2 S.D. dependent var 0.009592
S.E. of regression 0.007212 Akaike info criterion -6.990741
Sum squared resid 0.008375 Schwarz criterion -6.878718

Log likelihood 589.7269 Hannan -Quinn criter. -6.945273
F-statistic 26.52596 Durbin -Wats on stat 1.988521
Prob(F -statistic) 0.000000

Then, it was used ROE as a dependent variable and only one indicator from each category: CSR,
ROE from the previous year, logarithm of total assets, bank concentration and GDP per capita
growth . It can be observed that the coefficient, ROE from the previous year, logarithm of total
assets and GDP per capita growth are statistically significant, since the p -values are lower than
0.05. The CSR variable and bank concentration are not statistically significant. The R -squared is
0.2422, therefore the model explains 24.22% of the variability. The model is:

ROE =0.296485 + 0.318975 ROE( -1) -0.003351 CSR – 0.026261 LOG_TA -0.026432
BANK_CONCENTRATION + 0.302611 GDP_CAP_GROWTH

Dependent Variable: ROE
Method: Least Squares
Date: 05/20/18 Time: 14:57
Sample (adjusted): 2 198
Included observations: 197 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 0.296486 0.074879 3.959525 0.0001
ROE( -1) 0.318975 0.066720 4.780786 0.0000
CSR -0.003351 0.005096 -0.657528 0.5116
LOG_TA -0.026261 0.008513 -3.084794 0.0023
BANK_CONC -0.026432 0.032020 -0.825474 0.4101
GDP_CAP_GROWTH 0.302611 0.161292 1.876169 0.0622
R-squared 0.242274 Mean dependent var 0.095823
Adjusted R -squared 0.222439 S.D. dependent var 0.089769
S.E. of regression 0.079158 Akaike info criterion -2.204768
Sum squared resid 1.196793 Schwarz criterion -2.104772
Log likelihood 223.1696 Hannan -Quinn crite r. -2.164289
F-statistic 12.21403 Durbin -Watson stat 1.972609
Prob(F -statistic) 0.000000

The CSR variable did not present significance for any of the equations. Therefore, with these
variables, CSR did not have significance over the financial performance, for 2008 -2010.
In the end, it was introduced all the values into equation. For the ROA as a dependent
variable, the bank concentration and rule of law are statistically significant, the p value being
lower than 0.1. All the other v ariables are not statistically significant. The R squared value is
0.5175. Therefore, the model explains 51.7% of the variability.

Dependent Variable: ROA
Method: Least Squares
Date: 05/20/18 Time: 15:01
Sample (adjusted): 18 192

Included ob servations: 51 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C -0.011370 0.020553 -0.553179 0.5833
ROA( -1) 0.152054 0.128695 1.181505 0.2446
CSR -0.000101 0.000736 -0.137070 0.8917
STATUS -0.000843 0.001142 -0.737927 0.4650
EXTERNAL_ASSURANCE -0.003042 0.002589 -1.175145 0.2471
LOG_TA 0.000816 0.002179 0.374498 0.7101
LOAN_TA 0.013888 0.009515 1.459637 0.1524
LTD 0.001273 0.002381 0.534715 0.5959
BANK_CONC 0.014020 0.008147 1.720820 0.0932
NIM 0.082803 0.133568 0.619930 0.5389
GDP_CAP_GROWTH 0.003168 0.027302 0.116033 0.9082
RULE_LAW -0.004539 0.002454 -1.849427 0.0720
R-squared 0.517592 Mean dependent var 0.010976
Adjusted R -squared 0.381528 S.D. dependent var 0.006637
S.E. of regression 0.005219 Akaike info criterion -7.470618
Sum squared resid 0.001062 Schwarz criterion -7.016071
Log likelihood 202.5008 Hannan -Quinn criter. -7.296922
F-statistic 3.804035 Durbin -Watson stat 2.468712
Prob(F -statistic) 0.000941

For the ROE as a dependent variable, no variable was statistically significant. The R squared
value is also low, being 0.235. Therefore, the model explains 23.5% of the variability.

Dependent Variable: ROE
Method: Least Squares
Date: 05/20/18 Time: 15:04
Sample (adjusted): 18 192
Included observations: 57 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 0.581771 0.386624 1.504747 0.1394
ROE( -1) 0.017609 0.138018 0.1275 83 0.8990
CSR 0.001945 0.012815 0.151775 0.8800
STATUS -0.001802 0.020823 -0.086548 0.9314
EXTERNAL_ASSURANCE -0.013737 0.048346 -0.284131 0.7776
LOG_TA -0.048069 0.040765 -1.179174 0.2445
LOAN_TA 0.043168 0.170696 0.252893 0.8015
LTD -0.034473 0.044 873 -0.768235 0.4464
BANK_CONC -0.041571 0.143289 -0.290124 0.7731
NIM -1.492846 1.775301 -0.840898 0.4049
GDP_CAP_GROWTH -0.201283 0.492912 -0.408354 0.6850
RULE_LAW -0.009222 0.045129 -0.204355 0.8390
R-squared 0.235498 Mean dependen t var 0.091732
Adjusted R -squared 0.048620 S.D. dependent var 0.102001
S.E. of regression 0.099491 Akaike info criterion -1.592844
Sum squared resid 0.445427 Schwarz criterion -1.162728
Log likelihood 57.39606 Hannan -Quinn criter. -1.425687

F-statistic 1.260169 Durbin -Watson stat 2.701556
Prob(F -statistic) 0.278007

Therefore, for the years 2008 -2010, the CSR variable did not present any significant influence
over the financial performance.

2011 -2016
Then, it was d one the regressions for the years 2011 -2016. In the beginning, it was used
ROA as a dependent variable and only one indicator from each category: CSR, ROA from the
previous year, logarithm of total assets, bank concentration and GDP per capita growth. It c an be
observed that all the variables are statistically significant, since the p -values are lower than 0.05.
The R -squared is 0.4816, therefore the model explains 48.16% of the variability. The model is:

ROA =0.025628 +0.457819 ROA( -1) +0.000856 CSR -0.002 304 LOG_TA – 0.005769
BANK_CONCENTRATION + 0.133352 GDP_CAP_GROWTH

Dependent Variable: ROA
Method: Least Squares
Date: 05/20/18 Time: 15:07
Sample (adjusted): 2 394
Included observations: 202 after adjustments
Variable Coeffic ient Std. Error t-Statistic Prob.
C 0.025628 0.005848 4.382631 0.0000
ROA( -1) 0.457819 0.053032 8.632830 0.0000
CSR 0.000856 0.000362 2.363156 0.0191
LOG_TA -0.002304 0.000661 -3.485570 0.0006
BANK_CONC -0.005769 0.002497 -2.310144 0.021 9
GDP_CAP_GROWTH 0.133352 0.025736 5.181549 0.0000
R-squared 0.481664 Mean dependent var 0.009040
Adjusted R -squared 0.468441 S.D. dependent var 0.008783
S.E. of regression 0.006404 Akaike info criterion -7.234691
Sum squared r esid 0.008037 Schwarz criterion -7.136425
Log likelihood 736.7038 Hannan -Quinn criter. -7.194932
F-statistic 36.42662 Durbin -Watson stat 1.681241
Prob(F -statistic) 0.000000
Therefore, for the years 2011 -2016, the banks with a higher CSR rating present a higher
ROA with 0.000856, all the other variables being constant.
Then, it was used ROE as a dependent variable and only one indicator from each
category: CSR, ROE from the previous year, logarithm of total assets, bank concen tration and
GDP per capita growth. It can be observed that the ROE from the previous year, CSR variable
and GDP per capita growth are statistically significant, since the p -values are lower than 0.05.
The coefficient, logarithm of total assets and bank con centration are not statistically significant.
The R -squared is 0.546, therefore the model explains 54.6% of the variability. The model is:

ROE =0.052523 + 0.725570 ROE( -1) +0.013862 CSR – 0.008184 LOG_TA 0.017099
BANK_CONCENTRATION + 2.198276 GDP_CAP_GROW TH

Dependent Variable: ROE
Method: Least Squares
Date: 05/20/18 Time: 15:26
Sample (adjusted): 2 394
Included observations: 250 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 0.052523 0.103 817 0.505917 0.6134
ROE( -1) 0.725570 0.056177 12.91574 0.0000
CSR 0.013862 0.006999 1.980709 0.0487
LOG_TA -0.008184 0.012194 -0.671162 0.5028
BANK_CONC -0.017099 0.047684 -0.358597 0.7202
GDP_CAP_GROWTH 2.198276 0.481297 4.567396 0.0000
R-squared 0.546991 Mean dependent var 0.035812
Adjusted R -squared 0.537708 S.D. dependent var 0.199567
S.E. of regression 0.135690 Akaike info criterion -1.133184
Sum squared resid 4.492459 Schwarz criterion -1.048669
Log likelihood 147.6480 Hannan -Quinn criter. -1.099169
F-statistic 58.92420 Durbin -Watson stat 2.084102
Prob(F -statistic) 0.000000

The CSR variable presented significance for both equations. Therefore, with these variables,
CSR had significance o ver the financial performance, for 2011 -2016.
In the end, it was introduced all the values into equation. For the ROA as a dependent
variable, the coefficient, ROA from the previous year and logarithm of TA are statistically
significant, the p value being lower than 0.1. All the other variables are not statistically
significant. The R squared value is 0.465. Therefore, the model explains 46.5% of the variability.

Dependent Variable: ROA
Method: Least Squares
Date: 05/20/18 Time: 15:35
Sample (a djusted): 31 382
Included observations: 83 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 0.031018 0.019088 1.625008 0.1086
ROA( -1) 0.316297 0.099655 3.173904 0.0022
CSR 0.000602 0.000542 1.109659 0.2709
STATUS -0.000684 0.000795 -0.859917 0.3927
EXTERNAL_ASSURANCE -0.001716 0.001717 -0.999698 0.3209
LOG_TA -0.002973 0.001830 -1.624630 0.1087
LOAN_TA -0.005023 0.009188 -0.546674 0.5863
LTD 0.003394 0.002546 1.333073 0.1868
BANK_CONC 0.002477 0.005278 0.469266 0.6403

NIM -0.058327 0.080460 -0.724921 0.4709
GDP_CAP_GR 0.000857 0.000439 1.952275 0.0548
RULE_LAW -0.002119 0.001702 -1.245465 0.2171
R-squared 0.465267 Mean dependent var 0.008751
Adjusted R -squared 0.382421 S.D. dependent var 0.007063
S.E. of regression 0.005551 Akaike info criterion -7.416807
Sum squared resid 0.002187 Schwarz criterion -7.067095
Log likelihood 319.7975 Hannan -Quinn criter. -7.276312
F-statistic 5.616045 Durbin -Watson stat 1.606384
Prob(F -statistic) 0.000002

For the ROE as a dependent variable, ROE from the previous year, NIM and GDP per capita
growth are statistically significant. The other variables are not statistically significant. The R
squared value 0. 589. Therefore, the model explains 58.9% of the variability.

Dependent Variable: ROE
Method: Least Squares
Date: 05/20/18 Time: 15:38
Sample (adjusted): 31 382
Included observations: 92 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C -0.044824 0.462521 -0.096913 0.9230
ROE( -1) 1.231967 0.168770 7.299683 0.0000
CSR 0.012489 0.013335 0.936562 0.3518
STATUS 0.004367 0.018409 0.237246 0.8131
EXTERNAL_ASSURANCE 0.035751 0.039717 0.900155 0.3707
LOG_TA 0.000518 0.044514 0.011631 0.9907
LOAN_TA 0.133030 0.219446 0.606208 0.5461
LTD -0.007905 0.057210 -0.138173 0.8905
BANK_CONC -0.169295 0.116830 -1.449078 0.1512
NIM -2.584965 1.548021 -1.669851 0.0989
GDP_CAP_GR 0.024811 0.009800 2.531628 0.0133
RULE_LAW 0.038134 0.041005 0.929993 0.3552
R-squared 0.589041 Mean dependent var 0.053163
Adjusted R -squared 0.532534 S.D. dependent var 0.200430
S.E. of regression 0.137037 Akaike info criterion -1.016025
Sum squared re sid 1.502329 Schwarz criterion -0.687096
Log likelihood 58.73716 Hannan -Quinn criter. -0.883267
F-statistic 10.42423 Durbin -Watson stat 1.474435
Prob(F -statistic) 0.000000

Conclusion
For the period before the financial cr isis and in the financial crisis, the CSR rating did not
present any significant influence over the financial performance of European banks. However,

after the financial crisis, there was a positive and significant relationship between CSR rating and
ROA. There were not so much control variables in the significant equation. One motive for this
significance in 2011 -2016, but not before, could be that banks started being more serious and
accurate in publishing CSR reports with GRI standards. For the future, I would perform a
Principal Component analysis, since it may be a problem of multicollinearity in the regressions.
It would be tried more combinations of control variables and the dependent and independent
ones.

References

Anderson, A.M. and Myers, D.H. 2007. The Cost of Being Good. SSRN Electronic Journal , doi:
10.2139/ssrn.954703.
Andrieș , A. -M. et al. 2016. The Impa ct of International Financial Crisis on Bank Performance in
Eastern and Central European Countries. EuroEconomica 35(1), pp. 111 –126.
Athanasoglou, P. et al. 2006. Determinants of Bank Profitability in the South Eastern European
Region. Available at: https ://mpra.ub.uni -muenchen.de/id/eprint/10274.
Aupperle, K.E. et al. 1985. An Empirical Examination of the Relationship between Corporate
Social Responsibility and Profitability. Academy of Management Journal 28(2), pp. 446 –463.
doi: 10.2307/256210.
Barnea, A . and Rubin, A. 2010. Corporate Social Responsibility as a Conflict Between
Shareholders. Journal of Business Ethics 97(1), pp. 71 –86. doi: 10.1007/s10551 -010-0496 -z.
Beltratti, A. and Stulz, R.M. 2012. The credit crisis around the globe: Why did some bank s
perform better? Journal of Financial Economics 105(1), pp. 1 –17. doi:
10.1016/j.jfineco.2011.12.005.
Bolt, W. et al. 2012. Bank profitability during recessions. Journal of Banking & Finance 36(9),
pp. 2552 –2564. doi: 10.1016/j.jbankfin.2012.05.011.
Căpra ru, B. and Ihnatov, I. 2014. Banks’ Profitability in Selected Central and Eastern European
Countries. Procedia Economics and Finance 16, pp. 587 –591. doi: 10.1016/S2212 –
5671(14)00844 -2.
Carnevale, C. and Mazzuca, M. 2014. Sustainability report and bank val uation: evidence from
European stock markets. Business Ethics: A European Review 23(1), pp. 69 –90. doi:
10.1111/beer.12038.
Carroll, A.B. 1979. A Three -Dimensional Conceptual Model of Corporate Performance.
Academy of Management Review 4(4), pp. 497 –505. d oi: 10.5465/amr.1979.4498296.
Cheng, B. et al. 2012. Corporate Social Responsibility and Access to Finance.
Cochran, P.L. and Wood, R.A. 1984. Corporate Social Responsibility and Financial
Performance. Academy of Management Journal 27(1), pp. 42 –56. doi: 1 0.2307/255956.
Cornett, M.M. et al. 2013. Corporate Social Responsibility and Its Impact on Financial
Performance: Investigation of the U.S. Commercial Banks. SSRN Electronic Journal . doi:
10.2139/ssrn.2333878.
Demirguc -Kunt, A. and Huizinga, H. 1999. Det erminants of Commercial Bank Interest Margins
and Profitability: Some International Evidence. The World Bank Economic Review 13(2), pp.
379–408. doi: 10.1093/wber/13.2.379.

Dhaliwal, D.S. et al. 2011. Voluntary Nonfinancial Disclosure and the Cost of Equit y Capital:
The Initiation of Corporate Social Responsibility Reporting. The Accounting Review 86(1), pp.
59–100. doi: 10.2308/accr.00000005.
Dorasamy, N. 2013. Corporate Social Responsibility and Ethical Banking for Developing
Economies.
Dennis, O. and Tai sier, Z. 2011. Efficiency and bank profitability in MENA countries. Emerging
Markets Review 12(2), pp. 94 –110. doi: 10.1016/j.ememar.2011.02.003.
El Ghoul, S. et al. 2011. Does corporate social responsibility affect the cost of capital? Journal of
Banking & Finance 35(9), pp. 2388 –2406.
Forcadell, F.J. and Aracil, E. 2017. European Banks’ Reputation for Corporate Social
Responsibility: European Banks’ Reputation. Corporate Social Responsibility and
Environmental Management 24(1), pp. 1 –14. doi: 10.1002/csr. 1402.
Freeman, R.E. 1984. Strategic management: a stakeholder approach . Boston: Pitman.
Friedman, M. and Friedman, R.D. 1982. Capitalism and freedom . Chicago: University of
Chicago Press.
Goss, A. and Roberts, G.S. 2011. The impact of corporate social resp onsibility on the cost of
bank loans. Journal of Banking & Finance 35(7), pp. 1794 –1810.
Heffernan, S.A. and Fu, X. 2010. Determinants of financial performance in Chinese banking.
Applied Financial Economics 20(20), pp. 1585 –1600. doi: 10.1080/09603107.201 0.505553.
Hsu, K. -T. 2012. The Advertising Effects of Corporate Social Responsibility on Corporate
Reputation and Brand Equity: Evidence from the Life Insurance Industry in Taiwan. Journal of
Business Ethics 109(2), pp. 189 –201. doi: 10.1007/s10551 -011-1118-0.
Islam, Z.M. et al. 2012. Corporate Social Responsibility and Financial Performance Linkage:
Evidence from the Banking Sector of Bangladesh.
Kwan, S.H. 2003. Operating performance of banks among Asian economies: An international
and time series comparison. Journal of Banking & Finance 27(3), pp. 471 –489. doi:
10.1016/S0378 -4266(02)00384 -9.
McGuire, J.B. et al. 1988. Corporate Social Responsibil ity and Firm Financial Performance.
Academy of Management Journal 31(4), pp. 854 –872. doi: 10.2307/256342.
McGuire, J.W. 1969. The Changing Nature of Business Responsibilities. The Journal of Risk and
Insurance 36(1), p. 55. doi: 10.2307/251140.
Micco, A. et al. 2007. Bank ownership and performance. Does politics matter? Journal of
Banking & Finance 31(1), pp. 219 –241. doi: 10.1016/j.jbankfin.2006.02.007.
Molyneux, P. and Thornton, J. 1992. Determinants of European bank profitability: A note.
Journal of Ban king & Finance 16(6), pp. 1173 –1178. doi: 10.1016/0378 -4266(92)90065 -8.

Moore, G. 2001. Corporate Social and Financial Performance: An Investigation in the U.K.
Supermarket Industry. Journal of Business Ethics 34(3–4), pp. 299 –315. doi:
10.1023/A:101253701 6969.
Nelling, E. and Webb, E. 2009. Corporate social responsibility and financial performance: the
“virtuous circle” revisited. Review of Quantitative Finance and Accounting 32(2), pp. 197 –209.
doi: 10.1007/s11156 -008-0090 -y.
Orlitzky, M. et al. 2003. Cor porate Social and Financial Performance: A Meta -Analysis.
Organization Studies 24(3), pp. 403 –441. doi: 10.1177/0170840603024003910.
Parsons, T. 1991. The social system . New ed. London: Routledge.
Pasiouras, F. and Kosmidou, K. 2007. Factors influencing th e profitability of domestic and
foreign commercial banks in the European Union. Research in International Business and
Finance 21(2), pp. 222 –237.
Polychronidou, P. et al. 2014. Corporate Social Responsibility in Greek Banking Sector – An
Empirical Researc h. Procedia Economics and Finance 9, pp. 193 –199. doi: 10.1016/S2212 –
5671(14)00020 -3.
Scholtens, B. and Dam, L. 2007. Cultural Values and International Differences in Business
Ethics. Journal of Business Ethics 75(3), pp. 273 –284. doi: 10.1007/s10551 -006-9252-9.
Schultz, F. et al. 2013. The Construction of Corporate Social Responsibility in Network
Societies: A Communication View. Journal of Business Ethics 115(4), pp. 681 –692. doi:
10.1007/s10551 -013-1826 -8.
Shen, C. -H. and Chang, Y. 2009. Ambition Versus Conscience, Does Corporate Social
Responsibility Pay off? The Application of Matching Methods. Journal of Business Ethics
88(S1), pp. 133 –153. doi: 10.1007/s10551 -008-9826 -9.
Shen, C. -H. and Lee, C. -C. 2006. Same Financial Development yet Different Economi c Growth –
-Why? Journal of Money, Credit, and Banking 38(7), pp. 1907 –1944. doi:
10.1353/mcb.2006.0095.
Shleifer, A. and Vishny, R.W. 1997. A Survey of Corporate Governance. The Journal of Finance
52(2), pp. 737 –783. doi: 10.1111/j.1540 -6261.1997.tb04820.x.
Sufian, F. and Habibullah, M.S. 2012. Globalizations and bank performance in China. Research
in International Business and Finance 26(2), pp. 221 –239. doi: 10.1016/j.ribaf.2011.12.005.
The European Commission 2002. Communication from the Commission concer ning Corporate
Social Responsibility: A business contribution to Sustainable Development. Brussels. Available
at: http://trade.ec.europa.eu/doclib/docs/2006/february/tradoc_127374.pdf.
Tsoutsoura, M. 2004. Corporate Social Responsibility and Financial Perf ormance. Available at:
https://escholarship.org/uc/item/111799p2.
Vilanova, M. et al. 2009. Exploring the Nature of the Relationship Between CSR and
Competitiveness. Journal of Business Ethics 87(S1), pp. 57 –69. doi: 10.1007/s10551 -008-9812 –
2.

Waddock, S.A . and Graves, S.B. 1997. The Corporate Social Performance -Financial
Performance Link. Strategic Management Journal 18(4), pp. 303 –319. doi: 10.1002/(SICI)1097 –
0266(199704)18:4<303::AID -SMJ869>3.0.CO;2 -G.
Wood, D.J. 1991. Corporate Social Performance Revisi ted. Academy of Management Review
16(4), pp. 691 –718. doi: 10.5465/amr.1991.4279616.
Wu, M. -W. and Shen, C. -H. 2013. Corporate social responsibility in the banking industry:
Motives and financial performance. Journal of Banking & Finance 37(9), pp. 3529 –3547. doi:
10.1016/j.jbankfin.2013.04.023.

Appendices

Appendix 1. Number of banks in the sample by the country
Countries No. Banks
Austria 3
Belgium 1
Cyprus 2
Czech Republic 1
Denmark 3
Finland 1
France 4
Germany 2
Hungary 1
Italy 8
Netherlands 1
Norway 3
Poland 5
Russian Federation 2
Spain 6
Sweden 4
Switzerland 7
Turkey 7
United Kingdom of Great Britain and Northern Ireland 5
Total Banks 66

Appendix 2. Correlation matrix of the variables in the model

ROA ROE NIM Capital Risk Size Bank
Concen tration GDP Inflation CSR External
assurance
ROA 1.0000
ROE 0.7702 1.0000
NIM 0.0107 -0.1176 1.0000
Capital -0.1963 -0.2239 0.0742 1.0000
Risk -0.3616 -0.3898 0.2343 0.1646 1.0000
Size -0.4979 -0.2990 0.0916 0.0468 0.0005 1.0000
Bank
Concentration 0.1210 0.0893 -0.5991 0.0057 -0.3597 -0.1929 1.0000
GDP 0.2093 0.2436 0.0736 -0.0618 0.0856 0.0077 0.2253 1.0000
Inflation 0.0318 -0.0995 0.4682 -0.229 0 0.0687 0.1674 -0.4769 0.1124 1.0000
CSR -0.3235 -0.1664 -0.1618 0.1486 0.0508 0.3460 0.1221 –
0.0129 -0.1332 1.0000
External
assurance -0.3144 -0.2593 -0.0202 0.1334 0.2340 0.1976 -0.0031 –
0.1465 0.0138 0.5888 1.0000