1 Introduction

Corporate social responsibility (CSR) is a dramatically growing area of interest for both academicians and practitioners. Its origins, however, are not recent and date back to early 18th century when environmental concerns started to surface as significant business issues (Smith 1993). Nevertheless, there is still an ongoing debate about what exactly motivates firms to engage in CSR (Holder-Webb et al. 2009). Friedman’s thesis (1970, 1961) that the “social responsibility of business is to increase its profits” does not rescind the essence of CSR that explains that businesses should promote common wealth “even at the expense of their own profitability” (Maitland 2012: 223). From this perspective, CSR could be defined as “corporate behaviors that aims to affect stakeholders positively and that go beyond its economic interest” (Turker 2009: 413); it also refers to the general belief that contemporary businesses are active members of society (Lynn 1992) and have an obligation to it that extends beyond the stockholders or investors in the firm (Carroll 2012: 2).

The extent of fulfillment of such social obligations is usually relayed to society in many different ways (i.e. annual reports of corporations, newsletters, press release, the web, etc.) through corporate social disclosures (CSD). Broadly defined as the “manifestation of the practice by which organizations communicate their social and environmental impacts and responsibilities to different stakeholders” (Menassa 2010: 5), the quality and extent (quantity) of these disclosures differ across different industries and according to the characteristics of the disclosing party as argued by Branco and Rodrigues (2008). Accordingly, the banking industry, a customer-oriented sector, has social responsibility priorities in financing sustainable development of economic activities (Relano and Paulet 2012; Prior and Argandoña 2009; Scholtens 2009; Hermes et al. 2005; Murdoch 1999) through socially responsible investment and lending policies (Ullah et al. 2014; Simpson and Kohers 2002). The impact of such policies is of public interest since the banking industry has high public visibility (Menassa 2010; Branco and Rodrigues 2008; Miles 1987) and its services are considered as public goods (Miles 1987).

Socially responsible banking is therefore developing into a behavior of a strategic nature. The latest financial scandals have made obvious the vulnerability of this sector to moral hazard (Hortacsu and Gunay 2008). As a consequence, the international banking industry has responded with improved social responsibility levels. Studies of CSD related to this sector have therefore multiplied in number and coverage. Scholtens (2009) provided a framework to assess CSR of international banks based on publicly available information and found significant differences among countries and individual banks; Menassa (2010) considered the case of a developing economy and showed that even small family-owned banks with high public visibility can exhibit strong social and ethical awareness, and Branco and Rodrigues (2008) noted that higher visibility banks attribute more importance to CSD. Previous studies by El-Bannany (2007), Abdul Hamid (2004), Douglas, Doris, and Johnson (2004) and others investigated the main determinants and types of CSD of banks. El-Bannany (2007) reported that risk factors, market structure and investment in information technology have significant effect on CSR behavior in the UK, whereas Abdul Hamid (2004) observed that firm size and listing status are possible explanatory variables of this phenomenon. Similar to these studies, the chief objective of this paper is to investigate the type and quantity of CSD by German banks belonging to three different categories. The analysis contained herein assesses the potential relationships between these disclosures and a number of banks’ characteristics, in particular, size, age, financial performance, corporate form, listing status, and overseas presence. Only few studies have addressed CSD of German banks. Chief amongst these is Relano and Paulet’s (2012) typology of the German case. Using the case of three German banks, they claimed that banks attitude towards corporate social responsibility differs according to three distinct business models: banks whose social approach is based on what they say; banks whose social approach is based on what they are; and banks whose social approach is based on what they do. Another notable contribution by Herzig et al. (2012) who investigated the sustainability reporting practice of the ten largest German banks during the latest financial crisis and concluded that four different approaches to sustainability exist. Both studies considered a small number of observations and were directed towards assessing specific cases at specific times. In contrast, this paper adopts a generalization approach by assessing the CSD of around 10 % of the total number of banks operating in Germany. It also goes beyond the traditional body of literature in many other respects, in particular:

  • This study represents a first attempt to investigate if the particular corporate form of a bank has an effect on the type and quantity of CSD.

  • Furthermore, the relationship between CSD and the number of apprentices is analyzed as a new original variable. The involvement in apprenticeship could be regarded as a display of sustainable management against the background of the demographic change in Germany and hence could be an indicator of a larger quantity of CSD in general.

The German banking industry is an advanced sector contributing to the general welfare of the German economy. Despite its structural strengths and solid attributes, it had to deal lately with several financial uncertainties and increasing competition. In addition to the traditional strategic actions that a commercial bank could take to deal with these prevailing market conditions, CSR could be used as a tool to boost banks’ position in the market place. To do that, policy makers need to thoroughly understand the current level of CSD by banks, a necessary step to plan their future moves. To our knowledge, empirical studies examining the current CSD by German banks are rare, if any. Therefore, this study endeavors to fill this gap.

This paper is structured as follows: the following section provides a discussion of the different theoretical frameworks used to explain CSD. The third section poses the hypotheses to be tested, whereas section four presents a discussion of the methodology adopted in this paper and features the qualities of the sample used. This is followed by an interpretation of the findings. The last section presents some concluding remarks.

2 Theoretical foundations

CSR studies have proliferated as scholars have increasingly delved into describing its practices and consequences. Motivated by a line of inquiry seeking to appraise what “catalyzes organizations to engage in increasingly robust CSR initiatives”, Aguilera et al. (2007: 837) proposed a multilevel theoretical framework to examine CSR at the supra (transnational), macro (country), meso (organizational), and micro (individual) level, and observed that stakeholders of an organization, at these four different levels, have three main drives for pressuring the business to engage in CSR: instrumental motives which are based on self-interest, relational motives that are concerned with relationships among group members, and morality-based motives, concerned with “ethical standards and moral principles” (p. 839). From a similar perspective, Gray et al. (1995a) noted that there are three basic although somewhat overlapping groups of theories which are used to explain CSD: decision-usefulness theories, economic theory, and socio-political theory.

The decision usefulness approach argues that organizations disclose information that users find useful for investment purposes (Tilt and Symes 1999). In fact, earlier ranking studies by Firth (1984) and Benjamin and Stanga (1977) observed that the financial community (analysts, investors, bankers…) ranks CSD higher than other accounting information in order of perceived importance. Other studies (Holman et al. 1985; Aupperle 1984; Belkaoui 1980) attempted to investigate how CSD are used by stock market participants. The results of these studies remain inconclusive (Gray et al. 1995a) due to the difficulty of measuring CSD usefulness (Dierkes and Antal 1985) and because “interest in CSR is not motivated predominantly by a concern with the needs, wants and whims of financial participants” (Gray et al. 1995a: 51).

The economic theory overlaps with the decision-usefulness approach. The central assumption here is that a morally degenerate form of short-term self-interest motivates all actions and hence the disclosure (Gray et al. 1995a). From this perspective, “managers will disclose social information only if it increases their welfare, that is, when the benefits from disclosure outweigh the associated costs” or leads to a reduction of current and future agency costs (Ness and Mirza 1991: 212). According to Gray et al. (1995a), this proposition is both empirically implausible and highly offensive and has little or nothing to offer as a basis for the development of CSD.

While the decision-usefulness/economic angle did not provide consistent and plausible results, a different theoretical perspective, the socio-political angle, provided a better explanation of CSD (Tilt and Symes 1999). Central to this angle are the political economy theory, the stakeholder theory and the legitimacy theory. Although minor differences, related to the importance of the major societal variables that have an impact on the CSR activities of firms, exist between these theories (Waller and Lanis 2009), their central focus is the relationship between the firm and the society. While the political economy theory identifies various societal institutions, in particular the government, as having the biggest impact on the CSR activities of a company (Gray et al. 1995a), the other two theories—stakeholder and legitimacy—perceive that direct stakeholders of a firm exert more influence on its CSR activities (Waller and Lanis 2009), and therefore, organizations seek legitimization from these groups as well as society’s approval of their social strategies and activities to ensure the continuity of their business (Menassa 2010; Gray et al. 1995a); consequently, through CSD, firms prove that their activities are congruent with the values of society (Branco and Rodrigues 2008) and demonstrate that they fulfill the social contract (Waller and Lanis 2009; Patten 1991).

Notwithstanding the contribution of each of the mentioned theories in explaining CSD, this paper adopts a socio-political approach to examine the type and quantity of disclosures of German banks, and thus, is grounded in the social/legitimacy theory.

3 Hypotheses and operationalization

Deductive in nature, this paper aims to identify and categorize the types of CSR information disclosed by German banks and to report on the quantity of these disclosures and their relationship with the following variables: size, financial performance, corporate form, age, number of apprentices, overseas presence, and listing status. To the best knowledge of the authors, many of the variables, in particular the corporate form and the number of apprentices have not been tested before.

3.1 Bank size

Studies attempting to relate CSD to the size of the firm are still inconclusive (Menassa 2010). While scholars such as Menassa (2010), Branco and Rodrigues (2008), El-Bannany (2007), Belkaoui and Karpik (1989), Adams et al. (1998) and Choi (1999) observed a significant relationship between size and CSD, other studies by Roberts (1992), Singh and Ahuja (1983), Davey (1982) and Ng (1985, cited in Abdul Hamid 2004) found no relationship between these variables. According to proponents of this apparent relationship, larger corporations have more public visibility and tend to disclose more social responsibility information than smaller firms (Branco and Rodrigues 2008; El-Bannany 2007). Through these disclosures, visible firms attempt to minimize or avoid governmental interventions and legitimize their existence. Similar to these studies, this paper hypothesizes the following:

Hypothesis 1

The quantity of CSD of German universal banks is positively related to the size of the bank (proxied by total assets, number of branches, number of employees, and total shareholders’ equity).

3.2 Financial performance

The question of whether it is in a firm’s financial best interest to undertake CSR activities has been a long-running debate (McWilliams and Siegel 2000; Roman et al. 1999). Using a meta-analytic approach, Orlitzky et al. (2003) provided evidence of a significant positive effect of CSR activities of a firm on its financial performance. Nevertheless, findings of studies attempting to investigate the relationship between CSD and financial performance proxies (net profits, return on equity, and return on assets) are still inconclusive (Menassa 2010). Menassa (2010) attributed this inconsistency of results to a distinction between short-term and long-term corporate profitability. From this perspective, Abdul Hamid (2004) and Belkaoui and Karpik (1989) observed no apparent association between CSD and short-term profitability whereas Hackston and Milne (1996) and Patten (1991) provided evidence of a significant association between CSD and short-term profitability proxies. Similarly, Roberts (1992) observed a relationship with lagged profits, whereas Hackston and Milne (1996) found no evidence of a significant relationship between CSD and long-term profitability measurements:

Hypothesis 2

The quantity of CSD of German universal banks is positively related to the financial performance of banks (proxied by “short-term” net profit and ROE).

3.3 Bank age

Abdul Hamid (2004: 121) observed that the legitimacy theory predicts that “the age of a corporation is related to its reputation in the society and its history of involvement in CSD activities”. From this perspective, corporations “strive to develop good reputation throughout their years of existence” (Menassa 2010: 9). Nevertheless, the research results concerning the relationship between the age of a firm and its quantity of CSD are confusing. Whereas Choi (1999) and Abdul Hamid (2004) provided evidence of a significant relationship, El-Bannany (2007) and Menassa (2010) could only note a weak association between CSD and the bank age. In the present study, the age of the bank is measured by the number of years since establishment until the end of the examined fiscal year, in this case 2010.

Hypothesis 3

The quantity of CSD of German universal banks is positively related to the age of the bank.

3.4 Number of apprentices

From a different perspective, the relationship between CSD and the number of apprentices is analyzed as a new original variable. Apprenticeship in Germany is governed by The Vocational Training Act first adopted in 1969 and amended later. The resulting apprenticeship system (Dual System) was structured and organized in the 1970s and had received many improvements since then (Tremblay and Le Bot 2003). Investigating why German commercial banks invest in apprenticeship, Finegold and Wagner (2002) concluded the existence of two different logics leading to such investments: the first argues that apprenticeships help reducing costs related to recruitment and training of new employees (the logic of consequences), whereas the second logic (logic of appropriateness) explains that such behavior fulfills societal exigencies and supports “the strong reputational effect associated with training” (Finegold and Wagner 2002: 667). For them, these two logics are best seen “as complementary perspectives in the case of German banks apprenticeship” (p. 670). From a similar perspective, Backes-Gellner and Turo (2010) observed that apprentices training programs significantly improves recruitment success and signals “good long-term career prospects” (p. 273), thus contributing to social welfare. From this angle, Vives (2013) noted that contributing to entrepreneurship and reducing youth unemployment through offering apprenticeship programs, are part of business enterprises’ responsibility towards the society. It then becomes apparent that the involvement in apprenticeship could be regarded as a display of sustainable management practices against the background of the demographic change in Germany and hence could be an indicator for a higher volume (quantity) of CSD in general. This variable is measured as the percentage of apprentices to the total number of employees.

Hypothesis 4

The quantity of CSD of German universal banks is positively related to the number of apprentices employed by the bank (proxied by the percentage of apprentices to total number of employees).

3.5 Bank category and corporate form

According to the German Central Bank (2011a), deposit-taking banks in Germany are grouped under two main categories: universal banks and specialized commercial banks (1751 and 60 banks respectively excluding branches of foreign banks in Germany). This study is concerned with the quantity and type of CSD of universal banks and thus the specialized commercial banks are not included. From this perspective, this paper considers three different categories of universal banks: credit banks (including high street and regional banks), saving banks, and cooperative banks. This categorization is of prime interest for this study because the attitude of banks towards sustainable finance and corporate responsibility is not the same (Relano and Paulet 2012). For instance, commercial credit banks are different from cooperative banks. Because of their particular type, the latter emphasize “their legal status and their commitment to the values of the social economy” (Relano and Paulet 2012: 380). To the best knowledge of the authors, this study represents the first attempt to investigate if the particular type of a bank (and accordingly the corporate form) has an effect on the quantity of CSD.

Hypothesis 5

German universal banks of different categories/corporate forms exhibit different levels of CSD.

3.6 Bank international exposure

Corporations operating in global markets should be aware of the expectations of the society in which they operate (Adhikari et al. 2005). Consequently, “bank international exposure increases its visibility in both local and foreign markets” as argued by Menassa (2010: 10). Thus, similarly to bank size, this variable is related to the legitimacy theory (public visibility) and is tested as follows:

Hypothesis 6

German universal banks with overseas presence provide more CSD (quantity) than those with no overseas presence.

3.7 Bank listing status

The listing status of corporations is commonly associated with the corporation’s ability to access external financing opportunities (Menassa 2010). A good relationship between the firm and its stakeholders is therefore of prime importance. From this perspective, CSD can be used to improve public confidence in the cooperation, noted Zéghal and Ahmed (1990).

Previous studies attempting to relate this variable to the quantity of CSD generated inconsistent results. While El-Bannany (2007) and Menassa (2010) provided evidence of an insignificant association with the level of CSD (UK and Lebanon), other scholars such as Teoh and Thong (1984) and Abdul Hamid (2004) detected a significant relationship in the Malaysian market. Consistent with these studies, this paper hypothesizes that:

Hypothesis 7

German listed universal banks provide more CSD (quantity) than German non-listed universal banks.

4 Research design and strategies

The following paragraphs outline the main characteristics of the German banking industry and discuss the sampling procedures, the data collection tools as well as the measures used to test the hypotheses outlined previously.

4.1 The German banking sector

The German banking system consists of the German Central Bank (GCB) and 1919 commercial banks divided as follows (as at 31 December 2010): 1751 universal banks (excluding branches of foreign banks in Germany, #108) and 60 specialized commercial banks, depending on the scope of their business; universal banks (including branches of foreign banks) represent around 97 % of all German commercial banks and offers all kinds of banking services. The remaining 3 % are specialized banks like building societies and real-estate credit institutions. An on-going consolidation process coupled with increasing competitive pressure by direct and online banking, as well as recent plans of the European Commission to simplify significantly the shift for the customers from one bank to another (Anonymous 2012) reduced the total number of commercial banks by around 29 % between 2000 and 2011 (Tables 1 and 2).

Table 1 Business models of sampled banks
Full size table
Table 2 Sample composition
Full size table

Universal banks, central to this paper, are divided into three subcategories: credit banks, saving banks, and cooperative banks. The group of credit banks includes high street banks and smaller regional banks and other credit banks, as well as foreign banks with branches in Germany. It consolidates the operations of the four big players, namely Commerzbank AG, Deutsche Bank AG, Deutsche Postbank AG, and UniCredit Bank AG, and is regarded as the most heterogeneous group of all in terms of corporate form (AG and GmbH). The second group, saving banks, dates back to the 18th century with the establishment of Sparkasse Salem-Heiligenberg in Baden-Wuerttemberg in 1749. Later, the concept of public saving banks was born and is currently the dominating banking form in this group. Accordingly, more than 98 % of the banks included in this category have the corporate form of AöR (public institution) whose owners are cities municipalities, communities and administrative districts. In 2005, the ‘public guarantor liability’ and ‘maintenance liability of municipalities’ were cancelled due to request of the European Commission, which regarded these guarantees as a competitive disadvantage for all other banks in Germany (DSGV 2010a). Besides their alignment as universal banks providing all common services to individual and commercial customers, they usually have a special focus on small and medium sized companies and a strong regional attachment as well as a public welfare orientation (DSGV 2010b). The third group, cooperative banks, dates back to mid—19th Century (BVR 2012a) and are owned mainly by their members. Their main aim is not to make profits but to meet the objectives and interests of their members. The majority of institutions belonging to this group have the German corporate form eG (registered cooperative). Traditionally, cooperative banks are rooted in their particular region and offer, similarly to universal banks, a complete range of financial services (BVR 2012b).

4.2 Sample and sampling procedures

According to the register of the GCB, 1919 banks operated in Germany as at 31 December 2010 (German Central Bank 2011b). The population of interest for this study consists of 1739, thus excluding 108 branches of foreign banks, the ten Landesbanken, two regional institutions of credit cooperatives, and 60 other mortgage, building associations and special purpose banks (18, 24, and 18 respectively), which are not of interest to this analysis. The following table represents the sample composition:

The determining factor for the sample size was the size of the population related to the credit banks group. The homepages of all 172 banks belonging to this group were checked for the 2010 fiscal year annual reports. Only 58 were available of which 50 were selected (checked in 2012; eight banks which registered losses in 2010, including Commerzbank AG, were excluded to facilitate the logarithmic transformations needed for the multivariate regression model). Accordingly, the other two categories were set at near levels for comparison purposes and contained 60 saving and 59 cooperative banks which were selected via the random number function in Microsoft Excel. Hence, the final sample consisted of 169 universal banks representing around 10 % of the population in terms of numbers and accounted for 66.3 % of total assets of the population (1739 universal banks; German Central Bank 2011b).

4.2.1 Measuring CSR disclosure

Although the initial intent of the authors of CSD might not be apparent to stakeholders (Dion 2012), the quantity of disclosures related to a specific topic can be used as an indicator of its importance and weight (Menassa 2010; Krippendorff 2004; Deegan et al. 2002; Gray et al. 1995b; Zéghal and Ahmed 1990). From this perspective, content analysis, defined as a technique consisting of categorizing qualitative information using a set of codes in order to derive measures that could be used quantitatively (Abbott and Monsen 1979), appears as a suitable technique to appraise the quantity and significance of CSD. Dominant in CSD studies (see for example Menassa 2010; Branco and Rodrigues 2008; Holland and Foo 2003; Gray et al. 2001; Gray et al. 1995a), this technique makes replicable and valid inferences from data to their context (Krippendorff 2004) through a planned process which entails “the selection of the content/document to be analyzed, the determination of the selection criteria and measurement unit, and the codification of the text as well as the implementation of appropriate measures to enhance validity and reliability” (Menassa 2010: 10). Consistent with prior research (Menassa 2010; Branco and Rodrigues 2008; Gray et al. 1995a), this study uses the 2010 annual reports of German universal banks. Notwithstanding the importance of other media which may be used for disclosure (i.e. press news, internet, etc.), annual reports are fundamental documents dealing with the organization’s affairs as a whole (Gray et al. 2001) and enjoy a high level of credibility and authority not associated with other media (Guthrie and Parker 1989; Tilt and Symes 1999; Neu et al. 1998).

Following Menassa (2010) and Branco and Rodrigues (2008), the variety of CSR disclosures were categorized according to a set of codes related to environmental, human resource, product and consumer, and community involvement disclosures. Each of these categories included a number of elements as detailed in Appendix 1.

The question related to the measurement unit has long been debated (Abdul Hamid 2004; Unerman 2000). Three units of measures are usually used by this type of studies: number of pages (i.e. Gray et al. 1995b), number of sentences (i.e. Hackston and Milne 1996), and number of words (i.e. Zéghal and Ahmed 1990). While pages are viewed as the easiest measure for determining the quantity of CSD (Guthrie and Parker 1989), and reflect “the amount of total space given to a topic, and, by inference, the importance of that topic” (Gray et al. 1995b: 84), they might lead to bias due to different formatting of the annual reports and the use of graphics. In contrast, words, being the ‘smallest’ of these units allow the logging of disclosure in greater detail (Unerman 2000; Zéghal and Ahmed 1990). Nevertheless, the exclusive use of words must be critically approached as the interpretation of individual words may lead to meanings that are out of the examined context (El-Bannany 2007; Holland and Foo 2003; Unerman 2000). Sentences as a measurement unit overcome these problems (El-Bannany 2007; Hackston and Milne 1996) and “provide complete, reliable and meaningful data for further analysis” argued Milne and Adler (1999: 237). Motivated by this discussion and consistent with similar studies (Menassa 2010; Maali et al. 2006; Hackston and Milne 1996), this study uses the sentence count as the main unit of analysis to locate and examine the type and quantity of social disclosure provided in the annual reports of German universal banks.

The final step of this process of content analysis consisted of developing a checklist instrument (social disclosure index). Similar to Milne and Adler (1999), a single coder, who passed a sufficient period of training, coded the data. To account for the effect of the learning cycle, a second run of coding of all annual reports was performed by the same coder to certify the reliability of the results. Found discrepancies were then reconciled and resolved. Although not used as measurement units by this study, word count took place at the same time as sentence count.

4.3 The model

To test hypotheses one to four related to the association between the quantity of CSD and a selected number of variables, the following multivariate regression model is set out below:

$$\begin{aligned} {\text{CSD}}~ & = ~\alpha ~ + ~\beta _{1} {\text{SIZE}}_{{({\text{ASSETS}},\,{\text{ EQUITY}},\,{\text{ EMPLOYEES}},\,{\text{ BRANCHES}})}} ~ + ~\beta _{2} {\text{FINPER}}_{{({\text{PROFITS}},\,{\text{ ROE}})}} ~ \\ & \quad + ~\beta _{3} {\text{AGE}}~ + ~\beta _{4} {\text{APPRENTICE~}} + ~\varepsilon \\ \end{aligned}$$

where:

  • α = the intercept;

  • β = the regression coefficients;

  • ε = the error term;

  • CSD = quantity of corporate social disclosure by German universal banks (number of sentences);

  • SIZE = size of the bank measured by log10 total assets, log10 total equity, log10 total number of employees, and log10 total number of branches;

  • FINPER = financial performance of the bank measured by log10 net profits and square root ROE;

  • AGE = square root age of the bank;

  • APPRENTICE = square root percentage of apprentices.

Although the size and financial performance of the banks are measured by more than one proxy each, the aim is to test the appropriateness of each measure. In this respect, a stepwise regression is performed to account for issues of multi-collinearity.

Hypotheses five to seven are tested using nonparametric tests (Kruskal–Wallis H-test and Mann–Whitney U-test), thus accounting for concerns related to the unevenness of the number of observations (listing status of the banks and international presence; please refer to Table 3). Moreover, the corporate forms of the banks included four groups (nominal variables) which necessitate an extra N-1 dummy variables (three) to be represented in the ordinary least square regression model, thus the decision to adopt nonparametric tests instead.

Table 3 Description of the sample
Full size table

5 Analysis

The analysis of the data follows. First, descriptive analysis is used to describe the current situation related to CSD by German universal banks, followed by a test of the hypotheses declared in previous paragraphs.

5.1 Descriptive analysis

Table 3 below shows a detailed description of the sample:

The predominant form of saving banks is AöR (public law) whereas cooperative banks have an eG form. The credit banks group, consisting of three high street banks and 47 regional banks, have either the AG or the GmbH forms. Of these, only nine banks are listed on the stock market. Furthermore, 16 banks in total have branches outside Germany, 15 of them are credit banks.

Following Menassa (2010) and Gray et al. (2001), Table 4 exhibits the detailed descriptive statistics of the sample. Significant variability in most of the figures, as denoted by the standard deviation computation and the gaps between the minimum and the maximum values of tested variables, can be observed. This is coupled with positively skewed results with large Kurtosis coefficients (except for Apprentices), indicating that these variables are not normally distributed. Consequently, a log10 transformation was performed for all the variables. This resulted in acceptable skewness and kurtosis coefficients except for the Age and ROE variables. Another square root transformation was performed for these two variables which solved the problem. To illustrate, the average skewness and kurtosis coefficients improved from 6.69 and 70.89 respectively before transformation to 0.17 and 1.86 after transformation (refer to Appendix 2 for more details).

Table 4 Characteristics of the sample (untransformed)
Full size table

Tables 5 and 6 report on the type and quantity of social disclosures. Panel A shows that 82 % of the sample disclosed at least three of the four categories of CSD in their reports and 41 % of the sample (69 banks) disclosed information related to all four categories. Interestingly, the majority of saving banks and cooperative banks disclosed at least three of the four CSD categories (92 % of saving banks and 95 % of cooperative banks compared to only 54 % of the credit banks). However, there was no bank in the sample, which disclosed no CSD at all. It is noteworthy that 93 % of the saving banks and 86 % of the cooperative banks disclosed information regarding their community involvement. In comparison, only 50 % of credit banks disclosed this category. This is consistent with the mission of the different types of banks in particular the close regional attachment of saving and cooperative banks. Nevertheless, consistent with similar studies, environmental disclosure records the lowest score possibly due to a perception that banking operations have insignificant impact on the environment compared with the other sectors of the economy (Menassa 2010; Branco and Rodrigues 2008; Simpson and Kohers 2002). A closer look at the type of category disclosed by the banks shows that ‘human resources’ (99 % of all banks) closely followed by ‘product and consumers’ (96 % of all banks) were widely covered by the sampled banks. The third rank is occupied by the category ‘community’ with 78 %, whereas ‘environmental’ with 45 % comes in last. Results produced in Table 6 corroborate these findings. From a different perspective, and although not directly related to the quantity and type of CSD, a scan of the location of CSD in the annual reports reveals no apparent preference for the location when the sample is taken as a whole. Nevertheless, a few divergences are noticed when the different groups are examined separately (Appendix 3 provides a detailed view of the location of CSD).

Table 5 Type of CSD—categories disclosed by banks
Full size table
Table 6 Quantity of CSD per type and bank category
Full size table

5.2 Testing the hypotheses

Using the stepwise method, the final regression model includes four significant independent variables as shown in Table 7 below. The tolerance statistics and variance inflation factors (VIF) calculated for these remaining variables indicate low levels of multi-collinearity (Tolerance > 0.2 and VIF < 10), thus the variances of the estimated regression coefficients were not significantly increased because of collinearity (Neter et al. 1989).

Table 7 Results of the stepwise multivariate regression and collinearity diagnostics
Full size table

Table 7 indicates significant positive associations between the CSD quantity (sentence count) of German universal banks of all categories and a range of variables, namely the total number of employees (a size proxy), return on equity (a financial performance proxy), and percentage of apprentices to total number of employees. Table 7 also shows that the age of banks is negatively related to the quantity of CSD (relatively weak association compared with the other significant variables: p value = 0.044). Nevertheless, when considered separately, no single bank category exhibits a significant association between the age and the quantity of CSD (refer to Appendix 4). Consistent with Menassa (2010) and Abdul Hamid (2004), it appears that the age does not appear to be an influential variable (H3).

With respect to size variables (H1), only the coefficients of the total number of employees as a size proxy were highly significant for all the banks taken together. However, a closer look at the size proxies (Appendix 4) shows that the total number of employees was only significant for the credit banks category (p value = 0.000) Interestingly, the same Appendix indicates that assets show a high significant positive association with the quantity of CSD of cooperative banks, whereas only saving banks did not show any significant relationship with size measures. To complement these findings, the relationship between the quantity of CSD by thematic categories and size measures was assessed via Spearman Rho correlations (untransformed data—Panel A of Appendix 5). A strong positive association emerges with the number of branches and community involvement displaying the highest correlation coefficient. In brief, consistent with the findings of Menassa (2010), Branco and Rodrigues (2008) and Abdul Hamid (2004), larger banks with higher public visibility disclose more social information than smaller ones “as part of their reputation management strategies” (Branco and Rodrigues 2008: 175).

With respect to the relationship with financial performance measures (H2), Table 7 reveals that only ROE produced significant coefficients indicating a strong positive relationship with the quantity of CSD of all banks taken together (N = 169). Again, when considered separately, the different bank categories exhibited different levels of association with the tested financial proxies. From this angle, Appendix 5 shows that short-term net profits are significantly associated with the quantity of disclosure of saving banks, whereas this association is not significant in the case of credit and cooperative banks. In the case of cooperative banks, this result might be due to their business model aiming at improving the wellbeing of their members rather than making profits. Similarly to size variables, Spearman Rho correlations between the quantity of CSD by thematic categories and financial performance proxies were performed. Panel B of Appendix 5 shows that profits are strongly correlated with all four CSD themes, in particular products and consumers, while environmental disclosure exhibits the weakest association (although significant) as environmental issues can be regarded as peripheral concerns for banks (Branco and Rodrigues 2008; Simpson and Kohers 2002). In brief, findings pertaining to financial performance measures are slightly in agreement with those of Branco and Rodrigues (2008) who associated higher net profits with higher public visibility “which exposes banks to more intense social scrutiny” and result in “more pressure on banks to legitimize their existence and activities and to evade undesirable attention drawn by high profits through more CSD” (Menassa 2010: 16).

Interestingly, the new tested variable (H4) relating to the number of apprentices employed by German universal banks exhibits a strong positive association with the quantity of CSD (Table 7), particularly for credit banks (Appendix 4). Credit banks are commercial in nature and tend to recruit more apprentices, thus the positive linear relationship observed. Appendix 5 indicates that the found association mainly originates from community involvement disclosures of German universal banks.

Furthermore, the Kruskal–Wallis test results (Table 8 Panel A) indicate a high significant variation of disclosure levels for banks belonging to different categories or having different corporate forms (H5). Panel B reveals that the highest variation is between credit banks and saving banks, followed by weaker variations between credit and cooperative banks. No significant variation is observed between saving and cooperative banks. This is not unexpected as these two bank categories are not highly different in terms of characteristics and business models. Similarly, Panel C exhibits significant variations between GmbH and AöR, and GmbH versus eG, and a weaker but significant variation between AG and GmbH. Nevertheless, no significant variations are observed between AG and AöR, AG and eG, and AöR and eG.

Table 8 CSD quantity variation by bank category and corporate form
Full size table

Table 9 examined the hypotheses stating that banks with overseas presence and listed banks provide more CSD than banks with no overseas presence and non-listed banks. Mann–Whitney results for the whole sample suggest that there are no significant differences related to social disclosure between listed and not-listed banks. Similarly, no significant differences were observed for banks with and without overseas presence. These findings are misleading if considered in light of the descriptive statistics in Table 3. In fact, a close analysis of this table shows that only credit banks are concerned with the listing status (listed: 9 banks; not listed 41 banks) and have the largest overseas presence (overseas presence: 15 banks; no overseas presence: 35 banks). Another run of the Mann–Whitney test reveals a significant difference of CSD between listed credit banks and not-listed ones, but not for overseas presence (H6 and H7). These findings are not with the general view that internationally exposed corporations tend to disclose more CSR but are consistent with the findings of Teoh and Thong (1984) and Abdul Hamid (2004) who detected significant different disclosure for listed and not listed corporations. In spite of the concerns related to the sample (for these two variables) and based on the results of the descriptive and inferential statistics presented previously, it can be inferred that CSR is at the heart of the strategic decision of banks in Germany regardless of their listing status or overseas presence.

Table 9 CSD quantity variation—by overseas presence and listing status (Mann–Whitney U test)
Full size table

6 Conclusion

The empirical research work contained in this paper presented a comparative investigation of CSD by German universal banks and attempted to relate size, financial performance, age, and the number of apprentices to the quantity of CSD. It also tested hypotheses concerned with the different categories of banks and their legal forms, listing status and overseas presence, and their possible influence on the quantity of CSD of German universal banks. To the best knowledge of the authors, the number of apprentices and the legal form have not been considered before in the CSD context.

Grounded in legitimacy theory asserting that organizations seek to legitimate their existence through disclosing their CSR activities, multivariate regressions were performed to uncover these relationships. An interesting number of conclusions emerge. First, this research reveals that disclosing social responsibility activities is important for the German banking industry where 82 % of the examined sample disclose three or even all four categories in their reports. The highest quantity of CSD is concerned with the product and consumers category, followed by human resource. This is consistent with similar findings in the banking industry (Menassa 2010; Abdul Hamid 2004). It can also be assumed that these findings emphasize the importance of these categories for banks especially in times where the confidence of the society in banks is still damaged by the effect of the latest financial crisis (Lewis 2012).

Second, consistent with similar studies (Menassa 2010; Branco and Rodrigues 2008; Abdul Hamid 2004), this study found evidence of the relationship between the size of a bank proxied by the total number of employees and its quantity of CSD. Furthermore, the findings indicate that banks belonging to different categories have different levels of linear relationships with the quantity of CSD. This is particularly evident between credit banks and saving banks, and between credit banks and cooperative banks, which is not surprising as saving banks and cooperative banks are similar in their business orientation. They are both regional operating banks with a focus on small and medium-sized businesses and private households. One significant difference with regional operating credit banks is that saving banks as well as cooperative banks share a common market presence and product policies guided by their umbrella organizations. Hence, it can be stated that the bank category and accordingly the corporate form have an influence on the quantity of CSD.

Third, a further finding of this study is that the quantity of CSD seems to be almost independent from the bank age and overseas presence, whereas the number of apprentices, as an indicator for sustainable and socially responsible management, and the listing status show strong association with the quantity of CSD of credit and saving banks. Additionally and from a different perspective, short-term net profits (financial performance) are significantly associated with the quantity of disclosure of saving banks but not the cooperative banks, whereas ROE seems to be a suitable proxy to assess the relationship between the financial performance of German universal banks (all categories taken together) and the quantity of CSD.

Given the apparent importance of CSR reporting in conveying firms’ conduct and social performance to the different stakeholders, understanding the association between social disclosure and its possible explanatory factors is central. From this perspective, CSD can help in supporting lasting mutually beneficial relationships between the German banks and their stakeholders. As inferred earlier in this study, CSR is at the heart of the strategic decision-making of German universal banks. From that angle, the visibility of these banks becomes a prime factor. Notwithstanding the importance of the traditional visibility proxies used in this study and in earlier research e.g. age, assets, employees, profits, etc., the analysis presented herein made use of two original factors related to apprenticeship and legal forms in an attempt to advance the debate in this context. The promising findings related to these two factors transcend the German banking setting. Indeed, further longitudinal analyses and additional tests are warranted to confirm the validity of these encouraging results and mitigate the limitations related to the timeframe used in this study.