1 Introduction and research gap

With more than 90 % of firms, family firms account for a significant part of the German economy (Klein 2000). Family firms and their behavior can have a strong impact on local communities. Litz and Stewart (2000) find that family firms are particularly involved in charity projects of the local community. Prior research shows that family firms differ from non-family firms in their corporate social responsibility (CSR) behavior. Bassanini et al. (2013) and Block (2010) further show that family firms provide safer jobs and are less likely to downsize in times of economic crisis. Berrone et al. (2010) reveal that family firms pollute the environment less. Distinguishing between different dimensions of CSR, Block and Wagner (2014) find that family ownership is negatively associated with community-related CSR performance and positively associated with diversity-, employee-, environment- and product-related aspects of CSR. So far, most studies have investigated the situation in the U.S. Apart from a study by Tänzler (2013), we know little about CSR in German family firms. Our study addresses this gap in the literature by investigating the CSR mission statements of 714 listed German family and non-family firms using content analysis.

CSR mission statements describe firms’ CSR strategy and stakeholder prioritization. Based on legitimacy (Davis 1973) and signaling theory (Spence 1973), such CSR mission statements can serve as an effective means to signal firms’ commitment to CSR to gain stakeholders’ and society’s acceptance, approval and legitimacy for their business conduct (Mahoney et al. 2012). We chose to examine CSR mission statements, because they represent a communication medium that provides short and precise information about a firm’s contribution to economic, social and environmental issues, and helps to satisfy stakeholder expectations and to gain trust (Coombs and Holladay 2011; European Commission 2011). The importance of CSR mission statements is reflected by the Global Reporting Initiative’s (GRI) recommendation for firms to provide a mission and strategy statement as part of their CSR communication. Thus, the inclusion of such a mission statement in CSR reports, annual reports, or on corporate websites has become a standard among listed firms (GRI 2011).

Four research questions are at the core of our paper: (1) Do family firms differ from non-family firms in their extent of CSR communication? (2) Can we observe differences regarding the CSR dimensions covered in CSR communication? (3) Are there differences in the stakeholder groups covered? (4) Can we observe differences in the relationship between CSR and corporate culture or corporate strategy?

Our empirical analysis shows that family firms put a particularly strong emphasis in their CSR communication on the philanthropic dimension of CSR and are more likely than non-family firms to convey CSR as a part of their corporate culture. By contrast, non-family firms are more likely than family firms to communicate CSR as part of their corporate strategy.

2 Sample and coding

Our sample was constructed from 940 firms from the manufacturing, service, and retail sectors listed in the German Prime Standard (DAX50, CDAX, MDAX, GDAX, SDAX, TecDAX, DAX plus Family). We opted for listed firms, as prior research suggests that listed firms are more likely to report their CSR activities. In particular, legitimacy theory suggests that CSR reporting is highly prominent in listed firms, because they face political and public pressure from shareholders and the society (Hackston and Milne 1996; Kolk 2003; Owen 2007; Patten 1992). As not all of the 940 sample firms published a CSR mission statement, our estimation sample reduced to 714 firms, of which 438 are family firms and 276 are non-family firms.

To investigate the CSR mission statements, we employed content analysis, which has become a common method to examine corporate social and environmental disclosure (Gao 2011; Lamnek 2010). By codifying written text into various categories (Krippendorff 2004), content analysis enables the identification of otherwise unavailable information (Kabanoff et al. 1995) and ensures high levels of reliability and replicability (Potter and Levine-Donnerstein 1999). More precisely, we employed a directed content analysis, in which we used prior research and theory to develop a coding scheme (Hsieh and Shannon 2005; Mayring 2008; Weber 1990). Our coding scheme is derived from the five CSR dimensions of Dahlsrud (2008) and the six most prominent stakeholder groups according to Freeman (1984). In addition, we developed a criterion measuring the CSR approach. Table 1 shows our coding scheme which consists of 14 categories grouped into CSR dimensions (5 categories), stakeholder groups (6 categories), and CSR approach (3 categories).

Table 1 Categories and sub-categories of the coding scheme
Full size table

The 14 categories were coded using a binary scale:

0: The CSR mission statement does not provide information on the specific category.

1: The CSR mission statement provides information on the specific category.

We employed four coders and measured intercoder reliability through Cohen’s ĸ and Krippendorff’s α. For the latter, we obtained values from 0.69 to 0.91, indicating good to very high/excellent (ĸ > 0.75) consistency in the coding (Früh 2007; Wirtz und Caspar 2002). Finally, we calculated the mean of the results from all four coders for each category (Harris 1996; Mayring 2008).

3 Econometric model

We employed univariate t-tests and ran multivariate Probit regressions to compare family and non-family firms regarding their CSR communication.

The 14 categories (Table 1) constituted the dependent variables in our Probit regressions. The independent variable of main interest refers to the family firm dummy (FAM). To distinguish between family and non-family firms, we relied on the definition of the Institut für Mittelstandsforschung Bonn (IfM Bonn). The central criterion underlying the IfM Bonn definition is the unity of ownership and management. More precisely, “up to two natural persons or their family members have at least 50 % ownership of the company and these natural persons are also involved in the management of the company” (IfM Bonn 2015). Based on IfM Bonn database of the 500 largest German family companies (IfM Bonn 2007), we categorized the firms in our sample as either family or non-family firms. For those firms not included in the IfM Bonn database of the IfM, we widened the definition and coded them as family firms when at least one member of the family is either part of the management or supervisory board. The following variables were used as independent variables (Fifka 2013; Gamerschlag et al. 2011; Hackston and Milne 1996; Reverte 2009): firm size (SIZE), financial performance (return on equity: ROE), industry membership (manufacturing: IND_MANU and retail: IND_RETAIL), family management (FAM_MGM), family member in executive or supervisory board (FAM_BOARD), and firm age (AGE). In addition, we also controlled for the effect of the length of the CSR mission statements (LENGTH).

To reduce the risk of multicollinearity in our regressions, we kept the number of variables low and calculated variance inflation factors (VIFs).

We estimated the following regression model:

$$\text{CSRcategor}{{\text{y}}_{\text{i}\,}}={{\beta }_{0}}+{{\beta }_{1}}\text{FA}{{\text{M}}_{\text{i}}}+{{\beta }_{2}}\text{InSIZ}{{\text{E}}_{\text{i}}}+{{\beta }_{3}}\text{RO}{{\text{E}}_{\text{i}}}+{{\beta }_{4}}\text{IND }\!\!\_\!\!\text{ MAN}{{\text{U}}_{\text{i}}}+{{\beta }_{5}}\text{IND }\!\!\_\!\!\text{ RETAI}{{\text{L}}_{\text{i}}}+{{\beta }_{6}}\text{FAM }\!\!\_\!\!\text{ MG}{{\text{M}}_{\text{i}}}+{{\beta }_{7}}\text{FAM }\!\!\_\!\!\text{ BOAR}{{\text{D}}_{\text{i}}}+{{\beta }_{8}}\text{InAG}{{\text{E}}_{\text{i}}}+{{\beta }_{9}}\text{LENGT}{{\text{H}}_{\text{i}}}+{{\varepsilon }_{i}}$$

where the dependent variable CSR category represents our 14 categories from the coding scheme the independent variables: FAM: family firm dummy; lnSIZE: logarithm of firm size (employees); ROE: profitability (return on equity); MAN: manufacturing sector; RET: retail sector; FAM_MGM: family member in management board; FAM_BOARD: family member in executive or supervisory board; lnAGE: logarithm of firm age; LENGTH: length of the CSR mission statement.

Table 2 shows the correlations among the independent variables in our regressions. Multicollinearity should be of low concern, as all VIFs are below 4.

Table 2 Correlations among independent variables
Full size table

4 Findings from univariate analyses

With respect to our first research question, we find that family firms provide longer CSR mission statements than non-family firms (57.54 versus 54.45 words). Yet, this difference is not statistically significant (p = 0.292).

Table 3 shows the results regarding our second research question referring to the different CSR dimensions mentioned in the CSR mission statements. We find that family and non-family firms differ in their CSR communication only with regard to the philanthropic dimension: 51.4 % of family firms report philanthropic activities in their mission statements, whereas this number is only 37.7 % for non-family firms (p < 0.001).

Table 3 Univariate analysis
Full size table

The third research question is about the CSR communication regarding the six stakeholder groups that are (directly or indirectly) affected by the firm (Freeman 1984). Table 3 shows the findings. The stakeholder groups most often mentioned by family firms in their CSR communication are employees (47.0 %), society (43.6 %), and consumers (36.5 %). Investors and the government are least often referred to (17.4 % and 8.2 %, respectively). Similarly, non-family firm also put strong emphasis on society (49.3 %), employees (47.8 %), and consumers (34.4 %). However, compared to family firms, they put significantly more focus on investors (25.4 %, p = 0.012) and the government (12.3 %, p = 0.085).

With respect to the fourth research question, we evidence that 35.2 % of the family firms refer to CSR as being part of corporate culture. This number is only 25.7 % for non-family firms and the difference between the two groups is statistically significant (p = 0.007). Another difference between family and non-family firms can be observed in the relationship between CSR and corporate strategy. We find that 31.5 % (41.7 %) of family (non-family firms) report CSR as being part of corporate strategy (p = 0.006).

5 Findings from multivariate regressions

To analyze whether our findings from the univariate analyses also hold when controlling for a number of firm characteristics, we estimated several multivariate Probit regressions. Table 4 shows three regressions using (1) the philanthropic dimension of CSR, (2) CSR as part of corporate strategy, and (3) CSR as part of corporate culture as dependent variables.

Table 4 Multivariate Probit regression results
Full size table

Supporting the findings of the univariate analyses, the results from the multivariate Probit regressions show that family firms are less likely to communicate CSR as part of corporate strategy (Coef. = − 0.465, p = 0.019) but are more likely to convey CSR as part of corporate culture (Coef = 0.425, p = 0.041). The multivariate analysis also confirms that family firms are more likely than non-family firms to mention the philanthropic dimension of CSR in their CSR communication (Coef = 0.560, p = 0.004). The multivariate regressions, however, could not confirm our univariate results about the differences between family and non-family firms regarding the importance of investors and the government as stakeholder groups in CSR communication.

As regards our control variables, we find that firms in the manufacturing sector (variable IND_MAN) are less likely than firms in other sectors to regard CSR as part of corporate culture.

6 Discussion of results

Based on different ownership structures and differences in the goals of the owners, we expected strong differences in the CSR communication of family and non-family firms. Yet, we find remarkably few differences. Our multivariate regressions showed that family and non-family firms differed in only 3 out of 14 categories of CSR communication. This is surprising as prior research suggests strong differences in how stakeholders, particularly employees, are treated in family versus non-family firms (Block 2010; Bassanini et al. 2013; Tänzler 2013). We explain these strong similarities in CSR communication by the professionalism in the planning and reporting of CSR in listed firms. Most corporate CSR communication is prepared by consultants based on industry best practices (Hartman et al. 2007) and on specific guidelines (such as the GRI guidelines). The communication of CSR efforts and stakeholder prioritization of German listed firms is also heavily influenced and adapted to global standards (Weber and Marley 2012).

Next to these ‘non-findings’, our results show that family and non-family firms differ in regard to the philanthropic dimension of CSR and in the way CSR is embedded in the firm’s culture and strategy. How can these differences in CSR communication be explained and what practical implications can be derived from them? Family firms often have a strong organizational identity, which is shaped by the business-owning family (Zellweger et al. 2010). This leads to a strong corporate culture in family firms. In case CSR matters for the business-owning family, it is likely to be deeply rooted in the firm’s corporate culture and communicated as part of the firm’s identity rather than as a competitive strategy to succeed in the market. This also explains the particular importance of the philanthropic dimension for family firms. In line with prior research of Schröder and Westerheide (2010) and Campopiano et al. (2014), we also evidence that family firms put significantly more emphasis on the reference to the philanthropic dimension of CSR in their CSR communication. Unlike other firm owners, family owners care about their reputation as firm owners (Block 2010; Deephouse and Jaskiewicz 2013). They are visible to the public and the local community and are associated with their firms. Philanthropic activities of family firms thus help family owners to build a reputation as responsible and caring owners. This reputation spillover from firm to owners does not exist in non-family firms owned by anonymous shareholders.

Future research could expand the geographic scope of this study and investigate whether our findings replicate in other country settings. Family firms play an important role for the German economy and generally have a positive public image in Germany (Block et al. 2015). Family firms market themselves against their customers and employees as family firms by building a family firm brand (Krappe et al. 2011). Corporate culture, CSR, and the care for stakeholders’ needs helps family firms to build a family firm brand. In other countries, however, family firms sometimes have a less positive or even a negative image (Morck and Yeung 2004). It can be expected that the role of CSR and the way it is used to build a positive firm image is different for family firms in such countries. Another future area of research could be to interview executives as well as internal and external stakeholders of family and non-family firms about their perceived importance of CSR. Prior research shows that sometimes a gap exists between CSR communication and CSR efforts.