1

THE MEDIATING EFFECT OF ETHICAL CODES ON THE LINK B ETWEEN FAMILY FIRMS AND THEIR SOCIAL PERFORMANCE

Cuadrado-Ballesteros, Beatriz

Universidad de Salamanca

Rodríguez-Ariza, Lázaro

Universidad de Granada

García-Sánchez, Isabel-María

Universidad de Salamanca

Martínez-Ferrero, Jennifer

Universidad de Salamanca

Área temática : h) Responsabilidad Social Corporativa

Key words : family business; social performance; sustainability; ethical codes.

2h

2

THE MEDIATING EFFECT OF ETHICAL CODES ON THE LINK B ETWEEN FAMILY

FIRMS AND THEIR SOCIAL PERFORMANCE

Resumen

Este artículo vincula conjuntamente la investigación sobre el desempeño social, los

códigos éticos y la empresa familiar. Haciendo uso de una muestra de 547 empresas

internacionales y cotizadas para el período 2002-2010, se analiza si el uso de códigos

éticos formales actúa como factor explicativo de las diferencias en el desempeño

social entre empresas familiares y no familiares. La evidencia empírica obtenida pone

de manifiesto el menor desempeño social existente en las empresas familiares, y el

efecto mediador de los códigos éticos formales en dicha relación.

Abstract

This article brings together research on social performance, codes of ethics and family

firms. Using a panel dataset composed of 547 internationally listed companies for the

period 2002–2010, we test empirically whether the use of formal ethical codes could be

a reason to explain the differences between social performance in family and non-

family firms. We empirically show that family firms tend to present a lower social

performance than non-family firms, and the use of formal ethical codes mediate such

relationship.

3

1. Introduction

Sustainability issues are beginning to play a renewed role in society, and social

consciousness is gaining weight among citizens since it represents a way of integrating

corporate business with social and environmental welfare. In this regard, social and

environmental performance (i.e. corporate social performance) could be defined as a

company’s voluntary commitment to social development and environmental

preservation, developed within the company’s social sphere, as well as a responsible

commitment to the people and social groups with whom the company interacts. It

defines the company as a set of relationships, not just between owners and managers

but also with parties or groups interested in the evolution of the company: employees,

customers, suppliers, competitors, environment, and society (Adams, 2002).

Nonetheless, companies are profit-making entities, and very few would subscribe to the

idea that they can be persuaded to commit to environmental and social policies that

benefit the community at a cost to insiders. Because corporate aims, strategies,

management forms, and governance systems differ considerably between family and

non-family firms (Haalien & Huse, 2005), their social and environmental commitment

may also differ.

In this respect, some previous studies (Déniz & Cabrera, 2005; Burak & Morante, 2007;

López-Iturriaga & López-de-Foronda, 2011) agree, pointing out that family firms tend to

have a lower social performance than non-family firms. Since family members have a

relevant amount of investment in their own companies, they tend to be more committed

to achieving the greatest possible financial return, relegating environmental and social

commitment to the background.

However, little or nothing is known about the indirect determinants that may justify such

behaviour. From a wide range of possibilities, this study focuses on the use of formal

ethical codes in family firms because: (i) on the one hand, they are a common tool of

social and environmental performance designed for the explicit details of sustainable

commitment (Agatiello, 2008; Erwin, 2011); and (ii) on the other hand, the degree of

formal mechanism varies according to family firms and non-family firms.

Concretely, this exploratory study proposes a mediating effect of formal ethical codes

in the relationship between family ownership and social performance. This effect is

tested on a sample of 547 international non-financial listed companies from different

countries for the period 2002–2010. Our findings empirically confirm four statements: (i)

family businesses tend to present a lower social performance than non-family firms; (ii)

4

ethical codes play a key role in the effectiveness of achieving a good social

performance; (iii) family firms are more likely to follow informal codes; and (iv) the lack

of formal codes could be considered a reason for the lower social performance in

family businesses. In other words, the lack of formal codes of ethics exerts an indirect

impact on the family ownership–social performance relationship.

Therefore, this article is structured as follows: firstly, we describe the previous literature

related to our study and subsequently develop arguments to propose our hypothesis;

section 3 shows the methodological aspects, such as the sample, variables, and

models for the empirical analysis (moreover, we include an annex providing an

explanation of the measures of variables employed in our analyses); the results are

presented and discussed in section 4; and, finally, the contributions, implications, and

limitations are summarized in section 5.

2. Theoretical background and hypothesis research

Considerable research has been conducted on the question of how family firms behave

and, particularly, whether they behave differently from non-family firms. Significant

differences have been identified in terms of corporate governance, leadership,

performance, and succession (e.g. McConaughy et al., 2001; Klein et al., 2005; Brenes

et al., 2011). However, the literature until now has overlooked other topics, such as

corporate social responsibility (CSR) (Benavides-Velasco et al., 2013; Materne et al.,

2013).

Although there is no universal definition of a family business, one of the most accepted

definitions is the one proposed by Chen et al. (2008), who define a family firm as a

business in which family founders continue in a top managerial position, are present on

the board, or are able to act as blockholders. This means that they have great power

and hold fundamental positions that affect the management and decision-making

processes. Through their participation in management, family members seek to ensure

the company’s survival and vitality, as well as the transmission of its legacy (Singal,

2014) and goodwill (McVey & Draho, 2005) to their descendants.

One of the important management decisions nowadays is the level of commitment to

social and environmental practices, which determines the level of corporate social

performance. Traditionally, family firms have been characterized by non-financial aims,

such as identity, reputation, longevity, and the preservation of a positive image in the

public domain (Sharma et al., 1997; Anderson & Reeb, 2003; Berrone et al., 2010). In

5

order to ensure the survival of the firm in the market, family firms can carry out actions

approved by society, with the aim of satisfying stakeholders’ demand, gaining a

positive image, and legitimating the company.

However, these socially responsible actions pose a risk in relation to solid long-term

financial performance. As Virakul et al. (2009) suggest, among the main motives for

promoting socially responsible practices, it is necessary to note economically driven

motivations. If family members run their business with a profit maximization aim, the

conflict between social and economic goals may create a dilemma for the decision

maker. For example, the risk of investing in expensive pollution prevention beyond

compliance with regulations may not be compensated for by financial gains or the firm

may never achieve a reliable cost–benefit estimate of such actions (Margolis & Walsh,

2003).

This is especially relevant to family businesses, since family members usually have

large investments in their own firms, so they may be more interested in profitability and

financial performance than in environmental issues (Burak & Morante, 2007). Most

family businesses do not think that socially responsible practices generate competitive

advantages, although some assume that they have the resources to carry them out; as

such, they view these practices as a cost and not as an opportunity (Déniz & Cabrera,

2005). Therefore, we expect that family businesses tend to be less socially responsible

than non-family firms, as Burak and Morante (2007) and López-Iturriaga and López-de-

Foronda (2011) find. However, until now, we have not identified any studies explaining

the determinants of such behaviour.

In this respect, we expect that the existence of a formal ethical code may determine the

social performance of a company. It should constitute “a distinct and formal document

containing a set of prescriptions developed by and for a company to guide present and

future behaviour on multiple issues of at least its managers and employees toward one

another, the company, external stakeholders and/or society in general” (Kaptein &

Schwartz, 2008, p. 113). Ethical codes transmit ethical values to the members of the

organization (Wotruba et al., 2001), offering them moral guides or anchors when new

and confusing situations are encountered in the workplace (Chua & Rahman, 2011)

and in decision making (Urbany, 2005).

Codes of ethics are a common tool of social and environmental performance designed

for the explicit details of sustainable commitment, affecting positively the promotion of

such behaviour (Agatiello, 2008; Erwin, 2011). Therefore, the establishment of a formal

6

code of ethics, as an effective guide (Mijatovic & Stokic, 2010), is related to a positive

impact on the perceptions about the sustainable level within companies (Adams et al.,

2001). Mijatovic and Stokic (2010) support a positive influence of transparent codes of

conduct and corporate values on all types of sustainable issues and activities. Erwin

(2011) shows a positive link between the quality of ethical codes and the probability of

being included in the ranking of the most sustainable companies.

In the case of family businesses, they are likely to have a less formal mode of

operating; thus, they tend to adopt fewer formal policies, systems, rules, etc. The

organizational culture and climate tend to be informal in family businesses; we can

even say that the business is the “lengthening shadow” of the founder family (Hollander

& Elman, 1988). Thus, family firms appear not to rely primarily on formal codes of

ethics since they are likely to operate with a lower degree of formalization (Adams et

al., 1996), showing great trust in their values without needing formal rules or codes. It is

therefore less likely that have a formal code of ethics; thus, we expect that the

existence or not of a formal code may affect the level of social performance in family

firms. According to this, we hypothesize:

“The existence of a formal code of ethics mediates the relationship between

family firms and their social performance.”

3. Method

3.1. Sample and variables

The previously mentioned hypothesis is tested on a sample composed of 547

international non-financial companies listed for the period 2002–2010. We exclude

financial firms due to the different characteristics of their equity and because they are

not comparable with non-financial firms. The sample is unbalanced, consisting of 3,075

observations obtained from 12 countries (the USA, the United Kingdom, Canada,

Germany, the Netherlands, Denmark, Finland, Sweden, Norway, France, Italy, and

Spain).

There are three main aspects to the hypothesis: (i) social performance; (ii) ethical

codes; and (iii) family businesses. Social performance is measured by an aggregate

construct that represents the level of socially responsible commitment. It is called SP

and is determined from the non-weighted sum of 20 items related to environmental

issues, human rights, and relationships with stakeholders. In order to represent the

existence of a code of ethics and its level of application, the variable EthicCode is

7

defined. It is an ordinal variable that takes values between 0 and 4, such that 0

represents the absence of a code of ethics and 4 represents an advanced ethical code,

with the highest rating. These data (SP and EthicCode) are obtained from the Ethical

Investment Research Service (EIRIS) and the process of creation is explained in depth

in the Annex. Finally, we need a variable to represent the family businesses in the

sample. We denote Family as a dummy variable that takes the value 1 if the largest

shareholder is an individual or a family with more than 10% of the votes and 0

otherwise. Information about ownership is obtained from the Thomson One Analytic

database.

In addition, to eliminate bias from the results, we consider a set of control variables

previously shown to be effective in this respect: corporate size, leverage, and market

risk. Company size (Size) is measured by the logarithm of the total assets. The level of

firm leverage (Leverage ) represents the debt or non-compliance risk measures as the

ratio of total debt to total equity. Systematic risk is measured by the beta of the market

model (Risk ). These data are also obtained from the Thomson One Analytic database.

Furthermore, the results should be controlled by regional effects (Country k), industrial

effects (Industry m), and temporal effects (Yearn). In the three cases, we use dummy

variables to represent the different countries, activity sectors, and years.

3.2. Implementation of the mediating effect

To test the mediating effect, several steps are required, according to Baron and

Kenny’s (1986) procedure. Concretely, we propose three dependency models: “first,

regressing the mediator on the independent variable; second, regressing the

dependent variable on the independent variable; and third, regressing the dependent

variable on both the independent and on mediator” (Baron & Kenny, 1986, p. 1177).

Thus, it is necessary to carry out several regression analyses to see the influence of

family ownership on social performance and the mediating role of codes of ethics.

First Step . By estimating Model 1, we identify the effect of the independent variable

(Family ) on the potential mediator variable (EthicCode ), including control variables to

avoid biasing the results (Size, Leverage, Risk, Industry, Country, and Year).

(Model 1)

8

where i ranges from company 1 to bank 547 and t takes the values of the years from

2002 to 2010. The parameters β are the estimated coefficients from the constant and

each of the explanatory variables included in Model 1.

Second Step . Now we need to determine how the independent and mediator variables

(EthicCode and Family ) impact separately on the level of social performance (SP). To

achieve this, we estimate the following Model 2a and Model 2b, also entering the

control variables.

(Model 2a)

(Model 2b)

The parameters and are the estimated coefficients from the constant and each of

the explanatory variables included in Model 2a and Model 2b, respectively.

Third Step . Finally, it is necessary to consider how the two variables, EthicCode and

Family , impact jointly on the level of social performance (SP). This step requires

additionally that “the effect of the independent variable on the dependent variable must

be less in the third step than in the second step” (Baron & Kenny, 1986, p. 1177).

Moreover, it is necessary for the mediator to be statistically significant. To test these

conditions, we use Model 3.

(Model 3)

The parameters are the estimated coefficients from the constant and each of the

explanatory variables included in Model 3.

3.3. Technique of analysis

9

The models proposed are evaluated by taking into account that the dependent

variables (EthicCode and SP) are left- and right-censored, so an appropriate estimator

must be used. In this case, the Tobit technique is suitable, since it enables us to

address particular consideration to the extreme scores.

Therefore, unlike linear models, a Tobit regression for panel data models considers the

extremities of the rating scale (0 and 2000, for SP; 0 and 4 for EthicCode ) in a special

way. In this regard, by using the maximum likelihood method, Tobit models provide

efficient, consistent estimates of coefficients, because when the likelihood function is

maximized, it incorporates information from both censored and uncensored

observations. The basic Tobit model supposes that there is a latent variable (called yit*)

that can be explained by observable variable(s) (called xit). Specifically,

yit* = β xit + εit

εit ≈ N (0, σ2)

Then, the observable variable yit is defined as

yit = yit* if yL<yit

*<yU

yit = yL if yit* ≤ yL

yit = yU if yit* ≥ yU

where yL is the lower limit of the dependent variable (0 in our case for both variables)

and yU is the upper limit.

4. Results

The descriptive statistics for the main variable in the study are summarized in Table 2,

differing between family and non-family firms. The mean value of the social

performance index for non-family firms (726.286) implies that, on average, these firms

are more likely to engage in socially responsible actions than family firms, which have a

mean value of 663.507. Regarding the codes of ethics, our results again show that

family firms are more likely to adopt informal models of conduct to promote their ethical

behaviour and their mean value of ethical codes is lower (3.182) than the mean value

for non-family businesses (3.358). With respect to the control variables, for example,

the average size of the analysed non-family firms is 8.672, and the average size of

family firms is 8.889, showing that, in general, family firms are larger than non-family

firms. Table 1 also shows the absolute and relative frequency of family and non-family

firms, a dummy variable with values between 0 and 1. In this sense, 509 observations

(16.55% of the total) belong to family firms and the remaining 5,556 observations

(83.45% of the total) belong to non-family firms.

10

- Insert Table 2 about here -

Table 3 shows the bivariate correlations between the variables used in the model. In no

cases are high values obtained for the coefficients between dependent and

independent variables or between independent variables.

- Insert Table 3 about here -

The regression results for the hypothesis proposed are summarized in Table 4, which

provides the evidence regarding the possible mediation effect of ethical codes on social

performance following the three steps of Baron and Kenny (1986).

Firstly, as previously noted in Model 1, we test the effect of the independent (Family )

and control (Size, Leverage , Risk ) variables on the mediator variable (EthicCode ).

Specifically, Family has a statistically significant negative coefficient (coef. -0.486, p. <

0.05), meaning that family firms are less likely to use formal codes of ethics than non-

family firms. Family businesses operate in a more informal environment, with a lower

degree of formalization than non-family firms (Adams et al., 1996; Duh et al., 2010).

This affects the use of rules and norms, such as ethical codes. Values and ideas in

family businesses tend to be disseminated by informal mechanisms, instead of using

formal codes and reports.

In the second step, we test separately the impact of the mediator variable (EthicCode )

and the independent variable (Family ) on the level of social performance (SP), as we

can see in Models 2a and 2b. The results are shown in Table 4. Observing the

coefficients for Model 2a, we can see a positive effect of EthicCode on SP, which is

statistically relevant at the 99% confidence level (coef. 5.795, p. < 0.01). This result

supports a positive relationship between the presence of formal ethical codes and the

level of corporate social performance, suggesting that such codes act effectively in the

promotion of corporate social responsibility, increasing the level of social performance.

Ethical codes transmit values to all members, offering a moral and behavioural guide

that is useful in confusing situations (Urbany, 2005; Chua & Rahman, 2011; Wotruba et

al., 2011). The use of ethical codes affects positively the promotion of socially

responsible commitment, according to the findings obtained by Adams et al. (2001),

Mijatovic and Stokic (2010), and Erwin (2011).

The results for Model 2b show a statistically significant negative effect of the

independent variable (Family ) on the level of social performance (SP) (coef. -5.8604,

p. < 0.05). This result is very relevant to the family business literature, since it is

11

contrary to the traditional idea that supports the higher social orientation of family

businesses. Our finding is in accordance with those obtained by Déniz and Cabrera

(2005), Burak and Morante (2007), and López-Iturriaga and López-de-Foronda (2011),

suggesting that family firms tend to present a lower responsible commitment than non-

family enterprises. Arguments to support this could be understood as being due to the

amount of investment by family members in the company. The family usually has large

investments in its own firm, so economically driven motivations (Virakul et al., 2009)

are stronger than social and environmental actions. In other words, family members

may be more interested in profitability and financial performance than in environmental

issues (Burak & Morante, 2007). Our results agree with those of Déniz and Cabrera

(2005), who posit that family firms assume that they have the resources to carry out

socially responsible practices, but they consider that the risk is higher than the

economic benefit, viewing these practices as a cost and not as an opportunity.

Finally, the third step of the mediation analysis involves the joint testing of the impact of

the mediator variable (EthicCode ) and the independent variable (Family ) on the level

of social performance (SP). As previously noted, this step requires two assumptions.

First, the relationship between the mediator (EthicCode ) and the dependent variable

(SP) must be statistically significant. Our results meet this assumption, since the

variable EthicCode has a statistically significant positive coefficient in Model 3 (coef.

5.751, p. < 0.01). Second, “the effect of the independent variable on the dependent

variable must be lower in the third step than in the second step” (Baron & Kenny, 1986,

p. 1177). Our results are also in accordance with such an assumption, since in Model

3, Family impacts negatively on SP (coef. -5.057), although it is not statistically

relevant. This coefficient is lower than that obtained in Model 2b (coef. -5.8604, p. <

0.05). This confirms the existence of a mediating effect of ethical codes on the level of

social performance of family firms. In other words, the negative influence of family

ownership on their social performance appears to be justified by the absence of formal

ethical codes.

- Insert Table 4 about here -

As a consequence of the sample size and the existence of a more rigorous and

powerful test to analyse mediation, we use an additional test (Zhao et al., 2010).

Specifically, the “bootstrapping” test supports the significance of our previous mediation

results. This test is a non-parametric resampling method that calculates the indirect

effect in each sample and provides a confidence interval. If zero is in the interval, the

indirect effect is different from zero (Shrout & Bolger, 2002; Fernández-Gago et al.,

12

2014). The 99% confidence interval of the mediation role of codes of ethics falls

between -4.770 and -0.5803. This result confirms the mediation effect of codes of

ethics at the p < 0.01 level.

Our findings are summarized in Figure 1, suggesting that family businesses present a

lower social performance than non-family firms. This supports the findings of Déniz and

Cabrera (2005), Burak and Morante (2007), and López-Iturriaga and López-de-

Foronda (2011), breaking with the traditional trend that supports a greater social and

environmental orientation of family businesses. We find the lack of formal ethical codes

to be a reason for such behaviour. Family firms are more likely to rely on role modelling

to encourage ethical behaviour by the informal transmission of behavioural norms

among members. This is in partial agreement with Adams et al. (1996), since they

observed few differences in the ways in which ethical dilemmas are dealt with by

members of family and non-family firms, suggesting that the lack of formal codes

should not be used as an indicator of corporate ethical behaviour. Nevertheless, our

empirical results suggest that the use of formal codes of ethics may influence the level

of social performance in family firms.

5. Concluding remarks

Based on a sample of 547 international non-financial listed companies for the period

2002–2010, this paper examines the mediation effect of ethical codes on the

relationship between family ownership and the degree to which the management

achieves a strong social and environmental performance. Thus, in order to evaluate the

indirect link between family firms and their social performance through ethical codes,

we analyse the relationship between codes of ethics and social performance and the

existence of ethical codes in family firms. Finally, we examine whether such ethical

codes mediate the relationship between family ownership and social performance.

Our findings support the mediating effect of ethical codes on the association between

family-owned firms and their social performance. This suggests that family firms are

more shareholder-oriented and more committed to profit maximization of their

investments, relegating sustainable actions to the background. Regarding the

mediation, the results show that codes of ethics exert an indirect effect on the negative

association between family firms and their social performance due to, on the one hand,

formal ethical codes acting as an effective tool for the promotion of a greater corporate

social performance, but on the other hand, family firms being characterized by a lower

degree of formalization than non-family firms, leading them to rely on informal codes

based on family core values.

13

This study makes the following contributions to our knowledge of family firms, ethical

codes, and social performance. First, it enhances our understanding of family firms;

previous research tends to focus on leadership, ownership, and succession-related

topics, neglecting socially responsible issues (Benavides-Velasco et al., 2013).

Second, it examines the impact of the ownership structure on social performance. We

find that the business group and the ownership structure of a firm play a fundamental

role in determining the existence or otherwise of incentives to increase its social

performance. In contrast to the conventional wisdom concerning the benefits of

sustainable practices, we find that these can be neglected by family owners since they

are more oriented towards economic aspects than social ones. Nonetheless, the main

contribution of this paper is its analysis of the mediating effect of ethical codes on the

family ownership and social performance link. Moreover, in contrast to the majority of

studies, which focus on only one country, we use an international sample of 12

countries, thus obtaining potentially more powerful and generalizable results.

Moreover, the study’s consideration of the temporal dimension of data, particularly in

periods of great change, enriches its perspective. In this regard, the panel data

obtained enable us to control for year-on-year effects that may affect social and

environmental performance.

Our study results should be interpreted carefully, since this research is subject to

certain limitations. Due to the limited information available in the different databases,

the sample is restricted to specific countries. For example, the last year used for the

analysis is 2010. These limitations need to be addressed in future studies. Another

limitation concerns the definition of a “family firm”. In the present paper, a firm is

considered to be a “family firm” when a member of the founding family has more than

10% ownership. This is the most common approach, although there is no universal

definition. Ideally, family ownership and family management should be considered in

greater detail to characterize better the evidence discussed. In addition, in this

analysis, the variable “family firm” is dichotomized into family-controlled and non-family-

controlled firms. As Chen and Jaggi (2000) note, without a continuous measure of this

variable, the mediating effect of ethical codes on the association between family firm

and social and environmental approach may not have been properly and fully

evaluated. In this respect, a precise measure of family management, such as the

percentage of family members in senior management positions, could be a good

measure. Thus, future research could improve upon the measurement of the degree of

family control.

14

Finally, as future lines of research, it would be interesting to include other corporate

characteristics, such as board aspects, the existence of audit committees, and CEO

duality, as well as the existence of other owners (banks, government, etc.) as mediator

variables in the analysed relationship. Institutional and legal aspects should be

addressed too.

REFERENCES Adams, C. A. (2002). Internal organisational factors influencing corporate social and ethical reporting: beyond current theorising. Accounting, Auditing & Accountability Journal, 15(2), 223-250.

Adams, J. S., Taschian, A., & Shore, T. H. (1996). Ethics in Family and Non‐Family

Owned Firms: An Exploratory Study. Family Business Review, 9(2), 157-170.

Adams, J. S., Tashchian, A., & Shore, T. H. (2001). Codes of ethics as signals for ethical behavior. Journal of Business Ethics, 29(3), 199-211.

Anderson, R.C. and Reeb, D. M. (2003) Founding-family ownership and firm performance: Evidence from the S&P 500. The Journal of Finance, 58(3): 1301-1328.

Aoi, M., & Takehara, H. (2012). Family Businesses and Corporate Social Performance: An Empirical Study of Public Firms in Japan.

Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator variable distinction in social Agatiello, O. R. (2008). Ethical governance: beyond good practices and standards. Management Decision, 46(8), 1132-1145.

Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of personality and social psychology, 51(6), 1173.

Benavides-Velasco, C.A., Quintana-García, C., & Guzmán-Parra, V.F. (2013). Trends in family business research. Small Business Economics, 40: 41-57.

Berrone, P., Cruz, C., Gomez Mejia, L.R., and Larraza Kintana, M. (2010). Socio-emotional wealth and corporate responses to institutional pressures: do family-controlled firms pollute less. Administrative Science Quarterly, 55(1): 82-113.

Brenes, E., Madrigal, K., & Requena, B., (2011). Corporate governance and family business performance, Journal of Business Research. 64(3): 280-285.

Burak, A. and Morante, L.S. (2007). Corporate social responsibility and firm characteristics in Sweden: Who and what makes a firm a better corporate citizen?. Master´s Thesis in Finance Stockholm School of Economics.

15

Burak, A., & Morante, L. S. (2007). Corporate social responsibility and firm characteristics in Sweden: Who and what makes a firm a better corporate citizen. Stolkholm School of Economics.

Carroll, A. B. (1999). Corporate social responsibility evolution of a definitional construct. Business & society, 38(3), 268-295.

Chen, C. J. & Jaggi, B. (2001). Association between independent non-executive directors, family control and financial disclosures in Hong Kong. Journal of Accounting and Public Policy, 19(4), 285-310.

Chua, F., & Rahman, A. (2011). Institutional pressures and ethical reckoning by business corporations. Journal of business ethics, 98(2), 307-329.

Daft, R.L. (2003). Management, USA: Thomson South-Western.

Dahya, J., Dimitrov, O., & McConnell, J. J. (2008). Dominant shareholders, corporate boards, and corporate value: A cross-country analysis. Journal of Financial Economics, 87(1), 73-100

Déniz, M.C. & Cabrera, M.K. (2005). Corporate social responsibility and family business in Spain. Journal of Business Ethics, 56(1): 27-41.

Duh, M., Belak, J., & Milfelner, B. (2010). Core values, culture and ethical climate as constitutional elements of ethical behaviour: Exploring differences between family and non-family enterprises. Journal of Business Ethics, 97(3), 473-489.

Erwin, P. M. (2011). Corporate codes of conduct: The effects of code content and quality on ethical performance. Journal of business ethics, 99(4), 535-548.

Fernández-Gago, R., Cabeza-García, L., & Nieto, M. (2014). Corporate social responsibility, board of directors, and firm performance: an analysis of their relationships. Review of Managerial Science, 1-20.

Fuertes, A. T., & García, L. C. (2013). Analysis of the determinants of CSR disclosure in Spanish listed companies. Intangible Capital, 9(1), 225-261.

García de los Salmones, M.M., Herrero Crespo, A., & Rodríguez del Bosque, I., 2005, Influence of corporate social responsibility on loyalty and valuation of services, Journal of Business Ethics, 61: 369-385.

Haalien, L. & Huse, M. (2005). Board of directors in Norwegian family businesses. Results from the value creating board surveys. Research Report 7/2005, Norwegian School of Management BI, Oslo (Norway).

Hoopes, D. G., & Miller, D. (2006). Ownership Preferences, Competitive Heterogeneity,

and Family‐Controlled Businesses. Family Business Review,19(2), 89-101.

16

Jacobs, M. T. (1991). Short-term America: The causes and cures of our business myopia. Harvard Business School Pr.

Kaptein, M., & Schwartz, M. S. (2008). The effectiveness of business codes: A critical examination of existing studies and the development of an integrated research model. Journal of Business Ethics, 77(2), 111-127.

Kashmiri, S., & Mahajan, V. (2010). What's in a name?: An analysis of the strategic behavior of family firms. International Journal of Research in Marketing, 27(3), 271-280.

Klein, P., Shapiro, D. & Young, J., (2005). Corporate governance, family ownership and firm value: the Canadian evidence. Corporate Governance: An International Review. 13(6): 769-784.

Lam, K., Mok, H. M., Cheung, I., & Yam, H. C. (1994). Family groupings on performance of portfolio selection in the Hong Kong stock market. Journal of banking & finance, 18(4), 725-742.

Landry, S., Deslandes, M., & Fortin, A. (2013). Tax Aggressiveness, Corporate Social Responsibility, and Ownership Structure. Journal of Accounting, Ethics & Public Policy, 14(3), 611-645

López Iturriaga, F. & López de Foronda, O. (2011). Corporate social responsibility and reference shareholders: An analysis of European firms. Transnational Corporations Review, 3(3): 1-11.

Maignan, I. & Ferrell, O.C., 2001, Consumer perceptions of corporate social responsibility: A cross cultural comparison, Journal of Business Ethics, 30(1): 57-73.

Margolis, J. D., & J. P. Walsh. (2003). Misery loves companies: Rethinking social initiatives by business. Administrative Science Quarterly 48(2): 268-305.

Materne, C. F., III, Debicki, B. J., Kellermanns, F. W., and Chrisman, J. J. (2013). Family business research in the new millennium: An assessment of individual and institutional productivity, 2001–2009. In K. X. Smyrnios, P. Z. Poutziouris, and S. Goel (Eds.), Handbook of research on family business. Cheltenham: Edward Elgar Publishing.

McConaughy, D.L., Matthews, C.H. & Fialko, A.S. (2001). Founding family controlled firms: performance, risk and value. Journal of Small Business Management. 39(1): 31-49

McVey, H. & Draho, J. (2005). U.S. family-run companies - they may be better than you think. Journal of Applied Corporate Finance, 17(4): 134-143.

Mijatovic, I. S., & Stokic, D. (2010). The influence of internal and external codes on CSR practice: The case of companies operating in Serbia. Journal of business ethics, 94(4), 533-552.

17

Mok, H. M., Lam, K., & Cheung, I. (1992). Family control and return covariation in Hong Kong's common stocks. Journal of Business Finance & Accounting,19(2), 277-293.

Sharma, P., Chrisman, J.J. and Chua, J.H. (1997). Strategic management of the family business: Past research and future challenges. Family Business Review, 10: 1-35

Shrout, P. E., & Bolger, N. (2002). Mediation in experimental and nonexperimental studies: new procedures and recommendations. Psychological methods, 7(4), 422.

Singal, M. (2014). Corporate social responsibility in the hospitality and tourism industry: Do family control and financial condition matter?. International Journal of Hospitality Management, 36 (1), 81-89.

Turker, D., 2009, Measuring corporate social responsibility: a scale development study, Journal of Business Ethics, 85: 411-427.

Urbany, J. E. (2005). Inspiration and cynicism in values statements. Journal of Business Ethics, 62(2), 169-182.

Virakul, B., Koonmee, K., & McLean, G. N. (2009). CSR activities in award-winning Thai companies. Social responsibility journal, 5(2), 178-199.

Wotruba, T. R., Chonko, L. B., & Loe, T. W. (2001). The impact of ethics code familiarity on manager behavior. Journal of Business Ethics, 33(1), 59-69.

Zhao, X., Lynch, J. G., & Chen, Q. (2010). Reconsidering Baron and Kenny: Myths and truths about mediation analysis. Journal of consumer research,37(2), 197-206.

18

ANNEX. Measures of variables

Social Performance

Corporate social performance (SP) is measured using a multidimensional construct

addressing all the actions carried out, especially those taken in social and

environmental contexts (Carroll, 1999). In accordance with the Triple Bottom Line

approach, social performance can be defined as an aggregate construct composed of

three dimensions: economic, social and environmental. Although the economic

dimension has traditionally been assumed to form part of socially responsible practices,

some studies have concluded that it need not be included in the overall social

performance construct (Maignan and Ferrell, 2001; García de los Salmones et al.,

2005). According to Daft (2003), a business is an economic organisation that gains

profits by producing goods and services in a society. Since profitability is the basic

motive for its survival, the economic dimension should be considered the reason for its

existence, rather than a responsibility to society (Turker, 2009).

The EIRIS process starts with the information disclosed by the companies. Following

this, targeted questionnaires are sent to companies regarding areas where public data

are unclear. This results in a well-focused dialogue with companies, and also

encourages them to address issues of concern with investors and to improve their

public reporting. Sector specialists within each team review the research conducted by

colleagues before the results are published.

SP is determined from the non-weighted sum of environmental issues, human rights

and relationships with stakeholders based on 20 items. Each item is assigned a value

that represents the socially responsible performance of the company, scored as 0, 25,

50, 75 or 100, corresponding to very low SP, low SP, moderated SP, good and very

good SP, respectively. The first of these areas concerns items such as the company’s

environmental management system and policy, its impact on the environment, and

whether the company has published reports on this question. In addition, general note

is taken of the scope of the company’s strategy, policy, system and reporting in the

field of human rights. Other factors taken into account include the company’s

management systems, the quantitative information provided, the general level of

commitment with stakeholders, its policy and practices in support of equal opportunities

and diversity, the health systems and safety-at-work procedures that are implemented,

the support given to employee training and development, its relationships with

19

customers and suppliers, and the level of commitment with the community and with

social projects.

Figure 2 shows the composition of the SP index in detail.

Family firm

The explanatory variable of ownership concentration is Family , which is a dummy

variable (Kashmiri and Mahajan, 2010; Landry et al., 2013) that takes the value 1 if the

largest shareholder is an individual or a family with more than 10% of the votes, and 0

otherwise (Mok et al., 1992; Lam et al., 1994; Dayha et al., 2006; Aoi et al., 2012).

Code of ethics

Table 1: Social Performance Index Environmental in dex Environmental policies and commitment Environmental management systems

Environmental reporting

Level of environmental impact improvement

Human rights index Extent of policies addressing human rights issues Extent of systems addressing human rights issues

Extent of reporting addressing human rights issues

Stakeholders index Policies towards stakeholders overall Management systems for stakeholders overall

Quantitative reporting for stakeholders overall

Level of engagement with stakeholders overall

Policies on equal opportunities and diversity issues

Systems and practices to support equal opportunities and diversity issues

Health & Safety systems

Systems and practices to advance job creation and security

Systems to manage employee relations Systems to support employee training and development

Policies on maintaining good relations with customers - suppliers

Systems to maintain good relations with customers - suppliers

Level of commitment with community or charitable work

20

In order to represent the existence of a code of ethics and its level of application,

EthicCode was defined. It is an ordinal variable that takes values between 0 and 4,

such that 0 represents the absence of a code of ethics and 4 represents an advanced

ethical code, with the highest rating. A value of 1 represents limited inclusion, i.e., the

code refers to a very limited number of aspects, such as conflicts of interest, corruption

and bribery; a value of 2 is a basic level, incorporating, in addition to the first level,

recommendations on questions such as discrimination, occupational hazards, the work

environment and the confidentiality of information; a value of 3, the intermediate level,

incorporates, as well as the aspects addressed in the two previous levels, principles

and values related to relationships with customers, suppliers and competitors; a value

of 4, advanced, adds references to the sustainable use of resources, relations with

society and any other value that forms part of the corporate culture. The value of 0 is

assigned to companies that do not express any ethical commitment.

21

Figure 1. Mediation of codes of ethics on social pe rformance

22

Table 2. Descriptive Statistics

SP was defined to represent the corporate social performance through corporate social responsibility practices which is an index obtained from three areas: practices for society, practices for preserving human rights and practices for preserving the environment. EthicCode was defined to represent the existence of a code of ethics and its level of application and it is an ordinal variable that takes values between 0 and 4, such that 0 represents the absence of a code of ethics and 4 represents an advanced ethical code, with the highest rating. Family is a dummy variable that takes the value 1 if the largest shareholder is an individual or a family with more than 10% of the votes, and 0 otherwise. Controls include corporate size –as the natural logarithm of assets- and leverage-the ratio of total debt to total equity-, and systematic risk as measured by the beta of the market model. Non Family Firms Family Firms Mean Std. Dev. Mean Std. Dev. SP 726.286 442.846 663.507 431.362 EthicCode 3.358 0.965 3.182 1.042 Size 8.672 1.756 8.889 1.674 Leverage -124.977 4470.026 1.862 8.905 Risk 1.033 0.763076 0.927 0.523

Frequencies Absolute Relative (%) Absolute Relative

(%) Family 2,556 83.45% 509 16.55%

23

Table 3. Bivariate Correlations

SP was defined to represent the corporate social performance through corporate social responsibility practices which is an index obtained from three areas: practices for society, practices for preserving human rights and practices for preserving the environment. EthicCode was defined to represent the existence of a code of ethics and its level of application and it is an ordinal variable that takes values between 0 and 4, such that 0 represents the absence of a code of ethics and 4 represents an advanced ethical code, with the highest rating. Family is a dummy variable that takes the value 1 if the largest shareholder is an individual or a family with more than 10% of the votes, and 0 otherwise. Controls include corporate size –as the natural logarithm of assets- and leverage-the ratio of total debt to total equity-, and systematic risk as measured by the beta of the market model. Industry, Country and Year are dummies that represent sector group, country and year, respectively.

1 2 3 4 5 6 7 8 9

1. SP 1.000 2. EthicCode 0.303 1.000 3. Family -0.067 -0.080 1.000 4. Size -0.043 0.181 0.067 1.000 5. Leverage -0.011 -0.024 0.019 -0.070 1.000 6. Risk -0.125 -0.013 -0.057 -0.043 -0.037 1.000 7. Industry 0.026 0.001 -0.113 -0.139 0.042 -0.076 1.000 8. Country 0.190 -0.098 -0.047 -0.057 -0.013 0.002 -0.077 1.000 9. Year 0.333 0.174 -0.005 0.004 -0.022 0.017 -0.025 0.051 1.000

24

25

Table 4. The mediator effect of the code of ethics on the link between family firms and their social p erformance

This table shows the results of regressions proposed in model 1, 2a and 2b, and 3. SP was defined to represent the corporate social performance through corporate social responsibility practices which is an index obtained from three areas: practices for society, practices for preserving human rights and practices for preserving the environment. EthicCode was defined to represent the existence of a code of ethics and its level of application and it is an ordinal variable that takes values between 0 and 4, such that 0 represents the absence of a code of ethics and 4 represents an advanced ethical code, with the highest rating. Family is a dummy variable that takes the value 1 if the largest shareholder is an individual or a family with more than 10% of the votes, and 0 otherwise. Controls include corporate size –as the natural logarithm of assets- and leverage-the ratio of total debt to total equity-, and systematic risk as measured by the beta of the market model. ***, ** and * indicate significance at a level of 1%, 5% and 10% respectively. Model 1 Model 2a Model 2b Model 3 Dependent variable EthicCode SP SP SP

Coef Std. Error Coef Std. Error

Coef Std. Error Coef Std. Error

EthicCode 5.7947*** 1.046 5.751*** 1.070 Family -0.486** 0.188 -5.8604* 3.0667 -5.057 3.022 Size 0.024 0.030 -.294 0.527 -0.082 0.500 -0.333 0.525 Leverage 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 Risk 0.100 0.105 -11.436*** 1.651 -12.586*** 1.766 -11.588*** 1.635 _cons 4.185*** 0.335 106.677*** 6.570 105.183*** 5.700 108.617*** 6.842 Industry Controlled Country Controlled Year Controlled sigma_u 1.843*** 0.088 37.653*** 1.455 1.843*** 0.088 38.039*** 1.497 sigma_e 1.090*** 0.042 29.412*** 0.661 1.090*** 0.042 29.409*** 0.661 rho 0.741 0.022 0.621 0.021 0.741 0.022 0.626 0.021

26