Increasing the corporate social performance by intensive engagement with stakeholders.

1

Amsterdam Business School

Increasing the Corporate Social Performance by

Intensive Engagement with Stakeholders.

Date: January 29, 2014

By: Derk Jan van Bijsterveldt Student Number: 10506543

MSc in Business Studies

Master Thesis International Management

Supervisor: Alan Muller Second reader: Arno Kourula

2

Table of Content

Corporate Social Performance ... 16

Stakeholder Engagement and Corporate Social Performance ... 19

Multinationality... 21

Multinationality, Stakeholder Engagement and CSP ... 23

Methodology ... 26

Sample ... 26

Corporate Social Performance ... 27

Stakeholder Engagement ... 30 Multinationality... 33 Control variables ... 35 Industries ... 35 Regression Analysis ... 36 Results ... 38 Correlations ... 38

Stakeholder Engagement 2011 – CSP Total Strengths 2011 ... 40

Stakeholder Engagement 2011 – CSP Total Concerns 2011 ... 41

Stakeholder Engagement 2011 – CSP Total Strengths 2012 ... 42

Stakeholder Engagement 2011 – CSP Total Concerns 2012 ... 44

Reversed Causality: CSP – Stakeholder Engagement ... 46

Answering Hypotheses ... 47

Discussion ... 50

Conclusion ... 50

Limitations ... 53

Implications and Future Research ... 55

3

Abstract

For many managers it is important to perform both financially and socially well. As is the

case for financial performance, to have a good corporate social performance(CSP) a

thought-out strategy is needed to effectively and efficiently reach this goal. A relevant issue in this

matter is the relationship between stakeholder engagement(SE) and CSP. This paper

examines this relationship by analyzing 135 S&P500 companies. Results show a positive

relationship between the SE and positive CSP. More interesting, a positive relationship is

established between SE and negative CSP. Research is also done regarding the effect that

multinationality has on the proposed relationship. The result showed no significant effect

4

Introduction

Stakeholder theory and corporate social responsibility are two constructs that cannot be

totally separated. In literature and business the two theories are linked and both are highly

present in the other. Freeman (1994) states that stakeholder theory entails two important

questions. First, what is the purpose of the firm? Second, what responsibility does

management have to its stakeholders? As these questions are essential for stakeholder theory,

the answers to these questions denote largely what the position of the company is regarding

its corporate social responsibility. The discussion about what the responsibility of the

company is has been going on for several decades. Although it is not exactly clear when it

started, a good case can be made for the early 50s (Carrol, 1999). Since then, the literature

has grown exponentially and with it the variety of stances taken in this subject. For instance,

Friedman (1970) wrote a letter to the New York Times in which he decries the idea that

businessmen have social responsibilities in the following words:

“ The businessmen believe that they are defending free enterprise when they declaim that business is not concerned "merely" with profit but also with promoting desirable "social" ends; that business has a "social conscience" and takes seriously its responsibilities for providing employment, eliminating discrimination, avoiding pollution and whatever else may

be the catchwords of the contemporary crop of reformers. In fact they are--or would be if they or anyone else took them seriously--preaching pure and unadulterated socialism. Businessmen who talk this way are unwitting puppets of the intellectual forces that have been

undermining the basis of a free society these past decades……The key point that, in his capacity as a corporate executive, the manager is the agent of the individuals who own the corporation or establish the eleemosynary institution, and his primary responsibility is to them.”

As is shown above, a radical stance is taken; he basically says that the only responsibility for

5

Although Friedman (1970) still has a large group of followers, during the years

another perspective has seen a rise in support. This group points to the advantages that

corporate social responsibility(CSR) can have for the company, thus having a more strategic

approach regarding CSR. According to Porter and Kramer (2006), companies should use

CSR to establish a competitive advantage. They state that for several years, companies did

not apply their CSR properly for two reasons. First, managers pit business against society,

when clearly the two are interdependent. Second, they pressure companies to think of CSR in

generic ways instead of thinking which CSR strategy is most appropriate for them. Although

this stance shows more in favor of CSR, it is still very much connected to what it can bring to

the company; it thus has some connection to Friedman’s (1970) position that management should create shareholder value.

Another point of view that is present according to Van Marrewijk (2003) is that a

company is part of a larger system in which it operates. It is communion that stops freedom

when it interferes with the freedom of others. This entails that a company cannot just harm

other persons, or the living space or environment in which the other inhabitants of the system

live. Being an entity of a larger whole obliges the company to adapt to its environment and to

be accountable for its impact on others. In comparison to the previous two points of view, this

focusses far more on the moral obligation from a system approach, instead of an economic

reasoning from the perspective of the firm.

As is shown above, people differ in their opinion whether CSR is an instrument, a

goal or if it has no significance altogether. For those who consider CSR to be an important

asset to the company and want to perform well socially, it is important to know how the

corporate social performance (CSP) can be increased. In this case, the CSP is the outcome of

the CSR policy, as for all that one does, the outcome may not always be the same as the

6

One aspect that could influence CSP is stakeholder engagement (SE). SE is defined

by ISEA (1999) as “the process of seeking stakeholder views on their relationship with an

organization in a way that may realistically be expected to elicit them”. Enabling and facilitating SE may have many benefits. It can create strong social ties, as an embedded status

within the network of the organization makes that the company is ‘in sync’ with its stakeholder expectation. It makes the company able to mobilize multiple stakeholders in a

coalition to build a sustainable and responsible business (Maak, 2007). According to

Greenwood (2007) SE is much associated with corporate responsibility, but SE in itself is a

morally neutral activity and may not necessarily enhance the CSP of a company. Although

this could be the case, according to Valor (2005) corporate social responsibility and

stakeholders complement and reinforce each other. In addition to this, Carrol (1998) states

that corporate citizenship, which is viewed as closely related to CSR, entails for a great part

the involvement of relationships with all important stakeholders. Also, Noland and Phillips

(2010) state that SE is an important part of CSP and applying the right SE techniques can

raise the CSP. Some disagreement on this matter is thus present, and research regarding this

relationship would be desirable. I argue that for the sustainability of the SE process, a

strategic motivation is necessary in addition to a moral judgment, because it would create a

win-win situation, which will prolong and mature the stakeholders engagement process. If

more clarity is established with regards to the issue of SE, and the relationship between CSP

and SE, managers who want to raise the CSP of the company will be better able to

comprehend the means by which to do so.

In this matter, CSR is becoming more and more a way of doing business. Logically,

the geographical diversification can have a great effect on the process and outcomes of doing

business. According to Yang and Driffield (2012), multinational companies have

7

This thus entails that companies naturally want to get bigger, and expand to other countries.

But what does this mean for the CSP, and in particular to the effect that SE has on CSP.

Much research is being done regarding the effects of multinationality on the company’s

financial performance (Buckley and Casson 1976; Rugman 1986; Dunning 1988; Tallman

and Li 1996; in Yang and Driffield, 2012) and on the effects it may have on the internal

organization (Lu and Beamish, 2004; Hennart, 2007; Bartlett and Ghoshal, 1989). Yet, no

research has been done with regards to what happens to the relationship between SE and CSP

when a company internationalizes more. When a company internationalizes, logically it will

find itself dealing with more parties than it is used to. Dealing with more partners means that

the amount and variety of preferences and demands concerning how to behave as a company

will increase. When the relationship between SE and CSP is seen in the light of this

observation regarding multinationality it may have its implication. First, it could be that it

will be harder for the firm to interact with all the stakeholders. Second, as a result of the

internationalization process, some stakeholders could become less important, thus shifting the

focus and action of the company. Lastly, dealing with an increased amount and variety of

stakeholders may lead to opposing preferences, which may make it harder to formulate CSR

policy from SE. Multinationality may thus influence the relationship between SE and CSP in

many ways. It is important for managers to know if it will do so and how, because it could

mean that policies that were perfectly workable at a domestic level may no longer be

functional as the company starts to operate internationally.

In this study, the focus will be on the relationship between stakeholder engagement

and corporate social performance. When this relationship is established, it will be important

to determine what happens to this relationship when multinationality is added into the

equation. In accordance with the reviewed literature, as will be shown below, a positive

8

on this relationship. The analysis will be done in the following order. First, a theoretical

review will be performed; this will entail a clarification of what is important in the field of

stakeholder engagement, corporate social performance and multinationality. Alongside this,

there will be a review of what has been written about the relationships between the different

variables. Hereafter, the analysis will be performed using 135 companies from the S&P 500

list of 2012. In these sections the used methods will be explained and the results will be

shown. When results are collected and possible relationships established, the discussion will

be presented. In the discussion, the practical implications for managers and the implications

for the literature will be shown and possible limitations of this study will be put forward. It

will end with some suggestions of future research that may be of importance to the

9 Conceptual Model

H1

H2

This thesis will have a strong focus on corporate social performance (CSP), and what

influence the degree of SE will have on this. Expected is that intensive interaction with

stakeholders will raise the CSP, thus denoting a positive relationship. A second analysis will

be performed to show the effect that multinationality has on this relationship. As a result of

what is shown in existing literature, it is hypothesized that multinationality will have a

negative moderating effect on the relationship between SE and CSP.

Stakeholder

Engagement

Corporate

Social

Performance

Multinationality

10

Theoretical Review

Stakeholder Theory

Freeman (1984) defined stakeholders as “any group or individual who can affect or is affected by the achievement of the organization’s objectives”. Savage et al (1991) defined stakeholders as groups or individuals who “have an interest in the actions of an organization and … the ability to influence it”. According to Gao and Zhang (2006) these definitions reflect a two-way relationship; there is interdependence between the organization and its

stakeholders. During the 1980s, fundamental changes took place in business organizations.

External stakeholders demanded greater social responsibility of companies and stakeholders

obtained influence with respect to different organizational activities. These changes were

reflected in a 1987 cover story in business week (In Savage et al, 1991):

“Outside directors are asserting themselves. Other stakeholders – from employees…to communities- want a voice. The internal balance of power is beginning to shift… the days

when CEOs could neglect their big institutional owners and other corporate stakeholders are coming to an end…[N]ow managers will have to listen to – and learn from- other groups who are demanding a voice in the running of the company”

As shown above, and also expressed through the extensive attention given to within

the academic literature (Valor, 2005, Freeman, 1984, Rowley, 1997, Pedersen, 2006,

Donaldson and Preston, 1995), since the 80s increasing importance has been accorded to the

influence of stakeholders. According to Donaldson and Preston (1995) stakeholder theory

knows three different streams of literature, which are all interrelated as well as they are

distinct. The three different streams they denote are descriptive/empirical, instrumental and

11

stakeholder theory presents a model describing what the corporation is. “It describes the corporation as a constellation of cooperative and competitive interests possessing intrinsic

value”. The second stream, instrumental, refers to the establishment of a framework, which examines any potential connections between the practice of stakeholder management and the

achievement of various corporate performance goals. The final stream, normative, is the fundamental basis and involves acceptance of two ideas: (1) “stakeholders are persons or groups with legitimate interests in procedural and/or substantive aspects of corporate activity.

Their interests in the corporation identify stakeholders, whether the corporation has any

corresponding functional interest in them. (2) The interests of all stakeholders are of intrinsic

value”. Beside these different aspects, Donaldson and Preston (1995), supported by findings of others (Berman et al, 1999, Hillman and Keim, 2001, and Ogden and Watson, 1999), state

that the view of stakeholder management and favorable performance of the company going

hand in hand, has become more commonplace within both the professional and academic

management literature.

Stakeholder Engagement

A component of stakeholder theory is stakeholder engagement, which is defined by ISEA (1999) as “the process of seeking stakeholder views on their relationship with an organization in a way that may realistically be expected to elicit them”. According to Greenwood (2007) SE can be understood “as practices that the organization undertakes to

involve stakeholders in a positive manner in organizational activities”. There are various levels and ways of engaging stakeholders in the process of business (Gao and Zhang, 2001,

12

Pedersen (2006) talks about a continuum between low and high levels of engagement

on different dimensions: inclusion, openness, tolerance, empowerment and transparency. For

the first dimension, which is inclusion, a low level of engagement entails having a dialogue

with only a few privileged stakeholders, and a high level involves all relevant stakeholders.

Openness knows a low level of engagement when a dialogue is structured around a fixed set

of questions, as opposed to open questions. The third dimension is tolerance, whereby in a

low level of engagement, one position has priority over all the others. A high level of

tolerance entails that alternative and critical voices are respected. Empowerment is the degree

to which one stakeholder dominates the dialogue and decisions, as opposed to freedom and

equality in dialogue as well as in decisions. The final dimension in which the level of

engagement can differ is transparency. Within transparency, a low level of engagement

means there is no access to information about the process and outcomes of the stakeholder

dialogue. A high level of engagement concerning transparency entails full access to the

information. I will come back to these dimensions later, because the possibility to score high

on these levels of engagement may be more difficult if the degree of multinationality

increases.

Besides different dimensions and the related levels of engagement, Pedersen (2006)

talks about phases of stakeholder dialogue and related filters. Three filters precede the three stages. The first filter, the “selection filter” is concerned with the access to the dialogue. From all the possible stakeholders, a selection is being made regarding who can participate in the stakeholder dialogue, which is the first phase. The second filter, the “interpretation filter” precedes the decision phase. This filter is the transformation of multiple voices from the

dialogue into a limited number of decisions. Stakeholder dialogue is a difficult process and it

may not be possible to satisfy all stakeholders. The third and final filter Pedersen (2006)

13

decisions move out of the dialogue arena. There is a difference between preferred actions and

final outcomes of the SE process. The same is the case in this instance as it is for the levels of

engagement. The higher the level of multinationality, the harder it may be to satisfy all the

original stakeholders that were present before the selection filter. I will come back to this

later.

According to Noland and Phillips (2010) the topic of SE becomes more significant

and interesting. The difference between engagement and interaction needs a clear distinction.

Stakeholder interaction alone is no longer sufficient, interaction in this perspective can be

transacting without inquiring as to his or her wants, needs, wellbeing or capabilities. SE, on

the other hand implies, at minimum, recognition and respect of common humanity. Because

companies have an effect on persons and communities, companies must identify their

stakeholders and communicate with them.

In Table 1, different ways of conceptualizing SE are presented. All these possibilities

show some ranking of high or low SE, or attributes that are signs of higher SE. Gao and

Zhang (2001) divide stakeholder engagement into four clearly defined levels of engagement.

The first level is the ‘passive’ level of engagement, where stakeholders are merely given information. The number of stakeholders here is large and the information comes via public

media and through public reports. The second level they denote is ‘listening’, wherein, as the name suggests, stakeholders are consulted. This level includes a selected number of

stakeholders. The third level of engagement is the ‘two-way process’, which entails stakeholders engaging in dialogue with the company. Just as for the ‘listening’ level, there is

a limited number of key stakeholders. The fourth and maximum level of SE according to Gao

and Zhang (2001) is the “proactive” level. Within this level, stakeholders drive management. Examples of stakeholder approaches in this level are the creation of a stakeholder council,

14

allowing stakeholder representatives in management and asking stakeholder verification of

social report.

Levels of SE Critique

Gao and Zhang 2001.

They divide engagement into four levels: 1.passive (merely giving information), 2.listening (stakeholders are consulted), 3.two-way-process (stakeholders engage in dialogue with company), and 4.proactive (management is driven by stakeholder).

Clear distinctions between the different levels.

Question remains whether a level 0 could be present, which would mean that no information would be given.

Friedmand and Miles (2006)

Quality of SE(management) comprised by 8 categories: 1 and 2, manipulation and therapy. 3,4,5, informing, consultation and placation. 6, partnership, 7 delegated power, and 8 citizen control.

Is more applicable to established stakeholder engagement processes. Less concerned with the degree of SE.

Michelon (2011) Ordinal scale using the four criteria of GRI ranging from low to high engagement.

No specification what low engagement and high engagement entails. AA1000 According to AA1000 the proposed

engagement consists of five components: planning, accounting, auditing, reporting and embedding. Defines quality SE by some criteria that is must meet. Clearly defines scope, agreed decision-making process, transparent, etc.

These are useable guidelines for managers. Components do not show a clear distinction of low or high engagement.

Ayuso et al 2011 Two questions: one addressing the number of different external stakeholders the company regularly addresses through satisfaction surveys or perception studies (stakeholder scope) and one question about the number of mechanisms used by the company for engaging with these stakeholders (stakeholder process).

Score on stakeholder scope and stakeholder process could be dependent of many other factors like firm size, industry, country, policy, etc.

Manetti (2011) Uses GRI reports

Companies have a series of questions which they must answer using the GRI reports.

GRI reports are used on a broad scale and are

considered reliable because of independence of a possible GRI check.

All literature seems to denote some sort of ranking from high to low SE, although

some describe the different levels more specifically. In this study, the division as is shown by

Gao and Zhang will be used in the analysis, because with a raise in the level of SE, a clearly

defined extra attribute is added each time.

15

Behind the degree of SE from the company’s perspective may lay different motives.

A division of three kinds of motives is being made on the continuum between the normative

stream on the one side, and instrumental stream on the other side. As I will explain below, the

first two, being present in the philosophical reasoning of Jürgen Habermas represent moral

and strategic stakeholder engagement. The last motive denoted here, entails the ‘ethical strategist’. Moral and political philosopher Jürgen Habermas argued that engagement must be largely free of any strategic motivation, in order to ensure its moral legitimacy. This theory

lies within the normative stream of SE, as it denotes the importance and the justification of

SE. Scholars who comply with this reasoning, also called Habermasians, can be divided into

two camps: A school supporting moral stakeholder engagement on the one side, and a school

supporting strategic stakeholder engagement on the other. The first, moral engagement is

“marked by specific conditions of communication which ensure that the communication is uncorrupted by power differences and strategic motivation. The aim of this type of

engagement is agreement for the sake of agreement” (Noland and Phillips, 2010). The latter,

strategic stakeholder engagement is undertaken with strategic motivations. This does not

mean that these intentions are necessarily dishonest or malicious. Although this is part of the

normative stream, an instrumental aspect comes into play here, as it is interested in how SE

can enhance the achievement of a company’s goals. Whereas the followers of Habermas are

an important group within stakeholder theory, another school of thought consists of scholars

that can be called Ethical Strategists (Noland and Phillips, 2010). They state that honest, open

and fair engagement of stakeholders is necessary for a business to function properly. Whereas

the moral camp of Habermasians lies within the normative stream of SE literature, the Ethical

Strategists lies within the instrumental stream, because the SE should lead to the achievement

16

engagement and ethical strategists are both at the end of the continuum of why stakeholders

should be involved and that the strategic side of Habermasians is in the middle of these two.

As Rawls (2000, in Godfrey 2005) denotes, different stakeholders and individuals are

influenced by different doctrines, which provide them with definitions of what a good society

should be. This means that their beliefs on this matter may well be different from someone

else’s. In accordance to this, Godfrey (2005) states that a firm’s “public” consists of multiple communities, each representing different ethical values and value systems. Few ethical values

will be common across all communities, meaning that these values and preferences of

communities may very well conflict with each other. These different ethical values may pose

challenges when internationalizing. Besides this notion of multinationality, Greenwood

(2007) states that stakeholder engagement is a morally neutral action, though it may be

related to corporate responsibility. But SE in itself may be morally neither good nor wrong.

In line with the authors mentioned above, the goal of this paper is not to put a moral mark on

SE. The focus will be what the effect of SE will be, in this case on CSP, without making a

moral judgment regarding the justification of engaging stakeholders in the process of

business. Besides this, I argue that for the sustainability of the SE process a strategic

motivation may be preferable as this may entail a win-win situation, which will prolong and

mature the stakeholders’ engagement process. I will thus follow an instrumental approach,

and write this paper while slightly taking the role of an ethical strategist.

Corporate Social Performance

Stakeholders are increasingly becoming concerned about the corporate social

performance (CSP) of the company’s operation. Consumers may prefer eco-labelled products

17

screens are used by investors and corporations develop socially responsible purchasing

practices to promote more sustainable supply chains (Chen & Delmas, 2010). The

above-mentioned examples are just a part of the increasing number of expressions of CSP concerns

of stakeholders. CSP has been a topic in businesses and literature since the mid-1970s

(Wood, 1991), but the intellectual roots can be found, in the 1950s, in Boulding’s (1956, in Wood, 2010) view of complex organizations as open systems, intricately connected to their

larger environment. This was contradictory to the general opinion in those days, whereas

organizations were viewed of as closed systems. In accordance with this new view of the

organization, CSP is concerned with the harms and benefits generated by the organization’s

operations by interacting with its environment. This is reflected within different dimensions

of the surrounding; for instance social, cultural, political, economic and natural dimensions

(Wood, 2010).

According to Turban and Greening (1996), CSP can be defined as “a construct that emphasizes a company’s responsibilities to multiple stakeholders, such as employees and the community at large, in addition to its traditional responsibilities to economic stakeholders”. In addition to this definition, Mitnick (2000) states that the distinction should be set out

clearly between social performance and social action, whereas the first entails outcomes

(impacts) and the second represents outputs (activities). The definition that is given to CSP

by Turban and Greening (1996) seems to represent the latter of this distinction, the activities

of the company, not as much as the outcome. According to Clarkson (1995) a definitive

conclusion of what CSP exactly entails is not established yet, and Wood (2010) adds to this

that the CSP domain has remained controversial and fluid. But Clarkson (1995) states that

there has been some agreement with the definition of Preston (1988), which emphasises the

outcomes of corporate behaviour instead of the action. In this paper, CSP will thus be

18

Mitnick (2000, in Mattingly and Berman) in consideration. In line with this consideration, the

view of Preston (1988, in Clarkson, 1995) will be the focus or definition of CSP in this case;

it states that CSP is ‘the impact of business behaviour on society’.

A good CSP has been argued to have many effects. An important notification here is

the word ‘argued’, because for many of the relationships concerning CSP, opinions are divided into camps of believers and non-believers. First, it has been argued to enhance the

company’s performance, for instance because of the company’s engagement in cause-related marketing, corporate philanthropy, and green marketing. Marketers point to advantages it can

have on the company’s reputation, customer loyalty and customer-company identification

(Luo and Bhattacharya, 2009). Second, besides the somewhat obvious view from a marketing

perspective, a large part of the CSP literature focuses on the relationship between CSP and

firm financial performance. Unbelievers in CSP state that a company should not engage in

this because it demands resources that the company could better allocate otherwise. On the other hand, many scholars believe the adage of “Doing well by doing good”, and thus see a positive relationship between CSP and firm financial performance (Orlitzky, 2011). In

accordance to this finding, Wood (2010) carefully states that good social performance results

in a better bottom line for the company, and vice versa, bad performance is likely to result in

financial harm, but she denotes that it is hard to measure what the exact impact is of CSP on

firm financial performance. For instance, there is no direct payment from charity or CSR

initiatives because these may be translated into financial return of investment through

reputation, or acquisition of NGO alliances or pacification of potential belligerent claimants.

Next to this notion, Wood (2010) states that CSP can have an impact on a varied set of

variables, which can be seen as essential to the company’s survival. Some of these she denotes are reputation, innovation and access to needed resources.

19 Stakeholder Engagement and Corporate Social Performance

According to Greenwood (2007) stakeholder engagement is much associated with

corporate responsibility. It seems logical that if an organization shows more commitment to

its stakeholders and lets them get involved in the process of business that it has a responsible

attitude towards their stakeholders. Greenwood (2007) states that this is a myth and a

simplistic assumption that SE would necessarily increase the corporate responsibility. More

engagement is not the same as being more accountable and responsible towards the

company’s stakeholders. She also states, justly, that SE knows many goals, and can be used

as a mechanism for different desired outcomes. This assumption makes Greenwood’s (2007)

central argument that SE is a morally neutral practice. A distinction she makes in her model

is between SE (high engagement is where the stakeholder-involved activities are numerous or

of high quality) and stakeholder agency (whereas low stakeholder agency means involving a

single of few stakeholders). Different combinations and ways of engagement can thus be

characterized by quantity of the stakeholder-involved activities and quality of these

stakeholder-involved activities.

Wood and Jones (1995, in Wood, 2010) theorized, by examining 65 studies

concerning the relationship between CSP and financial performance, that within this

relationship, stakeholders play at least four different roles of importance. Stakeholders are the

source of expectations, thus constituting desirable and undesirable firm performance. They

experience the effects of corporate behavior, and evaluate these. And finally they act upon

their interests, expectations, experiences, and evaluations. Stakeholders thus notice much of

the CSP and enact upon these notifications. Taking the instrumental approach in

consideration it is thus of importance from a strategic stance that stakeholders are engaged in

order to perform in a way that positively influences them. By engaging with stakeholders, the

20

According to Valor (2005) corporate social responsibility and stakeholders

complement and reinforce each other. The needs and social demands of different stakeholders

vary and change over time. Consequently, good interaction and engagement with stakeholder

gives the company information about what the needs of the stakeholders are, and thus

increases the chance to perform socially well and improve the CSP.

Corporate citizenship entails for a great part the involvement of relationships with all

important stakeholders (Carroll, 1998). In accordance to this, both corporate citizenship and

CSR propose that companies should be controlled by society and not only by shareholders.

This is in conflict with the neoclassical view that the companies should only satisfy their

shareholders (Valor, 2005). Thus, if the responsibility of the company goes beyond

shareholders satisfaction according to CSR theory, it is undeniable that stakeholders

constitute a large part of the corporate social performance. The question remains, whether the

influence of SE has a neutral (according to Greenwood, 2007) or a positive influence

(according to Valor, 2005).

The process of SE does not stop at interaction, as is explained by Noland and Phillips

(2010), instead, it is a process that ends in outcomes. Taking this in consideration, the

outcome of this process of SE results partly in the outcomes of the company, which

determines the degree of CSP. From the perspective that the stakeholder’s social demands

and needs change over time, and are different per place, SE may enhance companies to

capture these changes. An important notion should be made here that CSP, as is stated earlier,

is regarded here as the outcome of the company’s performance, not just the outputs.

The conclusions of these authors lead to the assumption that SE will enhance the

company to positively influence the CSP outcomes. Thus, taking all of the above mentioned

21

consequence to this, the score on corporate social performance will increase. This leads me to

the following hypothesis:

H1: More stakeholder engagement leads to a higher score on CSP

Multinationality

There are indications that effects of stakeholder engagement may be complicated by

the exposure to a wide range of stakeholder views, which may be due to increased

internationalization. According to Yang and Driffield (2012), alongside many other authors,

multinational companies have opportunities to achieve greater returns from international

exploitation of intangible assets. Advantages resulting from being a multinational company

may be benefits of internalisation, including economies of scale and scope, and the ability to

relocate activities to reduce costs. Further, by being multinational a company has the

mechanisms at its disposal to allocate resources more efficient, through the creation of

intra-firm markets when intermediate markets are missing (Yang and Driffield, 2012). An effect of

these advantages is that costs can be lowered and productivity increased, leading to increased

financial performance of the company (Buckley and Casson 1976; Rugman 1986; Dunning

1988; Tallman and Li 1996; in Yang and Driffield, 2012). Of course, companies can also

experience disadvantages from multinationality, for example from increased coordination and

management costs and cultural diversity. In addition to this, according to Lu and Beamish

(2004) a multinational enterprise’s (MNE) performance will decrease when they reach a certain high degree of internationalization. They propose an S-curve, stating that with early

and mid-stage internationalizers the performance will increase, partly as an effect of the

22

their internationalization they experience a decline in their performance, as an effect of

cultural distance and coordination costs of very dispersed markets (Lu and Beamish, 2004).

Hennart (2007) states that the multinationality-performance literature has made two

predictions. First of all, the more internationally diversified an MNE is, the lower the risks.

Second, generally the more internationally diversified an MNE, the greater its profitability.

Another aspect of multinationality is the degree of local responsiveness or integration. MNEs

that choose to be locally responsive find it of importance to listen to local markets and along

with it, give subsidiaries in countries or areas thus some freedom to act on different and

changing demands within their markets. If, on the other side, a company chooses to have a

high degree of integration, this can positively influence cost reductions and there is much

control from HQ over subsidiaries. These choices are thus twofold, and lead to four types of

MNEs: (1) Centralized exporter, with a low degree of responsiveness and an high degree of

integration; (2) International projector, with both a low degree of responsiveness and

integration; (3) International coordinator, having a high degree on both areas and (4)

Multi-centred MNE, having a low degree of integration and a high degree of responsiveness

(Bartlett and Ghoshal, 1989 in Harzing, 2000).

All these aspects of multinationality have an effect on the company’s performance.

Although what the effects precisely are may be debatable, but that the aspects of

multinationality have any effect is universally accepted. Logically the different aspects and

effects of multinationality may also have its effect on smaller parts of the operation, in this

case, corporate social performance (CSP). Perrini et al (2007) studied the differences and

similarities of large companies and medium and small companies. They found that larger

companies find it more relevant to define and implement CSR strategies. Besides, small

companies make less use of CSR instruments than large companies do. Large companies are

23

and reporting strategies. Whether the degree of multinationality or other aspects of it

influence these differences is not clear. Further, the impact of CSP on a company’s

multinationality is studied (Bouquet and Deutsch, 2008), and also the relationship between

CSR practices and subsidiary autonomy is investigated (Muller, 2006), but these are not

comprehensive studies concerning the relationship between stakeholder engagement and

CSP, and the effect that multinationality has on this relationship.

Multinationality, Stakeholder Engagement and CSP

Multinationality or geographical diversification can have a great effect on the process

and outcomes of doing business. Multinational enterprises(MNEs) may have more difficulty

to identify the right issues. This stems from a few different attributes that come hand in hand

with doing business across national borders. First of all, the company has more actual and

potential stakeholders, just because it is present in more countries. Second, certain

transnational organizations must be taken into account which may not be the case when a

company only acts domestically. Also in this case, when operating in more countries,

logically more transnational organizations are stakeholder to the company. For example, if a

company operates in the different countries in Europe it has the European Union as a

stakeholder. When the company decides to go more abroad, for instance to the South

America, it will get an extra transnational institution to deal with, for instance the USAN, the

political and economic union of the South America’s. Third, even if a MNE engages in essentially the same basic activity in each of the various countries, it is very likely that each

country proposes a different set of stakeholders to the company (Nigh and Cochran, 1987).

According to Pedersen (2006) stakeholder influence or engagement goes through three filters. The first filter is the ‘selection filter’, as the name suggests, this is where the selection is being made regarding which stakeholder gets influence in the next phase, the ‘stakeholder

24

dialogue’ phase. When, in accordance to Nigh and Cochran (1987) the amount of countries a company operates in gets larger, the number of stakeholders increases,| this automatically

means that there is a larger pool to select from and it may be harder to distinguish which

stakeholders are of most important. The next filter is the interpretation filter; this concerns

‘the transformation of the multiple voices from the dialogue into a limited number of decisions’. If the multitude of voices is bigger, it is logically more difficult to translate these into decisions best suitable for most stakeholders. In accordance to this, Husted and Allen

(2006) state that diverse stakeholders and conflicting values between local and global social

needs, as a consequence of being more multinational, demands a more complex CSR

strategy. The final filter, called the ‘response filter’, is concerned with the activities that take place when the decisions move out of the dialogue area. It then is being influenced by local

interpretations, conflicting interests and organizational changes (Pedersen, 2006). Al of these

may influence and determine the outcomes of the SE and thus determine the corporate social

performance.

The notion that Godfrey (2005) made regarding different ethical values of stakeholders and stakeholder’s communities, also has its consequences when the degree of multinationality increases. When the degree of multinationality increases, logically more

stakeholders and stakeholder’s communities are involved by, or may involve, the process of doing business. When this amount thus increases, and alongside it the moral preferences will

become more varied, it will make it harder for a company to enact upon these more and more

varied opinions about what is the right way to do business.

As is described earlier, according to Lu and Beamish (2004), when MNEs continue to

increase their internationalization, they experience a decline in their performance, as an effect

of cultural distance and coordination costs of very dispersed markets. This may very well also

25

influence CSP, the presence of more stakeholders and a more diversified group of

stakeholders may have its negative effect on the corporate performance, just like the S-curve

theory of Lu and Beamish (2004).

When taken the above mentioned aspects of multinationality in consideration, the

degree of multinationality thus may, logically, influence the positive relationship between SE

and CSP. I argue, in accordance to these notions, that when the degree of multinationality

increases, and with it the number and diversity of stakeholders, it will be harder for

companies to retain the positive relationship between stakeholder engagement and corporate

social performance. This leads me to the following hypothesis:

H2: The degree of multinationality negatively moderates the positive relationship between stakeholder engagement and CSP.

26

Methodology

In this section, the methods of this study are being discussed. First, I will explain how

the selection of the sample took place, secondly, I will discuss several techniques to measure

the variables that are of importance in this paper. This section will end with the elaboration

on why certain analysis techniques are chosen.

Sample

To test the hypotheses, I developed a dataset based on 135 firms. In order to have a

useable sample large companies of which the most are active in several countries should be

included. Taken the variables into consideration a sample was collected from the S&P 500

list of 2012. This listing holds many multinationals and has a wide variety of industries. In

order to prevent that industry characteristic influence the outcome of the research I choose to

use a stratified sampling method. Nine of the most common industries are taken from the

S&P 500, being the following in random order: consumer discretionary, consumer staples,

energy, financials, health care, industrial, information technology, materials and utilities.

From each industry fifteen companies are randomly selected, which results in a sample of

135 companies. By applying this sampling method, certain industry related characteristics

27 Corporate Social Performance

Measurements of firm’s financial performance are readily available, think of return on assets or return on investment. The CSP counterparts measurements for performance in this

area are not this readily available. This notion can be ascribed in part to the more qualitative nature of CSP, whereas the focus is somewhat more on ‘soft’ measures instead of ‘harder’ measures. Several authors have described the challenges associated with measuring CSP

(Carroll 1999, Graves and Waddock 1994, Chen and Delmas 2010). Academic researchers

have measured CSP using many different approaches, for instance using survey

questionnaires, content analyses of annual reports, expert evaluations, and regulatory

compliance data. According to Graves and Waddock (1997), the difficulty with measuring

CSP lies in its multidimensional construct. Behaviors in this construct have a variety of

inputs, entail internal behaviors (treatment of employees) and external behaviors (community

behaviors). Besides this, it also occurs on a wide variety of industries, with different histories,

cultures and characteristics. Lastly, what makes it harder to measure is the absence of a

universally accepted definition of the construct CSP. In the past, reputation of the public was

used as an indicator of CSP (Aupperle 1991, in Graves and Waddock, 1997), though this may

be debatable taking into consideration that it actually only measures appearance, and not true

performance. At this moment, the most commonly used database for assessing CSP is the

KLD database (Chen and Delmas 2010, Turban and Greening 1997, Graves and Waddock,

1994, Johnson and Greening 1999, Griffin and Mahon, 1997), and it is widely used in

different prominent magazine such as Journal of Business Ethics, Business and Society, the

Academy of Management Journal, and many others (Chen and Delmas 2010). This widely

used and accepted way of measuring CSP by using KLD databases shows that it is an

accurate and sufficient measurement and will thus be used in this paper too. KLD rates

28

in this research the following dimensions will be of importance: Corporate governance,

community, diversity employee relations, environment, human rights and product. Although

using KLD to measure CSP is widely accepted, the way to measure it has differed in the past.

Whereas at first, researchers have measured it by detracting the strengths from the concerns

and thus obtaining one overall score for CSP (Turban and Greening 1996). Collapsing KLD

scores into a unidimensional index may mask the individual dimensions that are of relevance

to measuring CSP. It is thus not a proper way to compute all the different dimensions into one

single score (Johnson and Greening 1999). According to Mattingly and Berman (2008),

positive and negative social actions are both empirically and conceptually distinct construct.

For this reason, these constructs should not be combined. This means that detracting negative

from positive scores and thus creating one overall score for CSP is not accurate. In

accordance to these findings, and the findings of this research, which will be elaborated upon

later, the choice is made to keep the CSP strengths and concerns of the KLD database

separate, and thus perform analysis with these separately.

For each company the KLD-index divides scores in groups of strengths and concerns.

Taking into consideration that I use both KLD

scores from 2011 and 2012 this will both be

shown in this section. In the sample of 2011, the

number of KLD strengths a company has can

have a minimum of 0 and a maximum of 20 (M

=7.02, SD = 5,25). Figure 1 shows that the CSP

Total Strengths of the companies from the

sample are not normally distributed. The

Shapiro-Wilk test shows that this too (W(131) = .931, p < .05). A high frequency is noticeable on the

29

low scores, it could thus be stated that the distribution of the strengths in 2011 is skewed to

the right.

Regarding the CSP Total Concerns of 2011, the following is the case. The minimum

score a company from the sample of 2011 could

get was 0, the maximum CSP Total Concerns

would be 15 (M = 3.64, SD = 2.62). In both the

CSP Total Strengths and the CSP Total

Concerns, the variance exceeds the mean. This

will be of importance when analysis of the data

is being done. The CSP Total Concerns of 2011

are distributed as figure 2 shows. As was also the

case with the CSP Total Strengths of 2011, also here a distribution can be noted that is

skewed to the right. The Shapiro-Wilk test shows that there is no normal distribution (W(131)

= .882, p < .05). Taking in consideration that both are not normally distributed, means that

this should be taken into account when

analysing the data.

In the sample of 2012, the number of

KLD strengths a company can have a minimum

of 0 and a maximum of 17 (M =7.02, SD =

5,25). Figure 3 shows that the CSP Total

Strengths of the companies from the sample are

not normally distributed. The Shapiro-Wilk test

Figure 2: Distribution of KLD concerns. 2011

30

shows that there is no normal distribution (W(131) = .954, p < .05). Although less skewed to

the right than a year earlier, high scores are noticeable between 0 and 10, and hereafter, the

frequencies of the scores are lower.

Concerning the CSP Total Concerns of 2012 the

following can be stated. The minimum a

company can score here is 0, the maximum a

company can score is 13 (M = 2.15, SD = 2.4).

Figure 4 shows the distribution for the CSP

Total Concerns of 2012. The Shapiro-Wilk test

shows that there is no normal distribution

(W(131) = .814, p < .05). What can be noted here, as

was also the case in the previous distributions, that the distribution is skewed to the right.

This being constantly the case with the distributions, it may imply that similar analysis

technique would be sufficient in these cases.

Stakeholder Engagement

A clear and universally accepted measurement of SE is not present in the literature. Even

more, a quantifiable measurement of SE seems to be absent. Ayuso et al (2011) used two

questions to establish the degree of SE; one addressing the number of different external

stakeholders regularly addressed through satisfaction surveys or perception studies

(stakeholder scope) and one question about the number of mechanisms used by the company

for engaging with these stakeholders (stakeholder processes). Taking the time and

possibilities of this study in consideration makes this method not appropriate. Manetti (2011)

31

uses GRI reports. Although these reports are regarded as being trustworthy, the problem is

their availability. Some company’s GRI reports are available, but not all. A second point of

attention here is that this not necessarily entails a quantifiable measurement, for GRI reports

in itself do not give available usable scores of SE. Several authors(Gao and Zhang, 2006;

Friedman and Mills, 2006; Michelon, 2001), do make a ordinal scale out of SE, most of them

ranging from low to high SE. In this study, the ordinal scale of Ghao and Zhang (2001) will

be used as a guiding scale. What the different levels exactly denote will be explained below,

but it should be stated that this scale is not yet a measurement in itself. It entails a possibility

to qualitatively sort companies into one of these four levels of SE.

The strategy used in this paper was to make use of annual reports, CSR reports, other

reports or other information from the website. From these sources, clues of evidence were

being found of the degree of SE. Some examples that linked different quotes to SE are the

following:

SE level Exemplary quotes / explanation

SE level: 1 a company would fall into this category when there was no mentioning of stakeholder engagement or anything of its kind. This was the lowest

stakeholder category because companies in the S&P 500 are prohibited to

provide information about the achievements and actions of the company.

SE level: 2 “I would like to hear from even more voices throughout the company and

from outside stakeholders”

“Closing, I assure our stakeholders that we are committed to balancing ... challenges, and we welcome your thoughts and engagement on these”

32

suppliers, nongovernmental and nonprofit organizations and professionals in

industry, government, labor and education in a variety of informal and formal

communications."

"We routinely meet with external stakeholders to help us identify issues most

material to our business operations and to maintain our license to operate."

SE level: 4 "Positive relationships with employees and business partners help us to improve efficiencies, cost and quality, and allow us to develop and to

innovate. Effective two-way communication with our customers, dealers and

other stakeholders helps us to understand and deliver the products that

customers want"

“We believe active stakeholder engagement is a critical driver of performance and our continued improvement and success.”

Table 2: Examples of SE levels

Stakeholder engagement thus entails the independent

variable in this research. SE is being measured on the

above-described scale of 1 to 4. The score entail the following, 1: the

company has a passive attitude and merely gives information 2:

The company actively listens to its stakeholders, they are

consulted, 3: there is a two-way process; stakeholders are engaged

in dialogue with the company, and 4: a proactive attitude is taken

by the company and management is driven by stakeholder’s input.

By using annual reports, CSR reports, other reports and/or information on the websites,

companies are categorized into one of these four categories. The data is being collected for Category Frequency 1 39 2 13 3 66 4 15 Total 133

33

the year 2011, which means that information of 2012 was not in the data collection. The

following descriptive statistics are of importance to the SE. Of the total sample of 135,

information regarding SE could be collected for 133 companies. A minimum score could be

obtained of 1, the maximum score was 4 (M = 2,43 ST = 1,03), in accordance to the

categories explained above. In Table 3, frequencies are presented. What shows from this table

is that most companies engage with stakeholders in a two-way process, this is nearly half of

all the companies. Hereafter, the most common degree of SE is merely giving information.

What can be stated reasoning from this table is that companies most of the time seek input

from stakeholders. If they decide to actively interact with stakeholders, they do so in active

way but their decisions are most of the time not driven by stakeholders.

Multinationality

Multinationality can be measured in many different ways. Some examples are ratio of

foreign assets to total sales (Daniels and Bracker, 1989, Ramaswamy, 1989, in Kotabe,

Srinivasan and Alaukh, 2002), subsidiaries in foreign countries (Tallman and Li, 1996) or

foreign to total employment ratio (Kim, Hwang and Burgess, 1989, in Rugman, Yip and

Jayaratne, 2006). Though many legitimate ways of measuring multinationality exist,

according to Rugman, Yip and Jayaratne (2006), the majority of the studies since 1970 have

used foreign to total sales as an indicator of multinationality. Because of this widely used

mechanism, and the convenience of this measurement taking the time and possibilities of this

34

Multinationality is measured here, as is explained here above, by the ratio of foreign

sales to total sales. This measurement results in a continuum scale of 0 to 1, whereas 1 entails

maximum multinationality, whereby all sales are made in foreign countries, and 0 entails a

company that only sales in its domestic

country, in this case representing minimal

multinationality. In this sample, scores could

be taken from 131 companies, and

multinationality can score at minimum 0 and

at maximum 0.85 (M = .349, SD = .260).

When taking figure 5 into consideration, one

thing is stands out, this is the overrepresentation of companies with a multinationality score

of 0. When doing a normality test, the Shapiro-Wilk test, shows that the distribution is not

normal (W(131) = .916, p < .05). When the companies with a score of 0 are removed from the

sample, a normal distribution is present. The

Shapiro-Wilk test shows this to (W(101) =

.977, p = .079). In figure 6 it is also shown,

that when 0 scores are removed from the

analysis, the scores are normally distributed.

But in this paper, the option is being chosen to

include all 131 companies into the analysis.

Figure 3: Distribution of MN scores, 2011

35 Control variables

With the selection of the sample, the influence of the type of industry on CSP is being

is being minimalized, because the nine largest industries are equally present in the sample.

Other company characteristics that might influence CSP may be company size and profit

(Stanwick and Stanwick, 1998). These two are being taken into consideration. Company size

is measured by taking the total amount of assets of each company, profit is measured by

obtaining the gross profit of each company using WRDS as a ratio to sales. The first control

variable, size of the company, is being measured by Total Assets, this can take a minimum

score of 1311 and a maximum of 2129046 (M=81233, ST = 265239). This variable is not

normally distributed (W(134) = .282, p < .05). Because it is highly skewed, a log

transformation is needed.

The second control variable is the ratio proft/sales, this can take a minimum score of ,03 and

a maximum score of 1 (M = ,4108, ST = ,21737). Also this control variable is not distributed

equally (W(134) = .960, p < .05).

Industries

In the sample, nine different industries are taken into account. Table 4 shows the

different values for the dependent and independent variables per group of industries. The

values are shown for the mean and the standard deviation. What shows is that ‘Health Care’ scores the highest on SE. Health care also scores high on CSP strengths of 2011, but what is

odd, is that the difference with CSP strengths 2012 is big, whereas in 2012 the scores are

much lower for CSP strengths. Concerning CSP Total Strengths of 2011, the industry that scores the highest is ‘Consumer Staples’. The ‘Financials’ industry scores the lowest on SE, with almost half the level of SE as the ‘Health Care’ industry average. It should be mentioned

36

that the sample of each industry is not of a sufficient size (N is 14/15) to make complete

accurate judgments, but it could be an opening for further research.

Table 4: Variables per industry

Regression Analysis

In this analysis a binomial regression analysis will be used to test the relationship

between the variables. Taking into consideration that the data of the dependent variable is

counted, the choice will be best made between a ‘Poisson regression’ and a ‘Negative binomial regression’. The second, negative binomial regression analysis can be regarded as a

Industry N Mean Std. Deviation

Consumer Discretionary SEscore 15 2,27 1,223 TS2011 15 7,8667 5,95059 TS2012 15 4,7333 3,93640 Consumer Staples SEscore 15 2,87 ,915 TS2011 15 10,0667 4,41534 TS2012 15 6,8000 3,07525 Energy SEscore 15 2,33 ,976 TS2011 14 4,4286 3,87724 TS2012 15 5,0667 3,76955 Financials SEscore 15 1,73 ,961 TS2011 15 5,4667 5,12510 TS2012 15 4,6000 3,60159 Health Care SEscore 15 3,20 ,561 TS2011 14 9,0000 6,66795 TS2012 15 5,1333 4,08598 Industrials SEscore 15 2,20 ,941 TS2011 15 4,9333 4,36654 TS2012 15 5,2667 4,58984 Information Technology SEscore 15 1,93 ,961 TS2011 15 7,2667 6,04113 TS2012 15 5,1333 4,15532 Materials SEscore 14 2,71 ,994 TS2011 13 7,5385 5,14159 TS2012 14 7,2857 4,06540 Utilities SEscore 14 2,64 1,008 TS2011 15 6,6667 3,51866 TS2012 15 6,2667 3,28344

37

generalization of Poisson regression and it has an extra parameter to model over-dispersion.

Over-dispersion is the case when the conditional

variance exceeds the conditional mean. When this is

tested in SPSS, table 5 is the result. What can be seen in

the figure is that the variance does exceed the value of

the mean. This leads to the conclusion that a negative

binomial regression analysis would be most appropriate

here to measure the regression between the variables. In

accordance to all what is said above, the regression

analysis will be performed four times, these analyses will not differ from each other except

that the dependent variable will be different.

Hereafter, a test will be performed to check reversed causality. It could be the case

that SE and CSP only co-vary instead of any causality between the two. By testing reversed

causality any statements regarding the relationship between the two can be made with more

certainty. Mean Variance TS 2011 7.02 27.58 TC 2011 3.64 6.88 TS 2012 5.57 14.83 TC 2012 2.15 5.81

38

Results

To test the hypotheses, correlation and regression analyses are performed using SPSS.

The results of these tests will be discussed in the following sections. First, the correlation

between the variables will be shown, hereafter, the hypotheses will be tested in more depth

using regression analyses.

Correlations

In this section, the correlation between the variables will be tested. The variables for

which correlation will be tested are SE, CSP, MN and the control variables. A distinction will

be made between the two different dependent variables, namely CSP strengths and concerns

of 2011 and CSP strengths and concerns of 2012. The independent variable stays the same, as

is mentioned earlier, this is the SE of 2011. By testing it this way, a statement could possibly

be made if the effect of SE on CSP shows delay. Because the variables are not normally

distributed, and the independent variable is of an ordinal scale, a Spearman’s correlation will

be used. In Table 6, the correlation is shown.

Table 6: Correlation between variables ** Correlation is significant at the 0.01 level (2-tailed) * Correlation is significant at the 0.05 level (2-tailed) 1. 2. 3. 4. 5. 6. 7. 8. 1.SE . 2. CSP TS2011 .52** . 3. CSP TC2011 .102 .25** . 4. CSP TS2012 .56** .68** .40** . 5. CSP TC2012 .35** .51** .63** .54** . 6. MN -.03 .09 .06 .04 .06 .

7. Total Assets (log) .18* .56** .44** .57** .53** .03 .