Does the location of CSR and CS activities
affect firm value?
An event study of US-listed firms
1 Msc. International Business and Management
University of Groningen Faculty of Economics and Business
9700 AV Groningen The Netherlands
31 August 2011
Master Thesis
Does the location of CSR and CS activities affect firm value?
An event study of US-listed firms
Rijnstraat 1b 9725 EP Groningen
+31 (0)651804011 J.J.Vlaming@student.rug.nl
s1932691
1st Supervisor: Dr. A.A.J. van Hoorn
2
Abstract
This paper studies the effect that corporate social responsibility (CSR) and corporate sustainability (CS) announcements have on corporate financial performance (CFP). This study pays special attention to the development level of the location where the activity takes place. Using the event study methodology 95 cases of CSR and CS announcements are analyzed during the period of the 1st of January 2000 till the 31st of December 2010 of firms listed on US stock exchanges. The findings illustrate that CS, and in to a lesser extent CSR, announcements positively affect firm value. Furthermore, CSR and CS activities in developed countries are found to be positively related to firm value and negatively related in undeveloped countries.
3
Acknowledgements
You are about to read my master thesis which marks the completion of my International Business and Management master. However, I would not have gotten so far without the help and support of several people, which I would like to thank.
First of all my thesis supervisor Dr. A.A.J. van Hoorn, without his guidance and directions this thesis would not be the thesis it is today. Thank you André for always being available and providing quick and constructive responses to my questions or problems.
I would also like to thank Mr. S.F.M. Beckers, MSc. who informed me in the use and application of the event study.
And last but not least, I would like to thank my girlfriend Laura for her undisputed support and assistance on matters which I was not sufficiently knowledgeable of.
Jordy Vlaming
4
Table of Content
List of Abbreviations ... 5
1. Introduction ... 6
2. Literature Review ... 9
2.1 CSR, CS and the Triple bottom line ... 9
2.2 The History of CSR and CS ... 12
2.3 Social and Environmental Sustainability and CFP ... 13
2.3.1 Empirical Evidence ... 14
2.4 Location and Social and Environmental Sustainability ... 18
3. Theory and Hypotheses ... 19
3.1 Hypotheses ... 19
3.2 Conceptual Model... 22
4. Methodology and Measurement... 24
4.1 Methodology ... 24
4.2 Measurement ... 24
4.3 Sample Selection ... 27
4.3.1 Descriptive Statistics ... 29
4.4 Calculation of the Dependent Variable ... 30
4.5 Empirical Model ... 32
5. Results ... 34
5.1 Basic Results ... 34
5.2 Moderating Effect of Location ... 36
5.3 Moderating Effect of the Type of SES Activity ... 40
5.4 Regression Analysis... 41
6. Discussion... 43
6.1 Summary of Findings... 44
6.2 Discussion of Findings ... 46
6.3 Research Limitations and Suggestions for Further Research ... 51
7. Conclusion ... 54
References ... 56
5
List of Abbreviations
AR - Abnormal Return
AAR - Average Abnormal Return CAR - Cumulative Abnormal Return
CAAR - Cumulative Average Abnormal Return CFP - Corporate Financial Performance CS - Corporate Sustainability
CSR - Corporate Social Responsibility
6
1. Introduction
Over the last three decades the pressure on firms to take on socially and environmental responsible activities has increased due to the growth of the competitive global market and the increasing consumer awareness (Elkington, 1994). This pressure is put on firms by their stakeholders, stakeholders are a company‟s customers, employees, suppliers, local communities, governments and shareholders (Ahmed and Uchida, 2009). Hartman, Rubin and Dhanda (2007) argue that this pressure to act sustainable is especially true for American firms. This because US investors show particular interest in corporate communication and a majority of these investors review corporate citizenship prior to an investment or purchase. This indicates the importance of corporate investments in social and environmental activities and the communication to the public.
Zadek (2004) in his study on corporate citizenship displays the five stages of organizational learning in which firms adopt sustainable policies. First is the defensive stage in which firms often deny responsibility, the second is the compliance stage in which firms adopt practices seen as the cost of doing business, in the third managerial stage responsible practice is integrated in everyday business to sustain long-term gains, the fourth strategic stage implies the implementation of sustainable practice in the core business strategies, finally in the civil stage firms encourage industry wide participation (Zadek, 2004). The largest firms, under scrutiny of the public eye, are the first to adopt these stages and more and more firms follow in the implementation of sustainable activities. However, the terms under which these activities take place are converging (Willums 1999; Montiel, 2008; Portney, 2008).
7 The direct relationship between SES activities and corporate financial performance (CFP) has often been the topic of management studies with many different outcomes. The reason for these different outcomes is because corporate social responsibility is a vague and ever changing description and does not have a standard definition or fully recognized set of specific criteria (Ahmed and Uchica, 2009). This ever changing definition of CSR resulted in studies with many different proxies for CSR. Besides these mixed results multiple meta-analyses conclude that acting „good‟ is in the best interest of companies (Orlitzky, Schmidt and Rynes, 2003 and Margolis, Elfenbein and Walsh, 2007). This is particularly true for CSR and in a lesser extent for CS. However, much has happened in recent years and people today are more conscious to environmental issues than ever before. This study provides insight into investors‟ appreciation of both social and environmental announcements and will observe whether priorities have shifted over the past years.
When reviewing research papers on CSR and CS a vast information gap appears when the question is raised whether the influence of the location of the SES activity has effect on firm value. Most CSR and CS studies are conducted in the context of western regions with a focus area on these same countries, often excluding emerging and developing countries. The focus on where is the main topic of this research paper. This leads to the following research question:
Does the location of CSR and CS activities affect firm value?
8 existing around business conduct within developing countries and how this relates to firm value as well as the usage of SES as a strategic tool.
9
2. Literature Review
The following chapter reviews academic literature written on the subject of this research paper. It starts with several definitions of often used terms and provides a short history of these terms. Furthermore, this chapter will go in-depth on previous articles on the relationship between acting social and environmental sustainable and the effect is has on Corporate Financial Performance.
2.1 CSR, CS and the Triple bottom line
In the early 70‟s, when management literature on both corporate social responsibility and corporate sustainability picked up, the term CSR was preferred when taking social issues into account. When using environmental issues CS was more widely used Montiel (2008).
The definition of corporate social responsibility (CSR) varies among authors and has been intensively reviewed by a variety of researchers such as Montiel (2008). Since the 70‟s and 80‟s only a few authors (Maignan and Ralston, 2002; McGee, 1998 and McWilliams and Siegel 2001) paid attention to the essence of the phenomenon CSR, this can be observed by the definition formulation of CSR in Table 1. A possible reason that the term CSR has not been redefined is due to the emerging of a new term, Corporate Social Performance (CSP). The conceptual model of CSP further elaborates on the theory of CSR and ties this with social measurements. This model was introduced by Carroll in 1979, during the 80‟s and 90‟s most studies on CSR were measuring social performance indicators (CSP) and relating them to firm performance.
Table 1: Definition overview CSR by Montiel (2008: 253-254)
Author Definition
Elbing (1970) Social Responsibilities of businessmen. Describes the social responsibility framework (businessman has a responsibility more important than profit maximization), opposed to the economic framework (businessman has one singular responsibility to maximize profits of its owners).
10 requirements (p. 312).
Hay and Gray (1974) Social responsibility of business managers, responsibilities that extend beyond the traditional economic realm of profit maximization or merely balancing the competing demands of the sundry contributors and pressure groups (p. 137).
Mears and Smith (1977) Social Responsibility. Responsibility of the firm to the public, employee and consumer and responsibility of the employee to the firm.
Carroll(1979) Aupperle, Carroll and and Hatfield (1985); Tuzzolino and Armandi (1981)
Social Responsibility. It must embody the economic legal, ethical and discretionary categories of business performance because of the need to address the entire range of obligations business has to society.
McGee (1998) Corporate Social Responsibility. It states the ambiguity of the CSR concept, sometimes defined in purely economic profit-making terms or as socially oriented in a proactive social responsiveness view.
McWilliams and Siegel (2001)
Corporate Social Responsibility. Actions that appear to further some social good, beyond the interests of the firm and that which is required by law (CSR is beyond obeying the law) (p. 117).
Maignan
and Ralston (2002)
Corporate Social Responsibility. Conceptualized as motivating principles (driven by values, stakeholders, performance); processes (programs and activities aimed at implementing CSR principles and/or addressing specific stakeholder issues, including philanthropic, sponsorships, volunteer, code of ethics, quality, health and safety and managing environmental impacts); and stakeholder issues (community, customer, employee, shareholders and suppliers).
The definition of CSR which is most often cited is that of Carroll (1979:500) who describes CSR as: „the social responsibility of business encompasses the economic, legal,
ethical and discretionary expectations that society has of organizations at a given point in time.’
11 Nations. This report included a definition of sustainability which was credited to Gro Harlem Brundtland the chairperson of the WCED and also the Prime Minister of Norway. The definition of sustainability as stated in ‘Our Common Future’ is as follows:
‘development that meets the needs of the present without compromising the ability of future generations to meet their own needs’ (UNWCED 1987, p. 43). This definition of
sustainability has changed little since its emergence as can be seen in Table 2.
Table 2: Definition overview of CS by Montiel (2008: 256)
Author Definition
Gladwin and Kennelly (1995)
Sustainable Development. Process of achieving human development in an inclusive, connected, equitable, prudent and secure manner. Sustainable development components are 1. Inclusiveness (environmental and human systems, near and far, present and future); 2. Connectivity (world‟s problems interconnected and interdependent); 3. Equity (fair distribution of resources and property rights); 4. Prudence (duties of care and prevention); 5. Security (safety from chronic threats) (p. 878).
Starik and Rands (1995) Ecological Sustainability. Ability of one or more entities, either individually or collectively, to exist and flourish (either unchanged or in evolved forms) for lengthy timeframes, in such a manner that the existence and flourishing of other collectivities of entities is permitted at related levels and in related systems (p. 909).
Banerjee (2003) Sustainable Development. States that the Brundtland definition is not really a Definition but a slogan. Emphasizes that sustainable development is managed through ethnocentric, capitalistic notions of managerial efficiency (sustainable capitalism).
Sharma and Henriques (2005)
Corporate Sustainability. Refers to Brundtland definition: development that meets the needs of the present without compromising the ability for future generations to meet their own needs.
12 As is argued by Montiel (2008) contemporary research shows that the terms CSR and CS are converging. In CSR studies increasingly more value is attached to the environmental aspect and in CS studies an increasing importance of social matters can be observed. This converging effect of social and sustainable matters shows that firms balance the three elements of the so-called Triple Bottom Line. These elements are the economic, social and environmental elements. The ideology behind the triple bottom line dictates that firms are not only judged by their financial performance, but also how a firm performs when taking social and ecological issues into the equation (Elkington, 1994, 2004).
2.2 The History of CSR and CS
Blowfield and Frynas (2005) discuss the origin of CSR and find that although the term may be relatively new, the philosophy behind the term can be traced back for many centuries. The Islamic and Christian Church in medieval times condemned certain business activities, which could be considered unsocial. The largest distinction from previous social believes according to Fabig and Boele (1999) is that the current umbrella definition of CSR is a connection between social development, environment and human rights and is more global in scale. Blowfield and Frynas (2005) also argue that CSR can be approached from a cultural perspective, in which the term CSR can be found to be specifically Anglo-Saxon. This can also be found in the current research on CSR which are predominantly conducted in an Anglo-Saxon context. Nevertheless, this does not mean that Continental European, Asian and African societies do not had ties between business and the wider community. The term and means of their social activities were different as they were often tied to religion.
13 resources should be limited. The second wave was from 1984 till 1995, increased realization that sustainability requires new technology. According to Elkington (2004) we are currently in the third wave, which started in 1995. This wave brings forth growing acknowledgement that sustainable development needs change. These changes have to occur in the governance of corporations and the entire process of globalization. Governments should oversee these changes. Appendix I displays a figure and a more detailed description of these waves.
2.3 Social and Environmental Sustainability and CFP
Ever since the 1960‟s questions have been raised around the current term of CSR and what exactly is a firm‟s obligation towards social matters. Keith Davis posed two questions, „Can business afford to ignore social responsibilities?‟ (1960) and „What does
the businessperson owe society?‟(1967). According to Schwartz and Carrol (2003) these
two questions portray the two schools of thought on the field of corporate social responsibility and in a broader sense strategic management as a whole. The first group says that business has a wide range of obligations towards society. By taking into account others then just the owners of a company the term stakeholders emerged, a clear play on the word „stockholder‟ (Freeman, 1984). Freeman‟s definition of stakeholders is “any
group or individual who is affected by or can affect the achievement of an organization’s objectives” (1984:5). Adopting this stakeholders approach could lead to an agency
problem. The agency theory dictates that the principal, the shareholders, hires the agent, the managers, to act on their behalf and in their interest (Shleifer and Vishny, 1997). This separation of ownership and control could lead to conflicting interests when taking into account stakeholders other than the shareholders. Freeman (1984) argues, however that managers should be aware of the concerns of all stakeholders to ensure success on the long term.
14 a social „tax‟. This tax will eventually be paid by the shareholders, customers or employees. Friedman argues that such matters should be managed by either the individuals themselves or the state at large. Despite this negative perspective on CSR an increasing amount of firms adopt the stakeholders approach and invest more in socially responsible projects (Chatterji, Levine and Toffel, 2007). This increase in CSR activity can be observed in the growing amount of business publications on firms taking steps to become greener; by reducing their carbon footprint or pollution reduction of air and water; to invest in their workforce; by ensuring health and safety measures and more contributions, financially or otherwise in communities (Portney, 2008).
2.3.1 Empirical Evidence
The relationship between corporate social responsibility (CSR) and corporate sustainability (CS) and corporate financial performance (CFP) has been the topic of many studies with mixed results. The first empirical study was done by Bragdon and Marlin (1972). The motivation behind Bragdon and Marlin‟s study was to find out whether or not a company could do well by doing good, or that virtue must be its own reward. Their result was that CS was positively correlated with CFP, however they only focused on the environmental aspects of sustainability. The relationship between acting sustainable and firm performance was to be researched many times over with different measurements and definitions for CSR and CS (see 2.1). A reason for these different outcomes is due to the fact that corporate social responsibility is a vague and ever-changing description and does not have a standard definition or fully recognized set of specific criteria (Ahmed and Uchida, 2009). Ahmed and Uchida (2009) distinguished two common measures of financial performance, accounting returns and investor returns.
15 indicators often uses the event study methodology to assess the financial impact of CSR activity on the short term (Dam, 2008).
Studies which measure performance by accounting returns examines the relationship between some measure of corporate social performance (CSP), a measure of CSR or CS and long term firm performance, often using ratios such as return of asset (ROA), return on investment (ROI), return on equity (ROE), return on sales (ROS). The results from both types of studies have been both positive and negative, which creates vagueness on the relationship between SES and financial performance.
Dam (2008) performed a review of 68 studies on the relationship between CSR and CFP and distinguished three different types of CFP measurements. The first set of studies measure Market-to-Book value, the Tobin‟s q. The second set of studies measure accounting returns, the third set of studies review investor returns. The main findings of this review are portrayed in Table 3. For the full outcome of the literature review of Dam (2008) see Appendix II.
Table 3 Literature overview of CSR on CFP by Dam (2008)
CFP indicator Number of studies Positive relation Negative relation Mixed relation No relation Market-to-book 5 5 (100%) 0 (0%) 0 (0%) 0 (0%) Return on Assets 36 27 (75%) 0 (0%) 0 (0%) 9 (25%) Stock market returns 27 7 (26%) 9 (33%) 3 (11%) 8 (30%)
Total 68 39 (57%) 9 (13%) 3 (5%) 17 (25%)
16 Table 4: Recent event studies and their findings
Author Year Independent variable
Control variables
Location Abnormal Return Jones and Murrell 2001 CSR
announcements
Firm fixed effects
US Positive
Jones and Rubin 2001 CS
announcements Firm fixed effects US Positive Tsoutsoura 2004 CSR announcements
Risk, size and industry
US Positive
Curran and Moran 2006 CSR and CS Rating Index
N/A UK No
Relation
Nicolau 2008 CSR
announcements
N/A Spain Positive
Takeda and Tomozawa
2008 CS
announcements
Year, rating and industry Japan Negative and Positive Becchetti, Ciciretti and Hasan 2009 CSR and CS Rating Index Financial distress shocks and stock market seasonality US Positive
Arya and Zhang 2009 CSR
announcements Industry, age, size, leverage, Multinationality. South Africa Positive Cheung 2011 CS Rating Index Year, size, industry US No Relation
This increase in positive abnormal returns in event studies may indicate increased consumer awareness since 1997 and improved communication channels. Communication channels such as the internet, which allows companies to communicate more information cheaper and faster than ever before (Wanderley, Lucian, Farache and Filho, 2008).
17 analyzed 27 event studies, in which the markets‟ reactions on socially irresponsible or illegal behavior were measured. Frooman (1997) finds significant evidence that socially irresponsible behavior results in shareholders wealth reduction.
Ahmed and Uchida (2009) provide a conceptual framework which dictates how CSR is linked to CFP. This conceptual framework visualizes the line of reasoning of the meta-analyzes of Orlitzky, Schmidt and Rynes (2003) and Margolis, Elfenbein and Walsh (2007). This conceptual framework is displayed in Figure 1. In Ahmed and Uchida‟s (2009) framework the term CSR also includes environment, as is discussed in chapter 2.1, this is another indication of the converging of the terms CSR and CS in the academic literature.
18
2.4 Location and Social and Environmental Sustainability
Belal (2001) argues that most CSR studies are conducted in the context of developed countries, primarily the United States and Western Europe. A review of the geographical focus of the samples from the meta-analyses by Frooman (1997), Orlitzky, Schmidt and Rynes (2003) and Margolis, Elfenbein and Walsh (2007) confirms this.
The differences among the CSR policies within these regions also heavily differ. The philanthropic model is dominant in the US, where companies donate a certain share of their profits to charity. In the European model operating the core business in a socially responsible manner is prevailing, complemented by investments in communities (Albareda, Lozano and Ysa, 2007). This European model is believed to be more successful on a long term perspective (Belal, 2001), due to the fact that when economic difficult times arise donations are often the first to cut back on. The focus area of CSR studies are primarily on companies conducting CSR activities within these developed regions. Husted (2003) argues that firms tend to focus social action on the communities where they already have operations. This geographical concentration of CSR and CS activity is so the firm can optimally benefit from the image creation among the employees and local community.
From a marketing perspective, Russell and Russell (2010) find that consumer behavior is heavily dependent on egocentric tendencies. These tendencies can be recognized when the location of the CSR and CS activities is studied, according to Russell and Russell (2010) consumer‟s purchases reflect how the company is perceived. And these consumers have a stronger positive perception of firms which are active in social projects in their direct environment.
19
3. Theory and Hypotheses
The previous chapter reviewed academic articles on how CSR and CS relates to corporate firm performance. Much research has been done on this subject, but still some research gaps remain. This chapter provides the theoretical argumentation which explains the authors trail of thought and which led to the proposed hypotheses.
3.1 Hypotheses
The term „Social and Environmental Sustainability‟ (SES) is introduced due to the convergence of the terms corporate social responsibility and corporate sustainability. SES is viewed in this research as the complete set of policies, practices and programs that are integrated into business operations, supply chains and decision-making processes throughout the corporation and includes issues related to business ethics, community investment, environmental concerns, governance, human rights, the marketplace as well as the workplace (Environmental Management Centre, 2005). Such a broadly formulated definition can be assumed due to the fact that this research does not measure SES, it merely observes the announcement of social and environmental activities in corporate communication.
The first hypothesis is based on the theory that SES activities will lead to costs on the short-term, however these costs will pay off on the long-term (Baumol, 1970). This positive long-term return will be achieved through greater social legitimacy, customer and employee loyalty and less government interference. This in turn will lead t increased autonomy, which benefit corporations (Burke and Logsdon, 1996).My assumption is that investors foresee this long-term profitability and therefore value SES investments.
20 H1: SES announcements have a positive effect on firm value.
My assumption for the second hypothesis is that location has a positive moderating effect on the relation between SES announcement and firm value (Bennet, 2000), when announcements of SES activities take place in a developed location. I assume this, because consumers have a more positive association with firms who are socially and environmentally active in their own region. This is reinforced by Russell and Russell‟s (2010) results. These authors find, that people are egocentric of nature and thus react stronger to activities within a short geographical range. The study of Russell and Russell (2010) is performed with the main focus being on the consumer and regarding consumer purchasing behavior with respect to a company‟s CSR policy. When consumers are expected to act selfishly and will reward local or even domestic CSR activities and perhaps also CS activities, investors will also have a positive association with the announcement of local SES activities. This is due to the fact that investors speculate on future cash flows, thus consumer behavior. This results in a stronger positive reaction on announcements to projects in developed markets than announcements of projects which are located in underdeveloped or developing markets. This line of reasoning also follows that of Husted (2003), who argues that firms focus social activities in regions where they already have operations, to benefit most from the image creation.
A final reason why CSR or CS activities in undeveloped countries will be less effective is due to information disclosure. When conducting CSR or CS activities firms want the public to know, because this will often lead to an increase in reputation and thus an increase in future business. CSR and CS activities in undeveloped countries it is believed to have a less positive effect than when conducting similar activities in developed regions due to more advanced communication channels (Hamilton, 2003; Wanderley, Lucian, Farache and Filho, 2008). Also as noted by Jamali (2007) the media in undeveloped countries have a less controlling role than in developed countries.
21 findings may be explained by differences in the institutional environments. However, to the best of my knowledge no previous event study has considered the location of the SES activity to have a moderating effect on CFP. Summarizing these arguments leads to the following hypothesis:
H2a: Announcements of SES activities in developed countries have a stronger
positive effect on firm value than announcements of SES activities in undeveloped countries.
The research sample is separated in sub-groups of announcements in developed and undeveloped countries. To observe the individual effects of environmental and social announcements, the sub-groups are further divided in type of activity, namely CSR or CS. This results in sub-hypotheses H2b and H2c.
Milinski, Semmann and Krambeck (2002) find in their research that making public donations to charity programs which benefit people outside of one‟s social group is rewarded by people within the social group. This forms the foundation for sub-hypothesis H2b, which assumes that social initiatives in developing countries are rewarded stronger by the market than in developed countries.
The pollution haven hypothesis forms the basis for H2c. The pollution haven hypothesis, as is described by Levinson (2008:1), „is the idea that polluting industries
will relocate to jurisdictions with less stringent environmental regulations’. Foreign
direct investment from multinational operating companies in countries with low environmental regulation, does seem to be concentrated in high pollution sectors (Cole, 2003; Eskeland and Harrison, 2003; and Smarzynska and Wei, 2001). When polluting undeveloped countries is more accepted than polluting developed countries, it is likely that environmental activities in developed countries are higher rewarded on the market than activities in developing countries. This results in the following hypotheses:
H2b: Announcements of CSR activities in undeveloped countries have a
stronger positive effect on firm value than announcements of CSR activities in undeveloped countries.
H2c: Announcements of CS activities in developed countries have a stronger
22 If hypothesis H2b and H2c hold true then these hypotheses have an off-setting effect on the main effect of the country level of development on SES announcements.
As is discussed in the literature review, the term CSR was introduced in the early 1970s, while the term CS was introduced in the late 1980s. Firms and consumers alike had thus more time to get used to the term CSR than CS. This is reflected in the response to CSR announcements as the concluding words of the meta-analysis of Orlitzky, Schmidt and Rynes (2003:427) state: ’Corporate virtue in the form of social and, to a lesser extent,
environmental responsibility is rewarding in more ways than one.’ The meta-analysis
performed by Orlitzky, Schmidt and Rynes (2003) reviewed papers from 1972 till 1997. Now that the term CS had to the change to catch up for lost time it will be interesting to observe whether consumers and therefore investors, still value social over environmental activities or that perhaps this has shifted over the recent years.
There are reasons to believe such a shift has occurred due to the increase in environmental awareness driven by events such as the initiative for the Kyoto Protocol and the growth in green products (Cronin, Smith, Gleim, Ramirez and Martinez, 2010). This results in the following hypothesis:
H3: CSR announcements have a stronger positive effect on firm value than CS
announcements.
3.2 Conceptual Model
23 Figure. 2: Conceptual Framework hypotheses 1, 2a, 2b, 2c and 3
* The Developed location is used as a moderating effect in this model * * The CSR announcement is used as a moderating effect in this model
24
4. Methodology and Measurement
This chapter provides a description of the research methods used in this study by first explaining the general approach and then explaining the variables involved. After explaining how these variables are measured the data collection is addressed, here the conditions are given under which the sample is collected. Followed by the calculations of my dependent variable upon which my results are constructed in the following chapter.
4.1 Methodology
This research is an event study, as introduced by Fama, Fisher, Jensen and Roll (1969) and later developed by Brown and Warner (1985), an event study is a research method which helps to assess the financial impact of changes in corporate policy. The event study methodology is often used in financial studies. In an event study the ‘abnormal’ return on the stock price is associated with an unanticipated event. A return is considered to be abnormal when a stock outperforms its market, either positively or negatively. These abnormal returns are assumed to reflect the stock market's reaction to the arrival of new information (Geyskens, Gielens and Dekimpe, 2002). Financial theory dictates that a firm‟s stock price reflects the market‟s expectation of its future cash flow (Rappaport, 1987). The event study methodology relies on the presence of an effective market, which means that all available information is reflected in a firm‟s stock price. Shareholders update their share appreciation by buying or selling a share whenever new information enters the market. The significance of this event is determined by the strength of the rise or fall of the stock price (McWilliams and Siegel, 1997).
4.2 Measurement
Corporate Financial Performance. CFP can be measured in various different manners,
25 financial performance. Firm value can also be measured by estimating the expected future cash flows, measured by Tobin‟s q (Tobin, 1969). Tobin‟s q is defined as the ratio of the market value of an asset to the replacement cost of an asset, this measurement uses historical data and links them to current market valuation.
A third method of measuring a firm‟s financial performance regards the investor returns, this measurement takes the investor‟s perspective and is used in this research paper. This performance indicator makes it possible to study the short term effects of CSR or CS announcements on the net present value of a firm. Share price is the financial performance indicator which is normally used in event studies due to its highly fluctuant nature to corporate announcements (Clinebell and Clinebell, 1994; Hannon and Milkovich, 1996; Posnikoff, 1997; Worrell, Davidson and Sharma, 1991 and Wright and Ferris, 1997).
Location of SES activity. The Human Development Index (HDI) is used to measure the
26 Type of activity. By grouping the announcement as being either a social or environmental
orientated announcement it is possible to review the individual effects. Therefore, CSR activities are qualified as being a firm investment in social projects focusing on people and CS activities are ecological projects focusing on environmental issues (Zwetsloot, 2003) via dummy coding. Examples of social activities undertaken by the companies is my sample are; switching to fair-trade products, providing education, the donations of resources to help the needy etc. Some examples of environmental activities in my sample are; energy and/or waste reduction programs, recycling programs, water purification programs etc.
Control variables. The control variables are firm size and industry as commonly used in
previous research that related CSR and firm performance (McWilliams and Siegel 2000). Company size is an often used control variable since larger firms have more resources to implement SES policies, also larger firms have more incentive to act „good‟ since they draw more attention from stakeholders (Burke, 1986). Company size is measured by the number of employees and is continuous data. The number of employees is extracted from DataStream and includes both full and part time employees, but excludes seasonal employees. Due to the fact that the number of employees can drastically shift from one year to another and the announcements occur at any given time in a year, the average number of employees is used of the year in which the announcement takes place and that of the consecutive year.
27 are categorized as one of the following groups, Basic Materials, Conglomerates, Consumer Goods, Financial, Health Care, Industrial Goods, Services, Technology and Utilities. Industry grouping will be done through dummy coding. These data are retrieved from databases such as Orbis or Datastream.
4.3 Sample Selection
The sample selection for this study are all public firms, which are traded on US stock exchanges, who‟s SES activities can be found on Lexis/Nexis within the time period of the first of January 2000 till the thirty-first of 2010.
The choice for firms that are listed in the US is supported by the fact that companies in Anglo-American countries have the largest obligation to disclose firm information due to the dispersed ownership structures (Ferrarini and Moloney, 2005). American exchanges are commonly considered to be the most liquid markets with the strongest shareholder protection regulation of all financial markets (Roosenboom and van Dijk, 2009). Furthermore, research from Roosenboom and van Dijk (2009) about cross-listing on multiple stock exchanges demonstrated that companies listed on the US exchanges had significant higher announcement returns than companies that cross-listed on European or Asian stock exchanges. Because of these reasons a sample concerning US based firms is preferred over for example European based firms, where disclosure regulations are inconsistent and vague (Ferrarini and Moloney, 2005). Also, when the companies in the sample are all within the same country it excludes negative variations, which multi-country studies have to take into account, such as exchange rates and different institutional environments.
28 environmental nature, but are the core business of a firm. An example of this is when a waste management firm announces it will open a new recycling plant.
Thus, announcements which introduce the actions of a social or environmental nature will have the focus. Table 5 features an overview of the search words included in this study, it also displays the exclusion terms which would result in irrelevant announcements. These search terms and exclusion terms originate from the study done by Chabowski, Mena and Gonzalez-Padron (2011) on the structure of sustainability research in marketing, some alterations are made to ensure the focus of the announcements to this particular research.
Table 5: Overview of Search Terms
Inclusion search terms Exclusion terms Corporate social responsibility * (7) Socialism *
CSR project (10) Socialist *
Sustainability (2) Sustainable competitive advantage * Corporate sustainability (0) Corporate sustainability report Sustainability project (11) Index
CS project (0)
Triple bottom line * (0) People, planet, profit * (5) Environmental project (3)
Fair trade * (3)
Fair trade project (0)
Pollution * (3) Recycle (9) Recycling (7) Emission * (3) Donate (7) Donation (22) Charity (3)
* Original search terms from Chabowski, Mena and Gonzalez-Padron (2011)
Note: Values indicate the number of cases found per search word, adding up to the total sample.
29
4.3.1 Descriptive Statistics
The sample selection method described in chapter 4.3 provided a sample of 95 announcements of corporate social responsibility and corporate sustainability activities to be undertaken in both developed as undeveloped countries. Of these announcements 48 were of a social nature and 47 were of an environmental nature. Furthermore, 69 of these announcements were located in developed regions and 26 were located in underdeveloped regions as argued in chapter 4.4.1. An overview of the location and the type of announcements can be seen in Table 6, Appendix III provides the full sample.
Table 6: General overview of corporate announcements
Announcements CSR CS Total
Developed 30 39 69
Undeveloped 18 8 26
Total 48 47 95
The level of development of a country is determined by its HDI score. The average HDI is 0.784, with a standard deviation of 0.164. The lowest HDI score in the sample is 0.389, which is the average HDI for the continent of Africa. The highest HDI in this sample is 0.902 which is the United States, this is also the median. SES announcements in the US are observed a total of 52 times. This data is displayed in Table 7.
Table 7: Descriptive Statistics
Variable N Mean Std. D. Median Minimum Maximum
HDI score 95 0.784 0.164 0.902 0.389 0.902
Company Size* 95 140.670 285.035 56500 43 2.000.000
* Measured by the number of employees
30
4.4 Calculation of the Dependent Variable
This study collects the stock prices prior to the announcement of a SES activity and a certain time after the announcement. For collecting the historical stock prices DataStream is used. The selected timeframe, measurement model and data-analysis which are used in this research are in line with the explanation of the standard event study methodology (Brown and Warner, 1985; Geyskens, Gielens and Dekimpe, 2002; McWilliams and Siegel, 1997; McWilliams, Siegel and Teah, 1999).
The standard approach or the market model, relates the market index to the firm stock price over the same time period. According to the market model, the expected rate of return E(R it) would be the rate of return R it if the event would not have taken place. The normal or expected rate of return on the share price of firm i on day t is expressed as
E(R it) = αi + βi Rmt + εit (1)
where, E(R it) is the expected rate of return on the share price of firm i on day t, Rmt is the rate of return on a market portfolio of stocks (in this study the Dow Jones US market average index) on day t. The terms αi and βi are the regression coefficients for firm i during estimation period t and εit is the error term. The values for αi and βi are obtained by performing a linear regression of Rit on Rmt over an estimation period of 100 days. This estimation period starts 110 days till 10 days prior to the announcement date of the event. The estimation period and the event window are visualized in Figure 3.
Figure 3: Visualization of the estimation period and event window.
The abnormal rate of return of firm i on day t ARit is to be considered all values deviating from the expected rate of return E(R it). By using the calculated expected rate of return
Event day Event Day 0 Event Day -10 Event Day +10 Event Day -5 Event Day + 5 -110
Event Day Estimation Period
Event Day
31 the estimate of daily abnormal returns (AR) for firm i can be derived using the following equation:
ARit = Rit – E(R it) = Rit – (αi + βi Rmt + εit) (2)
The event window in this research is 10 days prior and 10 days following the occurrence of the event, in line with Mathur and Mathur (2000). In an ideal world the period prior to the event does not have to be included in the analysis. However, this is not an ideal world and information leakage does occur. By including a period prior to the event in the event window the possibility of information leakage is taken into the analyses (McWilliams and Siegel, 1997). The strongest deviation and within the shortest time period to the event date (day 0) is considered as the reaction to the announcement.
To analyze the significance of these abnormal returns the Patell (1976) t-statistic will be used. The Patell t-statistic is one of the more commonly used parametric test in event studies and is calculated by making use of average abnormal returns (Campbell, Cowan and Salotti, 2009). The average abnormal return for day t (AARt) is calculated by the sum of all ARit divided by the total number of observations (n) in the sample. The Patell t-statistic provides an overview on which day in the event window the effect of a CSR or CS announcement is the strongest. The expected return estimation period is indicated as
m and has a value of 100. The formula for the Patell t-statistic is the following:
(3)
32
4.5 Empirical Model
The calculations for the dependent variable mentioned in chapter 4.4 will have the main focus in this research. However, the Patell t-statistic is unable to include control variables and to provide an explanation on the interacting relations of the variables. For this reason the data mentioned above will also be analyzed using a linear regression. These regressions will provide the coefficients, which dictate the size and direction of the effect along with the significance of this effect. For the overall effect of SES on CFP the following regression equation is used:
ARit = α+ β1SES + β2 Industry + β3 Firm size + ε, (4)
where ARit is the most significant abnormal return of firm i on day t, ε is the error term and α is the intercept and β1, β2, β3 the coefficients for case i.
To test H2, which includes the level of development in which the activity takes place, and its sub-hypotheses the sample is split in two groups. This split is made in SES activities in developed and undeveloped countries. This results to the subsequent regression equation:
ARit = α+ β1 SES activity + β2 Location + (5) + β3 Industry + β4 Firm size + ε,
where ARit is the abnormal return of firm i on day t, ε is the error term and α is the intercept and β1, β2, β3, β4 the coefficients for case i.
The moderating effect of the type of activity will be included to test H3, which evaluates which type of announcements has a stronger significant effect on the correlation of SES activity and CFP.
ARit = α+ β1 CSR or CS activity + Β2 Industry + β3 Firm size + ε, (6)
33 The data is split by being either a CSR or a CS announcement and subsequent the variable location is included. This, to observe the effect of each type of activity in each of the two types of location. This results to the subsequent regression equations:
ARit = α + β1CSR activity + β2 Location (7) + β3 Industry + β4 Firm size + ε
and
ARit = α + β1CS activity + β2 Location + (8) + β3 Industry + β4 Firm size + ε
where ARit is the abnormal return of firm i on day t, ε is the error term and α is the intercept and β1, β2, β3, β4 the coefficients for case i.
34
5. Results
This chapter provides the results using the methods and data described in chapter 4. First the overall effect of SES on CFP will be shown and then the moderating effect of the location of the SES activity and the type of activity are demonstrated. This is done by tables in which the Patell t-statistic along with the average abnormal return is given as well as a graphic display of the AARs over the entire event window. The outcomes of my proposed hypotheses are provided shortly, in the next chapter I will further elaborate on these results. This chapter also holds the outcome of the regression analysis.
5.1 Basic Results
After having estimated the parameters (αi and βi) for each firm i, with Equation 1 using an estimation period of 100 days, the abnormal returns (ARit) can be calculated using
Equation 2. Table 8 and Figure 4 display the average abnormal returns (AAR) for the entire sample of 95 announcements on the event day, as well as 10 days prior and after the announcement date. Furthermore, the Patell t-statistic and the percentages of positive abnormal returns are displayed in Table 8.
On day -1 the percentage of positive abnormal returns is the highest within the event window with 60% of the 95 announcements providing positive returns. This correspondents with the AAR which is 0.25% on day -1 (p<.05). This percentage is absolute and relates to the previous trading day. Stronger effects do occur within the event window, such as on day -7 (p<.01) and on day +6 (p<.01). However, as is discussed in chapter 4.7, the statistically significant AAR closest to the event date will be assumed to be the arrival of new information to the market (McWilliams and Siegel 1997).
35 Table 8: Average Abnormal Return for SES announcements
Event Day Average Abnormal Return (%) Patell t-statistic Percentage of positive Abnormal Returns -10 -0.055 -0.532 49 -9 0.015 0.152 47 -8 0.042 0.413 49 -7 0.333 *** 3.221 58 -6 0.014 0.144 48 -5 -0.012 -0.123 52 -4 0.097 0.944 49 -3 0.015 0.149 46 -2 -0.056 -0.549 54 -1 0.252 ** 2.439 60 0 0.040 0.389 56 +1 0.052 0.509 46 +2 -0.063 -0.610 53 +3 0.167 1.620 58 +4 0.185 * 1.787 46 +5 -0.084 -0.816 42 +6 0.303 *** 2.926 52 +7 0.055 0.531 48 +8 -0.006 -0.067 52 +9 0.042 0.405 52 +10 -0.280 *** -2.703 37 * p < .10, ** p < .05, *** p < .01
Figure 4: Average Abnormal Returns for SES announcements
Note: Significance lines are approximations.
On top of the statistical relevance of these CAAR values one could consider the economic importance of the positive effects of SES announcements. To demonstrate this
36 effect the market value increase for a medium-sized1 firm in this research sample is given. The market value of Chevron would rise from 100.114.587 $ to 100.454.977 $ in 3 days with a cumulative abnormal return of 0.34%. Table 8 and Figure 4 indicate that the first hypothesis is supported, SES announcements have a positive effect on firm value.
5.2 Moderating Effect of Location
The overall effect of SES activities in developed and undeveloped regions combined are found to be positively related to firm performance. Dividing the SES announcements in activities in developed countries and activities in undeveloped countries enables us to observe the effect that the level of development of the location has on the relation between SES and CFP. Of the total sample of 95 cases, 69 cases were located in developed countries and 26 cases were located in undeveloped countries. Table 9 displays the AAR and Patell t-statistics for the sub-samples Developed and Undeveloped SES activities. Figure 5 provides a graphical presentation of the AARs of both sub-samples.
As can be observed in Table 9 and Figure 5 announcements of SES activities in developed countries have the expected positive effect on firm performance on several days in the event window. The most significant CAAR closest to the event date is that of the day prior to the event date till the event date (p<.01). This indicates that firms announcing a social or environmental activity in a developed country have an announcement return of .45% within these 2 days. Furthermore, Figure 5 also illustrates that this indicates stable growth since no significant negative AAR can be observed in the following days in which the market would adjust itself. In fact, more positive AAR can be seen on day +4 and +6.
1
37 Table 9: AARs and Patell t-statistics for SES announcements in developed and undeveloped countries
Event Day AAR SES Developed (%) Patell t-statistic AAR SES Undeveloped (%) Patell t-statistic -10 -0.116 -0.960 0.108 0.547 -9 -0.104 -0.855 0.334 1.686 -8 0.090 0.742 -0.083 -0.419 -7 0.346 *** 2.847 0.301 1.520 -6 -0.095 -0.783 0.307 1.552 -5 0.012 0.101 -0.079 -0.400 -4 0.217 * 1.792 -0.220 -1.115 -3 0.127 1.047 -0.281 -1.420 -2 -0.050 -0.415 -0.073 -0.372 -1 0.385 *** 3.171 -0.099 -0.503 0 0.069 0.571 -0.036 -0.186 +1 -0.015 -0.126 0.233 1.179 +2 0.075 0.622 -0.432 ** -2.180 +3 0.184 1.516 0.124 0.626 +4 0.287 ** 2.365 -0.086 -0.436 +5 -0.183 -1.509 0.178 0.899 +6 0.354 *** 2.910 0.168 0.852 +7 0.085 0.702 -0.025 -0.128 +8 -0.044 -0.369 0.093 0.473 +9 0.084 0.694 -0.070 -0.356 +10 -0.284 ** -2.342 -0.267 -1.352 * p < .10, ** p < .05, *** p < .01
Figure 5: Average Abnormal Returns for SES announcements Developed and Undeveloped
Note: Significance lines are approximations
Table 9 and Figure 5 further indicate that firms announcing SES activities in undeveloped countries experience negative abnormal returns on the second day after the SES announcement (p<.05). This negative AAR on day +2 represents the depreciation of the
38 stock by the market after the announcement. All other days within the event window do not yield any significant returns.
Hence, announcements of SES activities in developed countries gain overall positive returns and announcements of SES activities in undeveloped countries gain overall negative returns. Thus, hypothesis 2a is supported.
It is hypothesized that CSR has a stronger positive effect on CFP in undeveloped countries and CS has a stronger positive effect on CFP in developed countries. If this is true it would offset the relation between SES and CFP. To observe the individual effects of CSR and CS on CFP the cases are sub-divided. Each announcement is either a CSR or a CS announcement taking place in either a Developed or Undeveloped country.
Figure 6 illustrates the effect social announcements have on CFP in both developed and undeveloped countries. The AAR of CSR announcements in developed countries provided significant positive returns on day -7 and -4 (p<.01). The cumulative average abnormal return in this timeframe is 1.11%, this appears to be the time when the information entered the market. However, as earlier noted we follow McWilliams and Siegel (1997) and take the closest significant abnormal return from the event date. That is day +3, which yields weaker positive AAR, however still significant (p<.10).
39 Figure 6: Average Abnormal Returns for Developed/CSR and Undeveloped/CSR
Note: Significance lines are approximations
Figure 7 shows the effect CS announcements have on CFP in both developed and undeveloped countries. No significance lines are given in this figure due to severe differences in the degrees of freedom of the two sub-samples. The AAR of CS announcements in developed countries are significantly positive on day -1 (p<.01) and then again on day +4 (p<.10), however this last positive return is quickly counteracted by an equal negative return the following day.
Figure 7 further shows that the AAR of firms announcing CS activities in undeveloped countries experience significant positive returns on day -9 and day -7 (both p<.10). Also, an increase in AAR can be observed on day +1, however these AARs do not display any statistically significant returns till day +10, which is negative (p<.05). Thus, hypothesis 2c can be accepted.
Figure 7: Average Abnormal Returns for CS/Developed and CS/Undeveloped
Note: No significance lines due to severe differences in degrees of freedom
t t
40
5.3 Moderating Effect of the Type of SES Activity
Next to the development level of the location, the type of activity is also hypothesized to affect CFP. By again dividing the total sample into 2 sub-groups it is possible to observe this effect. This time the groups are not divided in Developed and Undeveloped, but in being CSR or CS announcements. Of the total sample of 95 cases, 48 cases were of a social nature and 47 cases of a environmental nature. Table 10 and Figure 8 display the Patell t-statistic and AARs and illustrate that CS announcements have a stronger positive correlation to CFP than CSR announcements. This holds true when considering the most significant abnormal return closest to the event date (McWilliams and Siegel, 1997). For CS announcements this is day -1 (p<.01) and for CSR announcements this is day +3 (p<.10). Accordingly, hypothesis 3 is rejected.
Table 10: Average Abnormal Returns and Patell t-statistics for CSR and CS announcements
41 Figure 8: Average Abnormal Returns for CSR and CS announcements
Note: Significance lines are approximations
5.4 Regression Analysis
A linear regression analysis is performed to further explain the effect the individual variables have on firm value. These regressions also include the control variables
Company Size and Industry, which are unable to include in the Patell t-statistics. The
results of the regression analyses are shown in Table 11, 12, 13 and Table 14 and follow the procedure as described by Geyskens et al. (2002). In the regression analysis the most significant AR value for the entire sample (day -1) is used as the dependent variable.
Table 11 illustrates Equation model 4, the main effect of SES announcements on firm performance. As can be observed none of the variables included in this regression equation yield a significant effect.
42 Table 11: Regression Analysis; Overall results (Equation 4)
Hypothesized Sign b t-Value
Intercept .129 .576 Control variables Company Size -.000 -.218 Basic Materials .450 1.258 Healthcare .426 .851 Industrial Goods .242 .568 Services -.107 -.325 Technology .223 .596 Utilities .172 .320 R² = . 038 R² (adjusted) = - .039 * p < .10, ** p < .05, *** p < .01, AR is day -1
Table 12, yields the outcomes of several regressions. The overall moderating effect of location (Equation 5), the overall moderating effect of type of activity (Equation 6) and a control regression for location. The overall moderating effect of location is, on average, not significant (b=.200, p>.10). The same insignificant result can be observed for the moderating effect of CSR (b= -.126, p>.10).
Table 12 also depicts the results for the control regression for location, but with the interval HDI value as a measure for development. This is done instead of categorizing the announcements in Developed or Undeveloped by dummy variable. Yet, this does not change the effect of the variable, which are all still insignificant.
43 Table 12: Regression Analysis; Moderating effects (Equation 5 and 6) and Control with HDI.
Hypothesized Sign b t-Value
Intercept -.015 -.051 Moderating effect Developed location + .200 .776 CSR + -.126 .687 HDI + .472 .648 Control variables Company Size -.000 -.266 Basic Materials .435 1.211 Healthcare .403 .804 Industrial Goods .232 .544 Services -.091 -.277 Technology .246 .653 Utilities .157 .291 R² = . 045 R² (adjusted) = - .044 * p < .10, ** p < .05, *** p < .01, AR is day -1
Table 13 illustrates the interaction effect of the 2 variables. The sub-group CSR in Developed countries is taken as the constant. Again nothing is significant, although the direction of the effect does relate to the earlier mentioned findings.
Table 13: Regression Analysis; Interaction effects (Equation 7 and 8)
44
6. Discussion
This chapter critically discusses the results from the previous chapter and relates these findings to previous research. Before doing so a summary of the findings will be provided. Furthermore, no research is without its limitations, after the discussion of the findings some limitations of this research are given. These limitations should be considered when making use of these results, from these research limitations arise the suggestions for further research.
6.1 Summary of Findings
The general results from this research indicate that corporate announcements of social and environmental sustainable activities is appreciated by investors. When regarding the total event window, of minus 10 and plus 10 days of the event, this is particularly true for day -7 and day +6. These two days portray strong positive abnormal returns (p<.01). However, as is discussed in Chapter 4.7, the statistically significant AAR closest to the event date will be assumed to be the arrival of new information to the market (McWilliams and Siegel 1997). For the overall effect of SES announcements on firm value this is day -1 (p<.05). On this day 60% of the 95 SES announcements provided positive returns, resulting in an average abnormal return of .25%. When considering the cumulative average abnormal return of day -1 till day +1 this is even higher, a 0.34% increase in 3 days. This significant positive reaction within such a short timeframe indicates that the market reflects and rewards the new information of firms announcing SES activities in the stock prices, thus forming the foundation to the support of my hypothesis 1.
45 The cases in which SES activities were announced in developed countries experienced significant positive returns over multiple days. The day yielding a strong significant return closest to the event day is -1 (p<.01). The cases in which SES activities were announced in undeveloped countries denote a negative abnormal return on day +2 (p<.05), all other days yield insignificant returns. This indicates that firms announcing SES activities in developed markets are appreciated by investors and firms announcing SES activities in undeveloped markets are depreciated. These findings support my hypothesis 2a.
Hypothesis 2b states that public CSR announcements taking place in undeveloped countries would be appreciated more positive than public CSR announcements taking place in developed countries. Hypothesis 2c states that announcements of CS activities in developed countries would be appreciated more than announcements of CS activities in undeveloped countries. When both sub-hypotheses would hold true they would off-set the main effect of the development level of the SES location on firm value. Thus, the two sub-groups Developed SES announcements and Undeveloped SES announcements are further sub-divided in social and environmental activities.
The exact opposite of H2b appears to occur on the market. Firms announcing CSR activities in developed countries are appreciated on day +3 (p<.10) and firms announcing CSR activities in undeveloped countries are depreciated on day +2 (p<.05). Thus, H2b can be rejected.
Firms announcing CS activities in developed countries experience strong significant returns on day -1 (p<.01). Firms which announce CS activities in undeveloped countries experience minor positive returns on day -9 and day -7 (both p<.10). These findings indicate that announcements of CS activities have a stronger positive effect when located in developed countries. Thus, H2c can be confirmed.
Since, H2b is rejected and H2c is supported no counter balancing events occur on the main effect of the level of development.
46 being a CSR or a CS announcement. This led to 48 CSR cases and 47 CS cases, analysis showed that CS announcements are stronger related to firm value increase than CSR announcements. For CS announcements this is day -1 (p<.01) and for CSR announcements this is day +3 (p<.10). The fact that CS is stronger positive correlated to CFP than CSR provides evidence to dismiss hypothesis 3 and assume the opposite.
To control for possible off-setting effects of the location of CSR and CS announcements, H3b and H3c are formulated which divides the sub-groups in location. However, the findings for these sub-hypotheses do not illustrate any deviation from the main hypothesis. For CS announcements in developed countries the strongest positive effect is on day -1 (p<.01) and for CSR announcements in developed countries this is day +3 (p<.10). Thus, in developed countries CS announcements are stronger correlated to firm value increase than CSR announcements. For CS announcements in undeveloped countries the strongest positive effect is on day -7 (p<.10). CSR announcements in undeveloped countries yield negative abnormal returns on day +2 (p<.05). Thus in undeveloped countries CS announcements are also stronger correlated to firm value increase than CSR announcements.
The regression analysis following the Patell t-test do not display any significant relations between the most significant AR for the entire sample, day -1, and the moderating variables. A possible reason for this is due to the occurrence of these effect on different days within the event window. The regression analysis is unable to display these different event days as well as the Patell t-test.
