The effects of being on the global 100 index on the stock returns of companies

FACULTY OF ECONOMICS AND BUSINESS,AMSTERDAM SCHOOL OF ECONOMICS

BACHELOR’S THESIS –ECONOMETRICS

T

T

G

100

Bekir Bekdur (10800247) Supervisor: Derya Güler

Statement of originality

This document is written by Student Bekir Bekdur who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

TABLE OF CONTENTS

1. Introduction 1

2. Theoretical Framework 3

2.1 Criteria for being socially responsible 3 2.2 The growth of socially responsible investments 5

2.3 The Global 100 list 5

2.4 Incentives for investors to invest in SRI 7 2.5 Incentives for companies to be socially responsible 8

2.6 literature reviews 9

2.7 hypotheses 11

3. Data 11

4. Methodology 13

4.1 The Capital Asset Prising Model 15

4.2 The FAMA French 3 factor Model 15

4.3 The FAMA French 5 factor Model 15

4.4 Hypotheses 15

5. Results 18

5.1 Testing the significance of the ACAR 18

5.2 The regression of the CAR on rank 19

5.3 The regression of the CAR on the overall score 19

5.4 The top 10 companies versus the rest 20

5.5 The results in comparison to the existing literature 21

1

1. Introduction

Global warming has become one of the most hotly debated issues this century and environmental consciousness among people has increased enormously in recent years. Environmental issues have started to become a regular part of many aspects of society, including the behaviour of investors in the stock markets. Millions of people invest in publicly listed companies’ stocks, but how are those companies and their performance related to our environment? Are these companies socially responsible, and does that matter? Socially responsible investments are investments made by investors who take into account not just financial returns, but other factors such as the environment, consumer protection, human rights and diversity. According to studies published by Lombardo and D’Orio (2012) investors do take into account social responsibility of companies as a factor for deciding on their investments.

Studies published by Statman and Glushkov and by Louche and Eurosif (2008) concluded that investors target both a maximized financial return as well as increased societal wellbeing in their investments. According to studies of López (2007), Louche (2004) and Cortez, Silva and Areal (2009), investing in socially responsible investments has been increasingly attractive for investors. Indicative of this increase is the fact that circa 12 per cent. of the total world assets are held in so-called socially responsible companies. But how does one measure the socially responsible performance of a company?

The Global 100 Index (or the G100) is an annual list of companies in which the top 100 companies globally are ranked in the way they address balancing an environmental, social and economic performance relative to their returns to investors. The index was introduced by Corporate Knights in 2005 and to create these lists, they use hundreds of environmental indicators and models and select the top 100 companies among a list of circa 6000 companies from 22 countries. The G100 is updated annually and its stocks are traded on nine different exchanges globally. As mentioned, previous studies have shown that social responsibility is a factor for investors in making their investments.

2

This thesis relies on older studies of Murguia and Lence (2015), in which the influence of the ranking on the G100 on the general value of the companies is analysed. In their research they also analysed the effect of being ranked one position higher on the list. Furthermore, they compared the top 50 of the G100 list to the bottom 50 companies. After computing if the components of the Green Score might be explaining the results better than the Green Score itself, they compared US traded stocks to non-US traded stocks.

In this thesis the 2018 Global 100 list which is published on 22tt January 2018 will be analysed by firstly calculating if there are any effects as a result of being on the Global index on the stock returns. Secondly, the effect of being ranked one position higher on the index will be analysed. Thirdly, having a one per cent. higher overall score will be analysed and this effect will be compared to the effect off being ranked one position higher. Finally, different sub-parts of the Global 100 index will be analysed. The top 10 companies will be compared as first to the rest of the index. This effect might not be significant enough, because the top 10 is a small sample. Therefore, it will be analysed from which point of the index the companies’ stock returns are statistically different to each other. The analyses in this thesis are calculated by computing the cumulative abnormal returns of the event of interest, which is being ranked on the Global 100 list of 2018. This will be calculated with different models. The Capital Asset Pricing Model, the FAMA French Three Factor model and the FAMA French Five Factor Model will be used for calculating the cumulative abnormal returns of this event.

This thesis is divided in five more parts. Firstly, the theoretical framework is discussed. Secondly, the way how data is collected, information about the variables and the descriptive statistics are discussed in the data section. Thirdly, the used econometric methods and the corresponding hypotheses are discussed in the methodology chapter. Fourth, the results of this research are described, followed by an analysis of these results. And finally, the conclusion of this thesis is presented together with the author’s remarks on the findings in this thesis.

3

2. Theoretical framework

Socially responsible investments are those investments in which the social, environmental and social aspects of the investments’ consequences are taken into account for making investment decisions about companies. The criteria for being a socially responsible company and the criteria of being on a special list which is called The Global 100 list is discussed further in this thesis. Based on the ranking on this list and the score of the company, the central question is answered in the following sections. The research in this thesis is based on event studies. This means that an event of interest is identified and the abnormal stock returns which are related to this event are analysed. In this thesis, the event is being on The Global 100 list. The effect of being on this list on a company’s stock returns will be analysed and compared to stock returns of the same company prior to being on the list. The expected returns of a company are measured first of all with the Capital Asset Pricing Model, which is: E(Ri,t) = Rf,t + ßi (Rm,t – Rf,t). In this equation, Rf,t is the risk-free return, Rm,t the market return and i is the coefficient for a company based on the risk of the stock. Then it is computed with the FAMA French Three Factor model, which is: E(Ri,t) = αi + Rf,t + ßi (Rm,t – Rf,t) + siSMBt + hiHMLt + et. Finally, it is computed with the FAMA French Five Factor model, which is: Ri,t–Rf,t= αi + ßi(Rm,t – Rf,t) + siSMBt + hiHMLt + riRMWt + ciCMAt + et. In these equations, SMBt stands for the Small Minus Big ratio at time t, HMLt stands for the High Minus Low ratio at time t, RMWt stands for the profitability factor at time t and CMAt is the calculation at time t of the difference in returns of conservatively investments and aggressively investments. Finally, Ri,t is the return of company which is calculated by hand with the daily stock prices.

2.1 Criteria for being socially responsible

In order to measure a company’s socially responsibility, different screeners such as ecological, social, corporate governance and ethical criteria are used. Several screens are used to mark a company as socially responsible, which are set out in table 1. Every screen is followed by a sign, based on their effect on social responsibility.

4

Table 1: The screens which are used to measure the socially responsibility

screen sign screen sign screen sign screen sign

tobacco - pornography/ adult - employment diversity + shareholder activism + alcohol - abortion/ birth control

- human rights + non-married -

gambling - labour relations and workplace conditions

- animal testing - healthcare/ pharmaceuticals - defence/ weapons - Environment + renewable energy + interest-based financial institutions - nuclear power - corporate governance

+ biotechnology + pork producers -

irresponsible foreign - business practice + community involvement +

In order to measure the social responsibility of a company, the first criteria used is negative screening in order to eliminate a-social, non-environmental or un-ethical companies, such as companies which have alcohol, tobacco or gambling in their list of activities. Secondly, positive screens such as companies’ social environments, labour relations, environmental relations, sustainability and their stimulation of cultural diversity are used in order to rank the companies on their social responsibility. The third screen is based on the so called ‘triple bottom line’, which is concerned with the efficiency of the balance of the company between three pillars, namely: people, planet and profit. The fourth screen which is used is based on the combination of sustainable investing with shareholder activism. These four screens determine the level of social responsibility of a company. Meeting these criteria is a requirement in order to be classified as a socially responsible company and to enter The Global 100 list. Consequently, these criteria are relevant to study the central question of this thesis,

5

which is about the effect of being present on The Global 100 list on a company’s stock returns.

2.2 The growth of socially responsible investments

Socially responsible investments have grown rapidly over the world in the past few years, especially in the United Stated and in Europe. The total value of socially

responsible investments is estimated around $2.3 trillion, of which circa $1.4 trillion is the European value. In both Europe and the United States, these numbers mean socially responsible investments represent circa 10% of the total market. The growth in socially responsible mutual funds increased in a similar manner, from 55 such funds in 1995 in the United States to 201 and from 54 in 1995 in Europe to 375 in 2005 (SIF, 2005; EUROSIF, 2006; SiRi,2005). This growth is expected to continue since peoples’ investment decisions are not only based on rewards they receive on the risks that they take. Their investments are also based on a set of personal and societal values (Bollen, 2007). According to the studies of Bollen, Renneboog, ten Horst and Zhang (2005), socially responsible investments will grow further even if the returns of these socially responsible investments would be lower than alternative investments. According to the same research, cashflows of socially responsible investments are becoming less volatile in comparison to their past performance.

2.3 The Global 100 list

Besides the importance of socially responsible investing, the way to announce and inform people about the existence and the degree in which a company is socially responsible is very important. In this way it is possible for investors to determine the social responsibility of a company, which is crucial for the investments decisions (Bollen, 2007). The 100 most sustainable corporations in the world are collected on a list, which is called ‘The Global 100’ list. This list is updated every year in January and so each year companies are added and deleted from the list or ascend or descend in the list. After the release at the World Economic Forum in Davos every year, this list is immediately published in leading media like the publishers’ magazine, ‘The Corporate

6

Knights Magazine’ which has a huge reach of 380,000 business and political decision-makers. The rankings on the list are based on publicly-disclosed data of companies from all industries. Every year, the companies which are on the list are contacted to verify the data that is used. The companies are informed prior to have the possibility to collect necessary data for the publicity or to pass small changes. The Global 100 list is based on the analyses of data and so this list is not ranked on judgement, which makes this list objective. Key performance indicators are used for the analyses, which are partially based on the speciality of the companies to rank the companies in the list. In this way the list is representative in the business sustainability and socio-economic context. In total there are 14 key performance indicators. From these 14 indicators, 4 are universal indicators while 10 are priority indicators. The four universal indicators are: leadership diversity, clean capitalism pay link, pension fund status and percentage tax paid. In this way the companies from different industries are compared to each other and so there is a possibility for diversity in the list. Expert stakeholders’ review and comments are also taken into account which makes the list more qualitative. The list is composed in three different parts.

First of all, the mid, large and mega-cap public companies are analysed from which some companies are eliminated through the screening process. In this screening process, the sustainability disclosure practices are checked as first. The companies below the 75% of the priority key performance indicators are eliminated.

Secondly, there is an extended financial health check. This financial check is done by a so-called F score. All companies with a F score below the five are eliminated from the list. The F score is determined by the sum of 9 subtests which all result in either a zero if the result is negative or in a one if the result is positive. The subtests that are used are as follows: the net profit is positive, the operating cash flow is positive, the net profit is the same as the previous year and the total asset is the same for previous year, the operation cashflow is larger than the net profit, the long-term debt have not increased and the average assets have not increased, the current ratio has increased, there is no

7

raising of ordinary equity over the previous year, the gross margin has improved over the previous year and finally the asset turnover has increased.

Thirdly, the product categories are inspected from which products with a GICS code that relate to tobacco products or armaments are eliminated. Fourth, the applicability of sanctions is analysed in such a way that companies that belong to the bottom quartile performers in the Corporate Knights sanctions screen are eliminated. Based on these four screens used it is possible to keep the most sustainable companies in the remaining list.

At the end The Global 100 list is created and ranked based on the overall percentage score. The first 100 scores are in the list and in this way companies in the list are comparable with the companies which are not included in the list. On this way the central question can be analysed. The sub-questions can be answered on the score base.

2.4 Incentives for investors to invest in socially responsible investments

The research of Bollen (2007) is already explained, in which he explains that investment decisions are also dependent on a set of personal and societal values besides their return on the risks they take. The criteria to become a socially responsible investment and to enter The Global 100 list are also explained. But, why should people find social responsibility of their investments important? There are three different explanations. The first explanation is that socially responsible investments can in the financial way perform better due to their socially responsible consumers.

The second explanation is that it is caused by the deontological investors, the investors whose reject non-ethical profits. In their decisions, they take the past environmental performance as well as the current management into account, to avoid scandals in the future. The last explanation is that consequentialist investors’ desire to reward good behaviour together with their desire to decrease environmental irresponsible firms’ market share (Murguia & Lence (2014)).

8

2.5 Incentives for companies to be socially responsible

Besides the question why investors would invest in socially responsible investments, it is interesting to analyse how interesting it is for companies to be socially responsible. Is it interesting for the companies to act socially responsible in order to maximize the utilization of the company? This question is relevant since being socially responsible brings more costs and together with more costs it is more difficult to maximize profits, according to established economic theory. Since investors of socially responsible investments do invest in these companies consciously and with the intention to invest in companies which are environmental friendly with a focus on social welfare, it is

interesting to analyse in which way the companies do respond to such a demand. According to Friedman’s (1970) arguments, companies should only take into account their own profits, in order to take care of their own shareholders. Following this logic, companies with a socially responsible policy with socially responsible investors do take into account social responsibility because this is in the interest of their own

shareholders. In this way it is possible to prevent the potential conflicts between the corporations and the stakeholders and the corporations and stakeholders do work

together more efficient (Heal, 2005). According to the same theory of Friedman (1970), if the profits due to the compromises with the stakeholders would decrease the firms’ value, these companies should not affect the policy in which these companies act socially responsible. According to Besley and Ghastak (2006) the socially responsible investments are always in line with the profit maximization strategies of the companies.

According to Barron (2001), ‘private politics’, where there is pressure of the lobbyists on the companies, is another reason why these companies should use such an

environmental friendly policy. This means that the companies do not only invest socially responsible in order to maximize their own utilisation or the utilisation of the shareholders, but also because of the political pressure they have on the policy of the company.

9

2.6 Existing literature

After the question whether investors should invest in socially responsible companies and after the question whether companies should be socially responsible, it is interesting to analyse how the returns of those socially responsible investments are in comparison to those investments which are not socially responsible.

Analyses on different portfolios made by Derwall, Bauer, Guenster and Koedijk (2004) based on equity portfolios which are composed by their sustainability show that a company which is classified as a sustainable company outperforms other companies significantly, by circa six per cent. per annum compared to a company which is not classified as a sustainable company. In this research they use the Capital Asset Pricing Model on the data they have from July 1995 until December 2003. They ranked the companies in their dataset by themselves based on the eco-efficiency. In these days The Global 100 list was not introduced yet.

Analyses of Gutowski, Murphy, Allen, Bauer, Bras, Piwonka and Wolff (2005) based on the impact of environmental news and the rankings on the stock returns result in a positive and significant relation between economic performance and environmental performance. They also found that the results of environmental news and rankings on the returns are different for Japan and for the United states. The researchers used over 50 different websites for their dataset. They collected data of Japan, northern Europe and the United states from July 1999 until April 2001.

Recent research of Murguia and Lence (2015) show that an increase of 10 positions in the rank of The Global 100 list improved the stock value by 113$ million dollars. Additionally, in their results they find that the average market capitalization is higher for the companies on The Global 100 list. Also, they found a stronger reaction for non-US-traded stocks in comparison to non-US-traded stocks. In their comparison of heavy sectors with heavy sectors they find that there is a stronger reaction in the non-heavy sector compared to the non-heavy sector. Furthermore, in their findings they find that US traded stocks are affected by past performance and managerial quality and the

US-10

traded stocks are interested in past environmental performance and managerial quality. This is not the same for non-US-traded stocks, where only managerial quality is relevant.

Comparable studies of Ameer and Othman (2012) resulted in that superior sustainable practices have higher performances in comparison to comparable companies which are not socially responsible. Their sample was The Global 100 List from 2008 and from this research they found significant different results for the companies that are selected in the list. The companies in the list tend to have, in addition to higher mean sales growth, also a higher return on their assets and a higher profit before taxation.

However, there are also studies in which there is not found any significant relation. An example of a study in which there was no significant relation is the research of

Goldreyer and Diltz (1999). This research analyses socially responsible mutual funds from the United States on their portfolio performance compared to conventional funds. Monthly returns data of 49 socially responsible mutual funds are used in their research with data from January 1981 until June 1997.

Likewise, the research of Anderson-Weir (2010), which is based on the Newsweek’s Green Ranking of 2009, did not result in a significant relation between The

Newsweek’s Green Ranking and the returns of the S&P 500 stocks. The Green Rankings is a list which is consisting of the 500 largest publicly traded corporations in the United States. Daily data of the companies from October 2008 until September 2009 was used for the research. The resulting coefficient in this research was also very small, 0.00005989, which means that one unit increase in the Green score should lead to a decrease of 0.006% in a companies’ cumulative abnormal returns.

11

2.7 Hypotheses

This thesis relies on older studies of Murguia and Lence in 2015 and expands the study by analysing the effect of having a 1% higher score. The studies of Murguia and Lence is one of the less studies about explicitly The Global 100 list, because the release of the Global 100 index just started in 2005. Most of the studies about this subject find a positive and significant relation between social responsibility and the stock returns. The study of Murguia and Lence is also expanded by comparing the top 10 of the list to the rest of the list. Furthermore, the effect of having a 1% higher score is compared to the effect of being ranked one position higher. The central question which is about analysing the effect of being present on the list is done with a different dataset, the dataset consisting of The Global 100 list 2018. The hypothesis of this thesis is that there is a significant relation between being on The Global 100 list and the companies’ stock returns. Also, a positive and significant effect of having a 1% higher score is expected. Finally, a positive and significant effect of being ranked one position higher is expected.

3 Data

The data for the event studies are collected by the Wharton Research Data Services (WRDS) via the library of the University of Amsterdam and, the DataStream application in excel in the library of the University of Amsterdam, the website of Corporate Knights and from Yahoo Finance. Firstly, the daily closing prices of indexes are collected from the WRDS website. Not all of these companies are in the WRDS website and these companies are therefore collected from the website of Yahoo Finance. From these daily closing prices, the daily returns are calculated by subtracting the price at time t with time t-1 and thereafter divided by the price at time t-1. The symbol for these returns is Ri, t. In later analyses this is done in the same way in the DataStream system. In formula form, this is: 𝑅𝑖,𝑡 =

𝑃𝑖,𝑡−𝑃𝑖,𝑡−1 𝑃_{𝑖,𝑡−1} .

The data which is necessary to calculate the expected returns in the different models that will be used, which are the Capital Asset Pricing Model, the FAMA French 3 factor model

12

and the FAMA French 5 factor model, are downloaded from the library of MBA Tuck Dartmouth. This is a list consisting of the variables Rf,t, Rm,t, SMBt, HMLt, RMWt and CMAt. The explanation of these variables are as follows. Rf,t is the risk-free rate at time t, Rm,t is the market return rate at time t, SMBt is the Small Minus Big ratio at time t, HMLt is the High Minus Low ratio at time t, RMWt is the profitability factor at time t and CMAt is the calculation at time t of the difference in returns of conservatively investments and aggressively investments.

From the website of Corporate Knights, a list with information about the listed companies is downloaded. These variables will be used in the regressions in order to try to explain the cumulative abnormal returns in a complete way. The variables which are downloaded are in table 2.

Table 2: The variables downloaded from Corporate Knights overall score

rank 2017&2018 energy productivity carbon productivity score employee turnover rate women executives supplier score

water productivity score waste productivity score R&D revenue 2014-2016 cash taxes paid ratio women on board sustainability pay link clean air productivity rate

CEO average worker pay ratio

pension fund status injury rate

fatalities number

For the calculations with the Capital Asset Prising Model, the variables which are needed are: the returns, a constant and the variable Rm,t-Rf,t. For the FAMA French Three Factor model the variables SMBt and HMLt are added to the variables in the CAPM. Finally, for the FAMA French Five Factor model RMWt and CMAt are added as well. Unfortunately, not all historical stock price data could be found. For 3 companies in the entire list, historical closing day stock prices couldn’t be found. These companies are Aberdeen Asset Management (number 58), Orsted (number 70) and Syngenta (number 91). Except these 3 companies, the returns for all other companies are calculated. The average returns of the entire list, the average returns of the top 10 companies and the average returns of the companies 11-100 in the list are listed in table 3 below as descriptive statistics. To

13

these results, the standard deviation is added for each category and all of these results are listed in table 3, the descriptive statistics.

Table 3: Descriptive statistics of the returns

entire list the top 10 companies companies 11-100

average returns 0,000502 0,001169 0,000426

Standard deviation of the average returns

0,001032 0,001005 0,001012

the maximum average return

0,003614 0,002538 0,003614

the minimum average return

-0,00217 -0,00066 -0,00217

4 Methodology

The research for this thesis is done by using event studies, in which the event of interest is analysed. In this thesis this, the event of interest is the release of The Global 100 list, which is released on 22th of January 2018. The effect of this release is calculated by analysing the abnormal returns around the announcement day. In order to do this, a test period needs to be chosen. This means that two dates, one before and one after the announcement day, need to be chosen in order to analyse the effect in the period between these two dates. To simplify, these dates are called t=T1 and t=T2, whereas the announcement day is called t=0. In this thesis T1 is chosen as 3 working days before the announcement day, as the 17th of January and T2 is chosen as 3 days after the announcement, which is the 25th of January. To analyse the abnormal returns in this period, an estimation period is selected first. This means that one more date is chosen before t=T1, which is called t=T0. In this thesis, 18th of May 2017 is used as T0. The period of time between T0 and T1 is called the estimation period. In this thesis, three

14

different models are used for this regression which are based on the estimation period of a company. This will be discussed later in this chapter. After these regressions, it is possible to calculate the expected return for every date. Then, the abnormal returns for the test period is calculated in the following way. The real values in the test period for every date are subtracted by the expected value for every company. In this thesis, these abnormal returns in the test period are summed up to a cumulative abnormal return for the entire test period. In figure 1 below the estimation period and the test period are explained with the exact dates used. The estimation period is used for the econometric regressions with which the expected returns of the stocks are calculated.

Figure 1 The estimation period and the test period which is used in this thesis

As mentioned, the expected returns are calculated with three different models. These are the Capital Asset Prising Model, The FAMA French Three Factor Model and the FAMA French Five factor model. The value of the cumulative abnormal returns (by summing up the daily abnormal returning the test period) resulting from these different models and the significance of them are calculated for these models. The effect of being listed one position higher, the effect of having a 1 per cent. higher overall score and the effect of being on the top 10 for these three models will also be compared to each other in the next sections. As a robustness check, a different period is used in order to check whether there are real different between the results. There, T0 is 10/5/2017, T1 is 29/12/2017 and T2 is 29/3/2018.

15

4.1 The Capital Asset Prising Model

The Capital Asset Prising Model (1963) is a very simple formula which calculates the expected returns by using the risk-free rate and the market return rate. For every

company of the Global 100 index the coefficient βi is calculated with the regressions of every company in the estimation period. This is shown in formula one below.

E(Ri,t) = Rf,t + ßi (Rm,t – Rf,t) (1)

4.2 The FAMA French Three Factor model

The FAMA French Three (1993) factor model is known as an expansion on the Capital Asset Prising Model, which does not only use the risk-free rate and the market return rate, but also uses a constant, the Small minus Big Ratio and the High Minus Low ratio. For calculating the expected returns, the coefficients ßi, si, hi and the constant are calculated for a company from a regression in the estimation period. The formula representation is given below.

E(Ri,t) = αi + Rf,t + ßi (Rm,t – Rf,t) + siSMBt + hiHMLt + et (2)

4.3 The FAMA French Five Factor model

The FAMA French Five Factor model (2014) is known as an expansion on the FAMA French Three Factor model. In addition, this model uses the coefficients for RMWt and CMAt as well in the formula. So for calculating the expected returns during the test period, the coefficients of ßi, si, hi, ri, ci and the constant are used. The model in formula form is as follows.

Ri,t–Rf,t= αi + ßi(Rm,t – Rf,t) + siSMBt + hiHMLt + riRMWt + ciCMAt + et (3)

4.4 Hypotheses

In this subsection, the tests and the hypotheses used in this thesis are discussed and explained with the necessary explanations. In the following parts all null hypotheses and alternative hypotheses are explained in detail for every econometric method which is used in this thesis.

16

4.4.1 Testing the significance of the average cumulative abnormal returns The main hypotheses in these thesis, which is calculated by the T-test, is as follows:

- H0: ACARi,t = 0 and H1 :ACARi,t  0

As mentioned before, the cumulative abnormal returns for every company are calculated for all three different models. For the test, the average of the cumulative of the abnormal returns needs to be calculated. Then, the standard deviation between these cumulative abnormal returns is calculated. This is divided by the square root of the number of companies, which is 97 in this thesis. By dividing the average cumulative abnormal return by the standard deviation between the cumulative abnormal returns, which is again divided by the square root of 97, a t-value is calculated. This t-value can be compared to the critical value of the t-distribution with α=5% and α=10% to conclude if the average cumulative abnormal return is significantly different from zero. In this thesis, this will be done for the Capital Asset Prising Model, the FAMA French 3 Factor model and the FAMA French 5 Factor model. The outcomes will be compared to each other.

4.4.2 Testing the influence of having a 1% higher overall score and testing the influence of being listed one position higher on the cumulative abnormal returns The cumulative abnormal returns (CARs) which are calculated as explained in this chapter for the three different models are collected in this thesis. These CARs are used for econometric regressions in order to analyse the effect of being listed one position higher and for having a one per cent. higher overall score. The CARs are used as the dependent variable and the overall score of the companies and the ranking of the companies are used as the independent variables. In this way, econometric regressions will be done in order to draw conclusions about the CARs of companies which are ranked one position higher and companies with a 1% higher overall score. The model is tried to be expanded by adding more explanatory variables from table 2, to make the model as complete as possible. The simple models, without adding extra variables, are better and therefore discussed in the next chapter. The T-test is used in order to determine whether an explanatory variable is significant or not, by dividing the coefficient by its standard

17

deviation. This is done for the CARs which resulted from the Capital Asset Prising model, the FAMA French 3 Factor model and the FAMA French 5 factor model. The results are in the next chapter. For all models the null hypotheses and the alternative hypotheses are the same, namely:

- H0: being listed one position higher does not result in higher CARs. This is the same as β=0 in equation 4.

H1: being listed one position higher does result in higher CARs This is the same as β<0 in equation 4.

- H0: having a one per cent. higher overall score does not result in higher CARs This is the same as β=0 in equation 5.

H1: having a one per cent. higher overall score does result in higher CARs This is the same as β>0 in equation 5.

For the test statistics, α=5% and α=10% are both used, and these results are compared to each other. The regression models are as follows.

CARi,t = α + β*RANKi + εi (4)

CARi,t = α + β*OVERALLSCOREi + εi (5)

From the regressions outputs, the significance of the coefficients β will be checked in order to draw conclusions.

4.4.3 The top 10 companies versus the rest

In this thesis the top 10 companies will be compared to the remaining companies in the list. In order to do this, the cumulative abnormal returns of these two groups will be compared to each other. This comparison will be made by creating a dummy variable for the top 10 companies. The T-test will be used in order to conclude if these groups are different. If the coefficient is significant, the groups are different. The null hypotheses and the alternative hypotheses are as follows:

- H0: the top 10 companies have the same CARS as the rest of the list. This is the same as β=0 in equation 6.

H1: the top 10 companies have a higher CAR than the rest of the list. This is the same as β>0 in equation 6.

18

This regression is done with the cumulative abnormal returns of the Capital Asset Prising Model, the FAMA French 3 factor model and the FAMA French 5 factor model. The results of the three different models are compared to each other. If the results are not statistically significant, a robustness check will be done in order to check which top leading countries are statistically different from the rest of the list. Again α=5% and α=10% are both used and the results of the different models with different α’s are compared to each other. The regression model which us used is below in equation 6.

CARi,t = α + β*DUMMYTOPTENi+ εi (6)

5 Results

In this section, the results for every regression is discussed in subsections. It is also discussed whether the corresponding null hypothesis is rejected or not. For all these results α=5% is used as well as α=10% to compare the significance levels to each other. The results of the robustness check are not discussed further since the first analyses were very similar to the previous results.

5.1 Testing the significance of the average cumulative abnormal returns

As described in the previous chapter, the main hypothesis in this thesis is to test whether the cumulative abnormal returns are significantly different from zero. The cumulative abnormal returns are the summations of the abnormal returns in the test period. The null hypothesis is, like discussed in the previous chapter, that these returns are equal to zero. According to the empirical research in this thesis, all of the models result in a negative and significant relation (for α=5% as well as for α=10%) between being listed on The Global 100 list and the stock returns during the test period. These results are summed up in the table below together with the standard deviations and t-values. The Average Cumulative Abnormal Returns are abbreviated as ACAR. The results, which is for the average of all companies in The Global 100 list, are below in table 4.

19

Table 4: descriptive statistics of the average cumulative abnormal returns

ACAR standard deviation t - value significance α=5% significance α=10%

CAPM -0,0095533 0,00351432 -2,7184048 Yes Yes

FAMA 3 -0,0073051 0,00327436 -2,2309957 Yes Yes

FAMA5 -0,008847 0,00326677 -2,7081726 Yes Yes

5.2 The regression of the cumulative abnormal returns on the rank

The effect of being ranked one position higher is done by regressions of the CARs just on the belonging rankings of the companies. This resulted in much significant and negative relations like expected, since the better the position of a company on the list is, the lower the ranking number is. The results are given in table 5 below together with the standard deviations and the t-values. For these test statistics, α=5% as well as α=10% is used in order to compare the results. All tests are one sided since the alternative hypothesis is that the Cumulative abnormal returns become larger when the ranking increases.

Table 5: The CARs regressed on the rankings coefficient standard deviation t-value significant for α=5% significant for α=10% CAPM -0.0001742 0.0001208 -1.44 No Yes

FAMA 3 -0.0002063 0.0001118 -1.85 Yes Yes

FAMA5 -0.0001971 0.001117 -1.76 Yes Yes

5.3 The regression of the cumulative abnormal returns on the overall score Testing the effect of having a one per cent. higher overall score is done by the same way as being listed one place higher in the list, by regressing the CARs on the overall scores of the companies as a single explanatory variable. With a significance level of α=5%, the results are not significant. With a significance level of α=10%, the results

20

are significant for all the models. The sign of the coefficients is in all cases positive like expected. The higher the overall score, the higher the cumulative abnormal returns are. Again, one sided critical values are used since the alternative hypothesis is that the CARs are increasing when the overall score of companies are increasing. The results are in table 6 below together with the standard deviations and the t-values.

Table 6: The CARS regressed on the overall score coefficient standard deviation t-value significant for α=5% significant for α=10% CAPM 0.0705663 0.0536168 1.32 No Yes FAMA 3 0.0816362 0.0497084 1.64 No Yes FAMA5 0.0774445 0.0496607 1.56 No Yes

5.4 The top 10 companies versus the rest

The top 10 companies are analysed by creating a dummy variable which is equal to one if a company belongs to the top 10 and which equals to zero if it does not. Again, the CARs are regressed on this dummy variable on his own. The results are strong, but not strong enough for α=5% and α=10%. The signs are positive like expected. The results are in table 7 below with the belonging standard deviations and t-values. The critical values are again the critical values for a one-sided test since the alternative hypothesis is that the CARs are bigger for companies listed in the top 10 of the list.

Table 7: The CARs regressed on a dummy variable for the top 10 companies coefficient standard deviation t-value significant for α=5% significant for α=10% CAPM 0.0127882 0.0115436 1.11 No No FAMA 3 0.011415 0.0107611 1.06 No No FAMA5 0.01207033 0.0107282 1.13 No No

21

Because the results were not significant for the top 10 companies, it has been interesting to analysing for which top leading companies the effect is significant, by doing a robustness check. Several dummy variables are made to check for which top leading countries the results are significant enough. This resulted in that the top 11 leading companies are, with a α=10%, significantly different from the rest of the list. The signs are positive for all the models. This means that the cumulative abnormal returns are bigger for the top 11 companies than for the rest of the list. These results are shown in table 8 below together with the standard deviations and the t-values.

Table 8 The CARs regressed on a dummy variable for the top 11 companies coefficient standard deviation t -value significant for α=5% significant for α=10% CAPM 0.0146777 0.0110392 1.33 No Yes FAMA 3 0.0140223 0.0102805 1.35 No Yes FAMA5 0.0147713 0.0102452 1.44 No Yes

5.5 The results in this thesis compared to the results of the existing literature These results are comparable with the results of Murguia and Lence (2015), except the result about the average cumulative abnormal returns. Murguia and Lence found an insignificant effect, while this effect is expected to be positive. In this thesis, this effect is negative and significant for all models and for both α=5% as well as for α=10%. A possible explanation for this will be given in the conclusion part of this thesis. Murguia and Lence found a positive and significant result for having a one per cent. higher overall score, which is the same in this thesis. The coefficient itself is however lower in this thesis. Furthermore, Murguia and Lence compared the top 50 list to the remaining list and they found that these two lists are significantly different from each other. In this thesis the top 10 companies are compared to the rest of this list. This resulted in insignificant results. However, regressing the top 11 companies result in a significantly difference between the two groups. So, the results of Murguia and Lence (2015) are in line with this thesis, while they are not completely the same.

22

6 Conclusion

The results of this thesis indicate that the stock market reacts to the announcement of The Global 100 list, since there is a negative and significant relation of the cumulative abnormal returns for the companies which are in the list. This negative sign could be explained by the so called ‘sell the news’ reaction. This means that investors do buy stocks of companies which are expected to be in the list, and sell these stocks again on the announcement day, or right after the announcement day. These investors bought the stocks with a short-term profit goal and sell it in the days the stock price is expected to be on the highest point. In fact, so many investors do think likewise, which results in a decline in the period close to the announcement day. In this thesis the test period is chosen as three days before the announcement until three days after the announcement, which could be a good explanation for the ‘sell the news’ reaction.

The effect of moving one position closer to the top of The Global 100 list resulted in an increase of (as an average of the three models) 0.01925 per cent. of the cumulative abnormal return. This means that there is statistically enough evidence to conclude that the market indeed reacts to a higher ranking of a company.

However, the result of having a one per cent. higher overall score seems to be larger. According to the average of the three models, having a one per cent. higher overall score results in a cumulative abnormal return which is 7.65 per cent. higher. These results are significant either. So, if the rank and the overall score are compared to each other, the overall score has larger effects on the cumulative abnormal returns and therefore larger effects on the stock values of the companies. Finally, the top leading listed companies are compared to the rest of the list. Unfortunately, there was not enough statistically evidence for a different top 10 leading companies. But trying the same for the top 11 countries resulted in a significant relation with α=10%. The results of the different models are very similar, which makes no big differences between the models.

For further research investors’ behaviour could be analysed in order to draw scientific conclusions about how investors buy and sell their stocks with The Global 100 announcements. This would lead to a better understanding of the results of the existing literature.

23

References

Ameer, R., & Othman, R. (2012). Sustainability practices and corporate financial performance: A study based on the top global corporations.Journal of Business Ethics,108(1), 61-79.

Anderson-Weir, C. H. (2010). How does the stock market react to corporate environmental news?Undergraduate Economic Review,6(1), 9.

Barron, K. E., & Harackiewicz, J. M. (2001). Achievement goals and optimal motivation: Testing multiple goal models.Journal of personality and social psychology,80(5), 706.

Benson, K. L., & Humphrey, J. E. (2008). Socially responsible investment funds: Investor reaction to current and past returns. Journal of Banking & Finance, 32(9), 1850-1859.

Blumenshine, N. T., & Wunnava, P. V. (2010). The value of green: the effect of environmental rankings on market cap.Technology and Investment,1(04), 239.

Bollen, N. P. (2007). Mutual fund attributes and investor behavior.Journal of Financial and Quantitative Analysis,42(3), 683-708.

Chatterji, A. K., Levine, D. I., & Toffel, M. W. (2009). How well do social ratings actually measure corporate social responsibility?Journal of Economics & Management Strategy,18(1), 125-169.

Cheung, A. W. K. (2011). Do stock investors value corporate sustainability? Evidence from an event study.Journal of Business Ethics,99(2), 145-165.

Derwall, J., Bauer, R., Guenster, N., & Koedijk, K. C. (2004). Socially Responsible Investing: The Eco-Efficiency Premium Puzzle.

24

Friedman, M. (1970). A theoretical framework for monetary analysis.journal of Political Economy,78(2), 193-238.

Geczy, C., Stambaugh, R., & Levin, D. (2005). Investing in socially responsible mutual funds.

Goldreyer, E. F., & Diltz, J. D. (1999). The performance of socially responsible mutual funds: incorporating sociopolitical information in portfolio selection.Managerial

Finance,25(1), 23-36.

Gutowski, T., Murphy, C., Allen, D., Bauer, D., Bras, B., Piwonka, T., ... & Wolff, E. (2005). Environmentally benign manufacturing: observations from Japan, Europe and the United States.Journal of Cleaner Production,13(1), 1-17.

Heal, G. (2005). Corporate social responsibility: An economic and financial framework.The Geneva papers on risk and insurance-Issues and practice,30(3), 387-409.

Jones, S., Van der Laan, S., Frost, G., & Loftus, J. (2008). The investment performance of socially responsible investment funds in Australia.Journal of Business Ethics,80(2), 181-203.

MacKinlay, A. C. (1997). Event studies in economics and finance.Journal of economic literature,35(1), 13-39.

Murguia, J. M., & Lence, S. H. (2015). Investors’ Reaction to Environmental Performance: A Global Perspective of the Newsweek’s “Green Rankings”.Environmental and Resource Economics,60(4), 583-605.

Renneboog, L., Ter Horst, J., & Zhang, C. (2008). Socially responsible investments: Institutional aspects, performance, and investor behaviour.Journal of Banking & Finance,32(9), 1723-1742.

25

Appendix I : The CARs for all companies with the 3 models

rank CAPM FAMA3 FAMA5 rank FAMA3 FAMA5

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 -0,0109 0,0118 -0,0178 0,0222 -0,0019 -0,0286 0,0213 0,0275 -0,0219 0,0174 0,0189 -0,0641 -0,0024 0,0249 -0,0335 -0,0217 -0,0082 -0,0463 -0,0094 0,0180 -0,0132 -0,0302 0,0386 -0,0251 -0,0370 0,0277 -0,0083 -0,0398 -0,0304 0,0405 0,0148 -0,0138 -0,0785 -0,0257 -0,0580 0,0063 0,0127 -0,0270 -0,0105 -0,0129 0,0393 -0,0417 0,0286 0,0277 -0,0059 -0,0011 0,0172 -0,0254 0,0315 -0,0173 -0,0018 -0,0104 0,0281 0,0040 -0,0255 0,0202 0,0144 -0,0130 0,0306 0,0271 -0,0529 0,0067 0,0268 -0,0334 -0,0282 -0,0008 -0,0121 -0,0067 0,0716 -0,0165 -0,0191 0,0353 -0,0204 -0,0402 0,0200 -0,0130 -0,0389 -0,0352 0,0393 0,0152 -0,0141 -0,0653 -0,0220 -0,0364 0,0241 0,0072 -0,0133 -0,0100 -0,0319 0,0522 -0,0551 0,0058 0,0244 -0,0025 0,0017 0,0048 -0,0246 0,0511 -0,0194 0,0037 -0,0156 0,0260 0,0037 -0,0258 0,0197 0,0136 -0,0146 0,0285 0,0269 -0,0582 -0,0057 0,0250 -0,0312 -0,0278 -0,0030 -0,0142 -0,0150 0,0095 -0,0155 -0,0237 0,0347 -0,0191 -0,0407 0,0309 -0,0144 -0,0379 -0,0358 0,0379 0,0157 -0,0158 -0,0579 -0,0239 -0,0384 0,0236 0,0039 -0,0109 -0,0157 -0,0328 0,0511 -0,0218 0,0112 0,0260 -0,0051 0,0061 0,0077 -0,0278 0,0590 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 92 93 94 95 96 97 98 99 100 -0,0066 -0,0165 0,0113 -0,0281 -0,0103 -0,0332 -0,0196 -0,0356 0,0757 0,0069 -0,0356 -0,0618 -0,0121 0,0038 -0,0089 0,0036 -0,0140 0,0244 0,0453 0,0017 0,0073 -0,0106 0,0446 0,0327 0,0032 -0,0034 -0,0119 -0,0034 -0,0231 -0,1300 0,0008 -0,0091 -0,0126 0,0002 0,0332 -0,0213 -0,0224 -0,0182 -0,0148 -0,0233 0,0105 -0,1680 -0,0438 -0,0773 0,0308 -0,0055 -0,0405 -0,0367 -0,0081 -0,0183 0,0172 -0,0215 -0,0024 -0,0304 -0,0168 -0,0132 0,0466 -0,0254 -0,0417 -0,0550 -0,0113 -0,0013 -0,0071 -0,0154 -0,0414 0,0280 0,0283 0,0154 0,0172 -0,0124 0,0343 0,0514 -0,0014 -0,0075 -0,0057 -0,0075 -0,0141 -0,1176 -0,0003 -0,0142 -0,0132 0,0112 0,0190 -0,0225 -0,0155 -0,0190 -0,0076 -0,0360 0,0083 -0,1576 -0,0502 -0,0632 0,0245 -0,0048 -0,0367 -0,0400