• No results found

The relationship between financial performance and corporate social responsibility An investigation to causality on Financial Performance and Corporate Social Responsibility.

N/A
N/A
Protected

Academic year: 2021

Share "The relationship between financial performance and corporate social responsibility An investigation to causality on Financial Performance and Corporate Social Responsibility."

Copied!
32
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

The relationship between financial performance and corporate

social responsibility

An investigation to causality on Financial Performance and Corporate Social Responsibility.

Final Version

Abstract:

This research investigates the causal relationship between Corporate Social Responsibility and Financial Performance in an U.S. context from 2007 till 2013. I constructed environmental scores using Principal Component Analysis based on carefully selected environmental indicators which are deemed to solely reflect actual environmental performance. Two regression analyses are conducted regarding environmental performance measurements and stock market returns. The findings confirm that causality runs predominantly from stock market returns to environmental measurements, pointing at risk mitigating steps. In this research some negative impacts of stock market returns on environmental scores are found, which is observed in the environmental index as well as in the environmental performance measurements. An explanation for this is that investments in operating activities cause environmental scores to drop in the short run. As the results show mixed results with different environmental performance measurements, it implies that only some aspects (underlying variables) of the environmental index contribute to stock market returns.

Keywords: Causal relationship; Corporate Social Responsibility; Environmental Performance; Financial Performance; Stock Market Returns.

JEL classification: G300; L210; L250; M140

(2)

1. Introduction

For several decades companies have been searching for ethical guidance to act in an acceptable manner in society. Milton Friedman (1970) proposed his shareholder view or so to say profit maximization view. He argued: “The social responsibility of business is to increase its profits”. According to Friedman’s theory a manager should only act in the interest of shareholders, which is seen nowadays as a very limited view. Freeman (1984) introduced the stakeholder view in which the group that has legitimate claims on companies is extended beyond shareholders. Every group and individual that affects the company, or is affected by the company, is a legitimate stakeholder. Jensen (2001) offers a proposal to the relation between stakeholders’ interest and value maximization, which he calls “enlightened stakeholder theory” or “enlightened value maximization”. It accepts the structure of stakeholder theory but also the criterion of maximization of long run firm value to make the tradeoffs between stakeholders. He captures most of the discussion points between these two views, by evaluating both as collaborating theories instead of opposing. In the field of finance issues like corporate social responsibility (CSR), corporate governance, and ethics are more soft measurements of a firm’s performance. Defining CSR is not that easy, as there are many different definitions pointing in different directions. A definition for CSR is given by Bénabou and Tirole (2010); they state that CSR is about sacrificing profits in the social interest by going beyond legal obligations on a voluntary basis. By looking at the ASSET4 the broad view of CSR is confirmed as their total aggregated score depends on four underlying pillars (Economic-, Environmental-, Social-, and Government-performance). Their indexes give an indication about the level of CSR and in this research specific environmental performance. Moreover, in this research the difference between environmental index and real environmental performance is discussed as well. To put this different, is the environmental index a good reflection of environmental performance of a firm?

(3)

being profitable for a firm and society, while profit maximization limits a firm in its ability to engage in CSR activities. This ongoing CSR debate will be broadly discussed further in the literature overview section. In this research the causality issue between CSR and financial performance will be elaborated on in more detail. Since earlier research is more dedicated to the impact of CSR on financial performance, the causal relation is fewer discussed. For practitioners in particular it is highly relevant to look at causality issues between these measurements of environmental performance and financial performance, in order to draw conclusions on what strategy will fit best. In this I try to contribute to existing literature in specifically looking at the direction of the causality between several environmental measurements and financial performance.

To create a useful dataset to test this specific relationship between financial performance, environmental indexes, and environmental performance, the availability and coverage of the data is a main driver of the choice to use U.S. firms. The data period from 2007 up to 2013 is as well a choice of gathering as much useful data as possible. In investigating the relationship between financial performance and environmental performance regression analysis is used. Besides this, a Granger Causality test is conducted to compare and confirm the results. The results suggest that different underlying components (environmental performance measurements) of the environmental index contribute differently to financial performance. My findings shed new light on the causality between financial performance and environmental performance. First, by looking at actual performance measurements, it shows that no clear conclusion can be drawn about the contribution to financial performance. Second, by investigating various lags, at least Granger causality can be identified, which points to causality running from financial performance to environmental performance.

(4)

2. Literature overview

By looking at CSR, one have to think of the idea behind CSR, it is about going beyond what is expected from a person/firm. Jensen (2009) argues about integrity. Integrity and doing what is expected from a person and doing what is expected by society is very close related to CSR. It is about what is expected from the firm and what it is obligated to do, to be able to act as a member of the society. In some way this is contradicting to the profit maximization theory of Friedman (1970), and the true drivers of CSR will lie somewhere in between these theories. In extension to what is stated by Jensen (2009), Shiller (2013) argues about the public’s negative perception of financiers (and to some extend companies) due to the financial crisis. He states that: “Finance should benefit society”, which is in line with general CSR goals. In the definition of CSR by Bénabou and Tirole (2010), CSR implies sacrificing profits. This would give a rather negative perspective on CSR, indicating that it always comes at a cost without further benefits. Many different studies show that this is not necessarily true and point out advantages of CSR. CSR is rather defined as a program to reduce externalized cost and to avoid distributional conflicts. Market failures are a big reason for firms to engage in CSR as is discussed by Heal (2005). Next different studies in past years on CSR and its effect on financial performance will be discussed. After these different kinds of CSR ratings will be elaborated on to get a good overview of the different CSR measurements and their characteristics. Finally a causal link between CSR and financial performance will be investigated on from previous literature.

2.1. Corporate Social Responsibility and Financial Performance

(5)

Margolis et al. (2009) asked themselves the question: “Does it pay to be green?” They performed a meta-analysis including 251 studies; the effect was positive but small. Although they find, on average, a positive effect for CSR on financial performance they have a rather skeptic view on CSR. They explain that it is not about generating profits but preventing the firm from future losses. Orlitzky et al. (2003) find in a meta-analysis that CSR and to a lesser extent environmental responsibility is likely to pay off. In their meta-analysis, they also try to shed some light on the differences in results from different researches. They attribute this to the different measurements of CSR and financial performance. They find different correlations for CSR with accounting-based measurements and market based measurements. This indicates that one has to be careful with picking measurements and moreover, be careful by interpreting results of earlier studies. Alloche & Larouche (2005) also find that CSR activities have a positive impact on financial performance and this is strongest in UK context. They specifically state that different aspects of CSR can have different impacts on financial performance. Moreover, Dam & Scholtens (2015) studied the impact of different financial performance measurements (Market-to-Book ratio, Return on Assets and stock market returns). They state that most of the observed relationships between CSR and financial performance of the last three decades are not contradicting, but rather it was aligned evidence of a correlation between CSR and various financial performance measurements. Interesting to point out in this relation is that Margolis et al. (2009) find that effects from the past decade (result from subsample of 106 studies out of 251) are even smaller than the small overall effect. They checked for different (stronger) relations under different conditions, but only find significant results for revealed misdeeds leading to negative influences on financial performance.

(6)

monitor and thereby have full disclosure of what happens within firms. The results are significant but economically modest, firms lower on CSR (firms that have concerns) pay 7 to 18 bps more than firms higher on CSR. So although the result is economically modest, firms higher on CSR are indeed “rewarded” with a lower cost of bank loans. While Goss & Roberts (2011) looked at the banking industry as a monitor of firms, Wu & Shen (2013) looked at the effects of CSR on banks. They find that banks high on CSR have better net interest income, non interest income, return on assets, return on equity, and lower non performing loans. So indeed banks who engage in CSR show better financial performance than others. Interesting to mention is that Wu & Shen (2013) state that the greenwashing argument can be ruled out since engaging in CSR increases cost, but revenues increase even more. Hence, engaging in CSR is always beneficial to financial institutions. Using a regression model Lioui & Sharma (2012) investigated the effect of Environmental Corporate Social Responsibility (ECSR) on financial performance. They found that both CSR strengths and concerns are negatively related to return on asset and Tobin’s Q. This negative effect is due to the fact that engaging in CSR as well as having CSR concerns is seen as potential cost. Additional value can still be created because CSR activity fosters R&D efforts. Dimson et al. (2014) showed that companies, after they engaged successfully in CSR activities, improved accounting and governance performance and increased institutional ownership.

Both Ribando & Bonne (2010), and Weber et al. (2005) have looked at the ASSET4, what it tells investors, and in a broader perspective what it tells about a firms’ performance. Ribando & Bonne (2010) find that ASSET4 scores add value on an absolute basis as well as a risk-adjusted basis, which is consistent with the theory that firms high on CSR measurements will create long term shareholder value. Next to that, Weber et al. (2005) find a positive correlation between CSR activities, the impact on sustainable development and financial performance. Both researches are in some way related to the owners/creators of the ASSET4, which have to be mentioned to interpret the results. There is a lack of more research on CSR activities using the ASSET4 as measurements.

(7)

positively affect financial performance (in terms of stock returns). However, general statements about the magnitude of this affect cannot be made at this level, as results are mixed.

2.2. Corporate Social Responsibility ratings

Over the past decades the field of CSR research has been widely studied upon. As Lemon et al. (2011) states, CSR has moved past definition and towards measurement and implementation. Pillarisetti & van den Bergh (2010) have looked at the nationwide aggregated indexes of sustainability. In particular they look at: the “World Bank’s ‘Genuine Savings”, the “Ecological Footprint”, and the “Environmental Sustainability Index”. They state that the indexes yield different results and after aggregating them only 29 countries in the world economy are considered to be sustainable by all three indexes. This in fact points out a major problem in research to CSR, which measurement should be used and how well does this measurement represents actual performance? In this Chaterji et al. (2009) try to answer the question whether ESG rating from KLD provide transparency about prior and future CSR. In general they find that KLD does a reasonable job in aggregating past environmental performances. Moreover, concerns indeed predict future environmental misdeeds, but strengths do not predict environmental outcomes. In his paper Jensen (2001) discussed the balanced scorecard in which he argues that it is impossible for managers to pursue more than one objective. This can be extended to CSR indexes, which also contain a lot of different measurements. Question is whether or not a lot of performance measurements smoothen out the index score between companies and whether or not it is about adopting instead of implementing CSR activities. Fiss & Zajac (2004) find in a corporate governance context in Germany that firms first espouse a strategy but do not implement it. In a CSR context this can lead to statements intending to engage in CSR but real activities lack behind.

2.3. Causality of Corporate Social Responsibility and Financial Performance

(8)

activities are helpful to become a bit more successful. Waddock & Graves (1997) labeled better financial performance leading to an opportunity to invest in CSR activities as “Slack Resources”. The other way round, CSR activities improving the relationship with key stakeholders resulting in better financial performance, is called “Good Management”. They find that CSR activities are positively associated with prior financial performance, supporting the definition of “Slack Resources”. They also find evidence for CSR leading to better financial performance and hence supporting the “Good Management” definition.

Scholtens (2008) found preliminary evidence that causation runs from financial performance to CSR, although the specific interaction patterns vary along different dimension. Nelling & Webb (2009) find, using time series fixed effects approach, that the relation between CSR and financial performance is much weaker than previously thought. Moreover, they indeed find that strong financial performance in terms of stock market measurements, leads to an increase in CSR engagement devoted to employee relations. In conclusion, they stated that CSR is not driven by financial performance but more by unobservable firm characteristics. In a Canadian study, Makni et al. (2009) find only a relationship for a composite CSR measurement and financial performance measured by stock market returns. However, by using an individual environmental CSR measurement, they find a negative impact on three financial performance measurements (return on assets, return on equity, and stock market return). The latter is in line with the negative synergy hypothesis, stating that firms high on CSR have lower profits. Makni et al. (2009) explain this contradiction in findings by the size of Canadian firms in comparison to U.S. firms.

(9)

3. Data &Methodology

I use all U.S. companies with an ASSET4 rating. These companies are gathered from the constituents list LA4CTYUS which consists 1,009 U.S. companies. The ASSET4 is a relatively new CSR rating. Started in 2002, ASSET4 has gathered more than 750 data variable in four different pillars. These four pillars are: Economic performance, Environmental performance, Social performance, and Corporate Governance performance. The focus in this study will be on the Environmental performance of the ASSET4. Due to the fact that the ASSET4 is relatively young and the inclusion of firms grew gradually over time, data will be used from 2007 up to 2013. Because the ASSET4 is reforming its database, 2013 was the last year which was available at this moment. In the next subsections the methodology used, financial performance, environmental performance, the control variables, and correlations will be elaborated on more specific.

3.1. Methodology

Because of the evaluation of the underlying variables of the environmental pillar some kind of data reduction method has to be used, to get a couple of components underlying the actual data, which can be used in further research. Field (2009) discussed two different approaches to reduce a dataset; Factor Analysis and Principal Component Analysis. However, Field (2009) cited Guadagnoli and Velicer (1988) who concluded by extensive literature review that the solutions from both data reduction methods differ little. Data reduction is just a way to obtaining useful data, it is nothing more than restructuring and obtaining underlying characteristics. Precise methods in obtaining the components with PCA will not be elaborated on. Briefly explaining the method used; first the variables which were useful to conduct a PCA where obtained. This means only variables with coverage of 50 percent or more. Next a PCA is conducted using Varimax as rotation method. Unfortunately, because of the many missing values that occur in the dataset, to conduct a PCA the missing values have to be replaced with means. This results in a bias to the output components, but because of the many missing values this cannot be resolved. As result the components are presented in the environmental performance section.

(10)

relationship. Two of these studies (Waddock & Graves 1997, and Nelling & Webb 2009) are using ordinary least square (OLS) regression analysis to analyze the causal relationship. Makni et al. (2009) however uses Granger causality. According to Bressler & Seth (2011) Weiner-Granger Causality or Granger Causality is a straightforward method in trying to look for a significant increase in explaining a future variable by adding a second variable other as the variable which one wants to predict. Scholtens (2008) uses both regression analysis and Granger Causality. Besides OLS regression analysis, Nelling and Webb (2009) also used a fixed effects model, to control for unobservable variables. In this research the methodology of the majority of previous mentioned researches will be followed. An OLS regression model will be tested as well as a fixed effects model, as Nelling and Webb (2009) do. Most of the research is done by taking one year lagged variables, only Scholtens (2008) varies from that in taking three and five years lagged variables. In this research three years lagged variables will be taken. Because following economic intuition by changing business strategy, it takes several years to measure a change in strategy in financial performance measurements. This might result in observing impacts not directly in one year lags but in more than one year lags. In addition to the regression analyses described before, a Granger Causality test will be conducted as well. This is in line with almost all previous research, and will give a good comparison with the regression analyses done. There will be two different analysis using different variables for environmental measurement. The first analysis uses the Environmental index as given by the ASSET4 as measurement. The second analysis uses the environmental performance measurements as are conducted by the PCA. Keep in mind that although these variables are nothing more than components underlying the environmental index, and hence another but rather more actual measurement for environmental performance. The different equations are presented below:

𝐸𝑁𝑉𝑚𝑒𝑎𝑠𝑢𝑟𝑒𝑚𝑒𝑛𝑡𝑖,𝑡 = 𝛽0+ 𝛽1𝐹𝑃𝑡−𝑘+ 𝛽2𝐿𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝑡−𝑘+ 𝛽3𝑆𝑖𝑧𝑒𝑡−𝑘+ 𝜀𝑖𝑡 (1)

𝐹𝑃𝑡 = 𝛽0+ 𝛽1𝐸𝑁𝑉𝑚𝑒𝑎𝑠𝑢𝑟𝑒𝑚𝑒𝑛𝑡𝑖,𝑡−𝑘+ 𝛽2𝐿𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝑡−𝑘+ 𝛽3𝑆𝑖𝑧𝑒𝑡−𝑘+ 𝜗𝑖𝑡 (2)

(11)

3.2. Financial Performance

For the measurement of financial performance there can be chosen for either accounting based measurements or stock market based measurements, each with own advantages and disadvantages. McGuire et al. (1988) stated that accounting based measures are subjected to biases from different managerial aspects and accounting procedures. Market returns as a stock market based measurement have several advantages over accounting based measurements. For instance, it is less subjected to managerial decisions and accounting procedures, moreover stock market based measurements represents investors’ evaluations. Scholtens (2008) also points out a weakness of market based measurements, as asymmetric information can turn evaluation of investors into wrong representation of fair value. Despite this drawback the stock market returns will be used to measure financial performance. From the return index the return is calculated by taking yearly log returns. For causality testing, one has to think about timing of the variables used. Going from the fact that scores on the ASSET4 are reported after a year is closed, one can assume that the scores reflect a whole year’s performance. For instance, the environmental index of 2007 represents the performance of the whole year 2007 from January 2007 until January 2008. To match this with financial data I take yearly financial data from January𝑡 toJanuary𝑡+1, in order to match the financial

data properly with the CSR data.

3.3. Environmental performance

(12)

and name are presented. Following Scholtens & Dam (2007), who try to look for similarities in output contributions to a specific component, the four components are labeled; Renewable Energy Products, Product Impact, Labels, and Emission.

To get the number of variables preferable for the input of a regression analyses, a cutoff is made different from the original output of the PCA. According to Field (2009) by looking at the point of inflexion one can obtain a correct number of components from the Principal Component Analysis. Keeping in mind that the total variance explained is 0.3708, one has to be careful in interpreting the results. Thomson Reuters, who provide the data for the ASSET4, are aware of this lack of coverage of the Asset4. In fact, they are imposing a new framework as this Thesis is written1, which should significantly increase the company’s coverage.

3.4. Control variables

For the control variables I rely on previous research, the most common control variables are related to size and leverage. Size is accounted for by Waddock & Graves (1997), Bernadette et al. (2001), De Haan et al. (2012), and Makni et al. (2009), by taking the market value of equity which is the current market price multiplied by the number of shares outstanding. Waddock & Graves (1997), and Makni et al. (2009) also uses debt ratio as a proxy for riskiness of a firm. Following this, the leverage is used as a control variable, which is calculated by the sum of long term and short term debt divided by common equity times hundred.

3.5. Correlation of variables

By looking at the correlation matrix (Table 2) there can be seen whether or not the variables relate to each other. As can be seen and as expected of course the environmental performance measurements correlate with the environmental index. However this correlation is positive, it is far from perfect. Almost all of the environmental measurements correlate positively with Size, indicating that a bigger firm will score better on environmental measurements. This can be due to the fact that environmental performance is less costly for bigger firms, as Makni et al. (2009) explain. None of the environmental measurements correlate with Return.

(13)

Table 1: Principle Component Analysis for ASSET4 environmental data of U.S. Companies, 2007-2013.

Components

1 2 3 4

Renewable Energy

Products Product Impact Labels Emission

ENVvariable1 .549 .098 .290 -.027 ENVvariable2 .354 .201 .149 -.184 ENVvariable3 .272 -.162 -.143 .320 ENVvariable4 .189 .057 -.171 .455 ENVvariable5 .372 .025 -.247 .488 ENVvariable6 .070 .012 -.104 .311 ENVvariable7 .051 -.166 -.014 -.145 ENVvariable8 .054 -.151 -.034 -.118 ENVvariable9 .363 -.344 -.158 .195 ENVvariable10 .140 .224 .592 .146 ENVvariable11 .357 .607 -.450 -.210 ENVvariable12 .356 .617 -.479 -.201 ENVvariable13 .206 .369 .213 .143 ENVvariable14 .082 .309 .368 .204 ENVvariable15 .376 .174 .544 .054 ENVvariable16 .742 -.507 -.094 -.159 ENVvariable17 .492 .078 -.005 -.229 ENVvariable18 .564 -.093 .211 -.207 ENVvariable19 .279 .281 -.160 -.202 ENVvariable20 .742 -.507 -.094 -.159 ENVvariable21 .234 .190 -.205 .286 ENVvariable22 .482 .107 .258 .202 ENVvariable23 .123 -.026 -.171 .520

In the table all the variables are labeled as ENVvariable plus a number. In total there are 23 useful variables retained from the ASSET4. The translation to the codes as the variables are labeled in the ASSET4 is presented in Appendix 1 table 1.1. For each variable the contribution to the components as they are conducted from the Principal Component Analysis are presented. The bold numbers are those who contribute 0,25 or more to the component, these variables are of particular interest for the contribution to the component. By looking at the variables with bold numbers a name for the component can be formulated.

Table 2: Correlation matrix of all the variables which will be used in the regression

Return Envscore REP

Product

Impact Labels Emission Leverage Size

Return 1,0000 ENVscore 0,0235 1,0000 REP 0,0021 0,3276 1,0000 Product Impact 0,0180 0,0443 0,0004 1,0000 Labels -0,0080 0,4186 0,0019 0,0002 1,0000 Emission 0,0160 0,1390 0,0007 -0,0002 -0,0025 1,0000 Leverage 0,0001 0,0174 -0,0031 -0,0037 0,0177 -0,0050 1,0000 Size 0,0694 0,3460 0,1059 0,0204 0,1915 0,2431 -0,0027 1,0000

(14)

4. Results

In discussing the two analyses as they are presented in the “Data & Methodology” section, first the analysis for the environmental index as constructed by the ASSET4 will be discussed. After this, the environmental performance measurements computed by PCA will be regressed on the returns, as presented in the second analysis. Then, the Granger Causality test will be presented and elaborated on, by discussing this test the findings are compared with findings in the OLS regression. Finally, the robustness of the results will be discussed.

4.1. Analysis of environmental index and financial performance

In running the regression analysis of environmental index and financial performance as presented in Table 3, first the returns are taken as dependent and the environmental index as explanatory variable. Only for a one year lag, statistical significant evidence (p<0.01) is found for environmental index influencing the returns. This result is economically insignificant, as its impact is only 0.0007. However, by running the regression with fixed effects, which means controlling for underlying firm specific effects, statistically significant results are found (p<0.01) on all three lags. Implying that, in line with earlier findings, environmental index indeed influence returns. Moreover, this implies that earlier year environmental indexes indeed impact returns in the future. However, looking at the economically significance, it can be stated that results are relatively small (0.0024 for one year lag, and 0.0036 for two year lags) and for three year lags the impact is even negative (-0.0027). This indicates that earlier year environmental indexes negatively influence stock market returns.

(15)

positively related to the environmental index (1.1280) and significant at a 1% level (p<0.01) . This can be explained by firms who are doing good for several years, tend to perform worse on environmental performance at first, but will overcome this lack of environmental activities if things keep on going good.

In summary, there can be concluded that in line with earlier findings environmental performance based on environmental indexes will lead to better financial performance, however these results are economical small. However, care is needed when interpreting the results for the relationship running from returns to environmental performance, because most of the findings are insignificant. It seems that returns are negatively influencing environmental performance. The causality of this relationship runs predominantly from returns to environmental index, which is in line with Scholtens (2008). As the signs are negative this implies that better financial performance will lead to a reduction in score on the environmental index.

(16)

Table 3: The relationship between environmental index and financial performance of U.S. firms, 2007-2013

Return ENVscore

(fixed effects) (fixed effects)

β β β β Constant 0.0229** -0.1539*** 38.4843*** 43.2590*** (0.0195) (0.0000) (0.0000) (0.0000) ENVscore 0.0000 0.0016*** (0.9677) (0.0024) Return -0.0350 1.1280*** (0.9677) (0.0024) Leverage 0.0000 0.0000 0.0003 0.0002* (0.9824) (0.7944) (0.1303) (0.0105) Size 0.0000*** 0.0000*** 0.0003*** 0.0000 (0.0000) (0.0000) (0.0000) (0.6308) Constant 0.0143 0.0805*** 39.0517*** 43.6479*** (0.1884) (0.0048) (0.0000) (0.0000) ENVscore(-1) 0.0007*** 0.0024*** (0.0024) (0.0001) Return(-1) -1.4627* -0.3121 (0.0980) (0.3616) Leverage(-1) 0.0000 0.0000 0.0001 0.0000 (0.5265) (0.9499) (0.6499) (0.8800) Size(-1) 0.0000*** 0.0000*** 0.0003*** 0.0000 (0.0000) (0.0000) (0.0000) (0.3506) Constant 0.0313*** -0.0473 39.9709*** 44.5377*** (0.0088) (0.1556) (0.0000) (0.0000) ENVscore(-2) 0.0004 0.0036*** (0.1071) (0.0000) Return(-2) -1.2500 -0.1253 (0.1723) (0.6920) Leverage(-2) 0.0000 0.0000 0.0000 -0.0001 (0.2643) (0.5193) (0.9283) (0.4009) Size(-2) 0.0000*** 0.0000*** 0.0003*** 0.0000 (0.0008) (0.0000) (0.0000) (0.6706) Constant 0.1329*** 0.1995*** 41.1069*** 45.4864*** (0.0000) (0.0000) (0.0000) (0.0000) ENVscore(-3) 0.0000 -0.0027*** (0.8823) (0.0000) Return(-3) -0.8702 -0.0996 (0.3681) (0.7199) Leverage(-3) 0.0000 0.0000 0.0002 0.0001 (0.7208) (0.3513) (0.6303) (0.2194) Size(-3) 0.0000 0.0000*** 0.0003*** 0.0000 (0.5157) (0.0000) (0.0000) (0.8739)

(17)

Switching the dependent and explanatory variable leads to running a total of eight regressions, each in which an environmental performance measurement is dependent on stock market returns, as presented in table 5. Contrary to findings for the environmental index, stock market returns has no statistically significant effect on any of the environmental performance measurements in the same year. In one year lag the measurements for Product Impact (p<0.1), Labels (p<0.01), and Emission (p<0.01) are statistically significant in which only Emission is confirmed to be statistically significant performing a fixed effects test. The measurements for Labels and Emission are negatively affected by stock market returns, which is in line with findings in the first analysis with environmental index. Product Impact however is positively related to stock market returns. The positive effect can be explained by the fact that Product Impact is related to R&D expenditures and Agrochemical revenues, hence increasing profits in the long run. However, increase in Labels and Emission comes at a cost. In two year lags and three year lags the results are rather mixed. For two year lags the variables Product Impact (p<0.05) and Labels (p<0.05) are positively statistically significant. For three year lags the variables Renewable Energy Products (p<0 .05), Labels (p<0.01), and Emission (p<0.1 for OLS and p<0.05 for fixed effects) are statistically significant and for Product Impact and Labels the sign remain the same, as for Emission the sign changes from negative to positive. This might be because emission lags behind on returns as is explained earlier.

(18)

Table 4: Relationship from environmental performance measurements to financial performance of U.S. firms, 2007-2013 Return (fixed effects) β β Constant 0.0307*** (0.0000) -0.0701*** (0.0000)

Renewable Energy Products -0.0023 (0.6876) 0.0068 (0.5517)

Product Impact 0.0073 (0.1979) -0.0110 (0.4096) Labels -0.0095 (0.1050) 0.0052 (0.6802) Emission -0.0004 (0.9488) 0.0074 (0.3490) Leverage 0.0000 (0.9700) 0.0000 (0.6464) Size 0.0000*** (0.0000) 0.0000*** (0.0000) Constant 0.0362*** (0.0000) 0.1696*** (0.0000)

Renewable Energy Products(-1) 0.0093 (0.1599) 0.0042 (0.7442)

Product Impact(-1) 0.0095 (0.1395) -0.0137 (0.4332) Labels(-1) 0.0059 (0.3891) 0.0184 (0.2304) Emission(-1) 0.0087 (0.2527) -0.0066 (0.5316) Leverage(-1) 0.0000 (0.5646) 0.0000 (0.7593) Size(-1) 0.0000*** (0.0000) 0.0000*** (0.0000) Constant 0.0324*** (0.0000) 0.0824*** (0.0000)

Renewable Energy Products(-2) 0.0181** (0.0200) 0.0670*** (0.0000)

Product Impact(-2) 0.0040 (0.5879) -0.0145 (0.5232) Labels(-2) 0.0196** (0.0132) 0.0604*** (0.0015) Emission(-2) 0.0025 (0.7752) -0.0032 (0.8025) Leverage(-2) 0.0000 (0.2033) 0.0000 (0.3785) Size(-2) 0.0000*** (0.0043) 0.0000*** (0.0000) Constant 0.1637*** (0.0000) 0.1170*** (0.0000)

Renewable Energy Products(-3) -0.0044 (0.5035) -0.0072 (0.5892)

Product Impact(-3) 0.0031 (0.6054) -0.0038 (0.8302) Labels(-3) 0.0092 (0.1599) -0.0050 (0.7529) Emission(-3) -0.0092 (0.1798) -0.0069 (0.4904) Leverage(-3) 0.0000 (0.6304) 0.0000 (0.2664) Size(-3) 0.0000 (0.3948) 0.0000*** (0.0000)

(19)

Table 5: The relationship from financial performance to environmental performance measurements of U.S. firms, 2007-2013

Renewable

Energy Products Product Impact Labels Emission

(fixed effects) (fixed effects) (fixed effects) (fixed effects) β β β β β β β β Constant -0,0383*** 0,0378*** -0,0072 0,0054 -0,0732*** 0,0084 -0,0962*** -0,0049 (0,0033) (0,0019) (0,5845) (0,6120) (0,0000) (0,4589) (0,0000) (0,7842) Return -0,0099* 0,0083 0,0349 -0,0109 -0,0425 0,0033 0,0006 0,0217 (0,0709) (0,5997) (0,1954) (0,4359) (0,1057) (0,8241) (0,9826) (0,3479) Leverage 0,0000 0,0000 0,0000 0,0000 0,0000 0,0000 0,0000 0,0000 (0,8225) (0,8349) (0,7728) (0,2557) (0,1449) (0,4909) (0,7197) (0,8484) Size 0,0000*** 0,0000*** 0,0000 0,0000 0,0000*** 0,0000 0,0000*** 0,0000 (0,0000) (0,0013) (0,1192) (0,7263) (0,0000) (0,7822) (0,0000) (0,5448) Constant -0,0257* 0,0674*** -0,0071 -0,0189 -0,0625*** 0,0085 -0,0864*** 0,0078 (0,0823) (0,0000) (0,6264) (0,1074) (0,0000) (0,4771) (0,0000) (0,7069) Return(-1) -0,0197 0,0121 0,0517* 0,0005 -0,0818*** -0,0160 -0,0754*** -0,0762*** (0,4963) (0,4404) (0,0714) (0,9728) (0,0040) (0,2740) (0,0092) (0,0026) leverage(-1) 0,0000 0,0000 0,0000 0,0000 0,0000 0,0000 0,0000 0,0000 (0,9334) (0,8554) (0,7297) (0,1096) (0,2576) (0,2284) (0,5383) (0,9821) Size(-1) 0,0000*** 0,0000*** 0,0000 0,0000* 0,0000*** 0,0000 0,0000*** 0,0000 (0,0000) (0,0001) (0,2111) (0,0506) (0,0000) (0,2225) (0,0000) (0,4250) Constant -0,0221 0,0517*** -0,0060 -0,0035 -0,0519*** 0,0156 -0,0733*** -0,0448* (0,1863) (0,0001) (0,7181) (0,7879) (0,0014) (0,2209) (0,0000) (0,0582) Return(-2) -0,0157 0,0003 0,0753** 0,0215 -0,0321 0,0339** 0,0168 0,0006 (0,6110) (0,9828) (0,0139) (0,1470) (0,2868) (0,0205) (0,5895) (0,9810) leverage(-2) 0,0000 0,0000 0,0000 0,0000 0,0000 0,0000 0,0000 0,0000 (0,5202) (0,9298) (0,2479) (0,9581) (0,9089) (0,9323) (0,7859) (0,9780) Size(-2) 0,0000*** 0,0000 0,0000 0,0000 0,0000*** 0,0000 0,0000*** 0,0000*** (0,0000) (0,1592) (0,3028) (0,5901) (0,0000) (0,1608) (0,0000) (0,0001) Constant -0,0158 0,0421*** -0,0012 -0,0056 -0,0281 0,0636*** -0,0472** 0,0615** (0,4035) (0,0009) (0,9498) (0,7097) (0,1300) (0,0000) (0,0180) (0,0261) Return(-3) -0,0457 -0,0295** 0,0288 -0,0254 -0,0904*** -0,0218 0,0626* 0,0628** (0,1638) (0,0325) (0,3667) (0,1230) (0,0051) (0,1499) (0,0705) (0,0375) leverage(-3) 0,0000 0,0000 0,0000 0,0000 0,0000 0,0000 0,0000 0,0000 (0,4943) (0,9050) (0,2472) (0,9379) (0,8646) (0,9360) (0,6520) (0,6538) Size(-3) 0,0000*** 0,0000 0,0000 0,0000 0,0000*** 0,0000 0,0000*** 0,0000 (0,0000) (0,8695) (0,6953) (0,6168) (0,0000) (0,3039) (0,0000) (0,8334)

(20)

4.3. Granger Causality

By performing a Granger Causality test, additional evidence is tried to be found on the direction of causality between financial performance in terms of stock market return and CSR in terms of environmental measurements. By comparing these results by the results stated in section 4.2 and 4.3, hopefully extra evidence can be found about the causality relation and the validity of earlier findings. As presented in Table 6, the output of the Granger Causality tests suggests that causality only runs from returns to environmental index statistically significant (p<0.05) by taking one year lag. This is in line with the results of the first analysis, in which there was an economic significant relationship running from stock market returns to environmental index, as was predicted based on earlier research. This Granger Causality test does not provide any information about the negative impact. For the environmental performances measurements, as are computed by the PCA, in line with findings of the second regression analysis results are mixed. With lags of one, three, and five years all variables and directions of causality are at least once statistically significant, except for the causality running from Renewable Energy Products to stock market returns. This is contrary to findings in the second regression analysis in which Renewable Energy Products seems to be one of the few statistically significant explanations of returns. For the other variables significance varies from p<0.01 to p<0.1, and causality runs both from environmental performance measurements to stock market returns as the other way round.

(21)

returns only drive part of the environmental performance measurements, the Granger Causality test indicates that causation runs both ways for all variables.

4.4. Robustness check

To check for robustness of this research and to some extent to use a dataset which is more in line with other research, the ESG index is used in a same analysis as presented above for the environmental index and environmental performance measurements. The ESG index is a total aggregated score of all aspects of the ASSET4 and is computed by the ASSET4 self, to give one score as a CSR measurement. As this robustness check is not part of the main research, results are presented in Appendix 3

(22)

Table 6: Output Granger Causality Tests

Lags: 1

Null Hypothesis: Obs. F-Statistic

ENVscore does not Granger Cause Return 4,984 2.4256 (0.1194) Return does not Granger Cause ENVscore 4.6560** (0.0310) Return does not Granger Cause Renewable Energy Products 5,521 0.2402 (0.6241) Renewable Energy Products does not Granger Cause Return 0.7875 (0.3749) Return does not Granger Cause Product Impact 5,521 2.9388* (0.0865) Product Impact does not Granger Cause Return 3.0268* (0.0820) Return does not Granger Cause Labels 5,521 1.1897 (0.2754) Labels does not Granger Cause Return 0.0354 (0.8507) Return does not Granger Cause Emission 5,521 6.1308** (0.0133) Emission does not Granger Cause Return 0.4471 (0.5037)

Lags: 3

ENVscore does not Granger Cause Return 3,088 1.0972 (0.3490) Return does not Granger Cause ENVscore 1.5155 (0.2084) Return does not Granger Cause Renewable Energy Products 3,611 1.6035 (0.1864) Renewable Energy Products does not Granger Cause Return 0.3067 (0.8206) Return does not Granger Cause Product Impact 3,611 1.9529 (0.1189) Product Impact does not Granger Cause Return 1.1107 (0.3433) Return does not Granger Cause Labels 3,611 3.7149** (0.0110) Labels does not Granger Cause Return 3.1940** (0.0226) Return does not Granger Cause Emission 3,611 4.5300*** (0.0036) Emission does not Granger Cause Return 2.1191* (0.0957)

Lags: 5

ENVscore does not Granger Cause Return 1,301 0.7664 (0.5740) Return does not Granger Cause ENVscore 0.5186 (0.7624) Return does not Granger Cause Renewable Energy Products 1,759 3.2976*** (0.0057) Renewable Energy Products does not Granger Cause Return 1.0556 (0.3833) Return does not Granger Cause Product Impact 1,759 1.5708 (0.1651) Product Impact does not Granger Cause Return 0.2902 (0.9186) Return does not Granger Cause Labels 1,759 1.5390 (0.1745)

Labels does not Granger Cause Return 2.0199* (0.0730)

Return does not Granger Cause Emission 1,759 0.8027 (0.5476) Emission does not Granger Cause Return 0.8054 (0.5457)

(23)

5. Conclusion & Limitations

This research is aimed at the causal relationship between Corporate Social Responsibility (CSR) and Financial Performance (financial performance). The ASSET4 database is used in measuring CSR, which differs from earlier studies which in almost all cases the KLD (Kinder, Lydenberg, and Domini) database is used. From the ASSET4 the Environmental pillar is used to study the relationship with financial performance. The other three pillars of the ASSET4 are left aside (Economic-, Social-, and Corporate Governance-performance). This focus allows to dig deeper into the underlying aspects of the score composition.

(24)

environmental performance measurements is found. Contributing to earlier research this research confirms that causality runs predominantly from stock market returns to environmental measurements. Indicating that CSR activities can be seen as risk mitigating steps. Moreover, in trying to unravel the environmental index in four components which measure actual performance, mixed results indicate that not every aspect of environmental measurement contributes to stock market returns. Firms scoring high on certain variables but low on others might generate a higher stock market return as firms high on overall CSR scores. This suggestion could be a good bases for further research in which firms with same CSR performance can be analyzed, and there financial performance can be compared. To check whether indeed aggregated scores can be smoothen out, resulting in less explanation of financial performance.

From this research there appears to be some causal relation running from environmental measurement to stock market return. Although one can argue whether this causation is based on a few aspects of environmental measurements, for the environmental index this direction of causal relation is not supported by the Granger Causality test. In contrast there seems to be a relation from stock market returns to environmental index. Direction of causality between environmental performance measurements and stock market returns is rather mixed as the results from both analyses and the Granger Causality test suggests (in Appendix 2, table 2.1 and 2.2 an overview of the results is given). This is in line with earlier research explaining the virtuous circle between CSR activities and financial performance. One important finding in this research is the negative impact of stock market returns on environmental measurements, which is observed in the first analysis and for some environmental performance measurements in the second analysis. Best way to explain this is by investments in operation activities causing environmental scores to drop in the short run. From the robustness test it can be seen that this may be the case for other scores as well (economic, social, and corporate governance), however this seems to be more a short run effect. This is partly in line with the slack resource theory, which in this context means that if there are investment opportunities, investment in CSR will lag behind.

(25)

be a useful measurement. As this research is done, the ASSET4 is improving, in which they try to increase the coverage. In conducting the environmental performance measurements by looking at the contribution, a name is giving to these components. This might be not completely correct, but does not influence any of the results. As in this research is stated, different CSR variables may have different contributions to stock market returns. This might also come from particular variables correlating with stock market returns, more as others. In this there are enough open ends to study CSR in relation to firm’s performance.

Acknowledgement

(26)

References

Allouche, J. and Laroche, P., 2005. A meta-analytical investigation of the relationship between corporate social and financial performance. Revue de gestion des resources

humaines, 57, 18.

Bénabou, R. and Tirole, J.,2010. Individual and Corporate Social Responsibility.

Economica, 77, 1–19

Blair, M., 1995. Ownership and control rethinking corporate governance for the twenty-first century.

Bressler, S. and Seth, A., 2011. Wiener-Granger Causality: a well established methodology. Neuroimage, 58(2), 323-329.

Chatterji, A., Levine, D. and Toffel, M., 2009.How Well Do Social Ratings ActuallyMeasure Corporate Social Responsibility? Journal of Economics & Management

Strategy,18(1), 125–169.

Crane, A. and Matten, D., 2010.Business Ethics: Managing Corporate Citizenship and Sustainability in the Age of Globalization. Oxford University Press.

Dam, L. and Scholtens, B., 2015. Toward a theory of responsible investing: On the economic foundations of corporate social responsibility. Resource and Energy Economics, 41, 103-121.

Dam, L., Koetter, M. and Scholtens, B., 2009. Why Do Firms Do Good? Evidence From Managerial Efficiency. SSRN working paper

De Haan, M., Dam, L. and Sholtens, B., 2012. The drivers of the relationship between corporate environmental performance and stock market returns. Journal of Sustainable

Finance & Investment, iFirst article, 1–38.

Dimosn, E., Karakaş, O. and Li, X., 2014. Active Ownership. Available at SSRN

2154724.

El Ghoul, S., Guedhami, O., Kwok, C. and Mishra, D., 2011. Does corporate social responsibility affect the cost of capital? Journal of Banking & Finance, 35, 2388-2406.

Field, A., 2009. Discovering Statistics Using SPSS. Sage.

Fiss, P. and Zajac, E., 2004. The Diffusion of Ideas over Contested Terrain: The (Non)adoption of a Shareholder Value Orientation among German Firms. Administrative

Science Quarterly, 49(4), 501-534.

(27)

Friedman, M., 1970. The social responsibility of business is to increase its profits. New

York Times,September 13.

Goss, A. and Roberts, G., 2011. The impact of corporate social responsibility on the cost of bank loans. Journal of Banking & Finance, 35, 1794-1810.

Gregory, A., Tharyan, R. and Whittaker, J., 2014. Corporate Social Responsibility and Firm Value: Disaggregating the Effects on Cash Flow, Risk and Growth. Journal of Business

Ethics, 124, 633-657.

Guadagnoli,E. and Velicer, W., 1988. Relation of sample size to the stability of component patterns. Psychological Bulletin, 103(2), 95-112.

Heal, G., 2005. Corporate Social Responsibility: An Economic and Financial Framework. The Geneva Papers, 30, 387-409.

Jensen, M., 2001. Value Maximization, Stakeholder Theory, and the Corporate Objective Function. Journal of Applied Corporate Finance, 14(3), 8-21.

Jensen, M., 2009. Integrity: Without It Nothing Works. Rotman Magazine: The

Magazine of the Rotman School of Management, 16-20.

Lemon, K., Roberts, J., Raghubir, P. and Winer, R., Measuring the Effects of Corporate Social Responsibility. Conference Board Incorporated, 3(7).

Lioui, A. and Sharma, Z., 2012. Environmental corporate social responsibility and financial performance: Disentangling direct and indirect effects. Ecological Economics, 78, 100-111.

Makni, R., Francoeur, C. and Bellavance, F., 2009. Causality Between Corporate Social Performance and Financial Performance: Evidence from Canadian Firms. Journal of

Business Ethics, 89, 409-422.

Margolis, J., Elfenbein, H. and Walsh, J., 2009. Does it Pay to be Good … and does it Matter? A Meta-Analysis of the Relationship between Corporate Social and Financial Performance. Retrieved from: http://ssrn.com/abstract=1866371.

McGuire, J., Sundgren, A. and Schneeweis, T., 1988. Corporate Social Responsibility and Firm Financial Performance. The Academy of Management Journal, 31(4), 854-872.

Nelling, E. and Webb, E., 2009. Corporate social responsibility and financial performance: the “virtuous circle” revisited. Review of Quantitative Finance and Accounting,

32(2), 197-209.

(28)

Pillarisettie, J. and Van den Bergh, J., 2010. Sustainable nations: what do aggregate indexes tell us? Environment, Development and Sustainability, 12, 49-62.

Ribando, J. and Bonne, G., 2010. A new quality factor: Finding alpha with ASSET4 ESG data. Starmine Research Note, Thomson Reuters.

Ruf, B., Muralidhar, K., Borwn, R. Janney, J. and Paul, K.,2001. An Empirical Investigation of the Relationship between Change in Corporate SocialPerformance and Financial Performance: A Stakeholder Theory Perspective.Journal of Business Ethics,

32(2),143-156.

Scholtens, B., 2008. A note on the interaction between corporate social responsibility and financial performance. Ecological Economics, 68, 46-55.

Scholtens, B. and Dam, L., 2007. Banking on the Equator. Are banks that Adopted the Equator Principles Different from Non-Adopters? World Development, 35(8), 1307-1328.

Shiller, R., 2013. Capitalism and Financial Innovation. Financial Analysts Journal,

69(1).

Waddock, S. and Graves, S., 1997. The corporate social performance-financial performance link. Strategic management journal, 18(4), 303-319.

Weber, O., Koellner, T., Habbegger, D., Steffensen, H. and Ohnemus, P., 2005. The relation between the GRI indicators and the financial performance of firms. Progress in

Industrial Ecology, an International Journal, 5(3), 236-254.

(29)

Appendix

Appendix 1. Environmental variables

Table 1.1: Translation of the environmental variables as used in the Principal Component Analysis to official ASSET4 codes.

Code in Table 1 Codes in ASSET4 ASSET4 name ENVvariable1 ENERDP073 ISO 14000 ENVvariable2 ENERDP074 EMAS Certified ENVvariable3 CENERDP123 Estimated CO2

ENVvariable4 ENERO02V Value - Emission Reduction/Biodiversity Controversies ENVvariable5 ENERO20V Value - Emission Reduction/Spills and Pollution Controversies ENVvariable6 CENERO23V Value - Emission Reduction/Environmental Compliance ENVvariable7 ENPIDP034 Environmental Assets Under Management

ENVvariable8 ENPIDP036 Equator Principles ENVvariable9 ENPIDP040 Nuclear

ENVvariable10 ENPIDP044 Labelled Wood ENVvariable11 ENPIDP052 Agrochemical Products ENVvariable12 ENPIDP053 Agrochemical Revenues (5%+) ENVvariable13 ENPIDP057 Animal Testing

ENVvariable14 ENPIDP058 Animal Testing Cosmetics ENVvariable15 ENPIDP062 Environmental Labels

ENVvariable16 ENPIDP066 Renewable/Clean Energy Products ENVvariable17 ENPIDP067 Water Technologies

ENVvariable18 ENPIDP068 Sustainable Building Products

ENVvariable19 CENPIO03V Value - Product Innovation/Environmental R&D Expenditures ENVvariable20 ENPIO07V Value - Product Innovation/Renewable/Clean Energy Products ENVvariable21 ENPIO21V Value - Product Innovation/Product Impact Controversies ENVvariable22 ENRRDP046 Renewable Energy Use

ENVvariable23 ENRRO13V Value - Resource Reduction/Environmental Resource Impact Controversies

(30)

Appendix 2. Overview of regressions and Granger Causality tests

Table 2.1: Overview of the relationship from environmental measurements to financial performance of U.S. firms, 2007-2013

ENVscore

Renewable Energy

Products Product Impact Labels Emission OLS regression (fixed effects) 0.0016*** OLS regression(-1) 0.0007*** (fixed effects) 0.0024*** OLS regression(-2) 0.0181** 0.0670*** (fixed effects) 0.0036*** 0.0670*** 0.0604*** OLS regression(-3) (fixed effects) -0.0027*** GRANGER(1) 3.0268* GRANGER(3) 3.1940** 2.1191* GRANGER(5) 2.0199*

This table provides an overview of the analysis conducted in this research on the relationship running from environmental measurements to financial performance. As can be seen, the analysis discussed in the Data & Methodology section are provided, in which OLS regression refers to the regression analysis with zero up to three lags and (fixed effects) referring to the fixed effects test associated with the OLS regression. GRANGER refers to the Granger Causality test for one up to five lags. For the OLS regressions the coefficients are presented. For the Granger Causality test the F-statistic is presented, as the Granger Causality test does not say anything about the sign. *,**, and *** refer respectively to the significance levels 10%, 5%, and 1%.

Table 2.2: Overview of the relationship from financial performance to environmental measurements of U.S. firms, 2007-2013

ENVscore

Renewable Energy

Products Product Impact Labels Emission

OLS regression -0,0099* (fixed effects) 1.1280*** OLS regression(-1) -1.4627* 0,0517* -0,0818*** 0,0754*** (fixed effects) -0,0762*** OLS regression(-2) 0,0753** (fixed effects) 0,0339** OLS regression(-3) -0,0904*** 0,0626* (fixed effects) -0,0295** 0,0628** GRANGER(1) 4.6560** 2.9388* 6.1308** GRANGER(3) 3.7149** 4.5300*** GRANGER(5) 3.2976***

(31)

Appendix 3. Regression of Returns and ESG score

Table 3.1: The relationship between ESG score and financial performance of U.S. firms, 2007-2013

Return ESGscore

(fixed effects) (fixed effects)

β β β β Constant -0.0116 -0.3516*** 49.3972*** 53.6652*** (0.3495) (0.0000) (0.0000) (0.0000) ESGscore 0.0007*** 0.0050*** (0.0013) (0.0000) Return 2.4965*** 2.6519*** (0.0013) (0.0000) Leverage 0.0000 0.0000 0.0001 0.0001 (1.0000) (0.8189) (0.6072) (0.2252) Size 0.0000*** 0.0000*** 0.0003*** 0.0000 (0.0001) (0.0000) (0.0000) (0.3525) Constant -0.0091 0.0233 49.7497*** 53.8075*** (0.5080) (0.5516) (0.0000) (0.0000) ESGscore(-1) 0.0010*** 0.0030*** (0.0000) (0.0000) Return(-1) -0.2807 -0.3254 (0.7239) (0.2916) Leverage(-1) 0.0000 0.0000 0.0000 0.0000 (0.5435) (0.8967) (0.8541) (0.7875) Size(-1) 0.0000*** 0.0000*** 0.0003*** 0.0000 (0.0000) (0.0000) (0.0000) (0.8803) Constant 0.0333** 0.0105 50.3837*** 54.4059*** (0.0290) (0.8172) (0.0000) (0.0000) ESGscore(-2) 0.0003 0.0018** (0.3385) (0.0306) Return(-2) -1.1510 -0.8405*** (0.1601) (0.0044) Leverage(-2) 0.0000 0.0000 0.0001 0.0000 (0.2637) (0.5472) (0.7134) (0.6886) Size(-2) 0.0000*** 0.0000*** 0.0003*** 0.0000 (0.0019) (0.0000) (0.0000) (0.6555) Constant 0.1320*** 0.2381*** 51.3612*** 55.1583*** (0.0000) (0.0000) (0.0000) (0.0000) ESGscore(-3) 0.0000 -0.0029*** (0.9849) (0.0000) Return(-3) 1.0236 1.3623*** (0.2319) (0.0000) Leverage(-3) 0.0000 0.0000 0.0003 0.0001 (0.7213) (0.3416) (0.3418) (0.3109) Size(-3) 0.0000 0.0000*** 0.0003*** 0.0000 (0.5431) (0.0000) (0 .0000) (0.5552)

(32)

the significance levels 10%, 5%, and 1%. In the Data & Methodology section the variables are discussed in more detail.

Table 3.2: Output Granger Causality Tests

Lags: 1

Null Hypothesis: Obs F-Statistic

ESGscore does not Granger Cause Return 4,984 11.4402*** (0.0007) Return does not Granger Cause ESGscore 29.5424*** (0.0000)

Lags: 3

ESGscore does not Granger Cause Return 3,088 11.2232*** (0.0000) Return does not Granger Cause ESGscore 16.8457*** (0.0000)

Lags: 5

ESGscore does not Granger Cause Return 1,301 0.9355 (0.4569) Return does not Granger Cause ESGscore 9.0053*** (0.0000)

Referenties

GERELATEERDE DOCUMENTEN

Schematic representation of the fabrication of micron-scale surface chemical gradients of the alkyne- functionalized thiol-sensitive probe 14 via electrochemically promoted CuAAC on

In Section 2, we confirm that the observed decay of wave modes in the Hele-Shaw laboratory tank, filled with water but without particles, is captured reasonably well by nu-

The meta-analysis tested the relationship between corporate social responsibility performance and analyst coverage (5.1), forecast accuracy (5.2), forecast error (5.3),

In order to examine the intervening effects of exploitation efforts on the relationship between corporate social responsibility and a firm’s financial performance,

In order to test if the impact of environmental and social dimension on CFP varies across industries, a model containing all interaction effects between the dimensions and

The regression is estimated using ordinary least squares with fixed effects including the control variables size and risk (Altman Z-score when using ROA and MTB, volatility of

Lastly, it should be noted that in this paper I used the result from content analysis as the comprehensive evaluation index to measure the performance of corporate social

Table 2 reports the descriptive statistics for all the variables used in the full sample, which are the Tobin’s Q-ratio, return on assets (ROA), ES (environmental and