The Corporate Social Responsibility Rating Accounting Essay

Published: 2021-06-18 07:55:05
essay essay

Category: Accounting

Type of paper: Essay

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

Hey! We can write a custom essay for you.

All possible types of assignments. Written by academics

In this thesis it will be studied how reliable CSR rating systems are by comparing the methodologies of KLD and Sustainalytics. This chapter will explain the possibility that ratings have different outcomes for companies. The main reason that can be given for these ambiguous results is the complexity of the definition of CSR. Section 2.1 will explain the intricate concept of CSR. It is also highlighted why it is important that CSR ratings represent the actual level of CSR as closely as possible. Section 2.2 discusses the convergence and divergence of CSR ratings. Should different CSR ratings give similar outcomes for companies? The concept of convergence or differentiation in CSR outcomes of different ratings is clarified. Thereafter, several research papers that have investigated the validity, transparency, and reliability of CSR ratings are discussed in section 2.3. In the sections 2.4 and 2.5 the main characteristics of the methodologies of KLD and Sustainalytics are described. Subsequently, the similarities and differences between the methodologies are discussed in section 2.6.
2.1 Complexity of CSR definition
‘‘The foundation of CSR is the acknowledgement that companies have responsibilities to society that go beyond shareholder wealth maximization’’ (Font et al., 2012). Although CSR is a globalized concept, there is no universally accepted definition (Freeman and Hasnaoui, 2011). Also Searcy (2011) states that ‘‘corporate sustainability is fundamentally a complex problem and there are no approaches that universally apply’’. Hence, ‘‘the assessment of a company’s CSR caliber is accordingly intricate’’ (Van den Bossche et al., 2010). Mainly for this reason Steurer et al. (2008) states that CSR ratings are ‘‘methodologically demanding judgments that are not always based on neutral criteria’’ and that they are ‘‘mostly characterized by rather subjective selection processes’’.
Based on these statements, it is not surprising that methods of different sustainability ratings can be completely different. It is useful to know how reliable CRS rating systems are since ‘‘stakeholder force companies to be transparent about their CSR practices’’ (Font et al., 2012). The rating systems are supposed to test the gap between social responsibility claims and actual practice. Ceton and Liston-Heyes (2008) show that this gap between actual and perceived CSP has an influence on the resource allocation decisions since ‘‘perceived corporate social performance ultimately dictates resource allocation decisions, firms conduct CSR investments strategically to enhance their reputation for good corporate behavior’’. They argue that this gap between actual and perceived practice of CSR is a source of market failure. Hence, it is of significant importance that CSR ratings represent the actual level of CSR as closely as possible. When a rating gives a transparent view of the sustainability performance of the company, the company can be rewarded or punished by the stakeholders (Chatterji et al., 2008).
The literature on how to measure CSR on a company level is growing rapidly. Nevertheless, there is still no generally established method which can serve as a basis for comparative studies (Gjølberg, 2009). Ceton and Liston-Heyes (2008) recognize the fact that CSR ratings measure different phenomena and cannot be used interchangeable. Any assessment of a company’s CSR performance will depend on how CSR is measured because CSR performance is a social construct and not an exact study (Font et al., 2012). Van den Bossche et al. (2010) even accept the fact that subjectivity in CSR rating is unavoidable and that it is impossible to obtain an exact CSR ranking of companies. Concluding, it is highly inconceivable that ratings are measuring the exact same phenomenon as long as there is no universally accepted definition for CSR. Therefore, research and comparative studies for CSR ratings are useful to discover the differences and similarities between these ratings. And maybe more important, it can be found out why these differences exist and what the consequences of these differences are for the CSR ratings that companies receive.
2.2 Convergence or divergence?
Should CSR ratings produce similar results for companies? What does it mean when a company is rated with different scores by CSR ratings? And what are the explanations for convergence or divergence in CSR ratings? This section will give answers to these questions.
Chatterji and Levine (2007) investigate why ratings of companies are similar or different. They use two views as reference points for their research, namely convergence or differentiation. The first view is that ratings from competing companies are strongly correlated. This correlation could be driven by ‘‘high-quality measurements of agreed-on standards or by neo-institutionalist forces pushing for imitation’’ (Chatterji and Levine, 2007). In the convergence view, a distinction is made between high validity and low validity. Convergence with high validity is the result of the use of a similar definition of social responsibility, high-quality measurements methods and data between raters. However, the authors state that it is also possible that there is high convergence without high validity. Convergence with low validity means that there is strong correlation between ratings but they are not representing the ‘actual’ level of social performance. The given reason for this is that ‘‘neo-institutionalist forces for imitation can lead raters to agree on a socially constructed measure of ‘responsibility’ even when that agreed-on measure is not valid’’ (Chatterji and Levine, 2007). Another theory that can explain convergence with low validity is herding behavior. When companies have limited information about the optimal CSR level, they can learn from observing the decisions of comparing companies. Even when the decision makers have information that advises different policies, herding behavior appears.
The second view that is discussed by Chatterji and Levine (2007) argues that there will be low correlation between the ratings. The given reasons for this low correlations are ‘‘raters attempting to inform stakeholders with different preferences or measurement errors’’ (Chatterji and Levine, 2007). The theories of differentiation are clarified. First, there can be differentiation with high validity. The authors explain that ‘‘theories of product differentiation will lead to low convergent validity, but that low convergent validity need not imply low true validity when sellers are appealing to diverse consumer preferences’’. It could be that the differences in the methodologies of the ratings are a result of different demands of the clients of the ratings. Secondly, differentiation with low validity appears when there are measurement errors. These measurement errors are caused by raters that are unclear about what determines the underlying social performance. Furthermore, it is unclear for some raters ‘‘how observable corporate behavior causally effects social outcomes’’. This means that it is difficult for the rater to construct a proxy to measure corporate behavior. In conclusion, raters can be uncertain about what subjects are relevant for measuring social responsibility and what the best way to assess these different subjects is.
The main conclusion of the research of Chatterji and Levine (2007) is that there is low correlation between major social ratings. This result is supporting the differentiation theories. When they adjust the differences between the ratings, the correlations does not increase. Chatterji and Levine (2007) make the conclusion that ‘‘the results of low convergent validity mean that all or most of the SRI ratings are not measuring ‘true’ social responsibility’’. Furthermore, they find that the chance that a company is involved in a major scandal is equal between high and low rated companies. This also points out limited validity.
2.3 Former research CSR ratings
So how reliable are CSR rating systems in reality? The KLD rating has already been investigated by many researchers. In contrast to the KLD ratings, there is no (public) research available for the ESG ratings of Sustainalytics. Some of the KLD research papers and the main findings will be described in this section.
The first paper that will be discussed is from Chatterji et al. (2008). The authors have investigated the transparency and validity of KLD’s environmental ratings. They use a sample of 588 US companies which are rated by KLD at least once during the period of 1991-2003. They make use of a single net environmental score, by subtracting the concerns of the strengths. The net environmental score is compared with the companies’ environmental performance [1] . Furthermore, they use the total sum of strengths and concerns as different indicators. The authors find that the net environmental score and the total sum of concern rating give a rather good summary of the past performance of a company. Therefore, ‘‘firms with more KLD concerns have slightly, but statistically significantly, more pollution and regulatory compliance violations in later years’’ (Chatterji et al., 2008). However, the sum of the strengths indicator is not performing that well. The authors state that the strengths are not accurately predicting the pollution levels or compliance violations. Furthermore, they find evidence that KLD is not optimally using data to rate the different indicators. For example, ‘‘there is no reason to summarize individual environmental sub scores such as emissions or regulatory problems as a one or zero indicator variable’’ (Chatterji et al., 2008).
Ceton and Liston-Heyes (2008) also contrast the ESG ratings of KLD with another CSR measurement. The authors compare the ‘actual’ CSR level to the ‘perceived’ lever of CSR. They argue that there is a gap between the actual and perceived level of CSR because proper instruments to measure corporate social responsibility are lacking. Ceton and Liston-Heyes (2008) try to investigate the causes for difference between actual and perceived CSR. To measure the actual level of CSR they use the KLD rating. They argue that this was the best instrument available for the social performance at that time. The CSR level of Fortune magazine [2] is the measurement used for the perceived level of CSR. The CSR level of Fortune magazine is based on the reputation of companies and this should be a ‘‘good reflection of underlying CSP values and behaviors’’ (Ceton and Liston-Heyes, 2008). They use different statistical analyses to analyze the datasets and the main conclusion is that there exists ‘‘a wide discrepancy between the perceived and actual CSP of S&P 500 firms’’ (Ceton and Liston-Heyes, 2008). The authors are aware of the fact that the assumption that KLD reflects actual CSR is questionable. However, they state that their results ‘‘highlight the extent to which the databases differ ’’ and ‘‘many studies do not specify at the onset whether the research is concerned with CSP reputation or with actual CSP’’ (Ceton and Liston-Heyes, 2008).
2.4 Methodology of KLD
This section will extensively discuss the methodology of the rating KLD. Kinder, Lydenberg, and Domini developed the KLD rating for CSR. KLD STATS (Statistical Tool for Analyzing Trends in Social and Environmental Performance) offers data of the environmental, social, and governance (ESG) performance of companies which are assessed by KLD Research & Analytics, Inc. (KLD Research & Analytics, Inc., 2006). It can be stated that ‘‘KLD is one of the oldest and most influential raters’’ (Chatterji and Levine, 2007). The clients of KLD are described as institutional investors and moneymakers who want to incorporate different aspects of CSR into their investment process (Chatterji and Levine, 2007). Chatterji and Levine (2007) mention that KLD compares their provided services with services of traditional financial research firms. In addition, many academic researchers have used the data of KLD for study and ‘‘scholars generally considered it the standard for measuring corporate social responsibility’’ (Chatterji and Levine, 2007). Nowadays, the KLD data provides an overview of the sustainability performance of over 3000 US companies. The dataset is published annually and this data is publicly available. In the beginning of 1991, the first dataset was generated and approximately 650 companies were included (KLD Research & Analytics, Inc., 2006). Only the S&P 500 Index® and the Domini 400℠ Social were evaluated in 1991-2000. In 2001 and 2002 KLD extended the coverage with, respectively, the Russell 1000® Index and the Large Cap Social℠ Index (KLD Research & Analytics, Inc., 2006). The Russell 2000® Index and the Broad Market Social℠ Index were introduced in 2003 and therefore the total coverage of KLD was raised to 3000 of US largest companies in 2010 (MSCI, 2011).
As mentioned before, KLD covers the three main categories for the measurement of sustainability, namely environment, social, and governance (ESG). In addition, information about controversial business issues is provided. Therefore, ‘‘the KLD ratings hold promise in several areas for CSP researchers’’ (Sharfman, 1996). Figure 2.1 illustrates the construction of the KLD rating system. The four different categories are shown and it becomes clear that the category social is divided in five subcategories. These five categories are community, diversity, employee relations, human rights, and product.
Figure 2.1 Categories of KLD
Source: KLD
In order to contain a full profile of the company’s performance, the data gathering is performed through several research processes (KLD Research & Analytics, Inc., 2006). According to numerous researchers, ‘‘KLD provides an objective, uniform and systematic assessment of the social behavior of firms’’ (Ceton and Liston-Heyes, 2008). The consistency of the evaluations increases because the firm evaluations are performed at the same time every year (Ruf et al., 1998). Furthermore, Ruf et al. (1998) expect the assessments to be consistent among the evaluators because the evaluations are based on objective rules and KLD employs a research staff for this task. Also Sharfman (1996) points out that ‘‘the data are evaluations done by individuals outside the focal firms so they are ostensibly more objective than data gathered via surveys or the content analysis of corporate documents’’.
KLD works with strengths and concerns to evaluate the different categories. The ESG categories include both positive and negative ratings (strengths and concerns) whilst the controversial business indicators only include concern ratings. Until 2010, KLD provides a total of 77 strengths and concerns for the different areas of sustainability. The indicators are mainly divided among the ESG categories. A total overview of these strength and concern indicators with the rating definitions is listed in Appendix A. In 2010 the KLD methodology underwent some changes. The construction of the rating remained the same but the amount of indicators decreased. These methodological changes will be fully discussed in section 2.1.3.
The KLD database is simply a binary overview of the positive and negative ratings. When a company meets the criteria for a strength indicator, this strength is ranked with a one. When the company does not meet the criteria, this strength is indicated with a zero. The same procedure accounts for the concern indicators. Not every category is ranked for the same quantity of strengths and/or concerns. Even within a category, the amounts of strengths and concerns evaluated can differ. For example, the category environment is assessed for six strengths and seven concerns whilst the category community is rated for seven strengths and four concerns. There are several corresponding strengths and concerns. For example, the strength and concern ‘Union Relations’, but also ‘Limited Compensation’ which is the opposite of ‘High Compensation’. The KLD dataset provides one variable with the summation of strengths and one for the concerns for every category. Remarkable is the fact that KLD does not include an overall score for every company based on their results. Hence, it is not possible to rank the different companies for their CSR performance based on the rating results. Section 2.4.3 will examine methods of how to create a composite index for the KLD ratings. Another drawback of the KLD database is the limited information provided of the companies. Besides the company name, only the company ID, CUSIP code and the ticker symbol are given.
2.4.1 Controversial ratings
The controversial variables differ from the ESG categories because they include only concern variables. In KLD Research & Analytics, Inc. (2006) is stated that these ratings are mainly used for exclusionary listing. Hence, these indicators are used for screening purposes for companies which are active in controversial business issues. The six controversial variables are alcohol, gambling, tobacco, firearms, military, and nuclear power. Examples of controversial business issues are that the company is producing a controversial product, the company owns a fraction of another company that is involved with a controversial issue, or the company derives revenues from activities related to one of the controversial business issues.
2.4.2 Methodological changes in 2010
In 2010, MSCI acquired RiskMetrics [3] and was therefore automatically absorbing KLD Research & Analytics since RiskMetrics bought KLD in 2009 (Nicholls, 2012). Nicholls (2012) states that ‘‘MSCI has been working to integrate its ESG research products into a single platform’’. For that reason, MSCI ESG Research introduced some significant changes in the KLD methodology in 2010 (MSCI, 2011). The structure of the rating systems remained the same but the scoring model was adjusted. There were introduced new indicators for each ESG category, whereas much more indicators were eliminated. The number of indicators decreased from 77 to 56. The original ratings will not be researched for a company if they are not relevant anymore (MSCI, 2011). There are a few general explanations conceivable to reduce the number of explanatory variables. One of the arguments is that this could increase the reliability of the data. It takes a lot of effort to investigate the ratings of a large amount of companies. When the number of indicators is decreased, in principle there will be more time for proper assessment of the indicators. This in turn increases the reliability of the acquired data. Furthermore, it could be possible that the 56 variables are equally capable to determine the CSR level as the old set of 77 variables. This would be the case when some of the variables of the old dataset measure different aspects of a same underlying, core variable [4] . The core variable could replace the variables which together utilize the same information.
Another explanation for the changes of the ratings could be that with the acquirement of KLD by MSCI the focus of the type of peer group/audience shifted. A different target group can demand different requirements for a CSR rating. MSCI defines its customers more clearly than KLD. It is stated that the clients include ‘‘institutional investors, asset managers, advisers, governments and government agencies, consultants, and NGOs’’ (MSCI, 2012). Furthermore, it is pointed out that the products and services of MSCI ESG Research help institutional investors and asset managers ‘‘to integrate ESG factors into their investment processes’’ (MSCI, 2012).
Appendix B provides an overview of the changed indicators for every category. The red colored indicators are removed from the dataset in 2010 and the green indicators are introduced with the methodological changes. Unfortunately, there is no publication available with the motivations of the methodological changes. Hence, only reasoning can explain the modifications.
The first notable changes are the definitions of the strengths and concerns that have not been replaced or eliminated with the methodological changes. For most of these variables, the definitions are adjusted and/or extended which made them less extensive. Furthermore, it can be noticed that four ‘Other Strength’ and four ‘Other Concern’ variables are removed from the dataset. The rating definitions of ‘Other Strength’ and ‘Other Concern’ variables are rather vague and broad. For most of these kinds of variables the definition holds ‘strengths or concerns that are not covered by the other ratings of the category’. Since it is very unclear what should be included in these ratings, this could be one reason to delete some of these variables from the dataset. Additionally, it can be noticed that some indicators have been replaced by a single indicator in 2010. For example, it could be justified that the strengths ‘CEO’ and ‘Promotion’ are replaced by the strength ‘Representation’ since the definition of this new strength could capture the descriptions of both former strengths.
Other changes could be explained by increase or decrease of relevance of certain subjects for CSR. For example, in the category ‘Human Rights’ several country-specific concerns are deleted and instead KLD introduced the concern for operating in the country Sudan. South Africa, Northern Ireland, and Mexico are not counted as sensitive countries as of 2010. Reports of frequent human rights abuses caused Sudan to become a sensitive country which is relevant for CSR. Other examples that can be given are deleted ratings such like ‘No-Layoff Policy’ and ‘Property, Plant, and Equipment’. Since 1994, the strength ‘No-Layoff Policy’ is not assigned to a company (KLD Research & Analytics, Inc., 2006). Similarly, the ‘Property, Plant, and Equipment’ strength has not been assigned to a company since 1995 (KLD Research & Analytics, Inc., 2006). This means that no company from the dataset has met the criteria for the strengths for several years. This could indicate that these variables are not relevant anymore for CSR and this could be the reason to remove the strengths from the dataset. Finally, there are indicators introduced because some subjects were not covered by the KLD data before 2010. The strength ‘Supply Chain Policies, Programs & Initiatives’ and the concern ‘Supply Chain Controversies’ are examples of uncovered issues in 2008 and 2009.
2.4.3 Constructing a composite index
Since the KLD database does not provide an overall score, a single score for every company has to be created. The idea of an overall score that aggregates the different categories of CSR is ‘‘that one number builds on the recognition that the very multidimensionality of CSR may impede an adequate comparison among companies’’ (Van den Bossche et al., 2010). Also Graafland et al. (2004) provide several advantages of benchmarking the CSR effort of companies. First of all, a composite index gives the opportunity to compare the level of CSR effort of different companies. Furthermore, the company itself can easily keep track of the internal improvement. In addition, a benchmark enhances transparency, especially when this benchmark is created by independent outsiders. The composite index also increases the accountability of the company with regard to the stakeholders since the stakeholders can confront the company with its actions. Another advantage of an overall score is that it improves the simplicity of interpreting the CSR effort of the company. In conclusion, Graafland et al. (2004) state that a benchmark offers a systematic approach to judge the CSR input, a benchmark provided by an independent outsider is more objective, and benchmarking forces companies to deliver information about their CSR effort.
There are various methodologies available for constructing a composite index since there is ‘‘lacking any clear theoretical guidance as to what unequivocally constitutes the best underlying CSR model’’ (Van den Bossche et al., 2010). Sharfman (1996) points out that there are different methods for creating a single score for the KLD data. The first method the author mentions, is simply adding all the ratings to form an overall score. This method suggests that every indicator is weighted equally. However, several researchers have shown that the categories are not equally important. This conclusion is supported by Graves and Waddock (1997) and Ruf et al. (1998). Hence, Sharfman (1996) also explains the methods of these researchers. They create the overall score by multiplying the ratings of every indicator by different weights. Graves and Waddock (1997) make use of a panel of three experts to evaluate the relative weights of every CSP attribute. They use the average normalized values of the panelists to compute the composite index for the level of CSR. Ruf et al. (1998) utilized the same process and ‘‘the developed indexes are statistically the same’’ (Graves and Waddock, 1997). Because there was not enough information available for the controversial category for the companies, these indicators were excluded and there was no weight listed for these ratings. As a final point, Graves and Waddock (1997) argue that ‘‘this weighting scheme deals with the problem of shifting of the relative importance of items in the KLD rating over time and with changing social standards’’.
2.5 Methodology of Sustainalytics
Subsequently, this section will describe the methodology of the rating Sustainalytics. Sustainalytics has developed an extensive understanding of trends and best practices in responsible investment and a process to assist institutional investors in integrating ESG considerations into their investment policies and strategies. This year, Sustainalytics is celebrating its ‘‘20th year in partnering with responsible investors’’ (Sustainalytics, 2012). The foundation of Sustainalytics began in 1992 with the creation of Jantzi Research Inc. The goal of Jantzi Research was to monitor ESG performance of publicly traded companies. In 2000, the Sustainable Investment Research International (SiRi) Group founded. This entity contained eleven ‘Socially Responsible Investment research organizations based in Europe, North America, and Australia’’ (Sustainalytics, 2012). The main members that marked the beginning of that collaboration are Jantzi Research, Triodos Research [5] , and Scoris [6] (Sustainalytics, 2012). 2008 is the year of the ‘‘dissolution of SiRi and the launch of Sustainalytics’’ (Sustainalytics, 2012). In 2009, Jantzi Research merged with Europe-based Sustainalytics which was the result of the long-standing partnership between the ratings and the company will globally operate under the name Sustainalytics (Sustainalytics, 2012). Sustainalytics works with a wide set of clients like NGO’s, corporations and investors. It is stated that they work together with ‘‘some of the world’s largest pension plans and investment managers, providing customized support to help them achieve their RI objectives’’ (Sustainalytics, 2012). Sustainalytics produces detailed information for different aspects of sustainability for over 2000 of the largest companies worldwide. However, this is private information and the dataset is not publicly available. The database consists of companies selected from different indices. The main source is the MSCI World Index and other examples are the S&P 500 Index®, the TSX Composite Index, the AEX, and the Jantzi Social Index. The companies are analyzed by local research partners. These research partners use a consistent and sound methodology for evaluating the companies. They utilize a large variety of sources, for example public reporting of a company, information from non-governmental organizations, international institutions, press, and governments. Sustainalytics also contacts experts and the company to gain information and the company profiles are continuously renewed.
The database contains detailed information about three different themes: environmental, social, and governance (ESG) sustainability. Additionally, Sustainalytics provides information about a number of controversial business issues, which are described by the products variables. Figure 2.2 shows the construction of the Sustainability rating system. It is shown from this figure that every ESG theme is divided in subcategories except the products variables. These subcategories are called topics by Sustainalytics.
Figure 2.2 Categories of Sustainalytics
Source: Sustainalytics
The topics include various indicators. The research is conducted at the indicator level. The Sustainalytics rating system consists of 160 indicators for sustainability. These indicators are divided into two different types, core variables and sector-specific variables. The core indicators are applied to every company in the database. The sector-specific indicators are only assessed for the companies for which they are relevant. Appendix C reports a list of all the indicators of Sustainalytics. The core indicators can be recognized by the three-digit code and the sector-specific indicators are indicated with a four-digit code. In total, there are 64 core indicators and 96 specific-sector indicators. Sustainalytics makes a distinction between senior listed companies and junior listed companies. The senior listed companies are the companies on the MSCI World Index. Sustainalytics makes use of a full template of indicators for these companies. The junior listed companies are mainly companies not included in the MSCI Index and there is a higher chance these companies are emerging. Therefore, junior companies are assigned a junior template of indicators, which include fewer indicators than the senior template. Specifically, the senior companies are assessed for about 65 to 80 indicators and the junior companies for 40 to 50 indicators.
The dataset of Sustainalytics provides more information about the companies besides the company name. For example, the country, the SEDOL code, the ISIN code, the ticker symbol, and various GICS codes are also listed for every company. GICS codes provide information about different kind of sectors of the company. Moreover, Sustainalytics uses its own variable, peer group, to categorize the companies. There are 42 different peer groups. Per peer group, a various amount of sector-specific indicators are relevant. This means that every peer group has its own set of sector-specific indicators. It is possible that not every peer group is assessed for the same amount of indicators. For example, for the peer group ‘‘Auto Components’’, sustainability is measured by using 67 different indicators, whilst for the ‘‘Oil, Gas, Coals and Consumable Fuels Producers’’ group 91 variables are available. The ‘products’ indicators are not sector-specific and therefore they indicate every sector.
Per indicator, a company gets a raw score with a minimum of 0 and a maximum of 100, depending on the performance of that specific indicator. Every indicator has a very explicit rating definition. The overall score per company is calculated by using weighted averages for every indicator. Nevertheless, weights can be modified by clients of Sustainalytics to meet their requirements. The overall company score is the weighted average of all raw scores of the relevant indicators. To determine the weights per indicators, Sustainalytics uses the default weight matrix. Section 2.5.1 provides an overview of the methodology of the default weight matrix and the use of weighted-averages of Sustainalytics.
Finally, Sustainalytics has included 10 indicators for controversies. Every topic, except ‘Philanthropy’, assesses the company for being involved in certain controversies or incidents. These indicators are very important and they have a relatively high weight compared to the other indicators for the overall score of the company. For clients, the main functions of the controversies are screening purposes.
2.5.1 Products ratings
In total, there are twelve products indicators. These indicators are used for screening purposes for companies which are active in controversial business issues. The products indicators include seven comparable indicators to the KLD database, namely alcohol, firearms, gambling, military contracting, nuclear power generation, nuclear power related services, and tobacco. However, Sustainalytics includes some additional indicators. The total list of products indicators can be found in table C4. The products ratings are not relevant for calculating the overall company score (section 2.5.2) and will therefore not be discussed more extensively in this research.
2.5.2 Default weight matrix
Sustainalytics uses customized weights for every indicator to calculate the total company score. These weights are uniquely defined for every peer group. The products variables are excluded from the overall company score since the weights of these variables are equal to zero in the default weight matrix. As mentioned before, every peer group is assessed for its own set of indicators. The overall ESG rating for every peer group is the weighted average of all raw scores of the relevant set of indicators.
Sustainalytics makes use of relative and absolute weights on indicator-, theme- and topic-level for every peer group for the senior companies. Since the junior and senior companies are evaluated for different sets of indicators, the weights of indicators will vary between these kinds of companies. Sustainalytics makes a distinction between three kinds of companies, senior listed companies, senior non-listed companies, and junior companies. This means that the weights can vary for these types of companies.
The total weight of the topic is simply the summation of the weights of the core and relevant sector-specific indicators for that peer group. Logically, the total weight of a theme is the summation of the weights of the topics. The total weights of the themes, topics, or indicators add up to 100%. Table 2.1 gives an impression of the relevance of the ESG themes for the total dataset of the senior companies. These ESG weights are different for every peer group, depending on the relevance of the theme for the total company score.
Table 2.1 Average weights ESG themes senior companies
Source: Sustainalytics
As mentioned before, the controversies are of relatively high importance for composing the overall company index. For the senior listed companies, every controversy indicator receives a weight of 3%. Each company is assessed for 10 controversy indicators. Hence, the total weight of these variables for determining the overall company score is 30%. Taken into account that every peer group is evaluated for an average of 78 indicators, these 10 controversy indicators have a relatively high input on the overall score.
2.6 Differences and similarities
In this section, the methodologies of KLD and Sustainalytics will be compared and the main differences and similarities are discussed. First, the most apparent similarity between the KLD and Sustainalytics rating system is the divisions of indicators. Both measurements cover the categories environmental, social, and governance sustainability. Unfortunately, there are no exact rules or guidelines for ESG reporting and therefore they are often inconsistent (Schäfer, 2005). This is also the case for KLD and Sustainalytics since the datasets utilize different kinds of information. The contents of the ESG categories are not similar and very different kinds of indicators are used to test for social responsibility.
Additionally, both KLD and Sustainalytics use screens to exclude some firms from the dataset. For KLD these variables are called ‘controversial’ ratings and Sustainalytics make use of ‘products’ ratings to detect firms which are operating in controversial business issues. There is a lot of overlap in the kind of issues which are included in these indicators, for example alcohol, gambling and firearms. For Sustainalytics, the ‘products’ ratings have no input in the overall company score.
A remarkable difference between the methodologies is that Sustainalytics does provide an overall score for every company, whereas KLD provides only an overview of values for every indicator. Sustainalytics puts relative weights to the indicators, themes, and topics to calculate the overall score. These relative weights are even adjusted for different peer groups. It can be justified that the relevance of ESG subjects varies between different sectors. KLD does not link weights to the indicators and evaluates companies for every indicator. Therefore, KLD does not judge the relevance of the different indicators. The advantage of overall company score are discussed in section 2.1.4. The main benefit of the data of Sustainalytics is the ability to compare the different companies for the level of CSR based on the benchmark. Though, there are also disadvantages attached to a given composite index. These disadvantages are also provided by Graafland et al. (2004). The authors argue that an overall score for the level of corporate social responsibility has a monistic nature; it is assumed that the different values can be put into one ranking as well as they can be compared. However, Graafland et al. (2004) state that the values of CSR have a pluralistic nature and benchmarking discounts the complexity of measuring CSR. Another drawback is that because of measurement problems, the outcome of the overall score could be rather subjective. As a final point, it is possible that benchmarking does not take into account the intentions of a company. By not linking weights to the strengths and concerns, KLD takes no risk providing subjective overall company scores. Moreover, this gives researchers or companies the freedom to determine the overall score using different methodologies.
Furthermore, the method of valuating the different aspects of sustainability is different between the ranking systems. Whereas the indicators of the KLD data can only result into zero or one, the raw outcomes of Sustainalytics’ indicators can vary between 0 and 100. Hence, the outcomes of the variables of Sustainalytics are less restricted compared to the KLD variables and can provide more information. Chatterji et al. (2008) conclude that the outcomes of the strengths and concerns of KLD could be represented much more accurate when KLD uses continuous or ranking indicators.
Next, there will be taken a closer look at the target groups of the ratings. Krajnc and Glavic (2005) state that there are three main target groups for which the requirements of a CSR rating differ. The groups are scientist, decision makers, and individuals. First, ‘‘scientist are interested primarily in statistically useable and possible not aggregated data’’ (Krajnc and Glavic, 2005). Secondly, decision makers need aggregated data and the relating data which contains information about the goals and criteria of the ratings. Finally, individuals favor aggregation of the data into an overall company score.
Based on this information, one would conclude KLD has a focus on scientist and Sustainalytics on decision makers. Since Sustainalytics does provide an overall score and KLD does not, it could be said that possible clients of Sustainalytics put a higher importance to actually ranking the companies. It can be concluded that the clients of Sustainalytics have a higher focus on selecting companies based on the position of the companies than the clients of KLD. It is more difficult to select the most ‘social responsible’ company based on the data of KLD than Sustainalytics. Furthermore, given the weight matrix of Sustainalytics and the clear criteria for the ratings, it is easy for the company or client to detect which CSR subjects are lacking attention by the company. For the ratings of KLD this could be less obvious since the criteria of the ratings are more general defined. Since KLD uses a zero/one approach, there is no distinction made in classifications of the ratings. Therefore, the information about the position of the company concerning ratings can be less specific than the information that is provided by Sustainalytics.
Furthermore, it must be noted that a lot of research has been done to the KLD ratings and datasets. Unfortunately, for Sustainalytics there are no public research papers available. As mentioned in the previous sections, KLD is known as a worldwide accepted measure for CSR for many years. It has been used by many academic researchers and scholars for study and KLD has been investigated by many researchers. The datasets of KLD are easy to obtain. On the contrary, the data of Sustainalytics is private information and cannot easily be obtained. As a result, less research has been done for the ratings of Sustainalytics. Again, based on this information, one might conclude that KLD has a focus on scientist as target groups.
However, it is difficult to make a distinction between the types of clients the ratings serve based on the publically available information. As mentioned in section 2.4 and 2.5, both ratings serve institutional investors and decision makers. Additionally, Sustainalytics also provides its services to some of the world’s largest pension plans. The pension plans are strongly connected to labor unions. One could imagine that the labor unions have a stronger focus on making a CSR statement than other institutional investors. This supports the statement of Krajnc and Glavic (2005) that decision makers prefer aggregated data and the relating data which contains information about the goals and criteria of the ratings. Finally, the scientific focus of the ratings of KLD cannot stay unnoticed. This could go back to the origin of the rating, since the founders of KLD have a scientific background.
To conclude, KLD does not provide very company-specific information in the spreadsheet. The information is limited to the company name, company ID, and ticker code. Sustainalytics gives additional variables with information regarding for example the country, sector, peer group, (sub) industry for every company. This information provides more opportunities when analyzing the data. Remarkably, KLD has a focus on evaluating companies from the US while Sustainalytics assesses companies from over 25 different countries. It must be taken into account that the focus of KLD on US companies leads to different accents in the indicators that are included since the concept of CSR in the US may differ from that in other countries.

Warning! This essay is not original. Get 100% unique essay within 45 seconds!


We can write your paper just for 11.99$

i want to copy...

This essay has been submitted by a student and contain not unique content

People also read