THE VALUE RELEVANCE OF SUSTAINABILITY ASSURANCE WHICH ASPECTS MATTER TO INVESTORS?
MSc Thesis, MSc Accountancy
University of Groningen, Faculty of Economics and Business
Joshua Pallencaöe Student number: S2725797
De Heurne 55-23 7511 GZ Enschede Tel.: +31 6 13573820
E-mail: j.j.pallencaoe@student.rug.nl Supervisor
Marcus Looijenga MSc EMA RA Second assessor
Prof. dr. D.A. de Waard Third assessor dr. R.B.H. Hooghiemstra
Word Count: 10.935
1 | P a g e ABSTRACT
This research examines the association between sustainability assurance and the market value, hereby measuring the value relevance. Investors have a significant influence on the stock price and their decisions should reflect on stock price movement. However, which information with regard to sustainability assurance is relevant in the investment decision making process is mainly unexplored. This study examined a sample of 389 listed UK and US firms which publicized sustainability information in 2016. The variables of interest are the aspects of assurance which are more visible to investors, i.e. the parts contained in the assurance report. These variables are the quality of the assurance report, the provider of the assurance report and the reported scope and level of assurance. The rationale behind this study leans on the agency theory and the concept of information asymmetry. Implying that lower information asymmetry leads to an increase in the value relevance. Regression analyses based on the model of Ohlson (1995), found a significant positive relation between sustainability assurance in general and the stock price. Further, there is a significant positive relation found between the quality of the report and the stock price. However, with regard to the other three variables, results did not turn out as expected. Based on these results, it could be inferred that investors incorporate sustainability assurance in their decision making process. However, not all aspects are considered with the same amount of attention.
.
2 | P a g e
Table of Contents
1. INTRODUCTION ... 3
2. THEORETICAL BACKGROUND ... 7
2.1 The value relevance ... 7
2.2 Quality ... 9
2.3 Scope, provider, level ... 10
2.3.1 Provider. ... 11
2.3.2 Scope/Level. ... 11
3. METHODOLOGIES ... 13
3.1 Sample ... 13
3.2 Dependent variable: Perceived value relevance to investors ... 14
3.3 Independent variable: Quality ... 15
3.4 Independent variable: scope, provider, level of assurance... 16
3.4.1 Provider ... 16
3.4.2 Scope ... 16
3.4.3 Level of assurance ... 17
3.5 Control Variable ... 17
4. RESULTS... 19
4.1 Descriptive statistics ... 19
4.2 Correlation ... 19
4.3 Regression analysis ... 20
4.4 Additional analysis... 22
4.4.1 Quality Scores. ... 22
4.4.2 Stock price date... 22
5. CONCLUSION & DISCUSSION ... 23
5.1 Interpreting the results... 23
5.2 Limitations ... 26
5.3 Recommendations for further research ... 26
References ... 27
Appendix ... 30
3 | P a g e 1. INTRODUCTION
World’s biggest investment and asset management corporation, Blackrock, is paying more attention to corporate strategies and practices. Blackrock’s CEO, Larry Fink, calls out companies to contribute to society in addition to attaining financial results.
This event is characteristic for the current uprising of providing sustainability information (Zorio et al., 2013). Increasingly more companies are publishing sustainability information. Issuing sustainability information is now a mainstream practice (Zhou et al., 2017). According to KPMG (2017), 77% and 88% of the examined companies in respectively Europe and North-America provided
sustainability information in 2017. Sustainability information is provided to satisfy stakeholders’ demands and to maintain a positive corporate image (Odriozola &
Baraibar-Diez, 2017; Brown & Kohlbeck, 2017; Zhou et al., 2017).
To give an impression why it is important to satisfy stakeholders’ demands and provide information regarding sustainable corporate practices, take a look at the recent Volkswagen scandal. After the so called ‘’diesel gate’’ in 2015, the well-known car manufacturers were recently causing a public outrage again. It became known that the Volkswagen group funded ‘secret’ research on toxic exhaust fumes. The problem with this research was that it used monkeys and living human beings as test subjects. After the news became public, stock prices plummeted with almost 10%.
Thus, satisfying stakeholders’, and in this case especially investors’, demands are meaningful to a company. However, it is no longer sufficient to merely provide sustainability information (Cuadrado-Ballesteros et al. 2017). According to Rao
(2017), comes the need for sustainability assurance along with the increasing interest in sustainability information. Third-party assurance is a way to increase the credibility and transparency of the sustainability information (Gurturk & Hahn, 2016; Zorio et al.
2013). The increase in interest for sustainability assurance, is also noted by the earlier mentioned report from KPMG. Their research found that more companies are investing in third-party assurance on their sustainability information. This amount has been growing steadily since 2005 (KPMG, 2017).
Sustainability assurance is structured to meet the demands of the market (Farooq &
de Villiers, 2017). However, what does the market demand? This research will shed light on this question, using the point of view of investors, since stock prices are based on their market choices.
What investors demand and perceive as meaningful is partly examined in studies
concerning the value relevance of sustainability assurance. In general, accounting
information is deemed value relevant if it has the ability to influence the decision-
making process of the investors and subsequently influences equity stock prices
(Purswani & Anuradha, 2017). In prior studies this is only examined to a superficial
extent, since research mainly focused on assurance as a whole (Peters & Romi,
2015; Kuzey & Uyar, 2017; Fazzini & Dal Maso, 2016). It is not clarified which
specific aspects of assurance and the accompanying report are valued by the
investors. Therefore, this research will examine which specific information investors
demand to be presented and perceive as value relevant.
4 | P a g e The need to examine investors’ demands, with regard to sustainability assurance, also comes forth out of prior studies. Recent studies found some interesting developments concerning current practices on sustainability assurance. It is
suggested that the freedom and voluntary nature which encompasses sustainability practices may diminish the transparency and credibility derived from the assurance report, which is accompanied with the sustainability information (Gurturk & Hahn, 2016). In addition to this, stakeholder’s expectations with regard to sustainability assurance seem to be increasing (O’dwyer, 2011). The possible ineffectiveness of current practices combined with the increasing expectations of stakeholders could incite a need for change. The question then again is, which direction should this change go? To answer this question and guide this possible change, it is necessary to examine which information of sustainability assurance is perceived as value relevant by investors.
Investors do not have access to all information of the firm, therefore this research focuses on the more visible parts of assurance. Namely, the quality, provider, scope and level of assurance. These aspects also arise out of prior studies. Nowadays, providing sustainability information and assurance hereon are mostly voluntary. This freedom causes differences in assurance reports across content, scope and
providers (Gurturk & Hahn, 2016). Further, sustainability assurance practices are still evolving and often more immature than financial assurance practices. This causes a variety in the quality between assurance reports (Perego & Kolk 2012). Hence, this research will examine the quality, provider, scope and level of the assurance report in relation to the perceived value relevance to investors. Current research regarding these aspects is scarce and often contrary. Also, current research is more concerned with the determinants of sustainability assurance and do not take into account what investors actually desire.
Take for instance the provider of the assurance report. Assurance can be provided by an audit firm or a consultancy firm (including boutique firms). From an academic point of view, it is unclear which of the two is preferred by investors. Results are contrary, some studies claim it does not matter who assures the sustainability information (Simnett et al, 2009; Perego & Kolk, 2012). Others, however, claim that sustainability assurance provided by an audit firm is related to several positive effects (Zorio et al., 2013; Cuadrado-Ballesteros et al., 2017; Martinez-Ferrero & Garcia- Sanchez, 2017). The findings of this study could give new insights in this discussion, since it takes the viewpoint of the investors and thereby demonstrates what is
preferred by the users of the report.
With regard to the level and scope of the assurance report literature is scarce. Only a few studies examine these aspects and the results are not conclusive (Cuadrado- Ballesteros et al. 2017). Further research regarding these aspects is therefore recommended (Cuadrado-Ballesteros et al. 2017).
Concerning the quality of the assurance report, quality itself has been examined by multiple researchers (Perego & Kolk, 2012; Gurturk & Hahn, 2016; Zorio et al., 2013).
These studies assume that a higher quality is more appreciated by the users of the
report. This seems common sense, however, is this actually the case? Or are the
users of the report indifferent towards the quality of the assurance report? Current
literature does not explore this. Therefore, this research examines the relation
5 | P a g e between the users, in this case the investors, and the quality of the assurance report.
Quality is examined here in a quantitative fashion through a content analysis. This analysis assigns a score to an assurance report. This is in line with prior studies, which used a similar method and suggested that further quantitative research is needed with respect to the quality (Gurturk & Hahn, 2016).
All these aspects are believed to have an impact on the perceived transparency and credibility and thus on the value relevance (Gurturk & Hahn, 2016; Junior et al., 2013;
Hodge et al., 2009; Cuadrado-Ballesteros et al., 2017). Therefore, the main goal of this study is to discover what specific information, in the assurance report, investors use in their decision-making process and thus perceive as value relevant and meaningful.
So, the main focus of this research will be the value relevance of sustainability
assurance to investors. The value relevance is a fairly abstract concept, so firm value will be used as a proxy to measure this. The rationale behind this can be deduced from the theory surrounding the value relevance of accounting information. As mentioned above, accounting information is considered value relevant if it has the ability to influence the decision making process of the investor (Purswani &
Anuradha, 2017). Subsequently, investors’ decisions should be reflected in the stock price movements and thus influence the firm value. This theory is supported by literature, which suggests that an increase in the perceived credibility and relevance of the assurance object, could lead to a greater willingness to invest in a company (Beisland, 2009; Cheng et al., 2015). This approach is also in line with prior studies which used firm value to measure the value relevance of sustainability assurance (Peters & Romi, 2015; Kuzey & Uyar, 2017; Fazzini & Dal Maso, 2016).
This research shall combine the above discussed subjects and examine which aspects are valued and perceived as value relevant by the investors. Which leads to the following main research question this study intends to answer:
To what extent do investors perceive the quality, provider, scope and assurance level of the assurance report on sustainability information as value relevant?
The findings of this study could give guidance to the development of auditing practices, standards and regulations. This will enable auditors to deliver assurance reports which are more valued by the users. This is also in line with Scott (2015), who suggests that empirical research could assist auditors in determining which information is valued and useful to investors.
In addition to the usefulness to regulatory bodies, as mentioned above, the findings
of this research could also prove to be useful to the companies self. Since, this
research demonstrates which aspects of sustainability assurance are valued by
investors. The findings could, for example, assist companies in making cost-benefit
decisions. For instance, the choice between reasonable and limited assurance. If, for
instance, investors’ perception proves to be indifferent towards certain aspects,
companies could decide to do the least expensive option. So, this research could
enable companies to tailor the assurance report to the investors’ wishes, which could
further increase investors’ satisfaction.
6 | P a g e From a scientific point of view, this research could provide new insights in unexplored areas of sustainability assurance research. This research has therefore a so called explorative character. Since current literature regarding these topics is not very
extensive and only a few exceptional studies make the connection with the investors.
The remainder of this research is structured as follows, chapter 2 will start to
consider the theoretical background and describe prior findings, after which the
hypotheses are developed. Subsequently, chapter 3 will elaborate the used research
methodologies and give some insight in the data collection. Hereafter, chapter 4 will
present the results of the various statistical tests and lastly chapter 5 will conclude
with an interpretation and discussion of these results. Finally, a number of limitations
and suggestions for further research will be provided.
7 | P a g e 2. THEORETICAL BACKGROUND
This section provides the theoretical structure of this research and discusses prior findings regarding the variables mentioned in the introduction. Subsequently, the hypotheses will be developed on the basis of this discussion. But first, some general theory and information regarding sustainability reporting and assurance will be provided to give an impression of the field of work.
Nowadays, a broader report is desired by multiple stakeholders. Companies are more often providing sustainability information, in addition to the usual financial information. According to KPMG (2017), sustainability information is ‘the new financial information’. Disclosure of sustainability information is currently only required for large public interest entities in Europe, through the EU-Directive EU/2014/95 on non-financial information. However, the requirements are quite general and can be broadly interpreted. Further, KPMG (2017) pointed out that more reporting regulation is to be expected surrounding sustainability reporting. It is
expected that voluntary guidelines are transitioning into mandatory reporting requirements (KPMG, 2017). Sustainability information can be disclosed in a
separate sustainability report or an integrated report, which combines sustainability information with financial information.
It is currently not required to assure sustainability information. Guidelines and standards exist for sustainability assurance engagements, but these are not as mature as their financial counterparts. Furthermore, it is not required to have
assurance carried out by an audit firm. It also happens that assurance is carried out by a consultancy firm in wide varieties.
Since both the provision of sustainability information and assurance hereon are mostly voluntary, the big question arises why companies engage in this exercise.
According to legitimacy theory, one of the possible motivations is the desire to
legitimize an organization’s operations (Deegan, 2002). However, as An et al. (2011) noted in their integrated framework, motivations could be diverse. This research considers the agency theory in more depth since this theory provides, as will be discussed below, a good basis for explaining the effect of value relevance from an investor perspective.
2.1 The value relevance
According to the theory concerning value relevance, if accounting information influences the decision-making process of an investor it is said to be value relevant (Scott, 2015). Investors examine both financial and non-financial information before deciding to invest in a company. If this information is perceived as value relevant, then it influences the investor’s perception of company value and thus influences their investment decision (Beisland, 2009). In other words, it demonstrates whether accounting information is utilized by investors in valuing the company (Purswani &
Anuradha, 2017). These investment decisions are subsequently reflected in stock prices. Hence, information is considered value relevant if it coheres with equity market value (Purswani & Anuradha, 2017).
Normative theory concerning the value relevance of accounting information suggests
that increased value relevance could lead to a lower cost of capital, which in turn
8 | P a g e could lead to increasing levels of investment (Beisland, 2009). This implies that, since the willingness to invest could increase if the value relevance of sustainability assurance also increases, one could deduce that investors base their decisions, among others, on the assurance report. Hence, it is important to provide investors with information which they consider to be value relevant, because this could influence the value of the firm.
The first significant evidence of the value relevance of accounting information was found by Ball & Brown (1968). A recent publication of Cuadrado-Ballesteros et al.
(2017), suggests that external assurance of sustainability information is a tool to ensure value relevance of the voluntary sustainability disclosures. This is supported by other studies. However, these studies are mostly conducted in differing corporate markets. Current literature, which is specifically focused on sustainability assurance in a western oriented market, is somewhat limited. Studies which examine the value relevance of sustainability assurance are discussed below.
The majority of these studies concur with each other and in general it was found that the value relevance of sustainability assurance is increased over the last few years (Peters & Romi, 2015; Kuzey & Uyar, 2017; Lee et al., 2016). This implies that, increasingly, the users of the sustainability information perceive the accompanied assurance report as useful and influences their decision-making process. Lee et al.
(2016), claim that an increase in value relevance is associated with sustainability assurance. In addition to this it was found that, market value of firms that implement environmental audits is 9 percent higher than those that do not. Their study was executed in Japan, therefore, the results of this study could provide corroborating evidence that these findings hold in a different corporate market. The same applies to the study of Kuzey & Uyar (2017), their study was performed in Turkey. However, to the contrary, Fazzini & Dal Maso (2016) did not find such a relation. This study was conducted in Italy and they claim that the market perception of assurance first needs to be developed further. This is supported by Junior et al. (2014), who claim that users of the sustainability report are often unaware of the different scopes and methodologies of the assurance providers.
The reason why sustainability assurance could be perceived as value relevant by investors, is deducible from the agency theory. The agency theory is one of the most well-known theories in this field of study. Jensen & Meckling (1976), define the agency relation as follows: ‘’a contract under which one or more persons (the principal(s) engage another person (the agent) to perform some service on their behalf which involves delegating some decisions making authority to the agent’’. The main problem in this relation is the fact that the agent may not always act in the best interest of the principal. To reduce management’s discretion and ensure that the goals of the principal and agent are aligned, the principal can impose certain controls. Sustainability assurance is one of the options to achieve this goal.
Providing merely sustainability information is not sufficient to decrease information asymmetry (Cuadrado-Ballesteros et al., 2017). More often is sustainability
assurance demanded by stakeholders. Prior studies found that sustainability
assurance is a way to reduce information asymmetry between the management
(principal) and investors (agent), which could subsequently lead to a decrease in
agency costs (Simnett et al., 2009; Kolk, 2008; Martinez-Ferrero & Garcia-Sanchez,
9 | P a g e 2017). Signalling theory hints at this too. Companies could signal their excellence to stakeholders in order to deal with information asymmetry problems (An et al., 2011).
The link between the value relevance theory and agency theory is discussed by various researches. As mentioned before, sustainability assurance adds credibility and transparency. Pervan & Bartulovic (2012) noted that an increase in
transparency results in an increase in the value relevance of information. This is due to the fact that an increase in transparency limits management’s discretion and reduces the information asymmetry. So, a reduction in information asymmetry leads to an increase in the value relevance. Rapp (2010), found similar results, claiming that information content will become more relevant if the information asymmetry decreases.
Cuadrado-Ballesteros et al. (2017) suggest something similar. Their research
suggests that sustainability assurance increases the credibility of the information and thus increases the value relevance. The rationale behind this claim, which is provided by Cuadrado-Ballesteros et al. (2017), is deducible from evidence found concerning financial reporting audits. Assurance delivered on accounting information reduces the dispersion of analysts’ forecasts concerning future earnings. This implies that
information becomes more accurate, which subsequently mitigates conflicts arousing from information asymmetry and thereby increasing its value relevance. This
rationale leans on the concepts of agency theory, pointing out that decreasing
information asymmetry is essential for increasing the value relevance of information.
The assumptions behind this rationale are supported by Casey & Grenier (2015), who suggest that sustainability assurance is associated with lower analyst forecast errors and dispersion.
Further, information asymmetry between investors and the management adversely impacts firm value (Fosu et al., 2016). So, if sustainability assurance is performed in such a manner that reduces information asymmetry between principal and agent, firm value could increase. This further supports the assumptions underlying the link between agency theory and theory surrounding value relevance.
Hence, taking the above discussed studies and accompanied findings into account, we arrive at the first hypothesis. This research expects that investors perceive the assurance report as value relevant. This is due to a reduction of information asymmetry and an increase in the credibility and transparency of the concerning sustainability information. The market value acts here as a proxy for the perceived value relevance. This leads to the following hypothesis:
H1: Investors perceive sustainability assurance as value relevant.
2.2 Quality
This research will provide insight in the value relevance of the quality of the
assurance report. Quality is a broad and often vague concept, this research intends to operationalize this concept by allocating a score to the quality. Thus, examining quality in a quantitative fashion. This score emerges out of a content analysis, more about this can be found in the methodologies section. The quality of
sustainability assurance is researched by multiple researchers (Perego & Kolk, 2012;
Junior et al., 2014; Gurturk & Hahn, 2016; Zorio et al., 2013). Zorio et al. (2013)
10 | P a g e reported that assurance reports are of a fairly acceptable quality, but quality seems to be differing significantly. This concern was also noted by Gurturk & Hahn (2016), who accounted this to a lack of standards and regulations.
The current situation is that the most common framework regarding sustainability reporting is the GRI framework (KPMG, 2017). However, not all organizations use some kind of guidance or framework for their reporting. In addition to the
sustainability information itself, the assurance on this information is also voluntary.
Which causes differences between the assurance reports as different assurors use different standards, if any are used at all (Perego & Kolk, 2012; Gurturk & Hahn, 2016; Zorio et al., 2013). As noted by Gurturk & Hahn (2016), is this voluntary nature considered as a cause for the diminishment of the transparency and credibility, which is normally derived from an assurance report. Perego & Kolk (2012), add to this that there is a transition of professional rules and procedures from financial into non- financial auditing. This may not be optimal since non-financial auditing may require a different method.
The extent to which users of the sustainability information value the quality of the assurance report is not thoroughly examined in existing literature. Perego & Kolk (2012) examined how the quality of assurance reports is shaped but did not link this with the users of the sustainability information. Junior et al. (2014), recommended that further research should be conducted regarding the quality of the assurance reports. Gurturk & Hahn (2016), attempted to examine the quality of assurance reports in a qualitative manner, and suggested that further quantitative research could reinforce the conclusions made. None of these studies examined the significance of the quality to the users of the sustainability information, which
ultimately use the assurance report. Perego & Kolk (2012) also noted that the quality of sustainability assurance is not improving much over the last years. It is stated that this is worrying from a stakeholder perspective since there is room for improvement regarding the quality. So, in their study it is assumed that stakeholders value the quality of sustainability assurance, but is this actually the case? This study aims to answer that question and therefore, this research shall focus on the relation between the quality of the assurance report and the users of the sustainability information.
It is presumed that a higher quality leads to a more value relevant and meaningful assurance report to investors. Therefore, the second hypothesis that will be tested is:
H2: A higher quality of the assurance report has a positive impact on the perceived value relevance to investors.
2.3 Scope, provider, level
Besides the quality of the content, this research shall focus on three other aspects of the assurance report itself. The aspects which are considered are the scope, provider and level of assurance. According to Junior et al. (2014) a detailed explanation of these aspects in the assurance report could be used to achieve a higher level of transparency. In the following part will these aspects be discussed. The provider of assurance will be discussed separately, and the scope and level of assurance will be discussed simultaneously since these topics intertwine.
11 | P a g e 2.3.1 Provider. Currently is the provider of sustainability assurance a
voluntary choice. Companies can choose between audit firms or consultancy firms, however, most companies seem to choose a Big four audit firm (KPMG, 2008; Zorio et al. 2013). Prior literature surrounding the assurance provider is not conclusive, there seems to be a lot of discussion. According to Zorio et al. (2013), an audit firm as provider positively affects the quality of the assurance report. However, Perego &
Kolk (2012), state that the quality delivered by an audit firm is not different from a non-audit-firm. Simnett et al. (2009) claim that it does not matter whether or not the assurance is provided by an audit firm or not. Cuadrado-Ballesteros et al. (2017) observed a greater reduction in information asymmetry when the assurance was provided by audit firms. Martinez-Ferrero & Garcia-Sanchez (2017) claim, that assurance provided by audit firms are related to a decrease in cost of capital. Casey
& Grenier (2015), support this claim and suggest that cost of capital reduction is significantly higher when assurance is provided by an audit firm, relative to non-audit firms. Casey & Grenier (2015), also suggest that the earlier mentioned reduction in analyst forecast dispersion caused by assurance is amplified when an audit firm provides the assurance. Pflugrath et al. (2011), found that the credibility increases when the provider is a member of the professional accounting profession. Hodge et al. (2009) however, suggest that sustainability specialists provide a more
comprehensive and fair report. Hodge et al. (2009), also found that users place more confidence in the report if the assurance is provided by a top-tier audit firm and of a reasonable level. Considering all the above-mentioned findings, it can be said that opinions differ regarding this subject. Though, evidence leans somewhat more towards audit firms as assurors.
This research will examine the situation from the users’ point of view. That is, it will be examined if the users of the sustainability report perceive the provider choice as relevant. This is not done in existing literature and therefore this could give new insights in the discussion surrounding the provider of sustainability assurance.
In this research it is expected that the choice for an audit firm will positively impact the perceived relevance to investors, since audit firms provide a greater credibility compared to non-audit firms (Cuadrado-Ballesteros et al., 2017). So, the third hypothesis that will be tested is:
H3: An audit firm as assurance provider has a positive impact on the perceived value relevance to investors.
2.3.2 Scope/Level. The scope and level of the assurance are both important aspects, since it is important for the readers to understand exactly how much
information in the sustainability report is being verified by the assurance provider (Junior et al. 2014). Literature concerning these topics is very scarce, so it is also interesting from an academic point of view.
Existing research regarding the scope and level of assurance is not very extensive.
Hodge et al. (2009) and Cuadrado-Ballesteros et al. (2017) consider only the level of assurance and Braam & Peeters (2018) consider both the scope and level of
assurance. Braam & Peeters (2018) link the scope and level of assurance to the
sustainability performance. Their study suggests that firms with a good sustainability
choose a broader scope, relative to firms with inferior performance. In the contrary,
12 | P a g e there is no significant relation observed between the level of assurance and the sustainability performance. According to Braam & Peeters (2018), the low number of companies that pay for reasonable assurance is a possible explanation. So, their study focused on the determinants of the scope and level, it is not examined which variant is actually preferred by users of the report. This low number of firms assuring their information with a reasonable level is also found by a prior study conducted in the Netherlands (De Waard & Kamp-Roelands, 2008). This latter study recommends that companies should pay more attention to assuring sustainability information to a reasonable extent.
As mentioned earlier, Hodge et al. (2009) found that users place more confidence in the report if the assurance is provided by a top-tier audit firm and of a reasonable level. These factors do not influence the users’ perception individually but only combined. With limited assurance for either type of provider is such a relation not found. This implies that, a higher level of assurance is only perceived as meaningful if the assuror is of an audit firm.
Cuadrado-Ballesteros et al. (2017), found that there is a slight effect of the level of assurance on information asymmetry. The reduction in information asymmetry is greater when the assurance output shows a reasonable level, which results in a higher credibility and lower informational risk. This implies that the users of the report perceive a higher level of assurance as more value relevant. However, it should be noted that Cuadrado-Ballesteros et al. (2017) state that these results are not
conclusive since sensitivity analysis do not hold up, so further research is suggested.
Since the limited amount of available literature, this research could provide new insights regarding the various aspects of the assurance report. Further, this research could follow up on the study of Cuadrado-Ballesteros et al. (2017) and provide
corroborating evidence.
This research expects a positive impact of a higher level of assurance on the
perceived value relevance to investors. Since, this decreases information asymmetry and increases credibility (Cuadrado-Ballesteros et al., 2017). Concerning the scope of the assurance report, no literature is found which hints at the nature and direction of the possible relation. A broader scope should provide more credibility and
transparency since a greater part of the sustainability information is assured. Hence, a broader scope is associated with higher costs. So, one could assume that a
broader scope is more favourable. Therefore, this research expects that a broader scope is perceived more value relevant by investors in comparison to a narrower scope. If it is found that investors seem indifferent towards the scope, then investing in a broader scope seems like an useless exercise. These expectations lead to the following two hypotheses:
H4: A broader audit scope has a positive impact on the perceived value relevance to investors.
H5: A reasonable/high level of assurance has a positive impact on the perceived
value relevance to investors.
13 | P a g e 3. METHODOLOGIES
This research will aim to establish whether there is a significant relation between the various concepts and variables. This will be done with the assistance of several statistical methods. In the remainder of this section will the sample, the variables and the various methods be discussed.
3.1 Sample
The sample is based on all the listed UK and US companies available in the GRI database. This ensures the availability of data. The reports, publicized in 2016, are used, since this leads to a larger sample. This is due to the fact that not all reports, publicized in 2017, are available in the database.
In the GRI database, information which complies with GRI guidelines and information which adheres to other guidelines, is included. The GRI guidelines are a multi-
stakeholder initiative and the most used guidelines currently available (KPMG, 2017).
Further, the GRI guidelines have been an important driver in improving the quality of sustainability reports (Perego & Kolk, 2012). Hence, using a sample, which is partly based on these guidelines, concerns a large part of the companies and it will provide a good overview of current practices. Further, it is expected that the now voluntary guidelines will become mandatory in the near future (KPMG, 2017). So, examining a sample which distinguishes between the guidelines, could also provide more future oriented insights.
Using a sample with a limited amount of countries ensures that the chances are reduced that the differences in the observed variables are merely coming forth out of the differences between countries. However, in order to obtain sufficient data from a limited amount of countries, it almost becomes a necessity to focus on the US and UK, since data is more adequately available for these countries.
In addition to this, Peters & Romi (2015) used a US based sample and demonstrated that the value relevance of sustainability assurance has increased over the last few years. Therefore, this study could indicate if this trend continues and reinforce their findings. Peters & Romi (2015) use almost the same method but the sample differs.
Further, existing literature regarding aspects which will be examined in this research, were focused on a global sample (Perego & Kolk, 2012; Cuadrado-Ballesteros et al., 2017). So, by examining a limited amount of countries this research could give new insights.
The total sample consists of 647 UK and US listed firm, which publicized
sustainability information in 2016. This are all the available listed UK and US firm in the GRI database. From this total sample, 120 firms assured their sustainability information. The final sample is smaller, since not all assurance reports were available for analysis and not all data concerning the variables is available in
databases. Further, a few outliers were deleted which also resulted in a reduction of the sample size. The final sample consists of 389 firms of which 101 reports are assured, a breakdown of the sample is provided at the end of this section in table 3.
How data unavailability will be handled, during the research, is discussed below.
14 | P a g e 3.2 Dependent variable: Perceived value relevance to investors
To measure the value relevance, this research will use a method devised by Ohlson (1995). This method examines statistical associations between information and market value to assess the value relevance of information. Ohlson’s (1995) method, or some sort of variation, is used in prior research regarding sustainability information which enhances its reliability (Peters & Romi, 2015; Fazzini & Dal Maso, 2017;
Cormier& Magnan, 2007).
For instance, Peters & Romi (2015) use the valuation model provided by Ohlson (1995) to examine the influence of the various assurance choices on firm value. And Fazzini & Dal Maso (2017), use Ohlson’s (1995) model to examine the value
relevance of sustainability assurance. In formula form the model, based on Peters &
Romi (2015), looks as follows:
(1)
Pi = β0 + β1BVEi + β2NIi + β3Assurance i + β4-25INDUSTRY+ ε, P = common stock price per share
BVE = book value of equity divided by the number of shares outstanding (Peters & Romi, 2015)
NI = net income before extraordinary items divided by the number of shares outstanding (Peters & Romi, 2015)
Assurance = Vector of Assurance Variables
INDUSTRY = industry classification based GRI database in combination with Barth et al. (1999)
ε = other value relevant information, contained in variables not included in this model (Lev & Zarowin, 1999).
Information is considered value relevant if it coheres with equity market value. This is determined through the extent to which the assurance variables explain the stock price (P component). This is expressed in a R
2level. Coherence between the variables of interest reflects consequences of investors’ actions, whereas methods based on interviews or surveys reflect investors’ beliefs and opinions (Lev & Zarowin, 1999).
Regressing stock prices on earnings and book value is a popular method in
accounting research to examine the value relevance of information (Lev & Zarowin, 1999). Using a popular and frequently used model further increases its reliability and comparability. To claim causality between variables, alternative explanations need to be reduced as much as possible. So, to further increase the internal validity of this research, in addition to the error component, the total sample consists of firms with and without assurance on their sustainability information. This research examines the differences in value relevance between these two groups. Due to this, the effect in the stock price movement can be accounted to the assurance on the sustainability information.
Peters & Romi (2015), use the common stock price per share measured at the end of
three months after the fiscal year-end. This research uses the common stock price
per share measured at the end of the month in which publication of the assurance
report took place. This choice was made during the collection of data, since it was
noted that many assurance reports were provided in differing months. So, using the
15 | P a g e stock price in one specific month does not fully capture the effect of sustainability assurance on stock prices. Therefore, the stock price at the end of the month of publication is chosen to ensure that the possible effect of the sustainability report and the accompanied assurance is included in the stock price. If publication takes place in the last week of the month, then the stock price at the end of the following month is taken. If the date is unknown or unavailable, then the method of Peters & Romi
(2015) is used. That is, the common stock price per share measured at the end of three months after the fiscal year-end. In the additional analysis will the effect of different dates be tested.
Financial data regarding the dependent variable is collected through Orbis. Not all data is available for all companies, which causes a reduction of the total sample size.
If financial data concerning a company is missing, then if possible additional data is collected through Yahoo Finance.
3.3 Independent variable: Quality
The quality of the assurance report will be examined with a construct, devised by Perego & Kolk (2012) and Gurturk & Hahn (2016). Perego & Kolk (2012), devised the initial setup for the content analysis, Gurturk & Hahn (2016) made some
alterations to this construct. The content analysis gives a score to the assurance report which makes it measurable. Companies can score up to a total of 34 points on this analysis. The content analysis used in this research is a mixture of both of the above mentioned constructs. The complete content analysis can be found in appendix 1.
Perego & Kolk (2012), based their research on the suggestions of O’Dwyer & Owen (2005). The latter one introduced minimum requirements of a high quality assurance report as indicated by leading initiatives in the area of sustainability reporting. This includes the GRI, FEE and AccountAbility.
In a number of cases it occurred that an assurance report was not available. If the assurance report was not available, then the average quality score of the specific assurance provider is used for the concerning company. This choice was made, due to the observation of quality consistency of assurance providers. In table 1 are the average quality scores provided, including the standard deviation. The standard deviation with regard to the overall quality is relatively higher. This supports the idea that the quality is to a certain extent consistent per provider. It should be noted that the standard deviation of the category ‘other’ providers is even higher, however, this is not strange since this category consists of multiple providers. The average of the category ‘other’ providers is not expected to play a major role in this research, since only three firms with missing assurance reports belong to this category.
In the additional analysis the companies with missing reports will be left out to
determine possible differences. In addition to this will the overall average quality
score be used to determine if the outcomes differ. Since the analysis is performed by
a single person and contains a couple of judgemental issues, certain items were
checked again, after completion of the data collection, to enhance the consistency.
16 | P a g e
Table 1: Average quality scores and std. deviations per providerAvg. q St. Dev.
Veritras 23,56 1,94
CC 27 2,83
Deloitte 20 0
DNV 26,6 2,07
ERM 22,86 1,95
EY 23,56 2,96
KPMG 21,8 3,03
Lloyds 28,5 2,12
Other 21,69 4,13
PWC 19,91 2,47
SGS 21 1
TruCost 25 2
Overall 22,78 3,41
3.4 Independent variable: scope, provider, level of assurance
The independent variables scope, provider and level are observable in the assurance report. Therefore, these are recorded simultaneously with the execution of the
content analysis which is performed to measure the quality. Next, the used definitions of these three variables will be explained. If data considering these variables cannot be collected, due to unavailability of the assurance report, then data regarding these variables is collected through the GRI database. To incorporate these three
variables into a regression analysis, dummy coding is used. The variable scope consists of three categories, which necessitates the use of two dummy variables. The reason why two, instead of three dummy variables are used is because the value of the third variable can be deduced from the outcome of the other two. Therefore, a third variable will result in multicollinearity. The dummy variable chosen as reference category is narrow scope. These three variables focus on the content of the
assurance engagement. The content analysis, described above, focuses on what is disclosed in the assurance report. So, for example, regarding the scope, if
specifically is disclosed what is assured, points are awarded according to the quality construct.
3.4.1 Provider. This research will examine whether an audit firm or a consultancy firm has performed the assurance.
3.4.2 Scope. A broad scope in this research is defined as assurance on the entire report, this is in line with prior research (Braam & Peeters, 2018). If not the entire report is assured, then this research will discern between average and narrow.
The scope is considered average if in addition to GHG data also other information is
assured. If only GHG data are in scope for the assurance assignment, then the
scope is considered narrow.
17 | P a g e 3.4.3 Level of assurance. This research will distinguish between two levels of assurance, a reasonable/ high level and a limited/ low level. If Certain information is assured with a low level and another part of the same information is assured with a high level, then the level of assurance is considered high. Since this only concerns a few companies, it is not expected to have significant influence and therefore no additional tests are performed to examine possible effects.
3.5 Control Variable
Simnett et al. (2009), demonstrated that the demand for sustainability assurance is higher among companies engaging in industries with a larger social impact, this should be considered during the analysis of data. Companies in mining, utilities and finance are more likely to assure their sustainability reports, due to the fact that these companies have a greater need to manage environmental and social risks (Simnett et al., 2009). Further, when examining the sample, it is also important to consider country specific factors. Fernandez-Feijoo et al. (2016), showed that the choice of assurance provider depends on country specific factors. In addition to this, Simnett et al. (2009), Braam & Peeters (2018) and Perego & Kolk (2012), found evidence that country specific factors influence the demand for sustainability assurance. Therefore, it is important to control these variables in order to increase internal validity, because investors could perceive sustainability assurance overall more relevant for
companies in these sectors or countries.
These variables are collected through the GRI database. This research, however, combines the industry specification of the GRI database and Barth et al. (1999). This resulted in 22 categories. To incorporate the control variables into a regression analysis, the same method regarding dummy coding, is applied.
Model 1 is used to examine the effect of sustainability assurance on stock prices, relative to firms who did not assure their sustainability information.
However, this model is not sufficient. Considering all the above mentioned variables, the complete model looks as follows (model 2):
(2)
Pi = β0 + β1BVEi + β2NIi + β3Assurance i + β
4Quality + β
5Scope + β
6Provider + β
7Level + β
8Country + β
9-30Industry+ ε,
In Table 2 are the variables summarized, including their definitions which are used in
this study.
18 | P a g e
Table 2: Summary of the variablesAn additional specification of the final sample is provided below in table 3. As mentioned above, this sample consists of listed firms from the UK and US which provided sustainability information in 2016. This table gives insight in how to initial sample of 647 firms is reduced to 389 firms. The main cause is the unavailability of financial data in Orbis.
Table 3: Breakdown of the total sample
19 | P a g e 4. RESULTS
This section discusses the results from the various executed analyses. First some descriptive statistics are provided. Hereafter, are the results of the Pearson
Correlation tests given. This test gives insight in the relation between independent variables. Subsequently, the regression analyses are discussed. These analyses are the main tests of interest in this study. These tests give insight in the exploratory power of the independent variables in relation to the stock prices. This chapter concludes with some additional analyses, these analyses examine the effects of the choices made in the methodology section.
4.1 Descriptive statistics
According to the descriptive statistics (table 4 & 5), 237 firms and 152 firms are respectively from the US and UK, this sums up to a total of 389 firms. Of this total amount, 101 firms assured their information. Assurance was performed in 60 instances by a consultancy firm, the other 41 assurance reports were provided by audit firms. The average quality of an assurance report is 22,8 (SD = 3,41); with a minimum and maximum of respectively 17 and 33. The scope is in 15 cases broad and 61 cases average. However, the level of the assurance engagement is in most cases low, only 15 reports were provided with a high level of assurance (including the reports with a combined level).
Table 4: Frequencies
Frequency %
External assurance 101 26%
Audit firm 60 15%
Consultancy firm 41 11%
Broad Scope 15 4%
Average Scope 61 16%
Narrow Scope 25 6%
High Level 15 4%
Low level 86 22%
US 237 61%
UK 152 39%
Total firms 389 100%
Table 5: Descriptive statistics quality
Minimum Maximum Average Std. Dev.
Quality 17 33 22,8 3,4
4.2 Correlation
The Pearson correlation matrix, provided in table 6, indicates possible correlations
between independent variables. These correlation values are all below the threshold
of 0,7. There is one significant correlation found, namely for the variables quality and
audit firm (p = 0,039). Further, no significant correlations are found. To examine this
relation in more detail, an independent samples test is performed, table 7. The
20 | P a g e results of this test are significant, t(101) = -2,094, p = 0,039. This implies that, based on these outcomes, the average quality provided by audit firms (M = 21,5; SD = 3,52) differs from the average quality provided by consultancy firms (M = 23,0; SD = 3,34).
So, the average quality provided by audit firms is significantly lower. Comparing the differences in the average scores on each point in the content analysis, it becomes clear where these differences originate. Audit firms seem to score lower on the provision of further recommendations (point 20, appendix 1), relative to consultancy firms. Further, Consultancy firms seem to score relatively higher on the explanation of the materiality (point 16, appendix 1). The rest of the points are on average somewhat similar.
Additionally, the country is included in the Pearson correlation matrix. The results show that country and assurance provider choice have a significant negative correlation. This implies that there is a relation between the choice of provider and the country in which the firm resides in. This relation is tested in more detail through a cross table with chi-square test. The Chi-square test is significant, Chi-Square(1) = 20,235, p < 0,001. This means that, based on this sample, US firms more often chose a consultancy firm and UK firms more often chose an audit firm.
Table 6: Pearson Correlation Matrix
Country Audit Firm Quality Score Level Scope
Country 1
Audit
Firm -0,455** 1
Quality
Score -0,109 -0,206** 1
Level -0,123 0,138 0,098 1
Scope 0,151 0,064 -0,042 0,042 1
***, **, * Significant at respectively p< 0,01; p<0,05; p<0,10.
Table 7: Independent Samples T-Test
F T Sig. (2-tailed) Mean Difference Std. Error Difference Quality
Score 0,537 -2,094 0,039 -1,4476 0,6913
4.3 Regression analysis
In this section are the results pertaining to the different hypotheses discussed. These
results are collected through regression analysis. More specifically, multiple linear
regressions are performed to analyse to which extent the variables explain the stock
price, this is expressed in a R
2. The results are summarized in table 8.
21 | P a g e
Table 8: Summary of the results of the regression analysisModel 1 Assurance
vs No Assurance
Model 2
Quality Model 3
Provider Model 4
Broad Model 5
Average Model 6
Level h/l Model 7 Avg.
scope Complete
Model 8 Broad scope Complete Intercept 13,249***
(0,002) -33,032
(0,137) 22,950*
(0,080) 19,217*
(0,100) 29,621**
(0,015) 18,253
(0,132) -18,756
(0,423) -24,579 (0,325)
BVE 0,371***
(0,000) 0,547***
(0,001) 0,512***
(0,004) 0,503***
(0,003) 0,473***
(0,004) 0,498***
(0,004) 0,522***
(0,002) 0,553***
(0,001)
NI 5,492***
(0,000) 1,264
(0,130) 1,465*
(0,097) 1,767**
(0,047) 1,652*
(0,051) 1,392
(0,114) 1,468*
(0,074) 1,489*
(0,093) Assurance 6,311**
(0,038)
Quality 2,162***
(0,007) 1,898**
(0,017) 1,857**
(0,032)
Provider -4,349
(0,492) 0,145
(0,981) -2,281 (0,713) Scope:
Broad 12,502*
(0,085) 6,909
(0,356)
Average -14,678**
(0,011) -14,470**
(0,013) Level:
High/Low 3,042
(0,654) 6,868
(0,286) 2,452 (0,684) Industry
Controls Yes Yes Yes Yes Yes Yes Yes Yes
Country
Controls Yes Yes Yes Yes Yes Yes Yes Yes
N 389 101 101 101 101 101 101 101
R2 0,681 0,739 0,710 0,721 0,737 0,709 0,760 0,744
Adjusted
R2 0,659 0,637 0,597 0,613 0,634 0,595 0,658 0,627
