Does Integrated Reporting facilitate transparency?
First empirical evidence from South Africa
Master thesis, MscBA, Accountancy
University of Groningen, Faculty of Economics and Business
August 23, 2013 Willem van Roekel Studentnumber: 1704389 Admiraal de Ruijterweg 227 1055 LR Amsterdam Tel: +31 (0)6-13017226 Email: w.g.van.roekel@student.rug.nl Supervisor/ university E. van de Mortel 2nd Supervisor/ university J.T. Degenkamp Supervisor/field of study E. van der Lee
2
Does Integrated Reporting facilitate transparency?
First empirical evidence from South Africa
Abstract
On the first of March 2010 the King III governance code (Integrated Reporting) was
implemented in South Africa. The implementation of Integrated Reporting in South Africa is a result of changing interests of stakeholders. Corporate social responsibility is an upcoming topic the CEO needs to address. The way a company operates and does its business is important to stakeholders. Stakeholder theory, legitimacy theory and voluntary disclosure theory describe how the CEO can attend to the company’s stakeholders. This thesis will discuss how the CEO communicates with stakeholders in annual reports and examines the influence of Integrated Reporting. This thesis will also discuss impression management and the use of textual complexity in CEO reviews. I have analyzed the textual complexity of CEO reviews from the top 40 companies of the Johannesburg Stock Exchange (JSE) between 2008 and 2011. For each individual year there is a high use of textual complexity in the CEO review. I did not find a relationship between the implementation of Integrated Reporting and the use of textual complexity in CEO reviews. The industry in which the company operates does not influence the use of textual complexity. It is important to note that my findings are based on a limited sample.
Keywords: Integrated Reporting, textual complexity, impression management, stakeholder
3 Table of contents
1. Introduction ... 5
1.1 The reason for Integrated Reporting standards ... 6
1.2 South Africa, the predecessor of Integrated Reporting ... 7
1.2.1 The King III governance code ... 8
1.3 The informing role of the CEO in the Integrated Report ... 9
1.3.1 The CEO review in different industries ... 9
1.4 Main research questions ... 10
1.5 Scientific contribution ... 10
2. Theoretical Framework ... 12
2.1 Identifying and informing stakeholders ... 12
2.2 Is transparency an issue? ... 15
2.3 Textual complexity as a tool for obfuscation ... 17
2.3.1 Textual complexity and firm size ... 17
2.4 Transparency through King III ... 18
2.5 Transparency in different industries ... 19
3. Method ... 20
3.1 Sample selection ... 20
3.2 Method for testing hypothesis 1 ... 21
3.2.1 The Fog index score ... 22
3.2.2 The Flesch reading ease ... 22
3.2.3 The Smog formula ... 23
3.3 Method for testing hypothesis 2 ... 24
3.4 Method for testing hypothesis 3 ... 24
3.5 Control variables ... 26
4
4. Results ... 27
4.1 Hypothesis 1 ... 27
4.2 Hypotheses 2 and 3 ... 30
4.2.1 Descriptive statistics and control variables ... 30
4.2.2 Hypothesis 2 ... 31
4.2.3 Hypothesis 3 ... 33
5. Conclusion ... 36
5.1 Discussion ... 36
5.1.1 Practical implications and suggestions for future research ... 37
5.2 Final remarks ... 38
5 1. Introduction
“When I hear the businessmen speak eloquently about the social responsibilities of business in a free enterprise system, I am reminded of the wonderful line about the Frenchman who discovered at the age of 70 that he had been speaking prose all his life. The businessmen believe that they are defending free enterprise when they declaim that business is not concerned “merely” with profit but also with the promoting desirable “social” end; that business has a social conscience and takes seriously its responsibilities for providing employment, eliminating discrimination, avoiding pollution and whatever else may be the catchwords of the contemporary crop of reformers. In fact they are- or would be if they or anyone else took them seriously –preaching pure and unadulterated socialism. Businessmen who talk this way are unwitting puppets of the intellectual forces that have been undermining the basis of free society these past decades” (Milton Friedman, 1970, p.1).
Roughly forty years later the discussion about social responsibility of organizations is still going on, although the tone is slightly different. Recent events like the oil spillage in 2010 by British Petroleum (BP), the tax avoidance by Starbucks in the U.K. (2012) and the news of the dreadful work-environment in Apple factories in China (2010) have led to companies with unsatisfied stakeholders. Angry customers, embarrassed business partners and workers on strike resulted in companies being more aware of their environment.
In 1984 R. Edward Freeman introduced the stakeholder theory. Stakeholder theory
contradicts the words of Milton Friedman who claimed that the only interest of management should be the shareholders interest. Donaldson & Preston (1995) summarize the essence of existing literature of stakeholder theory management as follows: “stakeholder management requires, as its key attribute, simultaneous attention to the legitimate interests of all
appropriate stakeholders, both in the establishment of organizational structures and general policies and in case-by-case decision making” (Donaldson & Preston, 1995, p.67).
However, identifying and reporting to the stakeholders of an organization tends to be difficult (Adams, 2004, p.732). She argues that a mandatory reporting guideline is necessary to
improve the “lack of completeness” of reporting on social, ethical and environmental issues.
I have shortly introduced the concept of social responsibility of organizations. Next I will discuss the reason for Integrated Reporting standards and the need for corporate social responsibility (section 1.1). Then I will continue with Integrated Reporting and the role of South Africa (section 1.2). Subsequently I will discuss the role of the CEO (section 1.3);
6
introduce the research question of my thesis (section 1.4); and will describe the scientific contribution (section 1.5).
1.1 The reason for Integrated Reporting standards
I believe that an important step in creating a mandatory reporting guideline on social ethical and environmental issues is the upcoming of Integrated Reporting. Integrated Reporting is partially introduced in Europe. For example, several Dutch listed companies like Akzo Nobel and Philips have published annual reports that show some marks towards a more integrated report (PWC, 2012). Akzo Nobel and Philips among several others try to show their social impact on society. The Global Reporting Initiative (GRI) and the International Integrated Reporting Council (IIRC) seek to a worldwide implementation of Integrated Reporting. But what is Integrated Reporting and why is there a need for an integrated report? Let’s start with the initiator: the IIRC.
The IIRC describes the need and essence of Integrated Reporting as follows:
“At the heart of Integrated Reporting is the growing realization that a wide range of factors determine the value of an organization – some of these are financial or tangible in nature and are easy to account for in financial statements (e.g. property, cash), while many are not (e.g. people, natural resources, intellectual capital, market and regulatory context, competition, energy security). Integrated Reporting reflects the broad and longer-term consequences of the decisions organizations make, based on a wide range of factors, in order to create and sustain value. Integrated Reporting enables an organization to communicate in a clear, articulate way how it is drawing on all the resources and relationships it utilizes to create and preserve value in the short, medium and long term, helping investors to manage risks and allocate resources most efficiently” (www.theiirc.org/about/aboutwhy-do-we-need-the-iirc).
I think the essence of this description of the IIRC is as follows: there is a need to report on value creating activities and decisions concerning people, environment and other intangibles in addition to existing reporting standards, which focusses mainly on financial and tangible results.
The need for reporting on social and environmental issues is a subject that has been studied and discussed by the academic world for a while now. Deegan (2002) describes the growing interest of researchers in this field. He names legitimacy theory introduced by Lindblom (1994) as the main driver for managers to report on social and environmental issues.
7
“Legitimacy is the manner in which the value system of an entity is congruent with the value system of the larger social system of the entity. When there is a potential or actual
discrepancy between these value systems there is a threat to the legitimacy of the entity” (Deegan, 2002, p.293).
Fig (2005) describes the social challenges and threats to the legitimacy of organizations in South Africa. He notes the following: “What passes for Corporate Social Responsibility is often greenwash, distracting the gullible into believing that business has a serious
sustainability agenda” (Fig, 2005, p.617).
Fig (2005) argues for political action to create world-class standards for social and
environmental compliance. Banerjee (2008) supports this notion and finds that the voluntary disclosures on Corporate Social Responsibility often do not serve their purpose. Fig (2005), Banerjee, (2008) and Adams (2004) all point out the difficulties in reporting to all
stakeholders properly. They call attention to the shortcomings of existing reporting standards and urge for a new style of reporting where organizations can address their values and legitimacy.
In South Africa a form of Integrated Reporting has been developed that tries to respond to the needs mentioned by Fig (2005), Banerjee (2008) and Adams (2004). I will discuss this further in the next section.
1.2 South Africa, the predecessor of Integrated Reporting
South Africa is a country torn apart by social problems. Carter & May (2001) describe the poverty trap from which many South Africans cannot escape. Kahn et al. (2007) describes the demographic developments, the spread of HIV/AIDS and the growing violence in Agincourt (South Africa). They emphasize the importance of good registration of facts and
developments in poor countries. Hamann (2004) argues that mining companies are an important aspect of the social problems that South Africa struggles with. Mining companies often operate in poor and distant mining areas, and have much impact in these areas. The existence of these problems and the presence of valuable natural resources are the reason why South Africa is the frontrunner in legislation on Corporate Social Responsibility Reporting. Government, in extension the people of South Africa find it important that companies take responsibility to improve the social issues the country faces in exchange for the (natural) resources the companies use (South Africa has large mineral reserves, e.g. gold, titanium and platinum).
8
1.2.1 The King III governance code
South Africa is the leading country on Integrated Reporting. The King III governance code is implemented on the first of March 2010. Companies listed on the Johannesburg Stock
Exchange (JSE) are obligated to comply to this code (comply or explain principle), so they have to publish an integrated report. The King III governance code (2009) and the
consultation draft of the International Integrated Reporting Framework (2013) show a lot of similarities (Deloitte SA, 2011). Mervyn King, the chairman of the Integrated Reporting Committee South Africa (IRC SA) is also the chairman of the IIRC. This may explain the overlap. Being the first country to actually make Integrated Reporting mandatory for publicly listed companies South Africa is an important indicator for its success.
The King III code is the successor of the King II code (2002) which was implemented and developed by the IRC SA. One of the key principles of the King III code is that the board communicates with the stakeholders in “clear and understandable” language (King III, 2009). Although this was already mentioned in the King II code a lot has changed since then.
King III emphasizes three key points of change in regard to King II. Leadership,
sustainability and corporate citizenship. In this thesis I will focus on leadership because I think sustainability and corporate citizenship originate from (good) leadership. Daily & Johnson (1997) describe the sources of Chief Executive Officer (CEO) power. The CEO has a great amount of influence on the direction an organization takes and which values it pursues.
I believe that once management, specifically the CEO, is aware of the benefits and necessity of sustainability and good corporate citizenship he will guide the company towards this direction. The King III code describes good governance and leadership as follows: “Good governance is essentially about effective leadership. Leaders should rise to the challenges of modern governance. Such leadership is characterized by the ethical values of responsibility, accountability, fairness and transparency and based on moral duties that find expression in the concept of Ubuntu*. Responsible leaders direct company strategies and operations with a view to achieving sustainable economic, social and environmental performance” (King III, 2009, p. 9).
9 1.3 The informing role of the CEO in the Integrated Report
The description of leadership from the King III code sounds like a tough job for the CEO. Explaining how he (the CEO) satisfies all the needs mentioned by the King III code in clear and understandable language is perhaps even harder. Is it even possible to be clear and understandable, and thus transparent on all these subjects towards the targeted stakeholders? What are the incentives for a CEO to be transparent? Or perhaps the opposite: Why would a CEO benefit from being unclear and be willing to obfuscate results? And how does he obfuscate? In this thesis these questions will be analyzed from both a practical and
theoretical perspective. Organizations need to report on social and environmental activities and decisions. The existing standards did not facilitate this need. Therefore the IRC SA and next the IIRC introduced Integrated Reporting.
Analyzing integrated reports consists of two questions; are all the relevant stakeholders addressed in the integrated report? And is there transparent communication in the integrated report (Is the report clear and understandable)? This thesis will focus on the second question; is the message clear and understandable?
1.3.1 The CEO review in different industries
The King III code acknowledges the fact that companies are diverse in their activities and goals and therefore one size doesn’t fit all. I expect that the industry-type in which an organization is active will have effect on the readability of the CEO review.
Brammer & Millington (2003) show an association between the industry type and the level of community involvement. This variation in community involvement can be traced back to Freeman (1984), the foundation of Stakeholder Theory. Who is the stakeholder and what does he want? Different industries have different stakeholders, e.g. a mining company will have to deal with environmental movements, government institutions and health and safety administrations, where a telecom company will have less to deal with these parties.
Frooman (1999) addresses different types of influence strategies. He shows which determinants lead to different choices of influence strategies by stakeholders. It is also possible for the organization to influence the stakeholders, e.g. obfuscation hypothesis (Courtis, 1998; Laksmana, Tietz & Yang, 2011) and textual complexity (Rutherford, 2003) which will be further discussed in the Theoretical Framework. Due to the variety of
stakeholders and the different ways they influence companies, and vice versa, organizations address different mixes of stakeholders in different ways. In order to keep all major
10
in the proper way. The industry in which an organization and its CEO operate is therefore an important factor for the way a message is conveyed.
In the next section I will name the main research questions. Section 1.5 will point out the scientific contribution of this thesis.
1.4 Main research questions
This thesis will research the following questions:
Main research question: Does Integrated Reporting facilitate transparency in South Africa? In the previous sections I have discussed the need for Integrated Reporting Standards, the upcoming of Integrated Reporting, the dominant and predictive role of South Africa and the role of the CEO. In order to answer my main research question I will research three sub questions.
Sub question 1: How transparent is the CEO review?
Sub question 2: Does the implementation of the King III code have a positive effect on the readability of the CEO review?
Sub question 3: Does the industry type influence transparency in South Africa?
1.5 Scientific contribution
Despite of the numerous publications on Stakeholder Theory (e.g., Freeman, 1984;
Donaldson & Preston, 1995) there is still a lot to be explored about Stakeholder Theory. For example Scott & Lane (2000) and Adams (2004) focus on the identification of stakeholders of organizations. In addition Perrini & Tencati (2006) underline the importance of reporting & evaluating corporate performance towards stakeholders. Jensen (2002) studied the proper relationship between value maximization and stakeholder theory. This led to enlightened stakeholder theory.
Briefly summarized; existing literature focuses on the identification of stakeholders,
prioritization of stakeholders, tools and objectives to measure performance and subsequently the necessity to report about this performance to stakeholders.
This thesis focusses on how CEO’s report on corporate performance towards stakeholders. It evaluates and measures the readability of the CEO review and examines which factors
11
influence this readability. To the best of my knowledge this is the first study that evaluates the readability of CEO reviews in combination with the King III governance code (Integrated Reporting). This thesis is explorative in nature and aims to give insight into how transparent (level of readability) CEO’s communicate with stakeholders. It also takes into account to what extent industry types influence transparency. However it is important to note that the available data at this point in time is very limited. The implementation of the King III code and the upcoming of Integrated Reporting are recent developments. Therefore there are not a lot of years to analyze. This is an important limitation of my research. Notwithstanding it will be interesting to analyze the data that is available and examine the impact of Integrated Reporting on transparency. At this point in time little research is done concerning this subject.
12 2. Theoretical Framework
As already mentioned in the introduction this thesis will examine the relationship between the readability of the CEO review (Hypothesis 1), the expected positive influence of the
implementation of the King III governance code (Hypothesis 2) and the influence of the industry-type (Hypothesis 3). I will explain the relationships, resulting in hypotheses. The next section will focus on the role of the stakeholder. Stakeholder theory, legitimacy theory and voluntary disclosure theory all contain reasons for management to report to stakeholders. In section 2.1 and 2.2 I will discuss all three theories and their relationship with reporting. In the following section (2.3) I will explain how textual complexity influences transparency and discuss previous research on the matter (H1). Section 2.4 focusses on the implementation of the King III code (H2) and section 2.5 discusses the influence of the industry type (H3).
Figure 2.A Theoretical Framework
2.1 Identifying and informing stakeholders
I have briefly discussed stakeholder theory, introduced by Freeman in 1984 and mentioned the definition from Donaldson & Preston (1995). Their definition comes down to the point that stakeholder management is “simultaneous attention to the legitimate interests of all appropriate stakeholders” (Donaldson & Preston, 1995, p. 67). This sounds rather easy but in practice it tends to be difficult. What is legitimate interest and how are appropriate
The readability of the CEO review H 1 The implementation of the King III governance code
H 2
Industry-type H 3
13
stakeholders identified? In order to explain this problem I will use legitimacy theory and voluntary disclosure theory. I will start with stakeholder theory.
Scott and Lane (2000) describe the relationship between managers and stakeholders as follows:
“Stakeholders have legitimacy when managers perceive their actions or claims to be proper and appropriate, relative to the prevailing standards of the institutional environment within which they both operate” Scott and Lane (2000, p. 54).
They however note that in reality this situation is far more complex. Therefore managers have to make a tradeoff between different stakeholder interests. The identification of important stakeholders by managers is very important. Government agencies, shareholders, customers, employees, suppliers and banks are all good examples of potential stakeholders who have an effect or are affected by an organization (Freeman, 1984).
Mitchell, Agle & Wood (1997) describe three stakeholder attributes that explain and classify relationships between managers and stakeholders. Power, legitimacy and urgency together result in one of eight stakeholder typologies. Figure 2.B shows the different stakeholders that Mitchell et al. (1997) have identified using the three stakeholder attributes.
14
Figure 2.B Stakeholder typology
(Mitchell et al., 1997, p.874)
This framework can be used by managers to identify and to prioritize stakeholders.
Legitimacy is a recurring element in stakeholder theory and therefore important to address. There are many definitions of legitimacy (e.g. Lindblom, 1994; Mitchell et al., 1997). Suchman (1995) defines legitimacy as follows:
“legitimacy is a generalized perception or assumption that the actions of an entity are desirable, proper, or appropriate within some socially constructed system of norms, values, beliefs and definitions” (Suchman, 1995, p.574).
In order for managers to inform the stakeholders there are a couple of important aspects to keep in mind. Who are the most important stakeholders? Managers need to identify and classify the important stakeholders of their organizations, possibly using the three attributes described by Mitchell et al. (1997). And what are their expectations of management? The manager must create legitimacy through actions that are congruent with the wishes of the different stakeholders, satisfying the essential stakeholders of the organization. When management has set out its direction there is another important question that arises. How do
15
we inform the stakeholders and what do we tell them? Naturally almost all countries have ground rules on the subject. However there is a lot of freedom for management while choosing topics they want to discuss.
The most important aspect of section 2.1 is the realization that identifying and prioritizing stakeholders is very difficult. I have used stakeholder theory and legitimacy theory to explain which difficulties managers face when reporting to stakeholders. To illustrate the potential stakeholders of an organization and their characteristics I have used the stakeholder typology (Figure 2.B) of Mitchell et al. (1997). However identifying and prioritizing stakeholders are not the only relevant aspects in the reporting process. The incentives that managers have to report are an important aspect as well.
Voluntary disclosure theory explains the different incentives management can have while reporting to stakeholders. I will further discuss this in section 2.2.
2.2 Is transparency an issue?
The King III code claims that leaders should be transparent and clear. This seems obvious but due to ulterior motives of management “getting the message across” is not always
management’s goal of reporting. To understand why management deliberately chooses to be unclear it is important to know why they are reporting in the first place, especially when it is not required by law. This often applies for Corporate Social Responsibility Reporting.
Although the King governance codes (I, II and III) try to facilitate and direct a more regulated environment for this kind of reporting a lot of it stays voluntary. This also applies for topics which a CEO wants to discuss in his review in the annual report.
Healy & Palepu (2001) describe and summarize several reasons for voluntary disclosure by management. Some reasons are in the best interest of the company and its shareholders, e.g. a manager who voluntary discloses to minimize the information risk (keep shareholders and stakeholders as well informed as possible) and thus increases the liquidity of the
organizations equity. But there are also reasons to disclose that are not in the best interest of stakeholders and shareholders. Some examples Healy & Palepu (2001) mention are the litigation cost hypothesis and the corporate control contest hypothesis. Litigation cost hypothesis: the manager discloses as much as possible so he will not get sued for not informing the stakeholders. As a result he can hide bad results as a needle in a haystack.
16
Corporate control hypothesis: when management is under pressure due to poor stock and earnings performance: “managers use corporate disclosures to reduce the likelihood of undervaluation and to explain away poor earnings performance” (Healy & Palepu, 2001, p.421).
This behavior of management does not improve transparency. On account of possible job loss or upcoming bonuses it can be beneficiary for management to hide or “explain away” (Healy & Palepu, 2001, p.421) bad performance and decisions.
In the previous paragraphs we have seen that management does not always benefit from a clear and transparent message which openly describes the performance and decisions made and taken by management. In some cases the actual performance and results are blurred and hidden in the message. Management is required to deliver information on performance and results, however they enjoy much freedom in the way they convey this information.
Lindblom (1994) identifies five communication strategies that management can use; educate and inform, influence perception, manipulating perception, change external expectations and ignoring. Where the first strategy is positive the other four are intended to distort the picture that a stakeholder has of an organization. There exists information asymmetry between management and the stakeholders and these strategies explain how management is able to exploit the gap of information between them and the stakeholders. Influencing and
manipulation of stakeholders perception have been subject to various studies.
Impression management explains why management uses e.g. textual complexity to influence transparency.
“Impression Management is a field of study within social psychology and is concerned with studying how individuals present themselves to others” (Hooghiemstra, 2000, p. 60). Schlenker & Weigold (1992) name three main motives for impression management: “self-glorification (self-esteem maintenance and enhancement), self-consistency (validating the self by confirming self-beliefs), and self-authentication (trying to learn the truth about self by pursuing diagnostic information” (Schlenker & Weigold, 1992, p.138).
These three motives describe in what way the manager wants to be perceived, sometimes contradicting and manipulating the facts reported in his story to achieve this goal.
To some extend these motives correspond with Lindblom (1994). The similarity is in the fact that management continuously is able to manipulate and influence the stakeholders of the
17
organization through the disclosures and annual reports they publish. This is a result of information asymmetry (agency problem) and the benefits/pressure for management to convey the message in the most favorable light (legitimacy theory, impression management, incentives, possible job loss etc).
To summarize section 2.2, there are several reasons for management to report. Voluntary disclosure theory and impression management explain that not all of the reasons to report benefit transparency. Transparency isn’t always the main goal of management and being unclear and vague could just as easily suit their purpose better. In section 2.3 I will explain how textual complexity can be used by management to influence transparency.
2.3 Textual complexity as a tool for obfuscation
We have seen that transparency does not have to be in the best interest of management. In some situations the facts do not favor management and in that case being vague about performance and results might suit management better. Self-glorification, self-consistency and self-authentication (Schlenker & Weigold, 1992) could be possible motives to manage the impression that stakeholders get from published annual reports and disclosures. Textual complexity is a tool often used for this purpose. Courtis (1998) shows that management uses textual complexity to obscure news/results in annual reports. E.g. air-brushing bad news in a story or hiding a strategic mistake behind a complicated stream of words. Rutherford (2003) finds that textual complexity is high in annual reports. Li (2008) shows a relationship between earnings of a public company and the readability of the annual report. Firms with higher earnings have an easier to read annual report. Firms with lower earnings have an annual report that is hard to read. These results of Li (2008) show that management
deliberately makes annual reports more complex/harder to read to hide bad results. This also applies when it comes to management compensation. Laksmana et al. (2012) show that when CEO pay exceeds the benchmark, the readability of the Compensation Discussion & Analysis Disclosure (CD&A) in the annual report becomes more difficult to read. In this case the attempted justification of high CEO pay in the CD&A is deliberately made vague and unclear by means of technical complexity.
2.3.1 Textual complexity and firm size
Previous readability research (e.g. Li 2008; Rutherford, 2003) also takes organization size into account while measuring textual complexity. The size of an organization has a positive effect on the amount of information being shared by an organization (Buzby, 1975). Achmed & Courtis (1999) also find a positive relationship between organizational size and the amount
18
of information that is disclosed by an organization. This is easily explained by the fact that with a certain size there are commitments and responsibilities to report towards stakeholders. Cho, Roberts & Patten (2010) show that the size of a company (measured in total assets) has influence on the language used in their environmental disclosures. Li (2008) also finds firm size (measured in market to book ratio) influence on report readability. In order to be
thorough and acquire a good view of the factors that could influence textual complexity I will include firm size as a control variable into my model for hypothesis 2 and 3. In section 3.5 I will describe the two ways I have measured firm size.
The work of Li (2008), Rutherford (2003) and Laksmana et al. (2012) have in common that they all conclude that annual reports in general are hard to read. But is this the case in South Africa? The King codes aim on transparency towards stakeholders. But when the
stakeholders can’t understand the reported information because the textual complexity is too high, there is no transparency. Not only shareholders but all stakeholders should be able to understand the reported information. For example customers, suppliers and employees could be important stakeholders for an organization.
Based on previous readability research (e.g. Li, 2008; Rutherford 2003; Laksmana et al., 2012) I expect the CEO review to be hard to read for stakeholders. This leads to my first hypothesis:
Hypothesis 1: The CEO review is hard to read for stakeholders.
2.4 Transparency through King III
With the introduction of integrated reporting through the King III code this research explores new territory. Adams (2000) argues that there is a gap between performance and reporting on social, ethical and environmental issues and that regulation could be the answer. Textual complexity to hide bad news, unreadable prose about the effects companies have on society and vague descriptions on remuneration are where we stand now. However awareness on the subject is growing, both from scientific and practical points of view. I have discussed several studies that established the low level of readability of annual reports and pointed out several reasons why this is the case. The media pointed out some errors in judgment of management on which they “forget to report” in previous years, or did so vaguely and hunts vigorously on scandals. The IIRC has the initiative from both government and business to build a
framework that can overcome the discussed difficulties. King III is the vanguard in this initiative being the first form of Integrated Reporting actually being mandatory. One of the
19
key points of King III is creating transparency towards stakeholders on subjects concerning the organization and its surroundings, but is this the case? This leads to my second
hypothesis.
Hypothesis 2: The implementation of the King III code has a positive effect on the readability of the CEO review.
2.5 Transparency in different industries
In the introduction I have briefly discussed that the King III governance code states that organizations listed on the JSE are very diverse and that they address different stakeholders. Identifying and addressing the right stakeholders tends to be difficult. Achterkamp & Vos (2005) describe the process of identifying stakeholders. Banerjee, Iyer and Kashyap (2003) show the different factors that influence corporate environmentalism and the explanatory role of industry type. This corresponds with Brammer & Millington (2003) who describe the relationship between stakeholder preference, industry type and community involvement. They show that the industry in which an organization is active affects the amount of
corporate community involvement. Cooke (1992) displays the influence of industry type on the amount of disclosure in the annual report. Due to the very diverse nature of the companies listed on the JSE, the different industries they are active in and as a result different
stakeholders they deal with, I expect varying ways the different industries address
stakeholders in the annual report. E.g. a food retailer will have other stakeholders, priorities and interests than an investment bank. The food retailer will probably identify (in addition to the shareholders) the customers and the health administration as primary stakeholders. The emphasis of the Investment bank will probably lay with (potential) investors, shareholders and financial agencies. These stakeholders demand different information and have diverse backgrounds and education levels. This leads to my third and last hypothesis.
20 3. Method
In this section I will explain the method I have used for analysing hypotheses 1, 2 and 3. I will discuss each hypothesis individually (section 3.2, 3.3 and 3.4). In section 3.5 I will explain the control variables I have used and section 3.6 shows my statistical model. I will start with my sample selection (3.1) and discuss the data I have used.
3.1 Sample selection
For my sample selection I have selected the CEO reviews from JSE listed companies in South Africa. Due to the implementation of King III governance code (2010) the JSE in South Africa it is the first and only country/stock exchange in the world where integrated reporting is mandatory (see section 1 and 2). I have selected the 40 largest companies of the JSE. Both Reuters and Bloomberg (well-known distributors of financial information) make the distinction between the top 40 and the rest of the JSE listed companies. The top 40 companies of the JSE are the only ones that are reasonable in size (measured in market capitalization, total assets and revenue) on a global scale and therefore internationally comparable.
I have selected the two years before the implementation of the King III code on the first of March 2010 (2008, 2009) and the two years after implementation (2010, 2011). Annual reports of 2012 were often not yet available. This results in 157 observations (company-year combinations). For these 40 JSE listed companies there were three year/company
combinations with a missing CEO statement due to diverse reasons. I have used Orbis, a database made available by the University of Groningen, to collect additional information (e.g. the industry codes, total assets, total revenue, market capitalization and market price year-end).
The small amount of suitable company-year combinations available at the JSE is an important limitation of my research and reduces the explanatory power of my model.
However there is no alternative data available at this point. On the other hand the strength of this sample is that it consists of internationally active firms who compete and do business all over the world and are therefore comparable to other multinational firms abroad. Which according to the IIRC will be the first to convert towards Integrated Reporting.
21 3.2 Method for testing hypothesis 1
Hypothesis 1: The CEO review is hard to read for stakeholders.
In order to analyze the CEO review on readability I have collected all the CEO reviews of the annual reports in the targeted years (2008-2011). In some cases there was an executive
Chairman or a joined statement from both Chairman and CEO in which case I collected that statement. In order to analyze the CEO review I copied the unformatted text and subsequently transformed this statement into a uniform standardized text document so I could measure the readability as objectively as possible.
There are several formulas to calculate the readability of a text in English. “A readability formula is a quantitative method of predicting whether prose passages are likely to be readable by a target audience. It attempts to provide the same kind of information about comprehension ease that a writer would have to judge through experience and feedback from readers, or measure through a comprehension test on the material. The success of a formula in providing meaningful predictive information depends on its ability to measure elements in writing that are related to reader comprehension” (Courtis, 1986, p.285).
The three most commonly used readability formulas are the Fog index score (Robert
Gunning, 1952), the Flesch reading ease (Rudolf Flesch, 1948) and the Smog formula (Harry McLaughlin, 1969) (Laksmana et al., 2012; DuBay, 2007). For hypothesis 1 I will analyze all three readability scores. For hypothesis 2 and 3 I will use the Fog index score as primary indicator of readability, this index is also used by Li (2008) and Courtis (1995). In order to test the outcome of the Fog index readability score I will compare it with the Flesch reading ease and the Smog formula. I will also calculate all the readability scores for each year individually. All three formulas are explained in the next three subsections. I have used the readability calculator from www.online-utility.org to calculate the readability scores. This is the standard readability calculator used by the University of Groningen.
22
3.2.1 The Fog index score
Fog index score = (words per sentence + percentage of complex words) * 0.4
Complex words are defined as words with three syllables or more. The Fog Index score indicates 5 levels of readability:
Table 3.A The Fog index scores
Fog index score Description of readability
> 18 Unreadable 14 – 18 Difficult 12 – 14* Ideal* 10 – 12 Acceptable 08 – 10 Childish * Norm (Li, 2008, p.225)
3.2.2 The Flesch reading ease
Flesch reading ease = 206.835 – (1.1015 * number of words / number of sentences) – (84.6 * number of syllables/ number of words)
The outcomes of the Flesch reading ease are values ranging from 0 to 100 and measure the readability of a text. They are classified into seven categories by Courtis (1995). He shows the following table (3B) which explains the readability scores, the style and the level of education required.
23
Table 3.B The Flesch reading ease ratings
Flesch reading ease rating Description of readability Education attainment level Typical style of magazine
00 – 30 Very difficult Postgraduate degree Scientific
30 – 50 Difficult Undergraduate
degree
Academic
50 – 60* Fairly difficult* Grades 10-12 Quality
60 – 70 Standard* Grades 8-9 Digests
70 – 80 Fairly easy Grade 7 Slick fiction
80 – 90 Easy Grade 6 Pulp fiction
90 – 100 Very Easy Grade 5 Comics
* Norm
(Courtis, 1995, p.7)
3.2.3 The Smog formula
Smog grading = 3 + square root of polysyllable count.
Polysyllable count is the total of words with more than two syllables (DuBay, 2007). Mc Laughlin (1969) describes three levels of readability measured in required education (table 3.C).
Table 3.C The Smog formula grading
Smog grading Required education
> 19 Professional qualification
17 – 18 Graduate training
13 – 16* College education
* Norm around 13
24 3.3 Method for testing hypothesis 2
Hypothesis 2: The implementation of the King III code has a positive effect on the readability of the CEO review.
The first explanatory variable is the implementation of the King III code on the first of March 2010. After analyzing the CEO reviews on readability (hypothesis 1) I will use the Fog index score as input (dependent variable) for my analysis. As discussed in the sample selection I have 157 company-year combinations divided over 4 years (2008-2011). First I will analyze the readability (using the Fog scores) of each year independently. Here I will analyze the trend of the Fog score (textual complexity) and look for significant change after the implementation of the King III code.
In addition I will divide the total distribution (157 company-year combinations) into two categories. Pre implementation of King III on the first of March 2010 and post
implementation of King III (table 3.D).
Table 3.D implementation King III
Pre implementation King (0) Post Implementation (1)
Annual reports from: Annual reports from:
2008 & 2009 2010 & 2011
Using linear regressions I will analyze the influence of the implementation of the King III governance code and search for a significant effect.
3.4 Method for testing hypothesis 3
Hypothesis 3: The industry type has effect on the readability of the CEO’s review.
The second explanatory variable is the classification of industry type. Several studies have used industry-type and in addition industry classification to explain certain interactions (e.g. Cooke, 1992; Banerjee et al., 2003; Brammer & Millington, 2003). In order to categorize the different JSE listed companies I will use the Fama & French method (1997). Bhojraj, Lee and Oler (2003) name the Fama & French method of industry classification as the one primarily used by academics.
25
The Fama French classification method acknowledges the following major industry types:
Table 3.E SIC codes
Codes Contains
SIC 1 Consumer (non)durables, wholesale, retail, some services SIC 2 Manufacturing, energy, utilities
SIC 3 Business equipment, telephone, television transmission SIC 4 Healthcare, medical equipment, drugs
SIC 5 Mines, construction, hotels, finance
All companies are classified in industries with a four digit SIC code. I have acquired these codes through Orbis. Using the codes I can divide the companies into five major industry types and use them as a dummy variable (0 or 1). Due to the large amount of mining and financial companies listed on the JSE I will also compare SIC 5 with the rest of the industry types, in order to gain a good view of the difference in textual complexity between sectors (Table 3.F). This is also needed because of the small sample size (157). SIC 5 is by far the largest industry group in South Africa and embodies half of the sample. By dividing the SIC codes into two major industry groups I can analyze if there is an industry effect (Cooke, 1992). In order to gain insight in the different industry classifications and their influence individually on textual complexity a larger sample is necessary. Due to the only recently introduced King III code there is limited data available. In my thesis I will investigate if there is an industry effect and not focus on the specific industry classifications.
Table 3.F Mining and financials versus all other industries
Codes Contains
SIC 5 Mines, construction, hotels, finance SIC 1,2,3 & 4 All other industries
26 3.5 Control variables
The control variables I will use for hypothesis 2 and 3 are market price year end and the total assets of a company. Previous research has shown (e.g. Li, 2008; Cho, Roberts & Patten, 2010) that the organization’s size can influence the textual complexity of an annual report. In order to be thorough I will include these variables in my statistical model. Due to the wide range of high values for total assets I will use the natural logarithm for the total assets of a company as a control variable (Cho, Roberts & Patten, 2010). In addition I will also look at the market price at the end of the year (Li, 2008).
3.6 Statistical model
The dependent and independent variables discussed in the previous section lead to the following statistical model:
Fog score = β0 + β1 Implementation King III + β2 Industry-type + β3 Total Assets + β4 Market Price + ε
27 4. Results
4.1 Hypothesis 1
Hypothesis 1: The CEO review is hard to read for stakeholders.
In order to test hypothesis 1 I measured the textual complexity of the CEO review in three different ways using readability formulas (The Fog index score, the Smog formula and the Flesch reading ease). Subsequently I combined the scores from all three readability formulas and linked the scores to the descriptions from tables 3.A, 3.B and 3.C. Finally I calculated the readability scores for each independent year (2008-2011) resulting in table 4.A. The complete explanation of these readability scores can be found in the method section.
Table 4.A Readability scores Readability
scores
Fog index score The Smog
formula The Flesch reading ease Norm 12-14 ≈13 50-60 Mean 2008-2011 15.234 14.666 35.772 2008 14.935 14.430 36.789 2009 15.367 14.835 35.973 2010 15.503 14.904 34.343 2011 15.140 14.574 35.976 Median 2008-2011 15.160 14.670 35.530 2008 14.860 14.330 35.850 2009 15.090 14.600 36.930 2010 15.450 14.910 34.900 2011 15.320 14.750 34.930 Std. Deviation 2008-2011 1.404 1.079 6.224 2008 1.302 0.943 5.422 2009 1.492 1.155 6.787 2010 1.058 0.858 4.786 2011 1.620 1.245 7.272
28 Variance 2008-2011 1.970 1.164 38.736 2008 14.935 14.430 36.789 2009 15.367 14.835 35.973 2010 15.503 14.904 34.343 2011 15.140 14.574 35.976
The results shown in table 4.A provide proof that the CEO review is hard to read for
stakeholders. All three scores indicate a high level of textual complexity in the CEO review. The Fog index score, the Smog formula (shown in Figure 4.A) and the Flesch reading ease (Figure 4.B) show the same change in textual complexity in the targeted years. Note that the Fog index score and the Smog formula show increased textual complexity with ascending scores (Figure 4.A). The Flesch reading ease shows increased textual complexity with descending scores (Figure 4.B). The year 2008 is the year with the lowest amount of textual complexity in the CEO review. The years 2009 and 2010 both show a small increase of textual complexity, followed by a small decrease in textual complexity for 2011. Because all three readability formulas indicate the same changes of textual complexity for the targeted years and draw the same conclusions: the textual complexity is high, I consider the calculated Fog index scores as reliable. Therefore I will use the Fog index score as dependent variable for hypothesis 2 and 3 (section 4.2).
Figure 4.A Fog index score and Smog Formula
13.000 13.500 14.000 14.500 15.000 15.500 16.000 2008 2009 2010 2011
Fog index score Smog Formula
29
Figure 4.B Flesch reading ease
Table 4.B shows a short summary of the results from table 4.A and interprets the scores. The Fog index score and the Flesch reading ease describe the texts as difficult, the Smog formula concludes that the reader must have a college education or higher to understand the CEO review. Hypothesis 1 is accepted.
Table 4.B Summary of readability scores
Readability index Score Description Norm
Fog index 15.234 Difficult 12-14
Smog formula 14.666 College education ≈13
Flesch reading ease 35.772 Difficult 50-60
33.000 33.500 34.000 34.500 35.000 35.500 36.000 36.500 37.000 2008 2009 2010 2011
30 4.2 Hypotheses 2 and 3
In this section I will show the results from testing hypotheses 2 and 3, explain my control variables and discuss some general findings of my research. Subsection 4.2.1 will discuss the descriptive statistics and the control variables I have used (Total Assets and Market Price). Subsection 4.2.2 will discuss hypothesis 2 (Implementation King III). Subsection 4.2.3 will discuss hypothesis 3 (Industry-type).
Fog score = β0 + β1 Implementation King III + β2 Industry-type + β3 Total Assets + β4 Market Price + ε
4.2.1 Descriptive statistics and control variables
The dependent variable for hypotheses 2 and 3 is the Fog index score. I have compared the outcomes of the Fog index score with the Flesch reading ease and the Smog formula in order to thoroughly measure the textual complexity of a text (hypothesis 1). Table 4.C contains the means, standard deviations and correlations of the main dependent variable (Fog index score), the independent (pre/post King III and industry-type) and control variables (market price year end and total assets). It contains the correlations between the variables to gain insight in interactions between the independent variables. In order to obtain this insight I will test for multicollinearity using the Pearson correlation. There are no significant outliers in the acquired data so there is no need for windsorizing. The control variable total assets however consisted of a wide range of high values (Total Assets measured in mil. ZAR). Therefore I used a natural logarithm to bring back these values to a more comprehendible number (LogTA).
Table 4.C Mean, std. deviation and multicollinearity
Variables Mean Std.
Deviation Fog score
King III Industry-type Total Assets Market Price Observed Fog score 15.234 1.408 King III 0.510 0.500 0.060 Industry-type 0.548 0.498 0.147 0.05 Total Assets 4.636 0.595 0.009 0.082 0.230** Market price 94.283 95.409 -0.184* 0.112 0.050 0.122
31
The outcomes in table 4.C do not indicate the presence of multicollinearity. The Pearson correlation score at which correlation becomes an issue is about 0.6 to 0.7. In table 4.C you can see that the highest correlation is 0.230 (between Industry-type and Total Assets). In order to determine the influence of my independent variables (King III and Industry –type) on the Fog index score I will test them separately.
Table 4.D shows the control variables that I have used for hypothesis 2 and 3. For each regression (1-5) I will show the R-Square (explanatory power of all independent variables combined) and the F-statistic (group variance indicator, which must be positive) of the regression. For each variable I included the Beta coefficient (the slope) and the t-statistics (coefficient and Sig.). The t-statistics show if the independent variable has a significant influence on the dependent variable (0.10 alpha level).
Table 4.D Control model
Regression 1 has a low R-Square (0.035) which indicates that both independent variables combined do not explain a lot of the variance of the dependent variable. I find a low negative Beta value (-0.076) that indicates that firm size (measured in Total Assets) has a small significant (0.021) effect on the textual complexity used in CEO reviews (Fog score). For market price I do not find a significant effect. In contrast with previous research (Li, 2008; Roberts & Patten, 2010) I did not find a strong significant firm size effect.
4.2.2 Hypothesis 2
Hypothesis 2: The implementation of the King III code has a positive effect on the readability of the CEO review.
The first explanatory variable in my thesis is the implementation of the King III code. In section 4.1 I have analyzed the textual complexity of the CEO review for each year
Regression 1 B t Sig. R-Square F-statistic
0.035 2.713
Fog score 15.125 17.245 .000
Total Assets -.076 -2.326 .021
32
individually (Table 4.A). The textual complexity is at its lowest point in 2008, increases in 2009 and 2010 and decreases in 2011. This indicates that the implementation of the king III code did not have an effect on the use of textual complexity in the CEO review in 2010. However in 2011 there is a decrease in textual complexity.
In order to be thorough I will also perform two linear regressions (table 4.E and 4.F) to analyze the effect of the implementation of King III. It is however important to note that due to the small amount of company-year combinations available (section 3.1), these regressions are a relative method for testing hypothesis 2 (the same applies for hypothesis 3). Table 4.E contains two control variables and the King III variable (2008, 2009 vs. 2010, 2011). Table 4.F (also used for hypothesis 3), shows my entire statistical model that takes all discussed variables into account.
Table 4.E King III variable and control variables
Table 4.F Complete statistical model
The results in table 4.E and 4.F indicate that the implementation of the King III code does not have an effect on the readability of the CEO review. The effect of King III on the Fog score is positive (Beta values of 0.259 and 0.261, which indicate an increase in textual complexity),
Regression 2 B t Sig. R-Square F- statistic
0.043 2.252
Fog score 15.075 17.184 .000
Total Assets .61 .324 .747
Market price .003 -2.434 .016
King III .259 1.149 .253
Regression 3 B t Sig. R-Square F- statistic
0.035 2.388 Fog score 15.206 17.362 .000 Total Assets -.011 -.058 .954 Market price -.003 -2.486 .014 King III .261 1.163 .247 Min/Fin(SIC) .379 1.649 .101
33
but it is not significant (0.253 and 0.247). In my regressions I have tested for a deviation in the trend. The years 2008 and 2009 (the two years before the implementation of the King III code) do not significantly deviate from the years 2010 and 2011 in regard to the readability of the CEO review. Even if we test both independent variables together and separately
(regression 2 and 3) there is no proof of a relationship between the readability of the CEO review and the implementation of the King III governance code. The implications of these results and the limitations of my thesis I will discuss in section 5. Hypothesis 2 is rejected.
4.2.3 Hypothesis 3
Hypothesis 3: The industry type has effect on the readability of the CEO’s review.
The second explanatory variable in my thesis is the effect of the industry type on the use of textual complexity in the CEO review (Fog index score). Table 4.G shows the readability scores for each individual SIC code. Table 4.H shows the readability scores of the mining and financial companies and all other industries combined (see section 3.4). Table 4.F (displayed in the previous subsection) shows the complete statistical model. Table 4.I shows the
different SIC codes and their effect on the Fog score. I have used SIC 1 as holdout year. Table 4.H shows the industry effect, measured as Mining and Financials (SIC) versus all other industries (see section 3.4).
Table 4.G Fog scores of individual SIC codes
Fog scores Norm 12-14
SIC 1 SIC 2 SIC 3 SIC 4 SIC 5
Mean 15.448 14.961 15.380 15.630 15.421
Median 15.210 15.120 14.195 15.625 15.360
Std. Deviation 1.430 0.804 1.354 0.835 1.434
34
Table 4.H Mining and financials versus other industries
Fog scores Norm 12-14
Mines, construction, hotels, finance (SIC 5)
All other industries (SIC 1,2,3, & 4)
Mean 15.421 15.007
Median 15.360 15.060
Std. Deviation 1.434 1.331
Variance 2.056 1.773
Table 4.I Industry variable using all SIC codes, hold out SIC 1 and control variables
Table 4.J Industry variable Min/Fin versus other industries and control variables
Regression 4 B t Sig. R-Square F- statistic
0.108 2.936 Fog score 15.399 17.595 .000 Total Assets .042 .222 .824 Market price -.002 -1.789 .076 SIC 2 -.349 -.765 .445 SIC 3 -1.142 -2.689 .008 SIC 4 .113 .234 .815 SIC 5 -.010 -.031 .976
Regression 5 B t Sig. R-Square F- statistic
0.052 2.727
Fog score 15.256 17.419 .000
Total Assets .004 .023 .982
Market price -.003 -2.376 .019
35
The readability scores in table 4.G show little variations for each individual SIC code. The readability score of the mining and financial companies (table 4.H) is a little bit higher than the other industries combined. All industries have a high readability score, well exceeding the norm. This indicates high use of textual complexity in CEO reviews.
The results displayed in table 4.F, 4.I and 4.J on the varying effect of the industry type on the readability of the CEO review do not show a relationship. I have measured this effect in two ways. In regression 4 (table 4.I) the SIC categories 2, 3, 4 and 5 were individually compared to SIC 1 in combination with the Fog index score. Here I found one significant (.008) negative effect of the SIC code 3 in combination with the Fog index score.
In regression 5 I have combined the largest SIC code 5(Min/Fin) and compared it with the total of SIC 1, 2, 3 and 4. Here we see an indication of an industry effect on the readability, however it is not significant. I have tested the industry effect both individually and in combination with the other variables. Both were almost significant (0.101 and 0.103), but almost isn’t enough. So although there is an indication of a varying industry effect on readability the effect is not significant enough. On account of this regression it is therefore not possible to claim that the industry type has a varying effect on the readability of the CEO’s review. Hypothesis 3 is rejected.
36 5. Conclusion
5.1 Discussion
My findings are explorative in nature and to some extend align with previous research. Previous studies (e.g. Li, 2008; Courtis, 1998 and Laksmana et al., 2012) find high textual complexity in annual reports and thus a lack of transparency. This thesis shows that for JSE listed companies high textual complexity also applies (hypothesis 1). To the best of my knowledge this has never been tested in South African context. All three readability scores (Fog index score, Smog Formula and Flesch reading ease) indicate a high textual complexity for the reviewed CEO reviews. Every single analyzed year (2008-2011) shows high textual complexity and vastly exceeds the desired norm (Table 4.A and section 3.2).
In section 2 I have discussed several reasons for management not to be transparent and explained that textual complexity is a tool often used to influence transparency. This brings us to the main research question of my thesis; does Integrated Reporting facilitate
transparency in South Africa?
I did not find any proof that obligating Integrated Reporting (through the King III governance code) improves transparency (hypothesis 2). The results shown in table 4.E and 4.F show that the implementation of the King III governance code did not have a significant influence on the textual complexity of the CEO review (measured with the Fog index score). However in 2011 there was a decrease in textual complexity. This is in contrast with the increase of textual complexity in 2008, 2009 and 2010. It will be interesting to see if this decrease in textual complexity continuous in the coming years. Due to the only recently implemented King III code the data available on the subject is limited. At this point in time there are only a few years that can be analyzed. Therefor an analysis of the average use of textual complexity for each year individually proved to be more suitable for my sample. For future research it will be interesting to see if there is a significant effect over a longer period of time (e.g. 7 years before the implementation versus 7 years after implementation). The limited data available at this point decreases the chance to find a significant effect and diminishes the explanatory power of my model. This is an important limitation of my research.
Notwithstanding I believe it is still a very interesting development to analyze.
In order to get a better picture of the different stakeholders of an organization and their effect on transparency I have compared industries. Different industries have different stakeholder interests and this can influence transparency. Although there were indications of industry
37
effect on transparency it was not significant (hypothesis 3). As discussed in section 3.1 an important limitation of my thesis is the relative small sample size. Although previous research had similar or even smaller samples (e.g. Courtis, 1998, 32 companies/2 years) it decreased the chance to find a significant effect for the industry-type effect. The same applies for my control variables.
5.1.1 Practical implications and suggestions for future research
One of the key principles of the IRC SA for reporting is that the board must communicate with the stakeholders in “clear and understandable” language (King III, 2009). However the results show that there has been no improvement since the implementation of the King III code. This is troublesome for the IRC SA and the IIRC. The IRC SA envisioned a transparent report where all relevant stakeholders were properly informed. The results of this thesis however show that the average CEO review is difficult to read due to its textual complexity. So there is still a lack of transparency in published reports, integrated or not. For the IIRC this will be an important aspect while developing Integrated Reporting on a international level.
Another important aspect to consider is that a lot of JSE-listed companies were not able to adapt completely to the new regulation (King III) in the two years after the implementation. There have been several reports from the Big-4 audit firms located in South Africa about this transition. They find difficulties in content that JSE-listed companies face while publishing an integrated report (e.g. Deloitte SA, 2011). How transparent can a company or CEO be without threatening their competitive advantage? Competitors will certainly look at each other’s annual reports. This is an example of one of the considerations the IIRC has to take into account.
It will be interesting to see how Integrated Reporting scores on transparency when companies get more accustomed with this way of reporting. And obviously there will be more years of data to analyze the effect of the implementation of Integrated Reporting.
The tone of the CEO review is an aspect of impression management that is not discussed in this thesis. It is however an interesting aspect of reporting. Researchers focus on the use of negative and positive tones by management to influence the reader (e.g. Schleicher &
Walker, 2010). Future research could focus on the use of biased tones in regard to Integrated Reporting.
38 5.2 Final remarks
An interesting development is the use of web-portals to replace the Portable Document Format (PDF) based reports. The web-portals made my search for CEO reviews a lot easier, especially when I was looking for specific information the web-portals proved to be a blessing. It will be interesting to see how stakeholders will respond to this new format of reporting. I hope that my Thesis will contribute to the development of a worldwide Integrated Reporting framework that facilitates transparency.
39 References
Achmed, K. Courtis, J.K. 1999. Associations between corporate characteristics and disclosure levels in annual reports: a meta-analysis. British Accounting Review, 31: 35-61
Adams, C.A. 2004. The ethical, social and environmental reporting- performance portrayal gap. Accounting, Auditing & Accountability Journal, 20: 731:757
Banerjee, S.B. Iyer, E.S. Kashyap, R.K. 2003, Corporate Environmentalism: Antecedents and Influence of Industry Type, Journal of Marketing, 67: 106-122
Banerjee, S.B. 2008. Corporate Social Responsibility: The Good, the Bad and the Ugly.
Critical Sociology, 34: 51-79
Bhojraj, S. Lee, C.M.C. Oler, D.K. 2003. What’s My Line, A Comparison of Industry Classification Schemes for Capital Market Research. Journal of Accounting Research, 41: 745-774
Brammer, S. Millington A. 2003. The Effect of Stakeholder Preferences, Organizational Structure and Industry Type on Corporate Community Involvement. Journal of Business
Ethics, 45: 213-226
Buzby, S.L. 1975. Company Size, Listed Versus Unlisted Stocks, and the Extent of Financial Disclosure. Journal of Accounting Research, 13: 16-37
Carter, M.R. & May, J. 2001. One Kind of Freedom: Poverty Dynamics in Post-apartheid South Africa. World Development, 29: 1987-2006
Cho, C.H. Roberts, R.W. Patten, D.M. 2010. The language of US corporate environmental disclosure. Accounting, Organizations and Society, 35: 431-443
Cooke, T.E. 1992. The impact of Size, Stock Market Listing and Industry Type on Disclosure in the Annual Reports of Japanese Listed Corporations. Accounting and Business Research, 20: 229-237
Courtis, J.K. 1986. An investigation into annual report readability and corporate risk-return relationships. Accounting and Business Research, 16: 285-294
Courtis, J.K. 1995. Readability of annual reports: Western versus Asian evidence.
40
Courtis, J.K. 1998. Annual report readability: tests of the obfuscation hypotheses.
Accounting, Auditing & Accountability Journal, 11: 459-472
Daily, C.M. Johnson, J.L. 1997. Sources of CEO Power and Firm Financial Performance: A longitudinal Assessment. Journal of Management, 23: 97-117
Deegan, C. 2002. The legitimising effect of social and environmental disclosures – a theoretical foundation. Accounting, Auditing & Accountability Journal, 15: 282-311
Donaldson, T. & Preston, L.E. 1995. The Stakeholder Theory of the Corporation: Concepts, Evidence, and Implications. Academy of Management Review, 20: 65-91
DuBay, W.H. 2007. Smart Language: Readers, Readability, and the Grading of Text. Costa
Mesa
Fama, E.F. French, K.R. 1997. Industry costs of equity. Journal of Financial Economics, 43: 153-193
Freeman, R.E. 1984. Strategic Management: A Stakeholder Approach,
Freemon, J. 1999. Stakeholder Influence Strategies. The Academy of Management Review, 24: 191-205
Fig, D. 2005. Manufacturing amnesia: Corporate Social Responsibility in South Africa.
International Affairs, 81: 599-671
Friedman, M. 1970. The Social Responsibility of Business is to Increase Its Profits. The New
York Times Magazine, 13 September
Hamann, R. (2004). Corporate social responsibility, partnerships, and institutional change: The case of mining companies in South Africa. Natural Resources Forum, 28: 278-290
Healy, P.M. Palepu, K.G. Information asymmetry, corporate disclosure, and the capital markets: A review of the empirical disclosure literature. Journal of Accounting and
Economics, 31: 405-440
Hooghiemstra, R. 2000. Corporate Communication and Impression Management – New Perspectives Why Companies Engage in Corporate Social Reporting. Journal of Business
