The effect of voluntary adoption of integrated reporting on firm value : an empirical study focused on European listed companies

(1)

1

The effect of voluntary adoption of Integrated Reporting on firm

value

An empirical study focused on European listed companies

Name: Sander van Tol Student number: 11399236

Thesis supervisor: Mr. dr. A. Sikalidis Date: June 25, 2018

Word count: 14238

MSc Accountancy & Control, specialization Accountancy Faculty of Economics and Business, University of Amsterdam

(2)

2

Statement of Originality

This document is written by student Sander van Tol who declares to take full responsibility for the contents of this document.

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

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

(3)

3

Abstract

The purpose of this study is to examine the effect of voluntary adoption of Integrated Reporting on firm value. Improved stock liquidity and reduced cost of capital are used as proxies for firm value. Prior research suggests that there is a negative association between voluntary disclosing information and the relative effective spread, used as a measure for stock liquidity. This negative relation means that the stock liquidity is improved, after voluntary information is disclosed, and prior research has found that improved stock liquidity leads to higher firm value. Prior research also suggests that there exists a negative relation between voluntary disclosing information and the cost of capital. Integrated Reporting, and its Framework, can be seen as a form of voluntary disclosure. Prior research also predicts that this relation is stronger for firms with superior performance. This paper’s focus is on European listed companies who adopted Integrated Reporting since 2014 and I compare that sample with firms who have no affiliation with Integrated Reporting. This has resulted in a total sample of 3,852 (2,068) for the first (second) hypothesis.

I find that there exists a negative relation between firms who have adopted Integrated Reporting and the relative effective spread. This indicates that firms who have adopted Integrated Reporting experience improved stock liquidity and that provides support for my first hypothesis. I find, however, no evidence that firms who have adopted Integrated Reporting experience a greater reduction in cost of capital or that the relation is more pronounced for firms with superior performance. In fact, this study finds weak, not significant, evidence that adopting Integrated Reporting leads to a higher cost of capital. A potential explanation for this contradictory result can be found in the critique on the IIRC’ Framework on Integrated Reporting. The Framework has received some debate whether it truly creates value over time. Critics argue that the Framework lacks impact and that it deviates from its original purpose of being the primary purpose and therefore does not lead to the creation of value relevant information. Thus, a possible explanation for not finding a negative association between adopting Integrated Reporting and the cost of capital could be that the Framework indeed lacks impact and therefore is not value-relevant. Another explanation might be the small test sample size for the second hypothesis (165 firm-years) because with a smaller sample size it could be harder to find statistically significant results. This limitation provides avenue for future research.

(4)

4

1. Introduction

In this study, I investigate whether adopting Integrated Reporting is related to firm value. More specifically, this thesis will investigate whether voluntary adoption of Integrated Reporting is associated with a higher level of firm value. I will attempt to answer the following research question:

RQ: Does the voluntary adoption of Integrated Reporting leads to a higher firm

value?

Providing an answer to this research question is important, because there is a growing attention to the use of Integrated Reporting. A recent survey of corporate social responsibility by KPMG shows big rises of Integrated Reporting in Japan, Brazil, Mexico and Spain (KPMG, 2017). However, there is also some debate whether Integrated Reporting truly creates value over time, as claimed by the Framework of the International Integrated Reporting Council (IIRC) (e.g. Brown and Dillard, 2014; Flower 2015). However, there is also criticism on traditional financial reporting. For example, Bobitan and Stefea (2015) argue that financial reporting online cannot provide the most accurate picture of a company to satisfy the needs of stakeholders. Furthermore, they argue that traditional financial reporting does not capture all relevant market information and therefore they challenge whether financial reporting is presenting an accurate overview of the present and future performance of a company. To overcome these shortcomings in traditional financial reporting, the IIRC published in 2013 the International Integrated Reporting Framework (IIRC, 2013a). However, also the IIRC itself has received some critique about whether Integrated Reporting truly creates value over time, or that it is more of an ideologically approach. Therefore, providing an answer to my research question can answer whether Integrated Reporting has actual improvements or that it remains an ideologically approach.

Prior studies on the subject of Integrated Reporting predominantly focus their research on South Africa, because in South Africa publishing an IR is mandatory with a ‘comply or explain policy’. Reimsbach et al. (2017, p. 3) summarizes that ‘’the few studies on the effects of Integrated Reporting have yielded heterogeneous results’’. This study contributes by providing empirical evidence whether voluntary adoption of Integrated Reporting is beneficial for listed European organizations.

(6)

6

To the best of my knowledge, focusing on the effects of voluntary adoption of Integrated Reporting on firm value has never been done before. Prior research has examined the economic consequences of mandatory adoption of Integrated Reporting on firm value, but those studies used samples from South Africa (Lee & Yeo, 2016; Barth et al., 2017). This study’s emphasis is on the voluntary adoption of Integrated Reporting using data from listed European firms. Focusing on voluntary disclosures is an important area to examine, because voluntary disclosure can be seen as a way to influence the minds of future stakeholders by providing financial and non-financial information (Brammer et al., 2006). Combining this with the concept of Integrated Reporting provides insights in whether voluntary adoption of Integrated Reporting leads to higher firm value or that the critique on the Framework is justified.

To examine the research question, I use three hypotheses:

H1: Firms that voluntarily adopt Integrated Reporting experience higher stock liquidity than firms that did not adopt Integrated Reporting, ceteris paribus.

H2a: Firms that voluntarily adopt Integrated Reporting experience a greater reduction in cost of capital than firms that did not adopt Integrated Reporting, ceteris paribus.

H2b: The negative relationship between Integrated Reporting and cost of capital is more pronounced for firms that have superior performance, ceteris paribus.

To test these hypotheses, I use two proxies for firm value, which are stock market liquidity and the cost of capital. I employ the model of relative effective spread as described by Nguyen et al. (2016) to measure stock liquidity and I use the model of Hillier and Clacher (2011) to calculate the cost of capital.

In this thesis, I study the impact of voluntary adoption of Integrated Reporting on firm value and to do so have I divided my sample in two reporting regimes: the pre-adoption period (2011-2013) and the post-adoption period (2014-2016). Furthermore, I use a test sample of companies who adopted Integrated Reporting in the post-adoption period and a control sample of companies who have no affiliation with the concept of Integrated Reporting. By doing this I can compare the reporting periods and analyze whether there is an actual change between the pre-adoption period and the post-pre-adoption period. If there is a change in firm value within the test group, between the pre-adoption period (without an IR) and the post-adoption period (with an IR), that doesn’t exist for the control group then that means that adopting Integrated Reporting is positively associated with firm value and the claims of the IIRC are justified.

(7)

7

I collect data on my sample by using the IIRC Integrated Reporting Examples Database to identify companies who have adopted Integrated Reporting since 2014. Furthermore, I use WRDS Amadeus and DataStream to collect the financial information needed to run the regression model and to provide an answer to my research question.

I make several contributions to the limited research on Integrated Reporting and to the ongoing debate about whether Integrated Reporting truly creates value. First, in line with theory, I find that firms who have adopted Integrated Reporting experience a higher stock market liquidity than firms who have no affiliation with the concept. Second, I show that firms who have adopted Integrated Reporting experience no greater reduction in the cost of capital. Furthermore, I demonstrate that having superior performance and adopting Integrated Reporting does not have a significant effect on the cost of capital. These findings contribute to the debate and criticism on the Framework. On the one hand, I show that adopting Integrated Reporting leads to providing more relevant information, which leads to a lower information asymmetry and, thus, a higher stock market liquidity. On the other hand, adopting Integrated Reporting does not lead to a lower cost of capital. However, I also find weak evidence that the control sample does not experiences a reduction in the cost of capital. Therefore, I have mixed results but I do provide evidence that Integrated Reporting has actual improvements (improved stock liquidity) and not just remains an ideologically approach. This finding can help persuade companies to adopt Integrated Reporting and therefore it can help the IIRC in becoming the new corporate reporting norm.

The remainder of the paper is organized as follows. Section 2 briefly explains the background of Integrated Reporting, discusses voluntary disclosure and develops the main hypotheses. The sample selection and research design is described in section 3. Section 4 presents the empirical findings and the discussion of these findings and section 5 summarizes and provides a conclusion. Section 6 contains the references used in this thesis. Finally, in the last section the Appendices, with variable definitions, are presented.

(8)

8

2. Literature review & hypothesis development

In this section, I discuss the underlying theoretical concepts of my research question. In section 2.1, I introduce the background of Integrated Reporting and its framework, in section 2.2 I discuss the theory on voluntary disclosure and its theoretical relation to firm value, and in section 2.3 I present my hypotheses.

2.1 Background

Bobitan and Stefea (2015) describe why financial reporting, which is prepared under generally accepted accounting principles (GAAP) or under International Financial Reporting Standards (IFRS), alone cannot provide the most accurate overview of a company to satisfy the needs of stakeholders. They argue that the traditional financial reporting is limited in the sense that it does not capture all relevant, financial and non-financial, market information and they therefore challenge whether it represents an accurate picture of the present and future performance of firms.

Corporate social responsibility (CSR) reporting can be the answer to this challenge. A definition of CSR commonly used in the management literature comes from Davis (1973, p. 312), who defines corporate social reporting as ‘’the firm’s considerations of, and response to, issues beyond the narrow economic, technical, and legal requirement of the firm to accomplish social and environment benefits along with traditional economic gains which the firm seeks’’. The adoption of CSR reporting has, over the past decades, grown from a small, undervalued notion into a complex, multifunctional concept which is increasingly important in much of nowadays’ corporate decision making (Cochran, 2007). However, recently, these CSR reports have been criticized for the fact that they are disconnected from the firm’s strategy, its business model, and financial performance (Barth et al., 2017). Furthermore, the current corporate reporting’s focus is on past financial performance and on a relatively narrow account of the value-creation process which has led to a gap between what investors need to know about businesses and what corporate reports are telling them (KPMG, 2016; IIRC, 2011). In recent years, there is also a growing concern about the increasing complexity and decreasing relevance of current corporate reports (FRC, 2009). The current reporting has shortcoming including the lengthiness, the average length of an annual report is 204 pages (KPMG, 2016, p. 12), and ‘’can obscure critical information rather than aid understanding’’ (IIRC, 2011, p. 4). The shortcomings also include the complexity, and that many reports clutter the reporting market (Feng et al., 2017). The FRC defines cluttering as

(9)

9

‘’undermining the usefulness of annual reports and accounts by obscuring important information and inhibiting a clear understanding of the business and the issues that it faces’’ (FRC, 2011, p. 2). The financial reporting has also been perceived as confusing and fragmented and as irrelevant or untimely (IIRC, 2011). To overcome these shortcomings the International Integrated Reporting Council (IIRC) has developed a new framework: The International Integrated Reporting Framework (IIRC, 2013a).

The IIRC was founded in 2010 at the initiative of two sustainability reporting organizations: The Global Reporting Initiative (GRI) and the Prince’s Accounting for Sustainability project (A4S) (Flower, 2014). The IIRC is a global coalition composed of regulators, investors, companies, standard setters, the accounting profession and NGOs (IIRC, 2013a). Their aim is to guide organizations on communicating different pieces of information that is needed by their stakeholders to assess the organization’s long-term prospects. The IIRC believes that Integrated Reporting brings together this material information about an organization and that it provides a clear and concise representation of how an organization creates and sustains value over time (IIRC, 2011). To be more specific, the IIRC believes that communicating about value creation should be the next step in the evolution of corporate reporting and defines an integrated report as ‘’a concise communication about how an organization’s strategy, governance, performance and prospects, in the context of its external environment, lead to the creation of value over the short, medium and long term’’ (IIRC, 2013a, p. 7). Key in this definition is that the focus is not only on the short term, but also on the medium and long-term of an organization.

Integrated Reporting has been created to better understand the range of measures that help organizations focus on the long-term value creation. Nowadays, organizations create value not only for their shareholder, but also for other important stakeholders, such as the society as a whole. Central to this idea is the proposition that value is increasingly shaped by factors other than financial factors (EY, 2014). The value creation concept is the core concept of the Integrated Reporting Framework and can be established through the activities and outputs of the organization. The IIRC aims to provide an integrated report, which provides insights about the resources and relationships, the so-called capitals, used in organizations. The six capitals categorized in the framework are: financial, manufactured, intellectual, human, social and relationship, and natural (IIRC, 2013a; IIRC, 2013b):

(10)

10

 Financial: the pool of funds available to a firm, including both debt and equity finance. The IIRC (2013b, p. 6) describes financial capital ‘’as a medium of exchange that releases its value through conversion into other forms of capital’’.

 Manufactured: this capital represents manufactured objects that are available to an organization that contribute to the production of goods or provision of services. This capital includes equipment, technology, buildings, machines and infrastructure.

 Intellectual: this capital represents organizational, knowledge-based intangibles that function as a key element in an organization’s future earnings potential. It includes intellectual property such as copyrights, software, patents, and organizational capital such as procedures and protocols, systems and tacit knowledge.

 Human: this capital represents the individual capabilities, competencies and experiences of the firm’s employees and managers.

 Social and relationship: the IIRC (2013a, p. 12) describes this capital as ‘’the institutions and the relationships within and between communities, groups of stakeholders and other networks, and the ability to share information to enhance individual and collective well-being’’. It includes shared norms, key stakeholder relationships, licenses to operate, and intangibles associated with the brand name and reputation that a firm has developed.  Natural: this capital represents resources, which can be used to provide a return such as all

renewable and non-renewable environmental resources.

(11)

11

Some of these capitals are captured within the current voluntary CSR reports, such as environmental issues, but as mentioned before, these CSR reports are criticized for the fact that they are disconnected from the firm’s strategy, its business model, and financial performance. Integrated Reporting combines these capitals into one report to picture the firm’s ability to create value in a broader sense (Barth et al., 2017) and therefore the capitals can be instrumental in improving an organization’s long-term financial performance. By reflecting how the organization, by using the capitals, creates value over time, the objective of the IR Framework of the IIRC is to foster the process of integrated thinking (IIRC, 2013a). In the Framework, the IIRC defines integrated thinking as ‘’the active consideration by an organization of the relationships between its various operating and functional units and the capitals that the organization uses or affects leading to integrated decision-making actions that consider the value over the short, medium and long term’’ (IIRC, 2013a, p. 2). The IIRC’s long-term vision is a world in which Integrated Reporting is the new corporate reporting norm, as a single report, whereas the cycle of integrated thinking and reporting will improve sustainability and financial stability (IIRC, 2013a). However, Brown and Dillard (2014) challenge this vision and argue that integrated reporting remains more an ideologically approach than that it provides actual improvements. They report that it seems that integrated reporting is designed to serve the interest of finance capital rather than that it serves in the public interest. Furthermore, Flower (2015) found some more weaknesses in the International Integrated Reporting Framework of the IIRC. He found that the Framework deviates from its original purpose of being the primary report by stating that ‘’an integrated report maybe either a standalone report or be included as a distinguishable part of another report’’ (IIRC, 2013a, p. 8)’’, that the Framework excluded sustainability, and that it has a lack of impact (Flower, 2015). Besides these weaknesses of the Framework, Integrated reporting is so far only mandatory in South Africa and Brazil (Robertson, 2015). So, when companies decide to publish an integrated report it means that they voluntary disclose information.

2.2 Voluntary disclosure and firm value

Over the years, there has been done a lot of research on the topic of voluntary corporate disclosure. Brammer et al. (2006) view corporate disclosure as a way to influence the perceptions of stakeholders regarding the future financial forecasts of the firm. They also argue that managers are motivated to make disclosures because they are afraid for the fallout if they do not (e.g. legislative

(12)

12

pressure and reduction in discretion over future investment opportunities). Healy & Palepu (1993) make another argument for voluntary disclosure. They argue that managers could improve the credibility of their financial reporting through voluntary disclosure, which includes discussing the company’s long-term strategy, providing non-financial information, and discussing the relation between leading indicators of the business and its future profits. However, they also argue that there is a trade-off between the amount a firm discloses to improve the firm value and the amount it withholds to maximize the firm’s competitive advantage (Healy & Palepu, 1993). In a more recent study of Healy & Palepu (2001), they argue that there is information asymmetry between managers and outside stakeholders on their firm’s expected performance. Information asymmetry means that there are information differences and conflicting incentives between managers and outside stakeholders that could result in lower firm performance. According to Verrecchia (2001) and Brammer et al. (2006), firms can reduce the information asymmetry problem by committing to the highest level of voluntary disclosures. Voluntary socially responsible behavior can also help firms avoid government regulation, litigation costs and therefore reduce the compliance costs. However, the current annual reports are sometimes very limited in providing socially responsible behavior whereas CSR reports and integrated reports provide much more information demonstrating the firms’ special effort and commitment to improving transparency, which could result in lower costs and higher firm value (Brammer et al., 2006). Merton (1987) describes another way in which corporate disclosure affects the cost of capital. He suggests that greater disclosure increases the investors’ awareness of the firm’s existence, which results in a larger investor base that improves risk-sharing and reduces the cost of capital. These arguments show the importance of voluntary disclosure in reducing the information asymmetry and thereby affecting firm value, which in turn can reduce the cost of external financing.

Voluntary disclosures have a signaling role. By voluntary disclosing information, the firm shows confidence in their performance, which sends a positive signal to investors. In the case of poor performance, voluntary disclosing information allows firms to offer explanations on how and why this happened and therefore voluntary disclosures improve transparency of long-term performance (Dhaliwal et al., 2011).

(13)

13

2.3 Hypothesis development

In this subsection, I discuss how liquidity and cost of capital affect firm value and how it is related to Integrated Reporting. I also present my hypotheses.

2.3.1 Improved stock liquidity

Diamond and Verrecchia (1991), and Kim and Verrecchia (1994) hypothesize that extensive voluntary disclosure reduces information asymmetry. They argue that the investors could be confident that stock transactions occur at fair value prices and therefore it could increase the liquidity in the firm’s stock. Healy et al (1999) find support for this hypothesis in that firms that increase voluntary disclosure experience significant increases in stock prices. They also find that the increase in stock prices is unrelated to current earnings performance. Gelb and Zarowin (2002) examine the association between voluntary disclosure and the informativeness of stock prices and they compare this association for firms with high versus low disclosures. Their findings indicate that firms with greater disclosure experience higher stock prices and future earnings relative to firms with lower disclosures.

Fang et al. (2009) examine the relation between stock liquidity and firm value and they find that firms with liquid stocks have better performance and thus higher firm value. Nguyen et al. (2016) extend the research of Fang et al. (2009) by examining the effect of stock market liquidity on firm performance in the Australian market. They also find a positive relation between stock liquidity and firm value.

The aim of the IIRC is to guide organizations on communicating different pieces of information that is needed by their stakeholders to assess the organization’s long-term prospects and to create and sustain value over time. Integrated Reporting is thus about providing more information in a broader sense and can be seen as a high level of disclosure. Based on the aforementioned findings I expect that voluntary adoption of Integrated Reporting, and thus an increased level of disclosure, results in higher stock liquidity compared to firms that do not adopt Integrated Reporting and therefore my first hypothesis is:

H1: Firms that voluntarily adopt Integrated Reporting experience higher stock liquidity than firms that did not adopt Integrated Reporting, ceteris paribus.

(14)

14

2.3.2 Lower cost of capital

Theory suggests that information asymmetry exists when some investors are better informed than others (Kim & Verrecchia, 1994; Verrecchia, 2001). Investors with less information are more likely to price-protect themselves and are less willing to trade. This results in an increase in the bid-ask spread and transactions costs, which ultimately leads to a higher cost of capital (Verrecchia, 2001). Voluntary disclosures could lower the firm’s cost of capital by reducing the information asymmetry between the firm’s management and investors. If the costs of capital are reduced then the market value will increase because the discount rate of unexpected future cash flows will be lower because that discount rate is based on the cost of capital (Gordon et al., 2010). Hail (2002) investigates the impact of voluntary corporate disclosures on the cost of equity capital for 73 Swiss firms and finds a negative association between voluntary disclosures and the cost of capital. Dhaliwal et al. (2011) also examine this potential benefit associated with voluntary CSR disclosure activities: a reduction in a firm’s cost of equity capital. They used a sample of 294 firms that issued a total of 1190 CSR reports. They hypothesize and find that CSR-disclosing firms with superior performance achieve a reduction in the cost of equity capital after they initiate a CSR report. However, they state that their results only apply to financial and non-financial information as long as the information is value-relevant. Francis et al. (2005) examined the disclosure consequences on cost of capital for a sample of 672 observations from 34 countries outside of the United States and find that higher disclosure levels will lead to a lower cost of external, both debt and equity, financing.

The IIRC (2013a) believes that an integrated report provides a clear and concise representation of how an organization creates and sustains value over time. Furthermore, an Integrated Report provides insights about six capitals of an organization. These capitals can be instrumental in improving an organization’s long-term financial performance and reflect how the organization has created value over time. By providing this information, Integrated Reporting can on the one hand reduce the information asymmetry, and thus reduce the cost of capital. On the other hand, publishing an Integrated Report with a long-term focus can lead to increases in the investors’ awareness of the firm and thus result in a larger investor base that improves risk-sharing among investors as described earlier.

Hence, I expect that Integrated Reporting reduces the information asymmetry between management and stakeholders, which improves the investors’ awareness and risk-sharing. I expect

(15)

15

that this leads to a lower cost of capital and a higher firm value. Furthermore, I expect that this result is more pronounced for Integrated Reporting firms with superior performance, following the results of Dhaliwal et al. (2011):

H2a: Firms that voluntarily adopt Integrated Reporting experience a greater reduction in cost of capital than firms that did not adopt Integrated Reporting, ceteris paribus.

H2b: The negative relationship between Integrated Reporting and cost of capital is more pronounced for firms that have superior performance, ceteris paribus.

(16)

16

3. Sample selection and research design

In section 3.1, I will discuss how I have selected my sample and how I collected the data. In section 3.2, I discuss how I measure firm value. The models for testing my first and second hypothesis are presented in section 3.3 and 3.4, respectively.

3.1 Sample selection

Integrated Reporting is a relatively new concept, the Framework only exists since December 2013 (IIRC, 2013a) and therefore there has not been much empirically data. This also means that the majority of data on Integrated Reporting is only available starting from 2014. This thesis’ focus is on archival data from the Integrated Reporting Examples Database and uses available data at WRDS Amadeus and DataStream. Furthermore, I focus on European countries that have published an Integrated Report. For a full list of countries included in this thesis, see Table 1.

In this thesis, I study the impact of voluntary adoption of Integrated Reporting on firm value and to do so have I divided my sample in two reporting regimes: the pre-adoption period (2011-2013) and the post-adoption period (2014-2016). By doing this I can compare the reporting periods and analyze whether there is an actual change between the pre-adoption period and the post-adoption period. If there is a change in firm value within the test group, between the pre-adoption period (without an IR) and the post-pre-adoption period (with an IR), that doesn’t exist for the control group then that means that adopting Integrated Reporting is positively associated with firm value and the claims of the IIRC are justified.

I started with collecting data of companies who published an integrated report according to the IIRC IR Examples Database. This database contains examples of firms that have published an integrated report around the world since December 2013, when the Framework was first introduced by the IIRC (IIRC, 2013a). The database contains 175 companies who, according to the IIRC, published an Integrated Report in Europe. I verified this sample by checking whether I could find their actual annual report/integrated report and whether the IIRC or Integrated Reporting is mentioned in the report. I excluded firms that were not within an European country and/or not listed on a stock exchange. This has resulted in finding 54 firms that met the initial criteria for test sample.

I obtained data for this sample, the test sample, from Wharton Research Data Service (WRDS) Amadeus (Bureau van Dijk). Amadeus contains information on around 21 million

(17)

17

companies across Europe and I used the information on ‘’Financials’’ and ‘’Stock Annual’’. ‘’Financials’’ provided me with financial information on companies such as balance sheet data, income sheet data or for example incorporation date. ‘’Stock Annual’’ contains annual stock data such as EPS, shares outstanding, share prices and market capitalization. Within Amadeus, these are two different databases and therefore I merged the datasets using Stata. If the dataset was incomplete, I hand collected the appropriate data to complete the dataset.

Next, I obtained data from DataStream for information on the cost of capital and the relative quoted bid-ask spread. Since the format of DataStream is different from the format of Amadeus, I used the VLOOKUP function in Excel to reshape the data in the appropriate format. Once I had done that, I merged the dataset from DataStream with the dataset from Amadeus by using Stata. Finally, I excluded firms with missing observations; mostly because the data was not available in DataStream and it was impossible for me to hand collect the missing data. Within the data, there were outliers and to exclude the extreme influence of a few values and to create a more normally distributed sample I trimmed the data at the 1st and 99th percentile of the distribution. My final test sample contains 208 firm-year observations.

To control for the impact of potentially confounding concurring events I used a control sample, which I also obtained from WRDS Amadeus Financials. I excluded unconsolidated observations and observations from countries that are not included in the test sample. I only collected firms from countries that are included in my test sample because this lowers the concern that any changes in firm value are driven by changes in sample composition, as country specifics can have an influence on my results (Byard et al., 2011). I also excluded observations that were mentioned on the IIRC IR Examples Database to make sure that my control sample has no affiliation to Integrated Reporting. Next, I obtained data from WRDS Amadeus on Stock Annuals and I merged the data using Stata. I dropped the observations that could not be matched with Amadeus’ firm observations on Financials. Just like my test sample, I also obtained data from DataStream for information on the cost of capital and the relative bid ask spread. Since the format of DataStream was different from the format of Amadeus, I undertook the same procedures as described before. After that, I merged the dataset from DataStream with the dataset from Amadeus by using Stata. Finally, I excluded firms with missing observations that are needed to analyze my hypotheses. I also dropped outliers, by trimming the data at the 1st and 99th percentile of the

(18)

18

distribution, to exclude the extreme influence of a few values and to create a more normally distributed control sample. My final control sample contains 3.644 firm-year observations.

Table 1, panel A, provides the distribution of my test and control sample per country and panel B of Table 1 provides the distribution of my test and control sample per industry. The test and control sample represents firms from 12 countries within 8 (9 for control sample) major industry groups, composed according to the Standard Industrial Classification (SIC) code (Siccode, 2018). The test sample contains heavy concentrations from the United Kingdom, Spain and the Netherlands whereas the control sample contains heavy concentrations from the United Kingdom, France and Germany. For the industry distributions, the test sample contains heavy concentration in industries such as Finance, insurance and real estate and Manufacturing whereas the control sample contains heavy concentrations in Finance, insurance and real estate, Manufacturing and Services.

(19)

19

Table 1. Sample description Panel A: Distributions of the test and control firms by country

Country Control sample (no. of firm-years) Test sample (no. of firm-years)

Belgium 82 6 Finland 102 4 France 1,079 24 Germany 375 4 Italy 108 2 Luxembourg 2 4 Netherlands 95 32 Poland 9 2 Spain 176 34 Sweden 228 2 Switzerland 161 6 United Kingdom 1,227 88

Total number of firm-years 3,644 208

Panel B: Distributions of the test and control firms by industry

Industry Control sample (no. of firm-years) Test sample (no. of firm-years)

Construction 154 11

Finance, insurance, real estate 1,297 76

Manufacturing 788 67 Mining 115 12 Retail 182 6 Services 643 15 Public utilities 263 19 Wholesale 177 2 Other 25 0

Total number of firm-years 3,644 208

Notes: The table shows the distributions of the test and control firms across countries and industries. The test sample consists of 208 firm-years of European listed firms that voluntarily adopted the IIRC’ Framework on Integrated Reporting published in December 2013. The control sample contains of 3,644 firm-year observations of European listed firms who did not adopt the IIRC’ Framework on Integrated Reporting. Both samples hold observations for the fiscal years 2011 to 2016. I identify my test and control firms based on firm’s financial data retrieved from WRDS Amadeus and DataStream.

(20)

20

3.2 Measuring firm value

To capture the impact of voluntary adoption of Integrated Reporting on firm value I use two measures: stock liquidity and cost of capital. I focus on these two proxies for firm value because prior research has found a direct impact of stock liquidity and cost of capital on firm value. For example, Nguyen et al. (2016) examined the effect of stock liquidity on firm performance and they found a positive relation between stock liquidity and firm value. In addition, Fang et al. (2009) find that higher stock market liquidity is associated with higher firm performance. Furthermore, Hail (2002) and Dhaliwal et al. (2011) examine the relation between voluntary disclosures, such as Integrated Reporting, and the cost of capital and they find a negative association between the two, meaning that voluntary disclosing information leads to a lower cost of capital.

My first proxy is the liquidity measure, which I measure as the relative quoted bid-ask spread. This spread is calculated as follows:

𝑅𝑒𝑙𝑆𝑝𝑟𝑒𝑎𝑑 = 𝐴𝑠𝑘𝑃𝑟𝑖𝑐𝑒 − 𝐵𝑖𝑑𝑝𝑟𝑖𝑐𝑒 1

2 ∗ (𝐴𝑠𝑘𝑃𝑟𝑖𝑐𝑒 + 𝐵𝑖𝑑𝑃𝑟𝑖𝑐𝑒)

According to Nguyen et al. (2016) this measures the arithmetic mean of the relative quoted spreads over firm i’s financial year. Fang et al. (2009) state that the effective bid-ask spread contains non-normality and therefore they use the natural logarithm of the relative effective spread (Log_RelSpread). This variable is constructed to be negatively related to stock liquidity. The ask and bid price data was collected from DataStream and then the relative bid-ask spread was calculated by hand, by using the aforementioned formula.

The second proxy is use is the percentage change in weighted average cost of capital (WACC). I obtain the percentage change in WACC (%∆WACC) by first calculating the WACC itself. The WACC, explained by Hillier and Clacher (2011) is calculated using the following formula:

𝑊𝐴𝐶𝐶 = ( 𝐸

𝑇𝑉) ∗ 𝑅ₑ + ( 𝐷

𝑇𝑉) ∗ 𝑅𝑑∗ (1 − 𝑇𝑐)

In this formula, E/TV and D/TV represent the total equity and debt, collected from Amadeus’ Financials, divided by total value (equity + debt) of the company, respectively. Re represents the

cost of equity, which is according to Hillier and Clacher (2011), calculated as: 𝑅ₑ =𝐷₁

(21)

21

In this formula, D1 means next year’s dividends per share and P0 represents this year’s stock price.

g is the dividend growth rate, which is the arithmetic average of the growth in dividend from year

t+1 – t (Thomson Reuters, 2010). This arithmetic average is then used for the entire period. Multiplying the cost of equity with the E/TV results in the weighted average cost of equity capital (Re). The cost of debt (Rd) is calculated by dividing interest paid, also collected from DataStream,

by total debt. Tc represents the corporate tax rates per year per country and are collected from

KPMG (2018). Multiplying D/TV with the cost of debt and the corporate tax rate results in the weighted average cost of debt. Summing these weighted averages totals the weighted average of capital (WACC). Next, I calculate the percentage difference in WACC for each firm for the years 2012-2016. I have excluded 2011 there are no values for WACCt-1, as that represents 2010. The

percentage change in WACC is calculated using the following formula:

%∆𝑊𝐴𝐶𝐶𝑡 =

WACC_t – WACC_t−1 𝑊𝐴𝐶𝐶𝑡

∗ 100%

3.3 Testing the effect of voluntary adoption of Integrated Reporting on stock liquidity (H1)

My first hypothesis examines if firms that voluntarily adopt Integrated Reporting experience higher stock liquidity than firms that did not adopt Integrated Reporting. I estimate the following regression model to examine the relationship between Integrated Reporting and higher stock liquidity, as a proxy for firm value:

Log_RelSpreadi,t = β0 + β1IRi,t + β2Log_Bvtai,t + β3Log_Agei,t +

Industryj + Countryi,t + Yeari,t + Leveragei,t + ROAi,t + ↋i,t

The dependent variable Log_RelSpread is explained in section 3.2. Based on prior literature and the fact that the natural logarithm of the relative bid-ask spread is constructed to be negatively related to stock liquidity, I would expect a negative coefficient on IR to support this hypothesis (Fang et al., 2009). A negative coefficient on IR means that if firms publish an Integrated Report the natural logarithm of the relative spread will decrease. Since Log_RelSpread is negatively constructed to stock liquidity (Fang et al., 2009) this means that a decrease in Log_RelSpread leads to a higher stock liquidity and, consequently, a higher firm value (Fang et al., 2009; Nguyen et al., 2016). This variable is the dependent variable used in Equation (1).

β0 is the intercept that captures the change around voluntary adoption of Integrated Reporting for the control firms. The key coefficient β1 captures the difference in stock liquidity between the

(22)

22

test and control firms and, thus, the impact of voluntary adoption of Integrated Reporting, where

IR is a dummy variable that equals 1 if firm i publishes an Integrated Report and 0 otherwise. Log_Bvta and Log_Age are the natural logarithm of the book value of total assets and firm age,

respectively. To control for confounding concurrent events I use the following control variables:

Industryj which is an indicator value that equal 1 if a firm belongs to a certain industry, and 0 otherwise, Countryi,t with an indicator value 1 if it is a certain country, 0 otherwise and Yeari,t with an indicator value 1 if it is a certain year and 0 otherwise. Leveragei,t is measured as the total book value of equity divided by the total book value of assets of firm i at fiscal year-end t (Fang et al., 2009). Return on assets (ROAi,t) is included as a measure of profitability (Dhaliwal et al., 2011). See section 7.2 for a full overview of the variable definitions and how they are collected.

3.4 Testing the effect of voluntary adoption of Integrated Reporting on cost of capital (H2)

My second hypothesis (H2a) expects that publishing an integrated report leads to a lower cost of capital. I estimate the following regression model to examine if adopting Integrated Reporting leads to a lower cost of capital, as a proxy for firm value:

%∆WACCi,t+1 = β0 + β1IRi,t + ∆Sizei,t + Betai,t + ∆Leveragei,t + ∆MBi,t

+ ∆LTGi,t + Countryi,t + Industryi,t + Yeari,t + ↋i,t

In this model %∆WACCi,t+1 represents the percentage change in firm i’s cost of capital from year

t to year t-1. The control variables also adopt the change form for the exception of Beta because in the data I retrieved from DataStream this variable is every year the same per company. A negative coefficient on IR would support H2 (Dhaliwal et al., 2011). β0 is the intercept that captures the change around voluntary Integrated Reporting adoption for the control firms. The key coefficient β1 captures the difference in cost of capital between the test and control firms and, thus, the impact of voluntary IR adoption where IR is a dummy variable that equals 1 if firm i publishes an Integrated Report and 0 otherwise. I expect a negative coefficient on IR, which indicates that publishing an Integrated Report leads to a lower cost of capital. The control variables are derived from the research of Dhaliwal et al. (2011). They include firm size (SIZE) and market-to-book ratio (MB) because prior research (Fama and French, 1992) finds that expected returns are negatively related with firm size and positively with the market-to-book ratio. BETA is included to control for systemic risk and Leverage is included because debt servicing provides a monitoring role and is important because debt holders demand greater disclosure. It is measured as the total

(23)

23

book value of equity divided by the total book value of assets of firm i at fiscal year-end t. LTG is the long-term annual growth rate in earnings per share and is included because prior research found that the long-term growth rate in earnings per share is positively associated with cost of equity capital (Gebhardt et al., 2001). Industry, Year and Country are included to control for the impact of certain industries, years or countries on my results.

My second hypothesis examines if the percentage change in WACC is related to publishing an Integrated Report and because of this change I have excluded year 2011 from this sample. Furthermore, the observations on the long-term growth rate were limited. I used DataStream to collect the data, but many observations were missing. This has resulted in a lower sample for my second hypothesis. My second hypothesis contains 2,068 firm-years, of which 165 (1,903) belong to the test (control) sample.

Furthermore, I also examine whether the negative relationship between firm performance and cost of capital is more pronounced for firms that have adopted Integrated Reporting (H2b). I estimate the following regression model:

∆%WACCi,t+1 = β0 + β1IRi,t + β2HighPerform + β3IR*HighPerform +

∆Sizei,t + Betai,t + ∆Leveragei,t + ∆MBi,t + ∆LTGi,t + Countryi,t

+ Industryi,t + Yeari,t + ↋i,t

The variables in this model are the same as in H2a for the exception of HighPerform, which is based on the return of equity. I use an indicator variable that equals 1 if a firm’s performance is higher than the industry mean and 0 otherwise. I expect a negative coefficient on the interaction and I expect that the interaction (β3) is more negative than Integrated Reporting alone (β1), which means that superior firms with an Integrated Report experience a greater reduction in WACC.

(24)

24

4. Empirical findings

In this section, I will discuss the empirical findings. I will start with the descriptive characteristics, mean comparison and correlations in section 4.1. Next, I discuss the multivariate regression analysis in section 4.2 and I conclude this section with a discussion of the findings in section 4.3.

4.1 Descriptive characteristics, mean comparison and correlations

Table 2, Panel A provides descriptive statistics for the variables included in Equation (1) for the full sample. Panel B of Table 2 provides descriptive statistics for the variables included in Equation (2) and (3). The mean and median are almost equal and the difference between p25 and the median and p75 and the median is similar indicating that the variables are normally distributed. Table 3 provides the mean comparison between the test and control sample. The table provides an overview of the mean for the test and control sample in the pre- and postadoption period and explores potential differences between the two samples. It also provides the total mean per variable for the test and control sample.

For the test sample the mean of Log_RelSpread and %∆WACC are -6.372 and -0.002 in the pre-adoption period and -6.980 and 0.012 for the post-adoption period, respectively. For the control sample the mean of Log_RelSpread and %∆WACC are -4.770 and -0.025 in the pre-adoption period and -6.106 and -0.004 in the post-pre-adoption period, respectively. The difference-in-difference results are significant (two tailed p ≤ 0.01) for Log_RelSpread, suggesting that firms in the test sample experience a significant greater reduction in the natural logarithm of the relative bid-ask spread. This reduction could mean higher stock market liquidity and thus higher firm value according to Fang et al. (2009) and Nguyen et al. (2016). However, this comparison does not take into account correlated factors that could potentially influence these results and therefore the results only provide preliminary evidence on the first hypothesis. These results should be interpreted with restraint.

The costs of capital for the test firms is negative (-0.002) in the pre-adoption period and increases in the post-adoption period (0.012), suggesting that adopting Integrated Reporting leads to a slight increase in cost of capital. However, the control sample also experiences an increase in the cost of capital between the pre- (-0.025) and post-adoption period (-0.004). Moreover, these differences between the test and control sample are not significant and do therefore not provide preliminary evidence on the second hypothesis.

(25)

25

In Table 3, I also compare other financial variables across the test and control sample to identify whether and how these samples differ. I find that the natural logarithm of book value (Log_Bvta), Leverage and Beta significantly (two tailed p ≤ 0.01) decreases for both the test and control sample. Return on assets decreases slightly for the test sample in the post-adoption period whereas ROA increases for the control sample in the same period. I also find that the market-to-book ratio significantly increases for the test sample. The change in long-term growth rate is not statistically different between the two test samples suggesting that they both experience a similar growth rate.

Overall, the results in Table 3 indicates that my test and control sample differ significantly. The table provides preliminary evidence on the first hypothesis and it slightly indicates that the cost of capital does not decreases in the post-adoption period of the test sample compared to the control sample (H2). However, the results on H2 are not significantly. Furthermore, the results in this table are not controlled for country, year and industry factors and should be interpreted with caution since these factors can have an impact on the results.

Table 4 presents the correlation between the liquidity measure (Log_RelSpread), the Integrated Reporting indicator variable and all control variables used in my model. Table 4 also presents the correlations between the cost of capital measure (%∆WACC), the Integrated Reporting indicator variable and all control variables used in my models. The correlation matrix includes the Spearman correlation (top right) and the Pearson correlation (bottom left) and shows that the results of both correlation methods are similar in value, sign and significance.

As shown in Table 4, the relative effective spread, Log_RelSpread has significantly negative Pearson (Spearman) correlations with IR, Log_Bvta, Log_Age, Leverage and ROA. In other words, firms with liquid stocks (a decrease in Log_RelSpread) tend to have adopted IR, have a higher book value, are older, have less debt in their capital structure and have a higher return on assets. These findings are consistent with the findings of Fang et al. (2009). The cost of capital measure, %∆WACC, has significant negative Pearson (Spearman) correlations with ∆Log_Size,

∆Leverage and ∆MB. These negative correlations might suggest that firms who experience a

reduction in their cost of capital have a greater change in their size, leverage and market-to-book value. The %∆WACC has no significant correlation with Integrated Reporting suggesting that adopting Integrated Reporting per se does not lead to a reduction in cost of capital.

(26)

26

The correlation matrix suggests that there exists correlation between the variables and therefore I test for multicollinearity by using the Variance Inflation Factor (VIF). The rule of thumb with the VIF is the rule of 10, which means that any VIF of 10 is a sign of severe multicollinearity (O’brien, 2007). When the VIF reaches that threshold, attempts to reduce the collinearity must be undertaken. Table 5 shows the Variance Inflation Factor, and thus multicollinearity, for my hypotheses and none of the variables have a VIF higher than 2. Therefore, even though the correlation among the variables are significant there exists no severe multicollinearity in my model. The VIF for the dummy variables on industry, year and country are not tabulated, but are all below 10.

(27)

27

Table 2. Descriptive characteristics Panel A: Descriptive characteristics of H1

Variables Mean Median St. Dev Minimum p25 p75 Maximum

Log_RelSpread -5.297 -5.422 1.616 -9.384 -6.466 -4.086 -0.233

Log_Bvta 20.591 20.687 2.259 13.159 19.019 22.162 27.878

Log_Age 3.310 3.296 0.913 0 2.708 4.025 5.793

Leverage 0.538 0.558 0.184 0.035 0.417 0.676 1.028

ROA 0.043 0.044 0.073 -0.484 0.017 0.076 0.278

Panel B: Descriptive characteristics of H2

Variables Mean Median St. Dev Minimum p25 p75 Maximum

%∆WACC -0.009 -0.005 0.271 -1.241 -0.116 0.070 1.797 ∆Size 0.106 0.111 0.263 -0.905 -0.042 0.271 0.984 Beta 0.709 0.690 0.474 -0.920 0.400 1.000 2.440 ∆Leverage -0.001 -0.004 0.045 -0.165 -0.025 0.019 0.207 ∆MB 0.046 0.036 0.599 -3.379 -0.117 0.237 3.500 ∆LTG -0.010 -0.005 0.147 -0.664 -0.063 0.044 0.664

Notes: The table provides summary statistics for the main variables for firm-year observations from fiscal years 2011-2016. Panel A reports summary statistics for the main variables for the sample of my first hypothesis (n = 3.852). Panel B reports summary statistics for the main variables for the sample of my second hypothesis (n = 2.068). The sample for my second hypothesis is lower because this hypothesis captures the change in variables and since fiscal year 2010 is not included, there are no changes measureable from fiscal year 2011 to 2010, and hence the lower sample size. Indicator variables are not included. All numbers are rounded up to third decimal place. Variable definitions are shown in the Appendix.

(28)

28

Table 3. Mean comparison between test and control firm-years Panel A: Mean comparison for H1

Variable Test firm-years Control firm-years t-value

(differences) Preadoption period 2011-2013 n = 108 Postadoption period 2014-2016 n = 100 Total mean n = 208 Preadoption period 2011-2013 n = 2,418 Postadoption period 2014-2016 n = 1,226 Total mean n = 3,644 Difference-in-differences between test and control sample Log_RelSpread -6.372 -6.980 -6.664 -4.770 -6.106 -5.219 3.49*** Log_Bvta 23.442 23.285 23.366 19.923 21.437 20.432 5.71*** Log_Age 3.517 3.647 3.580 3.219 3.445 3.295 0.75 Leverage 0.637 0.595 0.617 0.523 0.554 0.534 2.80*** ROA 0.041 0.034 0.038 0.037 0.056 0.437 2.51**

Panel B: Mean comparison for H2

Variable Test firm-years Control firm-years t-value (differences)

Preadoption period 2012-2013 n = 65 Postadoption period 2014-2016 n = 100 Total mean n = 165 Preadoption period 2012-2013 n = 677 Postadoption period 2014-2016 n = 1,226 Total mean n = 1,093 Difference-in-differences between test and control sample %∆WACC -0.002 0.012 0.007 -0.025 -0.004 -0.011 0.16 ∆Size 0.078 0.066 0.071 0.174 0.073 0.109 2.05** Beta 0.871 0.828 0.847 0.672 0.758 0.758 1.92* ∆Leverage 0.000 -0.003 -0.002 -0.006 0.002 -0.001 1.44 ∆MB -0.220 -0.099 -0.147 0.207 -0.017 0.063 3.53*** ∆LTG -0.009 0.004 -0.001 -0.022 -0.005 -0.011 0.19

Notes: This table provides mean differences between the test and control firms and the pre- and post-adoption periods for each group. The results are produced by using the difference-in-difference estimation of Stata and provide results in the differences between the test and control group and between the pre- and post-adoption period. The difference-in-difference results and their significance are reported in the last column. Indicator variables are not included in this comparison. The mean is estimated by a linear regression and the standard errors are robust. The test sample consists of 208 (165 for H2) firm-years of European listed firms that voluntarily adopted the IIRC’ Framework on Integrated Reporting published in December 2013. The control sample contains of 3,644 (1,903 for H2) firm-year observations of European listed firms who did not adopt the IIRC’ Framework on Integrated Reporting. Both samples hold observations for the fiscal firm-years

(29)

29

2011 to 2016, with the exception for H2, which measures the change between t and t-1. I identify my test and control firms based on firm’s financial data retrieved

from WRDS Amadeus and DataStream. The t-values are reported in the last column. *** indicate significance level at the 1% level,

** indicate significance level at the 5% level, * indicate significance level at the 10% level.

(30)

30

Table 4. Correlation matrix

Panel A: Correlations between stock market liquidity and the adoption of Integrated Reporting (H1)

Variables Log_Relspread IR Log_Bvta Log_Age Leverage ROA

Log_RelSpread 1 -0.178*** -0.691*** -0.131*** 0.150*** -0.179*** IR -0.175*** 1 0.196*** 0.058*** 0.046*** -0.024 Log_Bvta -0.689*** 0.201*** 1 0.246*** 0.359*** 0.007 Log_Age -0.129*** 0.062*** 0.243*** 1 0.124*** 0.027* Leverage -0.162*** 0.053*** 0.374*** 0.156*** 1 -0.271*** ROA -0.201*** -0.019 0.109*** 0.094*** -0.093*** 1

Panel B: Correlations between cost of capital and the adoption of Integrated Reporting (H2)

Variables %∆WACC IR HighPerform ∆Log_Size Beta ∆Leverage ∆MB ∆LTG

%∆WACC 1 0.033 0.013 -0.285*** -0.015 -0.424*** -0.148*** -0.045 IR 0.017 1 -0.059*** -0.046** 0.050** 0.002 -0.049** 0.030 HighPerform 0.004 -0.025 1 0.134*** 0.041* -0.124*** 0.034 0.013 ∆Log_Size -0.230*** -0.036 0.136*** 1 -0.042* -0.149*** 0.666*** 0.176*** Beta -0.014 0.050*** 0.070*** -0.044** 1 0.000 -0.061*** 0.007 ∆Leverage -0.297*** 0.011 -0.095*** -0.129*** 0.001 1 -0.176*** -0.037* ∆MB -0.114*** -0.056** 0.045** 0.568*** -0.079*** -0.151*** 1 0.141*** ∆LTG -0.024 0.015 0.734 0.146*** 0.010 -0.024*** 0.151*** 1

Notes: The table provides correlations for the main variable for firm-year observations from fiscal year 2011-2016. Panel A reports correlations for the main variables for the sample of my first hypothesis (n = 3.852). Panel B reports correlations for the main variables for the sample of my second hypothesis (n = 2.068). The sample for my second hypothesis is lower because this hypothesis captures the change in variables and since fiscal year 2010 is not included, there are no changes measureable from fiscal year 2011 to 2010, and hence the lower sample size. Top right shows the correlations according to Spearman and bottom left shows the correlations according to Pearson. All numbers are rounded up to third decimal place. Variable definitions are shown in the Appendix.

*** indicate significance level at the 1% level; ** indicate significance level at the 5% level; * indicate significance level at the 10% level.

(31)

31 Variables Variables IR 1.10 IR 1.92 Log_Bvta 1.72 HighPerform 1.17 Log_Age 1.24 ∆Log_Size 1.59 Leverage 1.32 Beta 1.10 ROA 1.11 ∆Leverage 1.05 ∆MB 1.57 ∆LTG 1.07

Mean VIF 1.30 Mean VIF 1.18

Notes: This table shows the Variance Inflation Factor, which is an indicator for multicollinearity. All values are below the threshold of 10 and therefore there exists no severe multicollinearity in the models. The values for the indicator variables on country, year and industry are not tabulated but are all below 10.

4.2 Multivariate regression analysis

In this section, I examine the effect of voluntary adoption of Integrated Reporting on stock liquidity and the weighted average cost of capital. I regress the natural logarithm of the relative bid-ask spread (Log_RelSpread) on a dummy variable indicating voluntary adoption of Integrated Reporting (IR), the Log_Bvta, Log_Age and the control variables Leverage, ROA and dummy variables for country, industry and year to control for confounding concurrent events. I also regress the percentage change in the weighted average cost of capital (%∆WACC) on a dummy variable indicating voluntary adoption of Integrated Reporting (IR) and the control variables ∆Log_Size, Beta, ∆Leverage, ∆MB, ∆LTG and dummy variables for country, industry and year. Next, I regress the same model but include a dummy variable for HighPerform to examine whether firms with a higher than the industry-mean of return on equity experience a greater reduction in cost of capital. Results are reported in Table 6.

In Table 6, Panel A, the regression analysis for my fist hypothesis is conducted. The R2 -adjusted is 0.587, indicating that 58.7% of Log_RelSpread is explained by the independent variables. Table 6, Panel A, shows that voluntarily adopting Integrated Reporting affects the natural logarithm of the relative bid-ask spread. More specifically, in Panel A, the estimated coefficient of interest is on IR, which is negative and significant at the 1% level. A negative significant relation means that firms who adopt Integrated Reporting experience a decrease in the natural logarithm of the relative bid-ask spread. Since the natural logarithm of the relative bid-ask spread is constructed to be negatively related to stock market liquidity (Fang et al., 2009), adopting Integrated Reporting leads to a decrease in Log_RelSpread. In turn, a lower

(32)

32

relative effective spread is associated with a higher market liquidity and a higher market liquidity means higher firm value (Fang et al., 2009; Nguyen et al., 2016). Thus, this outcome provides evidence that firms who voluntarily adopt Integrated Reporting experience higher stock liquidity than firms that did not adopt Integrated Reporting and, thus, experience an increase in firm value. As described in section 3.3, the intercept captures the change around voluntary Integrated Reporting for the control firms. The estimated coefficient on this is positive (3.268) and significant (p ≤ 0.01), indicating that firms who do not voluntarily adopt Integrated Reporting experience an increase in the Log_RelSpread, resulting in a lower stock liquidity and a lower firm value. Additionally, the positive coefficient on Leverage is significantly related to Log_RelSpread providing evidence that firms who have a lower leverage also have a lower relative effective spread and thus a higher firm value. When comparing this outcome with Table 3, it provides support that adopting Integrated Reporting (test sample) leads to a decrease in leverage (from 63.7% to 59.5%) as compared to the increase in leverage for the control sample (from 52.3% to 55.4%) and thus that adopting Integrated Reporting is beneficial for firms. Overall, the regression results in Table 6, Panel A, confirm the preliminary results of the mean comparison in Table 3 and provide evidence that adopting Integrated Reporting leads to higher stock liquidity.

In Table 6, Panel B, the regression analysis for my second hypothesis is conducted. The R2-adjusted is 0.182, indicating that 18.2% of %∆WACC is explained by the independent variables. Table 6, Panel B, shows that voluntarily adopting Integrated Reporting does not lead to a decrease in the cost of capital. More specifically, in Panel B, the coefficient on IR is positive (0.008) and not significant (t-stat = 0.34). This relation means that there exists no evidence that adopting Integrated Reporting is associated with a decline in the cost of capital and therefore I am unable to support hypothesis that firms who voluntarily adopt Integrated Reporting experience a greater reduction in cost of capital (H2). However, the intercept, which captures the change around voluntary Integrated Reporting for the control firms, is also positive (0.064) and that would indicate that the control sample also experiences an increase in cost of capital that is even greater than for firms who have adopted Integrated Reporting (0.008). However, these results are not significant and therefore do not provide conclusive evidence on this assertion. Regarding the control variables, ∆Size, and ∆Leverage are negatively related with the percentage change in cost of capital, which indicates that if a firms become larger, or the debt structure becomes smaller the cost of capital will decrease.

In Table 6, Panel C, the regression analysis for the sub-hypothesis of the second hypothesis is conducted. The R2-adjusted remains 0.182, indicating that 18.2% of %∆WACC is

(33)

33

explained by the independent variables. Table 6, Panel C, shows that firms who adopted Integrated Reporting and have superior performance don’t experience a greater reduction in the cost of capital. More specifically, the coefficient of interest is the interaction between Integrated Reporting and high performance (IR*HighPerform). The estimated coefficient on this interaction term is 0.011 (t-stat = 0.22) and that indicates that adopting Integrated Reporting and having superior performance does not have a significant effect on a decline in the cost of capital. This means that I am unable to support this hypothesis. However, the coefficient on

HighPerform is 0.006 (t-stat = 0.53), which indicates that having a high performance on its own

also does not lead to a significant decrease in the cost of capital. Moreover, the results even suggest, although not significant, that adopting Integrated Reporting and/or having superior performance lead to a small increase in the cost of capital.

(34)

34

Dependent variables Log_RelSpread (H1) %∆WACC (H2a) %∆WACC (H2b)

Explanatory variables Cf. Cf. Cf. IR -0.234** (0.095) [-2.46] 0.008 (0.025) [0.34] 0.004 (0.027) [0.16] Log_Bvta -0.458*** (0.010) [-45.31] Log_Age 0.112*** (0.019) [6.19] Leverage 0.654*** (0.109) [6.01] ROA -2.156*** (0.245) [-8.78] ∆Log_Size -0.277*** (0.036) [-7.72] -0.277*** (0.036) [-7.79] Beta -0.018 (0.017) [-1.02] -0.018 (0.017) [-1.01] ∆Leverage -2.005*** (0.154) [-12.98] -1.999*** (0.155) [-12.89]

(35)

35 ∆MB -0.011 (0.017) [-0.64] -0.010 (0.016) [-0.63] ∆LTG 0.012 (0.048) [0.25] 0.012 (0.047) [0.25] HighPerform 0.006 (0.013) [0.50] IR*HighPerform 0.011 (0.048) [0.24] Intercept 3.268*** (0.225) [14.53] 0.064 (0.036) [1.80] 0.059 (0.038) [1.59]

Country/Industry/Year dummies Yes Yes Yes

R2-adj. 0.587 0.182 0.182

N 3,852 2,068 2,068

Prob>X2 0.000 0.000 0.000

Notes: This table examines the effect of voluntary adoption of Integrated Reporting on stock liquidity (H1) and cost of capital (H2) and reports the result of the relevant OLS regression. For Panel A, the sample consists of 3,852 firm-year observations from fiscal years 2011-2016, where the dependent variable is the natural logarithm of the relative bid-ask spread. For Panel B and C, the sample consists of 2,068 firm-year observations from fiscal years 2011-2016, where the dependent variable is the percentage change in weighted average cost of capital (%∆WACC). Hypothesis H2b includes an interaction variable between superior performance and Integrated Reporting. All continuous variables are trimmed at the 1st_{and 99}th_{percentile of their distributions to mitigate the effect of outliers. The estimated coefficients are followed by robust standard errors}

() and t-statistics []. All numbers are rounded up to third decimal place, with an exception for the t-statistic. Variable definitions are shown in the Appendix. *** indicate significance level at the 1% level.

** indicate significance level at the 5% level. * indicate significance level at the 10% level

The effect of voluntary adoption of integrated reporting on firm value : an empirical study focused on European listed companies