The relationship between the level of integratedness of corporate reporting and financial performance. Exploring the moderating effect of investor protection

(1)

The relationship between the level of integratedness

of corporate reporting and financial performance

Exploring the moderating effect of investor protection

Master Thesis in Economics 2019/2020

Corporate Finance & Control

Date: 06-08-2020

Tjeu Mutsers - s4229266

(2)

2 Abstract

Over the years, companies have increased the disclosure of integrated information. Previous studies suggest that integrated reporting has a positive effect on financial performance, but the evidence is mainly based on samples of South African firms. In this study, the direct relationship between the level of integratedness of corporate reporting and financial performance in both the short and long term is investigated using a worldwide scope. Furthermore, the moderating effect of investor protection on the previously specified relationship is investigated. Panel data regressions are conducted in order to analyze a balanced dataset of 3,332 companies from 52 countries for the period of 2014-2018. This study is one of the first to provide empirical evidence for a positive relationship between the level of integratedness of corporate reporting and financial performance, in both the short and long term. Furthermore, the evidence shows that investor protection negatively moderates this relationship in the short term. Additional tests also provide evidence for the negative moderating effect of investor protection on this relationship in the long term. These results are valuable for corporate managers who intend to increase corporate value and for investors who seek for new investment opportunities.

(3)

3

3.2. Variables ... 19 3.2.1. Dependent variable ... 19 3.2.2. Independent variables ... 20 3.2.3. Control variables ... 21 3.3. Model ... 22 3.4. Testing assumptions ... 24 Chapter 4. Results ... 26 4.1. Descriptive statistics ... 26 4.2. Testing hypotheses... 27 4.3. Robustness checks ... 30

Chapter 5. Conclusion and discussion ... 33

References ... 37

Appendix 1: Breusch-Pagan / Cook-Weisberg test ... 44

(4)

4

Chapter 1. Introduction

This study investigates the relationship between the level of integratedness of corporate reporting and financial performance, and whether this relationship is moderated by investor protection. Over the years, firms behave more as responsible corporate citizens, implementing practices and actions linked to corporate social responsibility (Costa & Menichini, 2013). This trend has altered the field of corporate reporting. Corporate reporting has traditionally focused on financial measures. There has been a shift towards reporting both financial and non-financial information (Perrini, 2006). Until recently, both aspects of information were usually presented in separate reports. Since companies increasingly incorporate aspects of CSR in their strategy, financial and non-financial aspects have become intertwined with one another. Thus, publishing stand-alone reports of financial and non-financial aspects disconnects these intertwined aspects (Jensen & Berg, 2012).

The increased demand for the disclosure of non-financial information and its integration with financial information has led to the development of integrated reporting (Cohen & Karatzimas, 2015). Within integrated reports, financial and non-financial reporting is combined in a single document (Adams, 2017). It connects information on environmental and social aspects with financial measures, which potentially improves the representation of the relationship between a company and its stakeholders (Vitolla, Raimo, Rubino & Garzoni, 2019). Sustainability reporting is mostly conducted in the form of stand-alone reports, having weak connections to annual financial reports. Integrated reporting, in turn, is aimed at providing a single document in which environmental and social thinking are integrated in a company’s strategy in a comprehensive manner (Cortesi & Vena, 2019). According to the IIRC, an international coalition that developed the universally applicable Integrated Reporting (<IR>) Framework, integrated reporting is based on the concept of integrated thinking (IIRC, 2013). Integrated thinking is defined as “the active consideration by an organization of the relationships between its various operating and functional units and the capitals that the organization uses and affects” (IIRC, 2013). In turn, the way in which integrated thinking is embedded in the day-to-day decision-making of a company can be expressed in the level of integratedness of corporate reporting. The level of integratedness is not concerned with whether or not companies publish integrated reports (Busco, Malafronte, Pereira & Starita, 2019). Instead, it is concerned with the way in which integrated thinking is incorporated in an organization and how integrated thinking is reported in alignment with the <IR> Framework (Martinez, 2016; Mertins, Kohl & Orth, 2012). It reflects the capacity to communicate and

(5)

5

demonstrate that a company integrates financial, environmental and social dimensions into its everyday decision-making (Busco et al., 2019).

The concept of integrated reporting has increased in popularity considering the application in corporate reporting (Eccles, Krzus & Ribot, 2015). However, South Africa is yet the only country around the globe that requires listed companies to conform to integrated reporting on a so-called “apply or explain” basis (Dumay, Bernardi, Guthrie & Demartini, 2016). This makes it de facto mandatory. The <IR> Framework is principles-based, providing opportunities for individual judgment and flexibility in preparing integrated reports (Lee & Yeo, 2016). Except from South Africa, integrated reporting is only adopted on a voluntary basis (Loprevite, Ricca & Rupo, 2018). Arguing from previous research conducted on financial reporting standards, differences in the effects between mandatory and voluntary adoption might be expected. Christensen, Lee, Walker and Zeng (2015) showed that differences in incentives between the mandatory and voluntary adoption of IFRS results in differences in accounting quality. Voluntary adoption significantly improved accounting quality and mandatory adoption did not. Higher accounting quality, in turn, results in cheaper external financing (Christensen et al., 2015). Therefore, differences in the effects of the level of integratedness of corporate reporting on financial performance between mandatory and voluntary adopters can be expected.

According to Barth, Cahan, Chen and Venter (2017), integrated reporting is positively associated to corporate financial performance. In their study, integrated reporting is measured in terms of quality. Among other components, quality depicts the connectedness (i.e. integration) of reported information, which is similar to the level of integratedness of corporate reporting. However, this study only examined South African companies. In order to determine whether the findings of Barth et al. (2017) hold for companies that voluntarily adopt integrated reporting standards, the research scope should be extended. This thesis focuses on this gap and aims to answer the question whether the level of integratedness of reporting is positively related to corporate financial performance, in both the short and long term. Regarding financial performance, the distinction between short and long-term is made since shareholder theory suggests that companies tend to have a strong focus on short-term financial performance (Danielson, Heck & Shaffer, 2008). Besides maximizing shareholder value in the short run, the concept of integrated reporting also pursues long-term value creation (Adams, 2015; IIRC, 2013).

(6)

6

Despite the adoption of accounting and reporting standards such as IFRS, country-specific factors still have a vital influence on accounting and reporting in general (Glaum, Schmidt, Street & Vogel, 2013). In academic literature, it is widely argued that the institutional environment has a significant effect on accounting and reporting (Gebhardt & Novotny-Farkas, 2011; Holthausen, 2009; Morais & Curto, 2008). Therefore, institutional differences between countries likely affect the relationship between the level of integratedness of reporting and corporate financial performance. More specifically, this research focuses on investor protection as a component of the institutional environment. Studies suggest that investor protection potentially moderates the relationship between the level of integratedness of reporting and corporate financial performance. Companies headquartered in countries with strong investor protection face less problems regarding information asymmetry (García-Sánchez & Noguera-Gámez, 2017). Therefore, the potential value added by increasing the level of integratedness of corporate reporting in countries with strong investor protection could be less significant than in countries with weak investor protection (Durnev & Kim, 2005). By reducing information asymmetry, companies can decrease the cost of capital and improve financial performance (García-Sánchez & Noguera-Gámez, 2017; Zhou, Simnett & Green, 2017).

This thesis aims at filling the previously argued knowledge gaps. Academic literature suggests that the level of integratedness of corporate reporting is positively associated with financial performance (Barth et al., 2017; Lee & Yeo, 2016). However, conclusions are mainly drawn based on samples that focus on South Africa. Furthermore, it is suggested that investor protection moderates the relationship between integrated reporting and information asymmetry (Durnev & Kim, 2005; García-Sánchez & Noguera-Gámez, 2017). Though, the effect on the relationship between the level of integratedness of corporate reporting and financial performance seems to be underexposed and calls for future research. Therefore, the research question of this thesis is formulated as follows:

What is the effect of the level of integratedness of reporting on corporate financial performance and how is this relationship moderated by the level of investor protection?

By answering this question, the goal of this thesis is to determine whether increasing the level of integratedness of corporate reporting can actually enhance value for companies, and whether this relationship is affected by differences in the level of investor protection. In order to answer the research question, panel data regressions are performed using a worldwide sample of listed firms from 52 countries on a timeframe of 2014 to 2018. The data used in this

(7)

7

study is retrieved from the Refinitiv database, which can be accessed through Thomson Reuters Eikon, and the World Bank.

This study has several theoretical and practical implications. Based on a worldwide sample, it is one of the first studies to show the positive relationship between the level of integratedness of corporate reporting and financial performance, in both the short and long run. The results implicate that the claim of the IIRC that integrated thinking and reporting could enhance short and long-term value creation, is supported by empirical evidence. Furthermore, this study is the first to show the negative moderating effect of investor protection on the relationship between the level of integratedness of corporate reporting and financial performance. These results are useful for both corporate managers and investors. For managers who intend to improve corporate performance, the results of this study suggest that increasing the level of integratedness of corporate reporting could potentially contribute in achieving that goal. Regarding practical implications for investors, the results suggest that companies with high levels of integratedness of corporate reporting might be interesting opportunities to invest in. Due to these higher levels of integratedness of corporate reporting and their corresponding financial performance, investors could potentially generate increasing returns.

The remainder of this thesis proceeds as follows. In the next chapter, the theoretical framework underlying this thesis is discussed. It elaborates on the most important constructs of this thesis by reviewing current literature. Based on these constructs and their assumed relationships with one another, hypotheses are developed. The third chapter discusses the research method used in this thesis. In the fourth chapter, the results are presented. Finally, the fifth chapter elaborates on the conclusion and discussion.

(8)

8

Chapter 2. Literature review and hypotheses development

This chapter discusses the most important concepts of this thesis by means of reviewing literature regarding these concepts. It elaborates on integrated reporting, the institutional environment and more specifically investor protection. By connecting these concepts, and thus assuming potential relationships between the concepts, hypotheses are developed.

2.1. Integrated reporting

In order to cope with the increased interests of stakeholders in social and environmental information, the disclosure of non-financial information has increased within corporate reporting over the years (Perrini, 2006). This has resulted in stand-alone financial and non-financial reports at first. Due to the increased complexity and length of these stand-alone reports, companies and their stakeholders have looked for ways to combine financial and non-financial information in one single report. This resulted in practices which have become known as integrated reporting (De Villiers, Unerman & Rinaldi, 2014). Integrated reporting is focused on presenting a holistic view of a company, connecting the dots between financial, social and environmental issues and future targets. In contrast to the common short-term perspective of traditional financial reporting, the performance metrics in integrated reporting have a stronger focus on medium or long-term value creation (Jensen & Berg, 2012). According to the concept of integrated reporting, companies are required to adopt long-term value-creating strategies. Therefore, the proponents of integrated reporting claim that a sustainable society that meets the needs of both present and future generations could be achieved by complying to the <IR> Framework (Eccles & Serafeim, 2011).

In 2010 an international council, the International Integrated Reporting Council (IIRC), has been founded. This council advocates the adoption of integrated reporting as a worldwide reporting standard (De Villiers et al., 2014). The IIRC is a coalition of a wide variety of parties such as regulators, investors, companies and NGOs. They share the view that corporate reporting should evolve towards communication regarding long-term value creation (IIRC, 2013). In 2013 the IIRC published the <IR> Framework1, which serves as a guide that highlights the main principles of integrated reporting and the concepts underlying them (Flower, 2015). The IIRC’s mission is to encourage public and private organizations to adopt the combination of integrated reporting and thinking as the standard into their businesses (IIRC, 2020).

1

The IIRC is currently revising the <IR> Framework. However, the approval and launch of the revised

framework is set a date beyond the publication of this thesis. Therefore, this thesis takes the first publication of framework in 2013 as a starting point.

(9)

9

According to the IIRC (2013), an integrated report is based on six capitals. These are financial, manufactured, intellectual, human, social and relationship, and natural capital. Through their relationships with one another, long-term value creation can be achieved. The topics serve as the building blocks for a company’s core business model, in which their inputs are converted to outputs, through a company’s business activities. The higher the capacity of the capitals in terms of availability, affordability and quality, the more it can affect long-term value creation (IIRC, 2013). In Figure 1, the value-creation process is visually presented.

Figure 1: The value-creation process (IIRC, 2013)

For a report to be classified as an integrated report, it should not solely mention all six capitals, but also their linkages have to be considered (IIRC, 2013). By showing the linkages between the capitals, the integration between financial and sustainability performance becomes more visible (Eccles & Saltzman, 2011). In other words, the level of integratedness of a corporate report depends on the integration of financial and non-financial disclosure, which can be measured by analyzing the integration of the six capitals. By maximizing the integration of the six capitals, thus achieving a high level of integratedness, strategic decision-making resulting in long-term value creation can be achieved (IIRC, 2013).

(10)

10 2.2. Integratedness of reporting and financial performance

One of the main goals of companies and their investors is value maximization, which can be achieved when companies perform economically sound (Shleifer & Vishny, 1988). As mentioned before, integrated reporting aims at creating value for both the company and its shareholders (Busco, Frigo, Riccaboni & Quattrone, 2013). Corporate financial reporting and disclosing might help to solve information asymmetry and agency conflicts between companies and investors, which results in a decline in cost of capital (Healy & Palepu, 2001). Looking more specific into integrated reporting, Zhou et al. (2017) argue that aligning corporate reporting with the <IR> Framework developed by the IIRC, is related to a decrease in the cost of capital. Compared to other reporting standards, integrated reporting provides an additional amount of information to capital providers, lowering cost of capital, and thus increasing financial performance (Zhou et al., 2017).

Integrated reporting should create value in the short, medium, and long term (IIRC, 2013). However, it is questionable whether this statement holds in practice. By focusing on long-term performance, short-term goals might have to be sacrificed (Eccles & Serafeim, 2011). Churet and Eccles (2014) found that there is no significant relationship between integrated reporting and financial performance, measured as return on invested capital (ROIC). Integrated reporting practices did not correlate with companies achieving an increasing ROIC. On the other hand, companies that implemented these practices did not perform worse on average (Churet & Eccles, 2014). This measure is an accounting-based measure and therefore, it can be regarded as a short-term financial performance measure (Margolis & Walsh, 2001).

On the other hand, Lee and Yeo (2016) conclude that a higher score on integrated reporting, which represents a higher level of integratedness of corporate reporting, is positively associated with ROA. Thus, companies with a significant implementation of integrated reporting practices outperform companies with less integrated reporting practices, also in the short term. However, this effect became stronger over the years, which suggests that the long-term effects of integrated reporting outweigh the short-term effects (Lee & Yeo, 2016). A possible explanation for the weaker short-term effects could be the additional costs of implementation of integrated reporting practices. In order for integrated reporting to be implemented effectively, the educational level regarding integrated reporting and thinking within a company has to be improved (Vitolla, Raimo & Rubino, 2019). Therefore, the benefits of integrated reporting might only become visible after a certain period of time. A

(11)

11

similar effect was found previously in studies on CSR reporting and financial performance. The additional costs of implementation of CSR reporting are incurred directly, while the benefits are only visible after a certain period of time (Tsoutsoura, 2004). Comparing these results to integrated reporting, one could argue that the costs in the short term of implementing increased levels of integratedness outweigh the benefits in the short term. This suggests that companies with a high level of integratedness of corporate reporting perform worse in the short term compared to companies with low levels of integratedness. By having a strong focus on short-term financial performance, companies might be hampered in their ability to implement the changes needed to account for the value creation that is fundamental to integrated reporting (Cheng, Green, Conradie, Konishi & Romi, 2014).

Based on the previous discussion, it can be stated that existing literature is inconclusive regarding the relationship between the level of integratedness of corporate reporting and short-term financial performance. Empirical evidence suggests both a positive (Lee & Yeo, 2016) and a negative (Tsoutsoura, 2004) relationship. However, since the IIRC (2013) explicitly states that integrated thinking and reporting can also create value in the short term, it is assumed that the relationship between the level of integratedness and short-term financial performance is positive. Therefore, the following hypothesis is stated:

H1a: The level of integratedness of corporate reporting is positively related to short-term financial performance.

While investors are mainly interested in short-term performance (Danielson et al., 2008), integrated reporting also focuses on how long-term value creation and performance can be achieved (IIRC, 2013). The framework is a communication tool that should contribute to long-term performance (Churet & Eccles, 2014). Following Nichols and Wahlen (2004), it can be expected that future performance is a reflection of current performance. Current period earnings contain information on which the forecasts of future earnings are based. These forecasts, in turn, are reflected in the dividend expectations and the present value of these dividend expectations determines the share price. The share price is a market measure and reflects long-term performance (Nichols & Wahlen, 2004). In other words, the short-term effect of integratedness of corporate reporting on financial performance should also be reflected in long-term performance.

Looking more closely into the relationship between integrated reporting and long-term financial performance, some empirical evidence differs from Nichols and Wahlen’s (2004)

(12)

12

reasoning. For instance, Lee and Yeo (2016) found that integrated reporting is used to communicate a company’s ability to create value for potential providers of financial capital. They suggest that improved disclosure, arising from the implementation of integrated reporting, reduces information asymmetry as well as agency costs between companies and the providers of external financial capital. This, in turn, results in increased values of Tobin’s q, suggesting that long-term financial performance is improved with the implementation of integrated reporting (Lee & Yeo, 2016). Similar arguments are provided by Barth et al. (2017), who argue that integrated reporting is positively associated with expected future cash flows. Because integrated reporting improves the ability of investors to estimate future cash flows, investors’ forecasting is more accurate, decreasing the problem of information asymmetry. This increase in expected future cash flows results in an increase of Tobin’s q (Barth et al., 2017), suggesting once more that long-term financial performance is improved (Margolis & Walsh, 2001). Overall, the benefits of value creation in the long run should eliminate the additional costs that have to be made in the short run regarding the implementation of integrated reporting (Hoque, 2017).

This empirical evidence regarding the relationship between integrated reporting and long-term financial performance is mainly based on a sample of South African firms, where integrated reporting is de facto mandatory (Dumay et al., 2016). Based on the empirical evidence, a positive relationship between the level of integratedness of corporate reporting and long-term financial performance is suggested. Increased levels of integratedness of corporate reporting seem to result in companies increasing their performance in the long run. Therefore, the following hypothesis is formulated:

H1b: The level of integratedness of corporate reporting is positively related to long-term financial performance.

2.3. Institutional environment: investor protection

Empirical evidence arguing that economic outcomes, such as economic growth and investment behavior, can partially contribute to differences in national institutions has increased significantly over the years (Henisz, 2002). To the core of their existence, companies are embedded in their institutional environments (Scott, Scott & Meyer, 1994). Contrary to the paradigm that companies are autonomous and rational actors, institutional theory suggests that companies make decisions based on rules and structures that are developed in certain institutional environments. The institutional environment sets out the rules that determine how business is done in a certain country (Chan & Makino, 2007).

(13)

13

More specific to the field of corporate reporting, the effectiveness of reporting standards differs across institutional environments (Renders & Gaeremynck, 2007). The goal of adopting reporting standards around the globe, such as IRFS, is to have uniformity in standards and rules for financial reporting. This makes cross-country comparisons more feasible and transparent. However, it is unlikely that this could be achieved when institutional forces are significantly different across countries (Holthausen, 2009). Moreover, managerial incentives for voluntarily adopting such reporting standards depend on the characteristics of the institutional environment. For instance, companies in strong institutional environments are forced to implement globally acknowledged reporting standards to a greater extent (Renders & Gaeremynck, 2007).

One of the most widely reported components of a country’s institutional environment that partially explains the effectiveness of integrated reporting is investor protection (García‐ Sánchez & Noguera‐Gámez, 2018; Jensen & Berg, 2012). Investor protection is defined as the protection of minority shareholders against the expropriation of controlling shareholders (La Porta, Lopez-de-Silanes, Shleifer & Vishny, 1998). In financing their operations, companies can access external financial resources provided by outside investors. However, protection is necessary since these outside investors bear the risk that the returns on their investments might be kept in the organization by the controlling shareholders. In case of the complete absence of investor protection, controlling shareholders could confiscate corporate profits (La Porta et al., 1998). In countries with strong investor protection laws and enforcement, external investors are more willing to invest in companies. This results in more developed financial markets. On the other hand, in countries with weak laws and enforcement, the development of financial markets is hampered (La Porta, Lopez-de-Silanes, Shleifer & Vishny, 2002). It is found that in weak investor protection countries, companies are less incentivized to adopt globally used reporting standards such as IFRS, because of opportunism by management in those countries (Renders & Gaeremynck, 2007). High levels of investor protection are associated with an increase in the quantity of integrated reporting (Jensen & Berg, 2012). Moreover, investor protection seems to facilitate increased levels of integratedness of corporate reporting (Busco et al., 2019)

The legal protection of investors in a country is said to have a major influence on the development of financial markets (La Porta et al., 2002). Companies that are headquartered in developed financial markets have more opportunities to achieve increasing rates of economic growth (King & Levine, 1993), and at a higher pace (Haidar, 2009). Citing Himmelberg,

(14)

14

Hubbard and Love (2002), investor protection refers “collectively to those features of the legal, institutional, and regulatory environment - and characteristics of firms or projects - that facilitate financial contracting between inside owners (managers) and outside investors” (Himmelberg et al., 2004, p.2). Shleifer and Wolfenzon (2002) argue that, based on previous studies, investor protection seems to be related to a larger number of listed firms, higher valuation of these firms compared to their assets, and more valuable stock markets. In countries where laws protect outside investors and where enforcement of laws is at a sufficient level, companies face fewer difficulties in accessing external finance. This provides them with a greater opportunity to perform significantly well (La Porta, Lopez-de-Silanes, Shleifer & Vishny, 1997). Furthermore, in countries where investor protection is weak, the concentration of inside ownership is found to be stronger, due to less opportunities of accessing external finance. A higher concentration of inside ownership, in turn, is positively associated with cost of capital (Himmelberg et al., 2004). In other words, companies in countries with weak investor protection laws tend to have a higher cost of capital, harming their financial performance. Besides that, low levels of investor protection are associated with entrenched inside shareholders that are fixated on pursuing their own benefits, which harms external investors (Lan, Wang & Yang, 2012).

2.4. The moderating effect of investor protection

It is hypothesized that the level of integratedness of reporting and financial performance are positively related, both in the short and long term. This thesis also investigates the presence of a potential moderating effect, with investor protection being the variable moderating the relationship between the level of integratedness of corporate reporting and financial performance. Based on previous studies, it can be argued that a negative moderating effect of investor protection on the relationship between the level of integratedness of corporate reporting and financial performance might be expected.

García-Sánchez and Noguera-Gámez (2017) conclude that there is a negative relationship between integrated reporting and information asymmetry. Increased integrated disclosure, or a higher level of integratedness of corporate reporting, results in an improved distribution of private and public information among minority investors. By reducing information asymmetry, the cost of capital is reduced (García-Sánchez & Noguera-Gámez, 2017). This, in turn, results in improved financial performance (Zhou et al., 2017). Similarly, Martinez (2016) concludes that there is a statistically significant negative relationship between the level of alignment with the <IR> Framework and the bid-ask spread. This

(15)

15

suggests that reporting more in line with the <IR> Framework reduces information asymmetry (Martinez, 2016), which in turn can lead to a reduced cost of capital (Lambert, Leuz & Verrecchia, 2012).

When analyzing the level of investor protection, it has been concluded that strong and well-enforced minority shareholder rights give less incentives to corporate insiders to gain private control benefits. This results in improved disclosure of integrated information, since there are less incentives to hide information for outsiders (García-Sánchez & Noguera-Gámez, 2017). Companies from countries with a high level of investor protection have more opportunities in accessing external finance (Durnev & Kim, 2005). Where minority shareholders rights are well-protected and enforced, external investors are more willing to invest in companies. This results in a higher firm valuation and better corporate performance (La Porta, Lopez-de-Silanes, Shleifer & Vishny, 2000). Weak levels of investor protection provide greater incentives for managers to adopt corrupt accounting practices. This could potentially increase the cost of capital and harm financial performance (Houqe, van Zijl, Dunstan & Karim, 2012).

Considering the moderating effect of investor protection on the relationship between the level of integratedness of corporate reporting and financial performance, a negative effect is expected. In weak investor protection countries, the use of corporate governance mechanisms such as integrated reporting is scarce. Due to its scarcity, these mechanisms have more value in weak investor protection countries, than in strong investor protection countries where these mechanisms are more common (Durnev & Kim, 2005). Strong investor protection countries face less market frictions and information asymmetry. Therefore, the potential value of increased levels of integratedness in order to reduce information asymmetry and improve financial performance could be lower in strong investor protection countries, as compared to weaker investor protection countries (García-Sánchez & Noguera-Gámez, 2017). Since information asymmetry is less of an issue in strong investor protection countries, the value added of a higher level of integratedness of corporate reporting is expected to be lower than in weaker investor protection countries. This suggests that part of the effect of the level of integratedness of corporate reporting on financial performance is absorbed by the level of investor protection.

To the best of the author’s knowledge, current literature does not distinguish between short and long-term effects of the previously discussed moderating effect of investor

(16)

16

protection. Therefore, it is expected that the effect of the level of integratedness on financial performance is negatively moderated by the level of investor protection, both in the short and long run. This proposition is formulated in the following hypotheses:

H2a: The relationship between the level of integratedness of corporate reporting and short-term financial performance is negatively moderated by the level of investor protection.

H2b: The relationship between the level of integratedness of corporate reporting and long-term financial performance is negatively moderated by the level of investor protection.

(17)

17

Chapter 3. Research method

In this chapter the research method is discussed. Firstly, the data and sample used in this thesis are discussed. Then, the relevant variables are explained. Afterwards, the research model, in terms of a conceptual and econometric model, is presented. Finally, assumptions are tested to justify the method of analysis conducted.

3.1. Data and sample

As mentioned in the introduction, this thesis takes a worldwide scope regarding the relationship between integratedness of corporate reporting and financial performance. Widely cited literature focuses strongly on integrated reporting and company financial performance in South Africa (Barth et al., 2017; Lee & Yeo, 2016; Zhou et al., 2017), where integrated reporting is de facto mandatory (Dumay et al. 2016). This relationship is rather underexposed in other regions of the world, although the relevance and adoption of integrated reporting has been increasing since 2002 (Jensen & Berg, 2012). Therefore, as long as enough data was available, listed companies from around the globe were included in this research. Concerning the time frame, the period from 2014 until 2018 was chosen. Since the IIRC only published the <IR> Framework in 2013 (IIRC, 2013), it might be expected that a measurable increase in the levels of integratedness of corporate reporting is visible from 2014 onwards. Data after 2018 was incomplete and thus excluded from this study.

In order to construct the dataset of this study, the data regarding the level of integratedness of reporting was gathered first, since the data available for this variable is the smallest. This dataset, developed by Refinitiv (ASSET4), can be accessed through Thomson Reuters Eikon and provides data for over more than 8,000 companies worldwide. Based on companies in this dataset, the corresponding data on all other variables in this research, except for investor protection, was collected through Thomson Reuters Eikon as well. Finally, the country data for investor protection was collected through the open database of the World Bank. Since the data of companies in the dataset is expressed in different currencies, all currencies were converted to US dollars to make comparisons more feasible.

The initial dataset consisted out of 3,738 companies from 58 different countries, classified in 10 different industries. After omitting missing observations, inconsistencies and duplicates, a balanced dataset was constructed with 3,332 companies (16,660 firm-year observations) from 52 countries. By using a balanced dataset, the likelihood of having biased estimators is reduced in comparison to unbalanced datasets (Baltagi, 2008). The countries that are represented at the highest quantity were the United States with 734 companies, Japan with

(18)

18

382 companies, and Australia with 274 companies respectively. On the other hand, the least represented countries were Sri Lanka, Morocco, Peru, and Nigeria with one single company each. A table containing the full sample expressed in firm-year observations in each country is presented in Table 1.

Table 1: Firm-year observations per country

Country Freq. Percent Cum. Country Freq. Percent Cum.

AE 40 0.24 0.24 JP 1910 11.46 58.88 AT 70 0.42 0.66 KR 275 1.65 60.53 AU 1370 8.22 8.88 KW 15 0.09 60.62 BE 130 0.78 9.66 LK 5 0.03 60.65 BR 260 1.56 11.22 LU 15 0.09 60.74 CA 1055 6.33 17.56 MA 5 0.03 60.77 CH 300 1.80 19.36 MX 85 0.51 61.28 CL 95 0.57 19.93 MY 240 1.44 62.73 CN 440 2.64 22.57 NG 5 0.03 62.76 CO 45 0.27 22.84 NL 155 0.93 63.69 CZ 20 0.12 22.96 NO 90 0.54 64.23 DE 380 2.28 25.24 NZ 80 0.48 64.71 DK 115 0.69 25.93 PE 5 0.03 64.74 EG 30 0.18 26.11 PH 105 0.63 65.37 ES 190 1.14 27.25 PL 110 0.66 66.03 FI 125 0.75 28.00 PT 40 0.24 66.27 FR 365 2.19 30.19 QA 65 0.39 66.66 GB 1275 7.65 37.85 RU 115 0.69 67.35 GR 55 0.33 38.18 SA 25 0.15 67.50 HK 745 4.47 42.65 SE 220 1.32 68.82 HU 20 0.12 42.77 SG 200 1.20 70.02 ID 100 0.60 43.37 TH 155 0.93 70.95 IE 50 0.30 43.67 TR 90 0.54 71.49 IL 60 0.36 44.03 TW 530 3.18 74.67 IN 400 2.40 46.43 US 3670 22.03 96.70 IT 165 0.99 47.42 ZA 550 3.30 100.00 Total 16,600

Table 2 represents the sample distribution of firm-year observations per industry. The financial industry was represented most with 3,450 observations, where the utilities industry was the least represented industry with only 540 observations.

(19)

19

Table 2: Firm-year observations per industry

Industry Freq. Percent Cum.

Basic Materials 1920 11.52 11.52 Consumer Cyclicals 2535 15.22 26.74 Consumer Non-Cyclicals 1235 7.41 34.15 Energy 1275 7.65 41.81 Financials 3450 20.71 62.52 Healthcare 905 5.43 67.95 Industrials 2760 16.57 84.51 Technology 1295 7.77 92.29 Telecommunication Services 540 3.24 95.53 Utilities 745 4.47 100.00 Total 16,600 3.2. Variables 3.2.1. Dependent variable

In this research, the dependent variable is corporate financial performance. This variable is measured in both the short and long term. The distinction between short and long-term financial performance corresponds to the widely used distinction of firm performance measured as either accounting-based measures or market-based measures (Gentry & Shen, 2010; Margolis & Walsh, 2001). As a proxy for short-term financial performance, the measure Return on Assets (ROA) was used. Within business literature, ROA is one of the most widely reported profitability measures (Burton, Lauridsen & Obel, 2002). Following previous literature (Zhou et al., 2017), ROA was measured as follows:

𝑅𝑒𝑡𝑢𝑟𝑛 𝑜𝑛 𝐴𝑠𝑠𝑒𝑡𝑠 (𝑅𝑂𝐴) = 𝑁𝑒𝑡 𝑖𝑛𝑐𝑜𝑚𝑒 𝑇𝑜𝑡𝑎𝑙 𝑎𝑠𝑠𝑒𝑡𝑠

This measure can be accessed directly from the database of Thomson Reuters Eikon (Datastream). The higher the value of ROA, the more net income a company is generating over its assets, the better the company is currently performing.

The used measurement for long-term financial performance was Tobin’s q. Tobin’s q is a ratio of a company’s market value to the replacement costs of its assets. Similar to ROA, it is widely used in research regarding corporate financial performance (Lindenberg & Ross, 1981). Tobin’s q is a market measure that can be used to reflect future performance (Margolis & Walsh, 2001). It is assumed that the market value of liabilities is equal to the book value of liabilities. Therefore, the numerator is measured as the sum of equity market value and liabilities book value. Because of a lack of active markets in aged capital goods, it is difficult to correctly measure the replacement costs of assets, which is the denominator in the ratio of

(20)

20

Tobin’s q (Lindenberg & Ross, 1981). Therefore, the denominator in the ratio of Tobin’s q was specified as the book value of total equity and liabilities. Thus, the measure of Tobin’s q in this research was constructed as follows:

𝑇𝑜𝑏𝑖𝑛′_{𝑠 𝑞 =}𝐸𝑞𝑢𝑖𝑡𝑦 𝑚𝑎𝑟𝑘𝑒𝑡 𝑣𝑎𝑙𝑢𝑒 + 𝐿𝑖𝑎𝑏𝑖𝑙𝑖𝑡𝑖𝑒𝑠 𝑏𝑜𝑜𝑘 𝑣𝑎𝑙𝑢𝑒

𝐸𝑞𝑢𝑖𝑡𝑦 𝑏𝑜𝑜𝑘 𝑣𝑎𝑙𝑢𝑒 + 𝐿𝑖𝑎𝑏𝑖𝑙𝑖𝑡𝑖𝑒𝑠 𝑏𝑜𝑜𝑘 𝑣𝑎𝑙𝑢𝑒

In order to compute Tobin’s q, the measures equity market value, equity book value and liabilities book value were retrieved from Thomson Reuters Eikon (Datastream). Subsequently, the formula was applied. In case Tobin’s q has a value that exceeds 1, it is suggested that a company has intangible assets that are related to future growth opportunities (Sudarsanam, 2003).

3.2.2. Independent variables

In this study, two independent variables are of main interest. These are the level of integratedness of corporate reporting and investor protection. To measure the first independent variable, level of integratedness of corporate reporting, the data regarding the ‘CSR Strategy Score’ (TRESGCGVSS) from the Refinitiv database was used. This measure is described as “a score that reflects a company's practices to communicate that it integrates the economic (financial), social and environmental dimensions into its day-to-day decision-making processes” (Refinitiv, 2020). The score is scaled on a range from 0 to 100. The higher a company scores on this range, the better a company communicates or reports that it integrates financial, social and environmental dimensions into its decision-making processes (Refinitiv, 2020). Thus, a high CSR Strategy score corresponds to a high level of integratedness of corporate reporting.

This proxy is an updated version of the CGVS (corporate governance vision and strategy) measure (Refinitiv, 2020). Although this measure has its boundaries for being used as a proxy to measure components of integrated reporting (De Villiers, Venter & Hsiao, 2017), it is a measure that has been used in previous literature (Busco et al., 2019; Serafeim, 2015). Busco et al. (2019) used the CGVS score as a proxy to measure the company level of integration, a construct that is similar to the level of integratedness of corporate reporting. They define the company level of integration as the “management’s commitment and effectiveness towards integrating financial and extra-financial aspects” (Busco et al., 2019, p.279).

(21)

21

To measure the second independent variable, investor protection, the ‘Protecting Minority Investors’ measure from the World Bank was used. According to the World Bank, this measure is defined as “the strength of minority shareholder protections against misuse of corporate assets by directors for their personal gain as well as shareholder rights, governance safeguards and corporate transparency requirements that reduce the risk of abuse” (The World Bank, 2019). A score is calculated for, in case data is available, every country around the globe and these scores can vary over the years. The measure is constructed as two sets of indicators, one set indicating protection from conflicts of interest and the other indicating shareholders’ rights in corporate governance. In turn, each set consists out of three indicators. Summing the total of six indicators, a score for the protection of investors is calculated on a scale from 0 to 100. The higher the score, the higher the level of investor protection.

The usage of this proxy to measure investor protection has been widely reported throughout previous literature (Lin & Ewing-Chow, 2016). The measure is rooted in an article by Djankov, La Porta, Lopez-de-Silanes and Shleifer (2008), in which they presented a new measure of minority shareholder legal protection. Since then, the World Bank has been updating this measure annually (Haidar, 2009). Some widely cited studies that have used this measure are Enomoto, Kimura and Yamaguchi (2015), Haidar (2009), Houqe et al. (2012) and Jensen and Berg (2012).

3.2.3. Control variables

In order to avoid that the results are biased, this research controlled for a certain set of variables. By including such control variables, the justification of the findings is strengthened (Rehman, Orij & Khan, 2020). The control variables were chosen based on their usage in previous literature regarding integrated reporting and corporate financial performance (García‐Sánchez, Rodríguez-Ariza & Frías-Aceituno, 2013; Lee & Yeo 2006; Zhou et al., 2017). Following Zhou et al., (2017), the control variable size, which as measured as the natural logarithm of a company’s total assets (in US dollars), was included. With respect to size, it has been found that larger companies tend to disclose more, due to the more extensive usage of capital markets (García‐Sánchez et al., 2013). The second control variable was leverage, which is measured as the ratio of total debt to total assets (Barth et al., 2017). In case a firm is highly financed on debt, the agency of free cash flows can be tempered, which influences financial performance (Lee & Yeo, 2016). Furthermore, the control variable growth was included. Growth is measured as the percentage change of current sales compared to last year’s sales (Lee & Yeo, 2016). It is assumed that growth is an incentive for increasing

(22)

22

the disclosure of information in order to decrease issues regarding information asymmetry (García‐Sánchez et al., 2013). Moreover, this research controlled for the industry companies operate in, by including dummies that represent the industries based on TR1N codes, a 2-digit code that clusters industries. It is assumed that companies in the same industry express similar behavior regarding the disclosure of information (García‐Sánchez et al., 2013). Finally, country and year were included as control dummies. In Table 3, a summary of all the relevant variables in this study is included.

Table 3: Variables explained

Variable Description Data source

Dependent variables

Return on assets (ROA) Ratio of net income to total assets Datastream (Eikon) Tobin’s q (TQ) Market value of company divided by the book value of

total assets

Datastream (Eikon)

Independent variables

The level of integratedness of corporate reporting (IR)

The level of integratedness of corporate reporting on a scale from 0 to 100

ASSET4 (Eikon)

Investor protection (INV) The level of investor protection in a country on a scale from 0 to 100

The World Bank

Control variables

Size (SIZE) The natural logarithm of a company’s total assets Datastream (Eikon) Leverage (LEV) Ratio of total debt to total assets Datastream (Eikon) Growth (GRO) The net sales increase divided by the sales of last year Datastream (Eikon)

Industry (IND) Industry dummies Datastream (Eikon)

Country (COU) Country dummies Datastream (Eikon)

Year (YEAR) Year dummies Datastream (Eikon)

3.3. Model

In the previous chapter, the most important constructs and their relationships have been discussed. In short, this thesis investigates the relationship between the level of integratedness of corporate reporting and company financial performance in the short and long run, moderated by the level of investor protection. To make these relationships more convenient to grasp, a conceptual model is presented in Figure 2.

(23)

23

Figure 2: Conceptual model

Since the variables were measured taking into account multiple research entities and different moments in time, panel data regressions were the most suitable method of analysis. It provides multiple observations on each entity in a given sample (Hsiao, 2014). Thus, in order to test the hypotheses developed in chapter 2, panel data regressions were conducted estimating the following equation (1):

𝐹𝐼𝑃𝐸𝑅𝑖,𝑡 = 𝛼 + 𝛽1𝐼𝑅𝑖,𝑡+ 𝛽2𝐼𝑁𝑉𝑖,𝑡+ 𝛽3𝐼𝑅 ∗ 𝐼𝑁𝑉𝑖,𝑡 + 𝛽4𝑆𝐼𝑍𝐸𝑖,𝑡+ 𝛽5𝐿𝐸𝑉𝑖,𝑡 + 𝛽6𝐺𝑅𝑂𝑖,𝑡+

𝛽₇𝐼𝑁𝐷_𝑖 + 𝛽₈𝐶𝑂𝑈_𝑖 + 𝛽₉𝑌𝐸𝐴𝑅_𝑖 + 𝜖_𝑖,𝑡 (1)

where FIPER is short (ROA) or long-term (TQ) financial performance, IR is the level of integratedness of corporate reporting, INV is the level of investor protection, IR*INV is the interaction term of the level of integratedness of corporate reporting and the level of investor protection, SIZE is the size of the company, LEV is the leverage-ratio of the company, GRO is the annual growth of the company, IND is the industry the company is active in, COU is the country where the company originates from, and YEAR is the corresponding year.

It is hypothesized that the level of integratedness is positively related to both short and long-term financial performance. Therefore, it was expected that 𝛽₁ was positive in the ROA version of the equation(1) and positive in the TQ version of the equation(1). Regarding the other two hypotheses, it is hypothesized that the level of investor protection negatively moderates the assumed relationship between the level of integratedness of corporate reporting and both short and long-term financial performance. Therefore, it was expected that the interaction term 𝛽3 was negative in both the ROA version of equation(1) and the TQ version

of equation(1). Level of integratedness of corporate reporting Investor protection Corporate financial performance

(24)

24 3.4. Testing assumptions

Before the panel data regressions were conducted, some assumptions had to be considered. First of all, scatter plots were analyzed to check whether outliers in the variables could possibly influence the results. These scatter plots showed that ROA, Tobin’s q, Leverage and

Growth had some extreme outliers. To deal with the outliers, these variables were winsorized

at the 1% and 99% level. By winsorizing variables, one modifies or assigns lesser weights to outliers so these extreme values are closer to other values (Ghosh & Vogt, 2012). To check whether correlations between the variables might have been problematic, a correlation matrix including all the variables was constructed (Table 4). As can be noticed, all correlations have values below the critical value of 0.7 suggesting that collinearity was not an issue. To test whether multicollinearity was an issue, a VIF-test was conducted. As can be seen in Table 5, all variance inflation indicators are below the critical region of 5-10 (O’brien, 2007). Therefore, it can be concluded that multicollinearity was not an issue. Furthermore, to test whether heteroskedasticity was an issue, a Breusch-Pagan/Cook-Weisberg test was performed on two models, one where ROA is the dependent variable and one where TQ is the dependent variable. As the results in Appendix 1 show, in both models the zero hypothesis that there is constant variance had to be rejected (p=0.000 in both models). This suggested that heteroskedasticity was an issue in both models. Finally, to test for autocorrelation in both models, a Woolridge test was conducted. This test analyzes the zero hypothesis that there is no first-order autocorrelation. As the results in Appendix 2 show, in both models the zero hypothesis was violated (p=0.000). This suggests that autocorrelation was an issue in both models. To correct for both heteroskedasticity and autocorrelation, the standard errors of the coefficient estimates were adjusted to panel clusters standard errors (Vogelsang, 2012).

(25)

25

Table 4: Correlation matrix Matrix of correlations Variables (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (1) ROA 1.000 (2) TQ 0.441 1.000 (3) IR 0.031 -0.078 1.000 (4) INV 0.027 0.044 0.072 1.000 (5) SIZE -0.035 -0.272 0.365 -0.113 1.000 (6) LEV -0.086 -0.089 0.049 0.011 0.125 1.000 (7) GRO 0.154 0.110 -0.062 0.014 -0.035 -0.006 1.000 (8) IND 0.017 0.012 0.004 -0.034 0.131 0.074 -0.012 1.000 (9) COU 0.106 0.104 -0.044 0.083 0.174 0.089 -0.040 0.078 1.000 (10) YEAR 0.022 -0.021 0.098 0.027 0.019 0.008 0.029 0.000 -0.000 1.000 Table 5: VIF-test

Variable VIF 1/VIF

SIZE 1.270 0.787 IR 1.210 0.827 COU 1.070 0.934 INV 1.050 0.957 IND 1.030 0.974 LEV 1.020 0.976 YEAR 1.010 0.989 GRO 1.010 0.993 Mean VIF 1.080

(26)

26

Chapter 4. Results

This chapter discusses the results obtained in conducting the research. In the first section, the data used is described. Afterwards, the results of the panel data regressions are presented. Finally, the robustness of the results is tested by performing some robustness checks.

4.1. Descriptive statistics

Table 6 presents the descriptive statistics of the sample used in this research after missing observations, inconsistencies and duplicates have been omitted, and the variables have been winsorized. Since a balanced dataset is used (see chapter 3.1), all variables have an equal amount of observations (16,600). The average ROA in the sample is 0.051, indicating that the companies in the sample generate 5.1% in return on assets on average over the years 2014 to 2018. The average TQ in the sample is 1.704. This indicates that, on average, the value of the companies in the sample is more worth than the cost of its assets. Regarding the level of integratedness of corporate reporting (IR), the average of the sample is 41.386. This value is below the average mean of the benchmark, which is 50. This suggests that the average company in the sample has incorporated a relatively low level of integratedness of corporate reporting. Finally, INV scores an average value of 71.488. This indicates that the average company from the sample is from a country with a relatively strong level of investor protection, as compared to the benchmark average of 50. This claim is further substantiated by the fact that the average level of investor protection in the sample (71.488) is relatively close to the highest level of investor protection in the sample (88).

Table 6: Descriptive statistics

Variable Observations Mean Standard deviation Min Max

Dependent variables ROA 16,660 .051 .079 -.297 .304 TQ 16,660 1.704 1.177 .587 7.516 Independent variables IR 16,660 41.386 33.539 0 99.87 INV 16,660 71.488 9.024 28 88 Control variables SIZE 16,660 15.725 1.774 3.871 22.19 LEV 16,660 .256 .184 0 .815 GRO 16,660 .076 .205 -.452 1.149

(27)

27 4.2. Testing hypotheses

To test the hypotheses formulated in chapter 2, a random-effects model is used. The choice for a random-effects model is made since the variable INV is time-invariant for a significant part of observations in the sample. A fixed-effects model is not suitable for measuring time-invariant variables (Bell & Jones, 2015). Table 7 presents the results of testing all the formulated hypotheses. In hypothesis 1A it is argued that the level of integratedness of corporate reporting is positively associated with short-term financial performance. Hypothesis 1B argues that the level of integratedness of corporate reporting is positively associated with long-term financial performance. The second set of hypotheses, 2A and 2B, test whether there is a moderating effect of investor protection on the relationship between the level of integratedness of corporate reporting and company financial performance. It is argued that investor protection negatively moderates this relationship, in both the short and long term.

Model 1 and 2 test the direct relationship of the level of integratedness of corporate reporting and short and long-term financial performance respectively. In model 1 the results show a positive coefficient of IR of 0.00009 on ROA. The coefficient is significant at the 1% level (p=0.001). This suggests that hypothesis 1A is supported by empirical evidence. It seems that higher levels of integratedness of corporate reporting are associated with increased short-term financial performance. This result is in line with the study of Lee and Yeo (2016). Companies with a significant implementation of integrated reporting practices outperform companies with less integrated reporting practices (Lee & Yeo, 2016). The model indicates an R-squared of 0.1214. This means that 12.14% of the variance in return on assets is explained by the variables incorporated in model 1.

Regarding model 2, the results show a positive coefficient of IR of 0.00092 on TQ. Again, this coefficient is significant at the 1% level (p=0.003). This result thus provides support for hypothesis 1B. Based on the empirical evidence, it should be concluded that higher levels of integratedness of corporate reporting are associated with increased levels of long-term financial performance. This result is in line with previous studies of Barth et al. (2017), and Lee and Yeo (2016). The R-squared of 0.2613 indicates that 26.13% of the variance in Tobin’s q is explained by the variables incorporated in model 2. Based on the empirical evidence, it can be concluded that the level of integratedness of corporate reporting is positively related to company financial performance. Thus, the evidence suggests that increasing the level of integratedness of corporate reporting could result in value creation in both the short and long run, as claimed by the IIRC (2013).

(28)

28

Model 3 and 4 extend the previously tested models by controlling for investor protection. The results show that, when controlling for investor protection, IR is still positively associated with financial performance significantly. In model 3, investor protection (INV) is negatively associated with short-term financial performance (coef.=-0.00062; p=0.017). This result suggests that in countries with stronger levels of investor protection, companies perform weaker compared to companies from countries with lower levels of investor protection. Regarding the effect of investor protection on long-term financial performance, a positive coefficient of 0.0055 is reported. However, this coefficient is insignificant (p=0.118). These results partly contradict previous findings of Houqe et al. (2012) and La Porta et al. (2000), who suggest that investor protection is negatively associated with cost of capital, and thus positively associated with financial performance. The explanatory power of both models has barely changed compared to model 1 and 2, with overall R-squared values of 0.1214 and 0.2612 respectively.

Finally, in model 5 and 6, the moderating effect of investor protection on the relationship between the level of integratedness of corporate reporting and financial performance is tested. This moderating effect is tested by constructing an interaction term of

IR and INV, resulting in the interaction variable IR*INV. Before this interaction term is

constructed, both variables have been mean-centered. The results in model 5 report a negative coefficient of -0.00001 for the interaction term IR*INV. This coefficient is significant at the 10% level (p=0.063). Therefore, hypothesis 2A is supported by empirical evidence. This result suggests that investor protection negatively moderates the positive relationship between the level of integratedness of corporate reporting and short-term financial performance, and is in line with the findings of Durnev and Kim (2005), and García-Sánchez and Noguera-Gámez (2017). As argued in chapter 2.4, a higher level of integratedness of corporate reporting has less (more) value in stronger (weaker) investor protection countries. This seems to be case because information asymmetry is less (more) of an issue in strong (weak) investor protection countries. Regarding model 6, the interaction term also shows a negative coefficient of -0.00003. However, this coefficient is insignificant (p=0.206). Thus, the empirical evidence does not support hypothesis 2B. The relationship between the level of integratedness of corporate reporting and long-term financial performance does not seem to be moderated by investor protection. The explanatory power of both models is similar to the previously tested models, with an overall R-squared of 0.1215 and 0.2613 respectively.

(29)

29

Table 7: Panel data regressions

MODEL (1) (2) (3) (4) (5) (6)

Direct effect IR Direct effect IR

Direct effect IR, controlled for INV

Interaction effect IR and INV

VARIABLES ROA TQ ROA TQ ROA TQ

IR 0.00009*** 0.00092*** 0.00009*** 0.00092*** 0.00008*** 0.00091*** (3.28) (3.00) (3.28) (3.00) (3.19) (2.95) Predicted sign + + + + INV -0.00062** 0.00550 -0.00063** 0.00541 (-2.38) (1.56) (-2.45) (1.54) IR*INV -0.00001* -0.00003 (-1.86) (-1.26) Predicted sign - - SIZE -0.00113 0.31961*** -0.00112 0.32010*** -0.00112 0.32019*** (-0.97) (-11.29) (-0.95) (-11.30) (-0.96) (-11.31) LEV 0.09345*** 0.32811*** 0.09357*** 0.32625*** 0.09377*** 0.32725*** (-11.45) (-2.83) (-11.46) (-2.82) (-11.48) (-2.83) GRO 0.05567*** 0.20921*** 0.05573*** 0.20867*** 0.05580*** 0.20917*** (12.55) (7.54) (12.56) (7.52) (12.57) (7.54) Constant 0.05577** 6.40337*** 0.09848*** 6.02855*** 0.05815** 6.46267*** (2.07) (14.08) (2.99) (12.19) (2.14) (14.00)

Year Dummies YES YES YES YES YES YES

Country Random Effects YES YES YES YES YES YES

Industry Random Effects YES YES YES YES YES YES

Observations 16,660 16,660 16,660 16,660 16,660 16,660

Number of companies 3,332 3,332 3,332 3,332 3,332 3,332

Overall R-squared 0.1214 0.2613 0.1214 0.2612 0.1215 0.2613

(30)

30 4.3. Robustness checks

In order to test the robustness of the results, some additional analyses are performed. By conducting these robustness tests, it can be analyzed whether the results obtained are valid when changing the estimators in the regression equation (Lu & White, 2004). Firstly, the independent variable IR is lagged for one, two and three years respectively. By using lagged values of IR, the path of causality in the relationship between the level of integratedness of corporate reporting and financial performance could possibly be identified (Bellemare, Masaki & Pepinsky, 2017). Moreover, it could be argued that the effects of the level of integratedness of corporate reporting might not be visible yet in the year that a certain level of integratedness is implemented. Thus, by regressing lagged values of IR on financial performance, the relationship between the level of integratedness of corporate reporting and financial performance in consecutive years after implementation can be further explored.

In table 8, the results of the lagged independent variable analyses are presented in model 7, 8 and 9. As can be noticed, all lagged values of IR have a positive coefficient, suggesting that a higher level of integratedness of corporate reporting results in improved financial performance in consecutive years after implementation. However, where the coefficients in model 7 and 8 are significant at 1% and 5% respectively, the coefficient in model 9 is insignificant. These results suggest that the positive effect of the level of integratedness of corporate reporting on financial performance decreases over the years. This indicates that a positive effect on long-term financial performance is not supported by this empirical evidence, contradicting the findings in model 2.

As a second robustness test, the variable investor protection is transformed as a dummy variable by using a median split. Countries that score equal to or above the median of 71.6 are grouped into the category of countries with ‘strong’ investor protection. Consequently, firms from countries categorized as ‘strong’ score a 1 on the dummy variable, and 0 otherwise. In table 8, the results of this robustness test are presented in model 10 and 11. The direct effect of IR on both ROA and TQ holds with positive coefficients of 0.00012 (p=0.000) and 0.00161 (p=0.000) respectively. When analyzing the interaction term

IR*STRINV, both coefficients are negative with values of -0.00007 and -0.00116 respectively.

As argued in chapter 2.4., this suggests that strong investor protection negatively moderates the positive relationship between the level of integratedness of corporate reporting and financial performance. However, where the interaction term in the regular analysis was solely significant in the model with ROA as the dependent variable, the interaction term in the

The relationship between the level of integratedness of corporate reporting and financial performance. Exploring the moderating effect of investor protection