Corporate Growth: Effects on
Strategic Risk Disclosure Quality
University of Groningen, Faculty of Economics and Business
Master’s Thesis MSc Accountancy & Controlling, specialization Controlling
June 24, 2019
Author Jasper J. Spoler Student number S2766191
Address Jozef Israëlsstraat 43 9718 GC Groningen Telephone number +31 (0)610608119 Email address j.j.spoler@student.rug.nl
Supervisors Prof. Dr. J.A. Emanuels F.J. Bos, MSc
ABSTRACT
This study examines the association between corporate growth and the quality of strategic risk disclosures. Based on the contrasting expectations from the agency, stakeholder, and proprietary costs theory, a non-directional relationship is hypothesized. Additionally, the direct and moderating effects of investor protection are assessed. Using a sample of 361 firm-year observations from 122 firms for the period 2015-2017 and using a new, self-constructed index that measures the quality of strategic risk information, it is found that corporate growth does not significantly impact strategic risk disclosure quality. This finding, which is supported by additional robustness tests, suggests that growing firms have no incentive to either improve or deteriorate the quality of their strategic risk disclosures. Further, secondary tests show that the degree of investor protection is a positive determinant of this disclosure quality. The study theoretically and empirically contributes to the multi-faceted risk disclosure debates, and the findings provide practical insights for investors in their daily assessments of firms’ values, prospects, and risks.
Keywords: Corporate Growth, Strategic Risk Disclosure, Disclosure Quality, Investor Protection
TABLE OF CONTENTS
1. INTRODUCTION 4
2. THEORETICAL BACKGROUND 7
2.1. Strategic risk disclosure 7
2.2. Information asymmetry 8
2.3. Stakeholders’ influence 10
2.4. Costs of disclosing strategic risk information 11
2.5. Investor protection 13
3. METHODOLOGY 15
3.1. Sample and data collection 15
3.2. Strategic risk disclosure quality 16
3.3. Corporate growth 19 3.4. Investor protection 19 3.5. Control variables 20 3.6. Model of analysis 23 4. RESULTS 23 4.1. Descriptive statistics 23 4.2. Hypothesis testing 26
4.3. Disclosure index conversion 28
4.4. Skewness issues 29
4.5. Corporate growth robustness 29
4.6. Additional robustness tests 30
5. DISCUSSION AND CONCLUSION 31
REFERENCES 34
4 1. INTRODUCTION
The concept of effective risk management has received increasing attention since the beginning of the 21st century (Linsley & Shrives, 2006). This trend has mainly been triggered by rapidly changing business environments, such as increasingly complex business structures and regulations (Beretta & Bozzolan, 2004), increased dependence on information technology and international transactions (Dobler, 2008), and the globalization of financial markets (Shi, Magnan, & Kim, 2012). The urge for companies to identify and manage their risks has been strengthened even further as a result of a series of corporate scandals, the global financial crisis, and the recently emerged climate awareness (Quon, Zeghal, & Maingot, 2012; Frigo & Anderson, 2011; Guenther, Guenther, Schiemann, & Weber, 2016). In order to construct an aggregate risk profile, businesses nowadays often make use of Enterprise Risk Management, which helps them with identifying their risks, assessing the likelihood and possible impact of these risks, developing mitigation techniques, and monitoring the risks and mitigation processes (Berry‐Stölzle & Xu, 2018).
Information on a company’s risk profile is not only internally possessed, but also reported on in narrative disclosure sections of, for example, the annual report. The reason for companies to disclose information on their risks is partly regulation-based. Prior academic literature suggests that managers might be reluctant to disclose unfavorable information, as it may be harmful for the company’s position in the capital market, managers’ compensation or career prospects, and because it may provide competitors with proprietary information (Dobler, 2008; Campbell, Chen, Dhaliwal, Lu, & Steele, 2014). Investors and other stakeholders, however, need and demand information on risks as it enables them to more adequately assess the firm’s performance, value, and prospects (Solomon, Solomon, Norton, & Joseph, 2000; Campbell et al., 2014). As a result, several codes and standards, such as Security and Exchange Commission (SEC) requirements in the United States or the Dutch Corporate Governance Code in the Netherlands, have included sections that require companies to elaborate on risks that may reduce future firm performance.
Furthermore, incentives exist which may induce companies to disclose extended risk information on a voluntary basis. It is believed, and sometimes empirically found, that voluntarily disclosing information reduces information asymmetry (Shi et al., 2012), and, subsequently, the cost of capital (e.g., Campbell et al., 2014). Furthermore, extended reporting enables investors to better monitor managers, which in turn lowers agency costs (Hooghiemstra, Hermes, & Emanuels, 2015). These findings suggest that the disclosure of risk information could also provide benefits for the firm.
5 As a result of these opposing viewpoints, studies on the incentives for risk reporting have become more popular in recent years. A recent study by Elshandidy, Shrives, Bamber, and Abraham (2018) reviews the extant literature in detail. For example, company size (Linsley & Shrives, 2006), the number of executive directors (Abraham & Cox, 2007), board diversity and size (Ntim, Lindop, & Thomas 2013), and financial leverage (Deumes & Knechel, 2008; Taylor, Tower, & Neilson, 2010) have been found to be positive drivers or determinants for risk reporting. Another important factor that is often found to influence risk reporting practices is the type of industry. The general industry-related exposure to risks, the behavior of companies within an industry, social pressure to disclose industry-related information, or other industrial constraints are reasons why certain companies disclose more risk information than others (for an overview, see Beretta & Bozzolan, 2004).
Despite this growing interest, research on risk reporting is still somewhat in its infancy (Abraham & Shrives, 2014). A reason for this is that there is no universally agreed upon definition of risk, and that it entails several components (Linsley & Shrives, 2006). Moreover, Elshandidy et al. (2018) note that previous research has paid too little attention to the different concepts and types of risk and their meanings, as each firm’s environment and characteristics result in different (types of) risks. It therefore seems that previous literature has skipped a crucial step, namely that of identifying, separating, and studying the various components that the overarching concept of risk entails. Furthermore, studies on risk reporting have often been criticized for their techniques regarding the measurement of risk disclosure quality. Beretta and Bozzolan (2004) highlight that many researchers focus on the quantity of risk disclosures as a proxy for its quality, which they find to be insufficient. They propose a framework that not only pays attention to the volume of disclosure, but also to what is reported and how. Their work, however, has been criticized for being too subjective (Elshandidy et al., 2018).
Motivated by the recently emerging interest in risk disclosure practices and by the identified gaps in previous research, this study extends prior work by investigating the effect of a seldom addressed firm characteristic on one, specific component of risk disclosure. Specifically, this study examines the relationship between corporate growth and strategic risk disclosure quality. Growing firms have a competitive advantage, which means they are strategically ahead of their competitors (Stonehouse & Snowdon, 2007). Therefore, information on strategic risks is a particularly important part of non-financial disclosures for investors, competitors and other stakeholders of fast-growing firms. Moreover, strategic risks are the most important risks to a firm (Chockalingam, Dabadghao, & Soetekouw, 2018), making it even more interesting to study this type of risk. Furthermore, strategic risks “can
6 devastate a company’s growth and value” (Chockalingam et al., 2018: 229), which makes the concept particularly interesting when combined with corporate growth. Theoretical views, however, suggest an opposing association between corporate growth and strategic risk disclosure quality. On the one hand, the agency and stakeholder theory, through the demand of investors and the lower cost of capital, expect a positive relationship, while the proprietary costs theory suggests that costs related to the disclosure of strategic risk information become higher as a firm grows. Which of these contrasting theoretical expectations will prevail is therefore an empirical question, whose answering is the goal of this study.
Using data from 361 firm-year observations for 122 large firms from 14 countries between 2015 and 2017, it is found that there is no association between corporate growth and strategic risk disclosure quality. Several additional tests yield the same conclusion. This suggests that corporate growth is not a determinant of how a company handles the provision of strategic risk information. However, it could also suggest that the before-mentioned theoretical views exactly neutralize each other, or that disclosure of (strategic) risk information is so heavily regulated that any voluntary additions can only be marginal. Additional research could shed more light on the underlying reasons for this study’s results. Furthermore, investor protection, as proxied by a country’s legal system, was found to significantly impact strategic risk disclosure quality. That is, firms operating in countries with stronger investor protection scored higher on the self-constructed quality index than firms operating in countries with weaker investor protection. There was no evidence, however, for a moderating effect of investor protection on the relationship between corporate growth and strategic risk disclosure quality.
This study makes several theoretical and practical contributions. First, it adds to the emerging literature on drivers and determinants of risk disclosure by studying the effect of corporate growth. Second, the study focuses on strategic risk disclosure in particular. It therefore not only bridges the gap that previous research is too much focused on the overarching concept of risk disclosure in general; it also extends the poor literature on strategic risks and strategic risk disclosure. Third, by focusing on opposing theoretical expectations, this study enhances the understanding of why prior risk disclosure research has often found mixed results (Elshandidy et al., 2018), and, subsequently, adds to the understanding of why certain firms are more reluctant than others to voluntarily disclose (strategic) risk information. Fourth, this study uses an unweighted, self-constructed disclosure index. As a result, it uses a much more adequate proxy of disclosure quality as opposed to the widely used quantity measures, which fail to sufficiently embrace the quality aspect of
7 disclosed information, or the weighted indices, which can be criticized for being too subjective. Additionally, it adds to the literature of self-constructing (risk) disclosure indices. Finally, this study may provide investors with valuable information. Investors are generally highly interested in fast-growing firms due to high expected returns (Estrada, 2012), so knowledge of the association between company growth and strategic risk disclosure quality adds to investors’ assessments of the firm’s value and, subsequently, the expected returns.
The remainder of this study is structured as follows. The next section reviews extant literature and develops hypotheses based on relevant theoretical underpinnings. Section 3 then describes the methodology of the analyses, of which the results will be presented and discussed in section 4. Finally, section 5 discusses the conclusions, implications, and limitations of the research, and it provides guidance for future research.
2. THEORETICAL BACKGROUND
Several theories have been used in extant literature to explain why or why not companies voluntarily disclose non-financial information. Some of these theories, however, are contradicting and therefore lead to opposing expectations. In this section, I will discuss the theories that are deemed most important for this research and which will be used to develop testable hypotheses. First, however, some more insight will be given into the concept of strategic risk disclosure as being a major, yet complex component of risk disclosure as a whole.
2.1. Strategic risk disclosure
Studies on risk disclosure often do acknowledge that risk is a complex phenomenon that consists of multiple components (e.g., Linsley & Shrives, 2006), but also often take those different components together to construct one measure of total risk disclosure. Such an approach, however, fails to give an understanding of the entirely different nature of the several types of risk a company faces. Even though one could have endless discussions on how to categorize these different types, consensus seems to exist that risks related to the strategy of an organization is one such type. For example, based on a risk model by the ICAEW1, Linsley and Shrives (2006) distinguished six risk categories, with strategic risks being one of them. Similarly, Beretta and Bozzolan (2004) named strategic risks as one of
8 three categories, and Emanuels and De Munnik (2006) identified strategic risks as a separate category in a risk control system. Furthermore, in a study on internal control disclosures, Hooghiemstra et al. (2015) constructed an index that included three risk components, of which strategic risk, combined with operational risk, is one.
Strategic risks, however, are generally not uniformly defined (Chockalingam et al., 2018), as it is a complex concept that touches upon several aspects of an organization (Collins & Ruefli, 1992). Frigo and Anderson (2011: 83) define strategic risks as “those risks that are most consequential to the organization’s ability to execute its strategies and achieve its business objectives”. Such strategies and objectives can be thought of as the long-term goals of an organization and the courses of action and the allocation of resources necessary to carry out these goals (McGee, Wilson, & Thomas, 2010). Thus, strategic risks can be defined as risks that affect an organization’s long-term goals and their related activities. This is consistent with the definition of Chockalingam et al. (2018: 229), who state that a “strategic risk is the risk (…) that can devastate a company’s growth and value”. One may argue that these definitions are still susceptible to interpretation. For example, risks related to manufacturing and product processes are generally categorized as operational risks (e.g., Linsley & Shrives, 2006; McGee et al., 2010). However, if such risks occur and start to have such a large impact on an organization’s activities that its long-term goals are in jeopardy, those same operational risks could become strategic in nature as well. Yet, in this study it is assumed that strategic risks are only those risks that would have a direct impact on the firm’s strategic and long-term goals and objectives. Subsequently, strategic risk disclosure is the information on these strategic risks that a company discloses in its annual reports. Specific details on this construct are elaborated in the methodology section.
One further point this study would like to make is that risks in general and strategic risks in particular can change over time. For example, risks related to new technologies and scientific developments have gained increased attention in the last few decades (McGee et al., 2010). Similarly, the accounting scandals at the beginning of the century, the financial crisis that started in 2007, and the recently emerging climate awareness resulted in relatively new risks threatening a company’s long-term goals.
2.2. Information asymmetry
The agency theory is one of the most often applied theories in voluntary disclosure studies (Abraham & Shrives, 2014; Taylor et al., 2010). As Jensen and Meckling (1976) state, there is a separation between the owners (principals) and the managers (agents) of a firm,
9 where the latter, as daily decision-makers, act on behalf of the former. However, managers generally possess more firm-specific information than the owners, making it difficult for these owners to control the managers adequately (Hooghiemstra et al., 2015). As a result, managers may engage in opportunistic behavior (Taylor et al., 2010), which ultimately leads to an agency conflict.
Information on the nature, likelihood and possible impact of strategic risks is an example of information that is initially possessed by managers. One solution to reduce this information asymmetry problem is to share the information with investors (Healy & Palepu, 2001). The disclosure of such information could result in a more adequate assessment by the investor of the firm’s performance and prospects, and it aids in monitoring management’s behavior (Hooghiemstra et al., 2015). This would suggest a decrease of the cost of capital and, in turn, an increase in firm value, as empirically found by Francis, Nanda and Olsson (2008), Heinle and Smith (2017), and Berry-Stölzle and Xu (2018).
This line of argument is often used to predict that firms are, at least to some extent, eager to disclose information voluntarily (Clinch & Verrecchia, 2015). However, the beneficial effect of lowered financing costs differs across firms (An, Cook, & Zumpano, 2011). Firms with lower investment opportunities, for instance, are less likely to benefit from a higher level of risk disclosure as there is little need for external financing. In such case, the possible drawbacks of disclosing business-sensitive information, as will be elaborated in section 2.4, will outweigh the relatively small advantages. Similarly, firms with high investment opportunities benefit significantly more from a decrease in the cost of capital, making these firms more eager to disclose extended risk information.
In general, investments are accompanied by high levels of uncertainty, and are most effective when there is long-term commitment (Coad & Rao, 2008). Even though these factors to a large extent impact firms’ investment decisions, Coad and Rao (2008) empirically found that firms that are growing in sales generally increase their investment expenditures. Similarly, Coad and Grassano (2018) also found that sales growth is the largest determinant of investment growth, which is mainly because firms often devote a certain percentage of their sales to investments (Coad & Rao, 2008; Coad & Grassano, 2018). As a result of this larger increase in investment expenses, faster-growing firms are expected to benefit more from a lower cost of capital due to the lowered financing costs of these investments.
Moreover, the long-term commitment property of corporate investments makes investing a strategic company decision. Information on strategic risks will therefore result in a better assessment by investors of a firm’s investment prospects, opportunities and threats,
10 which lowers information asymmetry, agency costs and, in turn, the cost of capital even more. Furthermore, a positive relationship between corporate growth and general disclosure levels is empirically found by Khurana, Pereira, and Martin (2006), O’Sullivan, Percy, and Stewart (2008), and Chavent, Ding, Fu, Stolowy, and Wang (2006). Taken together above arguments and findings, the agency theory posits that faster-growing firms are more likely to disclose high-quality information on their identified strategic risks.
2.3. Stakeholders’ influence
The stakeholder theory shines more light on the discussion why companies are not necessarily reluctant to share sensitive information. The theory posits two views that both, albeit in different ways, explain the importance of stakeholders for an organization and the influence they have on an organization’s disclosure policies. The first view looks at moral stakeholders, which are all people and organizations that are affected by a firm’s activities (Frooman, 1999). It states that these stakeholders have a right to demand a certain standard of performance, such as in the case of a non-governmental organization (NGO) that represents the interests of animals or the environment. Furthermore, they have a right to be informed about the organization’s performance, actions, and prospects (Herremans, Nazari, & Mahmoudian, 2016). These rights can be a reason for companies to disclose risk information. The second view looks at strategic stakeholders, which is everyone who can affect or has influence on a firm (Frooman, 1999). This view suggests that strategic stakeholders can provide a firm with benefits, such as knowledge or legitimization to operate (Herremans et al, 2016). Furthermore, external stakeholders may have control over resources that are critical for the firm’s performance (Guenther et al., 2016). As a result of their supply of resources or any other contribution they make, stakeholders can, and often are, rewarded with information, as found by Qu, Leung and Cooper (2013).
There are many types of stakeholders, which are sometimes grouped based on the amount of pressure they can exert. For example, Rogers and Wright (1998) identified four groups: capital market stakeholders (e.g., equity and debt holders), product market or customer stakeholders, labor stakeholders (including potential employees), and political and social stakeholders (e.g., NGOs). The importance of these different groups and the influence they have has often been a subject of studies on information provision or disclosure. For example, Guenther et al. (2016) found that by disclosing environmental information, firms reacted to the demands of all analyzed stakeholder groups. Similarly, Chiu and Wang (2015) found that stakeholder power positively impacts social disclosure quality, while Qu et al.
11 (2013) state that in exchange for any contribution stakeholders make to the organization, firms voluntarily disclose information to satisfy and induce them.
Recalling the concise definition of strategic risks (i.e., risks that affect an organization’s long-term goals and their related activities), strategic risk disclosures may contain information that is relevant to various types of stakeholders. Linsley & Shrives (2006), for example, categorized strategic risks into several components that could be highly relevant for many stakeholders. Specifically, information on risks related to environmental, industry, planning, performance measurement, political, and regulatory issues is closely related to a firm’s strategy and is highly interesting for various stakeholders. For example, environmental risks are generally a high concern for stakeholders such as policy makers, public interest groups and local residents (Holland & Foo, 2003). Similarly, industry risks can be important for employees, suppliers and competitors; and regulatory risks can have an impact on trade unions, investors and employees.
The total number of stakeholders a firm has intuitively depends on the size of a firm. A larger firm has more employees, trading partners, customers and investors, and it also has more influence in regional politics, is more dependent on actors in the financial market, and is more likely to be watched by NGOs or other public interest groups (Buitendag, Fortuin, & De Laan, 2017). This means that faster-growing firms will experience a larger increase in the number of stakeholders they have as opposed to slow or negative-growing firms. Consequently, there will be both a larger demand for strategic risk information by stakeholders under the moral view, and a larger benevolence by firms to disclose such information under the strategic view. Regarding the former view, and as discussed above, information on strategic risks can be highly valuable for stakeholders, hence the increase in the number of stakeholders will pressure faster-growing companies to release such information. Regarding the strategic view, the increased number of stakeholders will be accompanied by an increase in the dependence of a firm on these stakeholders, whose contributions are vital for the firm. Therefore, the company will be more benevolent to reward the stakeholders with information on their strategic risks. Thus, the stakeholder theory posits that faster-growing firms are more likely to disclose information on their identified strategic risks.
2.4. Costs of disclosing strategic risk information
As often acknowledged in extant literature, the choice of voluntarily disclosing information is a matter of costs and benefits (e.g., Hooghiemstra et al., 2015; Shi et al., 2012;
12 Abraham & Shrives, 2014). Having identified possible benefits such as lower agency costs, a lower cost of capital, and satisfied stakeholders, the possible costs have yet to be explored. An important theory in this respect is the proprietary costs theory, which comprises two views regarding how disclosing firm-specific information can be harmful for both a company and its managers. First, by disclosing firm-specific information a firm provides its competitors with confidential or proprietary information. In this case, a firm runs the risk that new competitors are attracted to the market, or that existing competitors may use the information for their own benefit (Verrecchia, 1983; Hooghiemstra et al., 2015). In the case of the disclosure of strategic risk information, especially the existing competitors form a threat for the disclosing firm. That is, the disclosure provides them with information on how the disclosing company’s strategy might be threatened in the future and how it is going to manage these risks. When competitors have access to this information, it might reduce the disclosing firm’s competitive advantage or increase that of the competitors (Hooghiemstra et al., 2015).
The second way in which disclosing information might cause proprietary costs is in the case where the information is deemed inaccurate, incomplete or bad (Verrecchia, 1983, Abraham & Shrives, 2014; Hooghiemstra et al., 2015). In the case of inaccurate or incomplete information, managers or firms may face legal consequences (Hooghiemstra et al., 2015; Campbell et al., 2014). For example, many regulations, such as country-specific corporate governance codes, require companies to disclose risk information (Bailey, Karolyi, & Salva, 2006). Firms that disclose inaccurate or incomplete information may face non-compliance and, subsequently, legal consequences (Hooghiemstra et al., 2015). As a result, firms in general will be reluctant to share information as they run the risk of being non-compliant.
In the case of bad information, the costs of disclosing information are mainly related to discouraged investors (Abraham & Shrives, 2014). Especially when a company reveals information on its strategic risks, this provides investors with information about threats and uncertainties that may endanger the firm’s long-term goals and objectives. In their assessment of the firm’s value and prospects, this information is deemed negative. This might lead to lower stock prices or an increase in the required return, which is costly for the firm. Similarly, potential investors might be discouraged to start investing in the firm, which may endanger future growth possibilities.
As explained above, there are two ways in which the proprietary costs theory makes companies reluctant to disclose information on their strategic risks: it may provide competitors with valuable strategic information, decreasing (increasing) the disclosing firm’s competitive advantage (disadvantage), and it is disadvantageous in investors’ assessments of
13 the firm’s prospects and value, which may result in a deteriorated position on the financial market. In both cases, the proprietary costs will be higher for growing companies. First, according to Prencipe (2004), the proprietary costs related to the disclosure of information that can be useful for competitors, such as information on a firm’s strategy, will be particularly high for companies that are growing. For example, it may reveal to competitors the existence of a growing firm’s business opportunities, which they tend to have more of, and the accompanying threats. This revelation, in turn, is to the disadvantage of the disclosing firm (Prencipe, 2004). In this respect, Hope, Hu and Lu (2016) empirically show that companies with higher proprietary costs indeed provide less information on their risks. Second, as discussed in the section on agency theory, faster-growing firms will invest more as opposed to slower-growing firms and will therefore attach more value to a low cost of capital. As the disclosure of information on strategic risks may increase the cost of capital, these firms will be more reluctant to share this information under the proprietary costs theory.
Finally, Prencipe (2004: 4) states that “the higher the proprietary costs associated to the disclosure, the less negatively investors react to the withholding of relevant information”. The main argument in this case is that investors make a cost-benefit analysis themselves, concluding that the benefits of them possessing the information to assess the firm’s performance does not always outweigh the costs of competitors possessing the information. As a result, the arguments put forth by the stakeholder theory (i.e., stakeholders demand and deserve information) may lose their strength. Thus, the above arguments suggest that under the proprietary costs theory, faster-growing companies are less likely to disclose information concerning their identified strategic risks.
The three theories discussed above strongly suggest an association between corporate growth and strategic risk disclosure quality. However, the arguments put forth by the several theories are contradicting, making the assignation of a certain direction to the relationship difficult, if not impossible. Therefore, it is hypothesized that there is indeed a relationship between the two variables, but no direction can be assigned to this relationship. Formally:
Hypothesis 1: There is an association between corporate growth and strategic risk disclosure quality.
2.5. Investor protection
All previously discussed theories to some extent explain the possible relationship between corporate growth and strategic risk disclosure quality by looking at the role of
14 investors. Recalling, under the agency and stakeholder theory investors demand and deserve information of strategic risks, while the proprietary costs theory suggests that not publishing such information will prevent investors from being discouraged, or will be beneficial for investors as the firm’s competitors will not have access to the information. Moreover, it is commonly accepted that investor protection is an important determinant of company choices (Hooghiemstra et al., 2015), such as for the choice of whether to rely more on external or internal financing sources (Miroshnychenko, Bozzi, & Barontini, 2019). In the latter study, it is discussed that better investor protection lowers agency costs and increases monitoring effectiveness and efficiency, which lowers the cost of capital and, in turn, makes firms rely more on external financing sources. This lower cost of capital due better investor protection is, for example, found in Chhabra, Ferris, and Sen (2009).
The role of investor protection has also been investigated with regards to disclosure levels in annual reports. For example, Hooghiemstra et al. (2015) found empirical evidence that firms in countries with stronger investor protection disclose more information on their internal controls. Moreover, Bushman, Piotroski, and Smith (2004) state that investors’ demand for firm transparency increases with the degree to which they are protected. Especially this demand for transparency is an interesting feature. Under the agency and stakeholder theories, it was expected that the demand for strategic risk information increases with firm growth, for instance through the increased contribution of investors and their future role in capital accumulation. Thus, if this demand increases further when investor protection is high, it is expected that, under the agency and stakeholder theory, the level of investor protection contributes to the expected positive relationship between corporate growth and strategic risk disclosure quality.
The proprietary costs theory suggests that the decision to disclose information is much more a cost-benefit analysis, where it is expected that for growing firms the costs are likely to outweigh the benefits. However, the theory assumes that when managers wish to conceal negative information, they have the opportunity to do so (Gisbert, Navallas, & Romero, 2014). Yet, when investor protection is strong, there is severely more (legal) pressure to disclose strategic risk information, irrespective of the managers’ cost-benefit analysis. It is therefore expected that when the proprietary costs theory prevails, firms, despite their unwillingness, will disclose more information when shareholder protection is strong. Thus, under the proprietary costs theory, it is expected that the level of investor protection weakens the expected negative relationship between corporate growth and strategic risk disclosure quality.
15 Moreover, La Porta, Lopez‐de‐Silanes, Shleifer and Vishny (2002) found that investor protection is positively related to a firm’s sales growth and valuation. In addition, Chhabra et
al. (2009) found that investor protection is negatively related to the cost of capital of a firm.
Thus, in countries where investor protection is higher, companies tend to grow faster and the cost of capital tends to be lower. When this cost of capital is low, however, any small reduction or increase has a larger influence on operations than the same change for a higher cost of capital. Thus, keeping in mind the role the cost of capital takes in the relationship between corporate growth and strategic risk disclosure quality, it is expected that the degree of investor protection influences the previously hypothesized relationship positively.
Taking together above arguments, it is expected that investor protection has both a direct, positive impact on strategic risk disclosure quality and a moderating effect on the association between corporate growth and strategic risk disclosure quality by making the relationship more positive when investor protection is higher. Formally:
Hypothesis 2: There is a positive association between investor protection and strategic risk disclosure quality.
Hypothesis 3: The association between corporate growth and strategic risk disclosure quality becomes more positive as the degree of investor protection increases.
3. METHODOLOGY
This section provides information on the data and methods used in this study. In sequential order I describe the sample; the dependent, independent, moderating and controlling variables; and the regression model.
3.1. Sample and data collection
The initial sample consists of Fortune Global 500 companies, from which all 117 finance and insurance companies were excluded for several reasons. First, these companies should be studied independently as they focus significantly more on risk as opposed to non-financial firms (Linsley & Shrives, 2006; Barakat & Hussainey, 2013; Elshandidy et al., 2018). Second, it is expected that financial and non-financial firms have significantly different types of risks and, in turn, disclosures, which would create empirical bias (Khlif & Hussainey, 2016). Third, regulations regarding risk management are much more exhaustive for the financial industry than for other industries (Chockalingam et al., 2018).
16 Next, only companies from the United States (US), the United Kingdom (UK) or Europe were included, which led to the exclusion of another 192 companies. First, as opposed to only examining one country, the combination of these countries makes it possible to study the effect of investor protection by means of differences in legal systems (see section 3.4). Second, as opposed to including Asian and Latin American studies, the chosen countries have comparable disclosure regulations (Elshandidy et al., 2018), cultures (Zaman Mir, Chatterjee, & Shiraz Rahaman, 2009), and economies. Third, as opposed to including other Western countries, the chosen combination enhances comparability with other disclosure studies, which tend to focus mostly on the US and European countries (including the UK).
As a result of above exclusions, 191 companies remained in the sample. Due to time constraints regarding the data collection for the dependent variable, the sample was randomly narrowed down to 125 firms, giving an initial sample of 375 firm-year observations for the years 2015 to 2017. All companies are publicly listed, which is a necessity as the dependent variable will be hand-collected by means of assessing the companies’ annual reports (see section 3.2). The time period has been chosen for data availability reasons, and because possible time-specific effects will be largely avoided (Guenther et al., 2016). For example, corporate scandals (early 2000s) and the 2008 financial crisis have been found to dramatically influence disclosure practices (Quon et al., 2012).
Due to non-availability of data, another 14 observations were omitted from the sample. As a result, the final sample consists of 361 firm-year observations for 122 unique firms from 14 countries over the period from 2015 to 2017. The number of firm-year observations from the US, the UK, and Europe were 198, 42, and 121, respectively. All variables are explained and described below, and summarized in table 1. Unless otherwise described, all variables are extracted from the Orbis database or, in case of missing information, extracted from the company’s annual report.
3.2. Strategic risk disclosure quality
Disclosure studies often face difficulties in measurement approaches (Beattie, McInnes, & Fearnley, 2004) as “disclosure is an abstract construct that does not possess inherent characteristics by which one can determine its intensity or quality” (Kavitha & Nandagopal, 2011: 31). The two methods that are deemed most objective, and are therefore used most often, are content analyses and disclosure indices. The former method, however, is often criticized for being relatively unreliable (Kavitha & Nandagopal, 2011) and for particularly representing the quantity of information, which is not necessarily a proxy for its
17 quality (Beretta & Bozzolan, 2004; Beattie et al., 2004). Therefore, this study uses a self-constructed index to measure strategic risk disclosure quality (SRDQ), as such an index is far more suitable for capturing the usefulness of information (Scaltrito, 2015) and because it has been found to be a valuable research tool (Beattie et al., 2004).
The index used in this study was constructed in collaboration with four fellow students, with whom I also cooperated to collect the data for measuring SRDQ. In order to minimize the level of subjectivity, we used a dichotomous or unweighted method (Kavitha & Nandagopal, 2011), which means that each item could only score 0 (absence) or 1 (presence). Moreover, in order to increase the level of reliability, we discussed the coding instructions in detail and each item was coded by two different students. In case of inconsistencies, the specific item was discussed extensively by the two coders involved. SRDQ was constructed in a process that involves three steps and is similar to other disclosure studies (e.g., Deumes & Knechel, 2008; Hooghiemstra et al., 2015). The details are outlined below.
First, we identified several types of strategic risk to be included in the index. These were selected based on the definition of strategic risks as proposed in section 2.1 (i.e., risks that affect an organization’s long-term goals and their related activities); extant literature discussing types of (strategic) risks; and the condition that strategic risks cannot be hedged like, for example, financial risks (Miihkinen, 2012). Ultimately, this led to an index that comprises eight distinct risk types, as explained next and shown in the Appendix (Panel A). The first risk type is the environment, as environmental issues are increasingly important for companies and can have severe reputational or economic consequences (Holland and Foo, 2003). Moreover, environmental risks are viewed as an integral part of a company’s strategic management (Anderson, 2002), making them indispensable for the index. Two other important strategic risk types, which are also often specifically categorized as such (e.g., McGee et al., 2010; Linsley & Shrives, 2006; Miihkinen, 2012) are the industry and competition (risk types 2 and 3, respectively). In this respect, Byrne (2014) states that new entrants and rivalry among the competition are main threats to a firm’s strategy (see also, Porter, 2008).
Technological developments (risk type 4) have been described as relatively new risks (McGee et al., 2010), but are increasingly important to companies’ strategies due to companies’ high dependence on technology, both from an innovation and a security point of view. Next, according to Reger, Duhaime and Stimpert (1992), government regulations affect strategic choices and, in turn, a company’s risks and performance. Moreover, companies have to comply with an increasing number of standards and regulations, and non-compliance will
18 be noticed by auditors increasingly faster. In turn, such non-compliance may have large reputational or economic consequences for companies. In addition, Linsley and Shrives (2006) and Miihkinen (2012) have specifically characterized the impact of regulations as a strategic risk. Therefore, compliance with regulations (risk type 5) is added to the index.
Two other not to be underestimated forces that determine the long-term success of a company are its suppliers and customers. Michael Porter’s well-known five-force strategy model suggests that the bargaining power from both suppliers and customers are important to a firm’s strategy (Porter, 2008). Moreover, according to McGee et al. (2010), changing patterns of demand (by customers) and business relationships (with suppliers) are large strategic risk factors. In addition, Sadrieh and Voigt (2017) state that, due to information asymmetry, the relationship between a company and its suppliers is a major source of strategic risk. As a result, a company’s dependence on suppliers and customers (risk types 6 and 7, respectively) are included in the index (Miihkinen, 2012).
The eighth and last strategic risk type is organizational competencies. According to Civelli (1998), it is difficult to recognize, discover, identify and differentiate competencies and capabilities, although they are vital for success. Similarly, Dewett (2004) states that creativity of top managers is closely related to a company’s strategic posture, as it enhances their problem-solving skills. Moreover, organizations rely on creativity and innovation to maintain their competitive position (Dewett, 2004).
After having identified these eight strategic risk types, the second step in the process of constructing the index was measuring companies’ reporting quality by examining their annual reports. We used the risk sections of the annual reports as our only source of information for two reasons. First, restricting ourselves to information in annual reports increases the comparability with prior studies on (strategic) risk disclosure, and second, annual reports are the main source of information for investors and other stakeholders (Scaltrito, 2015; Hooghiemstra et al., 2015).
Where other studies often only look at the presence of the identified index items (e.g., Hooghiemstra et al., 2015; Deumes & Knechel, 2008), we believed the presence of information on a risk type is not the only determinant of the quality of disclosed information. For example, Arvidsson (2011) notes that investors especially appreciate disclosed information when the company explains what role the information plays in value-creation processes and the company’s strategy. Therefore, we looked at: (1) the presence of a risk that can be categorized as one of the eight risk types of the index; (2) the presence of information on the likelihood that such an event would occur; (3) the presence of information on the
19 possible impact for the company were the event to occur; and (4) the presence of information on the evolvement of the specific risk. This is also shown in the Appendix (Panel B).
Finally, firms were assigned one point for the presence of each of the four above-mentioned items for each of the eight strategic risk types. Each risk type and each item is equally weighted for reasons discussed before, so the final index score for any observation is the sum of all dichotomous scores. As a result, SRDQ can have values ranging from zero, reflecting poor disclosure quality, to 32, reflecting good disclosure quality. The reliability of the index measure was assessed by calculating Cronbach’s alpha, which judges the index’s internal consistency. The value of Cronbach’s alpha in this research is 0.7, which is exactly equal to the generally accepted threshold (Elshandidy, Fraser, & Hussainey, 2013). This means that the different items measure the same construct well and that the computed disclosure scores are sufficiently reliable. In addition, deletion of any of the eight risk types would result in a lower alpha, except for environment. However, the deletion of this variable would increase the alpha to 0.71, which is only a marginal change. Lastly, the average and maximum inter-item correlations are 0.24 and 0.52, respectively, suggesting that the items are independent from each other.
3.3. Corporate growth
With regard to the main independent variable, there are several methods of measuring firm growth (Miroshnychenko et al., 2019). Some studies take into consideration the change in total assets (e.g., Prencipe, 2004), while others measure firm growth by combining several distinct indicators, such as changes in the levels of employment, productivity, sales, research and development expenditures, and profits (e.g., Coad & Rao, 2008; Coad, Segarra, & Teruel, 2016; Audretsch, Coad, & Segarra, 2014). However, the most suitable measure of growth is considered to be sales growth, due to its ability to reflect performance from both the short and the long term (Miroshnychenko et al., 2019). Therefore, corporate growth (GROWTH) in this research is measured by taking the one-year percentage change of net sales. Specifically, the growth value in the current year is equal to the percentage difference of the reported net sales between the previous fiscal year and the current fiscal year.
3.4. Investor protection
The second independent variable, which also serves as a moderator in this study, is investor protection. Generally, this variable is proxied in two ways (La Porta, Lopez‐de‐ Silanes, Shleifer, & Vishny, 2002), either by using the anti-director rights index (e.g.,
20 Hooghiemsta et al., 2015) or by distinguishing between countries’ legal systems (e.g., Stulz & Williamson, 2003). La Porta et al. (2002) find evidence that there is a strong correlation between both measures, making the choice between both measures of less relevance. In this study, investor protection (INVP) is measured in the latter way.
A country’s legal system can be specified as either common or civil law (Stulz & Williamson, 2003). The main difference between these two types is that there is stronger political influence in civil law countries (Guenther et al., 2016; Stulz & Williamson, 2003). As a result, there is more demand for transparency in common law countries (Khlif & Hussainey, 2016), which results in better investor protection in common law countries than in civil law countries (Stulz & Williamson, 2003). In this light, several studies on company disclosures indeed capture a country’s legal system by distinguishing between common and civil law countries (e.g., Ruland, Shon, & Zhou, 2007; Shi et al., 2012; Guenther et al., 2016; Khlif & Hussainey, 2016). Consistent with these studies, INVP in this study is a dichotomous variable that takes the value 1 when the country’s legal system is based on common law and 0 when it is based on civil law. The distinction between the two law types and the corresponding countries is based on Porta, Lopez-de-Silanes, Shleifer and Vishny (1998), where common law countries have an English origin and civil law countries have a German, French or Scandinavian origin.
3.5. Control variables
Several control variables are added in the study as these variables have been found to significantly impact risk disclosure practices in extant literature. Taking these variables into consideration makes the study account for factors that are related to the dependent variable. First, the size of a firm is one of the most often researched firm characteristics that influence (risk) disclosure practices, and is nearly always found to be positively related with disclosure practices (Linsley & Shrives, 2006; Ahmed & Courtis, 1999; Prencipe, 2004). Reasons for this are that proprietary costs are lower for larger companies (Prencipe, 2004) and because large firms are more complex and are characterized by a greater level of public attention (Khlif & Hussainey, 2016). Consistent with previous research, the size of a firm (SIZE) is measured by total sales, which is converted to its natural logarithm due to non-linearity.
Second, Gaio (2010) found that the level of economic development has a positive impact on accounting quality, since a poor economic environment may provide a less developed infrastructure or may provide firms with too many costs to verify accounting information.Therefore, this study includes the natural logarithm of the gross domestic product
21 per capita (GDP) into the analysis. Third, LEV is a variable that denotes the leverage of the firm. As the rate of a firm’s leverage increases, companies are generally more willing to disclose information because it will reduce the increased levels of information asymmetry (Prencipe, 2004), and because these firms are more dependent on investors (Guenther et al., 2016). Fourth, financial performance and profitability are added as they are found to be positively related to disclosure quality. Firms with high profitability and good performance have greater incentives to signal that they are able to manage their risks successfully (Khlif & Hussainey, 2016), and they have the resources in place to manage risks (Guenther et al., 2016). Consistent with extant literature, these two firm characteristics are captured by the return on equity (ROE).
TABLE 1 Variable definitions
Variable Description
SRDQ Index that measures the quality of strategic risk disclosures. The index captures four items for each of eight distinct risk types, and can have values ranging from 0 (no disclosures, reflecting poor disclosure quality) to 32 (full disclosure, reflecting good disclosure quality). The eight risk types and four coding items are discussed and explained in the Appendix.
GROWTH Percentage change of sales from previous year to current year.
INVP Dummy variable that takes the value 1 if the company is based in a common law country and 0 when it is based in a civil law country.
SIZE Natural logarithm of the total level of sales.
GDP Natural logarithm of the gross domestic product per capita.
LEV Ratio of the total book value of long-term debt to the total book value of shareholders’ equity.
ROE Ratio of net income to the total book value of shareholders’ equity.
CASHF Natural logarithm of the total level of cash flow.
ORIGIN Dummy variable that takes the value 1 if the company has its origin in the United States and 0 otherwise.
Year dummies Y2015 and Y2016 are dummy variables that take the value 1 if the
annual report is from fiscal year 2015 or 2016, respectively, and 0 otherwise. Fiscal year 2017 (Y2017) is the reference group in the statistical analyses.
Industry dummies AMC, MANUF, TRANSPORT and TRADE are dummy variables that
take the value 1 if the company is active in the industry and 0 otherwise. OTHER is the reference group in the statistical analyses.
22 Fifth, this study controls for the size of a firm’s cash flows (CASHF). Several arguments in this study are related to the role of investments and the cost of capital. Recalling, faster-growing firms are expected to increase their investments, and hence benefit more from a lower cost of capital. Disclosing more and better strategic risk information could help these firms achieve a lower cost of capital through reduced information asymmetry. However, the availability of internal funds influences the dependence on external capital (Miroshnychenko
et al., 2019), and therefore the size of the benefits of a decrease in the cost of capital. That is,
when firms experience larger cash flows, they will use more internal funds and fewer external funds to finance their investments (Miroshnychenko et al., 2019). In addition, Heinle and Smith (2017) also note the importance of cash flows in relation to the cost of capital and risk disclosure practices. Therefore, the level of a firm’s cash flow is expected to influence the degree of information provision. CASHF is logarithmized due to non-linearity.
Other factors that have been found to influence (risk) disclosure practices are auditor size (Ahmed & Courtis, 1999; Ashbaugh-Skaife, Collins, & Kinney Jr, 2007) and cross-listings (Roosenboom & Van Dijk, 2009; Miihkinen, 2012). However, nearly all firms in the final sample are audited by one of the four biggest auditor companies, and all companies are listed on foreign exchanges, making it nearly impossible to control for these factors. Instead, this study controls for companies having their origin in the United States (US). The US SEC has mandated large domestic companies to present their annual report on Form 10-K, which includes more non-financial information on the company’s finances, opportunities, and performance (Campbell et al., 2014). In one of the first sections of the report, companies must extensively discuss the most significant risk factors, and this has been found to be valuable for investors (Campbell et al., 2014). Therefore, this study controls for companies’ origin (ORIGIN) by taking the value 1 if the company is from the United States and 0 otherwise.
Lastly, in order to control for time-series and industry differences, dummy variables are created for each of the three years (Hooghiemstra et al., 2015) and for industries based on their SIC codes. The variable equals 1 if the company belongs to a particular sector and 0 otherwise. The industry types are agriculture, mining and construction (AMC, SIC 0100-1799); manufacturing (MANUF, SIC 2000-3999); transportation, communication and utility services (TRANSPORT, SIC 4000-4999); trade (TRADE, SIC 5000-5999); and other (OTHER). In the statistical analyses, fiscal year 2017 and industry type OTHER are the reference groups. This prevents the model from having exact collinearity, which is often referred to as falling into the dummy variable trap.
23 3.6. Model of analysis
All three hypotheses are tested by means of an ordinary least squares (OLS) method, and are run through IBM SPSS Statistics 25 software. The empirical model is shown as model (1) and looks as follows:
SRDQi, j = β0 + β1·GROWTHi, j + β2·INVPi + β3·SIZEi, j + β4·GDPi, j + β5·LEVi, j
+ β6·ROEi, j + β7·CASHFi, j + β8·ORIGINi + β9·Y2015i, j + β10·Y2016i, j + β11·AMCi
+ β12·MANUFi + β13·TRANSPORTi + β14·TRADEi + εi, j (1) where SRDQi,j is strategic risk disclosure quality of firm i for year j; GROWTHi,j is equal to the percentage difference between the unlogarithmized values of SIZEi,j and SIZEi,j−1; ε is the error term; and all other variables are defined in table 1.
4. RESULTS
This section presents the results of the model that was described in the previous section. In sequential order, I present descriptive information of the gathered data, the results of the main analysis, and the meanings of several additional tests and analyses.
4.1. Descriptive statistics
Table 2 presents the descriptive statistics. The dependent variable, SRDQ, has a mean value of 15.5 and minimum and maximum values of 4 and 28, respectively. The median value is 16. Further, GROWTH values range from −67 to 131 percent, with a mean and median value of 3 and 1.6 percent, respectively. Further, the percentage of firm-year observations that take either side from any of the dichotomous variables can be derived from the column ‘Mean’ in table 2.
For eight of the firm-year observations, a negative cash flow was obtained. Since logarithmic values cannot be calculated for negative numbers, I added a constant to the cash flows of all observations. The amount of this constant was equal to exactly the largest negative cash flow. Ultimately, this made all values non-negative. Even though the addition of this constant does alter the mean of CASHF, it does not change the parameters or variances. Therefore, it is deemed a better option for dealing with negative values in logarithm transformations than using missing values (Hill, Griffiths, Lim, & Lim, 2008). It did, however, result in the omission of one value as the logarithm of zero is not defined.
24 TABLE 2
Descriptive statistics
Variable Observations Mean Std. Dev. Minimum Maximum
SRDQ 361 15.48 4.15 4.00 28.00 GROWTH 361 3.07 19.29 −66.76 131.19 INVP 361 .69 .46 .00 1.00 SIZE 361 24.60 .68 22.94 26.94 GDP 361 10.82 .25 9.44 11.54 LEV 361 1.28 2.88 −15.90 30.60 ROE 345 33.31 88.21 −134.03 942.09 CASHF 360 23.04 .52 21.32 24.80 ORIGIN 361 .55 .50 .00 1.00 AMC 361 .06 .25 .00 1.00 MANUF 361 .47 .50 .00 1.00 TRANSPORT 361 .17 .37 .00 1.00 TRADE 361 .19 .39 .00 1.00 OTHER 361 .11 .31 .00 1.00
Table 3 presents the correlation matrix for all variables as they are used in the analyses, and shows that none of the correlation coefficients is greater than the absolute value of 0.75. This, in turn, means that there is no indication of strong multicollinearity. However, there are five correlation coefficients that are greater than the absolute value of 0.5, which indicates at least some degree of multicollinearity. For example, ORIGIN, INVP, and GDP are relatively highly correlated with each other. This is intuitive because the United States constitute more than half of the sample, and in turn solely indicates the variable ORIGIN; for a large part indicates INVP (80 percent); and belongs to the group of countries with the highest GDPs. To verify that these moderate correlation values do not lead to multicollinearity issues in the following analyses, I obtained variance inflation factors. The largest value is 2.48, which is well below the commonly accepted threshold of 3 (Hooghiemstra et al., 2015). This proves that multicollinearity is not an issue in the following analyses.
25 TABLE 3 Correlation Matrix 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 1. SRDQ 1.00 2. GROWTH −.01 1.00 3. INVP .18 −.02 1.00 4. SIZE −.07 .01 .02 1.00 5. GDP .23 −.01 .69 .10 1.00 6. LEV −.04 .01 .11 .03 .07 1.00 7. ROE .01 .03 .16 .07 .11 .65 1.00 8. CASHF .04 .08 .03 .56 −.00 −.03 .10 1.00 9. ORIGIN .00 −.01 .74 .13 .68 .12 .18 .06 1.00 10. Y2015 −.04 −.28 .02 −.04 −.04 −.01 −.05 −.05 .02 1.00 11. Y2016 −.01 −.14 −.02 −.06 −.03 .05 −.02 −.38 −.01 −.50 1.00 12. AMC −.03 −.09 −.17 .07 −.21 −.07 −.07 .10 −.22 −.01 .01 1.00 13. MANUF .05 −.02 −.00 −.06 .12 −.10 −.00 .09 .03 .01 −.01 −.25 1.00 14. TRANSPORT −.03 −.07 −.18 −.03 −.27 .17 −.01 .06 −.27 .01 −.00 −.12 −.42 1.00 15. TRADE −.05 .05 .19 .14 .15 .09 .05 −.27 .19 −.01 .01 −.13 −.46 −.22 1.00
Correlations greater than the absolute value of 0.13 are statistically significant at the 0.01 level; correlations with absolute values between 0.10 and 0.13 are statistically significant at the 0.05 level. All variables are described in table 1, and the descriptive statistics can be found in table 2.
26 4.2. Hypothesis testing
Table 4 presents the results of the OLS data regressions. First, column (1) presents the results for the control variables, after which column (2) presents the results for hypothesis 1. In this latter column, no significant association is found between corporate growth and SRDQ (p = .60). This finding leads to the rejection of hypothesis 1, which in turn suggests that no association between corporate growth and strategic risk disclosure quality exists. Column (3) presents the results for hypothesis 2, which posited a positive association between investor protection and SRDQ. This association is indeed found significant and positive (β = 2.11,
p = .006), which leads to the acceptance of hypothesis 2. It should be noted, however, that the
95% confidence interval lies between 0.70 and 3.68. This relatively large range makes it difficult to define how large the impact of investor protection exactly is. Ceteris paribus, the degree of investor protection could result in a higher index score of not even 1 point or as much as almost 4 points. On our 32-point index, this is an important difference. Finally, column (4) assesses the moderating effect of investor protection on the association between corporate growth and strategic risk disclosure quality. This proposed relationship, as captured by hypothesis 3, is not found significant (p = .93), so this hypothesis is rejected.
Apart from these main findings, the results in table 4 also show some other interesting effects. For example, ORIGIN is highly significant throughout the analyses (all p = .00), but its sign is not consistent with expectations (β < −2.33). Firms originating in the US were expected to score higher on disclosure quality as they are mandated to present their annual report on Form 10-K, which obligates them to extensively report on their risk factors. A possible explanation for this inconsistency is that US firms do not use tables or graphs that denote quantitative impacts and likelihoods, while we noticed that firms from other countries regularly do use these to inform investors. As a result, these latter firms would, on average, score easier on the coding items impact, likelihood, and evolvement (see Appendix panel B).
Further, GDP is highly significant throughout the analyses (all p = .00), and its sign is in line with previous research. This suggests that a country’s economic development positively impacts a firm’s strategic risk disclosure quality. SIZE is also significant in the analyses, however its sign is inconsistent with prior research. Where most studies expect and empirically find a positive relationship, in my analysis company size adversely impacts disclosure quality. This may be due to the large average firm sizes in the sample, as this reduces the size effect on proprietary costs and, consequently, disclosure quality (Prencipe, 2004). Lastly, year and industry effects do not significantly influence a company’s strategic risk disclosure quality.
27 TABLE 4
Main regression results
(1) (2) (3) (4) CONSTANT −63.402*** (16.654)** −63.104*** (16.682)** −49.856*** (17.197)** −49.424*** (17.327)** GROWTH −.007*** (.013)** −.006*** (.013)** INVP 2.106*** (.760)** 2.108*** (.762)** MODERATOR .018 *** (.206)** SIZE −.928*** (.413)** −.947*** (.414)** −.689*** (.418)** −.706*** (.420)** GDP 7.226*** (1.175)** 7.178*** (1.180)** 5.801*** (1.272)** 5.743*** (1.284)** LEV −.070*** (.147)** −.064*** (.147)** −.082*** (.146)** −.077*** (.147)** ROE .001 *** (.003)** .001*** (.004)** .001*** (.003)** .001*** (.003)** CASHF 1.100*** (.563)** 1.136*** (.568)** .893*** (.563)** .629*** (.569)** ORIGIN −2.362*** (.628)** −2.375*** (.629)** −3.332*** (.714)** −3.342*** (.716)** Y2015 −.210*** (.539)** −.341*** (.593)** −.285*** (.534)** −.419*** (.592)** Y2016 −.050*** (.536)** −.157*** (.573)** −.088*** (.531)** −.199*** (.572)** AMC −.427*** (1.183)** −.544*** (1.204)** −.636*** (1.173)** −.751*** (1.197)** MANUF −.475*** (.804)** −.533*** (.812)** −.515*** (.796)** −.563*** (.811)** TRANSPORT −.159*** (.986)** −.256*** (1.004)** −.407*** (.981)** −.501*** (1.000)** TRADE −.363*** (.937)** −.387*** (.939)** −.671*** (.934)** −.693*** (.938)** R-squared .121*** .121*** .141*** .141***
Standard errors are shown in parentheses.
