Board composition and the influence on the
risk disclosure quality of European banks
MSc Accountancy & MSc Controlling – Combined Master’s Thesis University of Groningen, Faculty of Economics and Business
Lisanne Velthuis s2568608 Saturnuslaan 29 9742 EB Groningen +31 6 34 01 82 03 l.velthuis@student.rug.nl
Supervisor: J.G. Huttenhuis MSc EMA RA Co-assessor: Prof. Dr. R.L. ter Hoeven RA
Word count: 13.135 25 June 2018
Page | 1 ABSTRACT
The financial crisis of 2008 revealed the shortcomings in the banking industry. Basel III regulations were introduced, whereby banks had to comply with stricter requirements in terms of e.g. solvency, liquidity and disclosure, and corporate governance principles were revised. This research examines the influence of board composition on the risk disclosure quality of European banks. Using a sample of 75 European banks in 2017, the findings show a positive significant influence of supervisory board diversity, independence and activity on the quality of risk disclosures. No effect was found for management board diversity, supervisory and
management board size and audit committee activity. These results could be used by banks, shareholders, supervisors and regulators to increase the risk disclosure quality by recruiting new directors and setting new standards for the composition and activity of the supervisory board.
Keywords: risk disclosure quality, board composition, board diversity, board size, board
Page | 2 TABLE OF CONTENTS
1. INTRODUCTION ... 4
1.1. Introduction ... 4
1.2. Scientific contribution and practical relevance ... 5
2. INSTITUTIONAL CONTEXT ... 8
2.1. Overview of Basel III and CRR/CRDIV ... 8
2.2. Overview of IAS and IFRS ... 9
2.3. Governance structure ... 10
3. LITERATURE REVIEW & HYPOTHESES DEVELOPMENT ... 11
3.1. Quality of risk disclosures ... 11
3.2. Legitimacy theory ... 11
3.3. Stakeholder theory ... 12
3.4. Resource dependency theory ... 13
3.5. Agency theory ... 13
3.6. Voluntary disclosure theory ... 14
3.7. Hypotheses development ... 15
3.7.1. Board diversity ... 15
3.7.2. Board size ... 16
3.7.3. Supervisory board independence ... 17
3.7.4. Supervisory board and audit committee activity ... 18
3.8. Conceptual Model ... 20
4. RESEARCH METHODOLOGY ... 21
4.1. Data and sample collection ... 21
4.1.1. Original data and sample ... 21
4.1.2. Additional data and sample ... 22
4.2. Dependent variable: quality of risk disclosures ... 23
4.3. Independent variables ... 24 4.4. Control variables ... 24 4.4.1. Bank size ... 25 4.4.2. Leverage ... 25 4.4.3. Profitability ... 25 4.4.4. Legal system ... 26 4.5. Statistical Model ... 26 5. RESULTS ... 28
Page | 3 5.1. Descriptive statistics ... 28 5.2. Correlation matrix ... 30 5.3. Statistical analysis ... 30 5.4. Additional analysis ... 33 5.4.1. Robustness test ... 33
5.4.2. Additional analysis: sample 2015-2017 ... 34
5.4.3. Additional analysis: comparison of risk disclosure quality ... 36
6. DISCUSSION AND CONCLUSION ... 37
6.1. Conclusion ... 37
6.2. Theoretical and practical implications ... 38
6.3. Limitations and future research... 38
7. REFERENCES ... 40
Page | 4 1. INTRODUCTION
1.1. Introduction
The financial crisis of 2008 exposed the shortcomings in the banking system and seriously damaged the confidence in the banking industry (NBA, 2014). After the bankruptcy of Lehman Brothers in the United States, several large European banks, such as ING, Fortis bank/ABN AMRO and Royal Bank of Scotland, had to be nationalised or granted state aid to survive (NOS, 2013; The Guardian, 2008). Consequently, the demand and attendance for risk disclosure quality increased (Linsley et al., 2008). To restore the confidence in banks, increasing the quality of risk information is important. In addition, corporate governance improvements are also important to ensure effective risk management and well-functioning of the bank (BCBS, 2015a; Farag & Mallin, 2017). The necessity to improve risk disclosures and corporate governance, together with the lessons learned from the financial crisis, resulted in revised Basel regulation (Basel III) and establishment of the European Banking Authority (EBA) for the banking sector (Al-Maghzom et al., 2016; Official Journal of the European Union, 2010). To increase transparency, to regain trust and to create a resilient banking system, disclosing relevant information about (financial) risks is crucial to avoid a new banking crisis (Al-Maghzom et al., 2016; BCBS, 2017a; FSB, 2012). As the risk disclosure quality of banks is, amongst others, dependent on the decisions of the supervisory and management board, the corporate governance is of importance. Due to the demand to improve the corporate
governance, Farag and Mallin (2017) mention that board composition has been of greater attention lately.
In 2016, PricewaterhouseCoopers (PwC) investigated the board composition of several industries, including the banking sector and entities involved in capital markets. One of their findings was that on average 26% of all directors in the boards of the 21 largest banks and capital markets entities are female directors, which is above the average of the S&P 500 firms (20%) and one of the industries with the highest percentage. This is surprisingly, as the banking industry is mostly seen as a world that is dominated by men who are making the decisions (Wyman, 2016). Since there are currently no clear guidelines and accounting
requirements of which regulatory disclosures of the Basel III framework have to be included in the financial statements of the banks, voluntary disclosure of this information may be
dependent upon the composition of the supervisory and management board1.
In their daily operations banks have to deal with risks which are inherent to their business. Disclosure of risk information is therefore necessary for investors to assess the risk profile of banks (Ashfaq et al., 2016; Linsley & Shrives, 2005). The Basel III framework is developed
1
Hereafter referred to as ‘board’. Following the FSB (2013), “the term board refers to the oversight function and the management function in general and should be interpreted throughout the document in accordance with the applicable law within each jurisdiction, unless explicitly noted or otherwise indicated by the context” (p.2).
Page | 5 “in response to the financial crisis […] to strengthen the regulation, supervision and risk
management of the banking sector” (BCBS, 2017a, p.1). Basel III focusses, amongst others, on risk disclosures to promote market discipline and enhance transparency (BCBS, 2015b). Trust is an important factor upon which the banking industry is based, whereby effective corporate governance is a prerequisite to achieve and maintain the public confidence and trust in banks (Barakat & Hussainey, 2013; BCBS, 2006). Consequently, due to its risk-based nature and importance of trust, risk disclosures of banks should be separately viewed from the disclosures of non-financial institutions (Bessis, 2002). Informative risk disclosures are essential to
enhance market discipline and show sound corporate governance (BCBS, 2006, 2015b). However, Basel III permits the board of the banks discretion in the quality of reporting risks in the annual or risk report, as it is not mandatory to audit these disclosures nor to include them (BCBS, 2015b). This area for voluntary disclosure offers an opportunity to investigate the quality of risk disclosure decisions the board has taken and the influence of board
characteristics on those decisions.
The level of requirements for corporate governance is the highest in the banking industry compared to other industries (Al-Maghzom et al., 2016). An effective governance structure is essential for managing risks and the overall functioning of the bank, leading to more risk disclosures (BCBS, 2015a). Therefore, this paper will discuss what the influence of board composition is on the risk disclosure quality of European banks. The aim is to identify the characteristics of board composition that result in high quality risk disclosures by which valuable information is provided to users. The focus of this research will be on the supervisory and management board. The supervisory board is responsible for monitoring management, oversight of risk governance and accountable for approving the financial report (BCBS, 2015a; FSB, 2013). Hence, the supervisory board may prevent that information is withhold and
influence the level and quality of risk disclosures (BCBS, 2015a; Donnelly & Mulcahy, 2008). The management board is responsible for establishing and executing the risk management policy and preparing the financial report (BCBS, 2015a). Hereby, managers are able to influence the quality of risk information and might even withhold risk information for the supervisory board and other stakeholders. Using a combination of legitimacy, stakeholder, resource dependency, agency and voluntary disclosure theory, the influence of board diversity, board size, supervisory board independence and activity and audit committee activity on the risk disclosure quality of European Banks will be studied. Therefore, the research question is:
Does board composition have an influence on the risk disclosure quality of European Banks?
1.2. Scientific contribution and practical relevance
Previous research often examined the influence of company size, leverage, profitability and management ownership on risk disclosures (e.g. Albitar, 2015; Al-Maghzom et al., 2016; Ashfaq et al., 2016; Linsley & Shrives, 2006; Oliveira et al., 2011b). Prior studies examining the relation between board composition and quality of risk disclosures are limited. Barakat and
Page | 6 Hussainey (2013) found, e.g., that more outside directors in the supervisory board and an audit committee that is more active, contribute to a higher quality of operational risk disclosures by European banks. Other studies investigating board characteristics focus on the quantity of risk disclosures (e.g. Abraham & Cox, 2007; Al-Maghzom et al., 2016; Elshandidy et al., 2015; Ntim et al., 2013). According to Miihkinen (2012), the quantity of disclosures is partly explaining the disclosure quality. For this reason these studies are also of relevance for this research. In addition, board composition has often been examined in relation with level of risk disclosures in a single country (e.g. Al-Maghzom et al., 2016; Ashfaq et al., 2016) and/or for non-financial entities (e.g. Allini et al., 2016; Macchioni et al., 2014; Ntim et al., 2013; Oliveira et al., 2011a).
Consequently, the relation between board diversity and board size and the quality of risk disclosures of European banks is a novel research setting. Furthermore, this paper is also motivated by the paper of Barakat and Hussainey (2013), who suggest more investigation into the contributions of supervisory board directors in improving the risk disclosure quality in the banking industry. The goal of this research is, therefore, to elaborate on this and to further discover what the contributions of the supervisory and management board are for the risk reporting quality of European banks.
This paper contributes in five different ways to the corporate governance and risk disclosure literature, specifically the literature regarding risk disclosure quality of banks. First of all, where most studies concentrate on non-financial entities, this study focusses on financial entities, specifically banks. As mentioned before, risk disclosures of banks should be separately viewed from those of non-financial entities (Bessis, 2002). Secondly, where most studies examine risk disclosure quantity, this study investigates the quality of risk disclosures. Quantitative disclosures form a part of the qualitative disclosures and explain therefore only partially the risk disclosures (Miihkinen, 2012). Focussing on quality is thus important to get a complete overview of risk disclosures. Thirdly, with respect to Barakat and Hussainey (2013), this paper differentiates in that it takes into account the new regulations as included in the Basel III framework, such as the liquidity coverage ratio, minimum leverage ratio requirement and net stable funding ratio in addition to the already present solvency ratios of Basel I and II (BCBS, 1988, 2017c) and uses a comprehensive disclosure index not only related to one type of risk (i.e. operational risk) to measure the quality of risk disclosures. Fourthly, this study also investigates management board characteristics, where other studies only take supervisory board characteristics into account (e.g. Allegrini & Greco, 2013; Elshandidy & Neri, 2015). Finally, prior studies examined risk disclosures for one particular country, whereas this study
Page | 7 Not only does this paper contribute to the existing literature on corporate governance and risk reporting of banks, it also has practical implications. Banks are able to evaluate the
determinants of their quality of risk disclosures and, if necessary, improve it. The supervisor and shareholders might use the results of this research to change the composition of the bank’s board in favour of better risk disclosure quality. Higher quality risk information enables investors and other stakeholders to better assess the risk profile of the bank and adjust their decisions on that information. Differences in board composition and quality of risk disclosures can be used by regulators and standard setters to create best practices and serve as a guideline for setting new laws. For example, it could be interesting input for the new Basel IV
regulations, Corporate Governance Code for banks and regulations for gender balance on boards.
The remainder of this paper is structured as follows. The next section will give an overview of the regulations of risk disclosures and governance structure that are applicable to European banks. The theoretical framework and development of hypotheses will be explained in section three. The fourth section discusses the research methodology, whereas in section five the results are presented and discussed. Finally, the discussion and conclusions of this research will be given in section six.
Page | 8 2. INSTITUTIONAL CONTEXT
In this section, an overview will be given of the regulations of risk disclosures that apply to European banks. The main points of the Basel III framework, legally binding for European banks via the CRR/CRDIV, and IAS/IFRS standards will be discussed. Requirements of these regulations will be used for the construction of the disclosure index (explained in §4.2) to determine the quality of risk disclosures. Additionally, the two different governance structures (i.e. one/two tier) within the EU will be explained.
2.1. Overview of Basel III and CRR/CRDIV
In response to the global financial crisis, the Basel Committee on Banking Supervision (BCBS) issued the Basel III framework, which addresses the shortcomings of the previous Basel II framework. The aim of Basel III is to enhance the financial stability and create a resilient banking industry “to support the real economy” (BCBS, 2017a, p.1). Similar to Basel II, Basel III also applies the three pillar concept and further strengthens these pillars. Pillar 1 represents the enhanced minimum regulatory capital and liquidity requirements. Pillar 2 describes the risk management and supervision process. Pillar 3 requires regulatory disclosures to stakeholders to promote market discipline (BCBS, 2015b). Because of these requirements, the market gains insight into the key information on regulatory capital and risk exposure of banks, which
enhances the bank’s transparency and confidence (BCBS, 2015b). As the Pillar 3 disclosures in the (old) Basel II framework were insufficient in identifying the banks material risks and providing adequate information to the market to assess and compare the overall capital acceptability, a revised version was published in Basel III (BCBS, 2015b). Consequently, one of the main goals of the revised Pillar 3 of Basel III “is to improve the comparability and consistency of disclosures” (BCBS, 2015b, p.1). The BCBS has developed five guiding principles (e.g. comparable, clear disclosures) to achieve transparent, high quality risk disclosures that enable market participants to understand and compare the different risks of banks (BCBS, 2015b).
Through the Capital Requirements Regulation (CRR) and Directive IV (CRDIV) the Basel III framework is applicable to banks in the European Union (EU) as of 1 January 2014. The CRR applies to all EU member states and firms, whereas the CRDIV must be implemented through the national law of each country (Official Journal of the European Union, 2013a, 2013b). However, in fewer than half of the European countries the revised Pillar 3 disclosure requirements are implemented and enforced in national law in July 2017 (BCBS, 2017b). Moreover, Hossain et al. (2018) also state that the extent of implementation of the Basel III pillars differs from country to country, creating a divergent effect. The requirements of Basel III must be fully implemented, ultimately, in 2019 to enable banks to rebuild their capital and liquidity and jurisdictions to implement the requirements into their legislation.
Page | 9 The Basel III Pillar 3 report has to be published in a standalone document “that provides a readily accessible source of prudential measures for users” and shows users that they comply with the Pillar 3 disclosure requirements (BCBS, 2015b, p.2). It is permitted to include the Pillar 3 report in the financial report or a discrete section of it, as long as it is easy to identify (BCBS, 2015b). However, the Pillar 3 report is mostly not included in the financial report, as it is not required, resulting in none or only limited review of an independent auditor.
Additionally, information that seems meaningless to users should not be published by banks, as well as information that is considered as confidential (BCBS, 2015b). Besides that, banks are allowed to voluntary disclose risk information, in line with the five guiding principles, that they find relevant (BCBS, 2015b). As a consequence, through these exceptions and differences, the quality of risk disclosures may vary between banks. This is further strengthened by the fact that the five principles are only guiding, therefore subject to interpretation.
2.2. Overview of IAS and IFRS
In addition to the requirements of Basel III, the annual reports of listed entities in the EU are obliged to meet the International Financial Reporting Standards (IFRS) as of 1 January 2015. Since most banks are debt or shares listed, they are obliged to meet the requirements of International Accounting Standard (IAS) 32, IFRS 9 (or formerly IAS 39 up to 31 December 2017) and IFRS 7 standards. IAS 32 specifies the principles for presentation of financial instruments, IFRS 9 the recognition and measurement of financial instruments and IFRS 7 the disclosure of financial instruments2. For this research, IFRS 7 Financial Instruments:
Disclosures is particular of relevance due to the obligation for banks to share risk information in their annual report (IASB, 2005). The IASB issued IFRS 7 in 2005 with an effective date of 1 January 2007, applicable to all entities using financial instruments. IFRS 7 requires entities to disclose information about “the nature and extent of risks arising from financial instruments […] and how the entity manages those risks” (IASB, 2005, p.2). Moreover, also the
significance of the financial instruments should be disclosed (IASB, 2005). Entities are required to disclose qualitative and quantitative information regarding each type of risk resulting from the use of financial instruments (IASB, 2005). The main objective of these disclosures is to enable users to evaluate the financial position and performance of the entity. There is some overlap between the scope of IFRS 7 and Basel III disclosures. IFRS 7 mandates banks to disclose risks arising from financial instruments, whereas a part of Basel III also requires banks to disclose information about the risks of using financial instruments. Therefore, this overlapping information has to be included in the annual report, making it more efficient for banks to prepare these disclosures once based on IFRS 7. Moreover, including these disclosures on bases of IFRS 7 increases the comparability of the financial statements with the Basel III Pillar 3 report, thereby providing more useful information to stakeholders. Just like
Page | 10 Basel III, IFRS 7 also permits banks discretion in the disclosure of information due to the generic documentation of the standards.
2.3. Governance structure
There are differences in Europe regarding the governance structure of banks. In some countries, such as the Netherlands, a two tier structure is used, consisting of a separate supervisory and management board. The supervisory board fulfils the role of monitoring management and has no executive role, whereas the management board fulfils the executive role (BCBS, 2015a). In other countries a one tier structure is used where both roles are combined in one board, the board of directors, comprising of non-executive and executive directors. Hereby, non-executive directors have an oversight function and executive directors operate the business (BCBS, 2015a). Taking into account these differences, this research does not recommend a particular governance structure. In line with a two tier structure the terms supervisory and management board are used, whereby non-executive directors could be seen as the supervisory board and executive directors as the management board. Throughout this research, the terms used should therefore be interpreted in line with the applicable law within each country.
Page | 11 3. LITERATURE REVIEW & HYPOTHESES DEVELOPMENT
This section will explain the theoretical framework used in this study, starting with the risk disclosure literature. Thereafter, a combination of social theories (i.e. legitimacy, stakeholder and resource dependency theory), agency and voluntary disclosure theory will be explained. These theories will be used to explain how board composition influences risk disclosures, finally leading to the development of different hypotheses.
3.1. Quality of risk disclosures
Risk disclosures can be defined as “all information that firms provide in the risk reviews they present in their annual reports” (Miihkinen, 2012, p.442). This can include opportunities or threats that might affect the entity (Linsley & Shrives, 2005). The quality of risk disclosures is important for stakeholders to obtain the necessary information to gain insight into the bank’s risk profile and risk management (Linsley & Shrives, 2006). Prior studies show that the risk disclosure quality is low, presenting mostly qualitative, generic, historical and good news information (Ntim et al., 2013; Oliveira et al., 2011a; Saggar & Singh, 2017). Investigating Canadian and British banks, Linsley et al. (2006) found that disclosures include limited quantitative and forward-looking risk information, whereas research of UK and Portuguese entities show that disclosures of risk information lacks coherence and readability (Linsley & Lawrence, 2007; Linsley & Shrives, 2006; Oliveira et al., 2011b). Hence, the usefulness of risk disclosures for stakeholders can be questioned due to the existence of an information gap, leading to inadequately assessing the entities’ risk profile (Linsley & Shrives, 2006; Miihkinen, 2012). One of the goals of Basel III is to enhance the comparability and consistency of
disclosures, as explained in §2.1 (BCBS, 2015b). Further, the BCBS (2015) mentioned in Basel III that qualitative information must be used to clarify quantitative information, not replacing it. Studies of Miihkinen (2012) and Bischof (2009) documented that adoption and enforcement of regulation standards (e.g. IFRS 7) increased the quantity of risk disclosures. Consequently, laws and regulations may positively contribute to improve risk disclosures and reduce the risk information gap. One of the main factors of voluntary risk disclosure, according to Oliveira et al. (2011b), is stakeholder monitoring and entities reputation, referring to
legitimacy, stakeholder and resource dependency theory. Other influential factors are agency costs (Oliveira et al., 2011a) and proprietary costs (Linsley & Shrives, 2006), referencing to agency and voluntary disclosure theory. These theories and their incentives for risk disclosures will be discussed next, starting with the three social theories.
3.2. Legitimacy theory
Legitimacy theory states that entities are bound to a social contract with the society to operate its activities within the boundaries set by norms, values and beliefs. In return, the society legitimates the entities’ objectives and ensures its existence (Suchman, 1995). Hence, if the society views that the entity breached the social contract, the entity’s survival is at risk
Page | 12 (Deegan, 2002). Through public disclosures, entities are able to gain, maintain or restore
legitimacy (O’Sullivan & O’Dwyer, 2009). Barakat and Hussainey (2013) mention that banks voluntary disclose risk information to show its organizational legitimacy. Moreover, banks also disclose mandatory risk information to respect the institutional norms set by the society, i.e. the institutional legitimacy (Barakat & Hussainey, 2013). Even compliance with the minimum disclosure requirements enhances legitimacy (Oliveira et al., 2011b). An incentive for banks to disclose risk information is to build up and maintain a good reputation with the surrounding environment (Oliveira et al., 2011b). Linsley and Shrives (2005) mention the following: “if banks recognise that they need to disclose more risk information, then an incentive exists for them to improve their risk management capabilities as they will not want to be viewed as inferior to other banks in this respect” (p.206). Consequently, banks disclose risk information to legitimise their operations and to gain social acceptance to continue their business, otherwise the bank’s activities are at risk (Ntim et al., 2013).
The legitimacy theory can be more operationalised by taking into account the stakeholder and resource dependency theory (Chen & Roberts, 2010). Where legitimacy theory focusses on the society around banks as a whole, the stakeholder theory identifies influential stakeholders of banks within those society, which demand risk disclosures to analyse the amount and extent of risks the bank faces. A combination of different stakeholder perspectives in the bank’s board contributes, according to the resource dependency theory, to obtain essential resources for well-functioning and legitimacy of the bank. These theories also explain why banks (not) choose to disclose risk information and provide a narrower view than the legitimacy theory.
3.3. Stakeholder theory
As mentioned before, the stakeholder theory refines the legitimacy theory by identifying and taking into account the needs of specific influential stakeholders (Gray et al., 1995).
Stakeholder theory indicates that there are different stakeholders with different opinions, expectations and influences on the entity and that all of these stakeholders needs to be served, by creating value for them (Freeman, 1984). According to Freeman (1984), a stakeholder is “any group or individual who can affect or is affected by the achievement of the firm’s objectives” (p.46). The success and survival of the entity is dependent upon satisfying and balancing the various (conflicting) demands of influential stakeholders (Deegan & Blomquist, 2006; Roberts, 1992). Risk disclosures can be used as a tool to manage the different
information needs and to gain approval of influential stakeholders (Freeman, 1984; Freeman & Reed, 1983). Consequently, the risk disclosure policy is adapted to the expectations and needs of (influential) stakeholders (Deegan & Blomquist, 2006). Disclosure of risk information enables banks to better communicate and interact with their influential stakeholders (e.g. supervisors, regulators and depositors) and enables stakeholders to monitor the safe and sound operations of the bank (Bank Millennium, 2014; Barakat & Hussainey, 2013; BCBS, 2015a).
Page | 13 3.4. Resource dependency theory
Resource dependency theory complements stakeholder theory (Ruf et al., 2001) and states that an entity is dependent upon its external environment for resources to survive and grow and competes for these resources with other entities (Barakat & Hussainey, 2013; Pfeffer, 1972). Resources are “anything that could be thought of as a strength or weakness of a given firm” (Wernerfelt, 1984, p.172) and are, e.g., reputation, experience and knowledge (Branco & Rodrigues, 2006). Banks are particularly dependent upon the resources of, e.g., savers and depositors for their funding. The supervisory board can be used to manage the dependency for resources and provide essential resources to the entity, such as advice and counsel, access to channels of information and resources and legitimacy (Hillman & Dalziel, 2003; Pfeffer & Salancik, 1978). According to Carter et al. (2010), boards that are diverse consists of more different experiences, backgrounds, opinions and perspectives, resulting in more available resources. Hereby, directors represent the link between the external environment and the entity (Pfeffer & Salancik, 1978).
Another way to attract essential resources for entities is to disclose more risk information or risk information of a higher quality, thereby increasing transparency and building and maintaining their reputation (Branco & Rodrigues, 2006; Oliveira et al., 2011b; Pfeffer & Salancik, 1978). Increased risk disclosures can be used by banks to gain access to essential resources, such as attracting cheaper cost of capital, better internal risk management and improved reputation (Barakat & Hussainey, 2013; Branco & Rodrigues, 2006; Pfeffer & Salancik, 1978). Consequently, engaging in risk disclosures enables banks to attract and maintain resources and capabilities. To further narrow the incentives for risk disclosures, agency theory is discussed next which focusses on shareholders only.
3.5. Agency theory
Information asymmetry and agency conflicts trigger the demand for disclosures (Healy & Palepu, 2001). Different interests between shareholders and management and an information advantage for managers create information asymmetry and agency problems (Bosse & Philips, 2016; Eisenhardt, 1989; Healy & Palepu, 2001). Information asymmetry enables managers to take decisions in their own interest, which may deviate from the interest of the shareholders (Fama & Jensen, 1983). To align the interest of the shareholders and management, agency costs are incurred to monitor the management (Hill & Jones, 1992; Jensen & Meckling, 1976). Disclosure of information is a mechanism to mitigate information asymmetry and agency problems, resulting in less agency costs for management monitoring and a lower cost of capital (An et al., 2011; Deumes & Knechel, 2008; Fama & Jensen, 1983; Healy & Palepu, 2001; Jensen & Meckling, 1976). Entities could disclose information voluntary or mandatory, through law and regulations, to enable shareholders to monitor and, if necessary, discipline management (Healy & Palepu, 2001).
Page | 14 Risk disclosure quality is particularly relevant for banks due to the complexity of the banks’ activities, which increases information asymmetry and reduces shareholders’ ability to monitor the bank compared to other industries (De Andres & Vallelado, 2008). Given this higher information asymmetry and lower monitoring ability, disclosures of high quality risk
information are essential for shareholders to monitor the bank’s operations. Therefore, banks have a higher incentive to disclose risk information to provide insight into its activities and reduce agency and capital costs. Agency theory provides the narrowest view on incentives for risk disclosure quality due to its focus on shareholders (Abraham & Cox, 2007). There is yet another theory, somewhat overlapping with above mentioned theories, that explains why managers (not) voluntary disclose risk information.
3.6. Voluntary disclosure theory
Voluntary disclosure theory focusses on voluntary disclosure of information in the annual report. Voluntary disclosure can be defined as “disclosures in excess of requirements” and “represents free choices on the part of company management to provide accounting and other information deemed relevant to the decision needs of users of their annual reports” (Meek et al., 1995, p.555). The quality of disclosures is dependent upon the different incentives of managers to (not) voluntary disclose information that is useful for stakeholders (Meek et al., 1995; Miihkinen, 2012). A trade-off is made between the benefits and costs of voluntary disclosing risk information, whereby voluntary disclosure is expected if the benefits exceeds the direct and indirect costs of providing risk information (Healy & Palepu, 2001; Meek et al., 1995).
More voluntary disclosure of risk information decreases information asymmetry and lowers information risks, leading to more confidence of shareholders and a lower cost of capital (Healy & Palepu, 2001; Linsley et al., 2006). In addition, increased voluntary disclosure ensures that shareholders and supervisors are better able to monitor managers and discipline them if necessary (Hooghiemstra et al., 2015; Linsley et al., 2006). Another incentive for managers to voluntary disclose risk information is to establish a reputation for credible reporting (Graham et al., 2005). A requirement for credible reporting is that the information disclosed must be accurate, timely and complete, even if it is unfavourable information (Healy & Palepu, 2001; Hooghiemstra et al., 2015).
The costs incurred for disclosures are called proprietary costs and limit the incentives of entities to voluntary disclose risk information as this information is sensitive and potentially useful for competitors (Meek et al., 1995; Verrecchia, 1983; Wagenhofer, 1990). Other costs associated with voluntary disclosure are information collection, processing and litigation costs (Meek et al., 1995). Litigation risks reduce the incentives of entities to provide risk
information, because disclosure of inaccurate or incomplete information could lead to a lawsuit (Healy & Palepu, 2001; Hooghiemstra et al., 2015). This is particularly the case with
forward-Page | 15 looking risk information, which is in its nature uncertain, potentially leading to claims of
investors who traded based on the information (Linsley et al., 2006; Linsley & Shrives, 2005). All these costs may withhold managers to disclose detailed risk information.
3.7. Hypotheses development 3.7.1. Board diversity
Following from the financial crisis, board diversity attracts attention for improving corporate governance in the banking industry (Farag & Mallin, 2017). However, there is currently limited research of the influence of diversity on the functioning of the banks’ boards (De Cabo et al., 2012). Board diversity can be defined as “varying profiles that may exist in the board members and how the diversity can affect decision-making by the board” (Saggar & Singh, 2017, p.386). In this study, particularly the diversity of gender within the bank’s board will be investigated. Gender diversity is mostly referred to as “the presence of female directors on the boards of the company” (Saggar & Singh, 2017, p.386). Men and women are, according to Liao et al. (2015), “traditionally, culturally and socially different” (p.412) and demonstrate a variation of behaviours and skills (Allini et al., 2016). The BCBS recognizes the importance of diversity, as it states “the board should be comprised of individuals with a balance of skills, diversity and expertise” (BCBS, 2015a, p.13).
More diversity within the board contributes to a better monitoring function and reduces agency costs (Carter et al., 2003; Farag & Mallin, 2017). However, Allini et al. (2016) argue that, according to the agency theory, gender diversity has no clear effect on the board. On one hand, diversity may socially categorize board directors into groups that are similar and dissimilar to themselves, leading to decreased identification with the board (Nielsen & Huse, 2010). On the other hand, stakeholder theory argues that board diversity enables entities to understand and respond to the surrounding environment and its complexities, effectively managing the link with other stakeholders (Boyd, 1990; Carter et al., 2003; Pfeffer & Salancik, 1978). This increases transparency and helps entities to make clever decisions and to adopt the appropriate level and quality of risk disclosures (Carter et al., 2003).
Following resource dependency theory, female directors have different views and talents that lead to other benefits and resources for the entity (Carter et al., 2010; De Cabo et al., 2012). These different views and perspectives of female directors might lead to different and more effective solutions of board discussions and challenges, as more alternatives and their
consequences are evaluated (Barako & Brown, 2008; Carter et al., 2003). Furthermore, female directors can provide the board with unusual perspectives, ask other questions and are more empathic in understanding the dynamics of the market place compared to men, contributing to better decision-making (Carter et al., 2003; Daily & Dalton, 2003). Females also enhance collaboration and decrease interest conflicts between directors, leading to better
Page | 16 exhibiting therefore higher quality of information provision to avoid litigation and reputational risks, and show more discipline in reporting (Srinidhi et al., 2011). Accordingly, a diverse board contributes to better decision-making and alignment of their interest with stakeholders, possibly leading to a higher quality of risk disclosures.
There has only been a few studies investigating the influence of gender diversity on risk disclosures. These studies find a positive influence of diversity on the level of risk disclosures in an Italian, Saudi Arabian and South-African setting (Allini et al., 2016; Al-Maghzom et al., 2016; Ntim et al., 2013), whereas Srinidhi et al. (2011) shows that gender diversity results in higher earnings quality. Based on these prior studies and the discussion above, I expect a positive relation between board diversity and quality of risk disclosures. Therefore, the hypotheses are stated as follows:
Hypothesis 1a (H1a): The diversity of the supervisory board positively influences the quality of risk disclosures.
Hypothesis 1b (H1b): The diversity of the management board positively influences the quality of risk disclosures.
3.7.2. Board size
Board size affects the effectiveness of the board and is able to influence the disclosure of risk information (Allini et al., 2016; Elshandidy & Neri, 2015; Ntim et al., 2013). There has been attention in the literature whether adding directors is more effective in increasing risk
disclosures (Hassan, 2016) and whether a trade-off is present between costs and benefits of more board directors (Lipton & Lorsch, 1992). On one hand, it has been argued that larger boards lead to poor coordination, communication, director free riding, less effective monitoring and less flexibility in decision-making (Cheng & Courtenay, 2006; Hidalgo et al., 2011;
Jensen, 1993). However, there is in the literature no unambiguous explanation for these
arguments. On the other hand, larger boards increase the knowledge base for making decisions, enabling managers and supervisors to take better business decisions (Hou & Moore, 2010; Ntim et al., 2013).
3.7.2.1. Supervisory board
Taken agency and resource dependency theory together, larger boards have more diverse expertise and knowledge available, resulting in more effective monitoring and advising management (De Andres & Vallelado, 2008; Elzahar & Hussainey, 2012; John & Senbet, 1998; Moumena et al., 2016). This positively influences the disclosure policy and ensures reliable and timely disclosure of risk information (Adam et al., 2005; Bozec & Bozec, 2012). Furthermore, larger boards may reduce information asymmetry and agency costs by aligning the interest of the in- and outsiders, promoting voluntary disclosure (Chen & Jaggi, 2000; Elshandidy & Neri, 2015). Stakeholder and resource dependency theory assume that more directors on the board provide more access to the external environment by representing more
Page | 17 stakeholders and contribute to the securing of essential resources, motivating management to provide more risk information (Branco & Rodrigues, 2006; Jia et al., 2009; Ntim et al., 2013). Larger boards, consisting of wider expertise, experience and stakeholder representation, motivate directors to show shareholders their efforts made with respect to risk management (Elzahar & Hussainey, 2012). Moreover, directors with a financial and accounting background are likely to increase in number if the size of the board increases, possibly positively affecting the decisions on voluntary risk disclosures (Elzahar & Hussainey, 2012).
There have been several studies performed who found a positive effect between board size and level of voluntary risk disclosures of banks (e.g. Al-Maghzom et al., 2016; Elshandidy & Neri, 2015; Ntim et al., 2013). Elzahar and Hussainey (2012) also found a positive influence of board size on risk disclosures, although this relation is not significant. Based on these prior studies and the discussion above, I expect a positive relation between supervisory board size and quality of risk disclosures. Therefore, the hypothesis is stated as follows:
Hypothesis 2a (H2a): The size of the supervisory board positively influences the quality of risk disclosures.
3.7.2.2. Management board
The focus in the literature has been on the supervisory board and their effect on risk
disclosures. Hence, limited information is known about the effect of the management board on risk disclosures. Akhtaruddin et al. (2009) argue that a larger board possess more collective expertise and are therefore more competent to fulfil their responsibilities. Moreover, a larger management board possess better capabilities and resources for information processing and decision-making (Haleblian & Finkelstein, 1993). The increase in capabilities and resources, available on the larger management board compared to their smaller counterparts, enhances creativity and critical judgement and contributes to decisions of higher quality (Bantel & Jackson, 1989; Cummings et al., 1974), which may increase the disclosure of higher quality information. Based on the discussion above, I expect a positive relation between management board size and quality of risk disclosures. Therefore, the hypothesis is stated as follows:
Hypothesis 2b (H2b): The size of the management board positively influences the quality of risk disclosures.
3.7.3. Supervisory board independence
For effective monitoring and advising the management board, appointing more directors in the supervisory board is not sufficient (De Andres & Vallelado, 2008). Appointing independent directors are, according to the agency theory, key to effective monitor and control managers (Fama & Jensen, 1983). An independent director is defined as a director “who does not have any management responsibilities within the bank and is not under any other undue influence, internal or external, political or ownership, that would impede the board member’s exercise of
Page | 18 objective judgment” (BCBS, 2015a, p.1). BCBS (2015a) also states that “the board should be comprised of a sufficient number of independent directors” (p.1). The general perception is that independent directors ensure that managers act in line with the interest of shareholders, leading to more voluntary disclosures (Elzahar & Hussainey, 2012; Fama & Jensen, 1983; Frankel et al., 2011).
Career-wise, independent directors do not depend upon the CEO, resulting in less power the CEO is able to exercise over those independent directors (Core et al., 1999). As independent directors have no relation with the management board, less alignment and collusion with managers takes place and independent directors are able to take an objective viewpoint (Carter et al., 2003; Eng & Mak, 2003; Patelli & Prencipe, 2007). These objective viewpoints check and limit the incentives of managers to behave opportunistic and contribute to effective monitoring and advising the management board (Fama & Jensen, 1983; Frankel et al., 2011). Moreover, independent directors are prone to a higher reputational risk and, therefore, demand more transparency and accountability of managers to maintain and improve their reputation in the marketplace (Lopes & Rodrigues, 2007; Patelli & Prencipe, 2007). Hence, independent directors are more likely to comply with (voluntary) disclosure guidelines and provide more and higher quality risk information as a mean of showing and protecting their reputation of professional integrity (Agyei-Mensah, 2017; Allini et al., 2016). In addition, previous research shows a positive effect between the level of independent directors and quality of operational risk disclosures (Barakat & Hussainey, 2013).
The empirical findings are predominantly related to quantity of risk disclosures and independence of the board, generally resulting in a positive relation (e.g. Abraham & Cox, 2007; Agyei-Mensah, 2017; Elshandidy et al., 2015; Oliveira et al., 2011a). Based on these prior studies and the discussion above, I expect a positive relation between supervisory board independence and quality of risk disclosures. Therefore, the hypothesis is stated as follows:
Hypothesis 3 (H3): The independence of the supervisory board positively influences the quality of risk disclosures.
3.7.4. Supervisory board and audit committee activity
Attending board meetings is one of the responsibilities of a supervisory board director (Allini et al., 2016). The frequency of those meetings, used as proxy for activity, is important for the effectivity of the supervisory board and its committees (Karamanou & Vafeas, 2005; Vafeas, 1999). For more focus in a specific area and increase of the efficiency, the supervisory board is able to establish specialised committees, where the audit committee is one of them (BCBS, 2015a). An audit committee is “required for systemically important banks and is strongly recommended for other banks based on an organisation’s size, risk profile or complexity” (BCBS, 2015, p.16) and assist the supervisory board in monitoring the financial reporting
Page | 19 policy (Pincus et al., 1989). Therefore, the audit committee is to a greater extent able to
influence the quality of risk reporting.
Regular meetings enhance the monitoring and advising function for financial reporting through discussing and exchanging of opinions, improving the information quality of disclosures (Barako et al., 2006; De Andres & Vallelado, 2008). To assess the decisions made and behaviours conducted by the management board and align the interests of the managers with the shareholders, the supervisory board and audit committee particularly ensure more voluntary disclosure of information (Allegrini & Greco, 2013; Laksmana, 2008; Li et al., 2012). This is also shown by Barakat and Hussainey (2013), who found a positive influence of audit
committee activity on the quality of operational risk disclosures.
Previous research shows that more frequent supervisory board meetings increase the accuracy of forecasts and therefore the quality of information (Karamanou & Vafeas, 2005). Vafeas (1999) found that the frequency of supervisory board meetings has a significant relation with effective disclosure of risk information and De Andes and Vallelado (2008) found a positive, although not significant, influence of the number of supervisory board meetings on the
performance of banks. Prior research is also supportive of a positive effect between frequency of audit committee meetings and level of disclosures (Allegrini & Greco, 2013; Karamanou & Vafeas, 2005; Li et al., 2012), however research particularly for risk disclosures is limited. Based on these prior studies and the discussion above, I expect a positive relation between the frequency of supervisory board and audit committee meetings and quality of risk disclosures3. Therefore, the hypotheses are stated as follows:
Hypothesis 4a (H4a): The frequency of supervisory board meetings positively influences the quality of risk disclosures.
Hypothesis 4b (H4b): The frequency of audit committee meetings positively influences the quality of risk disclosures.
3
The frequency of management board meetings is not hypothesised as the management board meet on a daily basis, therefore there are no differences in frequency between the banks.
Page | 20 3.8. Conceptual Model
The above mentioned hypotheses are graphically displayed in the conceptual model below.
Page | 21 4. RESEARCH METHODOLOGY
This section will discuss which method is used to investigate the influence of board
composition on the risk disclosure quality of European banks. First, the data collection and sample used will be explained. Secondly, the measurement of the dependent variable, i.e. the quality of risk disclosures, through a disclosure index will be explained. Thereafter, the independent and control variables will be described. Finally, the statistical model to test the hypotheses will be constructed.
4.1. Data and sample collection
4.1.1. Original data and sample
In order to examine the quality of risk disclosures, an analysis of the 2017 annual reports of European banks is conducted. For banks whose financial year does not coincide with the calendar year, the annual report closest to 31 December 2017 is analysed. Since the annual reports of 2017 are the most recent and all available when executing this research, this provides the most recent insights into the current quality of risk disclosures. As the Basel III Pillar 3 report is published as a standalone document and not included in the annual report, it is not in scope of this research. Therefore, the risk information in the annual report is used to measure the quality of risk disclosures. A sample of 75 European banks, with a basis in the population of the European Banking Authority (EBA) Transparency Exercise, is used. The sample consist of 48 of the largest banks, measured by number of total assets, under the supervision of the European Central Bank (ECB) and is complemented by 27 banks under local supervision. Only European banks are chosen, because they are all subject to the same Basel III regulation. Also for collecting the independent variables the annual report is used or, if the information is not available in the annual report, the bank’s website is used. Table 1 shows the number of banks per country, whereas appendix 1 gives a detailed overview of all European banks per country included in this research.
Page | 22
4.1.2. Additional data and sample
Due to a previous research group, data for 2015 and 2016 is already available. This data is retrieved by other researchers, however the same scoring principles for the disclosure index are used, as well as the same reviewer to ensure comparability and quality of the data collection (Baigent et al., 2008). Taken together with data for 2017, this sample consist of 25 banks with one year of observations, 11 banks with two years and 44 banks with three years of
observations available, resulting in a total of 180 bank-year observations and 80 unique European banks. The total sample consist of 55 banks for 2015, 50 banks for 2016 and 75 banks for 2017. A review is conducted on the disclosure index scores of 2015 and 2016 to further ensure the same quality of data. The disclosure index scores of 16 banks out of the 80 banks are reviewed and if necessary adjusted, as either these banks were outliers with a large difference in risk disclosure quality between the years or decreased in score on disclosure quality compared to the previous year. Specifically, a changed score between the years on an item is reviewed, and if necessary adjusted, as banks often use the same annual report structure and text several years. There is no review conducted on the independent variables as these observations are more straightforward and do not need an interpretation of the researcher. The sample of 2017 is used to test all hypotheses, whereas the sample of 2015-2017 serves as an additional analysis. There is chosen for this option as there is missing data of two
independent variables in the sample for 2015 and 2016, which hinders the performance of a panel regression for those two variables and the complete model. As the results later will show (§5.3), supervisory board meetings is significant, where only data for 2017 is available for.
Country Number of banks Country Number of banks Austria 2 Italy 6 Belgium 4 Luxembourg 1 Bulgaria 1 Malta 1 Croatia 2 Netherlands 6
Czech Republic 2 Norway 2
Denmark 3 Poland 1 Estonia 2 Portugal 2 Finland 2 Romania 1 France 7 Slovakia 1 Germany 8 Spain 5 Greece 1 Sweden 5
Hungary 1 United Kingdom 6
Ireland 3
Total 75
TABLE 1
Page | 23 Therefore, the 2017 sample is chosen to test hypotheses to be able to test the influence of
supervisory board meetings on the complete model and the other variables, which would be impossible for the sample of 2015-2017 due to missing data to run a panel regression (which was favourable). Subsequently, the sample of 2015-2017 is used as an additional analysis to still test the other variables over multiple years to examine if the (significant) results remain unchanged and get a richer view on the characteristics and improvement in risk disclosure quality.
4.2. Dependent variable: quality of risk disclosures
To measure the quality of risk disclosures (QRD), a disclosure index is used. Risk disclosure quality depends on the extent to which stakeholders are able to form a proper opinion about the financial results and financial position of the bank (Hoogendoorn & Mertens, 2001; Knoops, 2001). If the stakeholders are better able to form a good opinion of the bank, the higher the quality of risk disclosures is (Hoogendoorn & Mertens, 2001; Knoops, 2001). There are two approaches to measure the risk disclosure quality: the normative and empirical approach (Knoops, 2001). According to the normative approach, assumptions of accounting principles, information needs of users and law and regulation determine how the reports of banks should look like (Knoops, 2001). The empirical approach analyses the response of the capital market on the release of reporting information or the association between reporting and capital market variables (Knoops, 2001). This research used the normative approach by means of using a disclosure index. A disclosure index is often used in studies and is seen as a validated instrument (Barakat & Hussainey, 2013; Beattie et al., 2004; Marston & Shrives, 1991). The disclosure index consists of a list of items that are ex ante determined and of relevance to stakeholders (Srinivasan, 2006). The disclosure items are based on the guidance from Basel III, IFRS, EDTF4 and ESMA5 guidance reporting. In total 30 items are included in the disclosure index, split into four categories: general (7), credit risk (8), solvency/financial strength (10) and liquidity risk (5). The items in the index are measured, in line with Miihkinen (2012), ordinal on a three point scale, where 0 stands for non-disclosure, 1 for limited and 2 for comprehensive disclosure. At the end, the scores of all items are added and divided by 60, i.e. the maximum score of all items. As unweighted items reduce subjectivity and are seen as common in studies using annual report data (Ahmed & Courtis, 1999; Meek et al., 1995), all items are equally important and no weighting is applied. The resulting number indicates the quality of risk disclosures of European banks, where a higher number indicates a higher quality of risk disclosures. The disclosure index is presented in appendix 2.
4 The Enhanced Disclosure Task Force (EDTF) is established by the FSB in 2012 and has as goal to develop
principles for and recommend improvements to risk disclosures and identify best practices (EDTF, 2012).
5
The European Securities and Markets Authority (ESMA) is an independent authority of the EU with the main goal of enhancing investor protection and promoting stability and orderly in the financial market (ESMA, 2017).
Page | 24
Reliability and validity
Using a disclosure index is seen as a semi-objective methodology, therefore this research is somewhat subjective (Beattie et al., 2004). For constructing a disclosure index, reliability and validity are important to take into account (Marston & Shrives, 1991). The QRD scores are reliable if the same score is achieved by other researchers (Beattie et al., 2004; Marston & Shrives, 1991). Furthermore, it is important for reliability to use clear instructions to score the index items (Marston & Shrives, 1991). A pilot test with different examples of risk disclosures is conducted before data collection to evaluate if all instructions were clear. If necessary, the instructions were adjusted. After the data collection process, the different coders reviewed each other’s data collection to further increase the reliability of the QRD scores. A valid disclosure index measures and represents what it intends to measure and represents (Beattie et al., 2004; Marston & Shrives, 1991), namely the quality of risk disclosures. The disclosure index is carefully constructed on bases of relevant laws and regulations, to guarantee the validity. A self-constructed disclosure index is a valid methodology, as there are no other reputable methodologies or universal accepted indexes available (Marston & Shrives, 1991).
4.3. Independent variables
In this research two characteristics of both supervisory and management board are examined, namely diversity and size. Supervisory board diversity (SBDIV) is measured as the percentage of female directors in the supervisory board (Allini et al., 2016; Farag & Mallin, 2017; Ntim et al., 2013). In line with supervisory board diversity, management board diversity (MBDIV) is measured as the percentage of female directors in the management board. The size of the board is measured as the total number of directors in the supervisory board (SBSIZE) and
management board (MBSIZE) (Allegrini & Greco, 2013; Allini et al, 2016). For the supervisory board two other characteristics are also examined: independence and activity. Supervisory board independence (INDEP) is, in the same way as diversity, measured as the percentage of independent directors in the supervisory board (Barakat & Hussainey, 2013; Elshandidy et al., 2015; Ntim et al., 2013). The activity of the supervisory board (SBM) is measured as the number of supervisory board meetings in the reporting year (Allini et al., 2016). In addition, there is also examined if the activity of the audit committee influences the risk disclosure quality. Audit committee activity (ACM) is measured as the number of audit committee meetings in the reporting year (Barakat & Hussainey, 2013).
4.4. Control variables
Previous studies often researched and concluded that bank size (SIZE), leverage (LEV), profitability (ROE) and legal system (LAW) influence risk disclosures, mainly the risk disclosure quantity (Khlif & Hussainey, 2016). Therefore these four variables are taken into account as control variables. Bank size, leverage and profitability are collected from the annual reports, whereas the legal system is derived from the classification of La Porta et al. (1997). All
Page | 25 variables are calculated in euros, if a bank used another currency than it is converted in euro’s using the exchange rate of 31 December 2017.
4.4.1. Bank size
Prior research shows that bank size has a positive influence on quantity and quality of risk disclosures (Allini et al., 2016; Barakat & Hussainey, 2013; Ntim et al., 2013). The cost of reporting risk information to stakeholders is lower for larger banks compared to smaller banks, due to economies of scale and more available resources (Deumes & Knechel, 2008; Elshandidy et al., 2015). Furthermore, larger banks are more complex and have a variety of operations, which leads to increased information asymmetry (Deumes & Knechel, 2008) and provides an incentive to disclose risk information (Watts & Zimmerman, 1983). Due to the increased complexity, it is presupposed that larger banks are also more likely to have more board directors. Following previous research, bank size is measured as the natural logarithm of the total assets at the end of the year (Barakat & Hussainey, 2013; Farag & Mallin, 2017).
4.4.2. Leverage
Risk disclosures could also be influenced by the bank’s leverage (Khlif & Hussainey, 2016). Highly leveraged banks disclose more risk information to show credit holders their ability to meet obligations and mitigate creditor’s concerns (Elzahar & Hussainey, 2012; Rajab & Handley-Schachler, 2009). However, highly leveraged banks could also provide less risk information to conceal their weaknesses related to risks (Miihkinen, 2012). Consequently, empirical evidence is inconclusive about the influence of bank’s leverage on risk disclosures. Prior findings show a positive influence between bank’s leverage and risk disclosures
(Elshandidy et al., 2015; Khlif & Hussainey, 2016), whereas other researchers find a negative or insignificant influence between those variables (Elzahar & Hussainey, 2012; Miihkinen, 2012). Although, in the banking industry the Basel III leverage ratio is often used, which takes into account the own capital of the bank (BCBS, 2017a), this study used another measure. Since disclosure of the Basel III leverage ratio was not required until 1 January 2018, not all banks did disclose this ratio leading to an otherwise reduced sample. Therefore it is chosen, in accordance with prior research, that the bank’s leverage ratio is measured as total debt divided by total assets at the end of the year (Allini et al., 2016; Ntim et al., 2013).
4.4.3. Profitability
Findings from prior research of the relation between profitability and risk disclosures are inconclusive. While Ntim et al. (2013) and Khlif and Hussainey (2016) show a positive relation, Miihkinen (2012), Oliveira et al. (2011a) and Elshandidy et al. (2015) find a negative or insignificant relation. The incentive to disclose risk information increases for high-profitable banks to signal their ability in successfully managing risks and showing their performance quality (Elshandidy et al., 2015). On the other hand, Skinner (1994) argues that poor
Page | 26 outlook and avoid litigation risks. In accordance with previous research, profitability is defined as the return on equity (ROE), calculated as profit divided by total equity at the end of the year6 (Elzahar & Hussainey, 2012; Eng & Mak, 2003). As the ROE is, contrary to return on assets, not asset dependent, banks with different asset structures could be compared with each other. Moreover, it also takes into account the funding structure of the bank, as regulatory
requirements require a larger amount of capital and shareholders a dividend pay-out, which both influences total equity.
4.4.4. Legal system
Banks from different EU countries, and therefore different legal environments, are used in this research. Consequently, there must be controlled for differences in the two legal systems that are applied in the EU: common law and civil law system. A common law system is
characterized by professionalism and transparency, while in a civil law system secrecy and statutory control are the properties (Stulz & Williamson, 2003). These different attributes of the accounting system might influence the quality of risk disclosures, as civil law countries often disclose less information (Dobler et al., 2011; Khlif & Hussainey, 2016). Barakat and Hussainey (2013) show a positive influence of legal system on quality of operational risk disclosures and Khlif and Hussainey (2016) found a moderating effect. A dummy variable is created that takes a value of 0 for banks headquartered in common law countries and 1 for banks headquartered in civil law countries (Farag & Mallin, 2017).
All variables, their proxies and operationalisation are summarized in table 2.
4.5. Statistical Model
The different hypotheses are tested by using the ordinary least squares method. The dependent variable is in each model the quality of risk disclosures. The basic model only contains all control variables. Thereafter, each hypothesis is separately tested in one model (i.e. model 2 till 8) by adding the corresponding independent variable to the basic model. Finally, model 9 contains all variables and is tested for the number of available observations. The formula for model 9 is stated as follows:
Quality of risk disclosures = β0 + β1 * SBDIV + β2 * MBDIV + β3 * SBSIZE + β4 * MBSIZE + β5 * INDEP + β6 * SBM + β7 * ACM + β8 * SIZE + β9 * LEV + β10 * ROE + β11 * LAW +εi The variables are defined in table 2, βi represents the coefficients and εi is the error term.
6 Negative profitability values are not adjusted, as a test shows that the results are the same whether the values are
Page | 27
Variable Proxy Operationalisation
Dependent variable
Quality of risk disclosures QRD Total score of the items from the disclosure index dividend by
the maximum score of all items (60), expressed in a percentage
Independent variables
SBDIV Percentage of female directors in the supervisory board
MBDIV Percentage of female directors in the management board
SBSIZE Total number of directors in the supervisory board
MBSIZE Total number of directors in the management board
Board independence INDEP Percentage of independent directors in the supervisory board
Supervisory board activity SBM Number of supervisory board meetings in the reporting year
Audit committee activity ACM Number of audit committee meetings in the reporting year
Control variables
Size SIZE Natural logarithm of the total assets at the end of the year
Leverage LEV Total debt divided by total assets at the end of the year
Profitability ROE Profit divided by total equity at the end of the year
Legal system LAW Common law countries = 0, civil law countries = 1
Summary of variables and their operationalisation
TABLE 2
Board diversity Board size
Page | 28 5. RESULTS
This section discusses the results of this research. The descriptive statistics of the variables will be explained first, followed by the correlation analysis. Thereafter, the results of the tested hypotheses and additional analysis will be discussed.
5.1. Descriptive statistics
Table 3 gives an overview of the descriptive statistics of all researched variables. For 5 variables the final sample is less than 75 banks, as several banks did not disclose information about these variables. The number of observations for each variable is presented in table 3. All outliers are examined first and to further minimize the impact of outliers, all continuous variables are winsorized at three times the standard deviation.
The average score of the disclosure index is 26,41, referring to that on average banks scored 26 out of the 60 points available. The maximum score is 42 for Standard Chartered Plc, whereas the minimum score 11 is for Volkswagen Financial Services AG. Appendix 1 shows the disclosure index scores per individual bank.
On average 28,91% of the supervisory board directors is woman, the average supervisory board size is 11 directors and 76,23% of the supervisory board directors are independent. The percentage of women in the management board is on average 16,33% and the average size of the management board is 7 directors. Furthermore, the supervisory board met on average 12 times during the year and the audit committee 8 times.
The average leverage and profitability (ROE) are 91,83% and 7,49% respectively. Most of the countries in the sample are from the civil law system, except the UK and Ireland which are common law countries. Therefore, the average law system (0,88) is scaled towards the civil law system.
Page | 29 In table 4 are the descriptive statistics of the disclosure index broken down per category. The highest average score, as percentage of the maximum points available, is for the credit risk category. Solvability/financial strength scored the lowest average percentage, which could be explained by the fact that these risk disclosures should be more extensive to score points relative to, e.g., liquidity risk, where for example providing a ratio is sufficient to score points. Banks are, namely, expected to be further in providing risk disclosures in the area of
solvability/financial strength than, e.g., liquidity risk. Both the Landesbank Hessen-Thüringen Girozentrale and the ING Group N.V. scored the highest score for the general risk category. The highest score for credit risk and solvability/financial strength is achieved by, respectively, The Royal Bank of Scotland and Groupe Crédit Agricole. Belfius Banque SA and Allied Irish Banks Plc scored both the highest on liquidity risk.
Category Maximum score
possible Average Average as % Minimum Maximum
General 14 6,64 47,43% 2 12 Credit risk 16 7,96 49,75% 1 15 Solvability/financial strength 20 7,45 37,27% 1 16 Liquidity risk 10 4,36 43,60% 1 9 Total 60 26,41 44,02% 11 42 TABLE 4
