Quality Risk Disclosures by European banks: the role of nationality in the Quality of Risk Disclosures

(1)

Quality Risk Disclosures by European banks: the role of

nationality in the Quality of Risk Disclosures

Master thesis Controlling

University of Groningen, Faculty of Economics and Business January 22, 2018 Bob Peeters Studentnumber: 3275221 Turfsingel 92c 9711 VX Groningen Tel.: +31 (0)6 83043679 E-mail: b.b.j.student@rug.nl Supervisor J. Huttenhuis

(2)

2

Abstract

This paper develops a model of national factors that influence risk disclosures of European banks. The quality of risk disclosure is examined using four nationality angles that influence the board of directors, whom are in charge of the disclosure policy. This model is tested using multiple regressions on a sample of 50 European banks. This paper contributes to disclosure literature by including variables that have not previously been tested in the European banking industry. This research uses a disclosure index based on IFRS 7, EDTF and Basel III. The results of this research imply that nationality is not a determinant of the quality of risk disclosure by European banks.

(3)

3 1.

2. Introduction

Jensen and Meckling (1976, p. 311) define a firm as: “legal fiction which serves as a nexus for contracting relationships and which is also characterized by the existence of divisible residual claims on the assets and cash flows of the organization which can generally be sold without permission of the other contracting individuals”. The firm is essentially a team whose members act solely out of self-interest. In order for the team to survive, however, they realize that they need to combine their efforts (Fama, 1980).

Each firm has stakeholders, and thus a need for information disclosure. As stakeholders want to be able to properly assess the firms’ situation. To do this stakeholders require both financial and non-financial information.

Risk reporting is defined by Hassan (2009) as: “The financial statements inclusion about managers’ estimates, judgments, reliance on market-based accounting policies such as the impairment, derivative hedging, financial instruments, and fair value as well as the disclosure of concentrated operations, non-financial information about corporations’ plans, recruiting, strategy and other operational, economic, political and financial risks.”

Research regarding risk disclosures has been a popular topic among researchers for a long time (Healy & Palepu, 2001). With a timespan of over 25 years, this topic has been extensively researched and will continue to be a topic among researchers due to economic environmental uncertainty and the vital role it plays in society. The banking industry plays a vital part in the current economic system. The payment system is centered on banks (Andres & Vallelado, 2008). The banking model involves being highly leveraged, due to the deposits of customers, which is in turn used to finance the bank’s own activities. This indicates the importance of banks to society, which also explains why it is subject to more intense

regulation than other industries. Banks are held responsible for maintaining a stable payment system and protecting the depositors’ funds (Andres & Vallelado, 2008).

Due to recent developments regarding increased supervision and additional regulations, this topic has become more interesting. The enforcement of Basel III by the European Union, and with Basel IV being recently published, has sparked interest in the quality of risk disclosures of European banks (Barakat and Hussainey, 2013; Bischof, 2009)

In recent years a lot of research has been done regarding corporate risk disclosures. This research direction has become increasingly more interesting due to globalization and recent economic events. To provide an overview of what corporate disclosure encompasses I will

(6)

6 quote Healy and Palepu (2001): “Firms provide disclosure through regulated financial

reports, including the financial statements, footnotes, management discussion and analysis, and other regulatory filings. In addition, some firms engage in voluntary communication, such as management forecasts, analysts’ presentations and conference calls, press releases, internet sites, and other corporate reports. Finally, there are disclosures about firms by information intermediaries, such as financial analysts, industry experts, and the financial press.” In this thesis I will mainly consider annual reports, as other disclosures are too

difficult to keep track off. The efficient capital market is functional because of these disclosures (Healy and Palepu, 2001). In such a market, firms’ value reflect all available information which allows investors to make the best decision (Malkiel & Fama, 1970).

Even after the failure of the banking system in 2008, trust has continued to drop1. With signs of recovery starting in to show in 20162. Trust in the banking system is essential to the industry as this, among others, reduces the run on withdrawal of funds. In the Netherlands the DSB bank went bankrupt after the chairman of the SOBI (Stichting Onderzoek Bedrijfs Informatie) called for a run3, indicating the importance of trust.

The subprime crisis in 2008 is not the first financial crisis to happen in world economy. The first crisis took place in 1720 which was caused by insider dealing of government debt. To this date crises have happened in the following years: 1720, 1792, 1825, 1837, 1857, 1873, 1907, 1929, 1973, 1987, 1997, 2001 and 2008. When examining these numbers a clear pattern arises. The rate of recurrence has increased significantly over time. This is caused by

globalization, which allows for so called “bubbles” to grow at a much faster rate. All over the world citizens and companies are able to take part in these economic commitments.

In regaining the trust of stakeholders communication plays a crucial role. Trust can be maintained or regained by justifying past actions or finding support for future actions (Palmieri, 2009). The impact of the subprime crisis affected many stakeholders. Bankruptcy of banks caused stakeholders to lose their funds, requirements for entrepreneurs and citizens

1_{https://www.theguardian.com/business/2012/aug/09/financial-crisis-anniversary-trust-in-banks}

2

https://www.nvb.nl/nieuws/1464/vertrouwensmonitor-banken-vertrouwen-in-sector-stabiel-maar-laag-in-eigen-bank-hoger-klantbelang-meer-gediend.html

(7)

7 in order to get loans were dramatically increased and banks themselves had difficulties

attracting capital. The complexity of the banking industry increases the difficulty of being transparent. As disclosures may not provide insights to stakeholders who are not active in the banking industry.

Opaqueness of banks played a significant role in the recent financial crisis (Jones, Lee and Yeager, 2013; Flannery, Kwan and Nimalendran, 2013). This opaqueness causes depositors to be unable to distinguish between healthy and sick banks, resulting in runs on withdrawal when bad news is published (Jones, Lee and Yeager, 2013). This unclearness of information can be explained by three reasons (Jones, Lee and Yeager, 2013). First, insufficient disclosure results in information asymmetry. Second, disclosed information can be questioned on

credibility resulting in conflicting interpretations. Lastly, due to the complexity of the banking industry investors can still be wrong about the implications of full and credible disclosed information. Investments in opaque assets were more profitable than investments in

transparent assets, this is reflected by the inflated mortgage market which led to the collapse of the financial market. (Flannery, Kwan and Nimalendran, 2013).

In this research I will focus on the impact of nationality on the quality of risk disclosures by European banks. This research is centered on the drivers of transparency of information, as the level previously was insufficient to properly assess the bank's financial situation; it played a major role in the Financial Crisis (Barth & Landsman, 2010). In line with agency theory, which addresses problems that arise due to differences between the goals or desires between the principal and agent, banks have to provide information regarding risk management in order to enable agents to thoroughly assess a firm's financial situation.

Studies show that the country in which a company is founded impacts the level of risk disclosure (Amran, Manat Rosli Bin, Che Haat Mohd Hassan, 2008; Linsley & Shrives, 2005). Based on these studies I expect that, within Europe, national differences affect the level of risk reporting.

The research aims to provide a valuable contribution towards the risk disclosure literature. Finding evidence regarding differences in the quality of risk disclosures between European banks based on their national characteristics, teaches us that this is another factor that influences the quality of risk disclosures.

(8)

8 The results of this research can potentially influence practitioners, as practitioners have

information available on what to expect, in terms of the quality of risk disclosures, when conducting business with banks in different parts of Europe. For example, the European Central Bank (ECB) can use this research to explain the attitudes of banks towards the level of their voluntary risk disclosures. Society can benefit from this research as people from all over Europe can access this information when conducting business with banks in other countries.

To my knowledge, a study that examines the effects of nationality within the European banking industry has not been conducted recently. This paper contributes to literature by testing whether nationality plays a role in the disclosure policy of European banks.

(9)

9

3. Theoretical framework

This chapter will provide an overview of the institutional as well as most recent literature that is applicable to corporate risk disclosure by European banks. I will start with some general theories, followed by literature that is more in-depth and related to the research topic.

3.1 Legitimacy theory

The legitimacy theory claims that each business has a social contract with society (Shocker & Sethi, 1973). Shocker and Sethi (1973) claim that a corporations growth and survival depends on its ability to conduct business in a manner that is socially desirable and shares the benefits of its business with those from whom it derives its power. As organizations derive their power from society, society expects that the benefits to society exceeds the costs for society (Mousa & Hassan, 2015). Legitimacy can be achieved by showing that the activities of the firm take into account the social values of society. As European banks are the subject of this research and the role banks play in society, it is can be claimed that society is directly influenced by the decisions and behavior of banks. While the banking system plays a key role within

society, it does not mean that individual banks are not held to their social contracts. Whenever the bank does not comply with the contract anymore, society can revoke this contract (Mousa & Hassan, 2015). In other words, a run on the specific bank can occur resulting in a

withdrawal of funds. Nearly all citizens are in some way related to a bank, causing banks to have a social contract with society. In recent years banks have started to experience

competitive threats outside of the banking industry. These threats vary from a reduction of the required services to eliminating a banks services by paying with, for example, cryptocurrency or google wallet. Reducing the market share the banking industry has. This urges the

importance of trust for the banking industry. Trust is established by guaranteeing protection of funds and services. A firm has five ways of obtaining legitimacy (Mousa & Hassan, 2015). First, it can choose to legitimate its activities. Second, the firm can alter its output, goals and methods of operating to comply with current perception of what legitimacy should be. Third, a firm can try to become related to symbols, values or other institutions who are known for their prominent position in relation to legitimacy. Fourth, the firm may attempt to alter society’s perspective through communication. Lastly, a firm can create the assumption that they are doing the “right” thing, or showing that they avert from “bad” behavior.

(10)

10

3.2 Stakeholder theory

The transparency of information is vital to stakeholders, this allows them to properly assess the firm. Donaldson and Preston (1995) define stakeholders as: “stakeholders are defined by

their legitimate interest in the corporation, rather than simply by the corporation’s interest in them.” Society can be considered a stakeholder as, due to the nature of the banking industry,

nearly all citizens are directly affected by material developments in the banking industry. This indicates that the stakeholder theory is applicable. According to stakeholder theory, banks incentives for disclosure lie in the enhancement of their ability to interact and communicate with important stakeholders (e.g. blockholders, supervisory and governments) (Barakat & Hussainey, 2013). The stakeholder model explains that a firm has the following stakeholders: governments, investors, political groups, suppliers, customers, trade associations, employees and communities. When a bank comes under resolution there are four possible tools that can be applied in order to safeguard stakeholders’ funds4. These are: the sale of business (SoB), bridge institution (BI), asset separation (AS) and a bail-in. In the case of a SoB the liabilities and markets assets are transferred to a purchaser at market price. The BI allows for temporary transfers of ownership. With a bail-in a bank is recapitalized, restoring financial soundness and long term viability.

3.3 Agency theory

An agency problem arises when there is a separation between ownership and control (Jensen and Meckling, 1976). The agency theory assumes that both the agent and the principal have their own agenda, causing a conflict when interests are not aligned. In a situation where it is difficult for the principal to verify what the agent is doing this problem is present (Eisenhardt, 1989). This phenomenon is also known as information asymmetry, which means that there is a disparity between the agent and principal in terms of access to the same information. The solution to this problem lies in contracts and regulations. Contracts enable the alignment of interests of both parties (Healy & Palepu, 2001). A corporation can also be seen as a nexus of contracts (Jensen and Meckling, 1976). Regulations force managers to disclose private information.

In agency theory two phenomena are present: adverse selection and moral hazard. Both are a form of information asymmetry. Adverse selection refers to the situation where one party has

(11)

11 information that the other party lacks. Adverse selections occurs before a transaction has been made. One example in the banking industry is that when firms’ with bad credit risks have an increased likelihood to acquire loans over firms that have good credit risks.

In a situation of moral hazard, one party is able to take additional risks that negatively affect the other party. In the banking industry a flawed incentive system was present, the system caused a moral hazard issue. The moral hazard issue was one of the causes of the financial crisis, as mortgage lenders had incentives to close as much mortgages as possible. These mortgages were then sold to investors, as they were perceived to be risk free, shifting risks to the investor. When the housing bubble collapsed, the home prices dropped dramatically in value. This led to mortgage defaults which was a significant factor in the subprime mortgage crisis.

3.4 Risk disclosure

Risk disclosure is dependent on the demand and supply of information. At the demand-side there are a variety of stakeholders: shareholders, investors, managers, employees, credit suppliers, customers, government, other industries and society as a whole. The supply-side is determined by legislation and cost-benefit considerations. Financial information plays a vital role in supplying information. These financial reports provide corporations with a way of sharing information with its stakeholders. Sharing this information reduces information asymmetry. Besides the fact that stakeholders want this information, there are reasons why a company chooses to disclose information. Risk disclosure can be used to: (1) monitor management, (2) reduce the market risk premium, (3) enhance legitimacy and (4) reduce litigation costs.

Risk disclosure can be considered a tool for monitoring management (Eng & Mak, 2003). Both on internal and external levels stakeholders want to assess the performance of the management. The board of directors plays a big role in monitoring management. Therefore, it is important that the composition of the board of the directors is of good quality. As research shows (Fama, 1980; Fama and Jensen, 1983), board characteristics can influence the quality of the monitoring. They find that outside non-executive directors positively influence the board’s ability to monitor management. An explanation can be found in the fact that independent outside board members have a higher demand for information, as they are not involved in daily management. With regards to risk, outside board members may also limit managerial opportunism. Outside board members are able to objectively judge information

(12)

12 because of their limited interest in the company. When foreigners are present in the board of directors, the board of directors independence increases. Foreigners view on matters can differ, this is due to experiences, thinking and heterogeneity (Hamzah & Zulkafli, 2014). Voluntary disclosure regards information that is not demanded to be shared by legislation. This information is relevant for stakeholders as information may alter their decisions with regards to the corporation. Depoers (2000) defines voluntary reporting as: “An item of

information is considered as discretionary whenever it goes beyond the compulsory

information for shareholders. Compulsory information has to be understood as all the items whose publication is duly required but also the items which firms must send to shareholders who ask for them. Whether its nature be qualitative, financial or anything else, voluntary disclosure cover all data which concern both the subsidiaries and the group itself.” In past

research, it has been shown that, in the banking sector, there are differences in the extent of disclosure among countries (Kahl & Belkaoui, 1981). As regulations regarding disclosure have changed, it raises questions whether these differences still exist. This is further nurtured by Einhorn (2005) who has found that, in countries with more mandatory disclosure, the level of voluntary disclosures increases.

Where large firms tend to be formal with set communication channels, small firms tend to be more informal. With set communication channels, not all information may be reported to the board of directors. In other words, there may be an information asymmetry between the board of directors and management. Information asymmetry comes at a cost. Stakeholders want to be compensated for the risk they are taking. That is to say, investors demand a risk premium for uncertainty. Disclosing information reduces the information asymmetry which leads to a reduction in uncertainty and therefore reduces the risk premium demanded by investors (Easley & O ‘Hara, 2004). Botosan and Plumlee (2002) find that the cost of capital decreases as the level of risk disclosure in the annual report increases.

Risk reporting enhances legitimacy (Oliveira, Rodrigues and Craig, 2011). It allows a firm to manage the stakeholders’ perception of the firm with respect to risk exposures. Managing stakeholders’ perception of the firm influences the firms’ image and informs stakeholders about the ability of the management to manage risks. (Iatridis, 2008) If done right, it reassures investors whom originally put trust in the management team.

Litigation costs are costs incurred by a firm to reduce the threat of shareholder litigation. Litigation costs originate from two sources. One source comes from the costs made to avoid

(13)

13 legal actions from shareholders against managers. The second comes from a reduction in disclosure with regard to future expectations; managers may choose to not share their

expectation as their predictions might deviate from the realization (Healy and Palepu, 2001). Publicly disclosing information can have negative side-effects, increasing the cost of equity capital (Botosan and Plumlee, 2002). Negative disclosures from business press lead to an increase in cost of capital (Kothari, Li and Short, 2009), this can cause corporations to

disclose information as it enables them to disclose them in a manner that is more favorable for the firm. Disclosed information is also accessible for competitors, this may lead to a

competitive disadvantage, leading to financial harm to the firm.

3.5 Legislation

When corporations keep growing they, at a certain point, become of importance to society. For these corporations legislations have been made, in order to protect society from

maladministration. Dutch law, for example, requires corporations who comply with certain standards to have their financial statements audited. Financial statements are, according to Art. 2: 361 BW, the statutory annual accounts which exists out of: the balance sheet, the income statement and notes.

Art. 2:398 BW explains that listed corporations have to be audited. According to art. 2:396 BW firms who comply with two of the following three requirements, for two consecutive financial years, have to be audited as well: First, the total value of assets is at least

€6.000.000. Second, the revenue during the financial year of the firm is at least €12.000.000. And lastly, the average amount of employees during the financial year is at least 50. Art. 2: 393 BW states that the financial statements should provide a fair representation of the

corporations’ financial position, financial performance and cash flow. Comparable legislation is enforced by the European Union. Regulation No 537/2014 specifies requirements for statutory audits of public interest entities, such as listed companies, banks and insurance undertaking5.

5

(14)

14

3.6 Regulations

3.6.1 International Financial Reporting Standards

As this research paper aims at the banking industry, financial instruments and their

regulations regarding disclosures play a major role in risk disclosure. International Financial Reporting Standards (IFRS) 7 is focused on the disclosure of financial instruments.

Regulation (EC) No 1606/2002 requires all listed companies to apply IFRS6, this makes company accounts understandable and suitable for comparison across countries. Which is also a reason for non-listed companies to apply IFRS. Information should be disclosed about the significance of financial instruments to an entity. Hereby specifying on the nature and scope of risks originating from these financial instruments, both in quantitative and qualitative terms7_{. This requires companies to disclose on the following categories}8_{: the statement of}

financial position, statement of comprehensive income and other disclosures. Of the financial statement position the following should be disclosed: financial assets measures at fair value through profit and loss, held-to-maturity investments, loan and receivables, available-for-sale assets, financial liabilities at fair value through profit and loss and financial liabilities

measured at amortized cost.

Besides IFRS 7, IFRS 9 will be an important reporting standard for the banking industry. IFRS 9 focuses on financial instruments9. This is a replacement of IAS 39 Financial

Instruments: Recognition and Measurement. These regulations will be effective as of the 1st of January 2018. But as IAS 39 Financial Instruments: Recognition and Measurement is still applicable to the banks in the sample I will elaborate on IAS 39 instead of IFRS 9. IAS 39 provides a framework for the requirements regarding the recognition and measurement of financial assets, financial liabilities, and some contracts to buy or sell non-financial items10. The financial assets are classified into four main categories: financial asset at fair value through profit or loss, held-to-maturity financial investments, loans and receivables and available-for-sale financial assets. IAS 39 is applicable to this thesis as it is (1) specifically enforced by Regulation No 537/2014 and (2) it focuses on financial instruments, which is a 6 https://ec.europa.eu/info/business-economy-euro/company-reporting-and-auditing/company-reporting/financial-reporting_en#ifrs-endorsement-process 48_{https://www.iasplus.com/en/standards/ifrs/ifrs7} 9_{https://www.iasplus.com/en/standards/ifrs/ifrs9} 10_{https://www.iasplus.com/en/standards/ias/ias39}

(15)

15 banks’ core business. Bischof (2009) has researched the influence of IFRS 7 on bank

disclosure in Europe. He finds that the quality in both financial statements and risk disclosures has increased. Thereby noticing a shift from market risk to credit risk in terms in disclosure. The finding regarding the quality of disclosure is seconded by Iatridis (2010).

3.6.2 Basel

The Basel Committee was founded in 1974 by the central bank Governors of the Group of Ten countries. This was a result of serious turmoil in the international currency and banking markets. The Committee’s objective was to improve financial stability by enhancing the quality of banking supervision. The first paper was issued in 1975 setting out principles for sharing supervisory responsibility.

Before Lehman Brothers collapsed it was clear that the Basel II framework was insufficient. Banks who followed the minimum requirements of this framework had too much leverage and insufficient liquidity buffers when the financial crisis hit. This effect was strengthened by poor governance, risk management and incentive structures11. This led to mispricing of credit and liquidity risk, and excess credit growth. As a result new capital and liquidity standards were endorsed, now known as Basel III. Basel III strengthened the elements of Basel II and added several guidelines. An additional layer of common equity, a counter cyclical capital buffer, a leverage ratio and liquidity requirements were introduced. These new requirements are phased in progressively, each with their own respective timeline.

In January 2012 the Regulatory Consistency Assessment Programme (RCAP) was proposed to more extensive monitor the implementation of Basel III. This monitors the timeliness, consistency and completeness of implementation. National authorities are allowed to set higher requirements than Basel III proposes. Periodically the Basel Committee on Banking Supervision publishes a progress report regarding the adoption of the Basel regulatory framework. In this progress report the risk based capital, liquidity standards, SIB, leverage ratio and disclosure the implementation status is presented for all 27 jurisdictions members. Belgium, France, Germany, Italy, Luxembourg, Netherlands, Spain, Sweden, and the United Kingdom all follow the EU process (Basel Committee on Banking Supervision, 2017). The

(16)

16 current state is that all the above mentioned countries comply with the adaptation status

determined by the Basel Committee.

The Basel Committee does not have any authority to enforce its propositions and thus no legal force12. This however, is enforced by the European Parliament through regulation No

575/2013. This regulation applies to credit institutions and investment firms in the European Union.

Basel III consists out of three pillars. Pillar 1 is concerned with minimum capital

requirements. These requirements are set in place to ensure a safeguard against the three major risks categories a bank faces: credit risk, operational risk and market risk. The second pillar is about supervisory review, a regulatory response. In addition it provides a framework for residual risk. Pillar 3 lies within the scope of this paper and thus will be elaborated on more thoroughly. Pillar 3 complements Pillar 1 and Pillar 2. Pillar 3 introduced new disclosure requirements, these requirements primarily concern the capital structure, capital adequacy, risk management and risk measurement of banks. In short, Pillar 3 requirements look at disclosing the following aspects from a qualitative and quantitative point13: Scope of application of the capital adequacy framework; capital structure and capital adequacy; credit risk; securitization; market risk; equities; interest rate risk in the banking book; and

operational risk. Which can directly be related to the research topic: quality of risk disclosures.

The Financial Stability Board, together with the Basel Committee, annually identify global systemically important banks (G-SIBs). Due to their importance they are subject to higher capital buffer, total loss-absorbing capacity (TLAC) (as of 1 January 2019) and resolvability requirements and higher supervisory expectations. In the Pillar 3 standards of Basel the specific TLAC requirements are presented in detail.

The information required for the disclosures of Pillar 3 is mainly collected from the information that has already been gathered for Pillar 1 disclosures. These changes could impact the financial system in the following manners14: reducing risk of systemic banking crisis, reducing lending capacity, reducing investor appetite for bank debt and equity and

12_{https://www.bis.org/speeches/sp161012.htm}

1415_{https://www.pwc.com/gx/en/banking-capital-markets/pdf/basel.pdf}

14_{https://www2.deloitte.com/content/dam/Deloitte/ie/Documents/FinancialServices/investmentmanagement/2014}

(17)

17 causing inconsistent implementation of the Basel III proposals leading to international

arbitrage.

The quality of financial information is as dependent on the enforcement as it is on the standards themselves (Ball, Kothari and Robin, 2000). If rules are not enforced they are deemed useless. This sets the question: can nationality explain the degree in which firms comply with these rules?

Basel III and IFRS both aim at measuring capital consumption. Where Basel III focuses on the resilience of banks, IFRS aims at transparency, accountability and efficiency. As both are applicable to banks, they also influence each other. For example valuation, which is

determined by IFRS, is critical for the measurement of capital which is essential for Basel III.

3.6.3 Enhanced disclosure task force

On the 12th of May 2012 the enhanced disclosure task force (EDTF) was founded. The primary objective of the EDTF was to develop recommendations for enhancing risk

disclosure by major banks and to identify leading practices on risk disclosures15. The findings of the EDTF were used to create the disclosure index for this research. On the 13th of May 2016 the Task Force was formally disbanded, due to the improvements made in bank disclosures.

3.6.4 The European Securities and Markets Authority

In 2011 the European Securities and Market Authority (ESMA) was founded. This

institutions’ members consist out of national authorities from each EU country. ESMA aims to improve investor protection and promotes stable, orderly financial markets. It aims to do so through four activities16_{: Assessing risks to investors, markets and financial stability,}

completing a single rulebook for EU financial markets, promoting standardization of supervision practices and directly supervising specific financial bodies.

15_{http://www.fsb.org/2012/05/formation-of-the-enhanced-disclosure-task-force/} 16_{https://europa.eu/european-union/about-eu/agencies/esma_en}

(18)

18

3.7 National culture

In this paragraph I will discuss several views on how cultural aspects can influence the level of disclosure. At first I will discuss Hofstede’s model of Cultural Dimensions. This model attempts to explain the culture of a country by using four dimensions. Namely: power distance (PDI), uncertainty avoidance (UAI), individualism and masculinity (Hofstede, 1983). In 1991 a fifth dimension was added: the Long-Term Orientation (LTO). In 2010 latest dimension was added: Indulgence.17_{These dimensions are elaborated on in paragraph 3.7.1.}

In order to describe the dimensions the definitions as provided by Hofstede18_{are used.}

Based on the cultural dimensions of Hofstede (1983), Gray (1988) developed four accounting values: professionalism, uniformity, conservatism and secrecy. Research has contradicting views on how the dimensions of Hofstede affect disclosures. Salter and Niswander (1995) argue that individualism and uncertainty are positively related to disclosures, whereas Zarzeski (1996) find that these dimensions are negatively related with disclosure.

3.7.1 Hofstede

In this research I try to relate the cultural dimensions of Hofstede to the quality of risk

disclosure by European banks. Research (Hope, 2003) has been done with respect to elements that explain varieties in accounting practices. From many elements studied, culture and legal system appear to play a vital part in explaining this. Gray (1988) claims that cultural values can be related to accounting information. This is seconded by Dobler (2008) who claims that culture plays a role in variations between countries with respect to risk disclosure. Although Hofstede’s model has been criticized, it is still frequently used in research regarding

disclosure (Dobler et al. 2016). Gray (1988) expects that power distance and uncertainty avoidance are negatively related to the level of risk disclosure, these are the opposites of individualism and masculinity. The underlying thoughts of these expectations are as follows. In countries with high PDI firms are likely to refrain from disclosing in order to protect their superiority in power. UAI should decrease the level of risk reporting as firms will want to avoid conflict and competition. At the time of this paper long-term orientation was not yet included in Hofstede’s model, researchers expect LTO to be positively related to the level of

17_{https://geert-hofstede.com/national-culture.html} 18_{https://www.hofstede-insights.com/models/national-culture/}

(19)

19 risk disclosure. Evidence regarding these beliefs is to this date inconclusive (Archambault & Archambault, 2003; Hope, 2003; Zarzeski, 1996; Salter & Niswander, 1995). PDI refers to the degree in which members of society accept authority and the unequal distribution of power. UAI is the extent to which members of society feel uncomfortable with uncertainty and ambiguity. The LTO score shows how society thinks about tradition and forward-looking thinking.

Elshandidy at al. (2015) find that firms from common law countries disclose significantly more risk information than code law countries. They find that legal as well as cultural factors are not as significant in explaining voluntary risk disclosure as they are in mandatory risk disclosure between firms across countries. This study used automated content analysis. In which the amount of certain (risk) words are related to the quality of risk disclosure. This is a quantitative approach, whereas my approach will feature a manual data collecting method.

3.7.2 Accounting values

Gray (1988) derived four accounting values: professionalism, uniformity, conservatism and secrecy. In his study he relates these values to the dimensions of Hofstede. In this paragraph I will elaborate on these accounting values.

From these accounting values Elshandidy (2015) argues that LTO and UAI are significantly related to conservatism. Whilst finding that secrecy is significantly related to UAI. Therefore, I choose to use these Hofstede dimensions in testing the hypothesis that certain national cultural aspects are related to the quality of risk disclosures.

Conservatism is related to optimism. Characterized by a careful approach to measurement in order to account for the uncertainty of future events. Optimism can be described as a more risk-taking, optimistic approach. This can lead to a demand for top-level systems in order to anticipate unforeseen risks and events.

Secrecy is related to transparency. Confidentiality and limited disclosure of information describes secrecy. Information is only shared with those who are closely related with the management and financing of the firm. This is in contrast to transparency, where the general approach is to be more open and publicly accountable. This is nurtured by the urge of restricting disclosure in order to avoid conflict, competition and preserving security.

(20)

20

3.7.3 Country of Origin

Many banks operate on an international levels and are heavily regulated. Despite these regulations differences can still be observed in the thirteenth progress report of the Basel Committee on Banking Supervision, where for example Germany has adopted several components of Basel III in its national law. In contrary to France, who only follows the EU process19. Höring and Gründl (2011) have examined risk disclosure in the insurance industry, which is also heavily regulated. In this study the level of risk disclosure is related to the home country of corporations.

Hope (2003) argues that there is a significant variation in financial reporting of firms across nations. Therefore, I hypothesize that the country of origin is related to the quality of risk disclosure.

3.7.4 Legal system

In Europe there are two main legal systems: Common law and Code law. While the legal system exerts coercive pressure, and thus influences the mandatory required reporting, it can also influence the level of voluntary reporting. If mandatory reporting standards are high the result may be that additional (voluntary) disclosure is relatively cheap. As information

systems are setup in a manner to comply with these high standards, whereas in countries with lower demands it may be more costly to disclose this additional information. This effect is strengthened by competitors who experience the same effects, resulting in increased peer pressure (Li & Yang, 2015).

19

Basel Committee on Banking Supervision. (2017). Thirteenth progress report on adoption of

(21)

21

4. Methodology

In this chapter I will describe the empirical part of the research. In paragraph 4.1 the sample will be described. Paragraph 4.2 will describe the process of data collection. In paragraph 4.3 the variables included in the research will be operationalized.

4.1 Description of the sample

The determination of the sample is subject to multiple reasons. First, I will discuss the years I take into consideration. Then I will elaborate on the sample that I will use.

4.1.1 Year of the sample

This research will mostly consider the year 2016. As part of the reason for this research lie within the increased demand for information, the most recent information will be the best indication for the current state of disclosures.

Due to the (phased) implementation of Basel III this year provides insights in the degree of disclosure. While differences in quality of risk disclosures may become smaller due to more regulations, this may not necessarily be the case. As firms have incentives to report more than demanded by regulations (i.e. voluntary disclosure).

4.1.2 Europe

The sample of this research consists out of European banks. More specifically, banks from: Austria, Belgium, Denmark, Finland, France, Germany, Hungary, Ireland, Italy, Netherlands, Norway, Poland, Spain, Sweden and the United Kingdom. The banks selected for this

research are connected to the European Banking Authority (EBA). The fifty banks of the sample participated in the EBA transparency exercise. These banks report to the EBA at the highest level of consolidation and therefore their data contributes to the EU averages. This provides us with a reliable and comparable sample. The determination of the sample size is related to the timeframe, and therefore a limited workload, of the program. Table 1 shows the amount of banks per country.

(22)

22 Table 1: Number of banks per country

4.2 Data collection

The annual reports are acquired by through the bank’s website. The data for the disclosure index are gathered via official (audited) annual reports, as the quality of the annual report can only be guaranteed by an auditor’s report. Besides the disclosure index additional data has been collected for our dependent and independent variables. The process of data collection has been conducted manually. Data required for the dependent variable, quality of risk disclosures, and independent variables have been collected through these annual reports. Missing data has been filled by using either the bank’s website or external databases (e.g. Bloomberg.com). In case data was no longer available, the most recent information was used. Due to no alternative methods, this procedure was the best option. For the independent variables stock exchange and national culture additional resources were used. Hofstede’s national scores were derived from Hofstede’s website. Data regarding the listing on stock exchanges was accumulated through either the banks website, annual report or the websites of stock exchanges. Due to a previous research group, data from 2015 had already been

collected. In order to obtain a comparable sample twelve additional annual reports from 2015 had to be reviewed. Whilst the data collection of 2015 has not been done by either van den Berg or me, the reviewer is the same. Ensuring the comparability and quality of data (Baigent et al., 2008). At the start of the data collection phase of this research a pilot phase was held. This ensured clarity with regards to the assessment of the annual reports. Extending the sample with an additional year provides more data and allows for comparability between

(23)

23 years. As Basel III is enforced as of 2014, an adjustment period from banks to these new requirements is present. Therefore, the quality of risk disclosure between the sample of 2015 and 2016 might show slight deviations.

4.3 Variables

Literature suggests that cultural differences affects beliefs, thoughts and behavior. This might indicate differences in the views on risk disclosure of banks. The dependent variable of this research is the quality of risk disclosures. The hypotheses cover four independent variables. These are: the composition of the board of directors with regards to nationality, the country of origination, the stock exchanges at which a bank is listed and Hofstede’s cultural dimensions. To examine the influence of these variables on the dependent variable, I propose a number of variables.

4.3.1 Dependent variable: Quality of risk disclosure

The quality of risk disclosures provides by banks will be examined using a disclosure checklist. A disclosure index assists in the assessment of risk disclosures. Creating a tailor-made checklist allows the user to conduct research in a precise manner. Using the disclosure index method, an index has been created ex-ante. The checklist consists out of thirty items, spread over the categories: general, credit risk, solvency and liquidity. The checklist is based on IFRS 7 and other guidance (Basel III, EDTF, and ESMA). Based on this checklist we analyze the quality of risk disclosure in a qualitative and quantitative manner. Each item will be rewarded either 0, 1 or 2 points. This is an ordinal method. This method has three possible scores limiting the possible amount of subjectivity, the same research done by another

practitioner should bear similar results. The checklist score per bank will be additive and unweighted. The items in this checklist are unweighted, as Meek et al. (1995) find that weighted and unweighted index lists are equal in terms of results. Meek et al. (1995) claim this for several reasons. First, assigning weights is subjective as different users will assign different weights. Second, firms disclosing “important” items better usually disclose “less important” items in a similar manner. Using an unweighted disclosure index therefore is common among researchers (Hassan, 2009; Meek et al., 1995). The QRD scores per bank are presented in appendix A.

(24)

24

4.3.2.1 Independent variable: Nationality board of directors

In the complex banking business one of the board of directors’ roles to monitor managers’ behavior and to provide them with strategic advice. The board members protect the interests of shareholders. The board also in charge of communication with regulators (Andres & Vaallelado, 2008). The underlying idea of Andres and Vaallelado (2008, p. 2571) is: “several

characteristics of the board of directors (size, composition or functioning) might reflects directors’ motivation and their ability to effectively monitor and advise managers.” This may

result in better governance of banks, and thus a higher quality of risk disclosure.

Previous studies have examined the influence of the presence of independent directors, board activity, board stock ownership and an extended board (Domínguez & Gámez, 2014). While the impact of nationality of board members on firm value has been researched (Carter, Simkins and Simpson, 2003), the impact on disclosure has, to my knowledge, not yet been investigated.

Oxelheim and Randøy (2003) find that foreign board membership is positively related to firm value. As in line with agency theory, Orens, Aerts and Lybaert (2009) find that disclosing information improves firm values. Disclosing information results in lower information asymmetry which causes investors to demand a smaller risk premium. Firms are more likely to voluntarily disclose information with more independent directors. (Cheng & Courtenay, 2006). As foreign directors are more likely to be independent, due to them not residing in the same country. One reason to assume this is the fact is that foreign board members could be used to a more demanding system, which should result in the board being more independent from management (Oxelheim & Randøy, 2003). This could lead to a higher level of

disclosure. Combined with evidence that independent directors are more willing to disclose information, and voluntary disclosure increases firm value. This leads me to believe that a board with foreign board members is more likely to voluntarily disclose information on a higher qualitative and quantitative level. I expect that this effect increases with the amount of foreign directors in the board, as this will give them more power. This variable is therefore measured relatively to the total amount of board members. The data was gathered either via the bank’s annual report or their website. This results in the following hypothesis:

H1: The presence of foreign board members in the board of directors is positively related to the quality of risk disclosure

(25)

25

4.3.2.2 Independent variable: Country of Origin

As discussed in paragraph 3.6.3 I expect that the country of origin is related to the quality of risk disclosure. Furthermore, in corporate social responsibility(CSR) literature it is found that disclosure of CSR information is significantly related to the country of origin (Wanderley et al, 2008). Although many banks in the sample operate on an international level, they still need to comply with regulatory pressures that stem from their country of origin (Kaymak & Bektas, 2017). This results in the following hypothesis:

H2: Ceteris paribus, the country of origin is related to the quality of risk disclosure 4.3.2.3 Independent variable: Stock Exchange

Firms do not always enter the stock exchange within their country of origin. One reason for firms to list their shares on a foreign stock exchange is varying financial disclosure

requirements (Saudagaran & Biddle, 1992). With the formation of the European Union, financial markets have become more uniformly regulated. However, this only covers the mandatory disclosure. For some corporations the financial disclosure requirements played a decisive role in their choice of exchange market. (Saudagaran & Biddle, 1992). As different stock market have different requirements in terms of mandatory disclosure. Voluntarily entering multiple stock markets requires a company to comply with multiple standards. This could be an indication of the banks willingness to voluntarily disclosure information.. This may have implications for risk disclosures as well. A reason for banks not to enter (multiple) stock markets is that investments in opaque assets are more profitable (Jones, Lee & Yeager, 2013), this might be an incentive for banks to keep their information nontransparent. As this would result in a higher firm value. This may be further fostered due to the complexity of the banking industry (Andres & Vallelado, 2008). This raises the expectation that differences between listed and non-listed banks with regards to the quality of risk disclosure may exist. This results in the following hypothesis:

H3: Ceteris paribus, the presence of a bank on the stock exchange is positively related to the quality of risk disclosure

4.3.2.4 Independent variable: National culture

As discussed in paragraph 3.6.1 and 3.6.2 I expect that the cultural aspects of Hofstede will influence the quality of risk disclosure. This results in the following hypotheses:

(26)

26

H4a: Ceteris paribus, banks from a country with high power distance are negatively related to the quality of risk disclosure

H4b: Ceteris paribus, banks from a country with high uncertainty avoidance are negatively related to the quality of risk disclosure

H4c: Ceteris paribus, banks from a country with high long term orientation are positively related to the quality of risk disclosure

4.3.3 Control variables

In the statistical model I will apply control variables. Control variables are implemented in order to help explain changes in the dependent variable quality of risk disclosure. If these variables are present in all statistical tests the effect of de independent variables should become apparent.

4.3.3.1 Size

One determinant of disclosure is firm size. Ahmed and Courtis (1999) find that firm size is positively related to disclosure levels of firms. Reasons lie within the demand for external capital by larger firms. Jensen and Meckling (1976) claim that the presence of external capital within a firm causes increased monitoring costs. One way of mitigating these costs is by disclosure of information in annual reports (Singhvi and Desai, 1971, p. 131). Large banks tend to have larger information systems, therefore it is easier for these banks to gather

information. This causes their accumulation costs to be lower. As larger banks’ securities are more liquid, the need for information is also larger. The size of a firm in this research is determined by the natural logarithm of the value of its total assets (Archambault & Archambault, 2003). In the sample multiple currencies are present. In Appendix B

descriptives are presented per currency, starting at a minimum presence in the sample of 3. Due to limited differences, I have chosen to not further control for the different currencies.

4.3.3.2 Profitability

Managers of profitable firms have incentives to disclose more information. As the agency theory describes the relationship between ownership and control, it can be assumed that a manager will want to show good news to its owners. Therefore, the claim that profitable firms disclose more information is logical. Research on this variable has had conflicting results. As profitability can incur political costs (Shehata, 2014). While Raffournier (1995) claims that

(27)

27 disclosing information when profitable allows managers to gain support with respect to their remuneration and position.

4.3.3.3 Leverage

As mentioned at the size control variable, agency theory predicts that firms with high leverage position need to disclose more information in order to suppress the cost of monitoring. In other words, agency costs rise when the amount of leverage in a firm is higher. Usually, leverage is the ratio of debt to equity. Leverage is measured by dividing the firms’ book value of debt to the book value of its total equity, as done by other researchers (Ben-Amar, Chang & McIlkenny, 2017). However, as this research is aimed at the banking industry, I have chosen to apply the Basel III specific leverage ratio. This ratio is defined as the capital measure divided by the exposure measure (Basel Committee on Banking Supervision, 2014).

4.3.3.4 Dividend

Dividend payments provide information to investors. This implicitly informs investors about future cash flows (Miller & Rock, 1985). The information that arises from dividend payments may cause a firm to disclose less information on other forms of disclosure, the dividend payments function as a substitute. In this research a binary system for dividends is used.

4.3.3.5 Legal system

In Europe two main legal systems exist. They are known as common law and code law. Li and Yang (2015) find that even though common-law countries voluntarily disclose more information, code-law is catching up since the adoption of IFRS in 2005. Code-law countries initially had low demand for disclosures from capital markets (Li and Yang, 2015). While differences are getting smaller, code-law countries still disclose less information than common law countries. This difference can be explained by the fact that common law is aimed at creating an environment where corporate disclosure is increased in order to satisfy the needs of individual corporations (Archambault & Archambault, 2003).

4.3.3.6 Foreign activity

In the search to explain whether nationality plays a role in a firm controlling for

internationally operating banks is important. Depoers (2000) argues that the foreign activity of a firm has significant influence on the level of voluntary disclosure. Although his research was aimed at the French listed market, and not the banking industry, I expect that the same trend is visible in the European banking industry. Internationally operating businesses tend to

(28)

28 have more sophisticated information systems due to their need for more information. For these corporations it is not as costly to voluntarily disclose more information, as the internal environment already requires this. This data was retrieved from the bank’s annual report.

4.3.3.7 G-SIB rating

As mentioned before, globally systematic important banks (GSIB) have additional requirements related to non-G-SIB banks. These banks are subject to higher supervisory expectations, which are aimed at risk management functions, risk data aggregation capabilities, risk governance and internal controls. This can affect the level of voluntary disclosure, as Einhorn (2005) finds that the degree of voluntary disclosure and mandatory disclosure are related to each other. In order to control for this effect, the G-SIB rating of banks is included as a control variable. The G-SIB rating is divided in five buckets, where each buckets means an additional capital buffer increase of 0.5% starting at 1.0%. This information is retrieved from the 2016 list of global systemically important banks. Based on this list the following scores have been allocated. These are available in appendix C.

(29)

29

5. Results

The results section exists out of two parts. In the first part I will provide the descriptives and correlations that are present in the model. In the second part the results of the statistical tests are presented. In this part I will elaborate on the type of test performed and make

observations.

5.1 Descriptives

Table 2: Descriptives 2016

(30)

30 Table 4: Correlation of the 2016 sample

Table 2 presents descriptive statistics of 2016 for the total sample. The mean, median, mode and standard deviation are presented. These banks are considered in the regression models. Variations in the quality of risk disclosure (QRD) are present, ranging from 8 to 42. The other descriptives show large variations as well (e.g. logarithm of assets, dividends, leverage ratio). In these tables the abbreviations mean the following: QRD = total disclosure score; LNASS16 = natural logarithm of total assets; PROF = profitability; DIV = dividends; LNRWA16 = the natural logarithm of the risk weighted assets 2016; LAW = legal system, civil law = 0

common law = 1; FORACT = sales in foreign countries / total sales; LEVRB = Leverage ratio calculated as proposed by Basel; FORBOD = foreigners in board of directors divided by total members in the board of directors; STEX = amount of stock exchanges at which the bank is listed; FORSTEX = amount of foreign stock exchanges at which the bank is listed; UAI = uncertainty avoidance, derived from Hofstede (1991); LTO = long term orientation, derived from Hofstede (2011); PDI = power distance, derived from Hofstede (1991); GSIB = G-SIB rating, derived from the 2016 list of global systemically important banks;

In table 4 a correlation matrix is presented which includes the variables used in the research. This table provides us with information about collinearity between variables. In order to test for significances outside the scope of the research, I have decided to test more variables than described in the method section. The LEVR, which represent the usual leverage ratio

(debt/assets), is high correlated with LEVRB. Therefore, I have chosen to exclude this variable. Comparable choices have been made for FORSTEX, which shows a high degree of collinearity with STEX, and LNRWA16, which shows high collinearity with LNASS16.

(31)

31 These correlations can also be explained by logic. For example, two leverage ratios are

expected to be related as they calculate comparable numbers. In testing these variables with the hypotheses this collinearity was present. In comparing total assets with risk weighted assets collinearity was found. This is logical as in absolute values a bank with a high amount of assets will also have a high amount of risk weighted assets. For example, a bank with one billion in total assets will have higher amount of risk weighted assets than a bank with one million in total assets. Table 4 also shows us that the quality of risk disclosures and the relative presence of foreigners in the board of directors are significantly related. This may be an indication that our first hypothesis will be supported.

5.2 Regression

Multiple models are used to test the different hypothesis, using multiple regression. This method is used because it enables me to test the relationship between multiple independent variables and the dependent variable. I start with a model that only considers the relationship between the control variables and the dependent variable. These are described in the method section. The following formula is the starting point of the regression:

Quality of risk disclosure = α + β1LNASS16 + β2PROF+ β3DIV+ β4LAW+ β5FORACT+

β6LEVR+ β7GSIB

In table 5 this resembles model two. Whereas models two till five are calculated by adding the following variables: In model two the independent variable foreign board members is added, this is inserted as a percentage of the total board members of the bank. Β8FORBOD is added

to model one. In model three adds the independent variable country of origin, these countries are transformed to a nominal scale in order to make it applicable in the multiple regression analysis. B8COUNTRY is added to model one. In model four the independent variable stock

exchanges is added, this is inserted as the amount of stock markets in which a bank is listed. B8STEX is added to model one. Model five adds the independent variables that are related to

Hofstede’s model. These are uncertainty avoidance, long-term orientation and power distance, as previously determined in the method section. B8UAI, B9LTO and B10PDI are added to

model 5.

(32)

32 Table 5: Regression models

5.2.1 The base model

This model is the starting point of the regression analysis. The included variables have proven to be significantly related to the level of voluntary disclosure, this is set out in paragraph 4.3.3. This model is significantly related to the quality of risk disclosure with an alpha of less than 1%. The explanatory value of this model is 74.8%. This value is very high for a base model. This will be discussed in chapter 6.

5.2.2 Foreign board members

The first hypothesis focuses on the presence of foreigners in the board of directors. Model 2 in table 5 present the results of our hypothesis. This model explains 74.9% of the quality of risk disclosure. The explanatory power of adding the presence of foreign board members has a significance score of .750, showing no significance in providing explanatory value. Contrary to my expectations, the degree of foreigners in the board of directors has not been proven to be significantly related to the quality of risk disclosure. As previously mentioned, the board of directors is responsible for the disclosure policy of the firm. This was combined with the fact that foreigners are more likely to be independent (Hamzah & Zulkafli, 2014) and that

independent directors are more likely to voluntarily disclose information (Cheng & Courtenay, 2006).

(33)

33

5.2.3 Country of Origin

The second hypothesis is aimed at the country in which a bank is originated. Model 3 in table 5 presents the results of this hypothesis. This model explains 74.8% of the quality of risk disclosure. The explanatory power of adding the country of origin has a significance score of .858, showing no significance in providing explanatory value. As Wanderley et al. (2008) find there are differences in voluntary disclosure between countries. Therefore the expectation was that differences in QRD would be observed between countries.

5.2.4 Stock exchange

The presence on stock exchanges is anticipated to be positively related to voluntary disclosures. Therefore, companies that are listed on multiple stock exchanges should, accordingly, disclose even more. This is the third hypothesis. In model 4 the presence on stock exchanges has been added. Adding this variable has not proven to increase the explanatory value of the model, as adding this variable has a significance score of .239. Model 4 explains 76.2% of the quality of risk disclosure. The expectation was that banks listed on stock exchanges would voluntarily disclose more information.

5.2.5 Hofstede’s cultural dimensions

As previously explained, I expect that the culture of a bank can be related to its country. This effect is primarily controlled by the magnitude of where a bank conducts its business.

Previous research has shown that PDI, UAI and LTO are significantly related to disclosure. Now the question raises whether this is also applicable to the heavily regulated banking industry. Model 5 explains 76.8% of the quality of risk disclosure. This model shows no signs of significance in improvements of explanatory value when relating the chosen Hofstede’s cultural dimensions to the quality of risk disclosures. UAI, LTO and PDI show significance scores of respectively .336, .212 and .666.

5.2.6 Control variables

The control variables that have been used in this research have been proven to be significantly related to voluntary disclosure in other studies. In the regression most of these variables do not seem significant in explaining the quality of risk disclosure. The two variables that are significantly related to the quality of disclosure are the G-SIB rating and the legal system that is applicable. This raises questions that will be discussed in the next chapter.

(34)

34

6. Discussion

In this chapter I will discuss the results of the statistical analyses. This paper attempts to explain the quality of risk disclosure by looking at national aspects. Based on multiple

regressions the national aspects are not found to be significant in explaining the quality of risk disclosure. Besides national aspects, previously found evidence on what influences disclosure is not found to be significantly related as well. These results suggest that determinants of risk disclosure, within the European banking industry, are not necessarily related to bank specific characteristics. The high degree of explanatory value indicates that, although not significantly related, several components of the model do have some explanatory value. Especially the inclusion of GSIB-rating has proven to be a significant influence, which entails about 30 percent of the explanatory value.

When analyzing table 5 it become apparent that the control variables alone have a high explanatory value (74.8%). This potentially reduces the explanatory value of the tested independent variables, resulting in a sub-optimal control model. Table 4 shows a significant correlation between the percentage of foreign board members in the board of directors and the quality of risk disclosure. This however, has not been proven to be a significant influence in model 2. This indicates that while the presence of foreign board members does influence QRD, other variables explain QRD better. As this papers focusses on nationality a limitation of this research is the limited scope it comprises. In table 1 the amount of banks per country is presented. Eight out of the fifteen countries in this sample have three or less banks present in the sample, whereas four of the fifteen countries are represented by one bank. With respect to the hypothesis that the country of origin may explain the level of risk disclosure, the limited sample size may cause a considerable limitation with regards to the reliability of the research. A larger sample might yield different results, as country of origin effects can become

apparent.

The sample has been discussed in terms of quantity, but the characteristics of the sample are also interesting. Research on disclosure has, for example, found that firm size is a predictor of the quality of disclosure (Ahmed and Courtis, 1999). The sample of this research is very specific, as this research targets European banks that are part of EBA transparency exercise. As the sample only includes European banks. This may not be the best sample for assessing the relationship between nationality and the quality of risk disclosure. In analyzing, for example, Hofstede’s scores on the cultural dimensions, differences can be observed between,

(35)

35 for example, European and African countries. Intercontinental differences in the quality of risk disclosures by banks is a potential avenue for future research.

Other disclosure research commonly targets multiple industries (Archambault &

Archambault, 2003), therefore reducing comparability as the banking industry is considerably more regulated than most industries. This can assist in explaining why the variables used in this paper are not significantly related. The increases in regulation can cause the explanatory variables to lose significance. Disclosure theory expects that mandatory and voluntary disclosure are related. As the mandatory disclosure requirements increase, it become less costly to voluntarily disclose additional information (Bischof, 2009). Resulting in more voluntary disclosure as mandatory disclosure increases (Einhorn, 2005). This is a potential avenue for future research. A longitudinal study can be conducted in order to examine whether the quality of (voluntary) risk disclosures by (European) banks increases as regulatory requirements increase. At first glance differences between 2015 and 2016 are present. When comparing the QRD of table 2 and table 3, considerable differences in mean and standard deviation are present. A longitudinal study may provide more insights in the relation between mandatory and voluntary risk reporting in the banking industry.

While the limitations of this research have been mentioned, it has not yet been mentioned that the quality of risk disclosure shows considerable variations. In table 2 the minimum of QRD is 8 while the maximum is 42, the standard deviation of 7.97 also suggests that deviations are common. Although the G-SIB rating has proven to be a significant determinant in explaining the quality of risk disclosure, there are numerous banks in our sample without a G-SIB rating that have high QRD scores. The variation in QRD scores for these banks has not been

explained. So far, either due to the nature of the sample or the fact that nationality simply plays no role, little explanation has been found for the differences in the quality of risk disclosure. This leaves a gap which is not yet explained. As the G-SIB rating appears to be a significant predictor, future research may choose to exclude G-SIBs to investigate what explains the variation between non G-SIBs.

Quality Risk Disclosures by European banks: the role of nationality in the Quality of Risk Disclosures