Introduction - Erasmus University Rotterdam Web viewExecutive summary: Since the financial ... p. 3). The set of developed accounting standards is called International Financial Reporting

Master Thesis

Risk Disclosure Practices of the European Banking Sector

Evidence from the 2007 Annual Reports

Author:

Jaap-Willem Don (288497)1

Date:

14 August 2010

Executive summary: Since the financial crisis, there has never been so much attention towards risk

disclosure by financial companies. It is important that stakeholders receive relevant information to

reliably assess the risk profile of a company. This study investigates risk disclosure practices of 40

sampled banks in the European banking sector by analysing the risk information and discussing its

nature. The study reveals an increase in risk disclosure compared to earlier studies, indicating that

risk disclosure practices has been improved. However, the usefulness is still questioned as the

information is mainly qualitative, neutral and biased towards the past. Furthermore, this study

relates the extent of risk disclosure to firm specific characteristics as size, profitability, leverage,

multiple listing statuses and governance structures. However, only size has been found significant in

explaining the variation in risk disclosure.

Keywords: risk disclosure, annual reports, banking sector, European Union

Master: Accountancy, Auditing and Control

Erasmus School of Economics, Erasmus University Rotterdam

Supervisor: drs. R. van der Wal RA

Co-reader: ....

1 E-mail address: [email protected]

Table of contents

1 Introduction............................................................................................................................4

1.1 Prologue.................................................................................................................................41.2 Research question..................................................................................................................51.3 Sub questions.........................................................................................................................61.4 Relevance...............................................................................................................................61.5 Structure.................................................................................................................................7

2 Institutional Framework..........................................................................................................8

2.1 Introduction............................................................................................................................82.2 The annual report...................................................................................................................82.3 Accounting standards.............................................................................................................9

2.3.1 IFRS.................................................................................................................................92.3.2 IFRS 7 Financial Instruments: Disclosures.....................................................................102.3.3 Regulatory risk reporting..............................................................................................11

2.4 Basel Accords........................................................................................................................122.5 Banking supervision..............................................................................................................13

2.5.1 EMU..............................................................................................................................142.5.2 Central Bank and Financial Authorities.........................................................................14

2.6 Regulation and supervision and the credit crunch...............................................................142.7 Corporate governance codes................................................................................................152.8 Conclusion............................................................................................................................17

3 Theoretical Framework.........................................................................................................19

3.1 Introduction..........................................................................................................................193.2 Positive Accounting Theory..................................................................................................193.3 Agency theory.......................................................................................................................203.4 Stakeholder theory...............................................................................................................213.5 The role of disclosure...........................................................................................................223.6 Efficient market hypothesis..................................................................................................233.7 Risk and risk management....................................................................................................243.8 Banks and their (specific) risks..............................................................................................263.9 Risk disclosure: development and discussion.......................................................................273.10 Conclusion............................................................................................................................28

4 Prior Research.......................................................................................................................31

4.1 Introduction..........................................................................................................................314.2 Initiation of risk disclosure research.....................................................................................314.3 Extent of risk disclosure........................................................................................................32

4.3.1 Surveys..........................................................................................................................324.3.2 Content analysis............................................................................................................344.3.3 Disclosure index............................................................................................................384.3.4 Summarizing table........................................................................................................39

4.4 Factors that influence risk disclosure...................................................................................404.4.1 Studies on financial firms..............................................................................................414.4.2 Studies on nonfinancial firms........................................................................................424.4.3 Other disclosure studies................................................................................................464.4.4 Summarizing table........................................................................................................46

4.5 Usefulness of risk disclosure.................................................................................................484.6 Conclusion............................................................................................................................48

2

5 Research Design....................................................................................................................49

5.1 Introduction..........................................................................................................................495.2 Sample selection...................................................................................................................495.3 Content analysis...................................................................................................................505.4 Hypotheses development.....................................................................................................53

5.4.1 Size................................................................................................................................535.4.2 Profitability...................................................................................................................535.4.3 Level of risk...................................................................................................................545.4.4 Multiple listing..............................................................................................................545.4.5 Credit ratings................................................................................................................555.4.6 Organizational structure...............................................................................................555.4.7 Board composition........................................................................................................56

5.5 Regression analysis...............................................................................................................565.6 Independent variable definition...........................................................................................57

6 Analysis and Interpretation of the Results.............................................................................60

6.1 Introduction..........................................................................................................................606.2 Descriptive statistics of the sample......................................................................................606.3 Part I: Risk disclosure practices.............................................................................................60

6.3.1 Risk disclosure score.....................................................................................................606.3.2 Risk categories..............................................................................................................626.3.3 Risk characteristics........................................................................................................646.3.4 Conclusion.....................................................................................................................66

6.4 Part II: Regression analysis...................................................................................................666.4.1 Data analysis................................................................................................................666.4.2 Multiple regression.......................................................................................................676.4.3 Interpretation of the results..........................................................................................696.4.4 Conclusion.....................................................................................................................71

7 Conclusion............................................................................................................................72

7.1 Introduction..........................................................................................................................727.2 Research conclusion.............................................................................................................727.3 Limitations............................................................................................................................737.4 Recommendations for future research.................................................................................74

Reference List................................................................................................................................75

Appendix A...................................................................................................................................80

Appendix B....................................................................................................................................81

Appendix C....................................................................................................................................82

Appendix D...................................................................................................................................83

Appendix E....................................................................................................................................84

Appendix F....................................................................................................................................85

Appendix G...................................................................................................................................86

Appendix H...................................................................................................................................87

Appendix I.....................................................................................................................................88

3

1 Introduction

1.1 Prologue

The recent financial crisis is clearly related to risk and risk behavior of individuals and companies as a

whole. The enormous and worldwide consequences the crisis revealed raises questions about the

awareness of risks companies are exposed to, and moreover about the techniques in place to

manage those risks.

In today’s financial system and economic world, banks fulfill an important position, as they allocate

funds from savers to borrowers. Performing this task in an efficient manner lowers the cost of capital

to firms, boosts capital formation and stimulates productivity growth. As a result, the functioning of

banks has ramifications for the operations of firms and the prosperity of nations (Levine, 2004, p. 2).

Since the evolution of financial derivatives, a market with a total value of $ 500 trillion in 2008

(Ormerod, 2008), banks became manufacturers and traders of complex financial products.

The stability of the financial system was heavily affected by the turmoil in financial markets in the

second half of 2007 in the second half of 2007 caused by the sub-prime mortgage loans crisis in the

US (Ryan, 2008, p. 1606). Investments in complex financial products turned out to have much greater

risk than assumed and indicated by credit rating agencies, especially towards the conditions on the

American housing market. As house prices in the US started to fall, banks worldwide were forced to

make major write-downs on the values of these investments. The effects of this turmoil were

transferred also into the operations of the European banks, as distrust between banks and

uncertainty in financial markets started to rise. Banks began to charge additional risk premium for

their interbank loans, resulting in a shortage of credit and illiquid markets.

The essence of any bank is that it is a risk taking enterprise, and therefore, it is expected that

relevant risk-related information will be released to the marketplace to reduce the uncertainty of

investors. Already, banks disseminate significant amounts of information to the marketplace.

However, there have been calls for even greater disclosure of information to ensure users are fully

able to assess the performance of a firm. If shareholders and other interested parties are to be able

to understand the risk profile of the firm, they need to receive information about the risks a firm

faces and how the directors are managing those risks (Linsley and Shrives, 2005a, p. 205).

4

Corporate risk reporting is a cornerstone of accounting. For corporations, information on risk can

help to manage change, lower the cost of capital and instruct on the future trajectory of business.

For investors, information on risk can help to determine the risk profile of a company, the estimation

of market value and accuracy of security price forecasts (Abraham and Cox, 2007, p. 227). Risk

reporting has gained interest in financial accounting practice, regulation and research in recent years

(Dobler, 2008, p. 185). Inevitably, this attention will increase due to the current financial crisis. The

crisis shows the urgent need for transparent risk reporting and must be used to learn lessons from,

to prevent current problems to occur again in the near future.

In 2005, Linsley and Shrives (p. 205) already noticed the opinion that at that time limited risk

disclosure occurred by banks. As a result banks were not transparent in this respect. The distrust

between banks, between banks and related parties and between banks and its customers during the

recent credit crunch supports this notion. Improved risk transparency could reduce uncertainty on

the financial markets and therefore tribute to a stable financial environment. Hence, it should

enhance the ability of shareholders and other related parties to manage their risk positions.

1.2 Research question

This thesis focuses on the risk reporting of banks, by first describing their risk disclosure practices,

and second investigating various firm-specific determinants related to the amount of this disclosure.

The following research question will be answered in this research:

In which way do European banks disclose risks in the annual reports of 2007 and which firm-

specific characteristics influence the extent of risk disclosure?

Using risk-related content in the 2007 annual report of 40 randomly selected European Union banks,

this research seeks to explore how risk-related information is communicated to the market by means

of the annual report. The primary objective is to examine the nature and characteristics of risk

disclosures. In addition, this research investigates which firm-specific factors are related to the extent

of this disclosure. In short, this study aims to influence the understanding and thus the practice in

relation to the extent of risk disclosure in the annual reports of the European banking sector and

could therefore be helpful for regulators. They could assess the effectiveness of their regulations and

ensure it has the best chance of achieving the desired outcomes.

The empirical research consists of two steps. First, the risk-related information in the annual reports

is investigated by means of content analysis, a method that has proven to be able to give insight in

5

(risk) disclosure practices in the annual report. This information will be categorized into major risk

areas and different attributes (form, qualification and period) will be distinguished. Second, the

gathered data is examined by means of regression analysis, to test for relationships between the

extent of risk disclosure and potential explanatory firm characteristics. These include variables such

as company size, profitability, risk level, multiple listing, credit rating and corporate governance

characteristics.

1.3 Sub questions

For the sake of clarity, the main research question is split up into five sub questions, which will be

discussed in the following chapters:

What is the relevant law and regulation regarding risk disclosures in the annual report?

(chapter 2)

Which economic theories can explain the occurrence and extent of risk disclosure?

(chapter 3)

What are the results of prior risk disclosure research? (chapter 4)

What are the risk disclosure practices of European banks in 2007? (chapter 6)

What firm-specific characteristics are related to the extent of risk disclosures? (chapter 6)

1.4 Relevance

This thesis is embedded in the developing debate as to how the different users of annual reports can

be provided with risk and risk management information that enables them to assess the risk profile

of firms. This debate about risk disclosure started with a Basel Committee paper in 1998, titled

“Enhancing Bank Transparency.” It discussed the significance of disclosure and transparency within

the context of market discipline and banking supervision. The Basel Committee on Banking

Supervision (BCBS, 1998, p. 15) defines bank transparency as “the public disclosure of reliable and

timely information that enables users of that information to make an accurate assessment of a bank’s

financial condition and performance, its business activities and the risks related to those activities”.

Noteworthy, it is argued that banks are generally more opaque than nonfinancial firms (Levine, 2004,

p. 2). For example, Morgan (2002) found evidence that bond analysts disagree more over the bonds

issued by banks than by nonfinancial firms.

Despite the developing debate, Linsley et al. noticed in 2006 (p. 268) that there has been relatively

little research into risk disclosure practices. Discussions concerning how risk reporting can be

developed are often unsupported with empirical evidence. Linsley et al. intend to initiate risk

6

disclosure research, and therefore this study responds to this call. Furthermore, the majority of prior

studies on (risk) disclosure have examined annual reports of nonfinancial firms. In addition, the

current financial crisis led to an increased attention towards the banking sector and risk management

in this sector. This further adds up relevance to this subject. Following Linsley et al. (2006) this study

therefore examines risk and risk management disclosure practices within the banking sector.

1.5 Structure

The structure of this thesis is as follows. Chapter 2 elaborates on the institutional framework in which

risk disclosure is embedded. Relevant law and regulation towards risk disclosure will be discussed. In

chapter 3 the theoretical framework is explained, entailing relevant economic theories and

approaches applied to the research subject. In addition, this chapter covers the concepts of risk and

risk management and describes the recent developments regarding risk disclosure. Chapter 4

enumerates prior research on risk disclosures and related aspects. Chapter 5 presents the research

design for the empirical study. In the first part hypotheses will be developed. Furthermore, it explains

the methodology and develops a conceptual model for the regression analyses. The results of the

empirical study will be presented in chapter 6 as well as an interpretation of these results. The thesis

ends with a conclusion, mentions limitations of this study and provides recommendations for future

research in chapter 7.

7

2 Institutional Framework

2.1 Introduction

This chapter elaborates on the institutional framework in which risk disclosure is embedded. The

following question will be answered:

What is the relevant law and regulation regarding risk disclosure in the annual report?

First this chapter will discuss the annual report, as a main element in financial reporting. The

subsequent paragraphs will reveal the relevant law and regulation for public listed banks in the

European Union regarding risk disclosures. This regulation subject banks to certain requirements,

restrictions and guidelines: the accounting standards IFRS and the Basel Accords. Furthermore

attention is given to banking supervision in the European Union and the role of central banks and

financial authorities, especially during the credit crisis. Finally, the role of corporate governance

codes is discussed.

2.2 The annual report

Risk disclosures can be found in the annual report, which is a document that is mandated to be

produced every year and expected to provide useful information to users for better decision-making

(Amran et al., 2009, p. 40). The annual report constitutes the main channel of information on the

company’s situation to external users and has been defined as the “main disclosure vehicle” (Beretta

and Bozzolan, 2004). It is an example of the concept of financial reporting, which is the

communication of a wide range of financial and other data to outside interested parties. Information

in the annual report is therefore intended to inform shareholders and other interested parties about

the company’s activities and financial performance.

Shareholders and other related parties have a need for information about the company to make

decisions, for example about buying or selling shares or to provide credit. There is however a conflict

of interests in determining the right information and the right amount of information that should be

communicated. From the company’s perspective as a supplier of information, there is an incentive to

keep information private as much as possible. Shareholders and other outside (related) parties have

the interest to gather as much information as they need to make economic justified decisions.

However, too much information clouds the understanding. Moreover, the demand side is not

8

homogeneous, which means that their interest can differ regarding their relationship with the

company (Klaassen and Hoogendoorn, 2004, p. 23-24).

As Klaassen and Hoogendoorn conclude, there needs to be an alignment of supply and demand of

information, which in most western countries with developed capital markets is done by law and

regulation. Deegan and Unerman (2006, p. 32) state that financial accounting tends to be heavily

regulated in these countries, with many accounting standards and other regulations governing how

particular transactions and events are to be recognized, measured and disclosed. The generated

financial statements, such as the balance sheet, income statement, statement of cash flows and

supporting notes are directly impacted by various accounting regulations.

Regulation is in this context the set of rules and standards that govern (financial) institutions. Their

main objective is to foster financial stability and to protect the customers of financial services. It can

take different forms, from information requirements to strict measures such as capital requirements.

Regulation can be distinguished from supervision, which is the process designed to oversee financial

institutions in order to ensure that rules and standards are properly applied (De Larosière, 2009, p.

13). The next paragraphs cover the relevant law and regulation regarding risk disclosures.

2.3 Accounting standards

Accounting standards are authoritative statements of how particular types of transaction and other

events should be reflected in the financial statements. These standards include specific principles,

bases, conventions, rules and practices necessary to prepare the financial statements (Stolowy and

Lebas, 2006, p. 13). The national governments or specific regulatory bodies across the European

Union and the supranational European Commission can be seen as the most important regulators of

the accounting standards. However, in many countries and also in the European Union, accounting

standards are not developed by the government, but created by independent professional bodies.

After a successful standard-setting process, standards are implemented by national governments.

Nowadays, the European Commission implements the accounting standards developed by an

international standard-setting board, the International Accounting Standards Board (IASB).

2.3.1 IFRS

The IASB, based in London, is committed to “develop, in the public interest, a single set of high

quality, global accounting standards that require transparent and comparable information in general

purpose financial statements” (IASB, 2006, p. 3). The set of developed accounting standards is called

International Financial Reporting Standards (IFRSs), applied throughout the European Union as of 1

9

January 2005. IFRSs and their interpretations are developed through a formal system of due process

and broad international consultation (IASB, 2006, p. 5-6). IFRS is meant to be an uniform reporting

system, to be applied globally and thereby to enhance the transparency and comparability of

companies throughout the world.

2.3.2 IFRS 7 Financial Instruments: Disclosures

An important standard for banks and related to the research subject of this paper is IFRS 7 Financial

Instruments: Disclosures. This standard incorporates the disclosures relating to financial instruments

required by IAS 32 and replaces IAS 30, a standard about disclosures that was intended only for

banks and similar financial institutions. This means that since the introduction of IFRS 7 all financial

instruments disclosure requirements are located in a single standard for all types of companies (Ernst

& Young, 2007, p. 1). The objective of IFRS 7 is “to require entities to provide disclosures in their

financial statements, that enable users to evaluate (1) the significance of financial instruments for the

entity’s financial position and performance; and (2) the nature and extent of risks arising from

financial instruments to which the entity is exposed during the period and at the reporting date, and

how the entity manages those risks” (IASB, 2009). Three main risks are mentioned: market risk, credit

risk and liquidity risk (PwC, 2006, p. 4).

The IFRS 7 disclosure requirements require both qualitative narrative descriptions and specific

quantitative data. The level of detail of such disclosures should not overburden users with excessive

detail, but equally should not obscure significant information as a result of excessive aggregation.

Risk disclosures, according to IFRS 7, do not have to be given in the financial statements, but may

either be provided in the financial statements or incorporated into the financial statements by

reference from another statement, for example the management commentary or a separate risk

report. However, as the risk disclosures are required by IFRS, they will be subject to audit (Ernst &

Young, 2007, p. 1).

The requirements to provide quantitative and qualitative market risk disclosures are new and may

represent a significant challenge for many companies (PwC, 2006, p. 2). The IASB defines the

qualitative disclosures as management’s objectives, policies and processes for managing those risks.

The quantitative disclosures has to provide information about the extent to which the entity is

exposed to risk, based on information provided internally to the entity’s key management personnel

(IASB, 2009). As a result, this standard should improve transparency and assist investors in assessing

how companies are managing their risks.

10

IFRS 7 applies to all entities, from a manufacturing company whose only financial instruments are

accounts receivable and accounts payable, to a financial institution most of whose assets and

liabilities are financial instruments (IASB, 2009). It is obvious that for banks the IFRS 7 disclosures will

be substantially significant, since they are heavy users of financial instruments.

2.3.3 Regulatory risk reporting

The status of current regulation of risk reporting under IFRSs is predominantly focused on the market

risk associated with the use of financial derivatives, as described in the previous paragraph.

According to Beretta and Bozzolan (2004, p. 267) and Cabedo and Tirado (2004, p. 181), this also

holds for the accounting standards system in the US, the so-called US GAAP (Generally Accepted

Accounting Principles). SFAC No. 133 establishes the compulsory disclosure of market risks arising

from the use of financial assets. Furthermore, Financial Reporting Release No. 48 (FRR 48, 1997)

requires SEC registrants to disclose both qualitative and quantitative information on market risks

arising from adverse changes in interest and foreign exchange rates and in stock and commodity

prices (Cabedo and Tirado, 2004, p. 183). This information needs to be disclosed in the notes to the

accounts and also in the management, discussion and analysis (MDA) section. The same regulatory

principles are applied in Canada. In the UK, the operating and financial review (OFR can be seen as

the equivalent of MDA), was introduced in 1993 for listed companies and is still non-mandatory. In

this review, companies are recommended to disclose its review of key risks and it strongly

encourages the inclusion of a clear discussion of trends affecting the future. In Germany, GAS5

requires information on risks to be presented in a separate section of the management report

(Amran et al., 2009, p. 41).

Although the basic disclosure requirements are comparable, additional regulation per country results

in significant differences in regulation regarding risk reporting. To illustrate this variation appendix A

shows a comparison of risk reporting requirements in the United States of America, according to

IFRSs and in Germany. Most regimes including IFRS and US GAAP follow a piecemeal approach. They

mandate selected risk-related disclosures referring to specific categories of risks and miss therefore a

comprehensive and systematic approach on the reporting of relevant risks (Meijer, 2003, p. 110). An

exception is Germany that has a separate standard (GAS 5) on comprehensive risk reporting. The risk

reporting requirements of US GAAP and IFRSs are however roughly comparable (Dobler, 2008, p.

185).

11

2.4 Basel Accords

The Bank for International Settlements (BIS), headed in Switzerland, is an international organization

which fosters international monetary and financial cooperation and serves as a bank for central

banks. One of the key objectives is to promote monetary and financial stability in the banking sector.

For this reason a special committee was appointed in 1974 to enhance understanding of key

supervisory issues and improve the quality of banking supervision worldwide, called the Basel

Committee on Banking Supervision (BIS, 2009).

The Basel Committee does not possess any formal supranational supervisory authority. Rather, the

committee formulates broad supervisory standards and guidelines. It also recommends statements

of best practice in banking supervision in the expectation that the national authorities of their

members will take steps to implement those (BIS, 2009). Their work resulted in the Basel Accords,

which are some of the most influential agreements in modern international finance. Drafted in 1988

and 2004, Basel I and II have ushered in a new era of international banking cooperation. Both accords

have helped harmonize banking supervision, regulation and capital adequacy standards (Balin, 2008,

p. 1). In January 2007 the Basel II Accord was implemented in the European Union via the Capital

Requirements Directive (CRD)2, which is legally binding for every member state of the EU. The CRD

gave financial institutions the option whether to apply the old Basel I rules or the revised Basel II

rules during 2007. However, as of 1 January 2008 institutions have to follow the revised rules of Basel

II (De Nederlandsche Bank, 2007).

The principal purpose of the framework is to create a more stable banking system, based upon three

complementary pillars: (1) capital adequacy requirements, (2) supervisory review and (3) market

discipline. Pillar 1 formulates guidelines for the calculation of a bank’s minimum capital requirements

and pillar 2 sets out the roles and responsibilities of supervisors in reviewing banks’ assessments of

their risks and capital requirements. Pillar 3 is the part of the new accord that has the greatest

significance for this thesis, as it sets out the risk disclosures required to ensure that the market

discipline mechanism can work effectively (Linsley and Shrives, 2005a, p. 207). Appendix B shows a

summary of the disclosure requirements in the 2004 framework, including quantitative and

qualitative disclosures.

The third pillar of Basel II provides a framework whereby banks’ risk exposures will be disclosed to

the marketplace under a common, standardized approach (Linsley and Shrives, 2005a, p. 210). Balin

(2008, p. 12) states, in line with Linsley and Shrives (2005a) that Pillar 3 intends to increase the

2 Consisting of the Directives 2006/48/EC and 2006/49/EC.

12

market discipline within a country’s banking sector. The Basel II Accord recommends that disclosures

of a bank’s capital and risk-taking positions, which were once only available to regulators, should be

released to the general public. Furthermore, a full description, including the assumptions, of the risk

mitigation approaches of a bank is recommended for quarterly release to the general public. Balin

(2008, p. 12) point out that Basel II hopes to empower shareholders to enforce discipline in the risk-

taking and reserve-holding methods of banks, where banks seen to hold too few reserves and take

on too much risk are punished by their own shareholders for doing so. This follows the notion of

Linsley and Shrives (2005a, p. 206), who state that, following the importance of transparency,

participants in the marketplace can sanction banks whose financial condition or risk profile is

considered, in some sense, unsatisfactory.

The final version of the Basel II framework was published in July 2006 and, as said before, became

legally binding for every member state of the EU as of January 2007. Basel II was developed to meet

some serious shortcomings in Basel I and greatly expanded the scope, technicality and depth of the

original Basel Accord (Balin, 2008, p. 6). However, Basel II was criticized too. According to the former

general manager of the BIS there are five major criticisms: (1) it is too complex; (2) it reinforces the

procyclicality of the financial system; (3) it gives too much weight to the judgments of rating

agencies; (4) it penalizes small and medium sized enterprises and (5) it does not take adequate

account of the special situation and concerns of emerging countries (Crockett, 2003). Officially there

is no sign of the development of a successor of Basel II. However, already in 2002 Crockett declared:

“The financial world changes all the time and those who supervise and regulate the financial world

have to be prepared to change too. So I do not have any doubt that soon after Basel II is in place, we

will have to start to think about Basel III” (Crockett, 2002). The current financial crisis, which led to an

impressive pressure on the financial system worldwide, even more shows an urgent need to develop

a new or revised framework. At a summit in March 2009, this was recognized by the leaders of the

European Union which led to an agreement that Basel II needs a thorough revision (Visser, 2009).

2.5 Banking supervision

As stated in paragraph 2.2, regulation can be distinguished from supervision. Supervision is the

process designed to oversee financial institutions in order to ensure that rules and standards are

properly applied (De Larosière, 2009, p. 13). This means that banks are monitored by central banks

and financial authorities, because of their important role in today’s economy. In some countries in

the European Union, banking supervision is carried out by the central bank. In other countries this

task is performed by another institution, sometimes in close cooperation with the central bank

(Eijffinger, 2005, p. 459). Together, central banks and financial authorities have the task to supervise

13

the payment system, ensure price stability and pursue the stability of banks (Padoa-Schioppa, 1999,

p. 296).

2.5.1 EMU

The European Monetary Union (EMU) is a single market with a common currency, where most (16

out of 27) European Union states are participating in, thereby adopting the euro as their currency.

The remaining countries still use their own currency, including the United Kingdom, Sweden and

Denmark. According to Padoa-Schioppa (1999, p. 296), the institutional framework of banking

supervision in the EMU consist mainly of two parts, national competence and cooperation.

First, national competence. According to the ‘single license’ principle, every bank has the right to do

business in the whole area, under the supervision and regulation of the authority that has issued the

licence. The full supervisory responsibility thus belongs to the home country. Second, cooperation.

The banking industry, a highly regulated single market that retains a plurality of national supervisors,

requires close cooperation among supervisors to safeguard the public good: namely, openness,

competition, safety and soundness of the banking industry (Padoa-Schioppa, 1999, p. 296).

2.5.2 Central Bank and Financial Authorities

Nowadays, most countries in the European Union follow a ‘separation approach’ in banking

supervision, which means that the banking supervision has been assigned to a separate institution

such as a financial authority, instead of the central bank (Padoa-Schioppa, 1999, p. 256). This trend

for central banks to retreat from supervisory functions can be explained as follows. First, banking is

becoming an increasingly complex business and less clearly defined. Second, new developments in

financial supervision increasingly emphasize the role of self-regulation and internal risk management

in financial institutions. Finally, there is increasing acceptance that the government, not the central

bank, should take responsibility for ultimate financial support (Eijffinger, 2005, p. 459).

2.6 Regulation and supervision and the credit crunch

Regulators and supervisors of the banking industry have met considerable critics on their role in the

emergence of the credit crunch. Two main critics can be distinguished, on the requirements of Basel

II and on the role of supervisors.

First, Basel II is a procyclical system. This means that banks facing defaults on credit payments need

more capital, which they can hardly attract because of their worsening liquidity position. Another

problem with Basel II is that banks are allowed to use their own internal risk models to value their

14

investments. This means that companies are encouraged to make lower risk estimates. In the

preceding years of economic growth, before the current credit crunch, this has led to a decline in the

capital hold by banks, making them more vulnerable to deteriorating economic circumstances. A

third problem is the scope of the Basel II Accords: banks moved their riskiest investments away from

the balance sheet and put them into special purpose vehicles (SPVs), thereby circumventing the

requirements (Visser, 2009).

Second, the supervision of the banking industry lies primarily at the national central bank and/or

local financial authorities. This approach has among certain advantages one major disadvantage

related to systemic risk. It was too much focused on the micro-prudential supervision of individual

financial institutions and not sufficiently on the macro-systemic risks of a contagion of correlated

horizontal shocks. Furthermore, strong international competition among financial centers

contributed to national regulators and supervisors being reluctant to take unilateral action, known as

‘regulatory competition’ (De Laroisère, 2009, p. 11).

However, not all supervisors were inadequate in their supervision. One example can be found in

Spain. The supervision, carried out by the Spanish central bank (Banco de España), is more

conservative and has more stringent requirements than Basel II. The additional requirements have a

countercyclical character, which means that in good times banks are forced to make more provisions,

to use them in bad times. Before the credit crunch, Spanish banks had therefore a much stronger

capital position than other European Banks (Visser, 2009). Furthermore, the central bank demanded

that banks post an 8 per cent capital charge against assets put in SPVs. That essentially killed this

lucrative business by removing incentives to create these creatures (Tett, 2008).

2.7 Corporate governance codes

A recent development in the worldwide corporate business world is the application of corporate

governance codes for public companies throughout many western countries. This development was

a response to some high-profile accounting fraud scandals, involving well-respected companies as

energy giant Enron (2001) and internet company WorldCom (2002) in the US. In Europe, accounting

fraud was discovered at dairy giant Parmalat (2003) in Italy and international food retailer Ahold

(2003) in the Netherlands (Mecking, 2008). In the wake of these corporate scandals there has been

an increased demand for more disclosures (Amran et al., 2009, p. 39).

Especially regarding these scandals, the integrity of financial reporting is a consistent concern among

regulators and practitioners (Chen et al., 2007, p. 3). As a consequence of these improper practices

15

attention has been focused on the need for strong corporate governance mechanisms within

companies (Davidson et al., 2005). Since 1992, reports, best practices, guidance and codes about

corporate governance for public listed companies were published and implemented in at least 67

countries worldwide (European Corporate Governance Institute, 2009). Cabedo and Tirado (2004, p.

183) pointed out that these documents, among other issues, emphasize on the need to disclose the

risks a firm faces, as part of their internal control systems.

Corporate governance is defined as the set of mechanisms (both institutional and market-based) that

induce the self-interested managers of a company to make decisions that maximize the value to its

owners, the suppliers of capital (Denis and McConnell, 2003, p. 2). The concepts of corporate

governance heavily rely on the necessity of internal controls. Internal controls help ensure that

processes operate as designed and that risk responses in risk management are carried out. The

guidance on the Turnbull Report (the UK corporate governance code) states: “A company’s system of

internal control has a key role in the management of risks that are significant to the fulfilment of its

business objectives” (Financial Reporting Council, 2005, p. 3). Transparency will be enhanced by

improving the disclosure of internal control. That should ultimately lead to improvements in the

communication links between investors and their investee companies (Solomon et al., 2000).

A wide variety of organizations in the EU Member States have issued corporate governance codes,

including governmental or quasi-governmental entities, committees or commissions organized or

appointed by governments, stock-exchange-related bodies and business, industry and academic

associations. Given the variety of the groups involved in developing codes, compliance mechanisms

and the “official” status of codes varies widely. However, all of the codes call for voluntary adoption

of their substantive recommendations. Some codes advocate or mandate (through linkage to stock

exchange listing requirements) disclosure by listed companies of the degree to which they comply

with code recommendations, together with an explanation of any areas of non-compliance. Such

disclosure against a code is referred to as disclosure on a “comply or explain” basis (Weil, Gotshal &

Manges LLP, 2002, p. 2, 16). Despite the divergence of corporate governance practices amongst EU

Member States, the European Commission reached the conclusion that a uniform corporate

governance code should not be adopted. Alternatively, the European Commission decided to

harmonize the codes’ enforcement mechanisms through Directive 2006/46/EC, which introduced the

comply-or-explain principle for the first time in European law. Although the concept of comply-or-

explain had already been in place in many EU Member States, the Directive makes the use of this

concept mandatory (RiskMetrics Group, 2009, p. 11, 27).

16

Also the Basel Committee on Banking Supervision (BCBS) focused attention towards corporate

governance and emphasized the urgency of good corporate governance in banks. The bank published

a paper in 1999 named: ‘Enhancing corporate governance for banking organizations’ (revised in

2006). This paper was issued to supervisory authorities and banking organizations worldwide to help

ensure the adoption and implementation of sound corporate governance practices by banking

organizations. One of the statements is that effective banking supervision is difficult if sound

corporate governance is not in place (BCBS, 1999, I.3). Hence, banking supervisors have a strong

interest in ensuring that there is effective corporate governance at every banking organization. The

crucial importance of strong corporate governance is underscored by the fact that banking is virtually

universally a regulated industry and that banks have access to government safety nets (BCBS, 1999,

II.7). The relation between corporate governance, transparency and risk disclosures will be discussed

further in paragraph 3.3 and beyond.

Another important notion in the BCBS 1999 paper is that there are no universally correct answers to

structural issues and that laws need not to be consistent from country to country, due to various

corporate governance structures in different countries. The paper therefore states that sound

governance can be practiced regardless of the form used by a banking organization (BCBS, 1999).

Regarding to this research, it is therefore complex to investigate relationships between risk

disclosures and corporate governance variables.

2.8 Conclusion

This chapter investigated the institutional framework of risk disclosures, in the context of relevant

law and regulation. For public listed banks in the European Union this includes the accounting

standards IFRS, in particular IFRS 7 (Disclosures of Financial Instruments), and the Basel Accords,

containing recommendations on banking laws, regulation and supervision. Furthermore, additional

regulation can vary per individual country. In addition, there has been increased attention towards

corporate governance in organizations including banks. One conclusion is that the regulation mainly

has been harmonized throughout the European Union. The execution of banking supervision is

however primarily national oriented, by the national central bank and/or financial authority. Using

the information in this chapter, a figure has been drawn which summarizes the situation of a public

listed bank in the European Union (see figure 1).

The recent credit crunch revealed two major weaknesses of this system. First, the regulation was

blamed to be too procyclical and second, the supervision was not sufficiently focused on macro-

systemic risks. Not surprisingly, the crisis evoked a discussion about the current status of the whole

17

banking regulation and supervision system, and therefore a reform (in the direction of stricter

regulation and more supervision) can be definitely expected.

Figure 1 The institutional framework of a public listed bank in the European Union

18

3 Theoretical Framework

3.1 Introduction

This chapter discusses the theoretical framework regarding risk disclosures in the annual report. The

following question will be discussed:

Which economic theories can explain the occurrence and extent of risk disclosure?

The Positive Accounting Theory is discussed in the second paragraph. Another applicable economic

theory is the agency theory, which is discussed in the third paragraph. The emphasis lies upon the

information asymmetry between the agent and the principle. Information that is disclosed in the

annual report is a major tool to overcome this information asymmetry. The role of disclosure and risk

disclosure in specific is therefore discussed in the fourth paragraph. Information plays an essential

role in financial markets and for that reason the efficient market hypothesis is discussed in the fifth

paragraph. The sixth paragraph starts with describing the concept of risk and the process of risk

management. The next paragraph explains the most common banking risks. In the eighth paragraph

a brief development of risk disclosure is given and attention is drawn to some issues arising out of

the risk disclosure debate. The chapter ends with a conclusion.

3.2 Positive Accounting Theory

The preparation of the financial statements is generally the responsibility of the (board of) directors

(Abraham and Cox, 2007, p. 231). Within the overall process of corporate governance, they make

accounting policy choices and determine to which extent information such as risk disclosure is

disclosed or not. The Positive Accounting Theory (PAT), popularized by Watts and Zimmerman,

studies managers’ accounting policy choices and is therefore important for this research.

According to Scott (2006, p. 240), PAT is “concerned with predicting such actions as the choices of

accounting policies by firm managers (…)”. Viewing a firm as a nexus of contracts, PAT asserts that

accounting policy choice is part of the firm’s overall need to minimize its cost of capital and other

contracting costs (Scott, 2006, p. 252). Many of these contracts involve accounting variables. Giving

management flexibility to choose from a set of accounting policies opens up the possibility of

opportunistic behavior. This means that managers may choose accounting policies from the set for

their own purposes, thereby reducing contract efficiency. The reason for that is the assumption that

managers are rational (like investors) and are therefore expected to choose accounting policies in

19

their own best interests if able to do so (Scott, 2006, p. 240-243). As a solution, PAT predicts that

contractual arrangements will be put in place to align the interests of the various self-interested

parties (Deegan and Unerman, 2006, p. 252).

The predictions made by PAT are largely organized around three hypotheses. First, the bonus plan

hypothesis predicts that managers of firms with bonus plans are more likely to choose accounting

procedures that shift reported earnings from future periods to the current period. Second, the debt

covenant hypothesis predicts that the closer a firm is to violation of accounting-based debt

covenants, the more likely the firm manager is to select accounting procedures that shift reported

earnings from future periods to the current period. Third, the political cost hypothesis predicts that

the greater the political costs faced by a firm, the more likely the manager is to choose accounting

procedures that defer reported earnings from current to future periods. This hypothesis introduces a

political dimension into accounting policy choice. Political costs can be imposed by high profitability,

which may attract media and consumer attention. Such attention can quickly translate into political

pressure on the firm as politicians may respond with taxes or other regulation (Scott, 2006, p. 243-

244).

The political cost hypothesis has been suggested to explain why firms make social disclosures.

Numerous empirical studies have directly sought to establish evidence for the political cost

hypothesis as an explanation for firms’ social disclosures (Milne, 2001, p. 1). Milne explains that,

according to Watts and Zimmerman’s theory, the answer why managers make social disclosures is

because it is in their interests to do so. In my opinion, this reasoning can also be applied to

(voluntary) risk disclosures. Large profits could for example be explained by disclosing information

about superior risk management.

3.3 Agency theory

To study the process of contracting (as emphasized by the PAT) the agency theory is very useful. The

agency theory models the process of contracting between two or more persons (Scott, 2006, p. 259).

The theory was first developed by Jensen and Meckling (1976) and proposes that within the joint

stock company the interests of managers diverge from those of the owners. The principals (owners)

will assume that the agents (managers) will be driven by self-interest. Therefore the principals will

anticipate that the agents will undertake self-serving activities that could be detrimental to the

economic welfare of the principals (Deegan and Unerman, 2006, p. 213-214). In this research, the

principals can be defined as the shareholders and/or users of the financial statements and the agents

are the managers of the banks.

20

According to Prat (2005, p. 862), there is a widespread consensus that principal-agent relationships

should be transparent as possible. If agency relationships are not transparent, some parties have an

information advantage over others, which is called information asymmetry. This is caused by the

position of the agents within the firm where they will have access to information not available to the

principals (Deegan and Unerman, 2006, p. 224). This information asymmetry between the agent and

the principal can be effectively mitigated if the firm is able to provide better levels of transparency

and disclosure (Chen et al., 2007, p. 658). In this theoretical framework, transparency is defined as

“the ability of the principal to observe how the agent behaves and the consequences of the agent’s

behavior” (Prat, 2005, p. 862). Prat argues that transparency improves accountability, which in turn

aligns the interests of the agent with the interests of the principal.

Related to the banking sector, Linsley and Shrives (2005a, p. 206) argue that the importance of

transparency is that if relevant information is released into the public domain, then participants in

the marketplace can sanction banks whose financial condition or risk profile is considered

unsatisfactory. On the contrary, banks whose financial condition or risk profile indicates that they are

well managed can benefit from disclosing appropriate information since the market’s disciplining

mechanism can offer incentives as well as sanctions.

3.4 Stakeholder theory

Another positive theory3 is the stakeholder theory. This theory provides alternative explanations

(alternative to PAT) about what drives an organization to make particular disclosures (Deegan and

Unerman (2006, p. 258). Furthermore, this theory is strong in explaining the interrelatedness of a

company and its stakeholders and has been used in multiple other disclosure studies (Amran et al.

(2009, p. 44). Developed by Freeman, a stakeholder is defined as “(…) any group or individual who

can affect or is affected by the achievements of an organisation’s objectives” (Freeman, 1984, p. 46).

Essentially, stakeholder theory is about the dynamic and complex relationship between the

organization and its surroundings (Amran et al., 2009, p. 44). The organization is considered to be

part of a wider social system. Moreover, it is assumed that the expectations of the various

stakeholder groups will impact on the operating and disclosure policies of the organization. The

organization will not respond to all stakeholders equally, but rather, will respond to those that are

deemed to be powerful (Deegan and Unerman, 2006, p. 289). As a result, the organization will tend

to satisfy the information demands of those stakeholders who are important to the organization’s

ongoing survival. The disclosure of information is considered to represent an important strategy in

3 A positive theory is a theory that seeks to explain and predict particular phenomena (Deegan and Unerman, 2006, p. 206)

21

managing stakeholders (Deegan and Unerman, 2006, 299). For example, Linsley and Shrives (2005b,

p. 302) point out that the positive relationship between company size and number of disclosures is

thought to be hold because larger companies have greater numbers of stakeholders to whom they

are accountable and as a consequence, they must provide more information.4

3.5 The role of disclosure

Corporate disclosure is critical for the functioning of an efficient capital market. Firms are providing

disclosure through regulated financial reports, including financial statements, footnotes and other

regulatory filings. The demand for financial reporting and disclosure arises from information

asymmetry and agency conflicts between managers and outside shareholders. The disclosure of

relevant information enables shareholders to evaluate the performance of the managers (Healy and

Palepu, 2001, p. 406, 412). This view is shared by the Basel Committee on Banking Supervision:

“Transparency of information related to existing conditions, decisions and actions is integrally related

to accountability in that it gives market participants sufficient information with which to judge the

management of a bank” (BCBS, 1999). In addition to mandatory disclosures, firms engage more or

less in voluntary communication. According to Dia and Zeghal (2008, p. 237), voluntary information

disclosure by management targets the lack of informational symmetry between managers and

shareholders.

Regarding the market’s disciplining mechanism as described in paragraph 3.3, it is argued that banks

that disclose greater amounts of useful risk information should benefit from a reduction in their cost

of capital. After all, the providers of funds will be better positioned to judge the bank’s risk level and

this will remove the need for them to incorporate a risk premium within the cost of capital. On the

other hand, banks that are reluctant to disclose risk information may not experience more expensive

funding but also that these are more difficult to access (Linsley and Shrives, 2005a, p. 206). Engels

(2008) confirms this view and states that higher disclosure quality and stricter disclosure

requirements (regarding the introduction of IFRS 7) could lead to less information asymmetry and

previous literature suggest that this will ultimately lead to a lower cost of capital.

Linsley and Shrives (2005a, p. 206) mention another aspect of risk disclosure. They argue that

increased risk disclosure can help bank supervisors to be more effective in their monitoring as they

are better positioned to foresee potential problems and therefore can act earlier. Furthermore, risk

disclosure has disciplining effects. If banks recognize that they need to disclose risk information, then

an incentive exists for them to improve their risk management capabilities as they will not want to be

4 See also Amran et al. (2009, p. 51).

22

viewed as inferior to other banks in this respect. This rise in risk management abilities across the

banking sector will, in turn, create more stability within the industry, hence reducing systemic risk.

Chen et al. (2007, p. 644) argue that financial transparency and information disclosure are crucial

elements of good corporate governance. Within firms that adopt poor financial transparency and

information disclosure practices, managers are more likely to use their information advantage to

pursue a private benefit of control. Ultimately this will lead to an increase in the agency cost, faced

by shareholders. According to Chen et al. (2007, p. 644), disclosure practices can be viewed as

effective mechanisms for the protection of the rights of outsiders. Therefore, an improved level of

transparency and disclosure practices ultimately leads to better corporate governance (Linsley and

Shrives, 2005, p. 303). Hence, good corporate governance is acknowledged to be a solution to the

agency problem (O’Sullivan, 2000). After all, better transparency and disclosure practices can help

shareholders to gain a better understanding of firms’ management practices and thereby help to

reduce the information asymmetry faced by shareholders.5

It is obvious to conclude that higher amounts of disclosures, whether or not combined with a higher

quality, lead to higher levels of transparency. However, Linsley and Shrives (2005a, p. 206) point out

that only disclosure of ‘useful’ information will create transparency. The Basel Committee (BCBS,

1998, p. 15) states that the fundamentally important qualitative characteristics that will contribute to

transparency are: timeliness, comprehensiveness, reliability, relevance, comparability and

materiality. Furthermore, Morgan (2002) states that high quality disclosures by banks may not

necessarily result in reduced levels of information asymmetry if these banks are not transparent in

the first place. However, Brown and Hillegeist (2007) found evidence that there is a negative

relationship between disclosure quality and information asymmetry.

3.6 Efficient market hypothesis

As said in the previous paragraph, disclosure is critical for the functioning of an efficient capital

market. According to Scott (2006, p. 88), efficiency implies that the information content of disclosure

is valued by the market. Theory about efficient securities market indicates what the primary reason

for the existence of accounting is, namely information asymmetry (see also paragraph 2.2). Market

efficiency does not guarantee that security prices fully reflect real firm value. It does suggest,

however, that prices are unbiased relative to publicly available information and will react quickly to

new or revised information. Deegan and Unerman (2006, p. 210) acknowledge the development of

5 This is also argued by Solomon et al. (2000, p. 450) and Dobler (2008, p. 187).

23

the Efficient Markets Hypothesis (EMH) crucial to the development of the Positive Accounting

Theory.

The EMH is based on the assumption that capital markets react in an efficient and unbiased manner

to publicly available information. The perspective taken is that security prices reflect the information

content of publicly available information, which is not restricted to accounting disclosures (Deegan

and Unerman (2006, p. 210). There are three forms of EMH and each is becoming successively

stronger in implication. As market prices correctly represent all the information contained in the

record of past prices and trading volume, then the weak form of EMH can be applied. If market prices

also rapidly adapt to new information and present public information currently widely known, the

semi-strong form of EMH can be applied. The strong form of EMH can be applied if, besides the

previous two, the market prices also accurately represent predicted future information, which

includes private information (Douglas, 2007, p. 1-2).

Scott (2006, p. 239) suggest the existence of an anomaly regarding the EMH. The efficient securities

market theory predicts no price reaction to accounting policy changes that do not impact underlying

profitability and cash flows. Consequently, if there is no share price reaction, it is unclear why

management and regulators should be particularly concerned about the accounting policies that

firms use. Efficient markets theory implies the importance of full disclosure, including disclosure of

accounting policies. However, once full disclosure of accounting policies is made, the market will

interpret the value of the shares in the light of the policies used and will not be fooled by variations

in accounting numbers that arise solely from differences in accounting policies.

3.7 Risk and risk management

Risk is an inescapable element of any business venture. The concept of risk relates to a distribution of

future outcomes. In a business context, risk can be driven by various external and internal factors or

sources. Risk factors comprise, for example, politics, regulation, market, finance, business process

and personnel. All of them potentially affect an entity’s future performance (Dobler, 2008, p. 187).

Since risk is mostly associated with losses, Amran et al (2009, p. 40) defines risks as conditions that

can adversely affect the conditions of a company and moreover the price of its securities. Hence, it is

in the interest of the shareholder that risk related information is disclosed in a timely manner.

Financial institutions operate within increasingly unpredictable and unstable external environments,

while at the same time being at the forefront of advances in risk management techniques, according

to Linsley and Shrives (2005a, p. 205). Amran et al. (2009, p. 40) acknowledge the increasing

24

complexity of doing business, which has evoked a situation in which risk management has become an

important and integral part of the company’s internal control and governance in order to achieve its

plans and objectives. Amran et al. (2009, p. 40) define risk management as “the methods and

processes used by organizations to manage risks (or seize opportunities) related to the achievement

of their objectives”. Risk management is, according to Solomon et al. (2000, p. 448), essential for the

maximization of shareholders’ wealth as it aims to maximize profitability while at the same time

reducing the probability of financial failure.

Risk management and internal control evidently overlap. The objectives of a company, its internal

organization and the environment in which it operates are continually evolving and, as a result, the

risks it faces are continually changing. A sound system of internal control therefore depends on a

thorough and regular evaluation of the nature and extent of the risks to which the company is

exposed. Indeed, profits are more or less the reward for successful risk-taking in business and the

purpose of internal control is to help manage and control risk appropriately rather than to eliminate

it (Financial Reporting Council, 2005, p. 3).

Lajili and Zeghal (2005) describe a typical risk management framework involving a few processes.

Firstly, there is the identification, measurement and assessment of risk types that a company might

face. Secondly, a response model has to be formulated to tackle risks. The risk bearing capacity, risk

reduction or mitigation procedures need also to be determined. Finally, the implementation of all the

actions planned as proposed by the response model needs to be monitored and checked. Amran et

al. (2009, p. 41) argue that by identifying and proactively addressing risks and opportunities, the

company protects and creates value for their stakeholders. Such risk management has been

mandated to be disclosed by the accounting standard boards in some developed countries. However,

risk management disclosure is still very much voluntary in many parts of the world.

In current risk management practices Value-at-Risk (VaR) is a major tool. VaR refers to the maximum

likely loss from an exposure for a given time horizon and solvency standard and result in information

that can be easily disclosed. However, a main problem with this tool is its scope: the tool applies only

to a small proportion of financial risk: market risk in trading activities. The assumption of a neutral

attitude towards risk, such that large losses with low probability are not penalized in capital

requirements more than small losses with larger probability of occurrence is problematic as

operational risks are underestimated (Sundmacher, 2006, p. 3).

3.8 Banks and their (specific) risks

25

Banks are in the risk business. In the process of providing financial services, they assume various

kinds of financial risks (Santomero, 1997, p. 2). The risks associated with the provision of banking

services differ by the type of service rendered. For the sector as a whole, risks can be classified into a

number of generic types. Based upon the classifications of Santomero (1997, p. 8-10) and Pyle (1997,

p. 3) the following major risks areas can be distinguished: (1) market risk; (2) credit risk; (3) liquidity

risk; (4) operational risk and (5) legal risk. All financial institutions face all these risks to some extent.

The next paragraph will explain these risks, as major sources of value loss, related to banking

activities.

Pyle (1997, p. 3) defines risk in the same way as Amran et al. (see previous paragraph). Risk in the

banking context can be defined as reductions in firm value due to changes in the business

environment. Market risk (1) is the change in net asset value due to changes in underlying economic

factors such as interest rates, exchange rates and equity and commodity prices (Pyle, 1997, p. 3).

Santomero (1997, p. 8) classifies market risk somewhat different under systematic risk. All investors

assume this type of risk, whenever assets owned or claims issued can change in value as a result of

broad economic factors. For the banking sector, Santomero mentions two factors of greatest

concern, namely variations in the general level of interest rates and the relative value of currencies.

Banks try to estimate the impact of these particular systematic risks on performance, attempt to

hedge against them and thus limit the sensitivity to variations in these factors (Santomero, 1997, p.

8).

Credit risk (2) is the change in net asset value due to changes in the perceived ability of

counterparties to meet their contractual obligations (Pyle, 1997, p. 3). It arises from non-

performance by a borrower and may arise from either an inability or an unwillingness to perform in

the pre-committed contract manner. The real risk from credit is the deviation of portfolio

performance from its expected value. As such, credit risk is diversifiable, but difficult to eliminate

completely. In fact, a portion of the default risk may result from the systematic risk outlined above.

Counterparty risk can be seen as a form of credit risk, as it comes from non-performance of a trading

partner. However, Santomero points out that counterparty risk is generally viewed as a more

transient financial risk associated with trading than standard creditor default risk (Santomero, 1997,

p. 9).

Leenaars (2003, p. 341) asserts that liquidity risk (3) has three perspectives: the inability of the bank

to (re)finance against normal costs, the liquidity risk of the market and the liquidity risk of the type of

assets. Santomero (1997, p. 10) describes liquidity risk as the risk of a funding crisis. Such a situation

would inevitably be associated with an unexpected event, such as a large charge off, loss of

26

confidence or a crisis of national proportion. The recent credit crunch is an example of a funding

crisis. Risk management activities need to focus on liquidity facilities and portfolio structure.

Operational risk (4) arises from costs incurred through mistakes made in processing, settling, and

taking or making delivery on trades in exchange for cash. It also arises in record keeping, processing

system failures and compliance with various regulations (Santomero, 1997, p. 10). Operational risk is

clearly a multitude of different risks that, if preferred, separately can be appointed (Leenaars, 2003,

p. 346).

Legal risks (5) are endemic in financial contracting and are separate from legal ramifications of credit,

counterparty and operational risks. New statutes, tax legislation and regulations can put formerly

well-established transaction into contention even when al parties have previously performed

adequately and are fully able to perform in the future. A second type of legal risk arises from inside

the company. Fraud, violation of regulations or laws and other actions by management or employees

can lead to catastrophic loss (Santomero, 1997, p. 10).

3.9 Risk disclosure: development and discussion

Though acknowledging the increase in risk management practices within financial institutions, Linsley

and Shrives (2005a, p. 206) argue that the increase in awareness of risk and risk management did not

(yet) led to an increase in the release of risk-related information. Furthermore, current annual

reports do provide some form of risk disclosure, but not in a comprehensible manner for the

stakeholders to understand (Linsley and Shrives, 2006). In addition, Engels (2008) states that

research on the quality of disclosures of the banking sector suggest that these disclosures, especially

risk disclosures are indistinct. This is also noticed at firms in common: the risk information as it

currently stands is too brief, not sufficiently forward looking and not wholly adequate for decision-

making purposes (Beretta and Bozzolan, 2004; Cabedo and Tirado, 2004, p. 183).

Research around the millennium change carried out by the Basel Committee on Banking Supervision

(2001, 2002, and 2003) observed a general trend that banks are disclosing more information.

However, according to Linsley and Shrives (2005a, p. 210) there is still a need for greater disclosure if

Pillar 3 compliance is to be achieved. A recent paper of KPMG (2008) investigates financial risk

disclosure of financial institutions among the European Union and emphasizes on the importance of

transparency. Companies operating in financial services are often exposed to complex risks that are

very hard for an investor to evaluate. Lack of transparency is therefore widely seen as a major reason

27

for valuation gaps between financial sector companies and other companies. The paper concludes

that risk disclosure can be improved significantly with limited effort. The overall level of disclosure at

most institutions is suffering due to “black holes”. In other words, a comprehensive approach on

disclosing risks is missing. Relevant pieces of risk information are missing in annual reports, according

to KPMG (2008, p. 5). The Basel II Accord has however led to an increase in the level of risk and

capital management related disclosure. The implementation of Basel II via the Capital Requirements

Directive (CRD) has resulted in some banks publishing a separate Pillar 3 document in connection

with the annual report (KPMG, 2008, p. 9). The next chapter will discuss the results of these studies

in more detail.

Opposed to the call for more risk-related information, some pose critical notes. The first issue is that

managers can be reluctant to provide detailed risk and risk management information if they consider

it to be commercially sensitive and therefore of potential value to competitors. Linsley et al. (2006, p.

269) mention this problem as the problem of proprietary information. The second issue concerns the

disclosure of forward-looking risk information. A paper of the Institute of Chartered Accountants in

England and Wales (ICAEW, 1998) states that companies should be providing forward-looking risk

information and not only historical information. Forward-looking information is much more useful for

decision-making. It is however understandable that managers can be reluctant to publish information

of this type, as it is inherently unreliable and could leave managers open to potential claims from

investors who have acted upon this information. Linsley and Shrives (2005a, p. 210) elaborates

further on this issue and point out that there is a tension between relevance and reliability. Forward-

looking risk information is potentially of great relevance, but also inherently unreliable.

3.10 Conclusion

This chapter discusses the theoretical framework and has focused on economic theories that can

explain the occurrence and extent of risk disclosure. The PAT studies managers’ accounting policy

choices and is therefore relevant to this research. The political cost hypothesis is suggested to explain

why firms make (voluntary) risk disclosures. Agency theory is useful in study the process of

contracting. Principal-agent relationships should be transparent as possible, but this transparency is

set back by information asymmetry between the agent and the principal. Providing better levels of

transparency and disclosure is proposed to be able to effectively mitigate information asymmetry

and improve corporate governance. Stakeholder theory provides alternative explanations about

what drives an organization to make particular disclosure. Large companies (which listed banks often

are) must provide more information because they have greater number of stakeholders to whom

they are accountable. Figure 2 outlines the situation as illustrated in this chapter.

28

Figure 2 The risk disclosure framework

Applying these theories to the banking sector, it can be concluded that increased transparency would

strengthen market discipline. Participants in the market place can sanction banks whose risk profile

are considered unsatisfactory but can also reward banks with appropriate risk management.

Furthermore, better levels of risk disclosure should banks enable to benefit from a reduction in their

cost of capital, paying lower risk premiums, as the providers of funds (often other banks) will be

better positioned to judge the bank’s risk level. In recent developments, it can be seen that this

analysis is followed by regulators, since market discipline is a key element in modern regulatory

frameworks in the banking sector. Figure 3 shows the effects of improved (quantity and/or quality of)

risk disclosure as discussed in this chapter.

Figure 3 The effects of improved risk disclosure

The last parts of this chapter covered risk(s), risk management and its development in the banking

sector. Risk management practices have increased and become even more an integral part of the

banks’ business. Surveys in the financial sector show an increase in the release of risk-related

information, but it is still too brief, not sufficiently forward looking and not presented in a

comprehensible manner for the stakeholders to understand. There have been calls for greater and

29

more relevant risk disclosure, but others pose critical notes towards these calls regarding the

problem of proprietary information and the chance of litigation.

30

4 Prior Research

4.1 Introduction

There has been extensive research in the developed countries to examine the corporate disclosure in

financial and non-financial companies. The focus of this chapter lies upon prior research on risk

disclosure in (financial) companies. The following question will be answered in this chapter:

What are the results of prior risk disclosure research?

The second paragraph introduces the research papers that have evoked the research on risk

disclosure by financial institutions. In the third paragraph studies on the extent of risk disclosures are

discussed, on both financial and non-financial companies. The fourth paragraph continues with

studies on the factors that influence risk disclosures. Research on the usefulness of risk disclosure is

discussed in the fifth paragraph. Studies examining the association between the volume of risk

disclosures made by banks within their annual reports and potentially relevant variables are very

scarce, and that is the reason that also studies examining risk disclosure in nonfinancial firms are

reviewed. The literature review in this chapter is used to develop hypotheses in the next chapter.

4.2 Initiation of risk disclosure research

A large number of disclosure studies have been performed in the last 30 years but their primary

focus has been analyses of corporate disclosures in general. Additionally, the majority of these prior

studies have examined the annual reports of non-financial firms. The study of Linsley et al. (2006)

specifically focused on risk and risk management disclosure within the banking sector and had

therefore an exploratory aspect. With their paper, they intend to initiate risk disclosure research in

the financial services arena (Linsley et al., 2006, p. 268). The results of their study will be discussed in

paragraph 4.3.2.

Before academic (empirical) research on the risk disclosure started, two papers were published in the

field of accountancy focusing on risk and risk reporting. In 1998, it was the Institute of Chartered

Accountants in England and Wales (ICAEW) that started a debate on the importance of risk reporting,

publishing a discussion paper entitled “Financial Reporting of Risk”. The paper stated that quoted UK

companies should voluntarily undertake to disclose risk information within a separate statement

contained within the annual report. Furthermore, it concluded that the annual report did not present

31

a coherent discussion of the risks that challenge the company and the action the directors are

undertaking to manage those risks (Linsley and Shrives, 2005a, p. 211-212).

The second paper (Schrand and Elliot, 1998) summarizes presentations and debates on an AAA/FASB

(American Accounting Association / Financial Accounting Standards Board) conference about “Risk

and Financial Reporting”. They noted fundamental inadequacies in regulatory requirements

regarding risk disclosures and concluded that companies have no incentives for voluntary disclosures

about risk. Concurrent to the previous papers, the Basel Committee on Banking Supervision (BCBS,

1998) issued a paper ‘Enhancing Bank Transparency’ related to the disclosure of risk information in

the banking sector. Already mentioned in the first chapter, this paper discusses the critical

importance of disclosure and transparency within the context of banking supervision. It

recommended banks to disclose six categories of information within their annual reports. Two of

these information categories are risk exposures, and risk management strategies and practices

(Linsley et al., 2006, p. 269-270).

4.3 Extent of risk disclosure

This paragraph discusses prior research on the extent of risk disclosures. First, the results of a few

surveys are shown. Second, studies using content analysis are reviewed. After that, studies using

other methods than content analysis are discussed. The paragraph ends with a summarizing table.

4.3.1 Surveys

The Basel Committee on Banking Supervision (BCBS) has published three studies to date (2001, 2002

and 2003) examining bank disclosures in general. The studies adopted identical research methods,

although the number of sampled banks differed slightly between the three years. The survey

instrument comprised a detailed list of 104 questions that the Basel Committee considered useful for

their own disclosure review purposes. These questions are grouped into 12 categories. The intention

of these surveys was to give an impression of the current disclosure practices for comparison with

the disclosure proposals of the Committee. Generally, there has been an increase in disclosure noted

by the bank supervisory authorities. A summary of the results of the three surveys are shown in table

1 (Linsley et al., 2006, p. 271).

32

Table 1 Summary of the surveys of the Basel Committee on Banking Supervision

Disclosure categories Number of questions per category

Percentage of ‘yes’ responses compared with ‘no’ responses

1999 (%) 2000 (%) 2001 (%)1. Capital structure 14 73 78 822. Capital adequacy 7 46 48 553. Market risk and internal modeling 16 64 66 684. Internal and external ratings 4 32 35 465. Credit risk modeling 5 32 33 336. Securitization activities 8 29 36 457. Credit risk 13 55 56 618. Credit derivatives, other credit enhancements

6 24 25 34

9. Derivatives 9 58 56 6210. Geographic and business line diversification 10 65 63 6511. Accounting and presentation policies 7 82 84 8412. Other risks 5 62 74 84Total 104

Although the BCBS studies above examined bank disclosures in general, their conclusion can also be

applied to risk disclosures. The 2003 paper states: “Banks continued to improve significantly the

disclosure on information on liquidity, operational and legal risks as well as interest rate risk in the

banking book. This information has now become as commonly disclosed as the basic information on

market risk or credit risk” (BCBS, 2003, p. 23). The rate of banks disclosing “other risks” (operational

and legal risks, liquidity risk and interest rate risk in the banking book) increased from 65% in 1999 to

84% in 2001. Linsley and Shrives (2005, p. 210) partly attribute this to the considerable increase in

awareness of operational risk matters that has occurred during that time. Although showing an

increase, less widespread were disclosures on credit risk modeling, credit derivatives and other credit

enhancements (from 24% in 1999 to 34% in 2001). Disclosure of information on internal risk models

was also much more common for market risk than for credit risk (68% versus 33% in 2001).

A recent paper of KPMG (2008) shows the result of a survey among financial institutions in Europe

about financial risk disclosure. In the banking area, KPMG (2008, p. 6) assesses six areas of risk

disclosure, namely credit risk (1), market risk (2), ALM (asset and liability management) and treasury

risk (3), operational risk (4), business risk (5) and overall risk related to capital strategy and

shareholder value (6). The findings of the report are that the disclosure of all areas, except business

risk, on average is on a mediocre level.

33

Credit risk (1) is the most developed area of disclosure in the banking sector. However, there still

exists room for improvement and KPMG relates that to the fact that not all surveyed banks did yet

apply the Basel II regulation. Indeed, the results show that financial institutions already applying

Basel II have a better quality in both quantitative and qualitative credit risk disclosure. Market risk (2)

and ALM and treasury risk (3) is generally mediocre. Gaps are found in the disclosure of risk

concentrations and in the disclosure on ALM practice and reporting. Business risk (4) is the least

developed risk disclosure area. About half of the banks provide inadequate disclosure of their

business risks. Many banks do not disclose whether they identify business risks at all and only a few

banks link these risks to their business strategies. Reputational risk, as stated by Basel II, is however

fairly well identified as a type of risk, indicating that regulatory requirements by some means drive

risk disclosure. The average operational risk (5) disclosure is also on a low level. Operational risks are

on average well defined, but the level of definition is often on a general level. Other deficiencies are

related to the disclosure of operational risk management, reporting and mitigation. The last

disclosure category, disclosure of issues related to risk and capital strategy and shareholder value (6)

is in general on a mediocre level. The dispersion of scores between institutions is fairly wide, which

means that all institutions disclose some information on the overall risk and capital management,

while the best ones provide good disclosure throughout the risk area (KPMG, 2008, p. 12-13).

4.3.2 Content analysis

Examining risk disclosures is often carried out by means of content analysis and more specific by

means of a sentence analysis. Content analysis is a method to study the content of communication.

In studies on risk disclosure, generally the written text in (parts of) the annual report is examined.

Studies on non-financial companies are Beretta and Bozzolan (2004), Lajili and Zeghal (2005), Linsley

and Shrives (2005b), Linsley and Shrives (2006), Linsley and Lawrence (2007), Amran et al. (2009) and

Michiels et al. (2009). There is only one study on financial firms using content analysis, by Linsley et

al. (2006). This is the most relevant study and a main fundament for this study and is first discussed.

Linsley et al. (2006)

The study of Linsley et al. (2006) examines risk disclosure practices within annual reports of Canadian

and UK banks. Their paper analyses and classifies the risk information communicated by the sample

banks and discusses the nature of the risk disclosure. A matched pair approach was adopted to

match nine Canadian banks listed within the ranking with UK banks of comparable size. Content

analysis was used to analyze and classify the risk disclosures within the annual reports. A total of

3,323 risk sentences were identified within the sample of annual reports. They were categorized in

credit risk, market risk, interest rate risk, operational risk, capital adequacy risk and risk management

34

policies. Furthermore they were characterized according to form (qualitative or quantitative), period

(past or future) and qualification (bad, neutral or good).

The highest numbers of disclosures fall within the credit risk area, followed by market risk and capital

adequacy risk. Market risk and credit risk are both financial risks, and it is therefore expected for

banks that these risks are dominant in risk disclosures. It differs however from the BCBS 2001 survey,

where four other disclosure categories were ranked more highly than credit risk. More consistent

with the BCBS 2001 survey are the rankings of market risk and capital adequacy risk. Furthermore,

operational risk disclosures are, on a relative basis, disclosed far less than market risk, credit risk and

capital adequacy risk. This result appears to be in contrast with the findings of the BCBS surveys

(2001, 2002 and 2003), where a noticeable increase occurred in the disclosure of operational risk.

The sentence characteristic that occurred most frequently is of the ‘qualitative/neutral/future’ type

(1,156 sentences). The majority of these disclosures consist of explanations of general risk

management policy. Such information does however not provide information about specific risks

that the bank faces and nor does it explain the actions the directors have taken to manage that risk.

Qualitative disclosures are much more disclosed than quantitative disclosures. Furthermore, there

are also more future disclosures than past disclosures. This could imply that there is greater

disclosure of future information. However, if we remove the general risk management policy

statements from the ‘qualitative/neutral/future’ type, this causes a reduction from 59% to 21%

future disclosures. The level of ‘quantitative/future’ type disclosures is very low and, according to the

authors, this is due to the fact that both quantitative and future disclosures are likely to be more

sensitive and therefore subject to higher levels of proprietary cost. Regarding the qualification

characteristic, 56% of the disclosures are neutral, 12% relate to bad news and 32% relate to good

news. Linsley et al. conclude that Canadian and US banks are willing to disclose some bad risk news,

but the difficulty is knowing whether there is further bad news that remains undisclosed. A final

noteworthy finding is that in a number of cases ‘quantitative and qualitative/bad news/past’

disclosure sentences are immediately followed by ‘qualitative/good news/future’ disclosure

sentences. The authors conclude that directors appeared to want to demonstrate that any ‘past/bad’

news had been acted upon and converted into good news.

Overall, the authors conclude that the usefulness of current disclosures is questioned as relatively

little quantitative risk information is disclosed and there is a very strong bias towards disclosing past

rather than future risk-related information.

35

Beretta and Bozzolan (2004)

Pioneers investigating risk disclosure were Beretta and Bozzolan (2004). They developed a framework

for the analysis of risk communication and an index to measure the quality of risk disclosure. Hence,

they applied their framework to a sample of nonfinancial companies listed on the Italian Stock

Exchange, focusing on the MD&A section only. Even though they identified 75 different risk items

being disclosed, they found that companies in general avoid communicating the expected impact in

quantitative terms of these risks and the economic direction of the firms. Furthermore, firms are also

reluctant to indicate whether the future risks disclosed will impact them, either positively or

negatively. They are more inclined to report past and present risks (Amran et al., 2009, p. 43).

Lajili and Zeghal (2005)

The study by Lajili and Zeghal (2005) examined risk information disclosures in the annual reports of

300 Canadian nonfinancial companies listed on the Toronto Stock Exchange. Employing content

analysis, they found a high degree of risk disclosure intensity, both mandatory and voluntary risk

management disclosures. However, the analytical power of such disclosures appeared to lack

uniformity, clarity and quantification, thus potentially limiting their usefulness. Related to the nature

of the disclosed risks, they found that financial risk was the most frequently disclosed risk and

include information relating to foreign currencies. Trailing behind financial risk is credit risk, followed

by market risk, which deals with companies’ reaction to competition. The authors conclude that

more formalized and comprehensive risk disclosures might be desirable in the future to effectively

reduce information asymmetries between management and stakeholders.

Linsley and Shrives (2005b)

In the UK, Linsley and Shrives (2005b, 2006) conducted two risk disclosures studies in the UK (United

Kingdom) context, by which the second paper was built upon the results of the first study. The first

paper examined risk information disclosed by UK public companies within their annual reports. Their

sample consisted of 79 nonfinancial companies listed in the FTSE 100 as on 1 January 2001. Using a

sentence-based approach they identified a total of 6,168 risk sentences. Consistent with the above

study by Lajili and Zeghal, financial risk was a dominant type of disclosure found in the sample, next

to strategic risk, which is followed by integrity risk. Most of the disclosure is qualitative in nature

which is again in line with the results of Lajili and Zeghal. An important conclusion is the fact that

directors appear to be willing to discuss external risks, but are more reluctant to discuss internal

risks. Overall, they found that the companies sampled are not providing a complete picture of the

risks they face. Minimal disclosure of quantified risk information is found and a significant proportion

36

of risk disclosures consist of generalized statements of risk policy. More useful is the released

forward-looking risk information.

Linsley and Shrives (2006)

In their second paper Linsley and Shrives (2006) continued with the results from the content analysis

of the above described risk disclosure research. They deepened their investigation into the

characteristics of the risk-related sentences, and disaggregated risk disclosure into good/bad,

future/past and monetary/non-monetary disclosure. They found very few monetary assessments of

risk information, but companies did exhibit a willingness to disclose forward-looking information.

Furthermore, statements of general risk management policy dominated and a lack of coherence in

the risk narratives was noticed. This implies, according to the authors, that a risk information gap

exists and consequently stakeholders are unable to adequately assess the risk profile of a company.

Linsley and Lawrence (2007)

In 2007, Linsley performed another risk disclosure study in the UK context together with Lawrence.

They used content analysis to identify risk disclosures in the annual reports of 2001 of the 25 largest

nonfinancial companies listed in the FTSE 100. From the 25 sample reports, they identified a total of

2,770 risk-related sentences. Hence, they tested these risk disclosures for readability and

obfuscation. In common, the level of readability of the risk disclosures is difficult or very difficult. The

authors conclude that, if the risk communication has to be improved, there is a need for greater

clarity when describing and discussing risks in the annual report.

Amran et al. (2009)

To date, research on risk disclosure has been undertaken mostly in the western setting. Amran et al.

(2009) provides a study on risk reporting by focusing on the Malaysian experience. Similar to

previous studies, they used content analysis to explore the availability of risk disclosure in a total of

100 listed nonfinancial companies’ annual reports. They identified a total of 2,023 risk-related

sentences, which consisted mainly of sentences about strategic risk (647) and operation risk (613).

Financial risk is a much less frequent disclosed risk, which is opposed to the findings of the studies of

Lajili and Zeghal (2005) and Linsley and Shrives (2005b). Furthermore, the total number of sentences

dedicated for discussion of risk information is very much less when compared to the study done by

Linsley and Shrives (2005b). According to the authors, this is due to the fact that disclosure reporting

by Malaysian companies is still at the infancy stage.

37

Michiels et al. (2009)

The most recent study on risk reporting is done in the Belgian context by Michiels et al. (2009). They

also investigated risk disclosures using content analysis. Their sample consisted of 46 Belgian listed

nonfinancial companies. A total of 4,037 risk-related sentences were derived from the annual reports

and they found financial risk the most frequent disclosed risk, followed by operational risk.

Disclosures on business and legal risk are the least frequent disclosed risks.

4.3.3 Disclosure index

Next to Linsley et al. (2006), two other studies have examined disclosure by financial firms. Helbok

and Wagner (2003) performed an empirical study on operational risk disclosure and Hossain (2008)

investigated the extent of disclosure in common in annual reports of banking companies in India.

Helbok and Wagner (2003)

They performed a comprehensive study on operational risk and focused on the annual reports of the

biggest banks (by total assets) in Europe, Asia and North-America from 1998 to 2001. Their sample

consisted of 59 banks and to analyze the annual reports the authors counted words and pages

related to operational risk. Next a disclosure index was developed to judge about the quality of the

disclosure. Results show that both the extent (quantity) and the depth (quality) of operational risk

disclosure had grown. The total number of banks disclosing operational risk showed an enormous

growth and rose from 41% (1998) to 92% (2001). Furthermore, the amount of disclosure increased

by a factor of four. The findings were in line with the results of the BCBS surveys.

Hossain (2008)

The study of Hossain focuses on the banking sector in India and investigates the extent of disclosure

(both mandatory and voluntary) within the annual report. His sample consisted of 38 listed banks

and he investigated the annual reports of 2002 and 2003. To measure the extent of mandatory and

voluntary disclosure an unweighted disclosure index was used. Mandatory items were derived from

laws and regulation, listing rules and guidelines in the Indian context. Voluntary items were derived

from prior research and papers (guidelines, recommendations) of international financial institutions

and accounting bodies. It is noteworthy that among the identified voluntary items, major risk

information areas are defined. Examples are disclosure about general risk management, exposure to

market, credit and liquidity risk, interest rate risk and currency risk. This implies that main parts of

risk disclosure were not mandatory over the research period.

38

On average, banks published 60% of the total disclosure items of which 88% were mandatory and

25% were voluntary items. The author concluded that Indian banks are very compliant with the rules

regarding mandatory disclosure, but far behind in disclosing voluntary items. However, some

voluntary information, such as corporate social disclosure, corporate governance and risk-related

voluntary information, had been disclosed in the annual reports to an acceptable level, in the

author’s view.

4.3.4 Summarizing table

Table 2 shows a summary of the above described literature on the extent of risk disclosures. The next

paragraph will discuss the literature that relates the extent of risk disclosures to company specific

characteristics.

Table 2 The extent of risk disclosure

Author(s) Object of study Sample(size, country, period)

Methodology Results

Surveys in the financial sectorBCBS (2001)

Banks public disclosures in the annual report.

57 banks;member states;1999

Survey by means of questionnaire (104 questions).

In general, an increase in disclosures is noticed. Disclosure on liquidity, operational and legal risks increased significantly. Less prevalent were disclosures on credit risk modeling, credit derivatives and other credit enhancements.

BCBS (2002)

Banks public disclosures in the annual report.

55 banks;member states;2000

Survey by means of questionnaire (104 questions).

BCBS (2003)

Banks public disclosures in the annual report.

54 banks; member states;2001

Survey by means of questionnaire (104 questions).

KPMG (2008)

Analyze and define components and quality of best practice risk disclosure.

25 banks and14 insurance companies;Europe;2007

Survey framework: The Best Practice Risk Disclosure framework, developed by KPMG.

Disclosure of all risk areas are on an average level, except business risk. Credit risk is the most developed disclosure area. Risk disclosure can be improved significantly with limited effort.

Studies on financial firmsLinsley, Shrives,Crumpton(2006)

Analyze, classify and discuss risk disclosures in annual bank reports.

18 public listed banks;UK and Canada;2001

Matched pair analysis to match UK and Canadian banks. Content analysis to identify risk disclosures.

Credit risk is the most disclosed risk, followed by market risk and capital risk. The usefulness of current disclosures is questioned as quantitative and future risk information is disclosed much less often than qualitative and past information.

Helbok, Wagner (2003)

Examine the extent and quality of operational risk disclosure in annual bank reports.

59 banks; Europe, Asia and North-America; 1998-2001

Counting words and pages and a disclosure index to assess quantity and quality of operational risk disclosure.

Quantity and quality of operational risk disclosure has grown during the period. The number of banks disclosing operational risk doubled during the period.

39

Hossain (2008)

Examining the extent of disclosure in annual bank reports.

38 banks;India;2002-2003

An unweighted disclosure index to assess the extent of disclosure.

Indian banks are compliant to mandatory disclosure, but far behind in voluntary disclosure. Risk-related information is however at an acceptable level.

Studies on nonfinancial firmsBeretta, Bozzolan (2004)

Propose a multidimensional framework for the analysis of risk communication and applying it.

85 nonfinancial companies listed on the Italian Stock Exchange;Italy;2001

Content analysis to analyze voluntary disclosure in the MD&A section.

Companies in general avoid communicating the expected impact of risks in quantitative terms. Firms focus upon disclosing information on past and present risk, rather than future risks. Firms are also reluctant to indicate whether future risks disclosed will impact them, either positively or negatively.

Lajili, Zeghal (2005)

Examine risk information disclosures in the annual reports.

300 nonfinancial companies;Canada;1999

Content analysis to analyze risk information disclosures.

High degree of risk disclosure intensity, but the usefulness is questioned. Financial risk was most disclosed, followed by credit risk and market risk.

Linsley, Shrives (2005b; 2006)

Examine risk disclosures in the annual reports.

79 nonfinancial companies;UK;2000

Content analysis by means of a sentence-based approach to identify risk disclosures.

Strategic risk and financial risk dominant, followed by integrity risk. Most of the disclosure is qualitative in nature. Statements of general risk management policy dominate and companies are not providing a complete picture of the risks.

Linsley, Lawrence (2007)

Examine risk disclosures in the annual reports and test for readability and obfuscation.

25 largest non-financial companies;UK;2000

Content analysis to identify risk disclosures and the Flesch Reading Ease formula to test for readability and obfuscation.

Level of readability of risk disclosures varies from difficult to very difficult. There is a need for greater clarity when risks are described and discussed in the annual report.

Amran, Bin and Hassan (2009)

Explore the availability of risk disclosures in the annual reports.

100 nonfinancial companies;Malaysia;2005

Content analysis by means of a sentence-based approach to identify risk disclosures.

Strategic risk and operation risk are dominant types of disclosed risk, and financial risk much less. Total risk disclosure is very much less compared to Linsley and Shrives (2005b).

Michiels, Vande-maele, Vergauwen (2009)

Examining risk disclosures in the annual reports.

46 nonfinancial companies;Belgium;2006

Content analysis by means of a sentence-based approach to identify risk disclosures.

Financial risk is most disclosed, followed by operational risk. Disclosures on business and legal risks are the least frequent disclosed risks.

4.4 Factors that influence risk disclosure

This paragraph discusses the literature that relates the extent of risk disclosures to company specific

characteristics. First, studies on financial firms (Helbok and Wagner, 2003; Linsley et al., 2006 and

Hossain, 2008) are discussed as these are most relevant for this study. Second, studies on

40

nonfinancial firms are analyzed. Some of these studies are already partly addressed in the previous

paragraph. Finally, prior studies are reviewed to search for additional variables that can have a

relationship with disclosure in general and risk disclosure in specific.

4.4.1 Studies on financial firms

Helbok and Wagner (2003)

Next to investigating the extent and quality of operational risk disclosure by banks, Helbok and

Wagner tried to answer the question if there a relationship exists between bank characteristics and

disclosure. They chose two ratios as bank characteristics, equity ratio (equity / total assets, capturing

leverage) and profitability ratio (net profit / total assets) and used descriptive statistics and a logit

model to test whether the financial situation of the sampled banks could explain variation in

operational risk disclosure. Results show that relatively less well-capitalized (low equity ratio) and

less profitable banks (low profitability ratio) tend to disclose more on operational risk over the 1998

to 2001 period. Less capitalized banks have less of a capital cushion to withstand any large losses

from operational risk and therefore seem to have stronger incentives to convey to the market that

their operational risks are well managed and under control.

Linsley et al. (2006)

The exploratory study of Linsley et al. seeks to examine the association between the volume of risk

disclosures made by banks within their annual reports and potentially relevant variables. Potentially

relevant variables were identified as company size, relative profitability, level of risk and the quantity

of disclosed risk definitions. They chose total assets and market capitalization to measure bank size.

All these variables were expected to have a positive association with the level of risk disclosure. Both

measures of bank size and the quantity of disclosed risk definitions indeed showed a positive

correlation with the number of risk disclosures. However, there did not appear to be an association

between levels of risk disclosure and either bank profitability or the level of risk within the bank.

Hossain (2008)

The study of Hossain also investigates the association between company characteristics and the

extent of disclosure. The scores on the disclosure index were then tested against age, size and

profitability. Furthermore, Hossain tested the scores against the complexity of business (number of

subsidiaries) and assets-in-place, which is the proportion of fixed assets. Hossain argues that a higher

proportion of fixed assets lead to lower agency costs and consequently to lower disclosure. Finally,

board composition (proportion of non-executive directors) and market discipline (proportion of Non-

41

Performing-Assets and level of Capital-Adequacy-ratio) were tested for an association with the extent

of disclosure.

To investigate the effect of each variable on the disclosure level, Hossain set up a (OLS) regression

model. The multiple regression model is significant and explains 41% (the adjusted R2 is 0.411) in the

variation of the disclosure level. Age, complexity of business and assets-in-place were however not

statistically significant (at 10%). Size and profitability were significant at the 1% level. The latter

indicates that more profitable banking companies disclose significantly more financial information

than do less profitable ones. Board composition also turned out to be significant. The proportion of

non-executive directors on the board was positively related with the disclosure level. The NPA and

CAR ratio (as proxy for market discipline) were both significantly negatively related to the level of

disclosure. The reason may be the managements’ conservative motives. In order to comply with

ratios as set by the guidelines, banks may pursue low return investments in the hope that a reduction

in risk may compensate for the lowering of returns. Therefore, banks will unwillingly limit their

voluntary disclosure of information.

4.4.2 Studies on nonfinancial firms

Beretta and Bozzolan (2004)

Both company size and industry variables were tested against the extent and quality of risk

disclosure. Company size was found to be significant, but type of industry was not statistically

significant in explaining the amount of disclosure. However, the authors emphasize on the

importance of the type of industry, because the technological and market constraints exerted by the

competitive, industrial environment on business models significantly influence the risk profile of

companies. Moreover, the types of risks a company faces are strictly related to both the unique

critical-success factors and to the typical business models of an industry.

Lajili and Zeghal (2005)

Next to examining compulsory and voluntary risk reporting for a sample of companies listed on the

Canadian stock exchange, Lajili and Zeghal also tested for relationships between the quantity of

compulsory and voluntary risk reporting and firm-specific characteristics. Performing bivariate tests,

no relationship is found between the quantity of compulsory and voluntary risk reporting and firm

size (total assets), profitability (profits), beta (β), or leverage (debt/equity and debt/total assets).

42

Linsley and Shrives (2005b)

Linsley and Shrives tested for relationships between the number of risk disclosures within the annual

report and the level of risk within the company or the size of the company. A major difficulty is

determining an appropriate measure for a company’s risk level. Therefore, the authors used five

measures although all have limitations and none of them can be justified as superior to any other

measure. These measures were: asset cover, gearing ratio, beta factor, ratio of book value of equity

to market value of equity and QuiScore6. However, none of the risk measures showed a significant

association with the number of risk disclosures. Therefore higher risk companies are not naturally

disclosing more information in an effort to better explain the causes of their risks or how they are

being managed.

Next to the level of risk, Linsley and Shrives tested for the relationship between level of risk

disclosure and the size of the company. This relationship is thought to hold because larger companies

have greater numbers of stakeholders to whom they are accountable and as a consequence they

must provide more information. To measure company size, Linsley and Shrives used market value

and turnover. For both measures there was a statistically significant correlation existing between the

number of risk disclosures and the size variables.

Linsley and Shrives (2006)

In 2006, Linsley and Shrives performed another analysis on the gathered data in their previous study

(2005b). Again they tested for relationships between the level of risk disclosure and both company

size and level of risk within the company. Next to the five measures mentioned in the 2005b

research, the authors added two new measures for capturing the level of company risk: Innovest

EcoValue‘21™ ratings and the BiE (Business in the Community Index of Corporate Environmental

Engagement) index score. Both EcoValue‘21™ ratings and BiE Index scores aim to capture the level of

environmental risk of individual companies. For both measures, the authors find a positive

correlation with the number of risk disclosures. The authors carefully conclude that this implies that

companies with lower levels of environmental risk are disclosing greater amounts of risk information.

Abraham and Cox (2007)

The study by Abraham and Cox analyzes the determinants of narrative risk information in UK FTSE

100 annual reports. Ownership, governance and US listing characteristics are related to the extent of

risk information. They contributed to the research on risk reporting by identifying ownership and

6 QuiScores are developed and maintained by CRIF Decision Solutions Ltd. and are measures of the likelihood of company failure in the year following the date of calculation.

43

governance as determinants to risk reporting. Ownership and governance factors may play a vital

role in firm’s risk reporting because the annual report is prepared by the directors for share owners.

The relationship between risk disclosure and corporate ownership and governance is of interest to

regulators because institutional owners and independent directors are expected to reduce agency

problems and thus lessen the need for regulatory intervention in corporate reporting. Finally,

Abraham and Cox look whether the fact that UK companies that are dual listed in the US also have

influence on the extent of risk information. UK firms with such a listing are required to reconcile

financial statements to US accounting standards and submit this via a 20F disclosure to the SEC. US

reporting standards require additional risk disclosures compared to the UK regulation and therefore

the authors expect that UK firms listed in the US disclose a greater amount of risk disclosure in the

UK annual report.

Their sample consisted of 79 nonfinancial companies in the UK FTSE 100 index and they investigated

the 2002 annual reports. They performed content analysis on three types of narrative risk

information, business risk, financial risk and internal control risk. The results indicated that corporate

risk reporting is negatively related to share ownership by long-term institutions. Therefore the

authors conclude that this important class of institutional investor has investment preferences for

firms with a lower level of risk disclosure. Concerning governance, Abraham and Cox find both the

number of executive and the number of independent directors positively related to the level of

corporate risk reporting. This association was not found with the number of dependent nonexecutive

directors. This notion supports the emphasis on the independency of non-executive directors for

good corporate governance. Finally, the study found that UK firms with a US stock exchange listing

do disclose more risk information within the UK annual report than non-US-listed UK firms.

Hassan (2008)

Hassan performed a study in the United Arab Emirates context and his study seeks to explore the

relationship between UAE corporations’ characteristics and the level of corporate risk disclosure

(CRD). The chosen firm-specific characteristics are size, level of risk, industry type and the corporate

reserve. Results are drawn upon a sample of 41 corporations which was divided in a financial sector

(24 companies) and a nonfinancial sector (17 companies). The level of corporate risk disclosure for

each corporation is captured by a risk disclosure index which is based on accounting standards,

professional requirements, prior literature and the UAE regulatory framework.

The empirical findings show that corporate size is not significantly associated with the level of CRD, in

contrast with results of prior research. The corporate level of risk and corporate industry

44

membership are however significant in explaining the variation of CRD. The latter is consistent with

previous research, but the level of company risk is only scarcely found to be significant for the extent

of risk disclosure. Finally, corporate reserve is not significantly associated with the level of CRD.

Amran et al. (2009)

Besides investigating the extent of risk management disclosure in Malaysian annual reports, Amran

et al. also aim to empirically test the sampled companies’ characteristics against the extent of risk

disclosure. The following firm-specific characteristics were identified: product and geographical

diversification, size, type of industry and company level of risk. Performing multivariate analysis,

Amran et al.’s chosen variables were able to explain quite a lot of the variability in risk management

disclosure (the adjusted R2 was 0.433). None of the diversification variables were found to be

significant, however they were positively correlated, which is consistent with the hypothesis. The size

variable was found significant but leverage was not. This is in line with the results of the previous

studies. Finally, two of the eighth industries (infrastructure and technology industries) were found

significant, which could be, according to the authors, due to the nature of these industries.

Michiels et al. (2009)

In the Belgian context, Michiels et al. investigated also firm-specific determinants related to the

extent of risk disclosures. The identified variables were company size, profitability, level of company

risk, quality of the auditor and the presence of a risk committee and/or manager. Furthermore they

tested against the proportion of the number of non-executive directors in the board of directors.

Finally they test the extent of risk disclosures against the situations of one-tier and two-tier system

regarding the organizational structure.

The variables that they found significantly related to the extent of risk reporting are sales (as proxy of

size), profitability and beta (as proxy of the company risk level). Between size and risk reporting and

beta and risk reporting the relationship was positive, which means that companies with big sales

volumes and a higher systematic risk beta also release more risk-related information. The

relationship between risk reporting and profitability is negative, which is somewhat surprising

compared to results from previous research.

All other variables were found insignificantly related to risk reporting. Although the presence of a risk

committee and/or manager was found positively correlated in the Pearson correlation, this relation

disappeared in the multivariate regression. According to the authors, this means that the univariate

correlation between risk reporting and the presence of a risk manager/committee is driven by

45

company size. The multivariate regression model does however explain 31% (the adjusted R 2 was

0,305) of the variance in the total risk reporting score.

4.4.3 Other disclosure studies

Literature on disclosure has developed into a distinct branch of economic and accounting research.

There has been extensive research in the advanced and developing countries to measure the

corporate disclosure in financial and nonfinancial companies. A few studies will be reviewed to find

some additional variables that could influence disclosure levels.

Wallace and Naser (1995) found type of auditor significant in explaining variation in disclosure

indexes, a result that was not found by Michiels et al. (2009). In addition, scope of business was also

significantly related to the variation in disclosure levels, which means that firms classified as

conglomerates provide more disclosure than firms which are classified as non-conglomerates.

Liquidity ratios and outside shareholders’ interests were less useful in explaining variation in

disclosure indexes.

Jaggi and Low (2000) examined the impact of culture, market forces and legal system on financial

disclosures. Their results indicate that firms from common law countries are associated with higher

financial disclosures compared to firms from code law countries. Cultural values have however an

insignificant impact in common law countries, and mixed results appear in code law countries.

Haniffa and Cooke (2005) investigated the impact of culture and governance on corporate social

reporting in the Malaysian context. The ethnic background of directors and shareholders was used as

a proxy for culture. Corporate governance characteristics include board composition, multiple

directorships and type of shareholders. Their results indicate a significant positive relationship

between social disclosure and boards dominated by Malay directors, chair with multiple

directorships and foreign share ownership. The proportion of non-executive directors was predicted

to be positive, but was found negatively related to social disclosures. Control variables (size, multiple

listing and type of industry) were also significantly related to corporate social disclosure.

4.4.4 Summarizing table

Table 3 shows an overview of the discussed literature in this paragraph (except 4.4.3), about firm-

specific characteristics related to the extent of risk disclosure.

46

Table 3 Firm-specific variables and risk reporting

Author(s) Object of study Sample (size, country, period)

Methodology Results

(Risk) disclosure studies on financial firmsHelbok, Wagner (2003)

Relationship between bank characteristics and operational risk disclosure.

59 banks;Europe, Asia and North America; 1998-2001

Descriptive statistics and a logit model.

Less well-capitalized and less profitable banks tend to disclose more on operational risks.

Linsley, Shrives, Crumpton (2006)

Association between volume of risk disclosures and potentially relevant variables.

18 public listed banks;UK and Canada; 2001

Mann-Whitney U-test and Pearson’s rank correlation.

Bank size and total quantity of disclosed risk definitions were positively correlated with risk disclosure. No association found between levels of risk disclosure and either profitability or level of risk within the bank.

Hossain (2008)

Association between bank characteristics and the extent of disclosure.

38 banks; India2002-2003

Multiple regression model (OLS).

Size, profitability, board composition and market discipline were significantly related, but age, complexity of business and assets-in-place were not.

Risk disclosure studies on nonfinancial firmsBeretta, Bozzolan(2004)

Investigating the size and industry effect regarding the quantity of disclosure.

85 nonfinancial companies listed on the Italian Stock Exchange, Italy; 2001

Regression model for relative quantity, using one size variable and 6 industry variables.

Company size was found to be significant positively related. Type of industry was not significantly related in explaining the amount of disclosure.

Lajili, Zeghal (2005)

Relationships between quantity of risk reporting and firm-specific characteristics.

300 nonfinancial companies; Canada; 1999

Bivariate tests on the selected variables.

No relationships found between quantity of compulsory and voluntary risk reporting and firm size, profit, beta or leverage.

Linsley, Shrives (2005b)

Relationship between number of risk disclosures and firm-specific characteristics.

79 nonfinancial companies;UK;2000

Pearson correlation coefficients with two variables for size and five for level of risk.

Both size measures were statistically significant related with the number of risk disclosures. For all risk measures, no association was found.

Linsley, Shrives (2006)

Relationship between number of risk disclosures and environmental risk.

79 nonfinancial companies;UK;2000

Pearson correlation coefficients with two measures of environmental risk

Both measures show a positive correlation with the number of risk disclosures, implying that companies with lower levels of environmental risk are disclosing more risk information.

Abraham, Cox (2007)

Analyzing determinants of narrative risk information in the annual report.

79 nonfinancial companies;UK;2000

Regression model with defined variables regarding ownership, governance and US listing characteristics.

Risk reporting is negatively related to share ownership by long-term institutions. Number of executive and number of independent directors is positively related with risk reporting. Furthermore, US listed UK firms tend to disclose more.

47

Hassan (2008)

Explore relationship between corporate characteristics and the level of corporate risk disclosure.

24 financial and 17 nonfinancial companies;UAE;2005

Regression model with the following variables: size, risk level, industry type and level of reserves.

Corporate size and reserve are not significantly associated with the level of CRD. The level of risk and industry membership are however significant in explaining the variation of CRD.

Amran, Bin and Hassan (2009)

Empirically test the sampled companies’ characteristics against the extent of risk disclosure.

100 nonfinancial companies; Malaysia;2005

Regression model (multivariate analysis) with the following variables: size, type of industry, company risk level and diversification.

Company size was found significantly but leverage was not. None of the diversification variables were found to be significant. Two out of eight industries were, due to their nature, significant.

Michiels, Vande-maele, Vergauwen (2009)

Firm-specific determinants related to the extent of risk disclosures.

46 nonfinancial companies; Belgium;2006

Regression model and Pearson correlation, with size, profitability, risk level, quality of the auditor and governance.

Sales (size), profitability and beta (risk level) were found significantly related to the extent of risk reporting. All other variables were found insignificantly.

4.5 Usefulness of risk disclosure

Besides investigating the extent and drivers of risk reporting, other studies look specifically to the

usefulness of risk disclosure. Examples of such studies are Linsmeir et al. (2002) and Venkatachalam

(1996). These studies investigate the relationship between risk disclosure and the interest rates,

foreign currency exchange rates and commodity prices. The study by Linsmeir et al. (2002) provides

strong evidence of the usage of risk disclosure by investors. The researchers found that the

disclosure of information about market risks is useful in taking investments decisions as it reduces

the level of investor uncertainty.

4.6 Conclusion

The results of prior research on risk disclosure in both financial and nonfinancial firms have been

discussed in this chapter. Two streams of literature are important for this research: literature on the

extent of risk disclosure and literature on the firm-specific characteristics that can explain the

(variation in) the extent of risk disclosure. The most relevant studies within these streams are

summarized in table 2 and 3.

48

5 Research Design

5.1 Introduction

In this chapter the research design is discussed, to begin with the sample selection in the next

paragraph. The research conducted in this study consists of two steps. First, risk disclosures in annual

reports are examined by means of content analysis to capture the extent and characteristics of

information about risk disclosed in the annual report. This method, its methodology and the reasons

why this method is adopted are therefore discussed in the third paragraph. Second, the gathered

data in the first step is examined by means of regression analysis, to test for relationships between

the extent of risk disclosure and potential explanatory variables. Therefore the fourth paragraph

develops hypotheses about these relationships, the fifth paragraph discusses the methodology of

regression analysis and presents the regression model. The last paragraph discusses the

measurement of the independent variables.

5.2 Sample selection

To select the banks to be examined in this research, the BankScope Financials database is used

(accessible via the Wharton Research database). Removing non-EU, non-listed and non-consolidated

bank (reports), a population of 217 banks remained. BankScope also recognizes different types of

specializations for each bank, for example commercial banks or investment banks. Different types of

banks can differ in their risk profile and in order to create a homogeneous sample, only three types

of specializations are considered: commercial banks (111), cooperative banks (25) and bank holdings

and holding companies (31). Other specializations were not taken into consideration to avoid that

they would affect the results (such as Mortgage banks, Investment banks, Islamic Banks,

Governmental Credit Institutions and Securities Firms). After that, twelve local Crédit Agricole

subsidiaries were removed from the sample, as they are subsidiaries of the French national Crédit

Agricole. Using the Ernst & Young sampling assistant tool in Excel, a number of 40 banks have been

selected from the number of remaining banks (155). The list of the final sample selection is showed

in appendix C. The EU area is chosen because of its harmonized accounting standards system (IFRSs).

The year 2007 is chosen because in that year the credit crunch occurred, resulting in turbulence on

the financial markets and volatility on the stock market, and thus may coincide with listed firms

placing greater value on risk disclosure. Furthermore, the increased regulation of IFRS 7 and Basel II

forces banks to disclose more risk information. All annual reports were with a year-end date of 31

December 2007, which ensures comparability of the annual reports that formed the basis for the

content analysis.

49

5.3 Content analysis

Accounting researchers have increasingly focused their efforts on investigating disclosure, in

particular the determinants of disclosure and the capital market consequences of disclosure.

According to Beattie et al. (2004, p. 2), two principal ways of measuring disclosure have been

employed. The first approach has been to use subjective analyst disclosure rankings. This approach is

not without conceptual problems, the practical problem is availability. Rankings are scarce and often

only available in the United States. The second approach has been to analyze the content of

narratives with either researcher-constructed disclosure indices or content analysis (see for a

discussion of these methods Beattie et al., 2004, p. 4-9).

Beattie et al. (2004, p. 10) argue that the extant approaches to the analysis of accounting narratives

in annual reports suffer from two fundamental limitations. First, they are essentially one-

dimensional, whereas disclosure is a complex, multi-faceted concept. To overcome this limitation,

this study captures multiple dimensions of risk disclosure. The first dimension to be considered is

topic, which in this study consist of the major risk areas. Hence, specific risk information is being

nested within the defined risk categories. In addition to this, three type attributes are defined, based

upon the following dichotomous descriptors: past/future, quantitative/qualitative and

good/neutral/bad.

The second limitation is that many extant approaches are partial, either because they examine only

selected sections of the annual report or because they focus on particular issues or pre-selected

items. In my view, this is not a limitation of the approaches but a limitation of the respective studies,

caused by the decisions of the researchers. This study is about risk disclosure, so obviously only

attention is paid to risk information. Previous literature (see chapter 4) shows that risk information

can be find throughout the entire annual report, for example in the notes but also in the

management report. Mandatory accounting information is usually disclosed in the notes to the

financial statements, but IFRS 7 leaves companies free to choose where to disclose the related risk

information. Hence, this study has chosen to analyze the entire content of the annual reports in

order to avoid biased results from a too narrow focus. Furthermore, the annual report has been

defined as the ‘main disclosure vehicle’, concluding that it is the most comprehensive financial report

available to the public (Beretta and Bozzolan, 2004, p. 276). In addition, Lang and Lundholm (1993)

showed that the disclosure level in annual reports is positively correlated with the amount of

corporate disclosure communicated to the market and stakeholders using other media.

50

Content analysis is a means of categorizing items of text. For valid inferences to be drawn, it is

important that the classification procedure is reliable and valid. Reliable means that different people

would code the text in the same way, and valid means that the variables generated from the

classification procedure represent what the researcher intended it to represent. Three types of

reliability can be identified: stability (the extent to which the same coder is consistent over time

when coding the same content); reproducibility or inter-coder reliability (the extent to which

different coders produce the same results when coding the same content); and accuracy (the extent

to which the classification of texts corresponds to a standard or norm) (Krippendorf, 1980, p. 130-

132). To improve reliability of content analysis, previous studies usually use more than one coder for

at least a part of the sample.

Weber (1985, p. 23-24) provides useful discussions regarding how to develop and test a coding

scheme. The basic steps are:

1. define the recording unit (e.g., word, sentence, theme);

2. define the categories (several key decisions);

3. test coding of a sample of text;

4. assess reliability;

5. revise coding rules (e.g., develop disambiguation rules);

6. repeat steps 3-5 until reliability is satisfactory;

7. code all text; and

8. assess achieved reliability.

Following step one, this research chose to measure the extent of disclosure of risk information by

counting risk and risk management sentences rather than words or pages. This is in line with usual

practice in previous disclosure studies using content analysis (e.g. Linsley et al., 2009 and others).

According to Milne and Adler (1999, p. 4), this unit of analysis is deemed more reliable as a coding

method than other units of analysis (e.g. words, paragraphs, pages). They criticize the use of words

since, by themselves, words do not convey any meaning unless referred to the sentences for their

proper contexts. Moreover, it is difficult to decide which words are considered as risk disclosure. Step

two is a step which involves several key choices to be made. The separate risk categories have to be

defined and the attributes that will be captured.

Prior literature is used to breakdown risks into relevant categories and to create an appropriate

classification scheme. The study of Linsley and Shrives (2005b, p. 298, on nonfinancial companies)

distinguished financial risk, operations risk, empowerment risk, information processing and

51

technology risk, integrity risk and strategic risk. The study of Linsley et al. (2006, p. 273, on banks)

split up financial risk into credit risk, market risk and interest rate risk, apparently because of the

significance of these risks to financial firms. Next to these risks, operational risk, capital adequacy risk

and a special category for disclosure about risk management policies were distinguished. Using this

classification and other schemes of Amran et al. (2009, p. 53-54) and Michiels et al. (2009, p. 4), a

new classification scheme is developed that best reflects the relevant risk categories. Furthermore,

risk categories from theory (paragraph 3.8) and in the BCBS studies and a KPMG paper (paragraph

4.3.1) are considered. In short, starting from the categorization of Linsley et al. (2006), interest rate

risk is categorized under market risk. Furthermore, the categories liquidity risk, business risk and

legal risk are added. The risk classification model is presented in appendix D.

Along with the categorization of the risks, certain attributes of the risk-related sentences are

determined. Linsley et al. (2005b, p. 296) noted the following attributes:

(1) whether the risk sentence provided qualitative or quantitative information;

(2) whether good news, bad news or neutral news was being communicated; and

(3) whether the information related to the future or the past.

These attributes gives greater insight in the characteristics and are therefore adopted in this study.

Together with the risk categories and the above defined attributes, a coding grid is created, which

can be found in appendix E.

Step one and step two are important to carry out properly in order to acquire reliable results.

Therefore coding or decision rules need to be developed, in order to achieve a consistent coding

process and to avoid ambiguous and biased results. Step five was used to revise the decision rules.

Initially adopted the same decision rules as used by Linsley et al. (2006), but the author refined and

updated these as they were found too inadequate in their description. The set of decision rules can

be found in appendix F. In short, sentences are coded as risk disclosures if it is considered that the

reader is better informed about risks that have already had impact upon the company, or may in the

future have an impact upon the company, or if the reader is better informed about risk management

within the company. The word “risk” did not have to appear within any given sentence for it to be

identified as a risk disclosure sentence. If a sentence had more than one possible classification, the

information was classified into the category most emphasized within the sentence.

As the author is the sole writer on this study, the reliability of the study cannot be improved by using

more than one coder. However, the author did the following to improve the reliability of the coding

process and to confirm the understanding of the decision rules. Before the coding of the sample, the

52

2001 annual reports from Northern Rock plc and LloydsTSB were coded. These annual reports were

included in the study of Linsley et al. (2006) and they coded respectively 77 and 181 risk disclosure

sentences. The author found respectively 81 and 189 risk disclosure sentences, resulting in a <5%

deviation from the results of Linsley et al. (2006), supporting the notion that the author has mastered

the coding process sufficiently.

Using the developed coding grid and decision rules, the author read each annual report and

identified all sentences providing risk or risk management information and coded each individual

risk-related sentence on this coding grid. The results can be found in paragraph 6.2. Appendix G

provides examples of risk sentences within the annual report and their categorization.

5.4 Hypotheses development

In this paragraph hypotheses are developed using prior research described in the previous chapter.

Firm-specific characteristics that are expected to influence risk disclosure levels are identified and a

prediction about the sign of the relationship is given.

5.4.1 Size

Many studies have managed to prove an association between company size and risk disclosure level

(Beretta and Bozzolan, 2004; Linsley and Shrives, 2005b, Linsley et al., 2006; Abraham and Cox, 2007;

Hossain, 2008; Amran et al., 2009 and Michiels et al., 2009). Lajili and Zeghal (2005) and Hassan

(2008) reported however a non-significant relationship between the two variables in their study. Size

is furthermore included in almost every disclosure study, either as a variable of interest or as a

control variable. As banks become larger, the number of stakeholders involved will also increase. To

meet the information needs of a larger group of stakeholders the bank’s burden of disclosure

becomes heavier. Therefore, the following hypothesis is developed:

H1: Ceteris paribus, there is a positive relationship between the size of the bank and the

extent of risk disclosure.

5.4.2 Profitability

Prior (risk) disclosure studies shows mixed results when testing for an association between the

company’s profitability and its disclosure level. No association is found by Linsley et al. (2006) and

Lajili and Zeghal (2005). Hossain (2008) found a positive association within banks but Helbok and

Wagner (2003) found a negative association. This is also found by Michiels et al. (2009) within

nonfinancial companies. According to Linsley et al. (2006) it could be argued that those banks that

53

are better at risk management will have higher levels of (relative) profitability and that they want to

signal their superior risk management abilities to the market place via disclosures in the annual

report. This potential link will be tested for in this study and therefore the second hypothesis is:

H2: Ceteris paribus, there is a positive relationship between the relative profitability of the

bank and the extent of risk disclosure.

5.4.3 Level of risk

Banks with higher levels of risk have a greater incentive to demonstrate that they are actively

monitoring and managing those risks and to ensure they are not penalized excessively by the market.

This can also be argued based upon stakeholder theory: a bank is expected to undertake more risk

disclosure in order to provide justification and explanation for what is happening in the firm.

Consequently, it may be argued that those banks with higher risk levels will disclose more risk

information in comparison to those with lower risk levels (Linsley et al., 2006, p. 274). Previous

studies did found a positive association between the level of risk and the extent of risk reporting

(Abraham and Cox, 2007; Hassan, 2008 and Michiels et al., 2009). The third hypothesis is therefore:

H3: Ceteris paribus, there is a positive relationship between the bank’s level of risk and the

extent of risk disclosure.

5.4.4 Multiple listing

Abraham and Cox found that UK listed companies with a US stock exchange listing do disclose more

risk information within the UK annual report than non-US-listed UK firms. According to them, this

supports two views. First, from an agency perspective it supports that greater disclosure is a method

by which managers can demonstrate to shareholders that they are acting optimally. Second, it

supports a transaction cost argument that firms have little reason not to disclose, for the additional

disclosure comes at zero marginal cost. Another study, by Haniffa and Cooke (2005) found a positive

relationship between the extent of corporate social disclosure and multiple listing as well. Banks that

are multiple listed could encounter a greater need for disclosure as their securities are distributed via

a more diverse network of exchanges. Consequently they have to deal with additional regulators (for

example the Securities and Exchange Commission in the US) and also with a larger group of

stakeholders (see also the first hypothesis). The fourth hypothesis is therefore:

H4: Ceteris paribus, the extent of risk disclosure is greater for banks with multiple listing

statuses.

54

5.4.5 Credit ratings

Credit rating agencies are firms that assess the likelihood that issuers of debt will make full and

timely payment of their obligation. The major agencies are Moody’s, Standard and Poor’s and Fitch.

These ratings estimate the credit worthiness of an individual, corporation or a country and are

typically assigned on an alpha or numeric basis (e.g. 1-5 or A-E) (Stern and Feldman, 2004). The

ratings are often interpreted as discrete indicators of quality. Prior research did not yet link credit

ratings to the extent of risk disclosure and therefore this hypothesis is based upon own postulations.

Higher credit ratings indicate that banks are more trustworthy than banks with lower credit ratings.

Moreover, higher credit ratings indicate that banks have a better risk management as they are

assumed to be safer to invest in than banks with lower credit ratings. To show their superior risk

management abilities and to maintain or gain additional market confidence, banks with higher credit

ratings are expected to release more risk disclosure. Therefore the fifth hypothesis is as follows:

H5: Ceteris paribus, the extent of risk disclosure is greater for banks with higher credit ratings.

5.4.6 Organizational structure

The next two hypotheses relate corporate governance to the extent of risk reporting. As said in

paragraph 2.7, due to the variety in corporate governance structures in the sample area (EU), it is

difficult to relate corporate governance variables to the extent of risk disclosure. Although

acknowledging this, the BCBS 1999 paper (‘Enhancing corporate governance for banking

organizations’) describes four important forms of oversight that should be included in the

organizational structure of any bank in order to ensure the appropriate checks and balances. The first

one is “(1) oversight by the board of directors or supervisory board (…)” (BCBS, 1999, II.11). Regarding

their organizational structure, companies can have a one-tier structure (board of directors with both

executive and non-executive directors) or a two-tier structure (board of directors with only executive

directors and an additional supervisory board with non-executive directors). The BCBS did not give an

indication about which structure should be preferred regarding corporate governance, but prior

research does. Fama and Jensen (1983) argue that in a company with a one-tier structure, there is no

separation between decision authorization and decision control. Agency theory further proposes that

within a one-tier structure the supervisory function of the board of directors is reduced, as argued by

Michiels et al (2009, p. 6). Therefore, it could be argued that a two-tier structure leads to stronger

corporate governance and hence, to more risk disclosure. The sixth hypothesis is formulated as

follows:

H6: Ceteris paribus, the extent of risk disclosure is greater for banks with a two-tier structure.

55

5.4.7 Board composition

Alongside the emphasis on the supervisory function within a firm, the BCBS paper underlines that

those individuals keeping oversight should not be involved in the day-to-day running of the various

business areas (1999, II.11). As the BCBS relates this aspect of board composition to strong corporate

governance, prior research has linked strong corporate governance to improved risk disclosure. Both

Abraham and Cox (2007) and Michiels et al. (2009) tested for board composition related to the

extent of risk disclosure. Since the approval of the annual report is done by the board of directors,

the composition of this board could influence the extent of risk reporting. Following the BCBS (1999)

notion, Abraham and Cox (2007, p. 231) hypothesize that the proportion of non-executive directors

in the board is positively related to risk disclosure. Executive board directors, as fulltime employees

of the firm, are corporate insiders. By working alongside managers, it may prove difficult for

executive directors to also monitor managers’ actions (Fama and Jensen, 1983). Agency theory

therefore proposes that executive directors have insufficient incentives regarding disclosure,

including risk disclosure. Michiels et al. (2009) state that due to the presence of non-executive

directors agency costs are likely to fall and this could be reflected in more extensive reporting.

Abraham and Cox (2007) did indeed find evidence for their reasoning, but Michiels et al. (2009) did

not. This hypothesis is only applicable for banks with a one-tier structure, having one board of

directors with both executive and non-executive directors. The seventh hypothesis is formulated as

follows:

H7: Ceteris paribus, there is a positive relationship between the proportion of non-executive

board directors and the extent of risk disclosure.

5.5 Regression analysis

This study uses multiple regressions in assessing the variability of the extent of risk disclosure. This

statistical method has been widely used in previous research (Haniffa and Cooke, 2005; Amran et al.,

2009 and Michiels et al., 2009). The regression model that is adopted to test the developed

hypotheses is the following:

RDSCORE=α 0+β1∗SIZE+β2∗PROF+β3∗LEV+β4∗ML+ β5∗CR+β6∗TIER+β7∗NONEXEC+ε

56

RDSCORE = number of sentences related to risks

SIZE = bank size in total assets (in millions US$)

PROF = bank’s profitability in return on (average) assets

LEV = bank’s level of risk (leverage as proxy)

ML = bank’s multiple listing (1 if a bank is multiple listed, 0 if not)

CR = bank’s long-term credit rating (scale variable)

TIER = bank’s organizational structure (1 if it is two-tier, 0 if it is one-tier)

NONEXEC = proportion of non-executive board directors divided by total board directors

In executing the regression analysis, the model needs to be slightly adapted. The reason is that the

variable TIER and the variable NONEXEC are dependent upon each other (NONEXEC has only a value

when the TIER value amounts 0). This problem is solved by first execute the regression analysis

without adding NONEXEC, but testing for all other variables. After that, to be able to test the variable

NONEXEC, the variable TIER is removed and the variable NONEXEC is added.

Before executing regression analysis, attention needs to be paid to the structure of the data. Cooke

(1998, p. 209). The problem of multicollinearity has to be determined. Transformation of the data is

possible to overcome non-normality. Another option recommended by Cooke (1998, p. 209) is using

Rank Regression. The data analysis in the next chapter deals with these issues.

5.6 Independent variable definition

The following independent variables need to be measured: size, profitability, level of risk, multiple

listing, organizational structure and board composition.

Size

Previous disclosure studies have measured size in different ways including turnover, total assets,

employee numbers and market capitalization. According to Hackston and Milne (1996) there is no

theoretical reason to favor one measure over another. Furthermore, prior research shows that

turnover, total assets and market capitalization are all found significantly related to the extent of risk

reporting. However, Linsley et al. (2006) argue that turnover is an inappropriate measure for the size

of banks as their profits do not derive from sales in the same way that the profits of, for example, a

manufacturing company derive from sales. Furthermore, in the author’s opinion, the accuracy of

market capitalization as a proxy for size may be reduced by turbulent market movements. Therefore

this study uses total assets to measure bank size. BankScope Financials database is the source of this

variable.

57

Profitability

A common proxy to measure relative profitability is the return on assets ratio. This proxy has been

used in all previous risk disclosure studies that examined the association of relative profitability with

the extent of risk reporting. BankScope Financials database contains the variable ‘return on average

assets’ and this variable is used.

Level of risk

Measuring the level of risk within a bank is problematical as there is no single measure that truly

encapsulates a bank’s risk level. Beta factor (from the CAPM model) and leverage (the ratio of total

debt to total assets) are the most used indicators of the level of company risk. Furthermore, asset

cover, gearing ratio, ratio of book value of equity to market value of equity and QuiScore are used

(Linsley and Shrives, 2005b). BankScope Financials database contains the following ratio describing

leverage: total equity divided by total assets. A low percentage indicates a high risk and vice versa.

Multiple listing

Manually the sampled banks will be checked whether they are multiple listed on other stock

exchanges within the European Union or elsewhere. For every bank this information was found

within the annual report.

Credit ratings

For the majority of the sample the credit ratings during 2007 could be gathered from the annual

report. In some cases, company websites (with the aid of Google) provided the information. The

credit ratings were found to be AA (highest) or BB (lowest) or in between. The rankings were

converted to a numeric scale. To match credit ratings of Standard & Poor’s with comparable ratings

from Moody’s and Fitch, an information table called “Long-term Rating Scales Comparison” provided

by the Bank of International Settlements was used (provided on their website).

Organizational structure

Based upon the annual report, banks will be checked if they have a two-tier structure (both board of

directors and supervisory board) or a one-tier structure (only a board of directors with both

executive and non-executive directors). This information was gathered manually from the annual

report or other documents, for example registration documents on company websites.

58

Board composition

The number of non-executive board directors compared to the executive board directors. Again,

annual reports, registration documents, press releases and other internet sources were used to find

this data.

59

6 Analysis and Interpretation of the Results

6.1 Introduction

This chapter analyzes and interprets the results of the empirical research. The following questions

are answered in this chapter:

What are the risk disclosure practices of European banks in 2007?

What firm-specific characteristics are related to the extent of risk disclosures?

6.2 Descriptive statistics of the sample

The sample consisted of 40 public listed banks from the European Union Member States. The

European Union consists of 27 countries, of which 21 are represent in the sample selection. The

countries not represent in the sample are: Luxembourg, Finland, Romania, Hungary, Estonia and

Latvia. Italy represents most banks (6), before United Kingdom, Germany, France, Denmark and

Greece (all 3). The biggest bank (measured in total assets) is Deutsche Bank AG from Germany and

the smallest bank is FIMBank Plc from Malta. For more information, see the list in Appendix C.

6.3 Part I: Risk disclosure practices

6.3.1 Risk disclosure score

A total of 10,329 risk sentences were identified within the sample of annual reports. Table 4 shows

the coding grid containing the risk sentence disclosures for the total sample of banks. Barclays plc

(United Kingdom) had the highest risk disclosure score of 753 sentences, where Bank of Attica SA

(Greece) had the lowest risk disclosure score, 40 sentences. Appendix H shows the risk disclosure

score for the individual banks of the sample and appendix G provides examples of risk sentences and

their categorization.

60

Table 4 Number of risk sentence disclosures for the sample of banks

Market risk

Credit risk

Liquidity risk

Oper-ational risk

Business risk

Legal risk

Capital struc..1

Risk mgt..2

Total

Text disclosure sentence characteristics

1 2 3 4 5 6 7 8

Quantitative Good Future A 29 4 0 0 2 0 0 0 35Quantitative Bad Future B 24 5 0 0 3 6 0 0 38Quantitative Neutral Future C 18 8 5 6 4 2 11 1 55Qualitative Good Future D 23 39 5 1 84 1 2 13 168Qualitative Bad Future E 18 22 3 0 56 3 2 0 104Qualitative Neutral Future F 1651 2030 702 597 184 203 333 875 6575Quantitative Good Past G 53 59 9 1 5 0 2 1 130Quantitative Bad Past H 48 63 9 24 34 1 4 0 183Quantitative Neutral Past I 546 781 177 63 28 4 186 15 1800Qualitative Good Past J 17 51 29 6 99 4 15 24 245Qualitative Bad Past K 24 25 7 8 65 18 6 0 153Qualitative Neutral Past L 174 183 80 98 71 40 99 98 843

Total 2625 3270 1026 804 635 282 660 1027 103291 Capital structure and adequacy risk; 2 Risk management frameworks and policies

The average risk disclosure score of the sampled banks is 258 risk sentences. Linsley et al. (2006)

found in 2001 annual reports of 18 English/Canadian banks on average 185 risk sentences. This

shows that the 2007 annual reports investigated in this study show a significant increase in risk

reporting measured in risk sentences. This can be explained by the increased demands of the

regulatory environment, the introduction of IFRS 7 and Basel II, and furthermore supports the notion

that there is an increase in the awareness of bank directors to disclose proper risk disclosures.

Other studies measuring risk disclosures sentences (see paragraph 4.3.2) varied from an average of

20 (Amran et al., 2009, Malaysian companies) to 111 (Linsley and Lawrence, 2007). These studies

were on non-financial companies, which explain the lower averages, compared to financial

companies. Banks are in the business of risk and are therefore expected to release more risk-related

information to the capital market.

61

6.3.2 Risk categories

Market risk Credit risk Liquidity risk Operational risk

Business risk Legal risk Capital structure and ade-

quacy risk

Risk mgt frameworks and policies

0

500

1000

1500

2000

2500

3000

3500

2625

3270

1026

804635

282

660

1027

Num

ber o

f disc

losu

res

Figure 4 Number of disclosures in the different risk categories

Figure 4 reports the defined risk categories into which the risk disclosures were coded. Credit risk is

the most disclosed risk category, which is a consistent finding compared to other studies. Linsley et

al. (2006) found credit risk to be the dominant risk category, amounting about 37% of total risk

disclosure, compared to 32% of this study. Banks are essentially lending institutions and therefore

this is a significant risk requiring appropriate disclosures. The KPMG survey in 2008 confirms that

credit risk is the most developed area of disclosure. In comparison with the BCBS 2001 survey, credit

risk is not found to be the dominant disclosure area. This can however be explained by the fact that

the BCBS 2001 survey focuses on disclosure in common, not specifically on risk disclosure.

Credit risk is followed by market risk and together they dominate the risk disclosures in this study

(for about 57%). This result is in line with Linsley et al. (2006) and KPMG (2008), where market risk is

the second most disclosed risk area. Interest rate risk in this study is categorized under market risk,

where Linsley et al. (2006) distinguished these risks from each other. However, if their respective

scores are combined, the rankings do not change. In studies on non-financial firms, usually financial

risk is the most disclosed area (Lajili and Zeghal, 2005; Linsley and Shrives, 2005b; Michiels et al.,

62

2009). Once again, this is consistent with these results, since financial risk in these studies includes

credit and market risk.

The remaining risks are far less disclosed than credit and market risk. Basel II requires significant

disclosures regarding operational risk (next to credit and market risks). Consequently, KPMG (2008)

defined the regulatory benchmark of operational risk relatively high, but found the average of

operational risk disclosure to be surprisingly low. This study also finds a low level of operational risk

disclosure, as it is only the fifth largest risk disclosure category. For example, there is more

information disclosed on liquidity risk. An explanation could be that banks are giving additional

attention to liquidity risk (management), to ensure readers that the effects of the credit crunch and

the resulting instability of the financial markets are under control and effectively mitigated. General

statements about risk management frameworks and policies amount to about 10% of the risk

disclosure, followed by risk disclosure regarding capital structure and adequacy risk. This type of risk

disclosure contains mainly information about capital ratios (and capital management), in essence an

indicator of the strength of a bank to deal with deteriorating economic circumstances. Compared to

the KPMG (2008) results, which reveal that risks related to capital strategy are on average on a

mediocre level, these results seem consistent.

The category business risk is absent in the study of Linsley et al. (2006), but generally acknowledged

as a separate risk area in the other risk disclosure studies. This type of risk disclosure contains

information about risks related to the macro-economic environment and strategic decisions. In this

study, information on the credit crunch and turbulence on the financial markets (in the second half

of 2007) mainly contributed to this risk disclosure category. The author found that almost all sampled

banks disclosed information about the credit crunch and some provided thorough analyses and

explanations what exactly occurred. However, these risk disclosures lacked detailed information

about what consequences these market circumstances would have upon the company. Instead,

readers were assured about proper risk management practices which were fully able to deal with

those risks. KPMG (2008) finds business risk to be the weakest disclosure area, as many banks do not

disclose whether they identify business risks at all, and only a few banks link these risks to their

business strategies. This study also finds banks to be reluctant to identify and distinguish business

risk and therefore confirm the KPMG findings. The least frequent disclosed risk is legal risk. One other

study, identifying legal risk as a separate risk disclosure area (Michiels et al., 2009) found the same

results.

63

6.3.3 Risk characteristics

Next to the categorization of risks, also certain characteristics were investigated. Table 5 provides the

characteristics of risk disclosures, their total numbers and proportion.

Table 5 Summary of characteristics of risk disclosures

Characteristic Total number of disclosures

Proportion (%)

Qualitative disclosures 8,088 78.3Quantitative disclosures 2,241 21.7 Future disclosures 6,975 67.5Past disclosures 3,354 32.5

Good news disclosures 578 5.6Neutral news disclosures 9,273 89.8Bad news disclosures 478 4.6

The split-up between qualitative and quantitative information shows that approximately 78% of the

defined disclosures were qualitative and only 22% quantitative. Linsley et al. (2006) found this ratio

to be 67% versus 33%. Quantitative risk information is recognized to be more useful than qualitative

information, since the disclosure of the potential size of a risk puts the reader in a better position to

understand its significance. The majority of the quantified risk information is from the credit and

market risk category, the most regulated disclosure areas. It is therefore a surprise that the

percentage of quantitative information dropped compared to the Linsley et al. (2006) study, since

the average risk disclosure significantly has increased and the introduction of IFRS 7 and Basel II

require specifically quantitative risk disclosure. A possible explanation is that the number of

qualitative disclosures increased faster than the number of quantitative disclosures. This would result

in a relatively lower ratio even if the absolute number of quantitative disclosures increased. Another

explanation for the lower scores of quantitative is related to the fact that risks are inherently difficult

to quantify. Hence, directors may be reluctant to provide quantitative information against which they

may eventually be judged and held to account. Additionally, quantified risk information may be

highly sensitive and therefore subject to higher levels of proprietary costs (Linsley et al., 2006, p.

276). Linsley and Shrives (2005b) found in their study on non-financial firms also minimal disclosure

of quantified risk information.

It can be seen from table 4 that the sentence characteristic occurring most frequently is of the

‘qualitative/neutral/future’ type (6,575). The majority of these disclosures contain information and

explanations about general risk management policy. They are coded as ‘future’ information on the

64

basis that they discuss internal control and risk management systems that will be remaining in place.

For example, the annual report of Allied Irish Banks states about risk management in general (p. 32,

coded as F8): “AIB has adopted an Enterprise Risk Management (“ERM”) approach to identify, assess

and manage risk”. Another example found in this report is about a risk management policy with

respect to an individual risk category (p. 38, coded as F1): “The Group uses a simulation tool to

estimate possible changes in future market values and computes the credit exposure to a high level of

statistical significance.” Linsley et al. (2006) found this type to be dominating as well, but questioned

the usefulness of this type of disclosure. They argue that, although these disclosures reassure the

reader that risk management systems are in place, it does not provide information about specific

risks that the bank faces and nor does it explain the actions the directors have taken to manage that

risk. The results in this study are however different interpreted, as the author did not have this

impression during the empirical research. On one hand, general and sometimes vague risk

management policies were indeed disclosed (category 8, table 4). For example, the annual report of

Bank of Attica states (p. 25): “Taking into consideration international best practices, Attica Bank uses

modern methods to manage and monitor the risks involved in its activities.” . On the other hand, in

many cases general statements were followed by explanations about risk assessment, risk

measurement/quantification, risk controlling and risk monitoring, regarding the different risk

categories.

Just as quantitative risk information is usually considered to possess greater value than qualitative

information, so is future (forward-looking) risk information generally considered to be more useful

than past risk information. Approximately 68% of the defined risk disclosures were categorized as

future information and 33% as past information. These results seem to show that companies exhibit

a willingness to disclose future information. However, Linsley et al. (2006) notice that the majority of

future disclosures result from the ‘qualitative/neutral/future’ type, consisting largely of general

policy disclosures. Therefore, if this type is removed from the total of future disclosures, the

percentage future information drops from 68% to 11%, showing that future information is far less

disclosed than past information. Again, future disclosures are likely to be more sensitive and hence

that may explain why directors disclose them less frequently. This is confirmed by looking at the total

of ‘quantified/future’ information: this amount to only 1.2% of the total risk disclosures.

Table 5 also reveals that the number of neutral risk disclosures forms the majority (90%) of the

disclosures, compared to bad and good disclosures. It may be expected that directors will prefer to

present positive information and therefore there will be more good news risk disclosures. These

results are similar compared to Linsley et al. (2006) and Linsley and Shrives (2005b). That neutral

65

disclosures are dominant is according to Linsley and Shrives (2005b, p. 301) symptomatic for

companies. They release significant amounts of rather insipid general policy statements to the

market place, concerned with internal controls and risk management systems. The level of good and

bad news disclosures are comparable, indicating that directors cannot just choose to withhold bad

news. For example, poor financial results may require them to explain where risks have arisen with

adverse impacts (Linsley and Shrives, 2005b, p. 301). Although the banks show willingness to disclose

bad risk news, the difficulty is knowing whether there is further bad news that remains undisclosed

(Linsley et al., 2006, p. 277).

6.3.4 Conclusion

Compared to earlier studies measuring the risk disclosure score (in sentences), this study shows the

highest average, indicating that banks are disclosing more risk information in the annual report than

a few years earlier. Analyzing the different risk categories shows that credit and market risk are the

most developed risk disclosure areas. Since these risk areas are heavily regulated by the introduction

of IFRS 7 and Basel II this indicates that regulatory requirements drive risk disclosure to a large

degree. The examination of the characteristics of the risk disclosures revealed that it can be

questioned whether the increase in disclosure has led to more useful disclosure. The majority of the

risk disclosures is still qualitative, past and neutral. Relatively little quantitative and future risk

information is disclosed.

6.4 Part II: Regression analysis

6.4.1 Data analysis

This paragraph analyzes the data of the variables used in the regression analysis (table 6). Before

using multiple regression analysis, Cooke (1998, p. 209) argues that scholars should pay attention to

the structure of the data and consider the appropriateness of transformations. Data should be

screened to assess the impact of distribution problems of skewness and kurtosis, as well as problems

of outliers and non-linearity. A common measure to test for non-normality is the so-called

Kolmogorov-Smirnov test. This test was for the risk disclosure score, size (total assets), profitability

(roaa) and leverage significant. Visual inspection of the data confirmed non-normality. Therefore, the

data was transformed using the natural logarithm. After the transformation, the problems of

skewness, kurtosis and the outliers had been greatly reduced. Performing the K-S test again showed

no significance anymore on the independent variables. Table 6 provides the descriptive statistics of

the original variables and the transformed variables.

66

Table 6 Descriptive statistics of the dependent and independent variables

Unit of measurement

N Min. Max. Mean S.D. Non-normality1

Skewness/ Kurtosis

Out-liers

Dependent variableRisk disclosure score (RDSCORE)

Sentences 40 40 753 258 176.50 Yes 1.5231.960

3

Risk disclosure score (LNRDSCORE)

Log of sentences 40 3.69 6.62 5.35 0.65 No -0.1150.308

1

Panel A: Independent continuous variablesSize (SIZE) Total assets US$ (in

millions)40 571 2,833,804 470,906 863,257 Yes 1.770

1.6098

Size (LNSIZE) Log of total assets US$ (in millions)

40 6.35 14.86 10.66 2.51 No 0.334-1.010

0

Profitability (PROF)

Return on average assets (ROAA)

40 -1.07 5,07 1.14 0.94 Yes 1.8417.740

3

Profitability (LNPROF)

Log of return on average assets

40 -1.83 1.62 -0.038 0.66 No -0.2420.781

2

Leverage (LEV) Total equity / total assets

40 0.02 0.33 0.08 0.06 Yes 2.5519.254

3

Leverage (LNLEV) Log of total equity / total assets

40 -3.87 -1.12 -2.69 0.60 No 0.0410.361

1

Board composition (NONEXEC)

Non-exec. directors / total directors

26 0.45 0.88 0.70 0.12 No -0.552-0.453

0

1 using the Kolmogorov-Smirnov test (using p. < 0.05)

Panel B: Independent categorical variablesMultiple listing (ML)

Dummy (1=yes, 0=no)

40 0 1 0.25 0.44

Credit Rating (CR_TF)

Scale 1-6 standing for AAA to B

40 2 5 3.05 0.96

Organizational structure (TIER)

Dummy (0=one-tier, 1=two-tier)

40 0 1 0.35 0.48

6.4.2 Multiple regression

In SPSS a multiple linear regression was performed to identify (in)significant relationships between

the dependent and the independent variables. Multiple regression analysis estimates the effect from

each individual independent variable upon the dependent variable, using the input data. Ordinary

Least Square (OLS) is a common used method in applying linear regression and is standard used in

SPSS. Cooke (1998, p. 4) presents Rank Regression analysis as an alternative for the conventional

OLS. The advantage of Rank Regression is that it yields distribution-free test statistics. Table 6 shows

that the original data did not meet the normal distribution requirements for linear regression. To

back-up results from the transformation using the natural logarithm, a Rank Regression analysis was

performed. In this regression technique, both the dependent and independent variables (excluding

TIER, ML and CR_TF, as these are dummy variables) are replaced with their corresponding ranks.

67

Table 7 gives the model summary and the coefficients table from both regression analysis. The

analytical power of both models to predict its outcomes seems very powerful. The R square amounts

to 0.720 for the OLS model and 0.734 for the Rank Regression model. This means that the models

explain respectively 72% and 73.4% of the variation in risk disclosure, using the data from the

independent variables as predictor. The adjusted R square gives an idea about how well the models

generalize. These numbers are very high, indicating the model is accurately explaining the variance in

risk disclosure.

Table 7 Regression analyses

Model 1 Model 2Regression using… natural logarithm ranked data

Model summaryR 0.849 0.857R Square 0.720 0.734Adjusted R Square 0.669 0.686Durbin-Watson 1.684 1.857

Independent variables1

Size 0.205(3.784)*

0.769(4.284)*

Profitability 0.134(0.790)

0.071(0.537)

Leverage -0.007(-0.029)

-0.034(-0.183)

Multiple Listing 0.133(0.665)

3.353(1.008)

Credit Rating -0.064(-0.631)

-0.263(-0.147)

Tier Structure 0.170(1.241)

1.738(0.738)

Added after removing TIERNon-executive ratio (n=26) -1.130

(-1.404)-0.171

(-0.779)1The upper figures for each variable are beta coefficients and the lower figures are the t-statistics.* significant at the 5% level.

To test the non-existence of autocorrelation (i.e. the assumption of independent errors), the Durbin-

Watson statistic was utilized. A value less than 1 or greater than 3 should pose a problem, and the

closer to 2 the value is the better the model. The Durbin-Watson statistics are 1,684 and 1,857,

indicating there are no validity problems of both models.

68

As stated in chapter 5, the variable NONEXEC is not included in the initial regression models. The

reason is that NONEXEC only has a value when the TIER variable (organizational structure) amounts

to 0 (i.e., has a one-tier structure). Out of the 40 cases, 26 banks have a one-tier system and

therefore there are only 26 values for NONEXEC. If the regression model was executed with

NONEXEC, the 14 cases missing NONEXEC values would be removed, and accordingly TIER would be

removed, as this variable would be constant (i.e, amounts to 0 in all cases). Therefore, NONEXEC is

not added in testing the models here described. However, to be able to test the hypothesis relating

to NONEXEC, the regression models are executed again without the TIER variable but this time with

the NONEXEC variable. Table 7 also shows the computed coefficient and t-statistic for the NONEXEC

variable.

The beta coefficients indicate the relationship between the dependent variable (risk disclosure score)

and each of the independent input variables. If this coefficient value is positive, this means that there

is a positive relationship between the dependent and independent variable. For example, in model 1

when size (LNSIZE) increases with 1, risk disclosure (LNRDSCORE) increases with 0,205. The

regression analysis also shows whether multicollinearity (correlation between independent variables)

is a problem. A common thumb rule in interpreting these statistics is that values of Tolerance smaller

than 0.1 and VIF values bigger than 10 are a strong indication of the existence of multicollinearity

(Hassan, 2009, p 677). The values in both models are within this range, indicating no significant effect

of multicollinearity (for the data, see the SPSS output in Appendix I).

6.4.3 Interpretation of the results

This paragraph interprets the output of the multiple regression regarding the computed beta

coefficients related to each independent variable (see Table 7).

The coefficients and the significance of the t-test show whether the model supports the pre-

developed hypotheses. The positive significant coefficient between size and risk disclosure confirms

hypothesis 1 (H1). Large banks disclose more about their risks than small banks. It supports the

positive accounting theory regarding the political cost hypothesis, when we apply this hypothesis

towards risk disclosure. Furthermore, it supports the stakeholder theory. To meet information needs

of a large group of stakeholders, the bank’s burden of disclosure becomes heavier. Finally, it confirms

earlier findings of Beretta and Bozzolan (2004), Linsley and Shrives (2005b), Linsley et al. (2006),

Abraham and Cox (2007), Hossain (2008), Amran et al. (2009) and Michiels et al. (2009).

69

No significant relationship is found between probability and risk disclosure. Therefore, there is no

support for hypothesis 2 (H2), although the positive beta coefficient is consistent with the theoretical

reasoning of the hypothesis. In earlier studies theoretical constructs about the relationship between

relative profitability and risk disclosure have been developed, but generally these are not empirically

confirmed. However, a few studies found significant relationships, either positive or negative

(Hossain, 2008; Helbok and Wagner, 2003; Michiels et al., 2009). A possible explanation that this

study did not find a significant relationship might be due to the consequences of the financial crisis.

Massive write-downs on certain portfolios led to major losses at certain European banks. Hence, this

influenced net income, one of the two elements in the profitability ratio return on assets. As a result,

the profitability ratio might be biased due to the volatility in net income.

Regarding the third hypothesis, no significant relationship is found between a bank’s level of risk and

risk disclosure. Therefore, it does not appear that the riskier banks are seeking to give the

marketplace confidence in their ability to manage risk through the disclosure of greater amounts of

risk information. Banks may not want to draw attention to their ‘riskiness’ and therefore may be

reluctant to voluntary disclose significant amounts of risk information. By contrast, banks with lower

levels of risk, perhaps because of their superior risk management abilities, may wish to signal this

through improved risk disclosure (Linsley and Shrives, 2006, p. 391). Linsley et al. (2006, p. 279)

suggest that it may be that the banks wish to keep discussions concerning their risk levels and their

risk management capabilities out of the public domain. This desire for privacy may be related to the

issue of proprietary costs. There is no empirical support for hypothesis 3. This confirms findings of

Linsley and Shrives (2005b), who used five proxies to capture the level of company risk but did not

found any of them significantly related to the extent of risk disclosure. Similar non-significant results

were found by Linsley et al. (2006), Lajili and Zeghal (2005) and Amran et al. (2009).

The fourth hypothesis (H4) emphasized that the extent of risk disclosure is greater for banks with

multiple listing statuses. The theoretical construct (see 5.4.4) seems very solid and is supported by

two earlier studies (Abraham and Cox, 2007 and Haniffa and Cooke, 2005). This study however shows

no significant coefficient on multiple listing statuses related to risk disclosure and therefore no

empirical support is found for the fourth hypothesis. On the other hand, the direction of the

relationship is positive, pointing towards a positive relationship that confirms the theoretical

construct. The relationship between credit ratings and risk disclosure is for the first time investigated

in risk disclosure research. It was argued that banks with higher credit ratings release more risk

disclosure than banks with lower credit ratings. However, the coefficient on credit ratings is negative

and not found to be significant. Therefore, the fifth hypothesis (H5) is not empirically supported.

70

Regarding the sixth hypothesis about organizational structure, it has been hypothesized that a two-

tier structure leads to stronger corporate governance, and hence to more risk disclosure. However,

the coefficient is found to be not significant, meaning that there is no empirical evidence to support

this hypothesis. In contrast, it confirms the BCBS notion (1999) that there are no universally correct

answers to corporate governance structures, as these structures vary in different countries.

For the seventh hypothesis, the proportion of non-executive directors divided by the number of total

directors is used. According to Abraham and Cox (2007, p. 231) non-executives in the board give

balance in the organization and a board with non-executive directors in greater numbers is in a

stronger position to fulfill shareholder preferences for accountability and transparency. Testing the

variable however leads to no significant relationship.

6.4.4 Conclusion

Although the outcomes show that the model is powerful in explaining the variation in the extent risk

of disclosure, a detailed analysis of the independent variables show that only size is significant.

Regression analysis using ranked data leads to the same results.

71

7 Conclusion

7.1 Introduction

This chapter presents the main conclusions from this study and furthermore gives attention to its

limitations and provides recommendations for future research.

7.2 Research conclusion

Never before has the importance of transparency and the role of risk disclosure been highlighted as

today. The recent credit crisis has emphasized the need for transparent and open disclosure of

financial information to all relevant market parties. When the risk management of an institution is

adequate, it is possible to obtain confidence in the markets by credible risk disclosure. On the other

hand, if the quality of risk management is poor, then it is likely that the risk disclosure is insufficient.

This study aimed to give insight in the risk disclosure practices and its determinants in the European

banking sector in 2007. The following research question has been formulated (chapter 1):

In which way do European banks disclose risks in the annual reports of 2007 and which firm-specific

characteristics influence the extent of risk disclosure?

Before the empirical research, first the institutional framework and theoretical framework related to

risk disclosure were described. Moreover, a literature review of risk disclosure has been provided

and discussed. The first part of the empirical research consisted of analyzing 40 annual bank reports

of 2007 by means of content analysis. Risk disclosure score was measured by the number of

sentences containing risk information. Furthermore, different attributes of the risk information were

taken into account: qualitative/quantitative, future/past and good/neutral/bad. The content analysis

showed how European banks disclosed risk in their annual reports.

This study shows the highest average of risk disclosure per annual report, compared to earlier studies

measuring the risk disclosure score in sentences. This indicates that banks are disclosing more risk

information in the annual report than a few year earlier. Analyzing the different risk categories shows

that credit and market risk are the most developed risk disclosure areas. This is in line with

expectations, as IFRS 7 and Basel II were introduced in 2007, supporting the notion that regulatory

requirements drive risk disclosure to a large degree. The examination of the characteristics of the risk

disclosures revealed that it can be questioned whether the increase in disclosure has led to more

useful disclosure. The majority of the risk disclosures is still qualitative, strong biased towards the

72

past and of neutral nature. Relatively little quantitative and future risk information is disclosed,

supporting the view that directors are not willing to disclose proprietary information due to its

sensitivity towards competitors or its high level of unreliability, increasing chances of possible

litigation. Directors could feel that they will be required to justify their estimates, when they disclose

future information which is inherently unreliable.

The second part of the empirical study tests for relationships between the gathered risk disclosure

score and potential explanatory firm characteristics. These include variables such as company size,

profitability, risk level, multiple listing statuses, credit rating and corporate governance

characteristics. The technique used is regression analysis (using both OLS and Rank Regression).

Although the outcomes show that the model is powerful in explaining the variation in the extent risk

of disclosure, a detailed analysis of the independent variables show that only the independent

variable size is significant. The significance of the size variable confirms theoretical constructs from

the stakeholder theory and findings of previous literature. As the bank grows bigger, it will have a

larger pool of stakeholders who would be interested to know the affairs of the company. Some

variables, found significant in earlier studies did not show significance in this study. The predicted

relationship between profitability and risk disclosure was not confirmed, but the positive coefficient

is consistent with the hypothesis. Moreover, leverage as a proxy of level of risk is not found

significant too. This supports the view of Linsley and Shrives (2006, p. 279) who suggest that directors

wish to keep discussions concerning their risk levels and their risk management capabilities out of

the public domain. This desire for privacy can be related to the subject of proprietary costs.

Additionally, the credit crunch might have had a major impact upon the profitability and leverage

ratios, and in that way possibly biased the results. The multiple listing and credit rating variables are

not found to be significant either and therefore seems not relevant for explaining risk disclosure.

Finally, this study takes two corporate governance variables into account. However, none of them

are found significant, confirming findings of Michiels et al. (2009) but contrasting with results of

Abraham and Cox (2007).

7.3 Limitations

An important limitation in this study lies in its method to analyze the annual report. Subjectivity of

content analysis cannot be fully eliminated because of the use of only one coder. However, chapter

five showed that several decisions were made to overcome major limitations of this method and to

increase the reliability of the coding process. Another limitation results from the fact that content

analysis only measures the extent of risk disclosure and not quality. Although the study strictly

considered focuses upon the extent of risk disclosure, the implicit underlying construct of interest is

73

generally the ‘quality’ of disclosure (as stated by Beattie et al., 2004, p. 4). Therefore, to analyze risk

disclosure more in-depth, characteristics of risk disclosure were examined. The study showed that

the qualification of past/future information was inadequate, as it might be argued that risk

information about current risk management practices cannot be considered as past nor future

information. Further research should distinguish another category next to past and future

information, namely “non-time-specific”. Finally, a limitation is found in sample size. Due to the

massive and time-consuming work of the content analysis method, only 40 cases (one annual report

of 40 European public listed banks) were analyzed.

7.4 Recommendations for future research

Since risk disclosure literature is scarce, further research into risk reporting would be beneficial. Risk

disclosure research over a couple of years could reveal more insight into patterns and practices of

risk disclosure by companies. Moreover, advanced techniques should be developed and used in

measuring quantity and quality of risk disclosure. Beattie et al. (2004) already introduced computer-

aided coding, to be able to analyze larger samples in an easy and reliable way. Another topic of

research that would enhance the understanding and effect of risk disclosure is to investigate in what

manner risk information would be of greatest use to annual report readers. These studies would be

beneficial and helpful to regulators to design and improve risk disclosure legislation.

74

Reference List

Abraham, S. and Cox, P. (2007), “Analyzing the determinants of narrative risk information in UK

FTSE 100 annual reports”, British Accounting Review, Vol. 39, pp. 227-248.

Amran A., Bin, A.M.R. and Hassan, B.C.H.M. (2009), “Risk reporting: An exploratory study on risk

management disclosure in Malaysian annual reports”, Managerial Auditing Journal, Vol. 24 No. 1,

pp. 39-57.

Balin, B.J. (2008), “Basel I, Basel II and Emerging Markets: A Nontechnical Analysis”, Unpublished

paper, The Johns Hopkins University School of Advanced International Studies (SAIS).

Bank of International Settlements (2009), via website www.bis.org, retrieved on: 14-08-2010.

Basel Committee on Banking Supervision (1998), “Enhancing Bank Transparency”, Bank of

International Settlements (BIS).

Basel Committee on Banking Supervision (1999), “Enhancing corporate governance for banking

organisations”, Bank of International Settlements (BIS), revised in 2006.

Basel Committee on Banking Supervision (2001), “Public disclosures by banks: results of the 1999

disclosures survey”, Bank of International Settlements (BIS).

Basel Committee on Banking Supervision (2002), “Public disclosures by banks: results of the 2000

disclosures survey”, Bank of International Settlements (BIS).

Basel Committee on Banking Supervision (2003), “Public disclosures by banks: results of the 2001

disclosures survey”, Bank of International Settlements (BIS).

Beattie, V., McInnes, W. and Fearnley S. (2004), “A methodology for analyzing and evaluating

narratives in annual reports: a comprehensive descriptive profile and metrics for disclosure

quality attributes, Accounting Forum, Vol. 28 No 3, pp. 205-236.

Beretta, S. And Bozzolan, S. (2004), “A framework for the analysis of firm risk communication”,

The International Journal of Accounting, Vol. 39, pp. 265-288.

Brown, S. and Hillegeist, S.A. (2007), “How disclosure quality affects the level of information

asymmetry, Review of Accounting Studies, Vol. 112 Nos. 2/3, pp. 443-477.

Cabedo, J.D, and Tirado, J.M. (2004), “The disclosure of risk in financial statements”, Accounting

Forum, Vol. 28, pp. 181-200.

Chen, W.P., Chung, H., Lee, C., and Liao, W.L. (2007), “Corporate Governance and Equity

Liquidity: analysis of S&P transparency and disclosure rankings”, Corporate Governance, Vol. 15

No. 4, pp. 644-660.

Chen, K.Y., Elder, R.J. and Hsieh, Y.M. (2007), “Corporate Governance and Earnings Management:

The implications of Corporate Governance Best-Practices Principles for Taiwanese Listed

75

Companies”, Paper presented at the 2006 Journal of Contemporary Accounting and Economics

Symposium, via website www.ssrn.com.

Crockett, A. (2002), “Deposit insurance and other current issues in international finance”, May 7,

via website http://www.iadi.org/speeches/050702.pdf, retrieved on: 14-08-2010.

Crockett, A. (2003), “Central banking, financial stability and Basel II”, February 13, via website

www.bis.org/speeches/sp030213.htm, retrieved on: 14-08-2010.

Davidson, R., Goodwin-Stewart, J., and Kent, P. (2005), “Internal governance structures and

earnings management”, Accounting and Finance, Vol. 45, pp. 241-267.

De Nederlandsche Bank (2007), “Bazel II / CRD”, via website www.dnb.nl

(http://www.dnb.nl/openboek/extern/id/nl/all/40-158788.html), retrieved on: 14-08-2010.

Deegan C. and Unerman, J. (2006), Financial Accounting Theory, McGraw-Hill, Berkshire.

Denis, D.K. and McConnel, J.J. (2003), “International Corporate Governance”, The Journal of

Financial and Quantitative Analysis, Vol. 38 No. 1, pp. 1-36.

Dia, M. and Zeghal, D. (2008), “Fuzzy evaluation of Risk Management Profiles Disclosed in

Corporate Annual Reports, Canadian Journal of Administrative Sciences, Vol. 25, pp. 237-254.

Dobler, M. (2008), “Incentives for risk reporting – A discretionary disclosure and cheap talk

approach”, The International Journal of Accounting, Vol. 43, pp. 184-206.

Douglas, N. (2007), “In The Light of Current Evidence, Critically Examine The Efficient Market

Hypothesis”, via website www.nedprod.com/NeoCapitalism/MN3101%20Essay.pdf, retrieved on:

14-08-2010.

Eijffinger, S.C.W. (2005), The European Central Bank and Financial Supervision, in: Masciandaro,

D., Handbook of Central Banking and Financial Authorities in Europe, Edward Elger Publishing, pp.

457-484.

Engels, C. (2008), “Has IFRS 7 increased the quality of disclosures?”, Master Thesis, Rotterdam

School of Management, Erasmus University Rotterdam.

Ernst & Young (2007), “IFRS 7 Financial Instruments: Disclosures”, second edition, via website

www.ey.com, retrieved on: 14-08-2010.

European Corporate Governance Institute (2009), via website www.ecgi.org, retrieved on: 14-08-

2010.

Fama, E. and Jensen, M. (1983), “Separation of ownership and control”, Journal of Law and

Economics, Vol. 26, p. 301-326.

Financial Reporting Council (2005), “Internal Control: Revised Guidance for Directors on the

Combined Code”, via website www.frc.org.uk, retrieved on: 14-08-2010.

Freeman, R. (1984), Strategic Management: a stakeholder perspective, Prentice-Hall, New Jersey.

76

Hackston D., and Milne M.J. (1996), “Some determinants of social and environmental disclosures

in New Zealand companies”, Accounting, Auditing and Accountability Journal, Vol. 9 No. 1, pp.

77-108.

Hassan, M.K. (2008), The level of corporate risk disclosure in UAE, University of Sharjah, paper

presented at the British Accounting Association Conference, Blackpool, April 1-3.

Healy, P.M. and Palepu, K.G. (2001), “Information asymmetry, corporate disclosure and the

capital markets: A review of the empirical disclosure literature”, Journal of Accounting and

Economics, Vol. 31, pp. 405-440

Helbok, G. and Wagner, C. (2003), “Corporate Financial Disclosure on Operational Risk in the

Banking Industry”, presented at the Austrian Working Group on Banking and Finance, November,

Graz.

Hossain, M. (2008), “The Extent of Disclosure in Annual Reports of Banking Companies: The Case

of India”, European Journal of Scientific Research, Vol. 23 No. 4, pp. 659-680.

International Accounting Standards Board (2006), International Financial Reporting Standards

(IFRSs), via website www.books.google.com, retrieved on: 14-08-2010.

International Accounting Standards Board (2009), “Technical Summary: IFRS 7 Financial

Instruments: Disclosures”, via website www.iasb.org, retrieved on: 14-08-2010.

Jaggi, B. and Low, P. Y. (2000), “Impact of Culture, Market Forces and Legal System on Financial

Disclosures”, The International Journal of Accounting, Vol. 35 No. 4, pp. 495-519.

Jensen, M.C. and Meckling, W.H. (1976), “Theory of the firm: Managerial behavior, Agency cost

and ownership structure”, Journal of Financial Economics, Vol. 3, pp. 305-360.

Klaassen, J. and Hoogendoorn, M.N. (2004), Externe verslaggeving, Stenfert-Kroese, Groningen.

KPMG (2008), “Financial Institution Risk Disclosure – Best Practice Survey 2008”, via website

www.kpmg.com, retrieved on: 14-08-2010.

Krippendorf, K. (1980), Content Analysis: An Introduction to its Methodology. Sage, New York.

Lajili, K. and Zeghal D. (2005), “A content analysis of risk management disclosure in Canadian

annual reports”, Canadian Journal of Administrative Sciences, Vol. 22 No. 2, pp. 125-142.

Lang, M. and Lundholm R. (1993), “Cross-sectional determinants of analyst’s ratings of corporate

disclosures, Journal of Accounting Research, Vol. 31 No. 2, pp. 246-271.

Larosière, J. de (2009), “Report”, The High-Level Group on Financial Supervision in the EU.

Leenaars, J.J.A. (2003), “Risicomanagement van banken”, Maandblad voor Accountancy en

Bedrijfseconomie, Juli/Augustus, pp. 340-347.

Levine, R. (2004), “The Corporate Governance of Banks: A Concise Discussion of Concepts and

Evidence”, World Bank Policy Research Paper 3404, September.

77

Linsley, P.M. and Shrives, P.J. (2005a), “Transparency and the disclosure of risk information in the

banking sector”, Journal of Financial Regulation and Compliance, Vol. 13 No. 3, pp. 205-214.

Linsley, P.M. and Shrives, P.J. (2005b), “Examining risk reporting in UK public companies”, The

Journal of Risk Finance, 2005; Vol. 6 No. 4, pp. 292-305.

Linsley, P.M., Shrives, P.J. and Crumpton, M. (2006), “Risk Disclosure: An exploratory study of UK

and Canadian banks”, Journal of Banking Regulation, Vol. 7 Issue 3/4, pp. 268-282.

Linsley, P.M. and Shrives, P.J. (2006), “Risk reporting: a study of risk disclosure in the annual

reports of UK companies”, The British Accounting Review, Vol. 38, pp. 387-404.

Linsmeier, T.J., Thornton, D.B., Venkatachalam, M. and Welker, M. (2002), “The effect of

mandated market risk disclosures on trading volume sensitivity to interest rate, exchange rate

and commodity price movement”, Accounting Review, Vol. 77 No. 2, pp. 343-377.

Mecking, E. (2008), Deflatie in aantocht”, Mets & Schilt, Amsterdam.

Meijer, J.W.M.K (2003), “Verslaggeving over risico’s”, Maandblad voor Accountancy en

Bedrijfseconomie, Maart, pp. 109-118.

Michiels, A., Vandemaele S. and Vergauwen, Ph. (2009), “Risicorapportering door Belgische

beursgenoteerde bedrijven”, Accountancy & Bedrijfskunde, No. 4, pp. 3-11.

Milne, M.J. and Adler, R.W. (1999), “Exploring the reliability of social and environmental

disclosures content analysis”, Accounting, Auditing & Accountability Journal, Vol. 12 No. 2, pp.

237-256.

Milne, M.J. (2002), “Positive Accounting Theory, Political Costs and Social Disclosure Analysis: a

critical look”, paper presented at the 2001 BAA Annual Conference at the University of

Nottingham.

Morgan, D.P. (2002), “Rating Banks: Risk and Uncertainty in an Opaque Industry”, The American

Economic Review, Vol. 92 No.4, pp. 874-888.

Ormerod, P. (2008), “Credit crunch: Risk-taking”, The Sunday Times, March 23.

O’Sullivan, N. (2000), “The impact of board composition and ownership on audit quality:

evidence from large UK companies”, British Accounting Review, Vol. 32, pp. 397-414.

Padoa-Schioppa, T. (1999), EMU and Banking Supervision, International Finance, Vol. 2 No. 2, pp.

295-308.

Prat, A. (2005), “The Wrong Kind of Transparency”, The American Review, Vol. 95 No. 3, pp. 862-

877.

Pyle, D.H. (1997), Bank Risk Management: Theory, Working Paper, Haas School of Business,

University of California, Berkeley.

PricewaterhouseCoopers (2006), “IFRS 7: Ready or not”, via website www.pwc.com, retrieved

on: 14-08-2010.

78

RiskMetrics Group (2009), “Study on Monitoring and Enforcement Practices in Corporate

Governance in the Member States”, via the website of the European Commission (ec.europa.eu),

retrieved on: 14-08-2010.

Ryan, S.G. (2008), “Accounting in and for the Subprime Crisis”, The Accounting Review, Vol. 83

No. 6, pp. 1605-1638.

Santomero, A.M. (1997), Commercial Bank Risk Management: an Analysis of the Process, Paper

presented at the Wharton Financial Institutions Center Conference on Risk Management in

Banking (October 1996), University of Pennsylvania.

Schrand, C.M., and Elliot, J.A. (1998), “Risk and Financial Reporting: A Summary of the Discussion

at the 1997 AAA/FASB conference”, Accounting Horizons, Vol. 12 No. 3, pp. 271-282.

Scott, W.R. (2006), Financial Accounting Theory, Pearson Prentice Hall, Toronto.

Solomon, J.F., Solomon, A., Norton, S.D. and Joseph, N.L. (2000), “A conceptual framework for

corporate risk disclosure emerging from the agenda for corporate governance reform”, British

Accounting Review, Vol. 32, pp. 447-478.

Stern, G.H. and Feldman, R.J. (2004), Too big to fail: the hazards of bank bailouts, The Brookings

Institution, Washington D.C.

Stolowy H. and Lebas, M.J. (2006), Financial Accounting and Reporting: A global perspective,

Cengage Learning (former Thomson Business Press).

Tett, G. (2008), “Why the pain in Spain has mainly been contained”, Financial Times, February 1.

Unerman, J. (2000), “Methodological issues: reflections on quantification in corporate social

reporting content analysis”, Accounting, Auditing & Accountability Journal, Vol. 13 No. 5, pp. 667-

680.

Venkatachalam, M. (1996), “Value-relevance of banks’ derivative disclosures”, Journal of

Accounting and Economics, Vol. 22, pp. 327-355.

Visser, M. (2009), “Spaans toezicht lichtend voorbeeld voor nieuwe kapitaaleis aan banken”, Het

Financieele Dagblad, March 23.

Wallace, R.S.O., and Naser, K. (1995), “Firm-specific determinants of the comprehensiveness of

mandatory disclosure in the corporate annual reports of firms listed on the stock exchange of

Hong Kong”, Journal of Accounting and Public Policy, Vol. 14 No. 4, pp. 311-368.

Weber, R.P. (1985), “Basic Content Analysis”, Quantitative Applications in the Social Sciences, No.

49, Beverly Hills, CA: Sage Publications.

Weil, Gotshal & Manges LLP (2002), “Comparative Study of Corporate Governance Codes

Relevant to the European Union and its Member States”, via the website of the European

Commission (ec.europa.eu), retrieved on: 14-08-2010.

79

Appendix A

Dobler (2008, p. 188)

80

Appendix B

Basel II - Summary disclosure requirements

Disclosure area Nature of disclosures in summary

Capital disclosuresCapital structure Main features of capital instruments

Amounts of tier 1, 2 and 3 capitalCapital requirements

Capital adequacy

Risk exposure and assessment disclosuresCredit risk General disclosure of risk management objectives and policies

for each separate risk area including:Strategies and processesStructure and organization of the relevant risk mgt functionsScope and nature of risk reporting and measurement systemsPolicies for hedging and mitigating risks

Specific disclosures for each separate risk area as prescribed

Market riskOperational riskEquitiesInterest rate risk

81

Appendix C

Sample selection

BANK NAME COUNTRY TOTAL ASSETS (USD, in millions, 2007)

BKS Bank AUSTRIA 5.752KBC Group BELGIUM 523.476First Investment Bank BULGARIA 3.156Bank of Cyprus Group CYPRUS 46.759Marfin Popular Bank CYPRUS 44.542Komercni Banka CZECH REPUBLIC 36.609Spar Nord Bank DENMARK 12.491Amagerbanken DENMARK 6.128Vestjysk Bank DENMARK 3.648BNP Paribas FRANCE 2.494.412Société Générale FRANCE 1.577.745Crédit Agricole FRANCE 2.081.883Commerzbank GERMANY 907.514Deutsche Bank GERMANY 2.833.804Comdirect Bank GERMANY 12.120Bank of Attica GREECE 5.747General Bank of Greece GREECE 6.381EFG Eurobank Ergasias GREECE 100.676Allied Irish Banks IRELAND 261.831Banca Ifis ITALY 1.877Banca Generali ITALY 6.199Banca Popolare di Milano ITALY 64.223Credito Valtellinese Soc Coop ITALY 25.362Banco Desio SpA ITALY 11.893Banca popolare dell'Etruria e del Lazio ITALY 12.587Siauliu Bankas LITHUANIA 875FIMBank Plc MALTA 571Van Lanschot NETHERLANDS 31.972ING Groep NETHERLANDS 1.932.151Bank Zachodni WBK POLAND 16.969Banco BPI PORTUGAL 59.688Tatra Banka SLOVAKIA 11.032OTP Banka Slovensko SLOVAKIA 2.167Probanka d.d. Maribor SLOVENIA 1.613Bankinter SPAIN 73.088Banco Bilbao Vizcaya Argentaria SPAIN 502.204Svenska Handelsbanken SWEDEN 289.915Barclays UNITED KINGDOM 2.459.149Schroders UNITED KINGDOM 13.778HSBC Holdings UNITED KINGDOM 2.354.266

82

Appendix D

Risk classification model

Risk category Subcategory1 Market risk interest rate risk

foreign exchange riskcommodity riskequity price risk

financial risks2 Credit risk counterparty risk

3 Liquidity risk liquidity risk

4 Operational risk failing of internal processesfailing of people; human error riskfailing of systems; IT riskinformation access and availability riskfraud riskinternal control weaknessescustomer satisfactionproduct and service failure risk

5 Business risk strategic riskcompetition riskreputation risk environmental risksystemic risk

6 Legal risk lawsuits, litigationchange in political environmentchange in legislationchange in tax law

7.

Capital structure and adequacy risk off balance structures

risk based ratiosrisk exposures of on- and off balance assets

8.

Risk management and policies risk management

actions and policies

83

* This model shows the most relevant risk areas for the banking industry and is based upon theory

and the risk classification models of Linsley et al. (2006), KPMG (2008), Amran et al. (2009) and

Michiels et al. (2009).

84

Appendix E

Coding grid

Market risk

Credit risk

Liqui-dity risk

Oper-ational risk

Business risk

Legal risk

Capital structure, adequacy risk

Risk management and policies

Sentence characteristics 1 2 3 4 5 6 7 8Quantitative Good Future AQuantitative Bad Future BQuantitative Neutral Future CQualitative Good Future DQualitative Bad Future EQualitative Neutral Future FQuantitative Good Past GQuantitative Bad Past HQuantitative Neutral Past IQualitative Good Past JQualitative Bad Past KQualitative Neutral Past L

85

Appendix F

Based upon Linsley et al. (2006) and Amran et al. (2009)

Decision rules for coding risk disclosures

(1) To identify risk disclosures, a broad definition of risk is adopted as explained below.

(2) Sentences are to be coded as risk disclosures if the reader is informed of any opportunity or

prospect, or of any hazard, danger, harm, threat or exposure, that has already impacted

upon the company, or may impact upon the company in the future or of the management of

any such opportunity, prospect, hazard, harm, threat or exposure.

(3) The risk definition just stated shall be interpreted such that “good” or “bad” “risk” and

uncertainties will be deemed to be contained within the definition.

(4) The type of risk disclosure shall be classified according to appendix E.

(5) If a sentence has more than one possible classification, the information will be classified into

the category that is most emphasized within the sentences.

(6) Tables (quantitative and qualitative) that provide risk information should be interpreted as

one line equals one sentence and classified accordingly.

(7) Any disclosure that is repeated shall be recorded as a risk disclosure sentence only when new

information is revealed or discussed.

(8) If a disclosure is too vague in its reference to risk, then it shall not be recorded as risk

disclosure.

86

Appendix G

Examples of risk disclosure in the sampled banks’ annual reports:

Commerzbank, Germany, p. 99: “Commerzbank has been able to respond quickly and strategically to

changing market situations using the liquidity risk measurement and control method that was

developed and established internally years ago – the available net liquidity (ANL) concept.”

Coded as J3 in the coding grid: liquidity risk, good news, qualitative and referring to the past.

FIMBank, Malta, p. 51: “As the Group carries out activities with counter-parties in emerging markets,

there are certain risk factors which are particular to such activities and which require careful

consideration by prospective investors since they are not usually associated with activities in more

developed markets.”

Coded as F2 in the coding grid. Counterparty risk is subcategorized under credit risk, it is neutral

news and refers to current practices that are in place and will be, and as such characterized as future

news.

Barclays, United Kingdom, p. 81: “Barclays Capital’s market risk exposure, as measured by average

total Daily Value at Risk (DVaR), increased to an average of £42m in 2007.”

Coded as H1 in the coding grid. Market risk, quantitative, refers to the past and can be interpreted as

bad news, taking the context into consideration.

ING, Netherlands, p. 83: “ING Bank has a framework of risk management policies, procedures and

standards in place to create consistency throughout the organization, and to define minimum

requirements that are binding on all business units.”

Coded as F8 in the coding grid. Disclosure about risk management policies, currently in place, neutral

and qualitative of nature.

Svenska Handelsbanken, Sweden, p. 54: “The economic capital model which is a component in this, is

designed to ensure that the Group has sufficient capital in relation to all its risks at any point in time.”

Coded as F7 in the coding grid. Capital structure and adequacy risk, qualitative, and future

information.

87

Appendix H

BANK NAME RISK DISCLOSURE SCORESvenska Handelsbanken 321Spar Nord Bank 163Amagerbanken 158Vestjysk Bank 59BNP Paribas 314Société Générale 288Crédit Agricole 518Commerzbank 461Deutsche Bank 506Comdirect Bank 158Banca Ifis 175Banca Generali 155Bank of Attica SA 40General Bank of Greece 104Allied Irish Banks 393EFG Eurobank Ergasias 122Banca Popolare di Milano 279Van Lanschot 239ING Groep 701Bankinter 211Banco Bilbao Vizcaya Argentaria 350Barclays 753Schroders 152Bank of Cyprus Group 182Marfin Popular Bank 168Komercni Banka 243Bank Zachodni 248Banco BPI 285HSBC Holdings 735Tatra Banka 142OTP Banka Slovensko 70Siauliu Bankas 117Probanka d.d. Maribor 189Credito Valtellinese 225FIMBank 112Banco Desio 247Banca popolare dell'Etruria e del Lazio 170BKS Bank AG 167KBC Group 300First Investment Bank 109

Appendix I

88

Output SPSS – OLS Regression

Model Summaryb

Model R R Square

Adjusted R

Square

Std. Error of the

Estimate Durbin-Watson

1 ,849a ,720 ,669 ,37605 1,684

a. Predictors: (Constant), TIER, LNLEV, ML, CR_TF, LNPROF, LNSIZE

b. Dependent Variable: LNRDSCORE

Coefficientsa

Model

Unstandardized

Coefficients

Standardized

Coefficients

t Sig.

Collinearity Statistics

B Std. Error Beta Tolerance VIF

1 (Constant) 3,252 ,931 3,493 ,001

LNSIZE ,205 ,054 ,786 3,784 ,001 ,197 5,086

LNPROF ,134 ,170 ,136 ,790 ,435 ,284 3,517

LNLEV -,007 ,249 -,007 -,029 ,977 ,163 6,152

ML ,133 ,200 ,089 ,665 ,511 ,471 2,125

CR_TF -,064 ,101 -,093 -,631 ,532 ,388 2,575

TIER ,170 ,137 ,126 1,241 ,223 ,825 1,213

a. Dependent Variable: LNRDSCORE

NONEXEC -1,130 ,805 -,204 -1,404 ,176 ,738 1,355

89

Output SPSS – Rank Regression

Model Summaryb

Model R R Square

Adjusted R

Square

Std. Error of the

Estimate Durbin-Watson

1 ,857a ,734 ,686 6,551275 1,857

a. Predictors: (Constant), TIER, Rank of LEV_ED, ML, CR_TF, Rank of ROA, Rank of

TA_USD_MIL

b. Dependent Variable: Rank of RDSCORE

Coefficientsa

Model

Unstandardized

Coefficients

Standardized

Coefficients

t Sig.

Collinearity Statistics

B Std. Error Beta Tolerance VIF

1 (Constant) 3,348 8,354 ,401 ,691

Rank of TA_USD_MIL ,769 ,179 ,769 4,284 ,000 ,250 3,997

Rank of ROA ,071 ,132 ,071 ,537 ,595 ,461 2,171

Rank of LEV_ED -,034 ,187 -,034 -,183 ,856 ,230 4,350

ML 3,353 3,326 ,126 1,008 ,321 ,517 1,934

CR_TF -,263 1,788 -,022 -,147 ,884 ,374 2,675

TIER 1,738 2,355 ,072 ,738 ,466 ,850 1,176

a. Dependent Variable: Rank of RDSCORE

Rank of NONEXEC -,171 ,220 -,110 -,779 ,446 ,723 1,383

90