Transparency of Derivatives: Risk and Disclosure in the Netherlands
M.B. Polstra
Faculty of Economics and Business Rijksuniversity of Groningen
Student number: s1775677 Master Thesis Finance1 Master Thesis Accountancy2 Supervisor (Finance): dr. P.P.M Smid
Supervisor (Accountancy): prof. dr. R.L. ter Hoeven Supervisor KPMG: Walter Berger
Date: 21-6-2013
Abstract
Previous research documents little about derivative usage in the Netherlands. Combining the use of derivatives with the quality of derivative disclosure provides insights in the manner companies are transparent about their use of derivatives. This is especially interesting given new accounting regulation (IFRS 7) to ensure a high level of transparency of the use derivatives. This paper finds for a sample of 22 Dutch listed firms no evidence to conclude in hedging or speculating purposes. In contrast to the lack of derivative intentions I find a negative relation between the firm’s derivative position and the quality of disclosure, indicating that firms with more derivatives construct less understandable derivative disclosures. Besides, the understandability is not influenced by IFRS 7.
Keywords: Derivative Financial Instruments, Derivative Disclosure, Hedging, Speculation, IFRS 7, Understandability
JEL classification: G3, G110, M48
1
Program code Finance: EBM866B20.
2
2
Preface
Writing my master thesis for the last five months at KPMG Amstelveen was an instructive and important ending phase of my fifth year as student at the Rijksuniversiteit Groningen. With this master thesis I’m graduating for two masters; the MSc. Accountancy and the MSc. Finance. I have experienced that writing one master thesis for two masters is a challenge and that discipline is indispensable.
To my knowledge the major challenge is to find an actual and relevant subject that combines both academic fields. With the subject disclosure of derivatives I combine two important subjects in the area of finance and accountancy. In this time derivatives are a much discussed issue. With recent failures of large companies due to problems with derivatives the disclosure of these financial instruments has become essential. Derivatives can be used for hedging or speculating purposes. In this paper I try to find an answer on the question what the intention of Dutch listed derivative users is and if their derivative disclosures are influenced by the degree of derivative position. I try to acknowledge the field of finance and accountancy and give the reader additional insights on the field of derivative (disclosures).
By writing this thesis I want to thank a few people. First, I want to thank my finance supervisor dr. P.P.M. Smid, who has taken the roll of primary supervisor. He gave me constructive but straightforward feedback on my work and was always enthusiastic to help me to tackle problems. Secondly, my attendant at KPMG, Walter Berger, has educated me to never forget my intentions and to do not lose sight on the purpose of writing my master thesis. Finally, I want to thank my accountancy supervisor, prof. dr. R.L. ter Hoeven, who was responsible for the function as second supervisor.
I hope that the reader of my thesis acquires some knowledge in the field of derivatives and derivative disclosures and reads the remainder of this paper with enjoyment.
Groningen, June 2013 Matthijs Polstra
3
Table of Contents
1.
Introduction
5.
1.1 Intro 5. 1.2. Scientific Contribution 7. 1.3. Practical Relevance 8. 1.4. Structure 9.2.
Theoretical Framework
10.
2.1. Derivatives 10. 2.2. Agency Theory 10. 2.3. Hedging or Speculating? 11.2.4. Disclosure of Derivatives in the Financial Statements 13.
2.5. Impression Management 15.
2.6. Disclosure Regulation – The Impact of IFRS 7 17.
3.
Research Design
19.
3.1. Data 19. 3.2. Descriptive Statistics 20. 3.3. Derivative Usage 22. 3.3.1. Exposure 22. 3.3.2. Hedging or Speculating 23. 3.4. Quality of Disclosure 25. 3.5. IFRS 7 26. 3.6. Robustness Tests 27. 3.6.1. Derivative Usage 27. 3.6.2. Quality of Disclosure 29. 3.6.3. IFRS 7 29.4.
Results
30.
4.1. Derivative Usage 30. 4.2. Quality of Disclosure 30.4 4.3. IFRS 7 31. 4.4. Robustness Tests 32. 4.5.1. Derivative Usage 32. 4.5.2. Quality of Disclosure 32. 4.5.3. IFRS 7 33.
5.
Discussion
35.
5.1.1. Derivative Usage 35. 5.1.2. Quality of Disclosure 35. 5.1.3. IFRS 7 37. 5.2. Further Research 38.6.
Conclusion
40.
7.
Bibliography
42.
8.
Appendices
48.
5
1. Introduction
1.1. Intro
The primary objective of this paper is to examine the intention with respect to the use of derivatives for firms listed on the Amsterdam Exchange Index (AEX) and to investigate if Dutch derivative usage has an influence on the quality of their derivative disclosure (given new accounting regulation) in annual reports. These subjects are addressed for the period 2005 to 2012. In this timespan the financial world has made some striking movements and therefore the subject of derivative disclosures is particular interesting.
With the financial credit crisis in mind, recent scandals have made the valuation and disclosure of derivatives a major topic in today’s news headlines. According to Buffet (2002, p15): ‘Derivatives are financial weapons of mass destruction’. For example Vestia, a major housing corporation, had substantial amounts of derivatives in 2011 that were prone to high risks. The accountant of Vestia approved the annual report over 2011 and chose for the cost hedge accounting method to record the derivatives in the annual report, thereby not disclosing enough information about the degree of risks associated with the derivatives. Consequence was that the user of the annual report was not informed those potential risks. As a result of the decreased interest rate in 2011, Vestia became exposed to high debts with respect to their derivative position. It is argued that the accountant of Vestia was to blame due to possible mistake in choosing the right method of accounting. The accountant should have disclosed more information about the risks of the derivatives Vestia was facing.
Derivative usage has frequently been examined in empirical risk management literature. Chernenko and Faulkender (2011) find that companies use derivatives for both hedging and speculating purposes and conclude that regulators need to make sure that it is not easy to disguise speculation as hedging. Therefore it is important to assess the intentions of using derivatives.
Referring to the problems that have occurred with Vestia, there is a major discussion going on about the two different methods to account for derivatives. These are cost hedge accounting and fair value accounting.
Cost hedge accounting prescribes that derivatives should be recorded at historical cost in the firm’s annual report. Forwards, futures and swaps have usually zero historical cost resulting in no recognition. Derivatives which have zero historical costs are commonly referred to as symmetrical derivatives, implying that the value of the derivative can become positive (negative),
6 resulting in an asset (liability). In contrast, options have the property to be exercised during the time to maturity and give the holder of the option the decision to not exercise the option when it has a negative value (Hull, 2012). Acquiring an option has initial costs in contrast to symmetrical derivatives. The fair value method states that derivatives must be recognized either as assets or liabilities at fair value in the balance sheet and recognizes unrealized gains or losses due to changes in fair value in their income statements (Zhang, 2008).
Cost hedge accounting is not the solution to the lack of transparency of derivatives due to the fact that the value of derivatives will not be recognized and users of the financial statements cannot perceive the derivative position of a company (Thinggaard, 1996). Although fair value hedge accounting, as is prescribed under the Integrated Financial Reporting Standards (IFRS), is a more useful method to record derivatives (Melumad et al., 1999) it is also not the solution in making the derivative position of a company completely transparent to the user of the financial statements. The solution may not lay in the method of accounting but in a sufficient quality of disclosure of derivatives in the financial statements (Hoogendoorn and van Santen, 2012).
With recent failures of large listed companies, standard setters are pressured to place more attention on the quality of corporate reporting (Beretta and Bozzolan, 2008). Banking organizations and the accounting profession have taken a number of steps in recent years to improve the quality of disclosure regarding derivatives. One of these is the introduction of ‘IFRS 7 – Financial Instruments: Disclosures’ by the International Accounting Standards Board (IASB), which became effective as of the first of January 2007. IFRS 7 ensures that a company discloses specified information about its derivative usage whereby the user of the financial statements can assess the firm’s derivative positions and the risks the company is facing. Especially during the financial crisis it is useful to examine the impact of IFRS 7 and study if this financial reporting standard has contributed to derivative disclosures as is intended by the IASB.
The main research question can be formulated as follows: Is the quality of Dutch derivative
7 To find an empirical answer on the main research question this paper tries to find answers to the following sub questions:
1) Do Dutch listed firms on the AEX use derivatives for hedging or speculating purposes?
2) Does the use of derivatives of Dutch listed firms on the AEX has an influence on the
quality of disclosure of derivatives?
3) Has IFRS 7 an influence on the quality of disclosure of derivatives by Dutch listed firms
on the AEX?
To find an answer on the formulated questions I use a sample of 22 Dutch listed firms on the AEX during the period 2005 to 2012. To examine if Dutch firms hedge or speculate with derivatives I measure the firm’s exposure to the interest rate, the exchange rate and the commodity price. 141 financial statements are examined to examine the firm’s derivative position and the quality of the derivative disclosure. Robustness tests are implemented to validate the results.
I find no evidence to conclude in hedging or speculating purposes. In contrast to this absence of derivative intentions I find a negative relation between the firm’s derivative position and the quality of disclosure, indicating that firms with more derivatives construct less understandable derivative disclosures. In addition, IFRS 7 does not have an influence on the understandability of the derivative disclosures.
1.2. Scientific Contribution
In the year 2007 and onwards some highly publicized financial losses were attributed to derivative contracts (Bartlett, 2010; Inkinen et al, 2010). These contracts were held by large corporations and municipalities. With these failures, public attention has focused more on derivative usage by companies. This leads to the need for more relevant information and greater transparency about an entity’s exposures arising from derivatives and how those risks are managed (McDonnel, 2007).
8 Following previous literature, extensive research is done to the purpose of using derivatives in the U.S. (Chernenko and Faulkender, 2011; Guay, 1999; Géczy et al., 2007; Hentschel and Kothari, 2001). Most of this research contributes to the risk management literature in the way that firms use derivatives for hedging purposes (Bartram, 2013; Guay, 1999; Zhang 2008). Limitation of this affluence of studies is the primary focus on firms in the U.S. Contrary to the affluence of literature about why firms use derivatives for hedging purposes there is relatively little evidence on a firm’s speculating activities. A reason for this lack of speculating evidence could be explained through the assumption that firms do not speculate with derivatives. Referring to the publicized financial losses due to derivative contracts makes this argument not easily acknowledged. Other reasons could be that the information published in the financial statements is not adequate to measure the extent to which firms use derivatives for speculative purposes. This can be clarified by insufficient accounting rules, firms applying the rules in a mistaken way or firms fraudulently applying the rules (Géczy et al., 2007).
Especially in this time, where recent corporate scandals have occurred due to the lack of transparency of derivatives, disclosure of derivatives is essential. Previous literature has examined disclosures of derivatives (Bischof, 2009; Woods and Marginson, 2004). However there is little literature about the intentions of managers to disclose information in the Netherlands. Firm characteristics are addressed (such as firm size) but derivative usage itself has not been combined to derivative disclosure. Therefore the reasons of disclosure have not been solved. This leaves open answers to the question if firms with a high derivative position disclose more or less information. In addition, and unquestionably important, research that includes the disclosure of derivatives in the Netherlands during and after the financial crisis can be useful. Combining Dutch derivative usage with the appurtenant disclosure in the financial statements, results to my knowledge in an unexplored area.
1.3. Practical Relevance
Derivatives are seen as efficient tools for managing risk, but confusion about their effectiveness exists among the public. A reason for this confusion stems from the increasing complexity of financial instruments (Edwards and Eller, 1995). Referring to the public debate about derivative usage by corporations, additional insights about what intentions corporations have with derivatives will lead to a better understanding for the user of the financial statements, referred to
9 as the stakeholder. The stakeholder can benefit from the elucidation of the effect of derivatives on firms risk (Zhang, 2008). In addition to this, awareness of the possible relation between derivative disclosure and the derivative position of a firm can provide insights if Dutch firms are (less) transparent about their derivatives.
1.4. Structure
The remainder of this paper provides the details of my analysis and is organized in the following manner. Section 2 describes the main concepts such as derivatives purposes, agency theory and impression management. Furthermore, results of empirical studies are used to formulate the hypotheses and give guidelines to the present research. Section 3 describes the data and provides insights on the design of the research for the hypotheses. Section 4 gives the results and Section 5 discusses the findings. The conclusion is presented in Section 6.
10
2. Theoretical Framework
2.1. Derivatives
Organizations are in their normal way of operating exposed to several risks. Examples are interest rate risk, foreign exchange risk and commodity price risk. These risks can result in income volatility. Organizations can aspire to mitigate these risks to generate more stable profits (hedging) or to expose themselves to these volatilities to benefit from movements in the value of the asset (speculating). One way to accomplish the reduction of risks is to enter into derivatives (KPMG, 2011). Derivatives are financial instruments that derive their value from the value of an underlying asset (Hull, 2012). Derivatives are mostly categorized as futures, swaps, forward contracts and options (Curley and Fella, 2009). Especially in this time the use of derivatives is a phenomenon with increased awareness due to importance of risk management. Risk management is currently part of modern corporate financial culture (Bodnar et al., 2001).
Derivatives usage can be separated into hedging and speculating. Derivatives used for hedging intent to reduce the exposure to the movements of certain variables. Hedging can be seen as risk management that reduces return volatility (Hentschel and Kothari, 2001). Speculating with derivatives suggests speculating on movements in the value of the underlying asset (Géczy et al., 2007). ‘To speculate generally implies that the derivative position is undertaken with the primary intention of making a profit or increasing risk and is accompanied with an increase in return volatility’ (Hentschel and Kothari, 2001, p93).
2.2. Agency Theory
Agency theory is a much dispersed topic in the literature. Jensen and Meckling (1976) define an agency relationship as a contract between two or more persons, where the principal engages the agent to perform certain tasks. The principal or shareholder delegates decision making authority to the agent, the manager, who can follow the tasks he gets assigned or can function in his own interests. Both parties want the highest benefit and therefore opposite interests can appear. In such a situation it is reasonably to argue that the manager will not always perform in the way the shareholder prefers. The manager will only act in the best interest of the shareholder when a proper incentive for the manager exists (Smith and Stulz, 1985).
Agency problems are universal (Ross, 1973). Relating agency theory to the field of derivatives, managers can use derivatives for their own interests and not only for reducing risk.
11 Managers decide the derivative policy of the firm, shareholders do not. Shareholders, however, decide on the compensation package for the manager in order to maximize the shareholders’ wealth and firm value (Stulz, 1984). Managers obtain private information about the firm’s future expected cash flows and risk through their proximity to operating activities (Nagar et al., 2002) and therefore if the use of derivatives can be valuable for the shareholder. The shareholder does not have the information to assess the derivative usage of the manager, which can result in the situation that the manager utilizes derivatives not in the most preferable way for the shareholder.
2.3. Hedging or Speculating?
Financial theory suggests that corporate risk management can lead to an enhanced firm value only when there are market imperfections (Bessembinder, 1991; Graham and Rogers, 2002; Modigliani and Miller, 1958). When a perfect market exists, shareholders possess the required tools to create their preferred risk profiles. This results in no motivation to hedge. The remainder of this paper assumes market imperfections whereby hedging can decrease volatility and thereby reduce costs.
When managers are risk averse it is common that they require a higher return for bearing risk. Therefore, managers have an incentive to reduce risk which can be accomplished with hedging (Guay and Kothari, 2002). First, a firm can use derivatives to reduce the volatility of taxable income which can result in lower expected taxes. This can result in an increase in firm value. Second, derivatives used for hedging purposes can decrease the required risk premium (Smith & Stulz, 1985), which results in a higher firm value. At last, hedging with derivatives can lead to a reduction of transaction costs. By decreasing the possibility of bankruptcy, the projected cost of financial distress is reduced. This will lead to an increase in firm value (Bartram, 2012).
However, firms can have motivations to expose themselves to additional risks. In this interpretation the agency problem alters to the situation where the shareholder is the principal and the debt holder is the agent. The shareholder is the holder of a call option and the debt holder is the writer of a put option (Hentschel and Kothari, 2001). By increasing volatility, shareholders of leveraged firms relocate wealth from bondholders to shareholders whereby the shareholders benefit at the expense of debt holders (Myers, 1997). Such situations can arise in times of financial difficulty when shareholders possess an out-of-the-money option and therefore have an incentive to increase the riskiness of the firm (Nguyen and Faff, 2010). Increasing the riskiness of
12 the firm can be achieved with speculative trades, which are not based on any underlying exposure (Nguyen and Faff, 2010). Derivatives can be used for these speculative trades (Hentschel and Kothari, 2001).
As can be inferred from the previous part, derivatives can be seen as effective and efficient tools in the context of financial risk management, but managers can also use derivatives for speculative intentions. It is useful to know for what purpose listed firms use derivatives. What are the consequences for the companies’ stakeholders; is the company increasing or decreasing its risk position? (Bartram, 2013).
The subject of derivative usage by firms has been widely examined in the literature. Various researches have been performed with respect to the amount of risk hedged by derivatives. Although data on derivatives usage have become available in the last two decades, detailed empirical evidence on the effects of derivative usage is widely mixed (Bartram, 2011)
Bartram (2013) examines derivative usage by firms in the U.S and concludes that firms primarily use derivatives to reduce risks. He compares pre-hedging exposures with post-hedging exposures to movements in an interest rate, exchange rate and a commodity price index and finds that pre-hedging exposure is higher. Hentschel and Kothari (2001) find that firms in the U.S. manage their exposures with large derivatives positions. These positions are primarily taken with the intention of hedging. Nonetheless, compared to firms that do not use derivatives, firms that use derivatives exhibit few measurable differences in risk that are related with the use of derivatives. Bodnar et al. (2001) perform a comparative study for the usage of derivatives for firms in the U.S. and in the Netherlands. They find that the Netherlands is more open with respect to their derivative usage and hedge more financial risks in comparison to firms in the U.S. The results of Zhang (2008) indicate that firms in the U.S. have lower exposures when firms use derivatives. Guay (1999) find similar results. He concludes that derivative use is more prevalent when U.S. firms bear high exposures to interest rate risk, exchange rate risk and commodity price risk. Firms with derivatives have lower values of risk, implicating the use of hedging. Tufano (1996) finds the same results for his specific sample in the Gold mining industry in North America.
Similar to previous literature (Bartram, 2013; Bodnar et al., 2011; Guay and Kothari, 2002; Zhang, 2008) I study the exposure to changes in the interest rate, the exchange rate and the commodity price. To examine if Dutch listed companies in the Netherlands use derivatives for
13 hedging or speculating activities, I study the possible relation between the firm’s exposure and derivative position. A firm with low exposures and a high derivative position indicates hedging, whereas a firm with high exposures and a high derivative position indicates speculating (Hentschel & Kothari, 2001). The following hypotheses can be formulated:
H1a: There is no relation between the derivative position of Dutch listed firms and their exposure to the interest rate, the exchange rate and the commodity price.
H1b: There is a relation between the derivative position of Dutch listed firms and their exposure to the interest rate, the exchange rate and the commodity price.
2.4. Disclosure of Derivatives in the Financial Statements
As can be inferred from section 2.2, an information asymmetry is present between the manager and the shareholder. By eliminating this information asymmetry with respect to the firm’s derivative position the manager has to disclose information regarding derivatives in the financial statements to enable the shareholder to verify the agent’s actions.
Managers disclose information in a variety of means and approaches such as financial reports, footnotes and other regulatory filings (Healy and Palepu, 2001; Roudaki, 2012). Disclosure can be separated in numerical and narrative disclosure. Numerical disclosure encompasses only numbers where narrative disclosure implies the clarification of quantitative financial measures (Lev and Zarowin, 1999; Cole and Jones, 2004). Narrative disclosures are an important aspect of the notes to the financial statement. They are viewed as the essential element in attaining the desired change in the quality of financial reporting and serve the information requirements of the market. Narrative disclosures provide the information required for corporate transparency and accountability (Beattie et al., 2004).
Other studies are in line with this. Gassen and Schwedler (2008) find that managers and investors treat footnote information in financial statements as valuable. Baumann and Nier (2004) and Chipalkatti (2005) conclude that there is a negative relation between a firm’s disclosures in general and the volatility of its equity returns. This implies that additional disclosure is useful for investors. This is in line with Gebhardt et al. (2004) who find that banks suggest new disclosure
14 standards that require a higher quality, implying that the current quality of disclosure is not perfect.
To assess the level of information with respect to derivatives in the financial statements, literature shows the concept of quality of disclosure. Quality of disclosure is a wide phenomenon and previous literature demonstrates several approaches to measure the quality of disclosure. Although it is comprehensible to use a single qualitative indicator to measure disclosure, the use of indices is more widely used in the literature (Bravo et al., 2009). A disclosure index has the property to proxy one or more information items. The most widely used information item is coverage or length of the disclosure (Cooke, 1989; Hossain et al., 2005; Singhvi and Desai, 1971). Quality of disclosure can be separated in a quantitative and a qualitative aspect. Several methods have been used to measure the quantity and quality of disclosure in the financial statements (Marston and Shrives, 1991).
The quantity aspect can be measured by the amount of words provided in the disclosure. However, by disclosing more information the firm will not create an enhanced transparency unless the readability of the disclosures increases (Linsey and Lawrence, 2006). Therefore the quality aspect measures the concept of understandability. ‘Understandable’ means that the user of the financial statements has a reasonable knowledge to comprehend the information. It measures if the user is able to perceive the significance of the information. To quantify understandability it is common to use readability formulas. A readability formula is an ex ante, quantitative method that gauges if a certain population can read the information (Courtis, 1986). The most common readability measures are the Flesch Reading Ease (Flesch Score) and the Fog Index (Li, 2008) which are both used in this research.
The Flesch Score measures the amount of syllables and the number of words per sentence (Linsey and Lawrence, 2006). It can be computed in the following manner:
Where FLESCHjt is the readability or referred to as the understandability of the derivative
disclosures for firm j, Wljt is the number of syllables per 100 words of derivative disclosure for
firm j and Sljt is the average number of words per sentence in the derivative disclosure for firm j.
15 An alternative for the Flesch Reading Ease is the Fog Index (Li, 2008). The Fog Index measures the complexity as a function of syllables per word and words per sentence. ‘The index indicates the number of years of formal education a reader of average intelligence would need to read the text once and understand that piece of writing with its word-sentence workload’ (Li, 2008, p225). It is calculated as follows:
Where FOGjt is the Fog Index for firm j, Wsjt is the amount of words per sentence in the
derivative disclosure for firm j and Cwjt is the % of complex words for firm j. Complex words are
words with three or more syllables. Table A.1.2 displays the different Fog Index categories.
2.5. Impression Management
Existing research indicates the relation between the characteristics of disclosure and the attention given from analysts. Lehavy et al. (2011) find that companies with a higher quality of disclosure get more attention from analysts in comparison to firms with a lower quality of disclosure. As negative organizational outcomes give rise to conflicts of interest between managers and shareholders, managers are prompted to manipulate outsiders’ perceptions of and decisions on financial performance and forecasts (Aerts, 2005). This in line with the results of Bloom (2002), who indicates that managers obfuscate information when the current firm performance is low.
This opportunistic managerial behavior has given rise to the obfuscation hypothesis (Courtis, 1998), which implies that managers are not neutral in presenting accounting narratives (Sydserff and Weetman, 1999). Managers tend to hide failures and emphasize successes (Adelberg, 1979). This sort of behavior is explained by impression management theory. Impression management can be defined as attempts ‘to control and manipulate the impression conveyed to users of accounting information’ (Clatworthy and Jones, 2001, p311). Managers use financial statements to ‘strategically manipulate the perceptions and decisions of stakeholders’ (Yuthas et al., 2002, p142). Relating impression management to derivative usage can provide insight if managers are transparent about their use of derivatives.
Lin et al. (2009) find that firms which use derivatives in the U.S. exhibit significantly less information asymmetry in comparison to firms without derivatives. This can be explained due to
16 the fact that those firms use derivatives for hedging purposes and therefore have less unpredicted shocks in their income volatility. Lower income volatility leads to a higher transparency (DeMarzo and Duffie, 1995). In addition, firms with derivatives, perform better in the long run compared to firms that do not use derivatives (Lin et al. 2009). It can be argued that this higher transparency and better performance would, based on impression management, result in longer and a better understandability of derivative disclosures.
While literature shows that derivatives can be used for an effective hedging policy it is also argued that derivatives are not the only solution for mitigating risk. There are alternative and less costly risk management activities that can be used as a substitute for the use of derivatives. One of these activities could be to maintain a higher liquidity. This would result in a lower insolvency risks through lower dividend payouts and a higher current ratio. With those results firms would be less forced to hedge their risks with derivatives (Amihud and Murgia, 1977). In this manner companies can choose other hedging instruments. By entering only in derivatives firms could deliver the wrong signal to investors by creating the assumption that the firm uses derivatives for speculating purposes. As stated before, firms are associated with a greater return volatility when they use derivatives for speculative purposes. Dobler (2008) argues that firms are in general reluctant to disclose information with respect to the risk the firm is confronting, because this could discourage the user of the financial statements. Therefore it can be argued that firms with derivatives are less transparent about their derivatives activities and disclose less information. In other words, the user of the financial statement should be less capable of understanding the information disclosed about derivatives. The aforementioned theories can be formulated as follows:
H2a: There is no relation between the derivative position of Dutch listed firms on the AEX and the quality of derivative disclosures in the financial statements.
H2b: There is a relation between the derivative position of Dutch listed firms on the AEX and the quality of derivative disclosures in the financial statements.
17
2.6. Disclosure Regulation – The Impact of IFRS 7
Regulation implemented to assure the transparency of derivatives is made by the IASB. As of the first of January 2007 the IASB implemented ‘IFRS 7 – Financial Instruments: Disclosures’ that ensures that firms ‘disclose quantitative and qualitative information about (a) the significance of financial instruments for the entity’s financial position and performance; and (b) the nature and extent of risks arising from financial instruments to which the entity is exposed during the period and at the end of the reporting period, and how the entity manages those risks’ (KPMG, 2012, p1793). In addition, IFRS 7 mandates the description of a firm’s risks, including but not limited to credit risk, liquidity risk and market risk. This reporting standard applies in the Netherlands to all listed entities and financial institutions. The reason for the introduction of this regulation is to enhance financial statements users’ understanding of financial instruments.
There is extensive literature regarding the impact of accounting regulation on the quality of disclosure. Woods and Marginson (2004) assess the quality of the derivative disclosures in the U.K. under Financial Reporting Standard (FRS) 13. They discuss the extent to which disclosure informs the user of the financial statements to 1) assess a company’s attitude to risk and 2) measure the financial impact of its derivative usage. They conclude that the user of the financial statements cannot get effortlessly any sense of the entity’s appetite for risk and total level of exposure with respect to derivatives.
Bischof (2009) examines the effects of IFRS 7 on bank disclosure in Europe and concludes that the quantity and quality of derivative disclosure have generally increased in the financial statements. This is in line with the results of Ernst & Young (2008) and Nelson et al. (2008). They conclude that the disclosures of the largest banks in Europe have increased in general due to the introduction of IFRS 7. Hodgeon and Wallace (2008) find that the disclosures of banks are improved but that they are not made to their own risk profile in the first year of the adoption of IFRS 7. They expect that this will evolve in the years after 2007.
Based on these findings derivative disclosure of Dutch listed firms should be longer and of a higher quality due to the presence of IFRS 7. In contrast, companies that use derivatives argue that regulators’ concerns about the dangers of using derivatives (speculation) are misplaced. They argue that the direct and indirect costs of unnecessary and disproportionate regulation will reduce the benefits of using derivatives (Hentschel and Kothari, 2001). Therefore
18 additional regulation could have a negative impact on the quality of disclosure. The previous described theories can be formulated as follows:
H3a: IFRS 7 has no influence on the quality of disclosure of derivatives in the financial statements.
H3b: IFRS 7 has a positive influence on the quality of disclosure of derivatives in the financial statements.
19
3. Research Design
3.1. Data
The sample consists of 22 Dutch firms that are part of the Amsterdam Exchange Index (AEX). The AEX is chosen because previous literature shows that large firms are more likely to use derivatives (Graham and Rogers, 2002; Nance et al., 1993). The firms are used for the financial years 2005 to 2012 which in principle results in 200 firm year combinations. There are a number of screening criteria. Firstly, several firms are not listed for the entire eight years on the AEX. Those firm years are excluded from the sample, because firms which are not listed are not mandatory to pursue IFRS. Secondly, two financial reports over the year 2012 (DE Master Blenders N.V. and Royal Imtech N.V.) are not yet available at the time of writing. Thirdly, some financial reports are secured, which means that they cannot be used for copying information. Finally, financial institutions and insurance companies (ING Groep N.V. and AEGON N.V.) are excluded from the sample due to the fact that they are more likely to have derivatives (Hentschel and Kothari, 2001). Besides, it is arguable that IFRS 7 has especially a strong influence on derivative disclosures by financial institutions, because financial instruments, including derivatives, are about 90% of financial institutions’ total assets and liabilities (Bischof, 2009). Therefore financial institutions are most likely to have the strongest reaction on the introduction of IFRS 7. Applying these criteria results in a total sample of 141 firm year combinations and is categorized as an unbalanced panel. Table A.2.1 displays the 25 firms listed on the AEX and the sample firms used for this paper.
The notes to financial statements and, when necessary, the chapter risk management are used to acquire data about a firm’s derivative position and provide the derivative disclosure (see section 3.3 for further details on disclosure). Both are collected by hand. Additional firm characteristics, such as stock price, firm size, the market to book ratio, the return on equity and leverage are included as control variables and are collected from DataStream. Data with respect to macro economic variables such as the Euribor interest rate, the Euro exchange rate index and the Standards & Poor Goldman Sachs Commodity price Index (S&P GSCI) originate from DataStream. Table A.3.1 presents an overview of all variables used in this paper and how they are measured.
As is used by Bartram (2013) data from DataStream is provided at daily frequency and is transformed (by taking the arithmetic mean) to annual data when required. To reduce the
20 influence of outliers all continuous variables are winsorized. This implies that all observations with a three times standard deviation above and underneath the mean are set equal to these boundaries.
I perform multiple Ordinary Least Squares (OLS) regressions to test the constructed equations. When I use data with both cross-sectional and time series dimensions, a pooled OLS regression is used. All equations are tested for the potential that the variance of the errors is not constant over time, which is also acknowledged as heteroscedasticity. In order to examine if heteroscedasticity is present, I perform for each equation a White’s test (output is not presented in the paper). When the errors are heteroscedastic, I correct the standard errors using heteroscedastic-consistent standard error estimates.
3.2. Descriptive Statistics
Table A.3.2 presents the descriptive statistics for all variables used in this paper. As can be seen are the Flesch Scores between the 11.431 and the 54.440. This implicates that the understandability of the derivative disclosures for firms listed on the AEX is between the difficulties ‘very difficult’ and ‘fairly difficult’ (based on table A.1.1). The Fog Index varies between 11.440 and 21.690, implying that the difficulty of the disclosures lies between ‘general circulation magazines’ and ‘technical books’ (based on table A.1.2). The derivative positions vary between the -2650 million (liability) and 1751 million (asset) euro. The normalized derivative position is corrected for firm size and varies between the values of -109.570 million and 72.030 million. It can be seen that the rescaling has narrowed the interval. The summed normalized derivative position has a wider interval. It ranges from -6.895 million to 3085.406 million.
The Jarque-Bera values give an indication if the variables are following a normal distribution. As can be observed are all variables, excluding the variables for quality and quantity of disclosure, the summed normalized derivative position and the returns of the interest rate, the exchange rate index and the commodity price index significant at the 1% level. This means that the null hypothesis of normality is rejected and that the significant variables do not follow a normal distribution. Variables that are winsorized and do not follow a normal distribution are insurmountable (Brooks, 2008) and therefore I use OLS regressions as intended.
21 Table A.3.3.1 and table A.3.3.2 present the correlation matrices for all variables used in this paper. The first table presents the correlation matrix which are obtained with a sample size of 141 firm year combinations. The second table displays the correlation matrix for equation 4 where one observation represents the average of the years 2005 to 2012. This results in a sample of 22 observations (see section 3.3.2 for further details).
There are problems with multicollinearity if one independent variable has a correlation beneath the -0.7 or above the 0.7 with one another. Table A.3.3.1 displays one problem with multicollinearity. The variable derivative position and the normalized derivative position are multicollinear. Therefore these variables are not used together in the same regression. Table A.3.3.2 presents a problem with multicollinearity between the exposure to the exchange rate index and the commodity price index. These variables are intended to be used in the same regression. Therefore, I calculate the variation inflation factor (VIF) to measure in what degree the variance of the estimated regression coefficient is increased because of multicollinearity. Based on these results, I conclude no increase in regression coefficient and therefore both variables can be used in the same regression.
Table A.3.3.1 displays negative correlations (at the 5% level) between the returns on the interest rate, the exchange rate index and the commodity price index and the stock return. Apparently, an increase (decrease) these three variables leads to a decrease (increase) in a firm’s stock return. This gives evidence that the firm’s stock return captures movements in the preceding variables.
The Fog Index correlates negatively with the Flesch Score (at the 5% level). This is as expected, due to the fact that both measures have an opposed scale. The correlations between the derivative positions (DP, NDP and NDP (A+L)) and the Flesch Score have a negative correlation. It could be inferred that firms with a higher derivative position construct less understandable derivative disclosures. Leverage shows the same relation (at the 5% level). Firms with more leverage construct less understandable derivative disclosures. There are less correlations between independent variables and the amount of words. NDP (A+L) shows a negative correlation (at the 5% level), implying that firms with relative more derivatives use less words in their derivative disclosure. Leverage shows the opposite effect (at the 5% level). Firms with more leverage use more words in their disclosure.
22 Table A.3.3.2 displays no significant correlations between independent variables and dependent variables.
3.3. Derivative Usage
To measure if a firm, listed on the AEX, is using derivatives for hedging or speculating purposes I combine two separate OLS regressions. Both regressions are described below.
3.3.1. Exposure
The first regression is used to estimate the exposure of a firm to the interest rate, the exchange rate index and the commodity price index. This method is used in previous literature (see, e.g., Bartram, 2013; Bodnar et al., 2011; Guay and Kothari, 2002; Zhang, 2008). The regression is performed for each firm separately over the period 2005 to 2012, which results in 22 OLS regressions. Daily data is used, as used by Bartram (2013), which implies that each firm has 2080 observations (8 years with 5 days a week). Performing the regressions I obtain 22 interest rate coefficients, 22 exchange rate index coefficients and 22 commodity price index coefficients, which represent the exposure for the years 2005 to 2012. The coefficients of the variables resulting from the OLS regression are a measure of exposure to the Euribor interest rate, the Euro exchange rate index and the S&P GSCI for the period 2005 to 2012. The equation can be stated as follows:
Where Rjt is the daily stock return in % of firm j, Rirt is the daily return in % on the three month
Euribor interest rate, Reit is the daily return in % on the exchange rate index, and Rcit is the daily
return in % on the commodity price index.
In this regression I use the daily stock return as dependent variable. Using a firm’s stock return as an aggregate measure of all relevant information is prescribed by the assumptions of an efficient capital market. A market in which prices always completely reflect available information is called ‘efficient’ (Fama, 1970). Different risk components of a firm are hard to decompose but using a firm’s stock return overcomes this problem as it can be seen as a measure of net exposure (Bartram, 2013).
23 The three month Euribor interest rate is used. This because a short term rate is a typical benchmark for short term and floating rate instruments such as commercial paper, floating rate bonds and interest rate swaps (Hentschel and Kothari, 2001). The daily three month Euribor interest rate is used for calculating the daily returns in %.
It is problematic to identify which currencies underlie exchange rate derivatives (Hentschel and Kothari, 2001). Therefore it is desirable to use an exchange rate index. In this paper the J.P. Morgan traded weighted euro exchange rate index is used which equates the year 2000 to an index of 100. It is calculated as a weighted average of exchange rates for domestic versus foreign currencies. The weight for each foreign country is equal to its portion in trade. The daily index rates are used for calculating the returns in %.
The Standards & Poor’s Goldman Sachs Commodity Index (S&P GSCI) is used for estimating the exposure to the commodity price. The S&P GSCI contains the 24 leading commodities from al commodity sectors and is a good representation of the combined commodity price (Goldman Sachs, 2013). It is a world-production weighted index based on the average production of each commodity in the index. Therefore it is a measure of investment performance and an economic indicator. The daily rates are used for computing the daily returns in %.
3.3.2. Hedging or Speculating
The second regression estimates the relationship between the derivative position of a firm and the exposure to the interest rate, the exchange rate index and commodity price index respectively (hypotheses 1) which is also performed by Hentschel and Kothari (2001). The firm’s derivative position is measured relative to firm size. Therefore I divide a firm’s total position in derivatives by the natural logarithm of total assets. A similar measure of a firm’s derivative position is also used by Hentschel and Kothari (2001). The normalized position in derivatives is calculated by netting the derivative position in assets and the position in liabilities for each year (a different method is used as robustness test). These values are generally provided in the financial statements in the note ‘Derivatives’ or ‘Financial Instruments’. For these positions the fair value amounts are used as this is the relevant valuation method prescribed under IFRS.
Figure A.3.4 displays the relation between the averaged absolute netted value of the derivative positions and the averaged normalized netted derivative positions for firms on the AEX during the period 2005 to 2012. As can be seen are the open dots more stable due to the fact
24 that these positions are controlled for firm size. E.g. SBM Offshore N.V. averaged derivative position is positive (609 million) for the sample period, but controlled for firm size SBM Offshore N.V. has approximately the same proportion of derivatives in comparison with other firms listed on the AEX. Therefore the normalized derivative position is a useful measure to see if firms have relative high derivative positions.
The exposures are measured with equation 1 as stated above. To examine the relation between exposure and derivative position I include several firm characteristics as control variables. These are firm size, the market to book ratio, the return on equity and leverage. Firm size is taken as control variable because literature shows that larger firms have more resources to acquire personnel with expertise in the field of derivatives and therefore are more likely to use derivatives (Shiu and Moles, 2010). The market to book ratio is involved due to the fact that firms with growth opportunities (high ratio) are more likely to use derivatives. This because derivatives that are used for hedging purposes can reduce the underinvestment problem associated with investment opportunities in the presence of financial constraints (Géczy et al, 1997). The return on equity is included because higher returns could forgo the opportunity to hedge risks with derivatives and would result in a lower derivative position. Leverage is taken as control variable due to the fact that firms with high leverage are more likely to use derivatives, because their interest rate risk increases as a function of leverage (Graham & Rogers, 2002). Table A.3.1 describes how these variables are measured. All variables are averaged by taking the arithmetic mean for the years 2005 to 2012. This because the exposures are estimated for this period (equation 3). Therefore I use a sample of 22 observations. The pooled OLS regression measures the following equation:
Where NDPj is the normalized derivative position of firm j, Eirj is the exposure to the interest rate
for firm j, Eeij is the exposure to the exchange rate index for firm j, Ecij is the exposure to the
commodity price index for firm j, LNTAj is the firm size of firm j, MTBj is the market to book
25 firm uses derivatives for hedging purposes I would predict that lower than average exposures are associated with above average derivative holdings. In this case the relationship is negative implying that the estimated coefficients ( are negative. In contrast, speculating would be associated with above average exposures and above average derivative holdings (Hentschel and Kothari, 2001) where one would predict positive estimated coefficients. Both predictions are captured in hypothesis 1b.
3.4. Quality of Disclosure
The notes to the financial statements give the user information about the firm’s derivatives and to which risks the company is exposed. The notes to the financial statements are therefore examined. Whereas IFRS 7 requires extensive disclosures about risk management, it does not prescribe a specific format or separate risk report (Nelson et al., 2008). Therefore, every firm discloses information in its own way, for example in separate notes or elsewhere in the financial statement. Consistency must be remained and therefore the financial statements are scanned to (a) the information about the firm’s derivative position, mostly disclosed in the note to the balance sheet ‘Financial Instruments’ or ‘Derivatives’ and (b) the risks the company is facing, disclosed in the note to the balance sheet ‘Financial Instruments and Risks’ or ‘Financial Risk Management’, but sometimes included in the chapter ‘Risks’ prior to the notes to the financial statements. In cases where the relevant information is provided in the firm’s risk chapter the information is collected from this section.
As stated in section 2.4 the quality of disclosure can be separated into a quantitative and a quality aspect. The quantity aspect measures the amount of words of the disclosure. The qualitative aspect measures the understandability (or readability) of the disclosure. The quantitative aspect is measured by the amounts of words provided in the notes to the financial statements with respect to derivatives combined, if necessary, with the amounts of words provided in the firm’s risk management chapter. A similar approach is used by Bischof (2009). The quality of disclosure is measured by the concept of understandability and is performed by the Flesch Score, as is used by Li (2008).
To measure the relationship between the amount of derivatives for firms listed on the AEX and the quality of their derivative disclosure (hypotheses 2) I perform two OLS regressions. The quality of disclosure is measured with the amount of words in the derivative disclosure and
26 the Flesch Score as dependent variables respectively. The normalized level of derivatives is used as independent variable. I include several firm characteristics as control variables. These are firm size, the market to book ratio, the return on equity and leverage. Firm size should be included because literature shows that large firms are more likely to disclose more information in general and have more complex financial statements (Courtis, 1995a, 1995b; Li, 2008). In addition the market to book ratio is used as a potential element of the readability of financial statements. Firms with high market to book ratios (growth firms) are considered to have more complex and uncertain business models and therefore could have more complex financial statements (Li, 2008). The return on equity is included due to the fact that firms with higher profits are supposed to signal this to the outside world with longer and better readable disclosures. Leverage is involved due to the fact that firms with more leverage are more likely to accelerate earnings and disclose less understandable information (Healy and Palepu, 2001). Besides, firms with more leverage are thought to have a more complex financial structure and therefore a more complex financial report (Li, 2008). The pooled OLS regression estimates the following equations:
Where DISCjt is the amount of words provided in the derivative disclosure for firm j, FLESCHjt is
the reading ease or referred to as the understandability of the derivative disclosures for firm j,
NDPjt is the normalized derivative position of firm j, LNTAjt is the firm size for firm j, MTBjt is
the market to book ratio for firm j, ROEjt is the return on equity for firm j and LEVjt is the
leverage for firm j.
3.5. IFRS 7
To measure the influence of IFRS 7 on the quality of disclosure of derivatives, as stated under hypotheses 3, it is necessary to compare the quality of disclosure of derivatives before and after the introduction of the reporting standard. As previous specified IFRS 7 is introduced at the first of 2007 implying that financial statements as of the year 2007 and onwards are composed with this reporting standard. This means that it is arguable that the quality of derivative disclosure of
27 the years 2005 and 2006 could be significantly different in comparison with the year 2007 and onwards.
To measure if there is a difference in quality of disclosure I use the quantitative and qualitative aspect. This results in two different OLS regressions with the amount of words and the Flesch Score as dependent variables respectively. I create a dummy variable to measure for year effects. This variable takes the value of ‘0’ for the years 2005 and 2006 and ‘1’ for the years 2007 to 2012. The same control variables as in section 3.4 are involved in both pooled OLS regressions. The equations can be formulated as follows:
Where DISCjt is amount of words provided in the derivative disclosure for firm j, FLESCHjt is the
reading ease of the derivative disclosures for firm j, YRjt is the dummy variable for firm j, LNTAjt
is the firm size for firm j, MTBjt is the market to book ratio for firm j, ROEjt is the return on equity
for firm j and LEVjt is the leverage for firm j.
3.6. Robustness Tests
I examine a variety of alternative specifications and undertake additional tests to validate the robustness of the empirical findings. For each set of hypotheses a robustness tests is performed. This is achieved by substituting the independent or dependent variable in each equation by another similar but different measured variable.
3.6.1. Derivative Usage
The OLS regression for testing hypotheses 1 (equation 4) is performed again but with another measure for the normalized derivative position. As stated under section 3.3.2 the normalized derivative position is calculated by netting the derivative assets and derivative liabilities. Another technique is that the derivative position of firm j consists of the summed derivative assets and derivative liabilities. With this method, liabilities are not be subtracted from assets but added. The first robustness test can be presented as follows:
28 Where NDP(A+L)j is the normalized summed derivative position of firm j, Eirj is the exposure to
the interest rate for firm j, Eeij is the exposure to the exchange rate index for firm j, Ecij is the
exposure to the commodity price index for firm j, LNTAj is the firm size of firm j, MTBj is the
market to book ratio for firm j, ROEj is the return on equity for firm j and LEVj is the leverage for
firm j.
A second robustness test for derivative usage (equation 4) is performed to test for the potential of endogeneity of derivative use. Previous literature indicates that there is a possibility that the estimated exposures to the three different variables are not only a function of the derivative position, but that they are influenced by the derivative position of the firm as well (see, e.g., Bartram et al., 2011; Bartram, 2013; Graham and Rogers, 2002; Hentschel and Kothari, 2001). In this approach a firm could have a higher (lower) exposure as a result of a lower (higher) derivative position. The robustness tests can be estimated with a Two-Stage Least Squares (TSLS) method. To test for the potential of endogeneity I formulate three equations with the exposures as dependent variables:
Where Eirj is the exposure to the interest rate for firm j, Eeij is the exposure to the exchange rate
index for firm j, Ecij is the exposure to the commodity price index for firm j and NDPj is the
29
3.6.2. Quality of Disclosure
To execute a robustness test on the possible relation between the value of the derivative position and the quality of disclosure (hypotheses 2), I substitute the summed normalized derivative position for the normalized derivative position. This can be quantified into the following equations which can be estimated with an OLS regression:
Where DISCjt is amount of words provided in the derivative disclosure for firm j, FLESCHjt is the
reading ease of the derivative disclosures for firm j, NDP(A+L)jt is the summed normalized
derivative position for firm j, LNTAjt is the firm size for firm j, MTBjt is the market to book ratio
for firm j, ROEjt is the return on equity for firm j and LEVjt is the leverage for firm j.
3.6.3. IFRS 7
As explained in section 2.3 another measure for the quality aspect of quality of disclosure is the Fog Index. The Fog Index measures the understandability of the disclosure with another scale in comparison with the Flesch Score. A higher Flesch Score indicates a better understandability or readability whereas this is the case with a lower Fog Index (see table A.1.1 and A.1.2). Therefore I expect opposed results with the Fog Index in comparison with the Flesch Score.
To validate the results on the third hypotheses, the OLS regression (equation 8) is performed again with the Fog Index as dependent variable. This robustness test can be presented as follows:
Where FOGjt is the understandability of the derivative disclosures of firm j, YRj is the dummy
variable for firm j, LNTAjt is the firm size for firm j, MTBjt is the market to book ratio for firm j,
30
4. Results
4.1. Derivative Usage
To estimate the different coefficients I perform the first OLS regression. Table A.4.1 presents the coefficient estimates with the appurtenant p-values for the firm’s stock return as dependent variable (equation 3). The coefficients are presented for the changes in the interest rate, the exchange rate index and the commodity price index. As can be seen are there only two interest rate coefficients significant (at the 5% level). There are 14 exchange rate index coefficients significant (at the 5% level) and 22 commodity price index coefficients significant (at the 5% level). Therefore it can be concluded that firms listed on the AEX are most exposed to differences in the commodity price index which is measured by the S&P GSCI.
A White’s test is conducted for equation 4. The results give no evidence for heteroscedasticity. Table A.4.2 presents the OLS regression output for this equation. The dependent variable is the normalized derivative position. The coefficients estimated in table A.4.1 are used for this regression. The exposure to changes in the interest rate, the exchange rate index and the commodity price index are not significant (at a 5% level). This means that the normalized derivative positions for firms on the AEX are not influenced by these exposures. The return on equity is significant (at the 5% level) with a negative coefficient, which means that firms with a higher return on equity have a lower normalized derivative position. Firm size, the market to book ratio and leverage are not significant. The explanatory power of the model (R2) is 0.492.
Based on these results, there is no relation between the firm’s exposures and the normalized derivative position. Therefore hypothesis 1a cannot be rejected.
4.2. Quality of Disclosure
The White’s test for equation 5 does not show any evidence for heteroscedasticity. Table A.5.1 displays the OLS regression results for this equation. The dependent variable is the amount of words in the derivative disclosure. The normalized level of derivatives is not significant (at the 5% level). This means that we cannot reject hypothesis 2a for the quantity aspect of disclosure. Firm size is significant (at the 5% level) and the coefficient estimate is positive, implying that larger firms listed on the AEX use significantly more words in their derivative disclosures. The market to book ratio, the return on equity and leverage are not significant (at the 5% level). The explanatory power of the model is 0.071.
31 Equation 6 is tested for heteroscedasticity with the White’s test. There is no evidence for heteroscedasticity. Table A.5.2 displays the OLS regression output for this equation. The dependent variable is the understandability of the derivative disclosure which is measured with the Flesch Score. The results show that the normalized derivative position is significant (at the 1% level). The coefficient is slightly negative implying that firms with relative more derivatives have significant less understandable derivative disclosures. Therefore hypothesis 2a can be rejected for the quality aspect of derivative disclosures. Leverage is significant (at the 5% level) with a negative coefficient. This implies that firms with more leverage have a significant less understandable derivative disclosure. Firm size, the market to book ratio and the return on equity are not significant (at the 5% level). The explanatory power of the model is 0.145.
There is no relation between the firm’s derivative position and the amount of words used in the derivative disclosure, but a negative relation exists between the firm’s derivative position and the understandability of the derivative disclosures. Therefore, hypothesis 2a cannot be rejected for the quantity aspect, but can be rejected for the quality aspect.
4.3. IFRS 7
The White’s test for equation 7 does not lead to any evidence for heteroscedasticity. Table A.6.1 presents the results of the OLS regression for this equation. The amount of words provided in the derivative disclosure is the dependent variable. As can be seen are the dummy variable and firm size significant at the 5% level. The dummy coefficient is positive (1107.050) meaning that from the years 2007 and onwards 1108 more words are used in the derivative disclosures for firms listed on the AEX in comparison to the years 2005 and 2006. This means that the quantitative aspect of quality of disclosure is improved with the introduction of IFRS 7. This favors the rejection of hypothesis 3a. The coefficient of firm size is positive, implying that larger firms listed on the AEX use significantly more words in their derivative disclosures than smaller firms. This result is in line with the results under 4.2. The variables market to book ratio, the return on equity and leverage are not significant (at the 5% level). The explanatory power of the model is 0.278. Figure A.6.2 presents the relation between the amounts of words provided in the derivative disclosure and years. It can be observed that the line peaks in the year 2007.
The White’s test for equation 8 shows evidence for heteroscedasticity. Therefore I use heteroscedasticity-consistent standard error estimates for the OLS regression. Table A.6.3
32 presents the output for the OLS regression for this equation. The dependent variable is the Flesch Score and the same independent variables are used as with the regression for quantity of disclosure. The dummy variable is not significant (at the 5% level), meaning that the year 2007 and onwards are not accompanied with better readable derivative disclosures in comparison to the years 2005 and 2006. Based on this result hypothesis 3a cannot be rejected (in contrast to the quantitative aspect of quality of disclosure). Leverage is significant (at the 5% level) with a slightly negative coefficient. This means that firms with more leverage have significant less understandable derivative disclosure. The variables firm size, market to book ratio and the return on equity are not significant (at the 5% level). The explanatory power of the model is 0.114.
4.4. Robustness Tests
4.4.1. Derivative Usage
The White’s test for equation 9 indicates that the errors are heteroscedastic. Therefore I use heteroscedasticity-consistent standard error estimates. These standard errors are used and the OLS regression is performed. Table A.7.1 presents the regression output for the robustness test (equation 9). The dependent variable is the summed normalized derivative position. The independent variables are the same in section 4.2. There is no difference in rejection of hypothesis. The exposures to the interest rate, the exchange rate index and the commodity price index are still not significant indicating that hypothesis 1a cannot be rejected. The explanatory power increases from 0.492 to 0.596.
Table A.7.2 provides the results for the Two-Stage Least Squares regression to test for the potential of endogeneity (equation 10, 11 and 12). As can be seen are none of the estimated coefficients for the normalized derivative position significant (at the 5% level) which results in the rejection of hypothesis 1a. The explanatory power of the three tests is between the 0.014 and 0.052.
4.4.2. Quality of Disclosure
The White’s test for equation 13 does not present any evidence for heteroscedasticity. Table A.7.3 displays the OLS regression results this for the robustness test on this equation (hypotheses 2). The dependent variable is the amount of words provided in the derivative disclosure. As can be seen is the summed normalize derivative position not significant (at a 5% level). This is in line
