Value relevance : related to banks’ lobbying behavior or not? : a study under the background of IFRS 9 project

(1)

Amsterdam Business School

MSc Accountancy & Control, track Accountancy Faculty of Economic and Business, University of Amsterdam

Value Relevance: Related to Banks’ Lobbying Behavior or Not?

A Study Under the Background of IFRS 9 Project

University of Amsterdam Accountancy & Control

2013-2014 Master Thesis in Accountancy

Author: Jin Zhang

Date of Final Version: 29 July 2014 Student Number: 10607692

First Supervisor: Dr. Sanjay Bissessur Second Supervisor: Dr. Ir. S. van Triest

(2)

1

Secondly, I would like to say I am very grateful to my supervisor, Dr. Sanjay Bissessur, who helped me a lot with my thesis and study. Definitely, he is the best teacher I have ever met. Besides, I want to say thanks to my classmates, friends and my teachers who ever offered me assistance during my master study. I feel so lucky to have you all.

At last, I want to offer my thankfulness to myself, who is some country girl with no

predominant background and had relative poor elementary education experience. Thanks for not giving up to the darkness in the way.

(4)

3

1 Introduction

The setting process of one accounting standard is associated with several procedures. In terms of IFRS (International Financial Reporting Standard), the usual procedure to issue a new standard is to follow the due process required by IASB (International Association Standard Board).

According to Scott (2011), accounting policies have economic consequences that can affect firm value. Under this circumstance, when the standard setters are about to change one particular standard, the value of the related firms would be affected. And the value of a firm is a concern of interrelated parties. Certainly the value of the firm would influence its

continuity, profitability and its future development. So when the firms conceive their interests are to be influenced towards the direction that they do not anticipate, they would exert some efforts to lobby.

In terms of accounting standards, the involving companies would like to give them opinions to the accounting standard setters to persuade them to modify the standards towards their interests. This kind of behavior is called lobbying. Lots of papers have studied the corporations’ lobbying behavior. Sutton (1984) and Watts and Zimmerman (1978) have discussed when the corporations will tend to lobby. Chen et al (2008) and Kosi and Reither (2014) investigated some determinants of the firms’ lobbying behaviors. The determinants they have observed involves ownership structure, major business, income volatility, capital structure and so on.

Relevance is regarded as of the basic qualitative characteristics of financial statements according to IFRS conceptual framework (IFRS Conceptual Framework). Value relevance can influence the cost of capital of s firm and the investors’ investing decisions. From this point of view, value relevance of financial reporting information should be of great

importance in influencing companies’ lobbying decisions. However, there is no such study before investigating the association between lobbying and value relevance, although there already are researches on income volatility and lobbying (e.g. Kosi and Reither.; 2014). The IASB’s recent project of replacing IAS 39 with IFRS 9 gives an opportunity to investigate the association of lobbying and value relevance. Lots of involved associations,

(5)

4

corporations, organizations and individuals have expressed their opinions on the new proposal. Due to the fact that IAS 39 and IFRS 9 are all about financial instruments and financial instruments are banks’ primary assets and liabilities, banks seem to be the most ideal study group.

In this paper, I study the banks’ value relevance and lobbying behaviors. 29 banks are identified in the comments letters submitted to the latest exposure draft of financial

instruments. And then 6 of them are deducted due to missing data, resulting 23 banks in the lobbying group. Using the data form DataStream, 5,645 firm-year observations are found for all the banks from year 1999 to year 2013. Among them, 5,363 firm-year observations are of non-lobbying group and 282 for lobbying group. Using these banks’ data, I obtained the value relevance data (in the form of Adjusted R-Square) for the two groups during year 1999 to 2013. By comparing the value relevance data for the two groups, I find that generally, the value relevance of lobbying group declines as of the new standard. And the value relevance of non-lobbying banks is obviously higher than lobbying banks since 2007. The results give supporting evidence to the first two hypotheses. Moreover, to get a more direct impression of the association between lobbying and value relevance, I run a regression using Adjusted R-Square and lobbying as well as other control variables. However, the result is insignificant, indicating that banks lobbying behavior is actually not related to the value relevance elements. In addition, I also try to find the relationship between lobbying intensity and value relevance, I use the comment letter page numbers as a proxy for lobby intensity. I find that the banks that submit comment letters are associated with higher value relevance, which means the higher the value relevance, the higher the lobbying intensity. I assume this is because that banks do not want their value relevance to decrease.

This paper has some limitations. Although some meaningful results are obtained from the empirical analysis, but I think there are other approaches to evaluate this assumed

relationship. Besides, although using the sample of banks are enough for this paper under the background of IFRS 9, the groups of other industry firms can also be used as testing samples to get an even more convincing result. To study the association of lobbying intensity and value relevance, the sample size can be expanded to examine the results.

Although the existence of limitations, this paper contributes to the study in the following ways. Firstly, it fills the gap in the study on the association between lobbying and value relevance, as there is no previous literature combine these two elements together to study.

(6)

5

Secondly, the research results are of importance to banks, investors and standard setters, as they can be aware of what consequences value relevance has in constituents’ lobbying decisions and the new standards’ impacts on lobbying corporations.

This paper will proceed as follows: section 2 illustrates previous literature with respect to value relevance and corporate lobbying. Besides, the introduction of institutional background of IFRS 9 is also provided in this section. In section 3 I put forward the research question, the developed hypotheses and research methodology. Section 4 provides information concerning data and empirical results. Section 5 concludes.

2 Literature Review

2.1 Lobbying

2.1.1 Lobbying and Standard Setting

“Without the compass and square, one cannot form squares and circles ( 不以规矩，不成方圆).” This is a historic quote said by Mencius; a great Chinese philosopher lived more than 2300 years ago. This quote tells us that everything in the world needs corresponding regulations to ensure successful operation. So is it in the accounting world.

As we have learned, accounting is an area in a great need of standards and regulations in order that stakeholders can obtain the relevant information directly and quickly. Scholars, i.e. Scott, consider accounting as a process to produce information of the certain organization. They have the opinion that the fundamental problem of financial accounting theory is how to determine the “right” amount of information. According to Allen (1990), information is an economic commodity. Thus, from the market perspective, the first-best information

production is when the marginal social costs to produce the information equal the marginal social benefits of the produced information. However, market does not always work well in the information production processes although managers have a lot of incentives to disclosure information. Due to the unavoidable existence of externalities, free-riding, adverse selection, moral hazard and unanimity, market could fail and that is why regulation is required in information production. That's the reason the modern researchers like Deegan (2006) argues that accounting standard setting is practically a political issue, which means standard setting is not as simple as it seems, there are complications associated in its root.

(7)

6

To regulate the production of financial information, a third party – standard setter – is needed. Accounting standard setting plays an important role in conveying users about how, when and where the financial information of a company is disclosed. In terms of international financial reporting standards (IFRS), the international association of standard board (IASB) is the mediating standard setter. In light of IFRS official website, IFRSs are developed through an international consultation process, the “due process”, which involves interested individuals and organizations from all around the world.1 The due process of IASB includes a formal process of inviting public to express their opinions during the whole process, and as Leuz et al. (2003) argue, the legitimacy of its standards depends on the participation of those affected by them. Therefore, lack of participation suggests failure in the process.

As we know that there are various users of financial information so it is obvious to have diverse and conflicting interests of users. Therefore, there is rarely an accounting standard, which is acceptable to all the users, exists. There are individuals, associations, or some corporate powers, which dominate the process of accounting regulations. Or in other words we can say that they lobby for or against a particular accounting standard to be imposed or already imposed by a standard setter.

Sutton (1984) offer the definition of lobbying as “All the actions, which interested parties take to influence the rule-making body”. The development and adoption of accounting standards is the result of a complex interaction among numerous parties. Watt and

Zimmerman (1978) note that interested parties, including corporate managers, accountants, auditors, and investors will expend considerable resources to influence the setting of accounting standards and the due process allows lobbyists to participate in a formal way (Giner and Arce, 2012). From the summary of the IASB staff paper, the constituents submitted comment letters to the IASB include preparer (31%), user/representative body (3%), preparer/representative body (26%), standard setter (12%), accountancy body (12%), regulator (7%), accounting firm (6%) and other (3%).

Georgiou (2004) illustrate two main kinds of lobbying method: direct lobbying and indirect lobbying. Specifically, direct methods include: (a) submitting letters in response to invitations to comment on discussion papers or exposure drafts; (b) speaking at public hearings; (c)

1_{The due process comprises six stages, with the Trustees having the opportunity to ensure compliance at}

various points throughout: 1. Setting the agenda 2. Planning the project 3.Developing and publishing the discussion paper 4. Developing and publishing the exposure draft 5. Developing and publishing the standard 6. After the standard is issued

(8)

7

having meetings with ASB members or technical staff (e.g., project directors); (d) having telephone conversations with ASB members or staff or speaking to them at conferences or in other settings; (e) having consultants working on particular projects. Indirect methods include lobbying through others such as: (a) commenting in the media; (b) trade or other

organizations representing the interests of the company; (c) company’s external auditors; (d) Financial Reporting Council (FRC) members; (e) the Consultative Committee of

Accountancy Bodies (CCAB); (f) government officials; or (g) sponsoring research studies. The analysis of the structure of constituent participation reveals that the most active lobbying group is preparers of financial statements while responses from users of financial statements are comparatively less (Tandy and Wilburn., 1992).

Prior researches on preparers’ participation in a due process of standard setting often investigate the comment letters written in response to accounting proposals in order to address the lobbying behavior. The analysis of comment letters is straightforward to observe the lobbying behavior of prepares and other constituents since these documents are publicly available and traceable. Besides, comment letters is considered to be a good proxy for lobbying behavior since there is a strong relation between the use of comment letters and the use of other lobbying methods, according to Georgiou (2004).

Sutton (1984) puts up with a cost-benefit theory of constituents’ lobbying behavior, that is a party only take lobby action when the expected benefits of lobbying, exceed the costs of lobbying. Watts and Zimmerman (1978,1986) argue that the benefits of lobbying, from the perspective of preparers, depend on the expected influence of the proposed accounting standards on the future cash flows of the preparer. Consistent with prior researches, they find that prepares submit significantly more comment letters relative to other constituent groups. However, these determinants cannot be taken as fully comprehensive.

According to Stenka and Taylor (2010), there are two main research lines identified in the literature concerning lobbying within the standard-setting process. One group of studies consider the motivations and incentives to lobby and their influence on lobbyists’ behavior (e.g. Watts and Zimmerman., 1978, Weetman et al., 1994) while the other studies examine the relationships of power within the standard-setting process and analyze regulations’ interactions with parties involved in lobbying (e.g. Hope and Gray, 1982.; Pong and Whittington., 1996; Weetman., 2001; Kwok and Sharp., 2005).

(9)

8

2.1.2 Determinants of Corporates’ Lobbying Behaviors

Chen et al. (2008) examine the effectiveness of preparers’ lobbying from a financial perspective. They test the association between firm’s lobbying expenditure and financial performance and find that a firm’s accounting earnings and cash flow from operation are both positively related to its lobbying expenses. They also test whether lobbying expenditures are value-relevant to its market return and obtain positive relationship. Besides, they compare the stock returns of firms that lobby with that of firms that do not and find evidence that firms with higher lobbying intensity outperform others. Despite they make new progress in this line of research, there is still one question existing: despite the growing numbers of lobbying firms, what is keeping even more firms from engaging in lobbying activities?

Consistent with Geogiou (2004), firms are not likely to take part in the due process when the proposed standard does not have a material impact on their financial statements. In line with the positive accounting theory (Watts and Zimmerman., 1978), companies’ participation in the due process is driven by the economic consequences of the proposed standard for its accounting numbers and cash flows. According to Zeff (1978), economic consequences mean the impact of accounting reports on the decision-making behavior of management,

government, associations, investors and creditors. The economic consequences have been linked to several corporate characteristics such as firm size, leverage and ownership structure (Durocher et al., 2011). In fact, previous empirical researchers conclude that negative effects of accounting information to be the most important drivers of corporate participation (e.g. Dechow et al., 1996; Hill et al., 2002; Jorissen et al., 2012).

Sutton (1984) argues that larger firms are more likely to lobby because they acquire relatively higher benefits from lobbying activities compared to incur costs. Besides, due to their

economic importance, large companies might have stronger influence over the outcome of the due process and in turn have more incentives to lobby (Koh., 2011). Consistent with these arguments, previous empirical researchers have discovered that larger firms indeed are more likely to take part in the due process of standard setting (e.g. Larson., 1997; Georgioou., 2005, Jorissen et al., 2012).

Similarly, more profitable firms are expected to have more resources to enable them to engage in the process of standard setting and thus to lobby more. Nevertheless, on the other hand, Jorissen et al. (2012) adopt profitability as a proxy for a firm’s dependence on external

(10)

9

stakeholders. Contrary to their predictions, the researchers find that more profitable firms lobby more although they are less dependent on external stakeholders.

A more recent study by Kosi and Reither (2014) investigate the determinants of financial firms’ lobbying behavior based on the replacement process of IFRS 4 Insurance contracts. They investigate firms’ lobbying decisions and their long-term lobbying intensity using an international sample of publicly listed financial firms. They examine the effect of income volatility, company characteristic, main insurance type of the company, debt contract covenants, the ability to raise new capital and ownership structure.

Elbannan and McKinley (2006) predict that an increase in perceived uncertainty of

accounting information is also a significant driver of corporate participation in lobbying. In Ang et al (2000) show that firms lobby against the proposed standards that would result in greater income volatility. Ernst & Young (2010) finds that the majority of lobbyers express concerns about the potential increase in income volatility of exposure draft on IFRS 4. Graham et al (2005) argue that firms’ preferences for smooth earnings are related to increased predictability of future earnings and the resulting increase in firm value as well as to the decrease in cost of capital due to lower risk perceived by investors. Risk-averse stakeholders also prefer constant payments from firms’ earnings and thus smooth consumption (Morduch., 1995). Managers’ motives to smooth earnings can be related to their incentives to increase private benefits or to prevent outside interventions by concealing true performance (Gaver et al., 1995; DeFond and Park,, 1997). These arguments would suggest that firms with high-income volatility are more likely to oppose any further increase in volatility and in turn lobby more. And firms with low-income volatility might have more incentives to lobby against increased volatility in order to retain the discretion to smooth their income.

Moreover, the expected increase in income volatility can lead to higher cost of capital

(Graham et al., 2005) and this could potentially limit firms’ ability to raise new capital (Koh., 2011). Limitations for firms to raise new capital would affect firms with higher future

financial needs and current financial constraints.

Dhaliwal (1980) presents evidence that the nature of loan covenants, is such that highly leveraged firms, will tend to choose those accounting methods which result in higher reported equity, higher reported earnings or lower volatility of reported earnings in order to reduce the probability of default.

(11)

10

In addition, the ownership structure also plays a part in firms’ participation in standard setting process. The anticipated decrease in earnings and other changes in financial statements due to the new ED would be perceived by outside investors as increased firm risk which can lead to potential decrease in firm value. Thus this could cause some financial problems. And such adverse consequences are more likely when a firm has dispersed outside investors, who cannot identify that the changes merely come from the change in accounting rules (Kosi and Reither., 2014). According to information asymmetry theory, block holders have obvious information advantage compared to disperse outside investors. Brockman and Yan (2009), as well as Edmans and Manso (2011), expressed the block holders’ information advantages mainly come from two reasons: (i) block holders’ large stakes (5% or more of common shares outstanding) might result in pressure on management and lead to access to private information and (ii) even when accessing the same information as other investors, block holders can spend more resources on the analysis of the information. Consequently, block holders are likely to link the adverse consequences to the new accounting rules and not to a change in firm performance.

Similar to the function of income volatility, Francis, Lafond and Olsson (2004) suggest that increased value relevance is associated with lower cost of capital. This is attributed to investors’ perception that value relevance contributes to lower information risk. Lower information risk can decrease the imprecision in the estimates of the investors’ future cash inflows from the company and thus lead to increasing level of investment. In general, value relevance, as same as income volatility, have real economic consequences.

2.2 Value Relevance and Accounting Standards 2.2.1 Objective of Financial Statements

According to IFRS conceptual framework, one of the basic qualitative characteristics of financial statements is relevance. “Relevant financial information is capable of making a difference in the decisions made by users. Information may be capable of making a difference in a decision even if some users choose not to take advantage of it or are already aware of it from other sources” (Conceptual Framework, 2010: A33).

Financial reporting ought to provide useful information for stakeholders of a company for them to make decisions about further activities concerning the company, especially for stock

(12)

11

investors as it is them that invest their capitals into the company. Thus, it is easily understood that they need to analyze the financial information to ascertain the fundamental value of companies. Through analyzing the financial statements, they can acquire necessary financial and even non-financial information to assess what the company is worth so that they can evaluate the respective stock prices.

In reality, one of the major objectives of financial reporting is to provide equity investors with information relevant to estimate the value of the company. The research about value relevance empirically evaluate whether this goal is met or not.

2.2.2 Efficient Market Theory

It should be noted that value relevance research is related to efficient market theory (Fama, 1970). When asking whether financial information is value relevant, one is asking whether stock investors use accounting numbers as an input for valuation. However, there is another question about if the investors’ use of accounting information is optimal. Aboody, Hughes and Liu (2002) document evidence suggesting that the market may not be completely efficient in its processing of public information. But value relevance researchers seem to implicitly draw the conclusion that stock market is efficient, at least in the semi-strong form. Assuming the validity of the results of Aboody, Hughes and Liu (2002) at a certain level, it is better to make an assumption for this research that the market is efficient, at least in a semi-strong level.

2.2.3 Value Relevance of Accounting Standards

Beisland (2009) argue that it should be noted that most standard setters view value relevance, along with other attributes, as an important characteristic of accounting information.

Empirical research on the relations between capital markets and financial statements is generally referred to as capital market-based accounting research (CMBAR). Modern CMBAR originated with Ball and Brown (1968). Although according to Barth, Beaver and Landsman, the concept of value relevance was not developed until 1993, some earlier researches, e.g. Ball and Brown, Beaver, also can be regarded as part of the value relevance research. Beisland (2009) denotes that value relevance research measures the usefulness of accounting information from the perspective of equity investors. For the last two decades, the most empirical value relevance researches focus primarily on the following three categories: (1) value relevance of earnings and other similar flow elements from the income statement

(13)

12

and cash flow statement (e.g. Beaver, McAnally and Stinson., 1997, Marquardt and Wiedman., 2004, Brown, He and Teitel., 2006); (2) value relevance of balance sheet measures, i.e., equity and other stock measures(Barth, Beaver and Landsman(1998), DEchow, Hutton and Sloan 1999, Caroll, Linsmeier and Petroni 2003); (3) value relevance development over time(Francis and Schipper., 1999, Hellstorm., 2006, Hail., 2013).

In addition, there is also a specific type of value relevance research focuses on the

distinguished value relevance of alternative accounting methods or standards. Ayers (1998) perform a quite typical study within the value relevance study. He compared the Statement of Financial Accounting Standards (SFAS) No. 109 Accounting for Income Taxes and

Accounting Principles Board (APB) opinion No. 11 Accounting for Income Taxes. His investigation on value relevance of the two principles suggests that SFAS No. 19 provides value relevant information above and beyond APB No. 11. Hope (1999) conducted another typical study on the value relevance differences of various accounting methods by

investigating the effects of introducing deferred tax accounting in Norway. He finds that this change in accounting legislation significantly increased the value relevance of earnings. Barth, Landsman and Lang (2008) examined whether the application of International Accounting Standards (IAS for short, denoted International Financial Reporting Standards since April 2001) is associated with higher accounting quality by investigating a sample of firms in 21 countries that adopted IAS between 1994 and 2003. They draw the conclusion that companies applying IAS displaying less earnings smoothing, less managing of earnings towards a more timely recognition of losses and a higher association of accounting amounts with share prices and returns. They also document higher value relevance for IAS firms. Since European law began to require all companies listed on European stock exchange to prepare their consolidated financial reports based on International Financial Reporting Standards (IFRS) in 2005, the value relevance effect of introducing IFRS in European countries has been a popular research topic. Some researches (i.e. Bartove et al., 2005, Van Tendeloo and Vanstraelen., 2005; Daske., 2006; and Barth et al., 2008) compare accounting amounts based on, and the economic implications of, non-US firms applying IFRS

voluntarily and domestic standards, using a variety of methods. Although some researchers suggest greater value relevance of net income and equity book value after voluntary adoption of IFRS, it is not easy to distinguish whether the findings are attributable to application of IFRS or to incentives for voluntary adoption.

(14)

13

A range of studies examine the changes in value relevance of the adjustments required to restate net income and equity book value from those based on domestic standards to those based on IFRS. Hung and Subramanyam (2007) find that, for a sample of 80 voluntary German non-financial IFRS adopters, the value relevance of the adjusted equity book value increases, while that of the net income adjustment does not. However, it is not clear whether the findings related with the net income adjustments generalize to financial firms, firms adopting IFRS mandatorily, firms in other countries, and firms adopting IFRS after the substantial changes made to IFRS by IASB. Nevertheless, Jermakowize, Prather-Kinsley and Wulf (2007) document increased value relevance of earnings after the adoption of IFRS for a group of thirty Germany companies with the largest market capitalization and turnover. Researchers in Finland report that the reconciliation of Finnish GAAP to IAS earnings does not provide significant relevance (Nisciliation, Kinnunen and Kasanen., 2000). Gjerde at al. (2008) find little evidence in terms of the difference in value relevance of earnings and equity book value based on IFRS and Norwegian standards by studying a sample of 145 Norwegian companies adopted IFRS mandatorily. What is unclear is that how this small sample findings generalize to other countries. Horton and Serafeim (2010) investigate the value relevance of earnings and equity book value adjustments of a sample of 297 large non-financial UK firms mandatorily adopting IFRS in 2005.Theyfind that the earnings adjustment related with two individual standards, share-based payment and goodwill, have incremental value relevance. Barth et al. (2012) reports greater value of net income and equity book value after mandatory IFRS adoption based on a sample of large European firms. However, this study does not take into consideration of the impact on value relevance associated with individual standards or whether the increased value relevance differs for financial and non-financial companies. Barth, Landsman, Young and Zhuang (2014) take a more comprehensive investigation about the value relevance of differences between earnings based on IFRS and domestic standards for European firms by testing a sample of 1201 firms from the 15 countries in Europe. Taken the limitations in the previous literature, they conducted three tests to address the value relevance between financial and non-financial firms, the influence of individual standard (especially IAS 39 Financial Instruments: Recognition and Measurement) on value relevance and the difference in value relevance of different country groups. Their findings provide evidence that earnings adjustments from mandatory 2005 IFRS adoption in Europe are value relevant for both financial and non-financial firms. For the individual standard, they find that the adjustments to earnings related with IAS 39 is value relevant for financial firms but not

(15)

14

non-financial firms. Despite some differences, investors view earnings measured using IAS 39 as more relevant that that measured under domestic standards, except for French/German non-financial firms. Taken all the findings together, their findings suggest that the

adjustments to earnings resulting from mandatory adoption of IFRS in Europe in 2005 are relevant to investors.

2.3 Institutional Background: IAS 39 V.S. IFRS 9

Accounting standards are in constant development and change since the day they were put into effect. That is, to a large extent, due to the development of the economic environment. Accounting for financial instruments has been a rather controversial topic owing to their complexity and volatility. The accounting standard set by the IASB for financial instruments has been subject to change. From the very beginning, it is IAS 25 that is in effect in 1985; and gradually IASB develops it into IAS 32 and IAS 39. Since the emergence of IAS 39, it has been undergone constant change throughout the years. IAS 39 has been the cause of much confusion and understanding. Under current stage, the new standard – IFRS 9 – is under discussion and is intended to replace the position of IAS 39.

In March 2008, a discussion paper was issued relating to reducing complexity in reporting financial instruments, which is the very beginning of the modification of financial instrument. As the Discussion Paper stated, complexity is one of the most important issues in financial reporting and financial instruments are among the most complexity things on which to report clearly (BD12). One cause of complexity is that financial instruments themselves are

complex and hard to understand even with full information of various terms and conditions (BD14). Another cause of complexity is that the standards for financial instruments contain many alternatives, bright lines and exceptions that often obscure the underlying principles (BD15). Besides, the many ways of measuring financial instruments is an important reason why today’s requirements are complex. They result in many accounting rules, for example, on how different financial instruments should be measured and when and how financial assets measured using a cost-based method should be impaired.

In practice, the concerns for the change of classification and measurement of financial instruments start from the financial crisis that began midway through 2007 and continued through the end of 2008. The financial crisis resulted in the collapse of substantive

commercial and investment banks, including Lehman brothers, Merrill lynch and Wachovia. Policy-makers, bankers and academics all have written a lot to offer their opinions related to

(16)

15

the causes of the crisis and most have reached a consensus that the crisis started following the bursting of house bubble in the USA. During 2007, financial institutions needed to recognize a drop in value of their financial assets. The financial press first focused on the amounts of impairment write-downs and the doubtful quality of the related assets. Fair value accounting began to be blamed by financial institutions. The CEO of AIG (American Insurance Group) was cited in the Financial Times in March 2008 as suggesting management estimate the likely losses and recognize those, rather that reflect market prices by measuring at fair value (Andre; 2009). Like what the European Banking Federation stated, the pro-cyclical nature of fair value measurement appears to have worsened the impact of the crisis on financial and non-financial corporations and it is vital that accounting standard-setters should take up the issue and seriously consider amending fair value accounting rules for financial instruments (EBF, 2008). IASB and FASB began to come under pressure to reconsider the fair value rules. That is why the IASB has been accused of having intensified the effects of the financial crisis. Besides, calls for action to prevent a repeat of the financial crisis have also paid

attention to accounting standard setting and the way in which accounting information contributed to the financial crisis.

Despite the critics on fair value accounting, it is true that during the years of economic prosperity in the past few years, measurement at fair value has improved the transparency of financial institutions and supported investors to understand their risk profile. And the

earnings volatility caused by the fair value accounting of financial assets was welcomed, since it was reflected in higher profits. And this kind of volatility also provides more relevant information for investors who are seeking better understanding of the risk profile of the entity (Fiechter, 2011). According to the Discussion Paper, fair value is regarded to be the only measure appropriate for all types of financial instruments (3.10).

Under IAS 39, financial instruments can be classified and measured in four categories (IAS 39.45): (1) Financial assets at fair value through profit or loss; (2) Available-for-sale financial assets; (3) Loans and receivables; (4) Held-to-maturity investments. Concerning financial liabilities, IAS 39 recognizes two classes of financial liabilities (IAS 39.47): (1) Financial liabilities at fair value through profit or loss; (2) Other financial liabilities measured at amortized cost using the effective interest method. 2

2_{The initial and subsequent measurement of financial instruments under IAS 39 is summarized table1 in}

(17)

16

In the exposure draft ED/2012/4 Classification and Measurement: Limited Amendments to IFRS 9 (Proposed amendments to IFRS 9 (2010)), the method of classifying financial instruments into four categories are changed and ED proposed that an entity shall classify financial assets as subsequently measured at either amortized cost, fair value through other comprehensive income or fair value through profit or loss on the basis of both: (a) the entity’s business model for managing the financial assets and (b) the contractual cash flow

characteristics of the financial asset. 3

Regarding the new classification and measurement approach, 173 comment letters were received from various interest parties. They provide their opinions concerning the limited amendments on the proposed classification and measurement standards considering their respective specific situation. IASB is currently in the balloting process for the new issuance of IFRS 9 Financial Instruments, incorporating the final requirements for impairment and the limited amendments to classification and measurement (IFRS website).

3 Research Question, Hypotheses Development and Methodology

3.1 Research Question

During the due process of a newly proposed IFRS standard, various related constituents would participate in it to share their specific opinions. Theoretically, companies, investors, auditors, associations and other parties have the intention to lobby if the proposed regulation has direct or indirect impact on their operation and financial performance. According to the positive accounting theory of Watt and Zimmerman (1978), the participation of corporates in the due process is driven by the economic consequences of the proposed standards for a firm’s future cash flows. So when the investors anticipate that the future cash flows of the entities they invested, they would expect different returns on their investment and then they would change their decision of intentional investments. If the application of a new

accounting standard ought to influence the investing decision of investors, it is of value relevance. Not considering other influencing factors, a company has the reason to take part into the due process to lobby standard setters to make the new standard be beneficial to it as much as possible, or at least not affect their current economic conditions negatively.

3_{The primary changes from IAS 39 to IFRS 9 (2010) to IFRS 9 (2012 ED) are attached in Table 2 in the}

(18)

17

In this paper, the main focus is on one single standard – IFRS 9, as the IASB is now still working on it. Due to the far-reaching effects of IFRS 9 on financial instruments, which comprise a quite large portion of the assets and liabilities in banks, this research primarily, focus on the constituents as banks. Therefore, I put forward the research question in this paper as follows:

Is value relevance related to the lobbying behavior of banks concerning IFRS 9?

3.2 Hypotheses

As referred in the literature review part, the study on the participation of constituents in the process of standard setting is extensive. Prior studies primarily employ a single-issue

approach, while some take a multi-issue or multi-period approach (Kosi and Reither., 2014). Under the precondition that previous researches on the lobbying constituents show that the prepares of the financial statements are the primary lobbying groups as their interests are directly related to the changes in the standards. The standards concerned in this paper are IAS 39 and IFRS 9 and both of them are concerned with financial instrument. The balance sheets of banks contain significantly more financial assets and liabilities than do the balance sheets of entities from other industries (Fiechter., 2011). Therefore, the main groups that are going to be studied in this paper will be banks.

Since one of the major goals of financial reporting is to assist equity investors to estimate the valuation of the company, value relevance of a company demonstrates whether this goal is met or not (Beisland., 2009). The higher the value relevance, the lower the cost of capital the firm is confronted. Therefore, the level of the value relevance is closely associated with a firm’s continuity, profitability and further development. Besides, referring to the contract theory and agency theory, the profitability of the company have influential effects on

management’s compensation, so management may decide to lobby when they feel their own utilities are affected. Given there exists a relationship between value relevance of corporates and cost of capital, the banks would lobby when they feel their interests are affected because of the new proposed standard. And they would anticipate to benefit more through lobbying than the resources they exert on lobbying the standard setter. So I would like to assume that the value relevance of the lobbying banks would decline because of the new standard proposal, which in the context of this paper starts in 2008.

(19)

18

Scott (2011) indicates that accounting standard has economic consequences. According to Zeff (1978), economic consequences mean the impact of accounting reports on the decision-making behavior of management, government, associations, investors and creditors. It is argued that the consequential behavior of these individuals and groups could be detrimental to the interests of other affected parties. And accounting standard setters must take these consequences into consideration when deciding on accounting questions. The lobbying behaviors are conducted since proposed new or changed accounting standards affect the accounting information in financial reports, which can in turn have accounting related economic consequences (Hartweg., 2012).

Although the changes in standards is associated with the interests of the banks, there is always the phenomena that only a small portion of the bank group would conduct lobbying behavior. Chen et al (2008) document that corporates’ expenditures on lobbying are value relevant to the firm’s market price and stock return. Combing this conclusion with signal theory, we can conjecture that corporates’ lobbying activities should be associated with the value relevance of the firms. When one firm’s value relevance is lower compared to its peer firms, its investors are confronted with higher risk of information and the firm is faced with higher cost of capital. This is definitely not beneficial for the certain firm’s business status in the industry. So here I want to make a hypothesis that:

H2: the lobbying banks have lower value relevance compared to the non-lobbying banks. Similarly, in line with the result of Chen et al (2008)’s research, among the lobbying banks, we can suppose that those lobbies more intensely should have lower value relevance than those lobbies not so intensely. So I want to describe the second hypothesis as follows: H3: the lower the value relevance of the banks, the higher the lobby intensity of lobbying banks.

3.3 Methodology

3.3.1 Methodology of Examining Value Relevance

The empirical research on value relevance is based on valuation theory. Traditional financial theory states that the theoretical value of a company’s equity is the present value of all future dividends or free cash flows to equity. Some models are developed to help equity investors to estimate company value, e.g. Feltham and Ohlson (1995)’s clean surplus model. However,

(20)

19

value relevance research does not pay attention to how accounting information is used in valuation. Instead, this line of research concerns whether accounting information is able ex post to explain variations in stock prices between companies. Value relevance researchers are interested in how accounting numbers influence market values of equity.

Beisland (2009) points out that the value relevance of balance sheet figures is sensitive to the valuation principles applied to the various asset and liability components. Under the

circumstances in this paper, the balance sheet numbers of financial instruments are subject to diversification when they are classified and measured under distinct accounting principles. In the meanwhile, the income statement figures would be affected due to different measurement methods. According to the exposure draft, when an asset is measured by amortized cost, its prospective expenses and profits ought to be the same financial asset that is measured under fair value option. Even through utilizing fair value option for two assets, the subsequent influence of them is possible to be different when it is measured at fair value through profit or loss and at fair value through other comprehensive income.

That is to say, the balance sheet numbers of financial assets and liabilities would change as the accounting methods change. The income statement numbers at the year-end should change as well. The value relevance of financial statement items should be subject to

fluctuate. The institutional background of this study, as stated in literature review, is the due process of issuing the new accounting standard IFRS 9: Classification and Measurement. Both the balance sheet and income statement numbers will be affected under this standard. Several empirical researches of balance sheet components compare the value relevance of historical cost estimates with that of fair value estimates. There is a conclusion that fair value estimates are more value relevant (118-120). But there are also findings indicating that for small banks, historical measures of loans and deposits are more relevant than fair values, e.g. Khurana and Kim (2003). But in general the fair value accounting measures provide more relevant information for stakeholders.

To test the first hypothesis, I want measure the relation between market value and accounting number of the selected firms using the model from Hail (2013). Hail (2013) establishes four model to test the association, two focusing on the income statement, one on the balance sheet, and one on the combination of the two elements. I would like to refer to three of the four regression models from her paper to investigate the value relevance of (1) earnings, (2)

(21)

20

balance sheet items, (3) book value of equity and earnings. So the three models I select is demonstrated here:

Earnings relation: RET =β0+β1 NI/P + β2ΔNI/P +ε (1.1) Balance sheet relation model: P=β0+β1AST+β2LIAB+ε (1.2)

Book value and earnings relation: P= β0+β1 BV+ β2 NI +ε (1.3) In these equations, RET (the annual buy- and-hold stock return) and P (the market value per

share at fiscal year-end) is used as the market measures. The accounting measures are NI (net income), ΔNI (the change of net income), AST (total assets) and LIAB (total liabilities) of the particular banks. According to Hail (2013), the explanatory power of the above regression (measured by the adjusted R2) will be used to represent the value relevance, meaning the percentage of variation in stock returns and price explained by the accounting variables. 3.3.2 Methodology of Examining Lobbying and Value Relevance Association

Previous researches state several factors that may induce preparers to lobby. Prior literature also contends the potential benefits and costs drive the decision of preparers not to lobby. For instance, Georgiou (2002) finds that UK listed prepares do not participate in the ASB

accounting standard setting process when they have the impression that the submission of comments is ineffective to influence the ultimate outcome of that standard setter. Schalow (1995) also observes a similar result for a sample of US preparers. However, there are also some other reasons to explain the non-participation of preparers. For example, prepares do not lobby when they agree with the proposals promulgated by the standard setters (Gavens et al., 1989).

In order to test the relation between lobbying behavior and value relevance, according to previous literature, I choose the following factors as the control variables in the empirical model. They are:

(1) Company size (SIZE) (2) Income volatility (VOL) (3) Profitability (ROA) (4) Leverage (LEV) (5) Free cash flow (FCF) (6) Close (CLOSE)?

(22)

21

Aware of these, to get a more obvious and direct image of the association between lobbying decision and the value relevance of the banks, I establish the following empirical model using these variables.

LOBBY = β0 + β1VR + β2SIZE + β3VOL + β4 ROA+ β5 LEV+ β6FCF + β7 CLOSE (2.1) LOB_ IN= β0 + β1VR + β2SIZE + β3VOL + β4 ROA+ β5 LEV+ β6FCF + β7 CLOSE (2.2)

The particular explanation of the variables and their predicted relationship are descried in

Table 1.

Table 1: Definitions for Variables

Dependent variables Definition

LOBBY Dummy variable that equals 1 if a firm submits a comment letter in response to the ED, and 0 otherwise.

LOB_IN Number of comment letter pages submitted by the lobbying banks to IASB Independent variable Definition

VR Value relevance measured by adjusted R2 of the firms obtained for every specific firm for the three relations

SIZE Natural logarithm of market capitalization (WC07210) in US dollars at the fiscal year-end

VOL Income volatility measured as standard deviation of return on assets (ROA) over the ten years prior to lobbying and the year of lobbying

(i.e. in the period 1999–2013)

ROA Profitability (return on assets) measured as net income before extraordinary items (WC01551) divided by total assets (WC02999) at fiscal year-end in %

LEV Long-term debt (WC03251) divided by the sum of long-term debt and market value of equity (WC08001)

FCF Sum of net cash flows from operating activities (WC04860) and investment

activities (WC04870) scaled by total assets (WC02999)

CLOSE Number of closely held shares (WC05475) divided by common shares outstanding (WC05301). Closely held shares include shares held by investors owning 5% or more of the outstanding shares

(23)

22 4 Data and Results Analysis

4.1 Examination of Value Relevance 4.1.1 Sample and Descriptive Statistics

In order to conduct test of the established model and in order to obtain relevant results, the data of the banks form all around the world are identified and sorted in DataStream. For the purpose of this research, two subsamples are required to identified in the total bank sample: the banks submitted comment letters and banks did not submitted comment letters. In short, they were codes as Lobby Banks and Non-Lobby Banks respectively.

To divide the two groups, I analyze the comment letters IASB received for the Exposure Draft: limited amendments to IFRS 9 Classification and Measurements. From all the 173 comment letter submitters for the exposure draft, 29 banks are identified in them. Then 6 of the 29 banks are deleted because of no date found. The name list of the lobbying banks is stated in Table 2.

Table 2. Lobby Banks List

Company Name Freq. Percent Cum.

AGRICULTURAL BANK OF CHINA LIMITED 7 1.27 1.27

AUSTRALIA AND NEW ZEALAND BANKING GROUP

16 2.90 4.17

BANCO BILBAO VIZCAYA ARGENTARIA SA 32 5.80 9.96

BANK OF CHINA LIMITED 16 2.90 12.86

BARCLAYS PLC 32 5.80 18.66

BNP PARIBAS SA 32 5.80 24.46

CHINA CITIC BANK CORPORATION LIMITED 10 1.81 26.27

CHINA CONSTRUCTION BANK CORP 12 2.17 28.44

CHINA MERCHANTS BANK CO LTD 13 2.36 30.80

CHINA MINSHENG BANKING CORPORATION LIMI

14 2.54 33.33

CITIGROUP INC. 48 8.70 42.03

COMMERZBANK AKTIENGESELLSCHAFT 32 5.80 47.83

CREDIT AGRICOLE SA 32 5.80 53.62

CREDIT SUISSE GROUP AG 16 2.90 56.52

DBS GROUP HOLDINGS LTD 16 2.90 59.42

DEUTSCHE BANK AKTIENGESELLSCHAFT 48 8.70 68.12

ERSTE GROUP BANK AG 32 5.80 73.91

(24)

23

LLOYDS BANKING GROUP PLC 16 2.90 82.61

NATIONAL AUSTRALIA BANK LIMITED 16 2.90 85.51

SOCIETE GENERALE 32 5.80 91.30

STANDARD CHARTERED PLC 32 5.80 97.10

UBS AG 16 2.90 100.00

Total 552 100.00

Using the data from the last 16 years during 1998 to 2013, total 7554 firm-year observations are found. And then 1909 observations are deleted due to missing data, resulting in 5645 observations left for the total sample group (See Table 3).

Table 3: Total Sample

fyearend Freq. Percent Cum.

1999 257 4.55 4.55 2000 273 4.84 9.39 2001 284 5.03 14.42 2002 301 5.33 19.75 2003 317 5.62 25.37 2004 361 6.40 31.76 2005 383 6.78 38.55 2006 408 7.23 45.78 2007 424 7.51 53.29 2008 448 7.94 61.22 2009 430 7.62 68.84 2010 451 7.99 76.83 2011 464 8.22 85.05 2012 453 8.02 93.07 2013 391 6.93 100.00 Total 5.645 100.00

(25)

24

The descriptive statistics of the total sample and the two subsamples are displayed in Table 4.

Table4-1: Descriptive statistics total sample (Lobby and Non-Lobby Banks)

variable N mean sd p25 p50 p75 NI 5645.000 0.076 0.096 0.048 0.075 0.107 RET 5645.000 1.148 0.440 0.881 1.072 1.321 AST 5645.000 5681.669 21018.739 60.829 303.639 2157.389 LIAB 5645.000 615.745 3804.748 8.449 44.021 217.349 BV 5645.000 365.213 1444.444 5.237 21.685 130.046

Table 4-2: Descriptive statistics total sample (Non-Lobby Banks)

variable N mean sd p25 p50 p75 NI 5,363,000 0.077 0.097 0.048 0.075 0.108 RET 5,363,000 1,151 0.443 0.881 1,073 1,324 AST 5,363,000 5,942,683 21,531,237 59,204 305,769 2,572,196 LIAB 5,363,000 635,031 3,900,608 8,154 43,745 215,622 BV 5,363,000 383,005 1,479,771 5,407 22,786 145,877

Table 4-3: Descriptive statistics total sample (Lobby Banks)

variable N mean sd p25 p50 p75 NI 282,000 0.071 0.081 0.053 0.076 0.103 RET 282,000 1,091 0.387 0.866 1,051 1,264 AST 282,000 717,771 1,103,264 77,914 226,488 832,123 LIAB 282,000 248,962 539,458 12,804 52,287 275,199 BV 282,000 26,848 40,816 3,897 11,622 29,321

NOTE: The total sample comprises 5,645 firm-year observations; the non-lobby banks sample comprises 5,363 firm-year observations and the lobby sample comprises 282 firm-year observations between 1999 and 2013, for which we have sufficient accounting data in DataStream. I use the following accounting measures: NI is the per-share amount of total NI before extraordinary items (WC 01551). RET is the annual buy-and-hold stock return inclusive of dividends and computed over a firm’s fiscal year. AST and LIAB are the per-share numbers of total assets (WC 02999) and total liabilities (WC 03351). BV equals a firm’s common equity per share (WC 03501). NIPS equals These accounting variables are used to run all three-regression analysis. I compute all the per-share amounts by dividing the total values with the number of common shares outstanding (WC 05301) and I truncate all the accounting measures at the first and 99th percentile.

The next Table 5 describes the Pearson and Spearman correlation to see how the different accounting measures are related to each other and therefore supports the quality of the

(26)

25

regression models. Their value can range from (-1) for a perfect negative linear relationship to (+1) for a perfect positive linear relationship. A value of 0 (zero) indicates no relationship between two variables. For the spearman correlation in this table above diagonal, I find a very strong positive correlation of 0.8898 between assets and liabilities, and a very strong positive correlation between book value and assets of 0.9731. There is also a strong correlation between liabilities and book value of equity, which is 0.8485. For the Pearson

correlation in this table below diagonal, I find a strong positive correlation between book

value and assets (0.9207). Finally, a moderate correlation of 0.6495 between book value and liabilities and a correlation of 0.6098 between assets and liabilities are also visible. All these correlation shows that the accounting variables are suitable for the models, thus improves the reliability of the results.

Table 5-1: Pearson Correlation

NI RET AST LIAB BV

NI 1 RET 0.0998 1 AST -0.0248 -0.0078 1 LIAB 0.0378 0.0372 0.6098 1 BV 0.0030 0.0059 0.9207 0.6495 1 N=5,363

Table 5-2: Spearman Correlation

NI RET AST LIAB BV

NI 1 RET 0.0728 1 AST -0.0295 -0.0366 1 LIAB -0.0056 -0.0397 0.8898 1 BV -0.0199 -0.0274 0.9731 0.8485 1 N=5,363 4.1.2 Empirical Results

As described in the third section, the value relevance of the proposed new financial instruments classification and measurement standard is measured referring to the value relevance model of Hail (2013). The first model (Model1.1) measures the earnings relation where return (RET) is the dependent variable and net income (NI) and changes therein (ΔNI)

(27)

26

are the independent variables. The second model (Model1.2) measures the balance sheet relation with stock price (P) as the dependent variable and assets (AST) and liabilities (LIAB) are the independent variables. The third model (Model1.3) measures the book value and earnings relation where stock price (P) is also the dependent variable and book value of equity (BV) and net income per share (NIPS) are the independent variables. Table 6-1 shows the regression results of the non-lobby banks and Table 6-2 displays the regression results of the lobbying banks group.

Table 6-1-1: Regression Results of Non-Lobby Banks (Earnings Relation)

fyearend N NI t-stat dNI t-stat Adj. R²

1999 241 0.178 -1 -0.023** (-2.06) 0.016 2000 257 0.225 -0.89 -0.017 (-1.55) 0.005 2001 268 1.032*** -5.12 -1.183*** (-5.52) 0.114 2002 284 1.280*** -6.86 -0.042 (-0.53) 0.142 2003 298 0.504*** -2.71 0.137*** -3.52 0.046 2004 342 0.869** -2.24 -1.004*** (-4.17) 0.076 2005 364 2.922*** -10.14 -3.136*** (-11.14) 0.271 2006 387 1.716*** -4.03 -1.976*** (-7.23) 0.12 2007 405 1.899*** -4.36 -3.743*** (-7.80) 0.128 2008 428 0.674** -2.29 -1.122*** (-3.95) 0.032 2009 411 1.882*** -9.11 -2.382*** (-10.93) 0.223 2010 430 1.259*** -4.12 -0.386*** (-3.57) 0.052 2011 443 1.043*** -7.02 -0.355*** (-4.15) 0.098 2012 432 0.814*** -6.41 -0.890*** (-7.23) 0.117 2013 369 0.015 -0.09 -0.091* (-1.70) 0.003

Table 6-1-2: Regression Results of Non-Lobby Banks (Balance Sheet Relation)

fyearend N AST t-stat LIAB t-stat Adj. R²

1999 241 0.040*** -12.58 -0.023* (-1.85) 0.417 2000 258 0.038*** -12.32 -0.012 (-0.94) 0.401 2001 268 0.036*** -31.51 -0.011* (-1.82) 0.809 2002 284 0.036*** -22.84 0.003 -0.64 0.776 2003 299 0.039*** -20.59 -0.014** (-2.24) 0.671 2004 342 0.037*** -11.97 0.102*** -5.07 0.682 2005 364 0.058*** -19.47 -0.01 (-0.69) 0.643 2006 389 0.077*** -10.64 0.121*** -2.95 0.521 2007 405 0.056*** -16.8 0.892*** -23.38 0.842 2008 428 0.057*** -14.54 0.461*** -12.61 0.709 2009 411 0.054*** -16.83 0.222*** -7.42 0.667 2010 430 0.028*** -7.78 0.707*** -28.89 0.863 2011 443 0.014*** -4.42 0.913*** -39.08 0.877 2012 432 0.020*** -5.79 0.701*** -29.46 0.811 2013 369 0.017*** -5.32 0.746*** -33.05 0.877

(28)

27

Table 6-1-3: Regression Results of Non-Lobby Banks (Book Value and Earnings Relation)

fyearend N BV t-stat NIPS t-stat Adj. R²

1999 231 0.439*** -12.27 4.267*** -4.91 0.594 2000 257 0.358*** -6.5 6.029*** -3.58 0.34 2001 265 0.325*** -14.03 2.896*** -5.94 0.62 2002 279 0.356*** -16.43 3.290*** -17.11 0.841 2003 298 0.469*** -21.45 3.215*** -13.97 0.848 2004 342 0.421*** -12.66 5.046*** -22.35 0.9 2005 362 0.699*** -15.9 3.964*** -11.88 0.86 2006 388 0.228*** -5.07 16.236*** -43.23 0.948 2007 405 0.384*** -18.22 10.356*** -66.74 0.981 2008 426 0.427*** -26.35 7.358*** -58.44 0.972 2009 402 0.580*** -32.65 3.911*** -29.4 0.957 2010 427 0.177*** -8.74 10.130*** -67.56 0.972 2011 441 0.004 -0.21 13.523*** -83.99 0.976 2012 431 0.091*** -4.2 12.195*** -64.69 0.971 2013 368 0.085*** -3.1 13.784*** -53.08 0.964

Note: These three tables represents the earnings relation (Table 6-1-1), balance sheet relation (Table 6-1-2) and the book value and earnings relation (Table 6-1-3) for the non-lobbying bank sample over the period 1999 to 2013 covering 5363 firm-observations. Where NI is the total net income scaled by outstanding shares and ΔNI represents the changes therein. AST and LIAB represent both total assets and total liabilities scaled separately by outstanding shares. BV is the total book value of equity scaled by share price and NIPS is the net income scaled by share price. T-stat, test the hypothesis that the coefficient is different from zero. To reject this, a t-value of greater than 1.96 (95% confidence) is needed. The P-value (not shown in the tables), test the statistical significance on the independent variable in explaining the dependent variable where this value has to be lower than 0.05 level. The Adj.R² represents the amount of variance of the dependent variable explained by the independent variable, thus emphasize the explanatory power of the models.

Table 6-2-1: Regression Results of Lobby Banks (Earnings Relation)

fyearend N NI t-stat dNI t-stat Adj. R²

1999 16 -0.996 (-0.33) 0.224 -0.09 -0.134 2000 15 8.015 -1.45 -13.649*** (-3.20) 0.422 2001 16 6.885*** -4.47 -4.661*** (-3.34) 0.558 2002 17 4.435*** -6.66 -4.186*** (-4.02) 0.728 2003 18 4.915*** -4.07 -4.886*** (-3.77) 0.486 2004 19 5.837* -1.94 1.346*** -3.36 0.346 2005 19 3.067 -1.7 -3.899** (-2.51) 0.193 2006 19 3.597 -1.52 -6.691*** (-3.52) 0.371 2007 19 -3.22 (-0.61) 2.734 -0.52 -0.099 2008 20 1.927 -0.93 -1.117 (-0.48) 0.021 2009 19 -0.198 (-0.20) -1.363* (-2.02) 0.174 2010 21 1.944 -0.56 -0.576 (-1.16) -0.025 2011 21 0.303 -0.42 -0.282 (-0.38) -0.1 2012 21 2.348*** -4.04 -2.211*** (-3.00) 0.48

(29)

28

2013 22 -0.023 (-0.03) 0.236 -0.78 -0.07

Table 6-2-2: Regression Results of Lobby Banks (Balance Sheet Relation)

fyearend N AST t-stat LIAB t-stat Adj. R²

1999 16 0.251*** -7.01 -0.334*** (-5.77) 0.817 2000 15 0.367*** -5.69 -0.501*** (-4.83) 0.775 2001 16 0.268*** -5 -0.370*** (-4.19) 0.686 2002 17 0.350*** -4.79 -0.607*** (-4.15) 0.635 2003 18 0.227*** -4.48 -0.416*** (-3.76) 0.586 2004 19 0.289*** -4.87 -0.515*** (-3.93) 0.639 2005 19 0.120** -2.87 -0.143 (-1.55) 0.463 2006 19 0.127*** -4.5 -0.134** (-2.55) 0.706 2007 19 0.084** -2.59 -0.063 (-0.92) 0.574 2008 20 0.026** -2.49 0.007 -0.27 0.655 2009 19 0.002 -0.26 0.056 -1.48 0.564 2010 21 0.030*** -3.45 -0.051** (-2.17) 0.518 2011 21 0.004 -0.46 0.079** -2.2 0.664 2012 21 -0.007 (-0.65) 0.111** -2.1 0.518 2013 22 0.009 -0.86 0.06 -0.86 0.383

Table 6-2-3: Regression Results of Lobby Banks (Book Value and Earnings Relation)

fyearend N BV t-stat NIPS t-stat Adj. R²

1999 16 0.186 -1.48 13.312*** -15.76 0.988 2000 15 -0.108 (-0.36) 16.969*** -9.29 0.978 2001 16 1.383*** -13.21 8.353*** -11.55 0.986 2002 17 1.624*** -11.75 8.332*** -8.93 0.981 2003 18 1.226*** -19.55 3.192*** -11.29 0.983 2004 19 0.406* -2.07 13.023*** -9.82 0.984 2005 19 -0.27 (-1.37) 12.610*** -10.66 0.988 2006 19 0.756*** -3.48 6.520*** -5.08 0.991 2007 19 2.420*** -33.63 -7.185*** (-8.42) 0.988 2008 20 0.767*** -15.27 -2.029*** (-10.91) 0.974 2009 19 0.351*** -3.16 1.485 -1.41 0.644 2010 21 -0.243 (-1.44) 11.534*** -3.8 0.674 2011 21 0.752*** -8.84 -0.929 (-0.80) 0.857 2012 21 0.433*** -6.85 3.291** -2.87 0.758 2013 22 0.312*** -3.44 4.965*** -2.98 0.747

Note: These three tables represent the earnings relation (Table 6-2-1), balance sheet relation (Table 6-2-2) and the book value and earnings relation (Table 6-2-3) for the non-lobbying bank sample over the period 1999 to 2013 covering 282 firm-observations. Where NI is the total net income scaled by outstanding shares and ΔNI represents the changes therein. AST and LIAB represent both total assets and total liabilities scaled separately by outstanding shares. BV is the total book value of equity scaled by share price and NIPS is the net income scaled by share price. T-stat, test the hypothesis that the coefficient is different from zero. To reject this, a

(30)

t-29

value of greater than 1.96 (95% confidence) is needed. The P-value (not shown in the tables), test the statistical significance on the independent variable in explaining the dependent variable where this value has to be lower than 0.05 level. The Adj. R² represents the amount of variance of the dependent variable explained by the independent variable, thus emphasize the explanatory power of the models.

In accordance with Table 6 above, the following three graphs (Figure1-1, 1-2 and 1-3) are generated to represent the comparison of Adjusted R2 between lobby banks and non-lobby banks over tome for three different value relevance relations.

Figure 1-1. Adjusted R2 of Earnings Relation over Time

-20,00% -10,00% 0,00% 10,00% 20,00% 30,00% 40,00% 50,00% 60,00% 70,00% 80,00% 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013

Adjusted R² (%) For Earnings Relation

(31)

30

Figure 1-2. Adjusted R2 of Balance Sheet Relation over Time

Figure 1-3. Adjusted R2 of Book Value and Earnings over Time

0,00% 10,00% 20,00% 30,00% 40,00% 50,00% 60,00% 70,00% 80,00% 90,00% 100,00% 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013

Adjusted R² (%) For Balance Sheet Relation

R² (%) for Non-Lobbying Banks R² (%) for Lobbying Banks

0,00% 20,00% 40,00% 60,00% 80,00% 100,00% 120,00% 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013

Adjusetd R² (%) For Book Value and Earnings

Relation

(32)

31

4.1.3 Empirical Analysis of Earnings Relation

To examine if value relevance is related to the banks’ lobbying behavior, the value relevance of two separate bank groups are identified and tested. The concrete data and more vivid representation of the tests results are displayed in Table 6 and Figure 1, respectively. First of all, the result of F-test (Prob >F) is overall statistically significant (not shown in the table), which means that the chosen models are sufficient to predict the value of the proposed financial instruments standard. And the value relevance models should be sufficient due to the models practically has been tested in Hail (2013)’s paper.

Adjusted R-Square represents the fitting of the model, usually when Adjusted R-Square is closer to one, the equation is more fitting for the test, which means that the change of independent variables can better explain the change of dependent variables.

For the earnings model, the average Adjusted R-Square is 9.63% for non-lobby banks, which is lower than that of lobby banks (22.55%). And these results show that the value relevance for the two groups is both quite low. We can also see clearly form Table 7-1 that the Adjusted R-Square of both lobby banks and non-lobby banks fluctuate strongly during the period of 1999 to 2013. And the Adjusted R-Square of lobby banks displays an unstable trend, with more than 70% in year 2002 and below zero in year 2007 and 2011.

The coefficients for the two variables, NI and the change of NI, represent the respective association between the certain variable and the market return. For non-lobbying group, in most of the years, the association between the two independent variables and dependent variable are statistically significant (marked with stars to represent the significance). To be more specific, for NI variable, only in year 1999, 2000 and 2013, the coefficients are insignificant. And for the change in net income, only in year 2000 and 2002 there is

insignificant result. This means that the income variables are significantly associated with the market return of the specific firms and investors can reply on the representatives of income for market return.

As to the lobbying firm, we find a comparatively slight significance during the examining years. Only five years (2001,2002,2003,2004,2012) display significant results of net income and nine years have significant results for change in net income (2000-2006.2009 and 2012). Another thing worth mentioning here is that the coefficient of change in net income for lobbying banks have a majority of below zero numbers which lead to the fact that when the

Value relevance : related to banks’ lobbying behavior or not? : a study under the background of IFRS 9 project