• No results found

Determinants of financial statement preparerâ s involvement in the IASB standard settings process relating to IFRS 9 Phase 2 and earnings management : are there significant determinants which are associated with the choice whether to participate in the s

N/A
N/A
Protected

Academic year: 2021

Share "Determinants of financial statement preparerâ s involvement in the IASB standard settings process relating to IFRS 9 Phase 2 and earnings management : are there significant determinants which are associated with the choice whether to participate in the s"

Copied!
33
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

Amsterdam Business School

Determinants of financial statement preparer’s

involvement in the IASB standard settings process relating

to IFRS 9 Phase 2 and earnings management

“Are there significant determinants which are associated with the choice whether to participate in

the standard setting process of IFRS 9 Phase 2 and is this associated with earnings

management?”

Thesis

Author: Name: Tjeerd Renkema Version: final version

First reader: dr. Sanjay Bissessur Second reader: dr. ir. Sander van Triest Student number: 10282475

Date: 21-08-2014

MSc Accountancy & Control, specialization Accountancy Faculty of Economics and Business, University of Amsterdam Nr of words: 12.173

(2)

Abstract

This thesis provides a sample of worldwide firms to examine the underlying determinants whether to participate in the standard setting process on the exposure draft IFRS 9 Phase 2 within the banking sector. Furthermore this thesis measures whether there is a difference earnings management of firms who participate opposed to firms who do not participate in the standard settings process. While much researchers have focused on the development of other IFRS standards, to my knowledge this research is the first research that will focus on the development of IFRS 9 Phase 2. Furthermore this research is the first research that investigates the link between writing a comment letter and earnings management within the banking sector.

Prior researchers (Huian, 2013; ) suggest that the involvement of preparers in the standard setting process captures the standard setting process. Furthermore (Sutton, 1998) states that people (in this case preparers of financial statements) participate based on the rational-choice model, when the benefits outweigh the costs, in other words participation is out of own interest. This could be in contrast with presenting a true and fair view. Based on this we expect that there are significant determinants which decide whether preparers of financial statements participate directly in the standard setting process by sending a comment letter. We also measure the degree of earnings management by earnings volatility and reporting small positive earnings, which are both related to earnings management (Barth et al., 2008).

The results suggest that earnings volatility and reporting small positive earnings are significant determinants for the decision whether to participate in the standard setting process of IFRS 9 Phase 2. Both determinants are associated to earnings management, and the results show that firms who participate are associated with less earnings management. Furthermore the larger the firms are, the more likely they participate in the standard setting process of IFRS 9 Phase 2.

The IASB can take this result into account by interpreting the comment letters and possibly change the current standard setting process (i.e. the involvement of preparers of financial statements), because people who participate in the standard setting process also capture the standard setting process (Huian, 2013; ).

Key words: accounting, IFRS 9 Phase 2, impairment, incurred loss model, earnings management, IASB standard setting process.

(3)

Table of contents Introduction ... 4 1.1 Background ... 4 1.2 Research question ... 5 1.3 Contributions ... 6 2. Theory ... 8

2.1 Standard setting process ... 8

2.2 Exposure draft IFRS 9 Phase 2 as replacement for IAS 39 ... 10

2.2.1 Reason for issuing IFRS ... 10

2.2.2 IAS 39 ... 10

2.2.3 Main proposal IFRS 9 Phase 2 ... 12

2.2.4 Increase in credit risk ... 13

2.2.5 Exemption ... 14

2.3 Earnings management ... 14

2.3.1 Definition earnings management ... 14

2.3.2 Earnings management incentives within the banking sector ... 15

2.3.3 Equity valuation of loan losses ... 16

2.3.4 Measuring earnings management ... 17

2.4 Hypotheses ... 17

3. Method ... 19

3.1 Methodology ... 19

3.2 Data and sample selection ... 21

4. Results... 26

4.1.1 Regression analyses earnings smoothing ... 26

4.1.2 Regression analyses managing towards small positive earnings ... 28

5. Conclusion ... 30

5.1 Conclusion... 30

5.2 Limitations... 31

5.3 Suggestions for future research ... 31

(4)

Introduction

1.1 Background

Is the involvement in the IASB standard setting process based on good intentions or self-interest?

The International Accounting Standards Board (IASB) and the Financial Accounting Standards Board (FASB) are using an incurred loss model1 to determine when financial assets and liabilities should be impaired2. This standard in which the incurred loss model is described is called ‘international accounting standard (IAS) 39’. The Report of the Financial Crisis Advisory Group (Financial Crisis Advisory Group, 2009) believes that applying the incurred loss model has delayed the recognition of costs as a consequence of impairment (which is referred to as loan-losses) during the crisis. The incurred loss model was initially devised to avoid the “big bath”3 provisions. However as the financial crisis began, the incurred loss model led to the delay of loan-loss provision. Under the old IAS 39 there should be evidence for impairment, which causes delay of recognition in loan-losses. Furthermore the site of the IFRS states that also the Group of Twenty Finance Ministers and Central Bank Governors, a group of finance ministers and central bank governors from 20 major economies (also known as the G-20), requests for a more forward-looking impairment model that reflects expected loan-losses. As such the IFRS started working on the replacement of IAS 39, called the IFRS 9 Phase 2, which includes an expected credit loss model instead of the incurred loss model. The expected credit loss model is further explained in section 2.

The standards are developed by the IASB through an international consultation process. The process knows fixed stages, which is further explained in section 2. The process of developing the new standard started in 2008, however due to the large number of comment letters the IASB decided to write 2 exposure drafts, and the process is still not finalized as per today. However the comment letters are all received and due to the transparency of the IASB these comment letters are also available to the public. As the standards are developed during an

1The current “incurred loss model” of international accounting standard 39 solely permits the recognition of credit

losses at the moment a triggering event of impairment occurs. A triggering event occurs when the entity is experiencing notable financial difficulties, has defaulted on or is late making interest payments or principal payments, is likely to undergo a major financial reorganization or enter bankruptcy, or is in a market that is experiencing significant negative economic change.

2 Reduction in the value of an asset because the asset no longer generates the benefits expected earlier as determined

by the company through periodic assessments.

3 Big Bath in accounting is a technique whereby a one-time charge is taken against income in order to reduce assets,

which results in lower expenses in the future and in lower net income for the current year. The objective is to ‘take one big bath’ in a single year so future years will show increased net income.

(5)

international consultation process the motivation of participation of preparers of financial statements is important to enhance our understanding of the IASB decision making (Barth, 2000). This paper helps to get insight in the different determinants which are associated with the choice whether to participate or not. Participation within the standard setting process can be done in several ways, however we build on the assumption that participation in the exposure draft results in the submission of comment letters (Kosi and Reither, 2014).

As the IASB states on their website the new model under IFRS 9 Phase 2 has the potential to incur loan-losses on a timelier basis. EY (March 2013) states in their report that under the new developed IFRS 9 Phase 2 exposure draft “the expected credit loss model will likely result in earlier recognition of loan losses compared to the current incurred loss”. The question is whether the preparers of the financial statements are pleased with the new regulation, as the new regulation implies that losses on financial instruments should be taken into the profit and loss account in a timelier manner.

Over the last few decades, much research has been performed on the standard setting process of the IASB and earnings management. Prior research suggests that the involvement of preparers in the standard setting process captures the standard setting process (Huian, 2013 ). Furthermore Sutton (1998) states that people (in this case preparers of financial statements) participate based on the rational-choice model, when the benefits (signaling) outweigh the costs (sending a comment letter). Prior research states that entities must have positive earnings and meet forecast in order to acquire capital (Degeorge et al., 1999), the new exposure draft could have a negative impact on earnings. We investigate two measures of earnings management, which are earnings smoothing and managing towards positive earnings. Prior research suggests that a higher degree of earnings smoothing is related to a lower degree of earnings volatility and that a higher degree of managing towards positive earnings (i.e. report more small positive earnings) is related to a higher degree earnings management (Barth et al. , 2008).

1.2 Research question

Based on the above background, my research question can be formulated as follows:

“Are there significant determinants which are associated with the choice whether to participate in the standard setting process of IFRS 9 Phase 2 and is this associated with earnings management?”

The research question investigates whether there are significant determinants in the choice whether to participate in the standard settings process and whether the degree of earnings

(6)

did not write a comment letter. This research question is interesting because it helps to get insight in the different determinants which are associated with the choice whether to participate or not. If companies who have more earnings management practices are involved with the standard setting process, and prior literature suggests that people who are involved capture the standard setting process, we can question whether this involvement option should be maintained. Furthermore if the opposite is true, that companies who are involved have a lower degree of earnings management, we can question if companies only participate due to signaling a positive sign to the market.

The research is done via 2 regression models. The first regression model measures whether there is a difference in earnings volatility between companies who wrote a comment letter opposed to companies who did not write a comment letter. The second regression model measures whether there is a difference in the degree of reporting small positive earnings between companies who wrote a comment letter opposed to companies who did not write a comment letter. By both regression models we also use 5 control variables consistent with prior literature. For the definition of the variables we refer to paragraph 3.2. The results show that there is a significant difference in earnings management measured by earnings volatility and reporting small positive earnings between companies who wrote a comment letter opposed to companies who did not write a comment letter. The difference is that companies who wrote a comment letter exhibit less earnings management than companies who did not write a comment letter. Furthermore consistent with prior research firms who are larger and have a higher ROA are more likely to participate in the standard setting process of IFRS 9 Phase 2 (Koh, 2011).

1.3 Contributions

Prior research has found why people participate in the standard setting process and how this influences the standards. Furthermore prior research has found why earnings management is used. “The loan-loss expense is typically a significant item in the financial statements of banks. The consequent potential for loan-loss provisioning, which is related to earnings management, to have significant economic effects, including by contributing to procyclicality, has been documented in a number of studies” (Hanlon, 2013). This research contributes to accounting literature in a number of ways. First, to my knowledge this study is the first of its kind that investigates the relationship between writing a comment letter and earnings management under IFRS 9 Phase 2 within the banking sector.

Furthermore the goal of IASB standards is that the financial statements become more transparent and of higher quality. With the current standard setting process the IFRS aims for an

(7)

open and transparent due process of which the publication of consultative documents, such as discussion papers and exposure drafts, for public comment is an important component for the achievement of higher quality standards and thus higher quality financial statements. Knowing whether determinants of preparers influence the participation choice is important for the IASB to improve the standard setting process. Prior research (Huian, 2013; ) suggests that the involvement of preparers in the standard setting process captures the standard setting process. This research finds that preparers which are involved in the standard process are associated with less earnings management practices. Furthermore the larger the firms are and the larger the ROA is, the more likely they participate in the standard setting process. With these results the IASB can interpret the comment letters in a better way and can think about future changes to the process to enhance the quality of the process.

Section 2 of this thesis discusses the main findings of prior literature and describes the hypotheses. Section 3 describes the data and selection method and also gives the regression model en estimation method. Section 4 provides the results. In section 5 the conclusion of this thesis is given.

(8)

2. Theory

The theory section of this thesis is as follows. Section 2.1 describes the standard setting process of the IASB. In chapter 2.2 the new regulation IFRS 9 Phase 2 is described and the differences with the old IAS 39 regulation. In section 2.3 the earnings management within the banking sector is described. We conclude the chapter with describing the hypotheses of this paper.

2.1 Standard setting process

The goal of IFRS standards is that the financial statement become more transparent and of higher quality. With the current standard setting process the IFRS aims for an open and transparent due process of which the publication of consultative documents, such as discussion papers and exposure drafts, for public comment is an important component for the achievement of higher quality standards and thus higher quality financial statements. Quality of financial statements depends on the extent to which the financial statements represent a true and fair view of the financial situation of the firm (Conceptual Framework for Financial Reporting, 2010). As for the paper the standard setting process is important, we elaborate the process in short below. The process knows 6 fixed stages:

1. Setting the agenda, based on the requirement of financial statement users with the aim to increase reliability, relevance and convergence;

2. Planning the project, the composition of the team will be determined (possibly with other institutions);

3. Developing and publishing the discussion paper, the discussion paper is not mandatory, the IASB normally publishes it as its first publication on any major new topic to explain the issue and solicit early comment from constituents;

4. Development and publication of an exposure draft, the exposure draft is mandatory and sets out a specific proposal in the form of a proposed standard, the public can give their comments in the form of a comment letter;

5. Developing and publishing the standard, step 4 could be re-done when the IASB decides so. When the proposal is finalized it must be approved by the IFRIC. (International Financial Reporting Interpretations Committee). Finally, after the due process is completed, all outstanding issues are resolved, and the IASB members have balloted in favor of publication, the IFRS is issued;

6. After the standard is issued, after an IFRS is issued, the staff and the IASB members hold regular meetings with interested parties, including other

(9)

standard-setting bodies, to help understand unanticipated issues related to the practical implementation and potential impact of its proposals.

This research focusses on step 4, which gives stakeholders the opportunity to comment on the proposed standard by writing a comment letter. Participation can be divided in direct participation and indirect participation. Direct participation is writing a comment letter directly on behalf of the firm, indirect participation is submitting comments to, for example, auditors or organizations who serve the interests of a sector. Both methods are used by preparers of financial statements, however in general prior research suggests that preparers of financial statements perceive the direct methods as more effective compared to indirect methods (Georgiou, 2010). Furthermore Kosi and Reither (2014) state: “Empirical studies investigating lobbying behavior of constituents focus primarily on formal participation via comment letters for several reasons: (i) letters are publicly available documents, (ii) formal participation became more important under the IASB’s regime with independent board members compared to the International Accounting Standards Committee (IASC) with members representing the views of particular constituent groups (iii) there is a strong correlation between submission of comment letters and other lobbying methods , and (iv) comment letters significantly influence the form of the final standard”. Our paper is also based on the assumption that participation in the exposure draft IFRS 9 Phase 2 results in the submission of comment letters by constituents.

We can divide the stakeholders in different groups, such as but not limited to the users of financial statements, the auditors of financial statements and preparers of financial statements. Within this research we focus on the preparers of financial statements as we research the relationship of earnings management by preparers and the determinants for writing a comment letter. All written comment letters are available on the website of the IASB.

Various research is performed with regard to the standard setting process (Begona Giner & Miguel Arce, 2012) show that the preparers of the financial statements are involved at the start of the standard setting process and this is in line with rational-choice model. Sutton (1998) developed the rational-choice model, which implies that participation of a group only exists when the benefits exceed the cost of lobbying. Prior research such as Huian (2013) and Helen Irvine (2010) suggests that the involvement of preparers in the standard setting process captures the standard setting process. With prior literature we can conclude that the involvement of preparers of financial statements can be questioned. They only participate when they benefit from it, the benefits should be well researched and it should be assessed whether this increases or decreases the quality of financial statements.

(10)

Our focus is determining which determinants have a significant effect on the choice whether to participate in the standard setting process of IFRS 9 Phase 2 by direct participation, writing a comment letter. To understand the potential drivers of participation we need to know what the new exposure draft indicates and what possible implications can be. We therefore elaborate on the exposure draft in paragraph 2.2.

2.2 Exposure draft IFRS 9 Phase 2 as replacement for IAS 39 2.2.1 Reason for issuing IFRS

Exposure draft IFRS 9 Phase 2 (hereafter the exposure draft) is the proposed replacement of IAS 39. The reason for issuing the exposure draft is the delayed recognition of credit losses4 which are associated with financial instruments during the financial crises under IAS 39. The current “incurred loss model” of IAS 39 solely permits the recognition of credit losses at the moment a credit loss occurs. The Financial Advisory Group5 recommended to search for alternatives with regard to the “incurred loss model”.

The objective of IFRS 9 Phase 2 is to enhance the insight in expected credit losses for users of financial statements and to simplify the determination of expected credit losses for the preparers of financial statements. This is further explained in paragraph 2.2.3, 2.2.4 and 2.2.5. 2.2.2 IAS 39

IAS 39 outlines the recognition and requirements for the recognition and measurement of financial assets, financial liabilities, and some contracts to buy or sell non-financial items. Under IAS 39 there are four categories to measure the financial instruments, which are explained below. The first group are “Financial assets at fair value through profit or loss”. This category has two subcategories, which are (1) Designated, this includes any financial asset that is designated on initial recognition as one to be measured at fair value with fair value changes in profit or loss and (2) Held for trading, which includes financial assets that are held for trading and all derivatives (except those designated hedging instruments). The second group is “Available-for-sale financial assets (AFS)”. This group includes any non-derivative financial assets designated on initial recognition as available or any other assets that do not fall within the other 3 groups. These assets are measured at fair value in the balance sheet, the changes in fair value are recognized directly in equity, through the statement of changes in equity, except for

4 Credit losses are referred to as short fall in cash flows.

5 In October 2008 the IASB and the FASB set up the Financial Crisis Advisory Group to deal with financial

(11)

impairment losses, interest and foreign exchange gains or losses. When the asset is derecognized the cumulative change in value which was recognized in equity is transferred to the profit and loss account. The third group “Loans and receivables” which are financial assets with fixed or determinable payments that are not quoted in an active market. The fourth group is “Held-to-maturity investments” which are non-derivative financial assets with fixed or determinable payments that an entity intends and is able to hold to maturity.

Initially measurement of financial assets and liabilities is fair value, including transaction costs for assets and liabilities not measured at fair value through profit or loss (IAS 39.43). Subsequently, financial assets and liabilities (including derivatives) should be measured at fair value. However there are some exceptions for the measurement at fair value (IAS 39.46-47). Loans and receivables, held-to-maturity investments, and non-derivative financial liabilities are measured at amortized cost using the effective interest method. Investments in equity instruments with no reliable fair value measurement (and derivatives indexed to such equity instruments) should be measured at cost. Financial assets and liabilities that are designated as a hedged item or hedging instrument are subject to measurement under the hedge accounting requirements of the IAS 39.

The fair value can be described as “the amount for which an asset could be exchanged, or a liability settled, between knowledgeable, willing parties in an arm's length transaction” (IAS 39.9). Furthermore ISA 39 prescribes a hierarchy to be used in determining the fair value for a financial instrument. In first instance quoted market prices in an active market should be used. If this market is not active, the entity must use a valuation technique that makes maximum use of market inputs.

If there is no active market for an equity instrument and the range market inputs is limited, then an entity must measure the equity instrument at cost less impairment.

Impairment has to be assessed at each reporting date. A financial asset is impaired, and impairment losses are recognized, only when objective evidence is obtained after the initial recognition. If such even evidence exists the enitity must perform a detailed impairment analyses to determine the amount of impairment ([IAS 39.64). Impairment must be reversed if events occurs after impairment which justifies this, with the exception of investments available-for-sale ([IAS 39.65). Using IAS 39 is also described as the “incurred loss model” which assumes that all loans will be repaid until evidence to the contrary (known as a loss or trigger event) is identified. Only at that point is the impaired loan (or portfolio of loans) written down to a lower value. The Report of the Financial Crisis Advisory Group (Financial Crisis Advisory Group, 2009), beliefs that applying the incurred loss model has delayed the recognition of costs as a consequence of

(12)

impairment (which is referred to as loan-losses) during the crisis. The incurred loss model only recognize losses when they actually occur. When management estimates that client can no longer repay their debts, management has to wait until the client is actually in default. From that moment the bank can write off the loans. The incurred loss model was initially devised to avoid the “big bath” provisions. However as the financial crisis begun, the incurred loss model led to the delay of loan-loss provision. Under IAS 39 there should be evidence for impairment, which causes delay of recognition in loan-losses. Furthermore the site of the IFRS states that also the Group of Twenty Finance Ministers and Central Bank Governors, a group of finance ministers and central bank governors from 20 major economies (also known as the G-20), requests for a more forward-looking impairment model that reflects expected loan-losses As such the replacement of IAS 39 is proposed by IFRS 9, of which we will explain IFRS 9 Phase 2 “impairment” in the next paragraph.

2.2.3 Main proposal IFRS 9 Phase 2

The exposure draft requires the entity to assess magnitude of possible credit losses as at the reporting date and to incorporate this in the financial statements as a loss allowance for financial assets and as a provision for commitments to extend credit, which I will refer to as credit losses on financial instruments. The main difference with IAS 39 is that credit losses are no longer based on events which occurred but on a broader assessment which includes past events, current conditions and future events with regard to expected credit losses.

The exposure draft distinguishes two situations for changes in credit quality. First, financial instruments which did not deteriorate significantly in credit quality since initial recognition, for these items a 12-month expected credit loss must be calculated. Second, financial instruments that deteriorated significantly since initial recognition, for these items a lifetime expected credit loss must be calculated. With these two situations the exposure draft distinguishes three stages, which differ in accounting method through 1) calculation of expected credit loss (allowance) and 2) for financial assets the calculation of interest revenue. The three stage are:

1) Financial instruments which did not deteriorate significantly in credit quality or that have low credit risk(refer to paragraph 2.2.3 for an explanation of credit risk);

2) Financial instruments which deteriorated significantly in credit quality (unless they have low credit risk > stage 1) and that do not have objective evidence of a credit loss; 3) Financial instruments that have objective evidence of impairment.

The figure below describes how the expected credit loss and the calculation of interest must be accounted for:

(13)

To simplify the calculation of credit losses exemptions are made, refer to paragraph “2.2.4 exemptions”.

2.2.4 Increase in credit risk

A significant increase in credit risk is an important variable in determining the credit loss, therefore this paragraph will explain how to determine a significant increase in credit risk. Credit risk is low if a default is not imminent and any adverse economic conditions or changing circumstances may lead to, at most, a weakened capacity of the borrower to meet its contractual cash flow obligations on the financial instrument. For example, a loan that has an internal credit risk rating equivalent to the external credit rating of ‘investment grade’6 would be considered to have a low credit risk (ED IFRS, March 2013).

Credit risk can best be assessed by comparing the probability in default for the remaining lifetime at reporting date compared relatively7 to the probability in default at initial recognition. Also the remaining lifetime should be considered, as the longer the remaining lifetime, the bigger the probability of default. For example, two financial assets which are purchased at the same date and assessed at purchase date + 2 years with different maturity date cannot be assessed in a similar manner. Furthermore the standard gives a quantitative measurement of 30 days overdue in payments which implies a significant increase in credit risk, however this can be rebutted when past information indicates that 30 days cannot be associated with a causal link between an event of default. The assessment must be performed each reporting date and can also result in a reversal of credit losses, for example when the criteria for a significant increase in credit risk are

6 The meaning of “investment grade” in this context can be described as counterparties which are judged by an

external credit rating agency such as Moody's Investors Service and Standard & Poor's “that are likely enough to meet payment obligations”.

(14)

no longer achieved and thus the credit losses should be calculated on a 12-month expected base. The amount of credit losses (or reversal) is recognized in the profit and loss account.

2.2.5 Exemption

Due to the costs of assessing whether trade and lease receivables encounter a significant increase in credit risk, the IASB simplifies this assessment:

1) Credit losses on short-term trade receivables are in general calculated following the lifetime expected credit losses, specifically for trade receivables that do not constitute a financing transaction in accordance with IAS 18 Revenue.

2) For credit losses on long-term trade and lease receivables, the entity can choose an accounting policy to always recognize a loss allowance following the lifetime expected credit losses, specifically for trade receivables that constitute a financing transaction in accordance with IAS 18 Revenue. It goes behind the aim of this research to elaborate in more detail about the exemptions.

2.3 Earnings management

2.3.1 Definition earnings management

Different definitions have been used in prior studies to define earnings management. Schipper (1989) defines earnings management as follows: “A purposeful intervention in the external financial reporting process, with the intent of obtaining private gain (as opposed to, say, merely facilitating the neutral operation)”, which can be interpreted as intentional influencing the financial reporting of a company.

A more detailed explanation: “Earnings management occurs when managers use judgment in financial reporting and in structuring transactions to alter financial reports to either mislead some stakeholders about the underlying economic performance of the company or to influence contractual outcomes that depend on reported accounting practices” (Healy and Wahlen, 1999).

Within the context of this research entities must show that they can achieve profits and this profits are not volatile. To do this entities must have positive earnings and meet forecast in order to acquire capital (Degeorge et al., 1999). Entities will not always meet forecast due to different reasons and as a consequence entities will try to manage earnings. Earnings consist of cash flows from operations and accruals, so to manage earnings managers can either adjust the real operations (real earnings management) or the accruals (accrual-based earnings management) (Rowchowdhury, 2006). We investigate two measures of earnings management, which are

(15)

earnings smoothing and managing towards positive earnings. Prior research suggests that a higher degree of earnings smoothing is related to a lower degree of earnings volatility and that a higher degree of managing towards positive earnings (i.e. report more small positive earnings) is related to a higher degree of earnings management (Barth et al. , 2008). This research does not aim to prove that participants in the standard setting process of IFRS 9 Phase 2 have lower or higher earnings management, it only hypothesis that there are determinants for participation, and the determinants are associated to the degree of earnings management.

2.3.2 Earnings management incentives within the banking sector

Prior literature has focused on the use of loan-loss provisions (LLP’s). Hanlon (2013) states that “the loan-loss expense is typically a significant item in the financial statements of banks. The consequent potential for loan-loss provisioning to have significant economic effects, including by contributing to procyclicality, has been documented in a number of studies”. This implies the importance for the loan-loss provisioning. The purpose of the provision is to adjust the loan-loss reserves so that the loan portfolios reflect the future expected losses.

Besides the aspect of reflecting future expected losses managers have incentives to manage earnings as managers try to meet forecasts in order to acquire capital (Degeorge et al., 1999). Furthermore managers can have incentives to manage the regulatory capital8 or have an incentive to engage in earnings management since reduced volatility is assumed to convey a signal of lower risk, less volatile earnings are fundamental predicate for stable stock prices (Ahmed et al., 1999).

In general managers can also use earnings management to gain their own benefit. This is due to the fact that they want to have job security or have gain from certain compensation plans. Prior research suggests that management engages more in earnings management when compensation is closely related to the value of the stocks (Cohen et al., 2014).

Beatty et al. (2002) examine the incentives by banks to use earnings management by comparing the use of discretionary provisions to avoid earnings decreases for publicly traded versus privately held banks. The findings suggest that publicly traded banks use more discretion in the loan-loss provision to achieve earnings targets compared to private banks. This findings suggest that the discretion in the loan-loss provision is related to earnings management incentives. Our research does not measure the degree of loan-loss accruals, we measure the earnings management in two other ways. As the new IFRS 9 Phase 2 is related to the loan-loss

(16)

provision and we measure earnings management in another way, we can also prove the association between earnings management and the incentives and writing a comment letter on IFRS 9 Phase 2, which is related to the use of loan-loss provisions.

2.3.3 Equity valuation of loan losses

There exists an information asymmetry between the managers and less informed investors (Beatty and Liao, 2013). This information asymmetry can be used by managers in their advantage by using this information to engage in earnings management so that the information reported is in line with investors’ expectations. Loan-loss provisions contain information which is investigated in several studies by examining the relationship between stock returns and the level of provisions. This is important for this research for the impact of the results. Because when loan-loss provisions are an important factor for investors, these loan-loss provisions are influenced by earnings management, it is important for investors whether companies are involved with earnings management. Furthermore it is important for the IASB, because the goal of the IASB is that the financial statements become more transparent and of higher quality and the boards propose that the objective of general purpose financial reporting is to provide financial information about the reporting entity that is useful to present and potential equity investors, lenders, and other creditors in making decisions in their capacity as capital providers (IASB).

Prior studies have mixed findings regarding the investor’s reaction on the loan losses. Some research (Beaver et al., 1998) finds a positive association between the market value and the loan-loss reserves. Other research (Liu et al., 1997) also finds a positive relation and assigns this to the fact that loan losses are a signal of the intention and ability to deal with bad performing loans. This research is very limited (Beatty and Liao, 2013), both the quantity and scope and the findings are based on data before the BASEL regulation was in place. The BASEL regulation is mentioned because prior to the BASEL regulation the loan-loss provision was decreasing the reported earnings by the loan-loss provision minus the applicable taxes. However the provision could be added back to the regulatory capital, which increased the regulatory capital net by the tax expense. The counter intuitive effect was solved within the BASEL Tier 1 capital. As such no positive addings (i.e. the tax effects) to the regulatory capital were allowed.

More recent research (Anwer et al., 2006) shows that the recognition of losses does have a negative impact on the valuation by investors of assets, more specific, the recognition has more impact than disclosures. This is also an important facet for this research, as the new regulations of IFRS 9 Phase 2 require recognition of the loan losses. As proven in prior research that this

(17)

has a negative impact on the valuation by investors, this could lead to incentives for managers to become involved with earnings management.

2.3.4 Measuring earnings management

This research investigates inter alia whether preparers of financial statements who wrote a comment letter on the new IFRS 9 Phase 2 standard have more earnings management incentives than the preparers of financial statements who did not write a comment letter. The comment letter could be written because preparers are concerned with the new regulation regarding IFRS 9 Phase 2 which includes the loan-loss provision as explained in paragraph 2.2.2.

A good way to measure the earnings management is the degree of earnings smoothing and earnings volatility. Prior research suggests that a higher degree of earnings smoothing is related to a lower degree of earnings volatility and that a higher degree of managing towards positive earnings (i.e. report more small positive earnings) is related to a higher degree of earnings management (Barth et al. , 2008).

2.4 Hypotheses

There is extensive research on participation in the standard process of other standards. Our focus is on the preparers of financial statements which participate in the standard setting process of IFRS 9 Phase 2. Following the rational-choice model of lobbying participants will only participate in the standard setting process when the benefits exceed the costs (Sutton, 1998). The benefits of participating could have several reasons. For example preparers of financial statements which exhibit a high earnings volatility could have a lower earnings volatility with the new exposure draft and comment through a comment letter noting they agree with the new standards. The opposite could also be true, when preparers of financial statements have a low earnings volatility and the exposure draft implications could lead to a higher/lower earnings volatility, they comment on the exposure draft that they disagree respectively agree with the exposure draft. The benefits of disagreeing would be that the IASB possibly changes the exposure draft which lowers the earnings volatility, the possible benefits of agreeing would by sending a positive signal to the market. Possible costs would be that preparers will have to recognize the loan-loss provisions earlier compared to the old standards IAS 39 and will have less opportunity to manage their earnings upward. Following the positive accounting theory (Watts and Zimmerman, 1978), corporate participation in the due process is driven by the economic consequences of the proposed standard for a firm’s accounting numbers or cash flows. These economic consequences have been linked to several corporate characteristics such

(18)

as firm size, leverage and ownership structure (Durocher et al., 2007). Other characteristics could be earnings volatility and management towards small positive earnings. We do not hypothesise in which way, negative or positive, these determinants are associated with sending a comment letter. We only hypothesise that these determinants have a significant relationship with sending a comment letter. Therefore my hypotheses are as follows:

H1: Income volatility affects a financial firm’s decision to lobby by writing a comment letter.

H2: Reporting small positive earnings affects a financial firm’s decision to lobby by writing a comment letter.

(19)

3. Method

3.1 Methodology

The hypotheses relating to determinants of participation in the standard setting process are tested with two metrics which measure the degree of earnings management. We compare lobbyers and non-lobbyers performing univariate tests on the main test variables. The two metrics are related to earnings smoothing and managing towards positive earnings. Prior literature suggests that a lower degree of earnings volatility is related to a higher degree of earnings management. Furthermore prior literature suggests that a higher degree of reporting small positive earnings is related to a higher degree of earnings management (Barth et al. , 2008). The earnings volatility is the change in net income9 between the years, for an exact definition of earnings volatility we refer to the legend of the regression models.

These two metrics to measure accounting quality, reflect effects attributable to the financial reporting system and those un-attributable to the financial reporting system. As our research focusses on companies within the banking sector and only the largest banks in the world we do not have to control for industry specific effects. As the sample only reflects countries which sent at least one comment letter, the effect of country-level differences is assumed to be limited. Consistent with prior literature (Pagano, 2002) we do use some control variables to mitigate other effects which are size, the degree of leverage, return on assets, the degree of change in cash flows and the market-to-book ratio. The multivariate analyses allow us to also describe other determinants in the participation of the standard setting process. We did not identify the use of a big-4 auditor as control variable, as almost all the largest banks in the world are audited by a big-4 auditor. The logistic model to measure the degree of earnings volatility is:

LETTER = β0 + β1EVOL + β2SIZE + β3LEV + β4ROA + β5CFVOL + β6MTB + ϵ Where:

EVOL = Average earnings volatility

SIZE = Average natural logarithm of end of year market value of equity

LEV = Average end of year total liabilities divided by end of year

equity book value

9 14 DataStream provides different definitions of operating income. The one we use does not include extra items or

preferred dividend. As all companies are reporting under IFRS we assume that this income does not differ across the sample.

(20)

ROA = Average income divided by end of year total assets CFVOL = Average Cash flow volatility

MTB = Average Outstanding share * price / common equity Please note that for an exact definition of earnings volatility we refer to the legend of the regression model in table 5. The coefficient β1 is of main interest in this model. A negative coefficient is consistent with the fact that sending a comment letter is associated with less earnings volatility (i.e. more earnings management). A positive coefficient is consistent with the fact that sending a comment letter is associated with more earnings volatility (i.e. less earnings management). We do not have an expectation of this coefficient, because this depends on the content of the comment letter. When firms strongly disagree with the content of the new exposure draft in their comment letter we would expect a negative relationship (i.e. companies are associated with earnings management). However when companies agree with the content of the new exposure draft in their comment letter it could be that companies exhibit a positive relationship.

Furthermore prior literature argues that larger firms are more likely to lobby because they have the resources to do so and the possible gains are larger than the costs of participation (Sutton, 1984). Also larger companies have more incentives because prior empirical studies show that larger companies have a stronger influence than small companies (Koh, 2011). These arguments are similar to the expectation of the ROA, as more profitable firms are expected to have more resources and the ROA is calculated by the total assets / profit. We expect a possible relationship between the direct lobbying decision by sending a comment letter and SIZE / ROA. For the other variables we do not have expectations.

The logistic model to measure the degree of small positive earnings is:

LETTER = β0 + β1SPOS + β2SIZE + β3LEV + β4ROA + β5CFVOL + β6MTB + ϵ Where:

SPOS = indicator that equals 1 for observations with annual earnings scaled by total assets between 0.00 and 0.01 SIZE = The natural logarithm of end of year market value of

equity

LEV = end of year total liabilities divided by end of year equity

book value

ROA = Income divided by end of year total assets CFVOL = Cash flow volatility

(21)

Please note that for an exact definition of earnings volatility we refer to the legend of the regression model table 6. The coefficient β1SPOS is of interest in this model. A negative coefficient is consistent with the fact that sending a comment letter is associated with less reporting of small positive earnings (i.e. less earnings management). A positive coefficient is consistent with the fact that sending a comment letter is associated with more reporting of small positive earnings (i.e. more earnings management). We do not have an expectation of this coefficient, because this depends on the content of the comment letter. When firms strongly disagree with the content of the new exposure draft in their comment letter we would expect a negative relationship (i.e. companies are associated with earnings management). However when companies agree with the content of the new exposure draft in their comment letter it could be that companies exhibit a positive relationship. For the expectations of the other variables we refer to page 16.

3.2 Data and sample selection

The data for the comment letters on the exposure draft IFRS 9 Phase 2 are obtained from the website of the IFRS. All comment letters can be downloaded from the website of the IFRS and include the name of the person who wrote the comment letter and the company on behalf of which the comment letter was written. We are only interested in the comment letters of the preparers of financial statements within the banking sector. After reviewing the comment letters we obtained 31 comment letters of preparers of financial statements within the banking sector.

To gain information about these banks and the control group (i.e. the firms who did not write a comment letter) we obtained data from Datastream of the 533 largest banks of the world between 2007 and 2013. All 31 banks who wrote a comment letter were part of these 533 largest banks, we labelled these bank with a dummy variable 1 in our dataset. We only used the firm’s years 2007-2013, as the financial crises began in 2007, to control for economic differences. Our sample collected initially included 3752 fiscal years’ observations. The sample includes both firms that did and did not write a comment letter. After deleting double records and records of which no data was available 1617 records remain for testing. To control for country differences we only used the countries of which at least one company wrote a comment letter, after filtering 617 firm years remained. We then computed our variables over all the firm years and took the average per firm of all firm years. Finally 105 companies with an average for the variables remained of which 31 wrote a comment letter. We ran the regression on these 105 companies of which 5 missed some variables, our final sample included 100 firm years.

(22)

Table 1 includes the country breakdown for our sample. The sample consists of many countries of which the greatest representation is of Switzerland and the United States. The United Kingdom and the United States included the firm years with the most comment letters.

Table 2 presents a breakdown of fiscal years with the count of fiscal year observations divided in fiscal years with a comment letter and fiscal years without a comment letter. Approximately all years have an identical number of firm years used in the regression.

Table 3 presents the descriptive statistics of the scaled variables used in the regression analyses. Note that the mean of LETTER is 0.295, which means approximately one third of our sample firms wrote a comment letter. The mean of SPOS is 0.743, which indicates that one fourth reported small positive earnings. Knowing that the financial crisis began during 2007, this doesn’t seem unlikely.

We also tested for multi-correlation between the control variables using a Spearman correlation test. We note that there is a strong significant positive (defined as >0,4) relationship between SIZE and LETTER (0,594) and MTB and ROA (0,575) and a strong significant negative relationship (defined as <-0,4) between SPOS and EVOL -0,597. However both values are not above 0,7 / -0,7 which indicates a higher than normal relationship, and as such we conclude that no multi-correlation exists which could impact the results between the variables.

Table 1

Descriptive Statistics Relating to Countries used in the regression

Financial year Number of Firm-Year Observations Percentage of Firm-Year

Observations Number of CL Firms Percentage of NCL Firm

AUSTRALIA 36 5,83% 6 3,23% AUSTRIA 30 4,86% 6 3,23% FRANCE 54 8,75% 18 9,68% GERMANY 36 5,83% 12 6,45% NETHERLANDS 12 1,94% 6 3,23% PORTUGAL 24 3,89% 6 3,23% SPAIN 41 6,65% 12 6,45% SWITZERLAND 118 19,12% 6 3,23% UNITED KINGDOM 47 7,62% 42 22,58% UNITED STATES 219 35,49% 72 38,71% Total 617 100,00% 186 100,00%

Column "Number of Firm-Year Observations" shows the total of firm years used within the regression. The column "Number of CL Firms" shows the number of firm years for which a comment letter is written, the definition of this is a comment letter is written in 2007, 2008, 2009, 2010, 2011 and 2012 when the firms wrote a comment letter in 2013 within a company of a country.

(23)

Table 2

Descriptive Statistics Relating to Frms Years used in the regression

Financial year Number of Firm-Year Observations Percentage of Firm-Year

Observations Number of CL Firms Percentage of NCL Firm

2007 99 16,05% 31 16,67% 2008 101 16,37% 31 16,67% 2009 103 16,69% 31 16,67% 2010 104 16,86% 31 16,67% 2011 105 17,02% 31 16,67% 2012 105 17,02% 31 16,67% Total 617 100,00% 186 100,00%

Column "Number of Firm-Year Observations" shows the total of firm years used within the regression. The column "Number of CL Firms" shows the number of firm years for which a comment letter is written, the definition of this is a comment letter is written in 2007, 2008, 2009, 2010, 2011 and 2012 when the firms wrote a comment letter in 2013.

(24)

Table 3

Descriptive Statistics Relating to Variables Used in Analyses

Variable N Mean Std. Deviation p25 p50 p75

LETTER 105 0,295 0,458 - - 1,00 EVOL 104 0,006 0,007 0,00 0,00 0,01 SIZE 103 15,448 1,590 14,36 15,14 16,83 LEV 105 0,424 0,421 0,18 0,30 0,52 ROA 105 0,006 0,013 0,00 0,01 0,01 CFVOL 101 0,013 0,011 0,01 0,01 0,02 MTB 103 1,250 0,643 0,81 1,15 1,51 SPOSS 105 0,743 0,439 - 1,00 1,00

Sample of firms from countries of which at least 1 firm sent a comment letter between 2007 and 2012. We define the dummy variable LETTER as 1 when the firm sent a comment letter. The variable EVOL is

computed as the net income divided by total assets based (= EARN) (1) average EARN per company of all available years (2) the difference between the average and the actual EARN in the year (3) squaring the difference (4) the average of the squared difference of Year 1 + year -1 + Year -2 and (5) the root of the average is the EVOL in Year 1 (6) the average of the outcome of all firm years. The variable SIZE is the natural

logarithm of end of year market value of equity per firm year and then taken the average of all firm years. The variable LEV is computed as the end of year total liabilities divided by end of year equity book value per firm

year and then taken the average of all firm years. The variable ROA is the income divided by end of year total

assets per firm year and then taken the average of all firm years. The variable CFVOL is computed as the

operating cash flow divided by total assets based (= CFVOL) (1) average CFVOL per company of all available years (2) the difference between the average and the actual CFVOL in the year (3) squaring the difference (4) the average of the squared difference of Year 1 + year -1 + Year -2 and (5) the root of the average is the CFVOL in Year 1 (6) the average of the outcome of all firm years. The variable MTB is computed as the

number of outstanding share multiplied with the share price divided by the common equity of each firm year and then taken the average of all firm years. The variable SPOS is computed indicator that equals 1 for

observations with the average of all firm years annual earnings scaled by total assets between 0.00 and 0.01 and 0 otherwise.

(25)

Table 4 Spearman correlation test N = 105

LETTER EVOL SIZE LEV ROA CFVOL MTB SPOS

LETTER Pearson Correlation 1

Sig. (2-tailed)

EVOL Pearson Correlation 0,132 1

Sig. (2-tailed) 0,181

SIZE Pearson Correlation ,594** -0,07 1

Sig. (2-tailed) 0 0,482

LEV Pearson Correlation 0,06 0,138 -0,171 1

Sig. (2-tailed) 0,54 0,164 0,084

ROA Pearson Correlation 0,079 0,046 -0,121 ,266** 1

Sig. (2-tailed) 0,42 0,641 0,223 0,006

CFVOL Pearson Correlation ,205* ,217* 0,185 ,299** ,385** 1

Sig. (2-tailed) 0,04 0,029 0,066 0,002 0

MTB Pearson Correlation -0,161 -0,106 -0,075 -0,035 ,575** 0,16 1

Sig. (2-tailed) 0,104 0,287 0,452 0,729 0 0,111

SPOSS Pearson Correlation -0,097 -,543** 0,007 -0,04 -0,107 -,237* -0,104 1

Sig. (2-tailed) 0,325 0 0,946 0,688 0,278 0,017 0,298

*, ** indicate 0,05 and 0.01 significant levels, respectively. We define the dummy variable LETTER as 1 when the firm send a comment letter. The variable EVOL is computed as the

net income divided by total assets based (= EARN) (1) average EARN per company of all available years (2) the difference between the average and the actual EARN in the year (3) squaring the difference (4) the average of the squared difference of Year 1 + year -1 + Year -2 and (5) the root of the average is the EVOL in Year 1 (6) the average of the outcome of all firm years. The variable SIZE is the natural logarithm of end of year market value of equity per firm year and then taken the average of all firm years. The variable LEV is computed as

the end of year total liabilities divided by end of year equity book value per firm year and then taken the average of all firm years. The variable ROA is the income divided by end of year

total assets per firm year and then taken the average of all firm years. The variable CFVOL is computed as the operating cash flow divided by total assets based (= CFVOL) (1) average

CFVOL per company of all available years (2) the difference between the average and the actual CFVOL in the year (3) squaring the difference (4) the average of the squared difference of Year 1 + year -1 + Year -2 and (5) the root of the average is the CFVOL in Year 1 (6) the average of the outcome of all firm years. The variable MTB is computed as the number of

outstanding share multiplied with the share price divided by the common equity of each firm year and then taken the average of all firm years. The variable SPOS is computed indicator

(26)

4. Results

4.1.1 Regression analyses earnings smoothing

Within this paragraph the results are shown for earnings management measured by earnings volatility. The results for this model are presented in table 5. We define the dependent variable in regression as a binary indicator. The objective of our research is to test whether income volatility is associated (i.e. affects a firm’s decision) to lobby by writing a comment letter. The regression model overall is significant, moreover the Nagelkerke R Square of 68,1% shows that 68,1% of the dependent variable LETTER is explained by the independent variables. The results of our regression show that there is a significantly positive coefficient for the variable EVOL 137,551 at the 0.05 significance level. These findings are in line with H1, which indicates that there is a significant relationship. The results of EVOL suggest that firms who wrote a comment letter have less earnings volatility, which is associated with less earnings management (Mary, 2008). Less earnings volatility is explained by the fact that when firms write a comment letter the change of earnings volatility increases. We would expect that companies who exhibit a higher degree of earnings management would respond to the exposure draft when they disagree with the content of the exposure draft. On the other hand we would expect that companies respond to the exposure draft when they agree with the content of the exposure draft to send a signal to the market. The question is thus what the content of the comment letters was. The summary of the comment letters (ED summary, 2013) states: “The vast majority of respondents support the proposals in the ED as an appropriate balance between faithful representation of credit losses on financial instruments, and the costs of producing that information. Most specified that they agree with the IASB that initial credit loss expectations are priced into assets when originated or purchased, and continue to support an approach that considers deterioration in credit quality in deciding the extent to which expected credit losses should be recognized”. As such the results shown in the table 5 are consistent with our expectations.

The variable SIZE in table 5 is a significantly positive coefficient 1,934 at the 0.01 significance level. This is in line with our prediction. The larger the company is, the more chance there is that the company is involved in the lobbying process. The variable ROA in table 5 is also a significantly positive coefficient 139,787 at the 0.01 significance level, which is in line with our prediction. The outcome of the ROA indicates that the larger the income divided by the total assets is, the larger the company is, the more chance there is that the company is involved in the lobbying process.

(27)

The variable MTB in table 5 is a significantly negative coefficient -3,277 at the 0.01 significance level. This indicates that firms with low volatility in cash flows are more likely to participate in the standard setting process. The variables LEV and CFVOL are not significant and as such we cannot draw conclusions about associations of this independent variables on the dependent variable LETTER.

Table 5

Results logistic regression for interaction LETTER on EVOL N = 100

Variable Predicted sign Coefficient S.E. Sig.

EVOL ? 137,551 69,519 0,048 * SIZE + 1,934 ,431 0,000 ** LEV ? 1,304 1,147 0,256 ROA + 139,787 51,144 0,006 ** CFVOL ? -59,732 49,012 0,223 MTB ? -3,277 1,025 0,001 ** Constant -29,436 6,626 0,000 ** -2 Log likelihood is 57,762 and the Nagelkerke R Square 68,1%.

The regression model used is LETTER = β0 + β1EVOL + β2SIZE + β3LEV + β4ROA + β5CFVOL + β6MTB + ϵ.

*, ** indicate 0,05 and 0.01 significant levels, respectively.

The Sample of firms from countries of which at least 1 firm send a comment letter between 2007 and 2012. We define the dummy variable LETTER as 1 when the firm send a comment letter. The variable EVOL is

computed as the net income divided by total assets based (= EARN) (1) average EARN per company of all available years (2) the difference between the average and the actual EARN in the year (3) squaring the difference (4) the average of the squared difference of Year 1 + year -1 + Year -2 and (5) the root of the average is the EVOL in Year 1 (6) the average of the outcome of all firm years. The variable SIZE is the natural logarithm of end of year market value of equity per firm year and then taken the average of all firm years. The variable LEV is computed as the end of year total liabilities divided by end of year equity book value per firm year and then taken the average of all firm years. The variable ROA is the income divided by end of year total assets per firm year and then taken the average of all firm years. The variable CFVOL is computed as the operating cash flow divided by total assets based (= CFVOL) (1) average CFVOL per company of all available years (2) the difference between the average and the actual CFVOL in the year (3) squaring the difference (4) the average of the squared difference of Year 1 + year -1 + Year -2 and (5) the root of the average is the CFVOL in Year 1 (6) the average of the outcome of all firm years. The variable MTB is computed as the number of outstanding share multiplied with the share price divided by the common equity of each firm year and then taken the average of all firm years.

(28)

4.1.2 Regression analyses managing towards small positive earnings

Within this paragraph the results are shown for earnings management measured by the degree of reporting small positive earnings. The results for this model are presented in table 6. We define the dependent variable in regression as a binary indicator. The objective of our research is to test whether reporting small positive earnings is associated (i.e. affects a firm’s decision) to lobby by writing a comment letter.

The regression model overall is significant, moreover the Nagelkerke R Square of 68,6% shows that 68,6% of the dependent variable LETTER is explained by the independent variables. The results of our regression show that there is a significantly negative coefficient for the variable SPOS 137,551 at the 0.05 significance level. These findings are in line with H2 which indicates that there is a significant relationship. The results of SPOS suggest that firms who wrote a comment letter have less earnings volatility, which is associated with less earnings management (Mary, 2008). Less earnings volatility is explained by the fact that when firms write a comment letter the change of reporting small positive earnings decreases. We would expect that companies who exhibit a higher degree of earnings management would respond to the exposure draft when they disagree with the content of the exposure draft. On the other hand we would expect that companies respond to the exposure draft when they agree with the content of the exposure draft to send a signal to the market. The question is thus what the content of the comment letters was. The summary of the comment letters (ED summary, 2013) states: “The vast majority of respondents support the proposals in the ED as an appropriate balance between faithful representation of credit losses on financial instruments, and the costs of producing that information. Most specified that they agree with the IASB that initial credit loss expectations are priced into assets when originated or purchased, and continue to support an approach that considers deterioration in credit quality in deciding the extent to which expected credit losses should be recognized”. As such the results shown in the table 6 are consistent with our expectations.

The variable SIZE in table 6 is a significantly positive coefficient 1,977 at the 0.01 significance level. This is in line with our prediction. The larger the company is, the more chance there is that the company is involved in the lobbying process. The variable ROA in table 6 is also a significantly positive coefficient 155,148 at the 0.01 significance level, which is in line with our prediction. The outcome of the ROA indicates that the larger the income divided by the total assets is, the larger the company is, the more chance there is that the company is involved in the lobbying process.

(29)

The variable MTB in table 6 is a significantly negative coefficient -3,648 at the 0.01 significance level. This indicates that firms with low volatility in cash flows are more likely to participate in the standard setting process. The variables LEV and CFVOL are not significant and as such we cannot draw conclusions about associations of these independent variables on the dependent variable LETTER.

Table 6

Results logistic regression for interaction LETTER on EVOL N = 100

Variable Predicted sign Coefficient S.E. Sig.

SPOS ? -2,286 0,958 0,017 * SIZE + 1,977 ,440 0,000 ** LEV ? 1,210 1,006 0,229 ROA + 155,148 52,173 0,003 ** CFVOL ? -52,070 47,924 0,277 MTB ? -3,648 1,079 0,001 ** Constant -27,266 6,155 0,000 ** -2 Log likelihood is 57,065 and the Nagelkerke R Square 68,6%.

The regression model used is LETTER = β0 + β1SPOS + β2SIZE + β3LEV + β4ROA + β5CFVOL + β6MTB + ϵ.

*, ** indicate 0,05 and 0.01 significant levels, respectively.

The Sample of firms from countries of which at least 1 firm sent a comment letter between 2007 and 2012. We define the dummy variable LETTER as 1 when the firm sent a comment letter. The variable SPOS is computed indicator that equals 1 for observations with the average of all firm years annual earnings scaled by total assets between 0.00 and 0.01 and 0 otherwise. The variable SIZE is the natural logarithm of end of year market value of equity per firm year and then taken the average of all firm years. The variable LEV is computed as the end of year total liabilities divided by end of year equity book value per firm year and then taken the average of all firm years. The variable ROA is the income divided by end of year total assets per firm year and then taken the average of all firm years. The variable CFVOL is computed as the operating cash flow divided by total assets based (= CFVOL) (1) average CFVOL per company of all available years (2) the difference between the average and the actual CFVOL in the year (3) squaring the difference (4) the average of the squared difference of Year 1 + year -1 + Year -2 and (5) the root of the average is the CFVOL in Year 1 (6) the average of the outcome of all firm years. The variable MTB is computed as the number of outstanding share multiplied with the share price divided by the common equity of each firm year and then taken the average of all firm years.

(30)

5. Conclusion

5.1 Conclusion

Our research question was whether there are significant determinants in the choice whether to participate in the lobbying process of IFRS 9 Phase 2. Especially we investigated two determinants which measure whether participation is associated with the degree of earnings management. We investigated whether earnings management is higher/lower for companies who wrote a comment letter opposed to companies who did not write a comment letter on the exposure draft IFRS 9 Phase 2 within the banking sector. Prior research such as Huian (2013) and Helen Irvine (2010) suggests that the involvement of preparers in the standard setting process captures the standard setting process. Furthermore Sutton (1998) states that people (in this case preparers of financial statements) participate based on the rational-choice model, when the benefits (delay of loan-loss expenses) outweigh the costs (costs due to new regulation). This is not in line with presenting a true and fair view, i.e. earnings management. The results show some interesting findings.

We measured the degree of earnings management by earnings smoothing. Earnings smoothing is measured by the degree of earnings volatility. We focused on the difference between the group that commented on the exposure draft via a comment letter and the group that did not respond via a comment letter on the exposure draft. We found a significant difference of earnings volatility, even with the use of 5 control variables. Our test on earnings volatility indicates that the chance of earnings volatility increases when writing a comment letter, i.e. less earnings management (Mary, 2008).

Furthermore we measured the earnings management by managing towards small positive earnings. Again we focused on the difference between the group that commented on the exposure draft via a comment letter and the group that did not respond via a comment letter on the exposure draft. We found a significant difference of managing towards small positive earnings, even with the use of 5 control variables. Our test on managing towards small positive earnings indicates that the chance of managing towards small positive earnings decreases when writing a comment letter. Consistent with prior literature a higher degree of small positive earnings is associated with a higher degree of earnings management (Mary, 2008).

The results above are consistent with our hypothesis that income volatility and reporting small positive earnings are significantly associated with the decision whether to engage in the lobbying process. When relating both results to earnings management we find that the results indicate that involvement in the standard setting process IFRS 9 Phase 2 is associated with less

Referenties

GERELATEERDE DOCUMENTEN

80 Als het syndicaat niet heeft voorzien in een eventueel faillissement van de security agent en de overdracht van de parallel debt in een dergelijk geval, betekent dit dat

The stimulation effect of the imposed electric field is investigated on multi-compartment models (NEURON 6.2) of three distinct neuronal populations in the subthalamic region:

[r]

On the contrary, systems such as the Gurney flap, the variable droop leading edge and the trailing edge active blade concept will modify the blade profile during the full rotation

I am not referring to military Unmanned Aerial Vehicles (UAVs) used for surveillance, communication relay and attacks on ground targets, but to Unmanned Cargo Aircraft (UCA)

When the Bosnian Serb forces attacked the UN protected enclave in Srebrenica in July 1995, most women, children and elderly fled to the UN compound to seek shelter.. However,

The expectation is still that firms that deliver high quality audits reduce earnings management more than firms that deliver less quality audits (refer to hypothesis one), only

Hypothesis 2c: The association between financial distress and real earnings management is stronger when the CEO of the firm is in the early years of its tenure than when the CEO of