Lobbying on accounting standards: a reflection of discretionary behaviour? : An empirical examination of lobbying activities in response to IFRS 15.

(1)

Lobbying on accounting

standards: a reflection of

discretionary behaviour?

An empirical examination of lobbying

activities in response to IFRS 15

Haala Aarab

Student number: 10003498

Final Version

June 23, 2014

Master Thesis

Msc Accountancy & Control

1st Supervisor: dr. S.W. Bissessur

2nd Supervisor: dr. G. Georgakopoulos

Universiteit van Amsterdam

Amsterdam Business School

(2)

1

ABSTRACT

In this thesis I examine the factors that explain why firms engaged in lobbying activities with the standard-setter in the development of IFRS 15. More specifically, I focus on whether discretionary revenues are an explanatory factor for writing a comment letter in response to the Exposure Draft Revenue from Contracts with Customers of IFRS 15. My results show that the likelihood of writing a comment letter increases with the amount of negative discretionary revenues. In addition, I provide further evidence for the strong positive relationship between writing a comment letter and firm size. Lastly, my results indicate that firms with a higher asset turnover and firms that are audited by a big-4 accounting firm are more likely to engage in lobbying activities.

Key words: lobbying, comment letters, IASB, FASB, standard-setting process, discretionary behavior, discretionary revenues, IFRS 15

(3)

2

Acknowledgements

First and foremost I would like to thank my supervisor, dr. Sanjay Bissessur, for sharing his knowledge with me and for his excellent support and feedback while I was working on this thesis.

Secondly, I would like to express my eternal gratitude towards my parents for supporting me throughout my studies, both morally and financially. I could not have made it here without you. I hope I made you proud.

And last but not least, a special thanks goes to Karim Belâbar. You have motivated me in ways you are unaware of.

(4)

3

LIST OF TABLES

TABLE3.1CATEGORIZATIONRAWTEXTINCOMMENTLETTERS ... 18

TABLE3.2DOMINANTTHEMESARISINGFROMTHECOMMENTLETTERS ... 19

TABLE 5.1PREDICTED RELATIONSHIPS ... 29

TABLE6.1SAMPLESELECTION ... 31

TABLE6.2SUMMARYSTATISTICS ... 33

TABLE6.3UNIVARIATETESTS ... 36

TABLE6.4LOGISTICREGRESSIONFORWRITINGACL(ABSDISCR_REV) ... 38

TABLE6.5LOGISTICREGRESSIONFORWRITINGACL(RESIDUAL) ... 39

TABLE6.6LOGISTICREGRESSIONFORWRITINGACL ... 40

TABLE6.7LOGISTICREGRESSIONFORWRITINGACL(YEAR >2002) ... 42

(6)

5

LIST OF ABBREVIATIONS

 CL = Comment Letter

 DISC_REV = Discretionary Revenues  DP = Data Point

 ED = Exposure Draft  LEV = Leverage

 MTB = Market-to-Book Ratio  ROA = Return on Assets

(7)

6

CHAPTER 1: INTRODUCTION

“Revenue is a crucial number to users of financial statements in assessing an entity’s financial performance and position. However, revenue recognition requirements in US generally accepted accounting principles (GAAP) differ from those in International Financial Reporting Standards (IFRSs) and both sets of requirements need improvement.”

– Exposure Draft, IFRS Foundation, November 2011

In this thesis I examine the motives of firms that engaged in lobbying activities in the development of IFRS 15, Revenue from Contracts with Customers. Revenue is a key performance measure used by shareholders to assess a firm. Therefore, new accounting standards which alter the measurement of earnings like the proposed revenue recognition model are typically of great concern to producers of financial statements. Empirical evidence based on the response of security prices to accounting information suggests that investors find accounting information decision useful (Scott, 2012). The presentation and content of financial statements are defined by accounting standards that apply to that information. Accordingly, to understand why firms report in the way they do, it is necessary to understand the process of standard setting.

One of the factors affecting the process of standard setting is corporate lobbying (Giner & Arce, 2012). Lobbying can be collectively described as “the efforts of individuals and organizations to promote or obstruct new accounting rules” (Sutton, 1984). Corporate lobbying is a complex process and involves a number of interrelated issues: whether to lobby in the first place, which methods to deploy and what arguments to use to support a viewpoint. In essence, firms engage in lobbying activities with the standard-setter to protect and promote their interests. An example of such an interest is discretion over financial reporting, i.e. managerial freedom of judgment and decision making. The greater the expected economic benefit from lobbying, the greater the incentive to interact with the standard-setter in the development of standards.

This thesis addresses the relationship between corporate lobbying activities in response to the Revised Exposure Draft Revenue from Contracts with Customers of the IASB and FASB and managerial discretion. The accounting for these kind of contracts is currently addressed in IAS 18 (Revenue) and IAS 11 (Construction Contracts). However, these existing standards for reporting revenues have been widely criticized as being inconsistent, resulting in economically similar transactions to be accounted for differently. Moreover, the disclosure requirements are perceived to be ineffective in enabling financial statements users to comprehend the company’s revenues and the associated judgments and estimates made in this context. Consequently, the IASB and FASB added the revenue recognition project to its agenda in 2002. This joint project aimed at improving financial

(8)

7 reporting by creating one common revenue recognition model for IFRSs and US GAAP that would clarify the revenue recognition principles and make them applicable to various transactions, industries and capital markets. At the heart of the new model lies the notion of identifying “a performance obligation” and the “satisfaction” of this obligation, in contrast to current accounting practices where revenue is recognized when it is both “realized or realizable” and “earned”. The proposed model for revenue recognition will impact business reporting for all companies that enter into contracts with customers, but certain industries in particular will experience significant changes, like those involved in engineering, real estate, software and telecommunications (KPMG, 2011).

In this study I attempt to answer the following research question: “To what extent does corporate lobbying in response to the Revised Exposure Draft Revenue from Contracts with Customers reflect discretionary behavior?” I address managerial discretion by looking at discretionary revenues, which represents the difference between the change in receivables as predicted by the model and the actual change in receivables (i.e. the residual). I make the assumption that abnormally high or low values of discretionary revenues are an indication of revenue management. I compare discretionary revenues between firms that wrote a comment letter in response to the Revised Exposure Draft (2011), the “lobbyists”, and firms who did not, the “non-lobbyists”, to see whether this is an explanatory variable for lobbying the standard-setter. In addition, I examine whether asset turnover, return on assets, firm size, leverage, the market-to-book ratio, operating under a loss and being audited by a big-4 accounting firm are explanatory variables for engaging in lobbying activities.

My results show that there exists a statistically significant relationship between discretionary revenues and the likelihood of engaging in lobbying activities in the standard-setting process. More specifically, the data seems to imply that the likelihood of writing a comment letter increases with the amount of negative discretionary revenues. This means that the more prudent a firm is in reporting revenues (i.e. it manages revenues downward), the higher the likelihood that it will engage in lobbying activities. My results also provide further evidence for the strong positive relationship between writing a comment letter and firm size. Furthermore, I find a positive relationship between the likelihood of lobbying the standard-setter and the variables asset turnover and being audited by a big-4 accounting firms.

The remainder of this thesis is organized as follows. Chapter 2 will discuss the theory and outline the main empirical studies in the field of corporate lobbying literature. Moreover, the development of the Proposed Revenue Recognition standard will be discussed in depth. Chapter 3 will continue with an overview of the main themes arising from the comment letters. In chapter 4, the hypothesis for this study will be developed. Chapter 5 will discuss the research methodology and chapter 6 will present the empirical results. Finally in chapter 7 the conclusion will be presented.

(9)

8

CHAPTER 2: THEORY/LITERATURE REVIEW

1. Introduction

In this chapter, I will first briefly introduce the theories of the standard-setting process. In particular, I will discuss the role of lobbying in the standard-setting process and outline the main empirical literature that evaluates corporate lobbying. Secondly, I will present the developments in the proposed revenue recognition standard by the FASB/IASB. The aim of this chapter is to explain why firms engage in lobbying activities and how the proposed revenue recognition standard will change accounting practices.

2. The role of lobbying in the standard setting process 2.1 Theories of standard-setting

Different theories of regulation have been put forward in existing literature to frame the standard-setting process. Main theories are the public interest theory, the interest group theory, positive accounting theory, and agency theory, which will be described in this section.

Public interest theory states that regulators make decisions that are in the best interest of society, i.e. those decisions that maximize social welfare (Scott, 2012). This regulation is required because of market failures: accounting information is a public good which, unregulated, will be under-produced. The result of the standard-setting process under this theory is driven by a societal cost- and benefit-analysis, whereas the benefits reveal itself as improved operation of the market. Nonetheless, determining the right amount of regulation required to maximize social welfare under this theory has proven to be difficult. Also, this theory suggests that lobbying to promote one’s own interests is useless because the standard-setter will always choose what’s best from a societal point of view (Kothari et al, 2010).

Interest group theory (also referred to as the “regulatory capture” model) takes on a more narrow perspective and views society as consisting of different constituencies which are all driven by the objective to maximize their own welfare (Scott, 2012). Each of these constituencies engage in lobbying activities with standard-setters to protect and promote their interests. Under this theory, the result of the standard-setting process is not driven by a societal cost- and benefit analysis but by competition amongst constituencies that lobby for or against regulation. This theory suggests that standard-setters are not fully independent and are “captured” by managers’, accountants’ and auditors’ interest to protect themselves against legal liability (Kothari et al, 2010). Somewhat related to the interest group theory is the agency theory, which offers insight in the incentives of preparers of financial statements (Eisenhardt, 1989). More specifically, it introduces the notion of self-interest in organizational thinking: managerial discretion in the preparation of financial statements might be used

(10)

9 opportunistically if this discretion is unverifiable (Ramanna, 2008). These incentives may drive the interaction in the standard-setting process with standard setters.

Finally, the ideology theory of regulation puts a combination forward of the public interest and the interest group theory and asserts that the standard-setting process is a joint outcome of public interest and special interest lobbying (Kothari et al, 2010).

Despite the various standard-setting theories put forward in the literature, none of the theories described above are fully able to explain the complicated standard-setting process (Gipper et al., 2013). The public interest theory raises the question if one can ever identify and understand the interests of all groups interacting in the standard-setting process, and even if this is possible, it does not tell us how to deal with possible conflicting interests. Moreover, theoretically standard setters are ought to be independent, while in reality they operate under direct oversight of the SEC and Congress, which possibly suggests standard-setters have agenda’s on their own.1 Concerning the ideology theory of regulation, this theory leaves the question of how large the external influence in the standard-setting process is unanswered, and thus only provides us with a general conceptualization.

2.2 Empirical literature on standard-setting

Research in the area of corporate lobbying on accounting standards has its origin in the 1980s (Watts & Zimmerman, 1978; Kelly, 1982; Holthausen & Leftwich, 1983; Sutton, 1984; Kelly, 1985 and Francis, 1987), but has become an active field of research in recent years again (Georgiou, 2004; Ramanna, 2007; Giner & Arce, 2012). Most of the literature on corporate lobbying has focused on analyzing comment letters written by constituencies in response to standard proposals, because comment letters are a readily observable and explicit form of lobbying (Francis, 1987; Georgiou, 2004; Sutton, 1983; Giner & Arce, 2012).

One of the earliest work on lobbying during the standard-setting process is that of Watts and Zimmerman in 1978. These researchers present a theory based on economic consequences arguments to predict the lobbying activities of firms. Their first prediction is the so called political cost hypothesis: firms that are subject to high regulatory and political pressure (large firms), are more likely to advocate proposed accounting changes that lower reported income because this reduces their exposure to regulatory and political actions. Secondly, under the bonus plan hypothesis, the authors argue that firms that are not exposed to regulatory and political pressure (smaller firms), are more likely to oppose proposed accounting changes that decreases reported income because bonus compensation plans have to be adjusted accordingly. To test these predictions, Watts and Zimmerman examine 53 corporate comment letters in response to a FASB discussion memorandum in 1974. First the expected effect of the new standard on net income was estimated for each company and then the

1

This refers to the extent to which standard-setters have other intents next to genuinely wanting to improve financial reporting by means of increasing transparency and decreasing managerial discretion. This is further described in Gipper et al. (2013).

(11)

10 authors evaluated if the firm voted for or against the proposed standard. Watts and Zimmerman find that the relation between firm size and the effect on earnings is highly significant, supporting their hypotheses. While the results generally supported the researchers’ predictions, a notable flaw is the relatively small sample size (53), making it hard to generalize the results.

A follow-up study of Kelly (1985) examines the characteristics of firms that choose to lobby on the FASB’s exposure draft on accounting for foreign currency translation. That is, the author focuses only on the decision to lobby, and not on the decision to lobby for or against the standard as Watts and Zimmerman (1978) did. The study adopts an agency theory framework to lobbying and addresses the question whether firm leverage and management’s percentage of ownership of the firm explain the decision to lobby. The sample used in the study consists of 195 firms that would have to change their accounting under the new standard, of which 55 firms engaged in lobbying by submitting comment letters and 140 did not. After dropping the firms for which no data were available, the researcher ended up with a final sample of 54 firms that lobbied and 116 that did not. A test to compare the lobbying group with the non-lobbying group showed that the lobbing firms had a significantly lower percentage of management ownership, which contradicted the expectation. The empirical results indicated that neither the leverage nor the percentage of management ownership were statistically significant. Here again, like with the study of Watts and Zimmerman (1978), the study is limited by a small sample size. Furthermore, the results of Kelly are highly ambiguous because of multicollinearity issues between the different variables.

Deakin (1989) studied firms’ incentives to lobby the FASB on oil and gas accounting proposals. In contrast to prior work, Deakin captures the whole lobbying process by examining different stages in which firms lobby: 1) the discussion memorandum of the standard in 1975, 2) the exposure draft in 1977 and 3) before the SEC public hearing in 1978. All firms affected by the new standard are classified and compared to each other based on whether they lobbied during the standard-setting process.The author adopts an economic consequences perspective to lobbying, similar to Watts and Zimmerman (1978) and Kelly (1985). Hypothesized is that lobbying is explained by management compensation plans incentives, debt contracting incentives and the oil and gas exploration activities of the firm. The paper provides evidence consistent with this hypothesis, and suggests that accounting methods matter in a way that they can affect the economic well-being of the company as well as its managers. The results of the Deakin study are relatively strong, as his research captures different stages in the lobbying process. Moreover, the oil and gas accounting proposals were an important economic and political issue at the time.

Another study that examined the characteristics of firms that lobby during the standard-setting process in the light of economic consequences, is that of Dechow et al. (1996). Comment letters submitted in response to the 1993 FASB exposure draft on accounting for ESOs (executive stock options) are analyzed to determine managements position towards the proposed standard (for or against). The results show that the use of stock options in executive compensation strongly increased

(12)

11 the likelihood of a firm submitting comment letters opposing the expensing of stock options. This finding supports the executive compensation cost hypothesis, which predicts opposition to the standard (which prescribes the expensing of stock options) because of managers’ concern about public scrutiny of their compensation.

One of the more recent studies on lobbying during the standard-setting process, is that of Ramanna (2008). This study examines the standard-setting process of SFAS 142, which uses unverifiable fair value estimates for acquired goodwill. Ramanna looks at four events (senate hearings, house hearings, house bill, senate letter) and determines 43 politicians in total who voted in opposition to the standard during at least one of those events. Subsequently, the author links these politicians to firms and industries that lobbied against the exposure draft on the standard. The researcher finds evidence that politicians who opposed the standard are linked to firms and industries that also opposed the new model. Moreover, he concludes that the standard is at least partly due to lobbying activities of firms that supported the standard because it granted them greater discretion on impairments. Specifically, firms with larger and more numerous business segments, higher market-to-book ratios and a higher proportion of net assets without observable market values increased the likelihood of lobbying for the proposed standard. The strength of the Ramanna study is that it takes on a more holistic perspective to lobbying on accounting standards by capturing several phases in the lobbying process.

Another recent study of Hansen (2011) investigates how the standard-setter develops standards when being influenced by constituents with differing interests. Hansen examined comment letters sent in response to five Exposure Drafts and found that lobbying success is positively related to the ability of the lobbyist to provide the standard-setter with relevant information. This relationship, however, is influenced by the credibility of the lobbyist. Furthermore, Hansen finds that lobbying success is associated with the impact the lobbyist has on the viability of the standard-setter2.

3. Developments Proposed standard Revenue from Contracts with Customers

Convergence between the US GAAP standards and IFRSs has been of high priority throughout the years for both the International Accounting Standards Board (IASB) and the Financial Accounting Standard Board (FASB) (EY, 2012). Ultimately, both US GAAP and IFRSs employ merely the same revenue definition. According to the FASB, revenue can be denoted as the “actual or expected cash inflows that have occurred or will result from the entity´s ongoing major operations”, whereas the IASB denotes revenue as “the gross inflow of economic benefits during the period arising in the course of ordinary activities of an entity ”. Furthermore, even though the criteria are not identical, both US GAAP and IFRSs rely on the transfer of risk to recognize revenue from the sale of goods.

2

(13)

12 Despite these similarities, differences in revenue recognition arise in practice because of different levels of accuracy between the two standards. In general, US GAAP includes more detailed and prescriptive guidance related to specific industries in comparison with IFRS (Bohusova, 2009). Even though IFRSs have fewer revenue recognition requirements, they have proven hard to comprehend and apply in practice (Exposure Draft, 2011). Nobes (2012) discusses the problems with the IFRS definition of revenue in IAS 18, and comes to the conclusion that this definition is seriously flawed due to several reasons, like the outdated term of “ordinary activities”.

In addition to this the standards significantly differ on some points, like for instance the deferred receipt of receivables (EY, 2012). Whereas US GAAP does not require discounting to present value, IFRS asserts the value of deferred revenue payments to be recognized is determined by discounting. Inconsistent standards permit managers to arbitrarily choose between different accounting methods and to exert judgment differently in the same situations. This means that a persistent application of accounting standards can only be reached if the accounting standards themselves are internally consistent (Wüstemann & Wüstemann, 2010).

On top of the mutual inconsistencies between IFRSs and US GAAP, the accounting standards on revenue recognition lagged behind with developments and changes in the organizational and reporting environment (Wagenhofer, 2014). During and after the Internet-bubble in the late 1990s, a lot of instances of revenue manipulation became public, which led to scrutiny about how revenues are being recognized. Companies also started to enter into complex and long-term contracts with their customers, which raised questions about how the revenues for these transactions should be accounted for. As for the reporting environment, investors called for more timely information about revenues which are also comparable across organizations.

Motivated by the need of standard improvements, the International Accounting Standards Board (IASB) and the Financial Accounting Standard Board (FASB) combined their strengths to develop a common standard for recognizing revenue from contracts with customers. The first exposure draft of the new standard Revenue from contracts with customers was made public in 2010 and a revised exposure draft was released one year hereafter. The new revenue recognition standard should accomplish the following main objectives:

1. Remove inconsistencies and weaknesses in existing revenue requirements; 2. Provide a more robust framework for addressing revenue issues;

3. Improve comparability of revenue recognition practices across entities, industries, jurisdictions and capital markets;

4. Provide more useful information to users of financial statements through improved disclosure requirements; and

(14)

13 5. Simplify the preparation of financial statements by reducing the number of requirements to

which an entity must refer. (Source: Exposure Draft, 2011)

The new revenue recognition standard would apply to all entities that enter into a contract with a customer, unless the contract is in the scope of another standard. Concretely, this means the new standard would replace IAS 11 and IAS 18 in IFRSs, and ASC Topic 605 in US GAAP.

The essence of the new revenue recognition requirements is that “an entity should recognize revenue to depict the transfer of promised goods or services to customers in an amount that reflects the consideration to which the entity expects to be entitled in exchange for those goods or services”(Exposure Draft, 2011). The proposal retains a single model with two ways to recognize revenues: over time (similar to the current stage of completion method), or at a point in time (similar to the current sales of goods accounting). All companies, under the proposals, will apply the single model. This means that a greater degree of managerial judgment will be required in applying the new standard.

The new standard proposes a five-step approach for recognizing revenue:

1. Identify the contract with the customer

The standard refers to a contract as an agreement between two -or more- parties which leads to enforceable rights and obligations. In order for a contract to exist, all of the following criteria must be met:

a) The contract must have commercial substance;

b) All parties involved in the agreement must have agreed to the contract;

c) The rights of all parties involved concerning the transfer of goods or services can be identified; and

d) The payment terms for the transferable goods or services can be identified.

The standard goes on to explain that there is no contract if transfer of the goods or services to the customer has not yet taken place and no consideration is received, nor does the right exist to receive this consideration. Furthermore, multiple contracts entered into around the same time with the same customer can be accounted for as one contract if several predetermined criteria are met.3 A change in the contract price is accounted for as an adjustment to the transaction price and a change in the transferable goods or services is accounted for as a (new) separate contract.

3_{Multiple contracts can be combined as one contract if the contracts serve the same commercial purpose, the consideration of one of the}

contracts depends on the price or performance of the other contracts, or if the goods or services in the contract belong to one performance obligation.

(15)

14 2. Identify the separate performance obligations in the contract;

The standard defines a performance obligation as an obligation mentioned in the contract to deliver a good or service. All activities that have to be performed in order to complete the contract, but that do not directly relate to the transfer of a good or service to the customer (e.g. administrative activities), do not meet the definition of a performance obligation. Each obligation in the contract to deliver a good or service is accounted for as a separate performance obligation if these goods or services are distinct. A performance obligation is deemed distinct if the transfer of the good or service differs from that of the other goods or services. Furthermore, at least one of the following criteria must be met in order for a good or service to be distinct:

a) The good or service is sold separately on a regular basis; or

b) The customer can benefit from the good or service separately or directly controls the resources to do this.

The standard goes on to explain that if a good or service is highly interrelated with a bundle of other goods or services, or if a bundle of goods or services is significantly modified in order to fulfill the contract, the bundle is accounted for as one single performance obligation.

3. Determine the transaction price;

The transaction price is the amount of consideration to which the entity expects to be entitled in exchange for a good or service. In determining the transaction price, the following points are taken into consideration:

a) Variable considerations – In case of a variable consideration, for example due to discounts, an estimate of the transaction price is made for every fiscal period. The estimate represents the amount of consideration the entity expects to receive or the most likely amount to which it is entitled.4 In estimating the transaction price, the entity considers historical, present and forecasted information.

b) The time value of money – The transaction price is adjusted for the time value of money if the contract contains a significant financing component. In determining if a contract contains a significant financing component, an entity considers amongst other things the following:

i. The period of time between transfer of the goods or services and receiving the consideration for these goods or services;

ii. Whether there is a significant difference in total consideration in case the consideration is paid directly in cash; and

iii. The interest rate included in the contract compared to the market interest rate.

4

(16)

15 If an entity decides that the contract contains a material financing component, an interest rate that reflects the credit risk of the customer or the entity is used to discount the expected consideration.5 If the expected contract term is less than one year, an entity is not required to consider the time value of money.

c) Non-cash considerations – the value of a non-cash consideration is determined by taking the fair value of the consideration, If no fair value can be reliably determined, the value of the consideration is estimated on the basis of the separate selling prices of the goods or services that are transferred in exchange for the non-cash consideration. d) Considerations payable to the customer – Any amount of consideration payable to the

customer is deducted from the transaction price.6 The implication of this is that less revenue is recognized when the goods or services are transferred to the customer, the consideration is paid or an obligation exists to pay the consideration to the customer. e) Credit risk – customer credit risk is not taken into account when determining the

transaction price.7 Instead, the estimated uncollectible amount is presented as a separate line item adjacent to revenue in the profit and loss statement.

4. Allocate the transaction price to the separate performance obligations in the contract; At commencement of the contract, the transaction price is allocated to the separate performance obligations (as identified in step 2) based on the standalone selling prices of the goods or services. When no standalone selling price is available, the entity makes a reasonable estimate of the standalone selling price.8 When a bundle of goods and/or services has a total selling price which is below the sum of the separate selling prices of the goods and/or services, this discount is allocated on a proportional basis to the separate performance obligations, unless the goods or services are sold separately on a regular basis and the stand-alone selling prices indicate that the discount must be allocated to one or more specific performance obligations. A change in the transaction price is allocated to the separate performance obligations.9

5

The credit risk of the entity is considered when the consideration is received from the customer before transfer of the goods or services, and the credit risk of the customer is considered when the goods or services are transferred before the consideration is received.

6

Except from payments for separate goods or services

7

This has to do with the fact that the transaction price ought to represent the amount of consideration an entity expects to be entitled to in exchange for a good or service, instead of the amount of consideration they expect to receive.

8

Suggested estimation techniques in the standard are the adjusted market assessment approach, which bases the estimation on what the customer would be prepared to pay for the good or service, “the expected cost plus a margin approach”, where an estimation is made of the costs plus an appropriate margin for the good or service, and the “residual approach” which estimates the stand-alone selling price of a good or service by taking the total transaction price and deducting the stand-alone selling prices of the other goods and/or services in the bundle from it.

9

(17)

16 5. Recognize revenue when (or as) an entity satisfies the performance obligation;

Revenue is actually recognized when the performance obligation is satisfied by transfer of control of the goods and/or services to the customer. The core principle of the model is thus that revenue is recognized when transfer of control takes place. Transfer of control to the customer has taken place when the customer is able to determine how the good or service is being used and he has the right to receive all the benefits of the good or service. This also includes being able to prevent third parties from using the good or service or receiving the benefits from it. The transfer of control can happen at a point in time or over a period of time:

a) Fulfillment of a performance obligation over time

When the customer obtains control over the goods or services over a period of time10, revenue is recognized over this period of time. The stage of fulfillment of the performance obligation is determined by using a method which most accurately reflects the transfer of control to the customer.11

b) Fulfillment of a performance obligation at a point in time

When a performance obligation is not fulfilled over a period of time, it is fulfilled at a point in time, being when transfer of control to the customer takes place. The standard suggests some indicators for determining when transfer of control takes place:

i. The customer obtains legal ownership;

ii. The good is in physical possession of the customer; iii. The customer is obligated to pay for the good or service;

iv. The customer is in possession of the significant risks and rewards; v. The customer has accepted the good or service.

The standard goes on to mention that the abovementioned indicators are not restrictive and no single indicator decisive.

If the consideration to be received for a good or service in the contract is variable, revenue is only recognized for the amount to which the entity can reasonably expect to be entitled. An entity can reasonably expect to be entitled to a consideration when it has past experience with similar performance obligations and this experience is of predictive nature for the consideration to be received.

10

A performance obligation is ought to be satisfied over a period of time if the performance creates an asset which the customer controls and it does not have an alternative function to the entity. Examples of this are construction services in the construction industry where the customer owns the work in progress and the rendering of consulting services with the right of payment upon termination of the contract.

11_{The standard proposes so called “output methods”, like the amount of produced of delivered products, of “input methods”, like the}

(18)

17

CHAPTER 3: DATA

1.Introduction

In this chapter, the data from the comment letters are condensed into dominant themes. The IASB/FASB received 357 comment letters in response to the revised exposure draft of 2011. Nearly half of these comment letters came from North America, and approximately two-thirds of the respondents were preparers of financial statements. In this thesis, only the comment letters of preparers are analyzed, which leads to a final sample of 155letters. Firstly in section two the approach for analyzing the qualitative data in the comment letters will be described. Section three then outlines the dominant themes identified in the comment letters. Section four contains a summary.

2.The inductive approach

An inductive analysis allows research findings to emerge from dominant themes inherent in raw data, without imposing a structured methodology on the research process (Thomas, 2006). One of the purposes of this approach is to condense varied and a large amount of raw data into a summary format. The inductive approach is particularly appropriate in this regard because the purpose of this thesis is to describe the actual effects of the proposed standard Revenue from Contracts with Customers on company reporting, and not just the planned effects.

In order to “let the data speak”, the stages of analysis as developed by Thomas (2006) for analyzing qualitative data are used as a framework. Specifically, these stages include the following12:

1. Preparation of raw data files (so called “data cleaning”) 2. Reading raw text in detail

3. Identifying and creating of categories 4. Overlapping coding and uncoded text13

5. Continuing revision and refinement of categories

Systematic analysis and comparison of the data leads to a set of “categories”, which again may be subdivided in “subcategories” or “dimensions”. In this thesis, matters raised in the comment letters are categorized in four parts: 1) the issue, 2) the reason, 3) the effect and 4) the –suggested- solution for the issue. Table 3.1 gives an overview of this categorization. While the examination of the raw text in the comment letters remains inherently subjective, particular terms allow for a more systematic approach to the identification of an issue and its related reason(s), effect(s) and solution(s). These “key words” are also included in table 3.1.

12

See also Appendix 1

13

This has to do with the fact that a segment of text may be coded into more than one category and other segments of text may not be coded at all because of their irrelevance to the research objective.

(19)

18 TABLE 3.1

CATEGORIZATION RAW TEXT IN COMMENT LETTERS

1. Issue 2. Reason 3. Effect 4. Solution

Matters raised in the comment letter

Reason for raising the matter

The –expected- effect of the raised matter

-Suggested- solution for the matter

Key words “We are concerned

about…”

“Because of…” “Does not reflect…”

“This would result in…”

“We recommend that…”

“We do not agree with...”

“We (do not) believe that...”

“We have worries about…”

“It is unclear how...”

“Does not provide…” “We (do not) believe that...”

“The proposed requirements would lead to…”

“As an alternative,…” “We suggest that…” “The Boards might consider…”

Ultimately, the categories are refined to a small number of “core categories” that describe the phenomena being studied. The intended result should be a small number of categories14 which reflect the key aspects of the themes identified in the data. These core categories are referred to as dominant themes for the remainder of this thesis.

Inherent to the examination of archival data such as comment letters are some notable methodological issues that will be mentioned upfront. Firstly, as mentioned by Gipper et al. (2013), most samples are restricted to those firms that choose to lobby as part of the comment letter process, without taking into account the other firms that will be affected by the new standard, creating a selection bias. In this thesis, this issue is controlled for by the use of a control group. Secondly, the data contained in the comment letters only represents a second best data set of the underlying issues, as a potential bias may arise when preparers attempt to display a certain image as part of the political process.

14

(20)

19 3. Dominant themes

An analysis of the data resulted in the identification of 3 dominant themes, which are grouped in three main categories: operational-, economic- and other concerns. These themes are displayed in table 1. This following section will discuss each of these themes separately.

TABLE 3.2

DOMINANT THEMES ARISING FROM THE COMMENT LETTERS

Operational concerns Economic concerns Other concerns

Inadequate information systems

Retrospective application

Disclosure requirements

3.1 Inadequate information systems

A lot of preparers raised the concern of having to acquire and implement new (accounting) information systems in order to comply with the new standard. The disagreement with amending existing or acquiring new information systems is dominated by a cost-benefit argument:

Current systems limitations and the prodigious effort required to retrospectively apply the new guidance would be extremely cost prohibitive and would not result in significantly more useful financial information for users (CL117, DP467, emphasis added).

[-] We believe the costs associated with designing and implementing systems to reliably capture the data required for the disclosures significantly outweigh any benefit to the investor or user of the financial statements (CL125, DP498, emphasis added).

[-] We believe the time, effort and cost to implement and maintain system changes to track this particular information greatly outweighs any potential benefits (CL121, DP487, emphasis added).

New systems and process changes will be needed in order to implement this requirement for very little benefit to the users of the financial statements (CL26, DP56, emphasis added).

(21)

20 While most preparers refer to the perceived gains for financial statement users, some preparers explicitly mention the lack of benefits for management:

If the Boards issue the disclosure requirement as is, global companies with thousands of contracts would likely be required to reconfigure their financial reporting systems at significant expense in order to meet this disclosure requirement which provides little value to management and financial statement users. (CL128, DP520, emphasis added).

In relation to what drives the need for a new or amended (accounting) information system, the following main requirements are mentioned:

1. The increase in volume of required disclosures (CL67, DP198; CL81, DP262; CL85, DP297; CL92, DP336; CL113, DP417; CL118, DP472; CL121, DP487; CL125, DP498; CL 128, DP520; CL130, DP546; CL145, DP619; CL164, DP720; CL22, DP20; CL35, DP33)

2. The full retrospective adoption requirements (CL55, DP127; CL76, DP231; CL90, DP325; CL111, DP413; CL115, DP432; CL117, DP467; CL154, DP672; CL158, DP690; CL159, DP718, CL26, DP45)

3. The requirement to capitalize and recognize as an asset the incremental costs of obtaining a contract, such as sales commissions (CL26, DP56; CL71, DP215; CL85, DP315; CL199, DP483; CL154, DP677; CL26, DP56; ).

4. The requirement to recognize revenue over time relating to each performance obligation for short term manufacturing processes (CL 56, DP132; CL98, DP364).

5. The stand-alone selling price allocation (CL95, DP351; CL119, DP477)

6. Accounting for the time value of money if the period between payment by the customer of all or substantially all of the promised consideration and the transfer of the promised services to the customer will be more than one year (CL127, DP514; CL26, DP54)

Interpreting the comments on having to implement or amend information systems, the concerns seem to be primarily driven by operational challenges (i.e. the required time, effort and expertise to successfully put a new system in place), but there is also an economic aspect to the concerns (the associated costs to acquire a new or amend an existing system). The disclosure requirements as stated in paragraphs 113-123 of the ED (2011) are the most often mentioned cause for having to implement new or amend existing information systems.

3.2 Retrospective application

Another prevailing concern in the comment letters is the matter of retrospective application. The original Exposure Draft (2010) proposed full retrospective disclosure, on which the standard setters

(22)

21 received a lot of commentary. The Revised Exposure Draft (2011) attempted to ease the full retrospective application requirement by allowing several practical expedients15, but nevertheless many respondents once again expressed their concerns about this requirement. The reason for disagreement with retrospectively applying the proposed standard is twofold. Firstly, almost all preparers raised the issue of the costs, complexity and effort associated with retrospective application:

We continue to believe that it is impractical and very costly to apply the Revised Proposed Update retrospectively to tens of thousands of complex long-term multiple-element contracts (CL118, DP473, emphasis added)

System changes and then re-accounting for transactions retrospectively through those systems or by alternate means would be extremely difficult, prone to error and costly to execute (CL115, DP432, emphasis added)

(-) we believe that the transition plan put forth by the Boards remains one of the most significant barriers to efficient adoption of the proposed standard (CL76, DP231, emphasis added)

In general, the perceived burden from having to retrospectively adopt the standard results from expected operational adjustments, changes to internal policies, having to establish process controls and procedures for transition and ongoing application, and additional reviews and audits of retrospective financial statements.

Secondly, preparers reasoned against retrospective application from the viewpoint of users of the financial statements:

We are strongly opposed to the full retrospective implementation approach advocated in the revised Exposure Draft on revenue recognition. (-) We do not believe the cost-benefit of a full retrospective application would provide more decision-useful, accurate, reliable, or comparable information for financial statement users (CL117, DP469, emphasis added)

15

The practical expedients under the Revised Exposure draft (2011) are as follows: 1) contracts completed before the date of initial application need not to be restated if the contracts begin and end within the same reporting period, 2) for contracts completed on or before the date of initial application which include a variable consideration, firms are allowed to use the transaction price at the date the contract was completed to determine revenue in comparative periods, 3) firms are allowed to not apply the onerous test to performance obligations before the date of initial application except when an onerous contract liability was recognized, 4) ) firms are allowed not to disclose the portion of the transaction price allocated to remaining performance obligations and the explanation of when that revenue will be recognized. When a firm opts to apply one of the expedients, this should be disclosed, together with a qualitative assessment of the estimated impact of the expedient.

(23)

22 Current systems limitations and the prodigious effort required to retrospectively apply the new guidance would be extremely cost prohibitive and would not result in significantly more useful financial information for users (CL117, DP467, emphasis added)

Preparers overall suggested the following main alternatives to retrospective application: 1. Allowing entities to choose between retrospective, modified retrospective and a prospective

application (CL34, DP96; CL67, DP199; CL115, DP432; CL117, DP467; CL127, DP516; CL128, DP524; CL145, DP626; CL154, DP671), potentially with disclosure of supplemental quantitative and qualitative information to aid in comparability (CL55, DP121; CL85, DP307) 2. Limiting presentation of comparative data to one year (only including income statement

disclosures) (CL34, DP96; CL158, DP690)

3. Expanding the number of permitted practical expedients (CL26, DP45)

4. Applying the ASU only retrospectively to those contracts still open at the time the ASU is adopted (CL127, DP516)

Interpreting the comments made about retrospective application, it is remarkable that almost all preparers repeatedly use amplifications such as “enormous”, “extraordinarily”, “unduly”, “significant”, “extremely” and “substantial” in order to emphasize their point of view. In the previous main theme described (i.e. inadequate information systems), preparers predominantly stated their concern about the matter. In sharp contrast, the retrospective application requirement led preparers to explicitly state their disagreement with the proposal. This could imply that preparers assign different weights to matters raised in their comment letters.

Current revenue recognition standards are dispersed and primarily contain industry-specific guidance for particular practice problems. A lot of the guidance has been developed on an ad hoc basis as new accounting issues emerged, resulting in situations where the guidance is inconsistent across pronouncements (Schipper et al, 2009). This dispersion may have created possibilities for preparers to select and apply certain guidelines to match their personal reporting objectives. Building on this, preparer’s strong expression of disagreement with retrospective application might display concerns about the manner in which revenue was recognized in previous years, i.e. consciously accelerated or delayed revenues in previous years will be reversed in the adjusted financial statements. I hypothesize that firms who have accelerated or delayed revenues in previous years are more likely to lobby the

standard setter in the form of submitting comment letters.

3.3 Disclosure requirements

The comment letters revealed that the specified disclosure requirements for annual as well as interim financial reporting are another significant concern to preparers. There is a general consensus amongst preparers that existing revenue disclosure requirements already provide the necessary information for

(24)

23 users to understand the nature, timing and uncertainty of revenues and cash flows from contracts with customers. According to the respondents, this information is sufficient and appropriately presented in, for instance, the Management Discussion section (MD&A) of their financial statement filings.

Consequently, the amendments to the disclosure requirements are repeatedly referred to as being “excessive”, “redundant”, “overwhelming” and so on, as management itself does not prepare or use the information in question to run its business:

(-) we are hard pressed to believe that many of the prescriptive requirements (e.g., contract asset/liability and onerous performance obligation roll-forwards) would provide financial statement users with necessary financial information by virtue of the fact that companies such as BMC do not otherwise prepare or utilize such information internally to run their businesses. (CL55, DP130).

(-) the users of financial statements do not receive more decision-useful information if management itself does not consider these issues to be of great importance (CL80, DP250)

Many preparers also raised the concern that the proposed disclosures would contain forward looking information that would require considerable judgment and may be misleading to users of the financial statements:

These disclosures remain estimations: we believe that readers of financial statements will be misled into believing that they do have full predictive value by virtue of the fact that they are included in the audited financial statements. (CL26, DP51, emphasis added)

As a result, to comply with the disclosure requirement, we will be required to describe a myriad of judgments involved in even more components for each customer contract.(CL57, DP159, emphasis added)

(-) would likely require the disclosure of potentially sensitive and confidential information and require a high degree of speculation as to the likely success of such assignments.(CL82, DP275, emphasis added)

In addition, while the intent of the extensive disclosure requirements is to enable users of the financial statements to better understand the nature, amount, timing and uncertainty of revenues and cash flows from contracts with customers, preparers unanimously agreed on the excessive nature of the disclosure proposals and repeatedly referred to the potential of “disclosure overload”:

(25)

24 (-) we believe that the prescriptive disclosure requirements in the Proposed Standard are in excess of what should be necessary to enable users to understand key financial statement information.(CL55, DP130, emphasis added).

This would have the effect of adding volume and complexity, while diluting more important disclosures and actually reducing the overall usefulness of our financial statements for many users. (CL92, DP335, emphasis added).

(-) critical information might be lost or hard to find amongst all the disclosures included (CL139, DP579)

Finally, preparers believe that the proposed disclosure requirements could potentially lead to the disclosure of firm-confidential information:

We believe that certain disclosure requirements would lead to the disclosure of information which companies have legitimate reasons to maintain confidential.(CL57, DP161, emphasis added)

May result in entities inadvertently breaking confidential and commercially sensitive information (CL60, DP176, emphasis added)

In order to make the disclosure proposals feasible, most respondents primarily suggested to simplify (CL25, DP29; CL113, DP417) or eliminate (part of) the disclosure requirements in the Exposure draft (CL32, DP37; CL57, DP159; CL65, DP186; CL80, DP250; CL101, DP376; CL121, DP487; CL125, DP494; CL145, DP615). Furthermore, preparers recommended a more principle-based framework towards financial statement disclosures (CL32, DP37; CL26, DP48; CL55, DP130; CL85, DP297; CL117, DP458; CL135, DP573; CL141, DP593; CL142, DP610; CL154, DP679)

Also, qualitative analysis describing the underlying drivers of material changes in contract liabilities are proposed as an alternative to the disclosure requirements in the exposure draft (CL35, DP33; CL26, DP51) With respect to interim disclosures, preparers are generally of the opinion that they should only be mandated when there is a significant change since the fiscal year-end annual disclosure (CL25, DP30; CL66, DP194; CP71, DP212; CL76, DP234; CL81, DP282, CL85, DP301; CL118, DP471; CL119, DP484; CL127, DP509; CL135, DP578; CL144, DP613).

Interpreting the commentary received on the topic of annual and interim disclosures, two main requirements are especially of interest. The first aspect of the proposed disclosure requirements on which a lot of preparers commented is the requirement to disclose in tabular format a reconciliation

(26)

25 from the opening to the closing aggregate balance of contract assets and contract liabilities. This reconciliation includes, amongst other things, cash received, amounts transferred to receivables and the effects of business combinations. The second requirement is included in par. 118 and requires an entity to disclose information about its performance obligations in contracts with customers, for instance when an entity typically satisfies its performance obligation, the terms of payment and obligations for returns, refunds and types of warranties.

4. Summary

In this chapter, I condensed the data from the comment letters into dominant themes by means of inductive analysis. As a result, three dominant themes in the comment letters of preparers are identified. First of all, a lot of preparers raised the concern of having to acquire and implement new (accounting) information systems in order to comply with the new standard. Secondly, the matter of retrospective application is another prevailing concern in the comment letters. Lastly, the analysis revealed that the specified disclosure requirements for annual as well as interim financial reporting are another significant concern to preparers.

(27)

26

CHAPTER 4: DEVELOPMENT OF HYPOTHESES

1.Introduction

In the previous chapter I discussed three main themes arising from the comment letters submitted in response to the Revised Exposure Draft (2011). In this chapter, I will build on these themes and translate them into testable hypotheses.

2.Hypotheses Development

I examine whether discretion potential explains a firm’s choice to lobby the standard-setter in the development of the new revenue recognition model. In other words: are firms with certain financial attributes (i.e. discretionary revenues) more likely relative to other firms to lobby in the standard-setting process? In order to examine this, I develop two testable hypotheses.

The comment letters revealed that, in general, preparers worry that the “one-size-fits-all” model contains explicit requirements that leave less room for managerial discretion in the reporting process. According to the respondents, at least some of the discretion that is no longer present under the new standard is necessary in order to reflect the true underlying economics of transactions. This implies that respondents interacted in the standard-setting process primarily to improve and make financial statements more useful and informative for users. Moreover, reputational costs and increased monitoring provide an incentive for managers to provide users with decision-useful information, restraining them from using their reporting discretion in an opportunistic manner. Accordingly, there should be no difference in financial characteristics (i.e. discretionary revenues) between lobbying firms and non-lobbying firms:

H0: There is no difference in discretionary revenues between firms that lobby the standard setter by submitting comment letters and firms who do not

However, if firm decisions are a reflection of management’s interest, agency theory suggests that unverifiable discretion can be used by managers to serve their own interests. Previous empirical literature has documented that unverifiable discretion can increase the likelihood of opportunistic disclosure. For example, Holthausen and Watts (2001) argue that fair values that are not based on actively traded market prices, increase the probability of opportunistic disclosure. Another study of Ramanna (2008) shows that firm’s lobbying support for SFAS 142 impairment tests increases in their discretion potential under the test in estimating fair values to account for acquired goodwill. These previous studies suggest that preparers of the financial statements might behave opportunistically in their revenue reporting choices while there is reporting discretion that allows them to. In this light,

(28)

27 lobbying activities of preparers in the standard-setting process may be perceived as an attempt to influence the potential level of discretion under the new standard. I hypothesize that firms that engage in revenue management, i.e. firms with greater discretionary revenues, have greater incentives to lobby the standard setter:

H1: Discretionary revenues differ between firms that lobby the standard setter by submitting comment letters and firms who do not

(29)

28

CHAPTER 5: RESEARCH METHODOLOGY

1. Introduction

In this chapter, I will first outline the regression model used to examine the likelihood of writing a comment letter. Moreover, I will define the variables included in the model. After that, I will describe how I define and measure discretionary revenues.

2. Model specification

My main explanatory variable of interest is discretionary revenues. Discretionary revenues can be broadly distinguished in two main forms. Firstly, revenues can be managed by real activities, for example by providing discounts, lenient credit terms, and so on. Secondly, revenues can be managed without the use of real activities, for example by creating fictitious revenues or aggressive revenue recognition. In this thesis, discretionary revenues are defined as the difference between the actual change in receivables and the predicted change by the model (i.e. the residual).

In order to capture firm discretionary revenues, a model based on the model of Stubben (2010) is used. Total reported revenues are equal to the sum of nondiscretionary revenues ( ) and discretionary revenues ( _):

If c is the fraction of nondiscretionary revenues which remains uncollected at year end, accounts receivable equals the uncollected nondiscretionary revenues ( _{) and discretionary revenues}

( _:

Because discretionary revenues increase accounts receivable and total revenues with the same amount, discretionary receivables are equal to discretionary revenues. As a consequence, ending receivables are expressed in terms of reported revenues because nondiscretionary revenues are not observable. The following equation is then derived by taking first differences:

A firm’s discretionary revenues for each industry-year with at least 10 observations are then expressed by the residuals from the following equation:

(30)

29 Where represents the annual change in accounts receivable divided by total assetsand is

the change in annual revenues divided by total assets. The absolute value of discretionary revenues calculated by equation (2) is then multiplied by -1 to reflect the fact that higher values of discretionary revenues (DiscRev) represent higher accrual quality.

The expectation with regard to the relationship between discretionary revenues and lobbying is, like described in the previous two chapters, twofold. On the one hand, the lobbying can be perceived as an attempt of managers to reflect the true underlying economics of firm transactions. In this case, we would expect a negative relationship between writing a comment letter and the

magnitude of discretionary revenues. On the other hand, managers lobbying the standard-setter can be seen in the light of opportunistic behavior: their primary concern when they interact in the standard-setting process is to secure (potential) revenue management.

Next to my main variable, I also explore 8 additional variables. Previous studies have documented that the probability of engaging in lobbying activities increases with firm size (e.g. Watts & Zimmerman, 1978; Ramanna, 2008). Likewise, I expect that larger firms have more resources to dedicate to writing comment letters et cetera, so I hypothesize a positive relationship between firms size and lobbying. Furthermore, I predict a positive relationship between being audited by a big-4 accountancy firm and lobbying the standard-setter, because big-4 accountancy firms are generally more up to date with developments in accounting standards relative to smaller accountancy firms. With regard to firms with net income less than zero, I hypothesize a negative relationship to my dependent variable, because these firms are expected to be more concerned with surviving on the long term than engaging in lobbying activities. Asset turnover, which equals the amount of revenues generated per dollar of assets, is predicted to be positively related to lobbying because firms who are more efficient in generating revenues per dollar of assets might have a stronger incentive to input in the development of a revenue recognition standard. Finally, with regard to the variables return on assets (ROA), leverage, and the market-to-book ratio no relationship is predicted. Table 5.1 provides an overview of the predicted relationships. Table 5.1 Predicted relationships Variable Expected Sign Variable Expected Sign DISCR_REV /+ BIG4 + assetturnover + LOSS  ROA ? LEV ? SIZE + MTB ?

(31)

30 In sum, the probability of a firm engaging in lobbying activities with the standard setter, i.e. submitting comment letters, is hypothesized to be a function of the following variables:

Where:

= discretionary revenues = the residual of the annual change in accounts receivable;

= revenues / total assets;

= return on assets = net income / total assets; = size as measured by total assets;

= firms with Deloitte, E&Y, KPMG, PwC or Arthur Andersen as auditor; = firms with net income < zero.

= total debt

(32)

31

CHAPTER 6: RESULTS

1. Introduction

In this chapter, I will portray the results of the empirical tests. Paragraph 2 first introduces the sample characteristics and gives a summary of the descriptive statistics. In paragraph 3 the results from the univariate analysis, the logistic regressions and the robustness tests will be presented and described.

2. Sample and Descriptive Statistics

In total, 164 comment letters of preparers that responded on the Exposure Draft Revenue from Contracts with Customers are analyzed. For each of these companies that lobbied the standard-setter (“lobbyers” hereafter), financial data is collected from the Datastream database. Subsequently, a control group of firms (“non-lobbyers” hereafter) with similar (financial) characteristics is selected for comparison with the lobbying-group. Annual data from the year 1998 to 2014 is included in the sample for statistical analysis. The initial sample consists of 53546 observations, of which 336 observations relate to lobbying-firms and 53210 to non-lobbying firms. The sample of non-lobbying firms (i.e. the matched sample) is matched to the sample of lobbying firms based on the firm

characteristics geography and industry. For every lobbying firm which has sent a comment letter in 201216, all other firms from the same country and same industry are included in the matched sample.

TABLE 6.1 SAMPLE SELECTION # of Observations Sample lobbyers 336 Sample non-lobbyers 53210 Total sample 53546

Observations – non-annual firm observations 158

Observations – duplicates in firm observations 11

Observations – firms with total assets ≤ 0 1763

Observations – firms with sales and revenue=0 137

Observations – firms with change in AR=0 13661

Observations – industry-years with < 10 observations 666

Excluded observations (16396)

Final sample used in regression analyses 37150

16

Officially, the four-month comment letter period ranged from 13 December 2011 until 13 march 2012. The year 2012 is used here because all letters sent by preparers are dated from 2012.

Lobbying on accounting standards: a reflection of discretionary behaviour? : An empirical examination of lobbying activities in response to IFRS 15.