Revenue recognition: determinants of the accounts ...

_________________________________________

Master thesis Department Accountancy, Faculty of Economics and Business Studies,

Tilburg University

Revenue recognition: determinants of the accounts receivable and the deferred

revenue account

Arthur Schothuis BSc Administration number: 672053 Master Accountancy and Management Control Supervisor Tilburg University: Dr. A. Yim Second Reader: Dr. Y. Zeng Internship: KPMG accountants, Den Bosch Internship supervisor: Drs. C.G.W. Laureijssen RA Date of completion: August 2010

REVENUE RECOGNITION: DETERMINANTS OF THE ACCOUNTS RECEIVABLE AND THE DEFERRED REVENUE

ACCOUNT

BY ARTHUR SCHOTHUIS*

(August 2010)

Abstract: This study investigates how opportunistic behaviour, as well as some other determinants, might have an impact on the normal changes in the deferred revenue account and the accounts receivable. I focus on opportunistic behavior arising from cash deficiency or more generally from financial constraints. From previous literature it is stated that the importance of revenue recognition is big for managers, standard setters, investors and auditors. The amount as well as the timing of revenue recognition is important, while it is a fact that some managers behave opportunistic regarding this recognition. Also in this study is tried to explain why it is important to know what the influences of opportunistic behavior are.

Previous literature has shown that if a firm is cash or financial deficient there is a tendency to influence the numbers in the financial statements. A sample of firms that have US-GAAP as accounting standard are being used in this thesis. I create ten regression models in which I also control for risk factors that were omitted in prior research. The results imply that behaving opportunistic, measured in characteristics of cash deficient firms, leads to a negative extent of changes of the accounts receivable. With respect to the deferred revenue account, cash deficiency leads to a positive extent of changes of the deferred revenue account.

Keywords: Revenue recognition, accounts receivable, deferred revenue, cash deficiency.

Data Availability: All data are available from public sources. * Arthur Schothuis is a Master’s student at Department of Accountancy, Faculty of Economics and Business Administration, Tilburg University. Email: [email protected].

Acknowledgements

While writing this thesis I got help and support of several people. First, I want to thank

Dr. Andrew Yim, who supervised me at Tilburg University, for his supervision and advices. I

appreciate his very quick responses on my questions and the inspiration he gave me. Second, I

want to thank Drs. Corneel Laureijssen, who supervised me during the period of the

internship at KPMG accountants Den Bosch, for his help and taking care of me. And also for

the nice time I had when doing the internship. Third, I want to thank my parents, brothers and

friends who had always been supporting me and keeping their faith in me.

Tilburg/Den Bosch, August 2010

Arthur Schothuis

Table of contents

Abstract

Acknowledgements

Table of contents

Section 1 Introduction Page 1

1.1 Motive Page 1

1.2 Problem definition Page 1

1.3 Research method Page 2

1.4 Goal of the thesis Page 3

1.5 Reminder of the thesis Page 4

Section 2 Literature Review and Hypothesis Formulation Page 5

2.1 Definition revenue recognition Page 5

2.2 Economic event occurring and the timing of recognition Page 6

2.3 The importance of a good revenue recognition system Page 7

2.4 Revenue recognition and earnings management Page 9

2.5 Formulating hypotheses Page 11

Section 3 Regression Models and Sample Construction Page 13

3.1 Research method Page 13

3.2 Data collection Page 19

Section 4 Results Page 20

4.1 Descriptive statistics Page 20

4.2 Results Page 20

4.3 Comparison with Caylor’s (2009) model Page 23

Section 5 Discussion Page 24

5.1 Conclusions and implications Page 24

5.2 Limitations Page 28

5.3 Possibilities for follow-up research Page 28

References Page 30

Appendix Page 34

1

REVENUE RECOGNITION: DETERMINANTS OF THE ACCOUNTS RECEIVABLE AND THE DEFERRED REVENUE

ACCOUNT Section 1 Introduction

1.1 Motive

Revenue is usually the largest single item in financial statements, and the issues

involving revenue recognition are among the most important and difficult ones that standard

setters and accountants face. In recent years, concerns related to the recognition of revenue in

accordance with accounting standards have heightened significantly. Quite often, companies

end up tweaking the revenue numbers.

An informed market recognizes the effects of economic events when they occur, but

revenue recognition must await compliance with formal accounting recognition criteria

(Warfield and Wild, 1992). The cause of this lag is a function of cross-sectional differences in

the application of accounting recognition criteria.

If the revenue recognition rules are not defined clearly some forms of earnings

management can appear and also the lag causes that the true economic substance is not

presented well. An example of revenue recognition rules can be found in (the differences

between) the IFRS and US-GAAP accounting standard. Because of the different revenue

recognition rules, the level of earnings management as well as the size of the lag is different.

There is some criticism on the IFRS as well as on the US-GAAP rules that these are not

sufficient and that these need to be revised. (Schipper et al., 2009, Sunder, 2009 and

Wustemann and Kierzek, 2005)

Healy and Wahlen (1999) investigated that accounting standards determine the value

of the financial statements. Leuz (2003) claim that increasing the level or precision of

disclosure should reduce the likelihood of information asymmetries between investors and

increase market liquidity. Investors add value to the disclosures that a company provides. So

differences in revenue recognition rules can lead to different value of reported performances

to the users of the financial statements.

1.2 Problem definition

Despite the accounting standards, some managers still see some opportunities to act

opportunistic. Burgstahler and Dichev (1997) show that managers try to influence profit

numbers to avoid that a lower profit or a loss is presented. The reason why managers want to

2

avoid this, is because firms with a consistent pattern of earnings increases command higher

price-to-earnings multiples. Also firms breaking a pattern of consistent earnings growth

experience an average of 14% negative abnormal stock return in the year the pattern is

broken. Burgstahler and Dichev (1997) have found evidence that two components of earnings:

cash flow from operations and changes in working capital, are used to achieve increases in

earnings. Because the changes in accounts receivable and deferred revenue are part of the

changes in working capital, it is interesting to investigate what the influence of opportunistic

behavior is on how those accounts are built.

Prakash and Sinha (2009) have found that small changes in the deferred revenue

liability can have a disproportionately large impact on future profitability. While Marquardt

and Wiedman (2004) have found that firms issuing equity appear to prefer to manage earnings

upward by lifting up accounts receivable to accelerate revenue recognition.

The main question that is being asked through the thesis is: What are the main

determinants of the accounts receivable and the deferred revenue account?

Previous literature (Chevalier and Scharfstein, 1996 and Fazzari and Petersen, 1993)

has shown that cash or financial constraint firms have greater incentive to cut prices to get

short-run profits. Also constrained firms will draw working capital down during low cash-

flow periods and accumulate it during high cash-flow periods. Therefore when a firm is cash

constraint, there is a tendency to act opportunistic.

The first hypothesis is that behaving opportunistic, measured in terms of

characteristics of cash deficient firms, leads to a negative extent of changes of the accounts

receivable between years.

With respect to the deferred revenue account, the second hypothesis is that behaving

opportunistic, measured in terms of characteristics of cash deficient firms, leads to a positive

extent of changes of the deferred revenue account between years.

1.3 Research method

The method of recognizing accounts receivable and deferred revenue is taken from

Caylor (2009), this method is a combination between the methods distracted from Dechow et

al. (1998) and Kothari et al. (2005). However Caylor did not include some risk factors and the

opportunistic behavior variable in his model. Therefore 47 Fama-French industry dummies

are included as controls for differences between industries, for both models. Also I include

some year variable dummies to control for changes in accounting rules between the years of

my investigation. Furthermore, I included some control variables, based on Richardson et al.

3

(2005) to control for changes of other accruals and its components. Finally I include a dummy

variable for opportunistic behavior in my model to see how opportunistic behavior is

influencing the change in accounts receivable and deferred revenue and a control variable or

the opportunistic behavior variable times the sales in a certain period.

The model claims that changes in gross accounts receivable depend on total assets,

changes in sales and changes in cash flows from operations. Changes in deferred revenue

depends on the same financial numbers, but then from different time periods. Both models are

controlled for the mentioned risk factors and the opportunistic behavior dummy variable.

To be consistent with Caylor (2009), I investigated US-GAAP firms in the years 2001-2005. 1.4 Goal of the thesis

This study can contribute to existing literature by giving the building blocks of how

accounts receivable and deferred revenue are recognized, given that some managers act

opportunistic regarding those accounts. Because I have tried to control for industry, year and

accruals, those risk factors are captured as well. While theory has shown that there are some

differences between those determinants, for example differences between industries, the

practical, real numbers of how these accounts are built has not been investigated so far. It is

important to know what the differences are because decisions of investors can be influenced

by differences in those determinants. It can be useful to see what the influence of behaving

opportunistic is on the measurement of those accounts. A financial number regarding the

accounts receivable and the deferred revenue account can therefore be better interpreted. This

thesis shows if the assumptions that Caylor (2009) make are correct under my model: is the

accounts receivable indeed influenced, or influenced in the same way by the factors he

mentioned, or do the factors that I include play a bigger role. It also answers the question:

does the introduction of more variables in recognizing the accounts receivable and deferred

revenue account lead to a more significant model.

I expect that this investigation will confirm my hypothesis regarding the recognition of

gross accounts receivable and deferred revenue. This means that opportunistic behavior, leads

to a negative change the extent of changes in gross accounts receivable between years and to a

positive change in the extent of changes in deferred revenue between years. This could be an

indication of earnings management with respect to these accounts. But this thesis will show if

my reasoning is correct.

4

1.5 Reminder of the thesis

After the introduction section, in the second section is shown what the theoretical base

is of this thesis. In this part, based on the existing literature and investigations, a frame is

created on which this thesis is based. Also certain points that are important for this paper are

mentioned. The hypotheses that follow from the introduction and the existing literature are

formulated. In section 3 the research method is further explained, as well as the data

collection. In section 4 the descriptive statistics and the results are mentioned. Finally in

section 5 the conclusions and implications of this paper are explained. Also there are some

limitations of this thesis and some possibilities for follow-up research mentioned.

In the appendix the tables are displayed.

5

Section 2 Literature Review and Hypothesis Formulation

2.1 Definition revenue recognition

The Financial Accounting Standards Board (FASB) and the International Accounting

Standards Board (IASB) both use a different definition concerning revenue: “Revenues are

inflows or other enhancements of assets of an entity or settlements of its liabilities (or a

combination of both) from delivering or producing goods, rendering services, or other

activities that constitute the entity’s ongoing major or central operations.” (FASB Concepts

Statement No. 6 Elements of Financial Statements, paragraph 78), and “Revenue is the gross

inflow of economic benefits during the period arising in the course of the ordinary activities

of an entity when those inflows result in increases in equity, other than increases relating to

contributions from equity participants.” (IAS 18, paragraph 7)

In the FASB Concepts Statement number 5, paragraph 58 you can read that

“Recognition is the process of formally recording or incorporating an item in the financial

statements of an entity as an asset, liability, revenue, expense, or the like. Recognition

includes depiction of an item in both words and numbers, with the amount included in the

totals of the financial statements.” Mentioned in the IAS 18 standard about recognition is:

“Recognition means incorporating an item that meets the definition of revenue in the income

statement when it meets certain criteria”

So, according to the concepts and standards mentioned above, revenue recognition is

recording revenue, which suffices certain conditions, in the financial statements.

Although these standards and definitions are notable dated (1984 and 1993), they are

still applicable. Only concerning the implementation of the revenue recognition process some

discussion is possible. As an addition to the FASB Concepts Statement, the Securities and

Exchange Committee (SEC) published in the end of 1999 some detailed rules on the area of

revenue recognition, in the Staff Accounting Bulletin 101 (SAB 101). The SAB 101 was

revised in 2003 and the SAB 104 was published as its substitute. The reason of publishing

more detailed rules was that the SEC expressed in public their concern about the big amount

of points of controversy and the problems that companies experience concerning the current

revenue recognition.

Schipper et al. (2009) describes some different conceptual models for revenue

recognition: the customer consideration model and the measurement model. These models

were proposed at an AAA/FASB Financial Reporting Issues Conference, in order to replace

the current revenue recognition models. Some participants believe that the notion of an

6

earnings process is insufficiently precise, to provide a sound conceptual basis for revenue

recognition standards. Therefore two new models were presented at that conference. Both

models are based on contract asset and contract liability, so they both recognize revenue when

contract asset increases or contract liability decreases. The difference between the two

proposed models were that, at the customer consideration model revenue is recognized based

on the prices that are mentioned in the contract, while at the measurement model recognized

revenue is based on what really is paid.

2.2 Economic event occurring and the timing of recognition

Accrual accounting distinguishes cash inflows from revenues and cash out-flows from

expenses, recognizing the differences between cash flows and income as liabilities or assets.

The principles which govern the recognition of revenues (and expenses) are the key

determinants of the properties of accrual accounting information. (Dhutta and Zhang, 2002)

The recognition of economic events in accounting revenues tends to lag that of the

market. An informed market recognizes the effects of economic events when they occur, but

revenue recognition must await compliance with formal accounting recognition criteria. The

application of these criteria involves basic concepts as reliability, objectivity, conservatism,

and verifiability. It affects earnings in two ways: (1) current earnings will include recognition

of certain prior periods' economic events, and (2) current earnings do not recognize all of the

current period's economic events until future periods. (Warfield and Wild, 1992)

The reason that this is a big issue is because events occur now, they are recognized

over some time and therefore future periods' earnings possess explanatory power for current

returns. The incremental explanatory power of future periods' earnings varies inversely with

the length of the reporting period. Warfield and Wild (1992) indicated that in certain

instances, the recognition lag is of such magnitude that the explanatory power of future

periods' earnings for current returns more than triples that of current earnings.

For investors and other stakeholders it is sometimes hard to say what the explanatory

power of current returns and earnings is. It is unclear in which period current returns will lead

to future earnings. And what the influence of previous returns on current earnings is.

Therefore it is important that there are good rules to determine when some item should

be recognized. In the following paragraph it is shown more extensively why it is important to

have good revenue recognition rules.

7

2.3 The importance of a good revenue recognition system

The importance of revenue recognition (rules) is big for managers, standard setters,

investors and auditors. The amount as well as the timing of revenue recognition is important.

Many decisions of investors depend on the effects that revenue recognition has on the

financial statements. Because of the big importance of revenue recognition it is important that

good quality accounting standards exist. So that it should be clear what the interpretation of

an amount of a certain account is to all the users of the financial statements.

In the current conceptual framework of the IASB two objectives of financial

statements are mentioned. Financial statements should “provide information about the

financial position, performance and changes in financial position of an enterprise that is useful

… in making economic decisions” (Framework 12). They also should “show the results of the

stewardship of management, or the accountability of management for the resources entrusted

to it” (Framework 14).

It is important that financial statements are reliable and they provide a true and fair

view of the economic reality. Revenue recognition is therefore a crucial part of the financial

statements. “Revenue is a crucial part of an entity’s financial statements. Capital providers use

an entity’s revenue when analyzing the entity’s financial position and financial performance

as a basis for making economic decisions. Revenue is also important to financial statement

preparers, auditors and regulators.” (Discussion paper FASB, 2008)

The importance of good revenue recognition cannot be shown better by the discussion

paper “Preliminary Views on Revenue Recognition in Contracts with Customers”. The two

biggest standard setters in the world (actually two competitors) decided to work together on

the revenue recognition area by developing a new revenue recognition model. They do that

because they both see that it is important to have good revenue recognition system. I will

discuss the current issues, which are going on in getting one main accounting standard, more

extensively in the discussion section.

Srivastava (2008) explain that “Revenue is typically the largest and most value

relevant item in firms’ financial statements” and Wustemann and Kierzek (2005) confirm this

view with “It is widely recognized that revenue is one of the most important items in financial

statements and that revenue recognition is one of the most difficult issues that standard-setters

and accountants have to deal with”

Users of financial statements attach much value to the revenue that is reported. Users

have the intention to make investment decisions based on the financial statements. On the

base of trends and growth-development numbers, they evaluate the companies’ past

8

performance and make predictions about the possibilities for companies to create future

company value or future cash flows.

The timing of revenue recognition has a direct or indirect effect on almost all the parts

of the balance sheet and the income statement. This timing has significant influence on the

stock price. Analysts compare actual results with the predicted results. The classification and

recognizing of certain items related to revenue activities can have an effect on the

interpretation of the financial statements. (Chlala and Landry, 2001)

Revenue recognition and classification decisions can be subjective if accounting

standards are missing, unclear, or any form of authorization is not present. Management can

pressure auditors so that they must accept the choice of their accounting policies. This has an

effect on the quality of the information provided by the company in the financial statements.

In an investigation from March 1999 the National Commission on Fraudulent

Financial Reporting has investigated that more than fifty percent of all the cases of fraudulent

reporting has to do with overestimating revenue. The most common abuses included

recording sales that never took place, shipping products before customers agreed to delivery,

and booking revenue up front from long-term contracts.

Dobler (2008) mentioned in his investigation, to rethinking of the revenue recognition

process that “revenue recognition is one of the most crucial issues in financial reporting

internationally and was the prevalent source for recent accounting scandals”.

Revenue recognition does not only have an effect on external stakeholders, also

decisions inside the company depend on revenue recognition. Revenue recognition has an

effect on the amount of the accounts receivable, on the deferred revenue, and also on the

amount of cash. In proportion of recognizing revenue earlier or later, the amount of the

accounts receivable and the deferred revenue account will change.

This in turn has an effect on the net working capital of a company. With the net

working capital you can calculate a couple of ratios, with the net working capital you can

calculate the profitability of the company, calculate the company’s risks, or even calculate the

whole firm value. (Smith, 1980) This is one of the main issues why there should be good rules

to determine the accounts receivable, the deferred income and the amount of cash that has to

be reported in the financial statements.

However it is a fact that managers behave opportunistic, so the calculations or

expectations of some working capital ratios may be wrong. Therefore in this paper the

accounts receivable and the deferred revenue account are unraveled so that the standard

setters, investors and auditors can see how these accounts are built, and what exactly the

9

influence of all the chosen determinates, especially the opportunistic behavior variable is, on

those accounts.

2.4 Revenue recognition and earnings management

Revenue is usually the largest single item in financial statements, and the issues

involving revenue recognition are among the most important and difficult ones that standard

setters and accountants face. In recent years, concerns related to the recognition of revenue in

accordance with accounting standards have heightened significantly. Quite often, companies

end up tweaking the revenue numbers, besides some other reasons. Recording revenue

improperly is also a commonly used earnings management technique. The ever evolving

business models and the growing online economy have only compounded the issue. Earnings

management/issues with revenue recognition have been the subject of headlines in the United

States and in the other parts of the world in the last few years.

From prior studies it is shown that companies and managers try to influence the profit

numbers is such a way that no earnings decreases or losses must be presented in the financial

statements. (Burgstahler and Dichev, 1997) In this study it is showed that the number of

companies that show a small profit is much higher than the number of companies show a

small loss. The main reason for managers to prevent that a small loss is presented is a lower

cost of capital that they can achieve, higher stock rates and obtaining a bonus in case of

certain profit targets.

One way of earnings management is influencing the revenue recognition process.

(Marquardt and Wiedman, 2004) According to this paper earnings management consist from

specific changes in accruals: the unexpected component in the changes of: accounts

receivable, inventory, accounts payable, accrued liabilities, depreciation expense, and special

items.

As a reaction on Marquardt and Wiedman (2002) and Burgstahler and Dichev (1997),

Caylor (2009) has investigated that managers try to influence the accounts receivable and the

deferred revenue in order to prevent that they must show negative earnings surprises. When

managers think they could not fulfill the profit-expectations of investors they try to upward

the revenue recognition by making the amount of accounts receivable higher, and deferred

revenue lower. The research method that Caylor mentioned to model normal changes in

accounts receivable and normal changes in deferred revenue, which determine the causes and

consequences of earnings management on the accounts receivable and the deferred revenue, is

the same method that I am going to use in my investigation. However I am going to extend

10

this model by putting some more variables, control variables and an opportunistic behavior

variable in the model.

Healy and Wahlen (1999) said before that “accounting standards can provide

corporate managers with a relatively low-cost and credible means of conveying private

information on their firms’ performance to external capital providers and other stakeholders.”

and “standards add value if they enable financial statements to effectively portray differences

in firms’ economic positions and performance in a timely and credible manner”

Hunton et al. (2006) said that greater transparency in reporting formats facilitates the

detection of earnings management. The size of the transparency depends on the demand of the

accounting standard. According to this paper, the importance of standards of good quality is

very high.

Martínez-Solano and García-Teruel (2007) have investigated that “managers can

create value by reducing their firm’s number of days accounts receivable and inventories.

Equally, shortening the cash conversion cycle also improves the firm’s profitability.”

Feroz et al. (1991) claim that 50% of the SEC enforcement actions between 1982 and

1989 are a function of overestimating the accounts receivable. The main cause of this is that

revenue of sold goods is recognized too early. The SEC pays much attention to good revenue

recognition.

Shortly, the accounting standards influence the intensity of earnings management, one

way of earnings management is influencing the revenue recognition, and accounts receivable

and deferred income is a part of the revenue recognition process. With other words: the

amount of the accounts receivable and deferred income and how you can influence this

amount depend on the accounting standards.

Burgstahler and Dichev (1997) confirm this reasoning by claiming they found

evidence that two components of earnings, cash flow from operations and changes in working

capital, are used to achieve increases in earnings. Cash flow from operations and changes in

working capital, are directly linked with changes in accounts receivable and deferred income.

At the calculation of earnings management is normally the abnormal change in gross

receivables calculated, however I am going to calculate what the normal change is. That is

because the goal of this paper is to find what the determinants of the accounts receivable are

under normal circumstances, and not if the expectations are met or not.

11

2.5 Formulating hypotheses

Despite the accounting standards, some managers still see some opportunities to act

opportunistic. Burgstahler and Dichev (1997) show that managers try to influence profit

numbers to avoid that a lower profit or a loss is presented. The reason why managers want to

avoid this, is because firms with a consistent pattern of earnings increases command higher

price-to-earnings multiples. Also firms breaking a pattern of consistent earnings growth

experience an average of 14% negative abnormal stock return in the year the pattern is

broken. Burgstahler and Dichev (1997) have found evidence that two components of earnings:

cash flow from operations and changes in working capital, are used to achieve increases in

earnings. Prakash and Sinha (2009) have found that small changes in the deferred revenue

liability can have a disproportionately large impact on future profitability. While Marquardt

and Wiedman (2004) have found that firms issuing equity appear to prefer to manage earnings

upward by lifting up accounts receivable to accelerate revenue recognition.

Because the changes in accounts receivable and deferred revenue are part of the

changes in working capital, it is interesting to investigate what the influence of opportunistic

behavior is on how those accounts are built.

A way of defining opportunistic behavior is looking at the cash deficient firms. The

accounts receivable and the deferred revenue account both have to deal with differences

between receiving cash and the recognition of revenue. If a firm is short on cash, or also said

cash deficient, especially when the competence between firms is high, firms can try to lift up

the amount of cash by creating huge deferred revenue accounts. Chevalier and Scharfstein

(1996) investigated that when the demand is high, firms have greater incentive to cut prices

because the short-run profits from stealing market share are high relative to the long-run

profits from collusion. For example firms can try to sell their product by giving big price

reductions if customers pay in advance. In that case the company has a lot of demand for their

products in the future and receives now a lot of cash. These firms want to avoid that

customers pay in a later period. Constrained firms will draw working capital down during low

cash-flow periods and accumulate it during high cash-flow periods (Fazzari and Petersen,

1993).

The explanation for the negative coefficient on working capital is that working capital

competes with fixed investment for the limited pool of finance. Thus, other things equal,

when firms choose to decrease (increase) working-capital investment, fixed investment should

rise (fall). Similarly, fluctuations in cash flow will affect the inventory investment of

constrained firms (Carpenter et al, 1994). Therefore this kind of behavior wants to decrease

12

the size of the accounts receivable and increase the size of the deferred revenue account. To

investigate how this behavior influences exactly the accounts receivable and the deferred

revenue, this variable is included to measure opportunistic behavior. 1994).

Therefore the main question that is being asked through the thesis is:

What are the main determinants of the accounts receivable and the deferred

revenue account?

While the following hypotheses are formulated:

H1: Opportunistic behavior, measured in terms of characteristics of cash

deficient firms, leads to a negative extent of changes of the accounts receivable between

years.

H2: Opportunistic behavior, measured in terms of characteristics of cash

deficient firms, leads to a positive extent of changes of the deferred revenue account

between years.

13

Section 3 Regression Models and Sample Construction

3.1 Research Method

For the first, as well as the second hypothesis the research method that was mentioned

in Caylor (2009) is used. This method is a combination between the methods distracted from

Dechow et al (1998) and Kothari et al (2005).

In the paper of Caylor two models are mentioned: a standard model and a model that is

adjusted for that paper. I am going to use the standard model. That model claims that changes

in gross accounts receivable depend on total assets, changes in sales and changes in cash

flows from operations. Changes in deferred revenue also depend on total assets, changes in

sales and changes in cash flows from operations, but then from different time periods. For

example at deferred revenue there is a connection between sales in next year and cash flows

from operations in this year. While at accounts receivable there is a connection between sales

this year and cash flows from operations next year.

However Caylor did not include some risk factors in his model. Therefore 47 Fama-

French industry dummy variables are included as controls for differences between industries

in both models. Also I include some year variable dummies to control for changes in

accounting rules between the years of my investigation. I included some control variables,

based on Richardson et al (2005) to control for changes in other accruals and its components.

Despite the accounting standards, some managers still see some opportunities to act

opportunistic. Burgstahler and Dichev (1997) show that managers try to influence profit

numbers to avoid that a lower profit or a loss is presented. Therefore a variable to account for

this kind of opportunistic behavior is included as well.

Hypothesis 1

Dechow et al (1998) has indicated that the relation between sales and cash flow from

operations is not one-to-one because some sales are made on credit. Specifically, they assume

that a proportion of the firm’s sales remain uncollected at the end of the period.

The accounts receivable accrual incorporates future cash flow forecasts (collection of

accounts receivable) into earnings. Because in the next period some of the current amount of

accounts receivable is being paid. Therefore current sales will lead to future cash flows.

Changes in accounts receivable should be positively correlated to changes in current sales and

changes in future cash flow from operations.

14

Because changes in accounts receivable also have an effect on the total assets, this is

part of the model as well. This parameter measures for which part the accounts receivable is

dependent on the total assets of period t-1.

Kothari et al (2005) use assets as the deflator to mitigate heteroskedasticity in the

model. If every part of the model is divided by the total assets this would lead to more relative

results.

Furthermore, a constant term that is based on Kothari et al (2005) is included. The

advantages of including a constant term are that it provides an additional control for

heteroskedasticity not alleviated by assets as the deflator. Second, it mitigates problems

stemming from an omitted size (scale) or risk variable. And discretionary accrual measures

based on models without a constant term are less symmetric, making power of the test

comparisons less clear-cut. Thus, model estimations including a constant term allows to better

address the power of the test issues that are central when doing the analysis.

The reason that is chosen for gross accounts receivable instead of net accounts

receivable is that the provision for bad debt and the amount that is mentioned on the balance

sheet in the accounts receivable should be both included to calculate the total gross accounts

receivable. According to Caylor (2009), if I use the net accounts receivable this could

influence the results.

Also some control variables for risk are introduced. I control for differences between

industries. Because Prakash and Sinha (2009) indicate that industry characteristics also affect

predictability, 47 Fama-French industry dummies are included as controls in both models.

Year dummy variables are included to control for changes in accounting rules or some other

influences between years. It could be the case that accounting rules are subject to change over

the time period of my investigation, so therefore there has to be controlled for them.

Prakash and Sinha (2009) also explain that the inclusion of accruals and its

components as controls is a good thing to do to control for the changes in accruals.

Richardson et al (2005) decompose total accruals into three components: changes in working

capital (∆WC), changes in non-current assets (∆NCO) and changes in financing (∆FIN). They

showed that ∆NCO has greater explanatory power for future earnings than changes in

operating accruals (measured as ∆WC). To be consistent with Caylor (2009), I divide the

accrual control variables by total assets.

By definition, a firm is said to be cash deficient if the amount of cash that a firm owns

is lower than the short term or current liabilities minus the short term or current assets. The

motivation to use this measure was mentioned in Fazzari and Petersen (1993). They

15

mentioned that previous literature explains why working capital is one of the key elements of

the firm: “Inventory components of working capital enter directly into the production

function. For example, firms stockpile materials to reduce the probability of stockouts that

could slow production. They also use work-in-process inventories to achieve economies of

scale by running large batch sizes. Other components of working capital such as trade credit

and finished-goods inventories facilitate sales. Accounts receivable, in particular, can affect

sales to customers who are themselves liquidity constrained. Finally, cash and equivalents and

current liabilities affect costs through the liquidity of the firm. For example, compensating

cash balances can reduce financing costs, and adequate cash stocks allow firms to take

advantage of discounts for prompt payment.” So the liquidity or the ability to repay short-term

debt of the firm is one of the main issues to determine if the firm is cash deficient or not.

In this paper it is mentioned that is has been investigated before that firms use liquid

assets as collateral for short-term borrowing, which reduces working capital through an

increase in current liabilities. Therefore the measure mentioned above is used to see if firms

are cash deficient or not.

Cash deficient firms will draw working capital down during low cash-flow periods and

accumulate it during high cash-flow periods. Therefore when a firm is cash deficient, there is

a tendency to act opportunistic.

A second way of determining if a firm is cash deficient or cash constraint or not is

looking at its characteristics. Kaplan and Zingales (1997) have investigated that firms that are

financial constraint have certain characteristics. For example, they have a low sales growth,

compared to firms that are not financial constraint. They have a high debt-to-capital ratio, the

ratio between investments and capital is relatively low and the relationship between cash flow

and capital is even negative. Therefore these four characteristics are taken as measure for

opportunistic behavior as well. So that not only by definition, but also in real numbers there is

a measurement for opportunistic behavior.

So the opportunistic behavior variable is measured in terms of cash deficiency. The

dummy variable is equal to 1 if firms are cash deficient by definition or have certain

characteristics that a financial constraint firm has and equal to 0 if firms are not cash deficient.

The five measures of opportunistic behavior are run in parallel to see what the influence of

every measure is on the determination of the accounts receivable. Therefore, I create five

regression equations.

16

As another control variable I multiply the five opportunistic behavior dummy

variables by the change in sales in the current period divided by the assets at the beginning of

the year.

The formula to measure this hypothesis is:

Hypothesis 1: ∆gart / At-1 = α0 + α1*(1/At-1) + α2*(∆St/At-1) + α3*(∆CFOt+1/At-1) +

α4*(∆WCt /At-1) + α5*(∆NCOt /At-1) + α6*(∆FINt /At-1) + ∑ δ*IND + ∑ φ*YEAR + α7*opp +

α8*opp*(∆St/At-1)+ εt

Whereby:

∆gart = change in gross accounts receivable during year t

∆St = change in sales during year t

∆CFOt+1 = change in cash flow from operations during year t+1

At-1 = total assets at the beginning of year t

∆WCt = changes in working capital during year t

∆NCOt = changes in non-current assets during year t

∆FINt = changes in financing during year t

∑ δ*IND = 47 Fama-French industry dummies depending on the SIC-code of the company

∑ φ*YEAR = year dummy variables for the years 2005, 2006 and 2007

opp = a dummy variable equal to 1 if a manager behaves opportunistic

α0 = constant term

εt = error term

Hypothesis 2

For hypothesis 2 the same assumptions can be made as for hypothesis 1. Except, the

sales and cash flow from operations behave just the other way round: there has been paid for a

product while there are no sales facing it. The changes in deferred revenue are linked to future

sales. Products which has been paid for now, and that may recognized as sales in the future

are expressed now in a deferred revenue.

Deferred revenue is therefore linked to current cash flow from operations, because in

this period the deferred revenue generates cash flow. Changes in deferred revenue should be

positively correlated to changes in future sales. This is because the amounts that are

recognized as deferred revenue now, are recognized as sales in the future. Because changes in

17

deferred revenue also have an effect on the total assets, this is part of the model as well. In

this model a constant term and the mentioned control variables are included as well.

As for the first hypothesis the way of measuring opportunistic behavior is looking at

the cash deficient firms. If a firm is short on cash, or cash deficient, firms can try to lift up the

amount of cash by creating huge deferred revenue accounts. For example firms can try to sell

their product by giving big price reductions if customers pay in advance. In that case the

company receives now a lot of cash. The amount of deferred revenue will in that case be

higher. The dummy variable is equal to 1 if firms are cash deficient and equal to 0 if firms are

not.

Again, the five measures of opportunistic behavior are run in parallel and five

regression equations are created, one for every measure/characteristic of opportunistic

behavior.

As for hypothesis one I use one other control variables: I multiply the opportunistic

behavior dummy variables by the change sales in the next period divided by the assets at the

beginning of the year.

The formula to measure this hypothesis is:

Hypothesis 2: ∆def revenuet / At-1 = α0 + α1*(1/At-1) + α2*(∆St+1/At-1) + α3*(∆CFOt /At-1) +

α4*(∆WCt /At-1) + α5*(∆NCOt /At-1) + α6*(∆FINt /At-1) + ∑ δ*IND + ∑ φ*YEAR + α7*opp

α8*opp*(∆St+1/At-1)+ εt

Whereby:

∆def revenuet = change in short-term deferred revenue during year t

∆St+1 = change in sales during year t+1

∆CFOt = change in cash flow from operations during year t

The remaining the variables are defined as mentioned at hypothesis 1.

Regression analyses

I make in total ten regression analyses (five measures times the two hypotheses) to see

what exactly the influence is of all the parameters of the model on the changes in gross

accounts receivable and deferred revenue. During the whole process there is controlled for

some risk factors. Some differences between industries and years are therefore captured. It is

important to see what exactly the influence is of all those building blocks on the determination

of the two accounts. Investors can better interpret the value of these two accounts if they

18

know perfectly how these accounts are built. They can see for example that there are

differences between industries. Especially it is interesting to see how behaving opportunistic

influences the results. Will the result of the Marquardt and Wiedman (2004) paper, that firms

issuing equity appear to prefer to manage earnings upward by lifting up accounts receivable,

be confirmed under this setting or not?

There is one main difference between doing my regressions and the way that Caylor

(2009) does his regressions. Caylor (2009) runs his regressions by industry and fiscal year

using all available firms with the requisite data. The coefficient estimates are based on means

of industry-years and t-statistics are based on the standard error of those means. In other

words he runs for each industry and each year the regression separately and takes a (weighted)

average of those regressions to come to his coefficients. I think the reason that Caylor (2009)

is doing that is because he does not have control variables to capture differences between

industry and firm years. By taking the mean, he takes the averages between industries and

years into account in the regression equation as well. Using this method requires some good

programming skills, which I am not capable of. If I want to perform it in the same way, I have

to do all the regressions, and take all the averages manually instead of letting the computer do

this automatically. This would take a lot of time. Therefore instead, I take a pooled sample in

which all the industries and firm years are taken into account at the same time in the same

regression, but with control variables. The reason that this is justifiable to do that instead of

Caylor’s (2009) method, is because creating dummy variables is also a way to capture the

differences between years and industries.

This thesis will also show if the assumptions that Caylor (2009) make are correct

under my model: is the accounts receivable indeed influenced, or influenced in the same way

by the factors he mentioned, or do, because of the variables that I include, the new variables

play a bigger role. If for example the parameter of 1 divided by total asset in the beginning of

the period is not significant in my model, this could imply that the assumption that accounts

receivable depend on total asset in the beginning of the period, that was correct in Caylor’s

model, is not correct in my model. This could be caused because in this model I use some

control variables that capture the differences between industries and years. In general, if the

factors that are omitted by Caylor are indeed important, the estimation of my normal changes

model should be different. A higher R-square will indicate if the introduction of more

variables in the determination of the accounts receivable and deferred revenue account will

lead to a statistically different and more significant model.

19

3.2 Data Collection

To be consistent with Caylor (2009), the investigated firms are US firms that use US-

GAAP as accounting standard. The data is collected from the Compustat database. In this

database I can collect all the data I need for my investigation. I have a number of all the items

that I need and that number I am using when looking up a certain item.

As Caylor (2009) did, the industry years 2001-2005 are investigated, therefore data

over the years 2000-2006 are collected. I have chosen not to take any furthermore years into

my investigation, because I do no want to take into account the effects of the financial crisis

that started late 2006/early 2007. The SIC-code for each company is looked up as well, so I

can see in which industry the companies are operating. I can use them as control variables. I

can control for and see if there are some differences between the industries. I start with a total

number of firm-year observations of 73,657.

In the paper of Caylor (2009) and in the paper of Burgstahler and Dichev (1997),

companies with a SIC company code between 4400 and 5000, between 6000 and 6500 and

companies with a code higher than 9000 are excluded from the investigation. This means that

utilities, financial institutions and firms related to public administration are deleted from the

investigation, because this could influence the results significantly. For my investigation I

also exclude them because the results can as well be influenced significantly if I include them.

It is also important that firms act in accordance with the US-GAAP standard. If I

include firms that act in accordance to a different standard this could influence my results. In

Compustat there is an option to display according to which accounting standard the numbers

are prepared. All the firms that have a different accounting standard than US-GAAP are

deleted from my sample.

For the first hypothesis it is important that the firms have an accounts receivable that is

not equal to zero, otherwise they are deleted for my investigation. The same counts for the

second hypothesis: firms must have a deferred revenue account that is not equal to zero,

otherwise they are deleted for my investigation. In the descriptive statistics it is shown how

many companies for each hypothesis per industry are taken in my investigation. Also the top

and bottom 1% of all the variables are deleted in order to take care that the outliers do not

influence the results.

This results in the total sample of 28,848 firm-year observations. From the total

sample 22,184 of them can be included is this investigation for the first hypothesis. While

7,296 firm-year observations of the total sample can be included is this investigation for the

second hypothesis.

20

Section 4 Results

4.1 Descriptive statistics

As under Caylor (2009), table 1 provides descriptive information for the Fama–French

(FF) industry groups for fiscal years 2001–2005, ranked in ascending order by percentage of

firms in the industry with non-zero short-term deferred revenue (Fama and French, 1997). I

have included all the firms in the descriptive statistics for this table, so the firms that are

excluded for my sample are included in this table to give a good oversight of all the

industries. For fiscal year 2005, 25.81% of all firms reported non-zero short-term deferred

revenue. All 48 FF industry groups have some firms with short-term deferred revenue on their

balance sheets. Table 1 displays the top industry groups in terms of percent of firms with

short-term deferred revenue. This group includes industries as Shipbuilding, Railroad

Equipment (27.42%), Healthcare (26.22%), Computer Software (24.81%), Entertainment

(24.79%), and Measuring and Control Equipment (24.16%). From the top 10 industry groups,

five are services that are provided, while the other five are industrial fabricated products.

Looking at the firms that actually have non-zero deferred revenue on their balance sheet, the

companies in the industries Pharmaceutical Products and Retail have the highest mean

deferred revenue-to assets.

Table 2 provides descriptive statistics for all firms that have either accounts receivable

or deferred revenue for fiscal years 2001–2005. Gross accounts receivable has a mean of

about 294 million dollars and a mean change of approximately 1.7% of beginning total assets.

Deferred revenue has a mean of more than 38 million dollars and a mean change in deferred

revenue of approximately 2.1% of beginning total assets. Table 2 also provides information

pertaining to the control variables, SIZE and BM, which Caylor (2009) used to test his

hypotheses. These control variables were used for the abnormal change model of the accounts

receivable and the deferred revenue account. Because I measure the normal change of those

accounts, I have not included them in my regression analysis. But to be consistent with the

descriptive statistics of Caylor (2009), I calculated and displayed them in my descriptive

statistics.

4.2 Results

Table 4 reports the results of performing the five regression analysis for the first

hypothesis. In this hypothesis the influence of certain variables on the determination of the

changes in the gross accounts receivable is explained. To take care of the differences between

21

years and industries, there are some dummy variables created. Because they are not of direct

interest for this study, they are not displayed. The results are consistent with the hypothesis

that behaving opportunistic, measured in terms of cash deficiency, would lead to a negative

change in the extent of changes in gross accounts receivable between years. Most of the

parameters are significant at the 1% level, the models have R-squares between 33.0% and

33.5%.

The results show that in three out of five ways to measure cash deficiency there is a

negative influence in determining the accounts receivable. Only one, the ratio between cash

flow and capital, shows a positive influence. This positive relationship was not expected,

because if firms are short on cash they would like to reduce the accounts receivable in order to

get as much cash flow as they can right now. But in general, looking at the parameters for

behaving opportunistic you can conclude that cash deficient firms try to lower the accounts

receivable. The explanation for this direction is that firms that are cash deficient want to

receive as much cash as they can now and not in the future, in the form of an accounts

receivable. So one part of the change of the accounts receivable can be explained by the fact

that some managers behave opportunistic, and therefore influence this account. Four

parameters are statistically significant at the 10% level.

What is notable about the other numbers in the table is, that the change in accounts

receivable are almost not dependent on the cash flow from operations in period t+1. While

was expected that, because the accounts receivable are paid in the next period, there was a

positive relationship on the cash flow from operations in period t+1, it turned out that there

was almost no relationship. This can be explained by the fact that I have included industry

dummy variables for all industries. Receiving cash for accounts receivable can occur in

different time periods per industry and is therefore captured by the industry dummy variables.

A different explanation can be found in the nature of the accounts receivable. If a big part of

them are short-term accounts receivable, i.e. they are paid within one year, they are only

influencing the current CFO and not the CFO in next year.

If I compare the expected with the actual signs, I notice that all the signs are the same

direction as what was expected. This implies that the signs of my model are consistent with

previous literature.

What also is notable, is that the parameters under the five models hardly change,

except for the measure of opportunistic behavior and the control variables linked to these

variables of for the measure of opportunistic behavior. The statistical significance also hardly

22

differs for each variable per model. This implies that the influence of opportunistic behavior

on the models is statistically significant.

In table 5 the results for hypothesis two are presented. It was tested how changes in the

deferred revenue account depend on several variables that are found in previous literature.

In this hypothesis the same conclusion regarding the opportunistic behavior variables

can be drawn as for the first hypothesis. The results are consistent with the hypothesis that

behaving opportunistic, measured in terms of cash deficiency, would lead to a positive change

in the extent of changes in deferred revenue. The models have R-squares between 10.9% and

12.3%

Four of the variables that where part of the model to measure potential opportunistic

behavior have a positive parameter. This implies that companies that are cash deficient indeed

try to create a positive extent of changes in the deferred revenue account. Because they are

short on cash, they have incentives to lift up that account. For example, they can try to lift up

sales by giving price reductions if customers pay in advance. It turns out that this expectation

is true. So one part of the change in the deferred revenue can be explained by the fact that

some managers behave opportunistic, and therefore influence this account.

What can be seen from the other parameters is that the parameters for ∆St+1/At-1 and

∆CFOt /At-1 are in all the five models almost equal to zero. And none of the parameters that

are equal to zero, except one, are statistical significant at the 1% level. So the influence of

those variables on the change in the extent of changes in deferred revenue is not as much as

was mentioned in the literature and what I expected that the influence would be. For the

changes in sales it can be explained by the nature of that account. If there is a lot of short-term

deferred revenue included in the deferred revenue, a big part of the deferred revenue is

influencing the sales now, and not in the future. However if that is the case, the parameters for

change in CFO should be positive, but this is not the case. This can all be explained by the

industry and year dummy variables. There can be a lot of differences between industry about

when to recognize a certain transaction or not. Therefore the differences between industries

can be captured by those variables, and not in the change of CFO.

All the parameters have the same sign as expected. This implies that those signs are

consistent with previous literature. Most of the other variables are significant at the 1% level.

As for hypothesis 1, the parameters under the five models hardly change, the statistical

significance also hardly differs for each variable per model. This implies that the influence of

opportunistic behavior on the models is stable for each of the measures.

23

Table 3 displays the Pearson correlation matrix for the variables of hypothesis 1 and 2.

4.3 Comparison with Caylor’s (2009) model

Because the basis of the model in this thesis is the model of Caylor (2009), a

comparison of his model with mine (extended) model is done, in order to see if there any

similarities and differences.

I cannot directly compare the results of Caylor (2009) with the results of this paper

because there are some differences in the sample size of the two papers. Although I have tried

to reproduce Caylor’s (2009) sample as much as possible, there are some differences. To

control for differences between industries and years, Caylor (2009) used industry averages to

determine the results for the regression analysis. Because I do not have the program skills to

let the computer determine those averages, I used dummy variables for year and industries.

Therefore, to do the comparison correctly, I compare the original regression model of Caylor

(2009), performed on my data, with the results of the extended model, based on the same data.

The formula that Caylor (2009) used to measure normal changes in accounts

receivable was: ∆gart / At-1 = α0 + α1*(1/At-1) + α2*(∆St/At-1) + α3*(∆CFOt+1/At-1) + εt. And

for the measure of the normal changes in deferred revenue the model was: ∆def revenuet / At-1

= α0 + α1*(1/At-1) + α2*(∆St+1/At-1) + α3*(∆CFOt /At-1) + εt.

If I compare the result of the extended model with the results by using Caylor’s

model, the first thing that is noticeable is the adjusted R-square of the models. For the first

hypothesis, the extended model have adjusted R-squares of about 33.3%, while Caylor’s

model has an adjusted R-square of 28.5%. The signs of the parameters, are the same, while

the parameters in the extended model are as statistical significant as those from the original.

This comparison implies that, based on the adjusted R-square, including more and control

variables do lead to a statistical better model for this hypothesis.

For the second hypothesis I have adjusted R-squares between 10.9% and 12.3%, while

the original model has an adjusted R-square of 10.5%. The signs of the parameters are all the

same, while the parameters of the original model are statistical more significant. For this

model it is hard to claim which of the models is better. Although the adjusted R-squares are

slightly higher for the extended model, the parameters are more significant at the original

model.

24

Section 5 Discussion

5.1 Conclusions and implications

In this paper is being tried to investigate what the effect of behaving opportunistic is

on the revenue recognition process. From previous literature it is stated that the importance of

revenue recognition is big for even managers, standard setters, investors and auditors. The

amount as well as the timing of revenue recognition is important. For example many decisions

of investors depend on the effects that revenue recognition has on the financial statements.

Previous literature also has found that some managers try to act opportunistic when it

comes to revenue recognition. They try to lift up the revenue in order to present good results.

It has been investigated that more than 50% of all the cases of fraudulent reporting has to do

with overestimating revenue. To measure if managers act opportunistic I look at the firms that

are cash deficient. It has been investigated before that cash deficient firms will draw working

capital down during low cash-flow periods and accumulate it during high cash-flow periods.

Also, some other papers have found similar results regarding influencing the financial

numbers at cash deficient firms. Therefore when a firm is cash deficient, there is a tendency to

act opportunistic.

This paper investigates what exactly the effect is of behaving opportunistic on the

determination of the accounts receivable and the deferred revenue account. To investigate that

influence, I extended a model to determine how the accounts receivable and the deferred

revenue account are established. I based this on some other papers. The model was extended

by some more determinants of these accounts, some risk factors to capture differences

between industries and years and the measurement of opportunistic behavior.

My first hypothesis was that behaving opportunistic, measured in terms of cash

deficiency, would lead to a negative change in the extent of changes in gross accounts

receivable between years. The results are consistent with this hypothesis.

The results are consistent with the second hypothesis that behaving opportunistic,

measured in terms of cash deficiency, would lead to a positive change in the extent of changes

in deferred revenue.

These results imply that behaving opportunistic, measured in two ways, indeed

influence, in either a positive or negative way, the determination of the revenue recognition

process. Users of the financial statements must be aware of this fact. The expectation is that

users of financial statements can better interpret the value of the two accounts if they know

that it is not only the other financial numbers that influence the revenue recognition process,

25

but also the opportunistic behavior of managers can influence the process. When making

decisions that are dependent on these accounts or make predictions of these accounts,

investors must know that opportunistic behavior of managers influences the extent of changes

between years of these accounts.

Auditors must also know that these accounts are being used to lift up cash-flows,

revenue or earnings.

My results are consistent with Burgstahler and Dichev (1997) and Marquardt and

Wiedman (2004): managers try to influence the amounts mentioned on these accounts. Also

the investigation that has been done in this paper gives some academical evidence for the

claim that 50% of the SEC enforcement actions are a function of overestimating the accounts

receivable, especially in cases where firms are cash deficient.

Because of the current discussion about getting one main accounting standard in the

world instead of two, it is interesting to mention theoretically what the implications are of the

differences between the two mostly used accounting standards in the world. I think the reader

of this paper will get some more theoretical background in the area of revenue recognition, by

mentioning the current issues regarding this area that are going on in the FASB and the IASB.

The results of this paper are being used in this discussion as well.

Previous literature has shown that there are some differences in IFRS and US-GAAP

regarding revenue recognition. The question that I ask myself when reading that literature is:

knowing the differences between IFRS and US-GAAP, how can this change the way

stakeholders make their decisions?

Historically, the US has required non-US firms listing on US exchanges to provide

reconciliations to US-GAAP of earnings and book value of equity. This requirement stems

from the belief that US investors can make better investment decisions regarding non-US

firms if the investors have access to information about these firms that is “similar” and of

“similar quality” to that available for US firms (Jenkins, 1999)

But now these firms aren’t required any more to make reconciliations. So the investors

must deal with the IFRS and US-GAAP existing next to each other. And literature has shown

that there are some differences in IFRS and US-GAAP. So the decisions of investors can

change when changing the accounting standard. (Pownall, Schipper, 1999)

Economic theory suggests that information asymmetries between potential buyers and

sellers of firm shares introduce adverse selection into share markets, and hence reduce market

liquidity. Information asymmetries are costly to firms, as investors adjust prices to

26

compensate for holding shares in illiquid markets. Increasing the level or precision of

disclosure should reduce the likelihood of information asymmetries between investors and

increase market liquidity. (Leuz, 2003) The level of disclosure depends on the accounting

standards, and therefore decisions of investors are influenced, as mentioned above, can

change.

Harris and Muller (1999) examine Form 20-F reconciliations from IAS to US- GAAP.

They find that, based on reconciliation magnitudes, IAS are closer to US- GAAP than other

foreign GAAP, but that reconciliation items are incrementally value relevant. They interpret

their findings as evidence that IAS and US-GAAP accounting measures are not substitutes. So

for investors it is relevant to know that the two standards exist next to each other and the

analyzing of differences between them give incrementally value.

Dobler (2008) mentioned that the accounting scandals and the proven incapable

accounting standards have motivated the international accounting standard setters to the revise

the recognition process.

Standard setters also find it useful to know what the differences between IFRS and

US-GAAP are. Gordon et al. (2008) state that US-GAAP differs from IFRS because of

differences in the enforcement and regulatory reporting environments. The need for users of

financial statements, to receive useful information, can be provided by preparers at a

reasonable cost, as a basis for making economic decisions. If there are differences in the rules

provided by the preparers, this will lead to different information that is provided to the users.

To take care that these differences on the area of revenue recognition will be as low as

possible the FASB and the IASB initiated a joint project on revenue recognition, primarily to

clarify the principles for recognizing revenue. (The discussion paper “Preliminary Views on

Revenue Recognition in Contracts with Customers”) They compared the two standards on the

area of revenue recognition and come to the conclusion that there are many differences

between the standards. Therefore the board, that deals with the joint project, make the claim

that a new model have to be developed to improve financial reporting by providing clearer

guidance on when an entity should recognize revenue.

That model should use one recognition principle, that is constant applicable to all kind

of transactions. The focus of the new revenue recognition model should be on changes in

assets and liabilities, because changes in these two parts of the financial statements can show

the revenue in the most distinct way. In the paper it is mentioned that “revenue is an increase

in assets, a decrease in liabilities or some combination of the two.” In the continuation of that

paper a whole new recognition model is explained. Another issue in this paper, and also of

27

some other literature, is the question if reporting in accordance with fair value solves the

problems that appear with the current revenue recognition processes. (The so-called Asset and

Liability Fair Value Approach).

Wustemann and Kierzek (2005) also take part into this collaboration. They

investigated three different conceptual frameworks that all are based on the asset and liability

view. (The Asset and Liability Fair Value Approach, the Asset and Liability Performance

Approach and the Asset and Liability Transaction Approach). Their main criticism on IFRS is

that the two current IFRS standards do not sufficient clarify how to deal with all kinds of

different transactions. More rules are needed to clarify revenue recognition for all those

different transactions. “Even though the general recognition criteria in the IASB Framework,

par. 83 apply to all sorts of income, the required level of probability and reliability of

measurement of the inflow of future economic benefits varies between different revenue-

generating transactions”. Their new proposed conceptual framework should remove these

objections.

This new model or framework should improve the comparability and

understandability of revenue for users of financial statements. If the standard setter knows

what the differences between two accounting standards are, he can find the inconsistencies

between the two standards and can take action to reduce the inconsistencies by adjusting

current rules or writing new rules. For example knowing the differences will show if the

criticism regarding that rule-based standards reflecting the true economic substance more

badly than principle-based standards on this area is correct. (Van der Meulen et al., 2007)

For auditors is it also useful to know what the differences between IFRS and US-

GAAP are. As said before, in an investigation from March 1999 the National Commission on

Fraudulent Financial Reporting has investigated that more than fifty percent of all the cases of

fraudulent reporting has to do with overestimating revenue. Therefore, differences between

IFRS and US-GAAP, especially on the area of revenue recognition, will lead to different

kinds or possibilities of frauds and therefore there are different important pitfalls that the

auditor needs to be aware of.

Taken the results of this paper in the perspective of the critique on the US-GAAP

accounting standard, and also on joint project mentioned before between the FASB and the

IASB on revenue recognition, it turned out that acting opportunistic is still possible under US-

GAAP. Because acting opportunistic is still possible under the US-GAAP, this confirms the

view that maybe the makers of these standards must clarify the principles for recognizing

revenue or a new model have to be developed to improve financial reporting. For example by

28

providing clearer guidance on when an entity should recognize revenue, in order to take care

that such opportunistic behavior is not possible in the future anymore.

5.2 Limitations

There are some limitations if you look at this investigation. This investigation has been

done by comparing firms that are using US-GAAP. But if you want to get a full view it can be

better to also include firms that are using a different accounting standard. For example

European firms that use IFRS or domestic American firms that use a different standard than

US-GAAP. You can maybe run the regressions in parallel for IFRS and US-GAAP and see

what the results are. Or you can create dummy variables for the accounting standard. If you

allow different accounting standards into your sample you get more companies/data and your

investigation will be more representative and generalizable.

To get a complete view of revenue recognition you can find more items in the balance

sheet that you can compare. Now the accounts receivable and the deferred revenue account

are compared but you can also compare some other items like for example inventory or the

value of fixed assets to see what the influence of behaving opportunistic in on those accounts.

Maybe you can draw conclusions regarding the influence of behaving opportunistic on

revenue recognition in general, and not only draw conclusions on the accounts receivable and

deferred revenue account.

To be consistent with Caylor (2009) I used firm data of the years 2001-2005. I do not

know if these years are representative for the years 2006-2009, or for any years in the past.

5.3 Possibilities for follow-up research

Following from the limitations there are some possibilities for follow-up research. A

follow-up investigation can be done by including more companies or by investigating more

items in the balance sheet or income statement. You can also focus more on differences

between two industries. You can include more years to see if the results are significant over

time, or that you can see a trend in the results, or there is maybe somewhere a temporary or

structural break.

Roychowdhury (2006) suggests that activities manipulation (in this case act

opportunistic to manage earnings) seems to vary positively with the stock of inventories and

account receivables. An example that he gives is that a firm with substantial credit sales to

dealers can more easily engage in accelerating the recognition of sales by shipping goods

early to its dealers and booking receivables. In his paper opportunistic behavior is defined as

29

“positive abnormal cash flow from operations (CFO)”. The author of this paper claims that

the abnormal CFO can differ from the normal CFO because of three reasons: sales

manipulation, reduction of discretionary expenditures and overproduction. Therefore it can be

interesting as well to use abnormal CFO as indicator for opportunistic behavior.

Following up on the discussion mentioned in paragraph 5.2 it can be useful if you can

investigate if there are differences between IFRS and US-GAAP regarding the determination

of the accounts receivable and deferred revenue account. As it has turned out, the importance

of a good revenue recognition system is high. Pownall and Schipper (1999) think that this

comparison might be assessed in any of four ways:

- two or more sets of standards might be directly compared to see if their requirements

differ

- accounting policies for a set of firms could be compared cross-sectionally, and

differences in accounting policies for demonstrably similar events/transactions would

be taken as evidence of noncomparability.

- accounting practices within a given accounting policy could be assessed

- the reported numbers themselves might be assessed

For example at the first one, you can give an exact description of how IFRS and US-

GAAP differ. You can write down how, in theory, what the differences are. With the second

research method you can use the 20-F reconciliation forms to see what the differences are.

The third method focus on differences in the notes that explain the financial statements. You

can determine what the real accounting practices of the rules are. You can see for example,

that the IFRS and US-GAAP rules are the same, but that in practice a different interpretation

is given to these standards. Or in the fourth case you can compare the reported numbers

themselves. You can compare the heights of the accounts receivable and deferred revenue

account and see how they differ.

30

References

Bauman, M., 2000, The Unearned Revenue Liability and Firm Value: Evidence from the

Publishing Industry, University of Illinois at Chicago

Briner, R., 2001, Subtle Issues in Revenue Recognition, The CPA Journal

Burgstahler, D., Dichev I., 1997, Earnings management to avoid earnings decreases and

losses, Journal of Accounting and Economics 24, pp.99-127

Carpenter, R. et al., 1994, Inventory (Dis)Investment, Internal Finance Fluctuations, and

the Business Cycle, Brookings Papers on Economic Activity, 2:75 123.

Caylor, M., Strategic revenue recognition to achieve earnings benchmarks, J. Account.

Public Policy (2009), doi:10.1016/j.jaccpubpol.2009.10.008

Chevalier, J., Scharfstein D., 1996, Capital-Market Imperfections and Countercyclical

Markups: Theory and Evidence, The American Economic Review, Vol. 86, No. 4

(Sep., 1996), pp. 703-725

Chlala, N. and Landry, S., 2001, Revenue recognition: More stringent rules on the

horizon, Ivey Business Journal

Dechow, P. et al., 1998, The relation between earnings and cash flows, Journal of

Accounting and Economics 25 133-168

Dobler, M., 2008, Rethinking revenue recognition – the case of construction contracts

under International Financial Reporting Standards, International Journal of Revenue

Management, Vol. 2, No. 1, pp.1-22.

Dutta, S. and Zhang, X., 2002, Revenue Recognition in a Multiperiod Agency Setting,

Journal of Accounting Research, vol. 40, issue 1, pages 67-83

Fama, E. and French, K., 1997, Industry costs of equity, Journal of Financial Economics

43, 153–193.

Fazzari, S. and Petersen, B., 1993, Working Capital and Fixed Investment: New Evidence

on Financing Constraints, The RAND Journal of Economics, Vol. 24, No. 3 (Autumn,

1993), pp. 328-342

Feroz, E. et al., 1991, The financial and market effects of the SEC’s accounting and

auditing enforcement releases, Journal of Accounting Research, Vol. 29 1991

Financial Accounting Standards Board, 1984, Statement of Financial Accounting

Concepts No. 5, CON5 Status Page, Recognition and Measurement in Financial

Statements of Business Enterprises

31

Financial Accounting Standards Board, 1985, Statement of Financial Accounting

Concepts No. 6, CON6 Status Page, Elements of Financial Statements a replacement

of FASB Concepts Statement No. 3 (incorporating an amendment of FASB Concepts

Statement No. 2)

Financial Accounting Standards Board, 2000, Statement of Financial Accounting

Concepts No. 7, CON7 Status Page, Using Cash Flow Information and Present Value

in Accounting Measurements

Financial Accounting Standards Board, 2008, Preliminary Views on Revenue Recognition

in Contracts with Customers, Discussion Paper

Fraudulent Financial Reporting: 1987-1997 - An Analysis of U.S. Public Companies

(1999)

Gordon, E. et al., 2008, Could IFRS Replace US GAAP? A Comparison of Earnings

Attributes and Informativeness in the US Market, Working Paper Series

Hail, L. et al., 2009, Global Accounting Convergence and the Potential Adoption of IFRS

by the United States: An Analysis of Economic and Policy Factors, Working Paper

Series

Harris, M. and Muller, K. (1999), The market valuation of IAS versus US GAAP

accounting measures using Form 20-F reconciliations, Journal of Accounting and

Economics, Vol. 26, pp. 285-312

Healy, P. and Wahlen, J., 1999, A review of the earnings management literature and its

implications for standard setting, Accounting Horizons Vol. 13 No. 4 December 1999

Hunton, J. et al., 2006, Financial Reporting Transparency and Earnings Management, The

Accounting Review, 81 (1), 135 (2006)

International Accounting Standards Board, 2008, Preliminary Views on Revenue

Recognition in Contracts with Customers, Discussion Paper

International Accounting Standards Board, IAS 11: Construction Contract

International Accounting Standards Board, IAS 18: Revenue

International Accounting Standards Board, Framework for the Preparation and

Presentation of Financial Statements

Jenkins, E., 1999. Financial Reporting in a Global Capital Market World, Financial

Accounting Series, No. 198-A (June 29), 2-6.

Kaplan, S. and Zingales, L., Do Investment Cash Flow Sensitivities Provide Useful

Measures of Financial Constraints?, The Quarterly Journal of Economics, Vol. 112,

No. 1 (Feb., 1997), pp. 169-215

32

Kothari, S. et al., 2005, Performance Matched Discretionary Accrual Measures, Journal of

Accounting and Economics, 39 (2005) 163–197

Leuz, C. et al., 2003, Earnings Management and Investor Protection: An International

Comparison, Journal of Financial Economics, 69, 2003 September 505-527

Leuz, C., 2003, IAS Versus U.S. GAAP: Information Asymmetry-based Evidence from

Germany’s New Market, Journal of Accounting Research 41: 445-427

Marquardt, C. and Wiedman, C., 2004, How are Earnings Managed? An Examination of

Specific Accruals, Contemporary Accounting Research, 21 (2), 461-491

Martínez-Solano, P. and García-Teruel, P., 2006, Effects of working capital management

on SME profitability, International Journal of Managerial Finance, 2007, vol. 3, issue

2, pages 164-177

Meltzer, A., 1960, Mercantile Credit, Monetary Policy and the Size of Firms, Review of

Economics and Statistics, 42: 429-37

National Commission on Fraudulent Financial Reporting, 1999, Report of the National

Commission on Fraudulent Financial Reporting

Pownall, G. and Schipper, K., 1999, Implications of Accounting Research for the SEC’s

Consideration of International Accounting Standards, Accounting Horizons, Vol. 13

(3), pp. 259-280.

Prakash and Sinha, 2009, Deferred Revenues and the Matching of Revenues and

Expenses, Boston U. School of Management Research Paper No. 2009-8, available at

http://ssrn.com/abstract=1316286

Richardson, S., et al., 2005, Accrual reliability, earnings persistence and stock prices,

Journal of Accounting and Economics 39: 437-485.

Roychowdhury, S., 2006, Earnings Management through Real Activities Manipulation,

Journal of Accounting and Economics, Volume 42, Issue 3, December 2006, Pages

335-370

Schipper, K. et al., 2009, Reconsidering Revenue Recognition, Accounting Horizons

Vol. 23, No. 1 2009 pp. 55–68

SEC, SEC Staff Accounting Bulletin No. 101 – Revenue Recognition in Financial

Statements

SEC, SEC Staff Accounting Bulletin No. 104 - Topic 13: REVENUE RECOGNITION

33

Smith K., 1980, Profitability Versus Liquidity Tradeoffs in Working Capital Management,

Readings on the Management of Working Capita, Ed. K. V. Smith, St. Paul, West

Publishing Company, pp. 549-562

Srivastava, A., 2008, Do Firms Use Flexibility in Revenue-Recognition Rules to Convey

Value-Relevant Information or to Manage Earnings?, Working Paper Series

Stubben, S., 2008, Discretionary Revenues as a Measure of Earnings Management,

Working Paper Series

Sunder, S., 2009, IFRS and the Accounting Consensus, Accounting Horizons Vol. 23,

No. 1 2009 pp. 101–111

Van der Meulen, S. et al., 2007, Attribute Differences Between US GAAP and IFRS

Earnings: An Exploratory Study, The International Journal of Accounting, 42 (2007)

123–142

Warfield, T. and Wild, J., 1992, Accounting Recognition and the Relevance of Earnings

as an Explanatory Variable for Returns, The Accounting Review, Vol. 67, No. 4 (Oct.,

1992), pp. 821-842

Wustemann, J. and Kierzek, S., 2005, Revenue Recognition under IFRS Revisited:

Conceptual Models, Current Proposals and Practical Consequences, Accounting in

Europe, Vol. 2, 2005

34

Appendix: Tables

On the next pages the four tables are displayed. The notes belonging to the tables can

be found after each table.

35

Table 1

Short-term deferred revenue by Fama–French industry group.

FF Industry N Proportion of firms with Mean deferred Median deferred

deferred revenue (%) revenue-to-assets (%) revenue-to-assets (%) Tobacco Products 36 11.11 1.49 1.27

Rubber and Plastic Products 319 15.36 11.91 5.31

Construction Materials 446 15.70 10.17 2.86

Beer & Liquor 102 17.65 5.12 1.91

Consumer Goods 458 18.12 8.96 2.97

Textiles 108 18.52 6.85 2.85

Automobiles and Trucks 438 18.95 7.78 3.50

Precious Metals 281 19.22 7.38 4.00

Shipping Containers 96 19.79 5.31 3.59

Aircraft 150 20.00 7.03 4.87

Electrical Equipment 504 20.24 7.91 3.00

Computer Hardware 815 20.37 9.33 2.81

Food Products 516 20.54 7.73 2.28

Agriculture 130 20.77 7.18 4.09

Non-Metallic and Industrial Metal

Mining 274 20.80 13.77 1.70

Candy & Soda 105 20.95 5.76 2.04

Utilities 1164 21.13 8.75 2.67

Business Supplies 293 21.16 8.56 3.86

Chemicals 694 21.18 7.80 2.87

Recreation 313 21.41 10.66 2.29

Petroleum and Natural Gas 1437 21.43 10.67 3.19

Printing and Publishing 242 21.49 5.99 2.70

Almost Nothing 1251 21.66 9.55 3.24

Pharmaceutical Products 2655 21.69 52.67 3.14

Medical Equipment 1292 21.75 8.68 3.07

Insurance 1290 21.78 7.33 2.76

Coal 78 21.79 5.14 1.76

Banking 10967 21.89 14.02 3.03

Steel Works Etc 502 21.91 10.49 3.66

Construction 332 21.99 9.70 3.83

Retail 1730 22.02 39.51 2.92

Transportation 967 22.13 9.31 2.49

Wholesale 1180 22.37 8.42 3.11

Defense 62 22.58 12.26 1.72

Personal Services 446 22.65 11.25 2.69

Real Estate 388 22.94 8.09 3.33

Restaraunts, Hotels, Motels 692 22.98 10.35 3.72

Electronic Equipment 1914 23.20 7.85 2.92

Machinery 950 23.37 8.23 2.39

Communication 1232 23.38 7.96 2.89

Apparel 483 23.40 8.47 2.30

Trading 2404 23.50 7.98 2.89

Business Services 2065 23.54 8.93 3.96

Fabricated Products 89 23.60 17.64 8.35

36

Measuring and Control Equipment 592 24.16 8.27 2.94

Entertainment 710 24.79 7.23 2.25

Computer Software 3221 24.81 47.50 2.73

Healthcare 675 26.22 8.30 3.20

Shipbuilding, Railroad Equipment 62 27.42 10.93 2.97

Total and average 47150 22.22 16.41 3.04

Note to table 1: this table reports the Fama and French (1997) industry name, total number of non-missing and non-zero observations for the ratio of short-term deferred revenue-to-total assets, percentage of firms in that industry with non-missing and non-zero short-term deferred revenue, as well as the mean and median of the ratio of short-term deferred revenue-to-total assets. Ratios are multiplied by 100% to convert to percentages for expositional purposes.

37

Table 2

Descriptive statistics

Mean Std. dev. 25% 75% Firms with accounts receivable Gross A/R (in $ mil) 293.538 2988.251 2.720 90.658 ∆Gross A/R 0.017 0.115 -0.018 0.032 SIZE 2.105 1.118 1.281 2.787 BM 0.468 1.182 0.199 0.791

Firms with deferred revenue Deferred revenue (in $ mil) 38.291 260.469 0.351 12.752 ∆Def Rev 0.021 0.627 -0.005 0.018 SIZE 2.125 1.049 1.405 2.723 BM 0.328 0.749 0.139 0.597

Note to table 2: this table provides descriptive statistics for firms with accounts receivable and firms with deferred revenue. All variables are scaled by lagged total assets, except for the raw values of gross accounts receivable and deferred revenue, book-to-market ratio, and log of size. ∆Gross A/R is defined as the change in gross accounts receivable. ∆Deferred revenue is defined as the change in short-term deferred revenue. SIZE is the natural logarithm of a firm’s size using beginning of the year market value of equity. BM is the beginning of the year book-to-market ratio.

38

Table 3

Panel A: Pearson correlation matrix for the variables of hypothesis 1

∆gart / At-1

1/At-1 ∆St/At-1 ∆CFO t+1/ At-1

∆WCt / At-1

∆NCOt

/At-1 ∆FINt / At-1

Cash deficiency

It/Kt+1 Sales growth

Debt to capital

CFt/Kt+1

∆gart / At-1 Pearson 1,000 Sig. (2-tailed)

1/At-1 Pearson ,026*** 1,000

Sig. (2-tailed) ,000 ∆St/At-1 Pearson ,522*** -,012 1,000

Sig. (2-tailed) ,000 ,085 ∆CFOt+1/At-1 Pearson ,014** -,107*** -,011 1,000

Sig. (2-tailed) ,044 ,000 ,099 ∆WCt /At-1 Pearson ,209*** -,267*** ,125*** ,098*** 1,000

Sig. (2-tailed) ,000 ,000 ,000 ,000

∆NCOt /At-1 Pearson ,233*** -,062*** ,265*** ,011 ,071*** 1,000

Sig. (2-tailed) ,000 ,000 ,000 ,104 ,000 ∆FINt /At-1 Pearson -,051*** -,244*** -,035*** ,017** ,138*** ,038*** 1,000

Sig. (2-tailed) ,000 ,000 ,000 ,015 ,000 ,000 Cash deficiency Pearson -,085*** ,261*** -,085*** -,037*** -,186*** -,093*** -,238*** 1,000

Sig. (2-tailed) ,000 ,000 ,000 ,000 ,000 ,000 ,000

It/ Kt-1 Pearson ,107*** -,001 ,094*** -,005 ,037*** ,031*** -,013 -,034*** 1,000

Sig. (2-tailed) ,000 ,839 ,000 ,454 ,000 ,000 ,075 ,000 Sales growth Pearson -,174*** ,069*** -,348*** -,011 -,061*** -,161*** -,039*** ,059*** ,381*** 1,000

Sig. (2-tailed) ,000 ,000 ,000 ,107 ,000 ,000 ,000 ,000 ,000 Debt to capital Pearson -,009 -,051*** -,027*** ,012 ,024*** -,019*** -,260*** ,178*** -,020*** ,044*** 1,000

Sig. (2-tailed) ,192 ,000 ,000 ,070 ,000 ,006 ,000 ,000 ,004 ,000 CFt/ Kt-1 Pearson ,027*** ,173*** -,073*** -,001 -,008 -,072*** -,062*** ,064*** ,228*** ,215*** -,016** 1,000

Sig. (2-tailed) ,000 ,000 ,000 ,886 ,251 ,000 ,000 ,000 ,000 ,000 ,017

***,**,* denotes statistical significant at the 1%, 5% and 10% two-tailed levels respectively

39

Panel B: Pearson correlation matrix for the variables of hypothesis 2

∆def revt / At-1

1/At-1 ∆St+1/At-1 ∆CFOt/ At-1

∆WCt / At-1

∆NCOt / At-1

∆FINt / At-1

Cash deficiency

It/Kt+1 Sales growth

Debt to capital

CFt/Kt+1

∆def revt / At-1 Pearson 1,000

Sig. (2-tailed) 1/At-1 Pearson ,226*** 1,000

Sig. (2-tailed) ,000

∆St+1/At-1 Pearson -,030 -,077*** 1,000

Sig. (2-tailed) ,052 ,000 ∆CFOt/At-1 Pearson -,008 -,008 -,001 1,000

Sig. (2-tailed) ,617 ,599 ,967 ∆WCt /At-1 Pearson -,030 -,008 ,044*** ,025 1,000

Sig. (2-tailed) ,052 ,611 ,005 ,117

∆NCOt /At-1 Pearson ,123*** ,193*** -,027 -,005 ,470*** 1,000

Sig. (2-tailed) ,000 ,000 ,091 ,746 ,000 ∆FINt /At-1 Pearson -,144*** -,419*** ,049*** -,022 -,122*** -,356*** 1,000

Sig. (2-tailed) ,000 ,000 ,002 ,175 ,000 ,000 Cash deficiency Pearson -,118*** -,124*** -,024 -,026 -,003 ,004 ,054*** 1,000

Sig. (2-tailed) ,000 ,000 ,117 ,099 ,851 ,803 ,001

It/ Kt-1 Pearson ,030 ,026 -,023 -,039** ,008 ,037** -,050*** ,007 1,000

Sig. (2-tailed) ,057 ,102 ,139 ,014 ,623 ,022 ,002 ,657 Sales growth Pearson ,021 ,127*** -,027 -,021 -,021 -,002 -,098*** -,151*** ,367*** 1,000

Sig. (2-tailed) ,186 ,000 ,088 ,183 ,173 ,886 ,000 ,000 ,000 Debt to capital Pearson ,003 ,002 -,015 ,035** ,041*** ,009 -,157*** -,018 ,025 ,058*** 1,000

Sig. (2-tailed) ,863 ,878 ,346 ,027 ,009 ,587 ,000 ,250 ,112 ,000 CFt/ Kt-1 Pearson ,050*** ,077*** -,011 -,010 -,016 ,001 -,092*** -,384*** ,213*** ,174*** ,016 1,000

Sig. (2-tailed) ,001 ,000 ,492 ,539 ,318 ,969 ,000 ,000 ,000 ,000 ,294

***,**,* denotes statistical significant at the 1%, 5% and 10% two-tailed levels respectively

40

Table 4

Model parameters for changes in gross accounts receivable

Independent Variables Expected Sign Cash deficiency Investmentst/ Sales growth Debt to total Cash Flowt/

definition Kapitalt-1 Capital ratio Kapitalt-1

Intercept ? 0.010 0.014* 0.010 0.010 0.008 (1.304) (1.739) (1.282) (1.292) (1.046)

1/At-1 ? 0.024*** 0.020*** 0.021*** 0.022*** 0.018*** (12.159) (10.117) (10.290) (11.367) (12.159)

∆St/At-1 + 0.111*** 0.112*** 0.110*** 0.110*** 0.105*** (73.233) (45.653) (61.508) (72.198) (65.846)

∆CFOt+1 /At-1 + 0.000 0.000 0.000 0.001* 0.000 (1.559) (1.065) (0.868) (1.829) (1.015)

Opp - -0.007*** -0.004* -0.002** 0.000 0.008***

(-4.918) (-1.779) (-2.105) (-0.524) (6.552)

Opp*(∆St/At-1) ? -0.021*** -0.008*** -0.014*** -0.013*** 0.009*** (-6.575) (-2.868) (-4.589) (-4.257) (3.107)

∆WCt /At-1 + 0.059*** 0.063*** 0.063*** 0.061*** 0.063***

(25.123) (26.853) (26.732) (26.137) (25.123)

41

∆NCOt /At-1 + 0.031*** 0.034*** 0.034*** 0.032*** 0.035*** (15.628) (16.725) (16.604) (16.009) (17.146)

∆FINt /At-1 ? -0.003*** -0.003*** -0.003*** -0.003*** -0.002***

(-7.769) (-6.836) (-7.076) (-7.081) (-6.302)

Industry dummy variables Included Included Included Included Included

Year dummy variables Included Included Included Included Included

Adjusted R-square 33.3% 33.3% 33.0% 33.2% 33.5%

***,**,* denotes statistical significant at the 1%, 5% and 10% two-tailed levels respectively Note to table 4: this table provides parameter estimates for the model to measure the normal changes in gross accounts receivable. Opp is the variable that indicates if the managers act opportunistic measured in terms of characteristics of cash deficient firms. This variable is defined in five different ways. Firms with a SIC company code between 4400 and 5000 (utilities), between 6000 and 6500 (financial institutions) and companies with a code higher than 9000 (public administration) are deleted from the sample. The sample is based on 22,184 US-GAAP firm-years. Between brackets the T-statistic for that variable is mentioned. The industry dummy and the year dummy variables are not displayed in this table. The top and bottom 1% of all the variables are deleted in order to take care that the outliers do not influence the results.

42

Table 5

Model parameters for changes in deferred revenue

Independent Variables Expected Sign Cash deficiency Investmentst/ Sales growth Debt to total Cash Flowt/

definition Kapitalt-1 Capital ratio Kapitalt-1

Intercept ? 0.005 0.002 0.006 0.006 0.006 (0.416) (0.175) (0.536) (0.545) (0.505)

1/At-1 ? 0.013*** 0.014*** 0.014*** 0.014*** 0.014*** (9.146) (10.165) (9.996) (10.346) (10.115)

∆St+1/At-1 + 0.001*** 0.000 0.000 0.000 0.000 (4.837) (0.260) (0.864) (-0.868) (-0.855)

∆CFOt /At-1 - 0.000 0.000 0.000 0.000 0.000 (-1.508) (-0.777) (-0.777) (-0.846) (-0.790)

Opp + 0.009*** 0.004** 0.000 0.002 0.002

(4.555) (2.002) (-0.206) (1.355) (1.099)

Opp*(∆St+1/At-1) ? 0.000 0.000 0.000 0.000 0.000 (-0.768) (-0.303) (-1.474) (1.089) (1.455)

∆WCt /At-1 + -0.006*** -0.006*** -0.006*** -0.006*** -0.006*

(-4.132) (-4.356) (-4.361) (-4.512) (-4.361)

43

∆NCOt /At-1 + 0.005*** 0.005*** 0.005* 0.005*** 0.005* (6.181) (6.003) (6.009) (6.277) (5.949)

∆FINt /At-1 ? 0.000 0.000** 0.000** 0.000** 0.000**

(-1.583) (-2.216) (-2.255) (-1.964) (-2.346)

Industry dummy variables Included Included Included Included Included

Year dummy variables Included Included Included Included Included

Adjusted R-square 12.3% 11.0% 10.9% 11.1% 11.0%

***,**,* denotes statistical significant at the 1%, 5% and 10% two-tailed levels respectively Note to table 5: this table provides parameter estimates for the model to measure the normal changes in deferred revenue. Opp is the variable that indicates if the managers act opportunistic measured in terms of characteristics of cash deficient firms. This variable is defined in five different ways. Firms with a SIC company code between 4400 and 5000 (utilities), between 6000 and 6500 (financial institutions) and companies with a code higher than 9000 (public administration) are deleted from the sample. The sample is based on 7,296 US-GAAP firm-years. Between brackets the T-statistic for that variable is mentioned. The industry dummy and the year dummy variables are not displayed in this table. The top and bottom 1% of all the variables are deleted in order to take care that the outliers do not influence the results.

44